Government Matters: Welfare Reform in Wisconsin 9780691222479

Citation preview

Government Matters

Government Matters WELFARE REFORM IN WISCONSIN Lawrence M. Mead

PRINCETON UNIVERSITY PRESS PRINCETON AND OXFORD

Copyright  2004 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, 3 Market Place, Woodstock, Oxfordshire OX20 1SY All Rights Reserved. Library of Congress Cataloging-in-Publication Data Mead, Lawrence M. Government matters : welfare reform in Wisconsin / Lawrence M. Mead. p. cm. Includes bibliographical references and index. ISBN 0-691-11646-6 1. Public welfare—Wisconsin. 2. Wisconsin—Politics and government—1951– I. Title. HV98.W6M4 2004 361.6′8′09775—dc21 2003051767 British Library Cataloging-in-Publication Data is available. This book has been composed in Sabon Printed on acid-free paper. ∞ www.pupress.princeton.edu Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

This book is dedicated to the welfare administrators of Wisconsin and the nation.

Contents

Preface List of Acronyms CHAPTER 1 Introduction

ix xiii 1

CHAPTER 2 The Development of Reform

18

CHAPTER 3 The Politics of Reform

37

CHAPTER 4 Implementing Work Requirements

63

CHAPTER 5 Local Variations

79

CHAPTER 6 The Emergence of W-2

107

CHAPTER 7 Implementing W-2

132

CHAPTER 8 Paternalism

157

CHAPTER 9 The Decline of Welfare

175

CHAPTER 10 The Effects of Reform

193

CHAPTER 11 Welfare Reform and Good Government

214

CHAPTER 12 Origins of Excellence

245

CHAPTER 13 Implications

261

Notes

279

Index

341

Preface

THIS BOOK RECOUNTS how Wisconsin led the nation toward welfare reform. In this state, more than in any other, welfare was transformed to condition aid to families on work by the adult recipients. I explain not only the dramatic changes Wisconsin made, but also how. It is an inspiring story of politicians who faced up to difficult challenges, and of administrators with the talent to implement waves of reform programs. Together, leaders and officials rebuilt the welfare state around work. Wisconsin did not avoid all errors, but it still struck the most telling blow against family poverty that government has managed in forty years of struggle. I also make the broader argument that Wisconsin’s success reflected its good government qualities. This is a state where, at least compared to the national norm, politics is high-minded and public administration is highly developed. Here we see government actually doing what American voters want it to do everywhere—tackle social problems with political resolution and determined programming. Case studies of welfare reform elsewhere show that Wisconsin’s story is not unique. Most of the other states that led reform also have good government traditions. Strong institutions were their leading resource in the struggle against poverty. The book was written using mainly data, documents, and interviews from the state itself. My first debt is to the many officials who gave of their time to generate or explain these sources to me, and to give their own accounts of reform. Among state officials, I particularly thank Jim Bates, Ginevra Ewers, Bob Korb, John McPeek, Joanne Rowe, Madelyn Scheer, Joe Stafford, Jason Turner, and Peter Van Ness. Special credit to Jan Van Vleck, who answered more questions than anyone else. I also spoke to local officials too numerous to name. I especially credit the leaders of several county welfare departments who helped to shape the state reform, notably Jon Angeli and his staff in Grant County, and Clark Earl, Larry Jankowski, and George Leutermann in Kenosha. I am also grateful to the Rockefeller Institute of Government and the Urban Institute for case studies of welfare reform at the state level. I use these to establish a link between good government and successful reform that stretches beyond Wisconsin. Dick Nathan and Tom Gais at RIG allowed me access to unpublished reports from their study states that were invaluable. My able research assistants (most of them graduate students at NYU or Princeton University) included David Dodenhoff, Kendal Elliott, Ian

x

•

Preface

Gold, Kevin Kosar, Matt Mercurio, and Adam Shrager. I am grateful to the Department of Politics at NYU for allowing me much time off from teaching to write this book. This time and other costs of the project were generously funded by the Lynde and Harry Bradley Foundation, the Annie E. Casey Foundation, the John M. Olin Foundation, the Smith Richardson Foundation, and the Wisconsin Policy Research Institute. I drafted much of the book during academic 2001–2002 when I was on sabbatical at Princeton as a visiting fellow in the James Madison Program in American Ideals and Institutions. Created by Robert George, McCormick Professor of Jurisprudence at Princeton, the Madison Program seeks to explore in depth the nature and implications of the American regime. The program involved me in wide-ranging discussions with the other fellows and many visiting speakers. An attentive staff dealt with our every need. This was an ideal setting in which to finish my project. I received valuable comments on all or parts of the manuscript from Tom Corbett, Jason DeParle, Tom Kaplan, Jean Rogers, Jason Turner, and Michael Wiseman. Many of my readers are themselves published experts on Wisconsin and welfare reform. A team of officials headed by Mary Rowin from the state welfare department also provided valuable comments and corrections. I benefited as well from expert anonymous reviews commissioned by Princeton University Press and two other publishers. All this input saved me from many errors and oversights. However, I did not adopt all the suggestions made, and responsibility for the final product remains my own. At Princeton University Press, my editor Charles Myers has been a pleasure to work with. He encouraged this project from an early point. He improved the book by astute comments and by requiring me to cut the length substantially. Kevin McInturff, Gail Schmitt, and other staff dealt capably with the production of quite a complicated volume. I acknowledge permission to reprint the following articles. Each title is followed by the journal in which it appeared, the volume and issue, and the publisher: “The Politics of Welfare Reform in Wisconsin,” Polity 32, no. 4 (Summer 2000), Northeastern Political Science Association; “Implementing Work Requirements in Wisconsin,” Journal of Public Policy 21, no. 3 (June 2001), reprinted with the permission of Cambridge University Press; “Welfare Reform in Wisconsin: The Local Role,” Administration and Society 33, no. 5 (November 2001),  Sage Publications, reprinted by permission of Sage Publications; “Optimizing JOBS: Evaluation versus Administration,” Public Administration Review 57 no. 2 (March/April 1997), American Society for Public Administration; “Welfare Case load Change: An Alternative Approach,” Policy Studies Journal 31, no. 2 (2003), Blackwell Publishing, Ltd.; and “Welfare Reform: The Institu-

Preface

•

xi

tional Dimension,” Focus 22, no. 1 (Special Issue 2002), Regents of the University of Wisconsin System on behalf of the Institute for Research on Poverty. I emerged from this study with deep respect for many people in Wisconsin, from Tommy Thompson, the dynamic governor who led most of the reform, right down to the recipients, who adjusted to radical changes with remarkably little complaint. But especially I credit the state’s welfare administrators. As one of them said to me, with an immodesty atypical of the state, “It’s because of us. We are welfare reform.” That is an overstatement; reform is a national movement that has arisen from localities across the country. But it is not untrue. Wisconsin administrators asserted their own responsibility for solving the welfare problem, often before political leaders did. What they did had epic consequences. They became quite literally world statesmen and stateswomen. Ambitious local welfare officials first transformed welfare at the county level. They then sold their programs to the state, which in turn helped to sell work-based reform to the nation, and beyond. And so the deeds of these local and state administrators finally echoed to the ends of the earth. Today, welfare officials from New Zealand and Europe travel to Grant, Kenosha, Madison, and Milwaukee to learn how they too might get a handle on their welfare problem. These Wisconsin officials exemplify, not only effective social policy, but their state’s intense faith in the public enterprise. I also know that they are not alone. Others like them labor to improve welfare in cities and states across the country. They may do so more obscurely, with less support and resources than their Wisconsin counterparts enjoyed. Nevertheless, they progress more rapidly than anyone would have guessed only a few years ago. Since the enactment of the radical national reform of 1996, welfare has changed in many states with a speed that astonishes longtime observers of American government. As in the Badger State, reform is a high-stakes gamble that by requiring, as well as helping, adult recipients to work in return for aid, they and their children may enter more fully into American life. In orchestrating that revolution, I believe, America’s welfare administrators hold the nation’s future in their hands. With admiration and gratitude, this book is dedicated to them. My final debt is to my family—my wife, Robin, and our daughter, Nora, who is the light of our lives. In addition to running the household, taking care of Nora most of the time, and pursuing her own career, Robin found time to read every word of this book, making valuable suggestions and catching many errors. She is my reality test. If it makes sense to her, and provokes no outrage, then I believe my audience will hear it. She also

xii

•

Preface

provided steady support during my long labors, first to do this research and then to get it accepted by journals and publishers. Nora, for her part, is too young to read, but her delight in “Dada” sure gives me a further reason to come home at night. Both of them remind me—and I do sometimes need reminding—that social policy is not an end in itself. Rather, it is a means toward what the Founders called “the pursuit of happiness.” Lawrence M. Mead New York University New York, NY August 2003

List of Acronyms

ACORN AFDC AFDC-UP ANF AW CARES CETA CFWA CRN CSAS CSJ CWEP DHFS DHHS DHSS DILHR DWD EITC ES ESAP ESI FEP FSA GOP GR IRP JOBS JS JTPA MEP MFIP MIS

Association of Community Organizations for Reform Now Aid to Families with Dependent Children Aid to Families with Dependent Children—Unemployment Parent Assessing the New Federalism study (Urban Institute) America Works Client Assistance for Reemployment and Economic Support computer system Comprehensive Employment and Training Act Congress for a Working America Computer Reporting Network Child Support Assurance System Community Service Job Community Work Experience Program Wisconsin Department of Health and Family Services U.S. Department of Health and Human Services Wisconsin Department of Health and Social Services Wisconsin Department of Industry, Labor, and Human Relations Wisconsin Department of Workforce Development Earned Income Tax Credit Employment Service (Job Service) Employment Skills Advancement Program Employment Solutions, Inc Financial and Employment Planner (W-2) Family Support Act of 1988 Grand Old Party (the Republican party) General Relief Institute for Research on Poverty (University of Wisconsin—Madison) Job Opportunities and Basic Skills Training Program Job Service (Employment Service) Job Training Partnership Act Management and Evaluation Project Minnesota Family Investment Program Management Information Systems

xiv

•

List of Acronyms

NEWWS OIC OJT PFP PFR PIC PRWORA RFS RS SCS SSF SSI SWIM TANF UMOS UW W-2 WEJT WEOP WIN WNW

National Evaluation of Welfare-to-Work Strategies Opportunities Industrialization Center On-the-Job-Training Pay for Performance Parental and Family Responsibility Demonstration Project Private Industry Council (JTPA) Personal Responsibility and Work Opportunity Reconciliation Act of 1996 Right of First Selection (W-2) Resource Specialist State Capacity Study (Rockefeller Institute of Government) Self-Sufficiency First Supplemental Security Income Saturation Work Initiative Model (San Diego) Temporary Assistance for Needy Families United Migrant Opportunities Services University of Wisconsin Wisconsin Works Work Experience and Job Training Program Wisconsin Employment Opportunities Program Work Incentive Program Work Not Welfare

CHAPTER 1

Introduction

IN THE MID–1980S, welfare was in chronic crisis in Wisconsin, as in much of America. “Welfare” here means the controversial aid program for needy families once called Aid to Families with Dependent Children (AFDC) and now named Temporary Assistance for Needy Families (TANF). The welfare rolls in Wisconsin reached 100,000 families and 300,000 people in 1986—in a population of only 5 million. The state, like the nation, had struggled with rising welfare dependency since the 1960s, and no solution seemed in sight. But in little more than a decade, the situation was transformed. The Badger State became home to the most radical welfare reform in the nation. Welfare in its traditional form was virtually abolished. The vast majority of recipients were driven off the rolls, mostly into jobs, while those who remain must work at some level to get any cash aid. At the same time, unprecedented new benefits were offered to the entire low-income working population. This revolution transformed the welfare state and the lives of many former recipients, mostly for good. I ask two questions in this book. First, what did Wisconsin do? I show in some detail how the Wisconsin reform developed. Starting in 1986, the state instituted a series of experimental programs in welfare. Most of them stiffened requirements that adult recipients work in return for aid. The changes culminated in the radical Wisconsin Works (W-2), a new aid system implemented in 1997. The effects of the reforms seem remarkably positive. Not only did dependency plummet, but work levels among lowincome adults soared while poverty fell. Many former recipients remain poor, but incomes are rising. After a generation of defeatism about uplifting the inner city, Wisconsin exemplifies a new and more promising antipoverty policy. The solution to poverty is no longer seen as doing more or less for the poor, the traditional debate among liberals and conservatives. Rather, it requires restructuring welfare so as to promote recovery from poverty. That means supporting families and helping needy adults work and stay in school. But it also requires clearer expectations that recipients function in minimal ways in return for support. In Wisconsin, we see for the first time, and on a large scale, what such a policy means in practice. Wisconsin is also a showcase for an innovative style of programming that I call paternalism. The state not only makes greater demands on its

2

•

Chapter 1

recipients. It supervises them closely to make sure that they fulfill those expectations, as well as claim benefits they qualify for. Paternalism combines benefits with demands, help with hassle. Alongside conduct requirements, this intense administrative style is a promising trend in social policy. It became a theme in Wisconsin in several areas of social policy besides welfare. My second and more important question is: How did Wisconsin accomplish this? The simple answer is: good government. The Wisconsin reform is the achievement of legislators and administrators who were unusually able and conscientious by national standards. Welfare reform is a test for government. Elected leaders must agree on what demands will be made of the poor in return for support, questions that have bitterly divided politicians in the past. And then officials must create complex arrangements so that families are aided while, at the same time, parents are inducted into work programs and given the child care and other services they need to work. To do this well was a triumph for government. Thus, ironically, a reform that most would call conservative finally enlarges our sense of what antipoverty policy might achieve. The Wisconsin story makes clear how dependent successful social reform is on strong institutions. I will show that, not only this state, but others with good-government traditions have led welfare reform across the country. Their success is inspiring, but also cautionary. Only some states perform at this level. Wisconsin finally directs our attention beyond welfare to the capacity of the American regime.

WHY WISCONSIN? Of course, successful welfare reform is hardly confined to Wisconsin. The welfare rolls fell nationwide starting in 1994. The fall was accelerated by the passage of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in 1996. Drafted by the conservative Republicans who took control of Congress in 1994, PRWORA was the most radical welfare reform ever enacted in Washington. Besides changing AFDC into TANF, it capped federal funding for state welfare programs, limited most families to five years on the rolls, and sharply raised the share of welfare parents who were expected to work. Coupled with superb economic conditions and expanded wage and child care subsidies, welfare reform drove the national caseload down by 60 percent through June 2002. Virtually all states shared in that progress. In Wisconsin, however, both the what and the how were still special. First, the state’s reform was unusually thoroughgoing. Most states have simply added stronger work requirements and incentives to a welfare sys-

Introduction

•

3

tem that, in other respects, changed little. Wisconsin totally redesigned family aid from the ground up. W-2 is, by common consent, the most innovative aid system in the nation. It demands immediate work with uncommon severity, but it also promotes work with its lavish support benefits, and it has other special features. Second, the consequences of reform in Wisconsin seem unusually favorable. Like no other urban state, Wisconsin virtually destroyed traditional welfare. Figure 1.1 contrasts trends in the Wisconsin and national caseloads during 1980–2001. The decline in Wisconsin began unusually early—in 1986—and the state avoided almost entirely the sharp run-up in recipients that hit most of the country between 1989 and 1994. In the average state, that jump was a startling 34 percent, an increase that helped put welfare reform on the national agenda, leading to PRWORA. In Wisconsin, the caseload fell almost continually from 1986, and by 2001, the decrease was around 80 percent from 1986, with fewer than 10,000 families remaining on W-2 cash benefits.1 One might imagine that Wisconsin merely reversed a level of dependency that was high for the nation. Figure 1.2 plots trends in the percent of population on AFDC/TANF in Wisconsin and nationally. It is true that Wisconsin rose well above the national norm in the 1980s—but by 2001 it was well below. Table 1.1 shows how much each state and the District of Columbia reduced the number of recipients on their rolls between 1994, when the national caseload peaked, and 2001. At 82 percent, Wisconsin comes behind only Wyoming and Idaho and well ahead of other urban states paying high benefits. Even in Milwaukee, a heavily nonwhite and depressed city, traditional family welfare almost disappeared. The table also understates Wisconsin’s achievement. By 1994, the state had already driven its more employable recipients off the rolls—yet after that it still managed almost the greatest caseload fall in the country. Another, perhaps better measure of change is how fully a state required work for the recipients remaining on welfare. Table 1.2 ranks the states by the percentage of their recipients who satisfied TANF’s work participation standards in fiscal 2000. Those rules required that adults be active in work activities (defined to mean largely actual work, but permitting some training) for at least thirty hours a week. Again, Wisconsin ranks almost first in the country, ahead of all other urban states. This despite the fact that, due to its extreme caseload deflation, its remaining recipients were probably among the less employable in the country. The Wisconsin reform also excels by a still more important measure— the effects on former recipients and the society. Nationwide, research on adults who have left welfare finds the effects to be largely positive. Threefifths or more of the leavers are working, and although some are not, there is little evidence of hardship. Work levels among the poor are

4

•

Chapter 1

Figure 1.1. Trends in Wisconsin and national AFDC/TANF caseloads. (Source: U.S. Administration for Children and Families)

Figure 1.2. Percent of population on AFDC/TANF, Wisconsin and the nation. (Sources: U.S. Administration for Children and Families, U.S. Bureau of the Census)

Introduction

•

5

TABLE 1.1 States Ranked by Percent Reduction in AFDC/TANF Caseloads (in persons), 1994–2001 State Wyoming Idaho Wisconsin Florida Mississippi Colorado Illinois Louisiana Oklahoma North Carolina South Carolina Michigan Ohio Georgia Maryland Oregon Massachusetts Alabama South Dakota Virginia New Jersey West Virginia Pennsylvania Connecticut Kansas Utah

Percent −94.0 −90.3 −82.3 −81.4 −77.5 −77.2 −74.4 −73.6 −73.5 −71.9 −71.2 −71.0 −70.9 −69.4 −69.3 −67.9 −67.3 −67.0 −66.7 −66.6 −66.2 −65.8 −65.1 −63.9 −62.0 −61.5

State National average Kentucky Arkansas Montana Maine Arizona Texas New Hampshire Alaska California Delaware Missouri North Dakota Washington New York Iowa Nevada Tennessee Vermont Nebraska Indiana New Mexico District of Columbia Minnesota Rhode Island Hawaii

Percent −61.5 −60.7 −60.0 −59.8 −59.4 −58.9 −55.6 −55.5 −55.3 −55.1 −55.1 −53.9 −52.7 −51.5 −51.1 −51.0 −49.0 −48.3 −47.5 −47.4 −46.5 −45.1 −41.8 −39.7 −33.8 −33.1

Source: Data from the U.S. Administration for Children and Families.

sharply higher, and poverty rates are falling, albeit more slowly. The effects on families and children also appear largely favorable.2 The recipients Wisconsin expects to work might be more disadvantaged than elsewhere, yet the proportions of leavers who are working and getting by are similar to those in other states. And work levels among poor parents on and off welfare are unusually high. Like no other state, Wisconsin has moved poor families not only off welfare but into jobs. These achievements brought Wisconsin extraordinary publicity. By the early 1990s, the innovations in the Badger State, and its ability to reduce the rolls when they were rising elsewhere, drew national attention. Jason DeParle of the New York Times chronicled the Wisconsin reform for a national audience, while other journalists wrote it up for their own states.3

6

•

Chapter 1

TABLE 1.2 States Ranked by Percent of Recipients Satisfying TANF Work Participation Standards in Fiscal 2000 State Kansas Wisconsin Indiana Massachusetts Montana Oregon Illinois Wyoming South Carolina New Hampshire Ohio Washington Idaho South Dakota Virginia Connecticut Alaska Iowa Maine Arizona National average New Jersey Alabama Nevada New Mexico Colorado

Percent 77.4 73.4 72.3 69.2 68.2 64.0 59.2 59.0 54.0 53.1 52.9 52.8 47.7 46.5 44.9 43.0 42.1 41.8 40.0 39.7 38.8 37.8 37.7 37.4 36.9 36.6

State Michigan North Dakota Tennessee Minnesota Missouri Oklahoma Louisiana New York Florida Utah Hawaii Delaware California Kentucky Texas Rhode Island District of Columbia Nebraska Arkansas North Carolina Mississippi West Virginia Georgia Pennsylvania Maryland Vermont

Percent 36.4 35.7 35.4 34.7 34.0 33.9 33.5 33.2 33.0 31.1 29.7 27.6 27.5 25.6 25.6 25.0 24.4 22.6 20.8 19.2 17.8 17.1 12.2 11.2 6.3 NA

Source: Data from the U.S. Administration for Children and Families.

Officials from around the country and from foreign countries flocked to the state to see for themselves, as did an army of researchers.4 And from publicity sprang influence. As the principal leader of the reform, Wisconsin Governor Tommy Thompson became a national power broker on welfare. Without Wisconsin’s example, PRWORA would have been less radical. THE MANNER OF THE REFORM The how of the Wisconsin reform, however, is as important as the what. The American public has expressed rising cynicism about government since the 1960s. A leading reason was the inability of government to solve

Introduction

•

7

chronic national problems, among them entrenched poverty and dependency. Impatience with the liberal social programs of the 1960s and 1970s helped drive the nation’s politics to the right, but Republicans initially had little more success dealing with poverty than did Democrats. And of all social programs, welfare was the most unpopular. Into this disconsolate scene, the Wisconsin success blew like a fresh wind. Here we see a popular government, in response to the public will, truly facing up to welfare and mastering it, at least in part. The governor and earnest legislators focused on solutions, downplaying partisanship. They quickly settled on a work-based approach, followed it for some years, then boldly enacted a completely new system. They delegated the details to talented administrators who implemented the new programs masterfully, if not without error. They even saved money in the process. To see how unusual Wisconsin is, we have only to look elsewhere. All states claim to have reformed welfare, but in many, there is less real change than meets the eye. While the caseload fall is substantial everywhere, few of the more urban, Northern states rank high in table 1.1. That reflects in part their deep divisions about how to change welfare. On the other hand, few southern states rank high in table 1.2. They are willing and able to restrict aid, but as yet lack the bureaucratic capacity to build work into welfare. Only a handful of states have been able to enforce work in welfare seriously without stumbling. Of these, Wisconsin is the most remarkable. It has actually done what other states talk about. It has recently been fashionable in political science to assume that behavior in politics is driven by self-interest. But often in politics, people act out of moral convictions or loyalty to the community, even though doing so does not serve their own interest in any material sense.5 Such behavior was conspicuous during welfare reform in Wisconsin: • Governor Tommy Thompson and Democratic opponents in the legislature abandoned long-held positions in order to craft reforms that are simultaneously big-government and demanding toward the recipients. Only a minority of either party dissented. • The move to abolish traditional welfare was led by a black Democrat representing inner-city Milwaukee. His own constituents pressed him to do this. Reform never became a race issue. • Community groups, although reflexively liberal in the past, accepted the end of entitlement and confined their criticism to details. • Business groups supported reform by launching experimental reform programs of their own. • Some counties forged their own welfare reform programs without waiting for the state government in Madison, and some of these became models for the state.

8

•

Chapter 1

• Counties competed to host various Thompson experiments in welfare reform, despite the political dangers and the trouble of changing their administrative routines. • Welfare staffs throughout the state implemented statewide reforms diligently, even though the reduction in the caseload would ultimately cost some of them their jobs. • The recipients endured a revolution in their lives. The vast majority left welfare and shouldered more responsibility for themselves with considerable resourcefulness and little complaint.

To explain conduct like this, some appeal to virtue is simply unavoidable. This fact alone compels our admiration. I do not deny that self-interest also operated. Given the forceful public will to change welfare, legislators of both parties, and administrators, had every incentive to support reform. Nor were altruistic motives lacking in less gifted states. I say only that the reform process in Wisconsin was highminded relative to the norm in American politics. That quality helped policymakers avoid divisions and other problems that prevented change elsewhere.

AN INSTITUTIONAL APPROACH Most scholars understand social programs in economic terms. Welfare is seen as a source of income, training programs as a source of human capital. Instead, this book takes an institutional approach. I view programs more as authority structures that do or do not govern the society. Mostly, welfare reform is not about changing the economic scale of welfare or enriching the skills of recipients. Much more, it means attaching requirements for good behavior to existing benefits. Authority is even more important if programs have the paternalistic character that they developed in Wisconsin, where case managers closely oversee clients. I investigate not only program structures but the government institutions standing behind them. Other scholars ask whether social programs are sufficient to overcome poverty. I ask whether government is even able to enact and administer such programs. To do so, politicians must resolve sensitive issues about personal responsibility that earlier they delegated to social workers. Administrators must implement complicated programs. They also must do so in an era when public service is less admired than it was before 1960. In the last generation, government has lost prestige compared to the private sector. There is a widespread desire that it be “reinvented” on a business model.6 That demand shaped welfare reform even in big-government Wisconsin.

Introduction

•

9

This focus on institutions is unusual in poverty studies, but fruitful. Most experts who study poverty attribute it to social factors such as the labor market or the family structure of the poor. They call for doing more to help the poor, but their research says little about government itself.7 A focus on social barriers is plausible for groups that were working already or are not expected to. Their constraints often are economic. Higher wages and benefits have indeed reduced poverty among working families and the elderly and disabled since the 1960s. Most of the working-aged poor, however, lack all steady employment. That fact is not explained well by social or economic conditions; a culture of poverty and permissive public policies appears to matter more.8 To overcome this sort of poverty requires helping people but also changing lifestyle. That is a task for politics and administration, and these are the forces driving change. In Wisconsin, reform built up the authority structure surrounding welfare far more than it altered benefits. Work enforcement was principally responsible for driving the rolls down, although the economy and new benefits also helped. As reform succeeded, discussion shifted away from background conditions and toward institutions. There was less and less talk about a lack of opportunity and more and more about reshaping programs to promote work. Mostly, government held the solution to welfare in its own hands. While other states have not transformed welfare in the same manner, most of the changes they have made are also governmental. Work requirements are being implemented with a speed that belies the idea that public agencies cannot change. States and localities are also reorganizing social service programs, combining welfare with employment and other agencies in novel ways.9 To date, welfare reform has principally been an exercise in statecraft. This might suggest that political scientists like myself have a better chance to understand reform than do economists. But in political science, past research is of scarcely more help. We have convincing studies of the political movements that sometimes give rise to new social programs.10 We have useful models of how demands for programs come onto the political agenda, so that government acts on them.11 We have a literature about policy innovation at the state level,12 and about the implementation of policy.13 But there is little about administrative statecraft—the translation of politics into policy and routines that work on the ground. I know of no research that follows social policy innovation from initial proposals and politics all the way through implementation and effects, as I do here. Nor do I know of earlier research that connects welfare policy, not only to implementation, but to the general capacity of government. Further, political science research on public policy takes little interest in the merits of policy per se. There is a salutary focus on institutional

10

•

Chapter 1

process, but only that. The question is why a certain policy was enacted and administered as it was, but not whether it was good or bad.14 Just as poverty researchers tend to tell government what to do while ignoring it, most political scientists study government without telling it what to do. Government is first of all the servant of the society. Political participation can be an edifying experience, prized by political theorists. But for most people, government’s purpose is to solve public problems that they cannot handle themselves. So political science must take policy analysis seriously. Ask first what government should do to solve some important problem, such as poverty. Then discuss the institutional constraints that impede realizing that policy ideal. In political science, to link policy argument to political analysis like this is rare. But this is the sort of reasoning that policymakers themselves do. Only if students of government sit in the same seat can political science recover its heritage as the master science.15 In my earlier books, much in this spirit, I advocated work requirements as a strategy for welfare reform. I also pointed to the political and administrative obstacles, and I doubted whether government could in fact enforce work.16 In Wisconsin, however, a government actually has enforced work in welfare. Our problem is to explain success rather than failure. But as I note later in the book, limits on government capability still confine what many states can achieve.

SOURCES Most of the research now going on about welfare reform focuses narrowly on social conditions. The dominant question is whether the poor have been helped or hurt by the revolution flowing from PRWORA. Some of this inquiry focuses on Wisconsin; I draw on it heavily in chapter 10 and elsewhere. For the how of reform, I have much less to go on. Most of the literature on welfare work programs consists of evaluations of their effects, mostly by the Manpower Demonstration Research Corporation. These studies were essential to establishing work requirements as a feasible approach to reform. But they say little about the political and administrative processes that create effective programs, or fail to do so. There are also studies of welfare politics, but they focus mainly on public opinion or on reform at the national level. Many of these studies cover only one stage in the reform story, or are out of date.17 There is some prior research on the politics and administration of welfare reform at the state and local level.18 These projects generated most of the case studies that I use to compare Wisconsin to other states. But these inquiries are much more limited than what I offer here for Wisconsin.

Introduction

•

11

Within Wisconsin itself, there is considerable research on the effects of reform. But past writing about the politics and administration of change is surprisingly limited, given the state’s prominence. It is confined to one important journal article, several book chapters, and a Ph.D. dissertation.19 I have drawn gratefully on all, but my account is based mostly on original sources. I relied heavily on interviews with state policymakers, government documents, and data from state programs. Here in brief is how I assembled the study; later chapters go into more detail. For the politics of reform, I gleaned an initial impression from the secondary sources, then talked to senior state administrators in 1994–95; in 1997–98 I interviewed a score of leading policymakers in the bureaucracy, the state legislature, and community groups. I also had research assistants review journalism about welfare reform in state newspapers for the period 1985–99. The interview respondents were people whom the other sources indicated were key players in welfare decisions from 1985 through the implementation of W-2.20 I asked them what had happened during the reform struggles and why the outcome was so consensual. For the administrative side of the story, I relied chiefly on government documents. The most important program was the Job Opportunities and Basic Skills Training Program (JOBS). A national program, JOBS was the primary welfare work structure in Wisconsin from the late 1980s up through the implementation of the more radical reforms in the mid– 1990s. I interviewed about this and other programs in Madison and in ten counties around the state. These visits were mostly in 1994–95, followed by further visits to Milwaukee in 1995 and 1999. I asked officials running welfare and JOBS how they processed clients, what welfare work policies they associated with good JOBS performance, and why caseloads had fallen. I then tested this input using program data for JOBS. The results indicated that the counties generally did treat their clients in the ways they claimed to do. I also built statistical models that related the policies and other features of the counties to their performance in JOBS, as measured by the program’s own indicators. The results showed that counties that stressed high participation and actual work over training placed more people in jobs, but these policies had less effect on the quality of jobs. Later I did similar analyses relating county JOBS policies and other factors to caseload changes. Again, the more demanding counties had the most caseload decline. I also compared Wisconsin to other states. Several analyses show that the connection of good government to successful welfare reform is seen nationwide, not just in the Badger State. I used the case studies of TANF development mentioned above to show that good-government states like Wisconsin were able to enact and implement new welfare policies more

12

•

Chapter 1

smoothly than states of other traditions. I also built models of caseload change across the states, parallel to the county-based analysis just mentioned. These traced variations in caseload change in the periods 1989– 94 and 1994–98 to the welfare policies and other features of the states. In the first period, caseloads rose in most states, while in the second they fell. I found that states like Wisconsin that had demanding JOBS and child support policies typically had less caseload increase, or more decline. I also linked these policies, in turn, to the governmental features of the states. The states with good-government traditions like Wisconsin were the most able to fuse generous benefits with strong work requirements. That is the combination that seems to work best and that the public supports. In the cross-state comparisons and elsewhere in the book, I use a scheme of state political cultures developed by Daniel Elazar to characterize and measure what good government means. Wisconsin is one of a number of “moralistic” states where policymaking tends to be focused on problem-solving rather than partisanship, and public administration is strong. With some exceptions, these moralistic states emerge as the most effective welfare reformers. Finally, my research assistants and I studied Wisconsin’s past to discover the deeper springs of its competence. These include the state’s early settlement, its leadership of the Progressive movement, and its tradition of leadership in national social policy. All those antecedents prepared the state to show leadership again in the current era. We also researched the history of federal attempts to promote capable welfare administration. Throughout the research, I benefited from close ties to Wisconsin. I first encountered the state in 1987, through a visiting appointment at the University of Wisconsin at Madison. I immediately became involved in the early stages of Tommy Thompson’s welfare reform efforts. Along with other experts from the university, I advised Kenosha County on its own local welfare reform, which turned out to be a model for the state. I continued to advise the state welfare department on research and policy issues.21 These contacts explain the wealth of program data, public documents, interviews, and other inside information that I use to tell the Wisconsin story. Few scholars of social policy have had such entre´e. I recognize that such close involvement poses dangers to objectivity. While I do think well of the Wisconsin reform and the regime behind it, I have tested my presumptions. The evidence for Wisconsin’s success is strong, and success is clearly tied to governmental quality. I also make clear several reservations about the state’s performance: Its early reform programs were casually planned, W-2’s initial design was too severe, and its implementation was troubled. Welfare reform was a testing challenge, even for Wisconsin.

Introduction

•

13

THE MEANING OF SUCCESS To show that Wisconsin succeeded because of good government, I have to show that reform was in fact a success. Whether it was depends in part on one’s goals. I adopt as the chief aim of reform raising work levels among the adult recipients. Most poor adults, on and off welfare, work irregularly if at all. To cause them to work more steadily in available jobs is the best thing one can do to relieve family poverty, although improving low-skilled wages and benefits is also needed.22 Raising work levels is the chief goal that Wisconsin and other states have set. The public strongly endorses the work goal, as do most recipients.23 Heeding the voters, so do most politicians. Still, there are influential dissenters. On the right, they include antigovernment conservatives who oppose the whole idea of welfare. On the left, they include some liberal and Democrat politicians along with many community organizations and most academic experts on poverty. They would have preferred reform to focus on reducing poverty and inequality. They also wanted to promote work among the recipients, but they wanted to do it through new benefits, such as training or stronger work incentives, not by enforcing work as a condition of aid. In short, they defend entitlement, or the traditional idea that welfare should be given to families on the basis of need alone, with no questions asked about lifestyle. That is exactly the principle that reform in Wisconsin has rejected, as in most states. It is also controversial to say that reform has satisfied the work goal, in Wisconsin or elsewhere. Some observers think the rise in work levels is chiefly due to good economic conditions. Or they think that a recession more serious than the recent one may land millions back on aid. I return to this dissent in my conclusion.

CONNECTING GOVERNMENT TO SUCCESS Assuming that work is the goal, I also have to show that governmental quality contributed to successful reform, in both Wisconsin and the nation.24 This requires that I establish that Wisconsin already had a good government prior to reform, that this regime helped it enact policies thought to be effective, and that these policies in fact had good results. Table 1.3 summarizes these propositions. “Wisconsin’s background” means the formative experiences that shaped the Wisconsin regime. “Governmental excellence” connotes a legislative process focused on problem-solving, rather than partisan rivalry, and strong public administration. These are the qualities that most people

14

•

Chapter 1

TABLE 1.3 Connecting Governmental Quality to Successful Welfare Reform Propositions Wisconsin’s background → Governmental excellence Governmental excellence → Effective welfare policies Effective welfare policies → Successful welfare reform.

Addressed in chapters 12–13 2–8, 11 5, 8–10

mean by good government, and which are stressed in Elazar’s idea of the moralistic style. “Effective welfare policies” means policies that the best information at the time indicated could promote successful reform. Finally, “successful welfare reform” means programs that in fact had good consequences. Table 1.3 asserts that Wisconsin derived a good government from its past, used it to craft the best policies known at the time, and obtained good effects from them. Importantly, all these terms are defined independently, not in terms of each other. “Wisconsin’s background” refers to historical events that occurred long before welfare reform. “Governmental excellence” refers to the high-minded and effective manner in which Wisconsin government operated during reform, as judged by myself and other close observers. I mean that Wisconsin was “high-minded” and “effective” relative to other states, a comparison that I make both judgmentally and in more rigorous ways. “Effective welfare policy” concerns the substance of policy; it means that the measures enacted were those that research at the time indicated would succeed. “Successful welfare reform” connotes that the programs actually did succeed, as determined by later research and evaluation.25 How do I show these connections rigorously? If I merely inferred them post hoc in light of the Wisconsin story, the argument that good government promotes successful welfare reform would not be very persuasive. How could we be sure that this connection was even typical for Wisconsin, let alone other states? One cannot consult only an exemplar like Wisconsin in a period of triumph, where good government and policymaking success coincide, and infer that the one caused the other. We have to show that the association is sustained, and that it occurred in other states. We have to show, not only that governmental virtue produces success, but that governmental weakness does not. In short, we have to broaden the empirical basis of our inference. One way I do that is by deriving my hypothesis in part from welfare reform prior to Wisconsin. I did not hit on good government as Wisconsin’s secret before my research began. But as soon as that hypothesis emerged,26 I realized that it made sense of much of the prior research on welfare reform, at either the national or state level. From the beginning

Introduction

•

15

of the welfare issue in the 1960s, the chief impediment to reform had been deep political division about how to handle the problem. Should welfare be expanded or restricted? Should work be demanded or not? In Washington, those rifts frustrated all fundamental reform until PRWORA in 1996, despite the public’s forceful will to change. Much the same was true in many states, especially those with large urban caseloads.27 And the chief problem with welfare reform programs was their administrative demands. One can put welfare adults to work only if one pays them aid while also inducting them into work programs and providing child care and other support services. Implementation research made clear that localities that solved these problems placed more recipients in jobs, and reduced welfare, compared to those that did not.28 My own prior studies showed that the performance of these programs chiefly reflected the proportion of clients that participated in them and looked for work. That in turn depended on a program’s ability to assert a work demand and solve the logistical problems of actually moving clients into jobs.29 Governments that could do these things were good governments. Therefore, good government was the real solution to the reform problem. A second way of broadening the inference is to show that the link between good government and success is sustained in Wisconsin over time. Where other accounts of the state cover mainly W-2, I cover reform well back into the 1980s. This allows me to show that the same governmental virtues appeared during the entire reform process. Throughout the struggle, Wisconsin’s policymakers focused principally on problem-solving, rather than partisan maneuver, and administrators were creative. The third, and most crucial, way of broadening the argument is to escape from the limitations of a case study. If I considered only one state, even over time, it would be difficult to prove a general connection between governmental excellence and reform success. There would be no way to obtain enough variation—the good as well as the bad—on the variables in Table 1.3. It is impossible as a practical matter for me to research other states in the same depth I do Wisconsin. I could not get permission to do it, let alone gather and analyze all the same information. Simply to know what all states are doing as they reform welfare is impossible. As Dick Nathan, a veteran intergovernmental scholar, remarked, “We can’t count all the sand in the sea.”30 However, I can exploit variation within the state. This I do by comparing the seventy-two Wisconsin counties, ten of which I visited and studied in some depth. The cross-county statistical analyses mentioned above establish that localities that instituted work requirements well achieved higher performance in JOBS than those that did not. They also had more caseload decline. These counties had stronger leadership than the others—politicians who supported change, and enterprising administrators

16

•

Chapter 1

able to innovate. The more backward counties in terms of policy—especially Milwaukee—also had the worst institutional problems. The counties thus dramatized in microcosm the association of successful reform with strong government. The other way to escape a case study is to compare Wisconsin to other states, to the extent feasible. Throughout the book, I compare Wisconsin informally to the many other states where I have had research or consulting experience, especially New York.31 Further, I make more rigorous comparisons by analyzing case studies of TANF implementation covering about half the states. I also use my national caseload change models, mentioned earlier, to associate successful welfare reform policies with governmental qualities across all the states. All these analyses show that good-government states in Elazar’s sense generally have excelled at welfare reform.

A LOOK AHEAD Table 1.3 lists the chapters in which I especially address each of the book’s three key propositions. Most of the book focuses on the second proposition—the tendency of a good-government style to promote effective welfare policies. I lead the reader through what Wisconsin did, including the politics and the implementation of reform. Finally, I connect reform to good effects. I do all this once for the early stages of reform, then again for W-2. Chapter 2 summarizes the Wisconsin reforms up until W-2, and chapter 3 covers the surprisingly bipartisan politics behind these changes. Chapter 4 covers the early implementation of work requirements in the state, chiefly through the JOBS program. This was the administrative foundation of the more demanding work policies to come. In chapter 5, I closely examine several counties to show the role of leading and lagging localities in shaping the state’s reform. This chapter also uses the first cross-county analysis to show that counties that were demanding about employment achieved the highest performance in JOBS, even controlling for other factors. Chapter 6 returns to the legislative arena to describe the drafting of W-2 and its enactment. Again, we see a political process marked more by agreement about fundamentals than partisan combat. Chapter 7 describes the dramatic implementation of W-2, which included an adventurous “reinvention of government.” Chapter 8 discusses the paternalistic character of many of the state’s reform programs. Turning again to effects, chapter 9, on the dramatic decline in the welfare caseload, argues that this decline was more due to welfare reform than other factors, including the good

Introduction

•

17

economy. Here appear several of the statistical analyses, comparing counties within Wisconsin and states nationwide. Chapter 10 summarizes research on the effects of the later stages of reform, including surveys and other analyses showing that the economic and social effects were largely favorable. Chapter 8 contains some evaluation information as well. As described earlier, Chapter 11 broadens the argument connecting successful reform to good government from only Wisconsin to all the states. Here is where I show, using both case studies and statistical analysis, that the moralistic states as a group, not just Wisconsin, have been the leaders of reform nationwide. I postpone the historical origins of Wisconsin’s competence—the first proposition—until chapter 12. I do this partly because the state’s origins become interesting and important only if one first establishes that its performance was indeed unusual. Much that I say here applies to the other good-government states as well, which enter the picture in chapter 11. Finally, chapter 13 addresses the national implications of Wisconsin. If good government is essential to successful welfare reform, how may it be promoted? I conclude that the nation’s civility, not its wealth, is its leading resource against poverty.

CHAPTER 2

The Development of Reform

THIS CHAPTER DESCRIBES the stages of Wisconsin’s welfare reform up through Wisconsin Works. Reform might have had various goals. Some of these are liberal, tending to the expansion of welfare. Indeed, the development of Wisconsin’s liberal welfare regime in the 1970s and early 1980s reflected that impulse. But dependency grew. The pioneering reforms in child support enforcement that the state undertook in the early 1980s failed to stem it. Then, in the middle 1980s, came a reckoning. A consensus for workbased change gelled quickly. Benefits were trimmed, new welfare work programs enacted, and a series of experiments launched based on waivers of normal federal welfare rules. The innovations became more and more radical until, by 1996, welfare recipients faced serious pressures to work, or to avoid welfare entirely. The sequence of initiatives can appear haphazard, but it conceals an underlying rationality. For lack of guidance from earlier policy or research, Wisconsin simply tried out different ideas, until mandatory work programs became dominant. Even the lesser programs helped to educate the government and the society about the need for change, which then produced more change. That process accelerated until welfare was transformed.

THE GOALS OF REFORM Welfare reform is so dominated today by putting recipients to work that we forget that it might have other goals. Since its inception in 1935, AFDC has drawn criticisms on several grounds. The liberal perspective was that AFDC was insufficiently generous while also generating incentives against escape from poverty. States had the power to determine benefit levels, and virtually all set payments well below the federal poverty line. Levels were particularly low in the South. AFDC also denied coverage to most twoparent families, however needy they were. The fact that eligibility was ordinarily limited to single-parent families also discouraged marriage, a frequent route out of poverty, for if a welfare mother married she usually lost her benefits.

Development of Reform

•

19

Welfare discouraged work as well, because benefits were typically reduced in line with any earnings a mother had. If she decided to leave welfare for work anyway, the system did little to “make work pay”; crucial supports such as child and health care that were available on the rolls might be unavailable off. To fix these problems, liberals said, required raising benefits, broadening coverage to the two-parent poor, adding work incentives (allowing recipients to keep some benefits if they worked), and improving services and wage subsidies on or off the rolls. In this way, one might minimize dependency while still assuring aid. For many liberals, such steps entailed strengthening federal control, even nationalizing the program, as well as spending more money.1 The conservative perspective was that welfare was too costly and centralized as it was. Welfare made the poor prefer dependency to employment. It caused single mothers to marry the state rather than the fathers of their children. The system also promoted fraud and dishonesty, as applicants for aid often claimed they were needy and without husbands when they were not. Some conservatives wanted to “clean up” the system by confining aid to the “truly needy” and policing eligibility closely. Others wanted to abolish AFDC entirely. Rather than federalize the system, they said control should be devolved to localities and perhaps to churches and other private bodies. Finally, along with some liberals, conservatives wanted to toughen child support enforcement. Not government but runaway fathers should be held responsible for poor children.2 The largest issue dividing liberals from conservatives was whether aid should be given to applicants on the basis of entitlement, without behavioral conditions. That is, should administrators ask only whether families were needy? Or should they insist as well that applicants help themselves overcome dependency? Should the focus of aid be on overcoming poverty, or on reversing behaviors, such as unwed pregnancy and nonwork, that often kept families poor? While all regretted the conduct issues, liberals typically gave first priority to giving aid, conservatives to changing lifestyles. At the national level, liberal goals dominated the early era of welfare reform in the 1960s and 1970s, while conservative goals came to the fore after that. One reason was the general turning of the nation’s politics to the right—a process that was itself abetted by rising welfare and other social problems linked to poverty. But another cause was that liberal reformism turned out to be impractical. How could one simultaneously aid the needy and reduce dependency? Liberals counted heavily on work incentives (allowing working recipients to keep part of their benefits) to raise work levels without limiting access to support. But then, recipients who took jobs could emerge better off than comparable families that

20

•

Chapter 2

never went on aid, raising fairness concerns. More seriously, work incentives that were added to AFDC in 1967 failed to show any clear effects on work levels.3 Partly for this reason, they were cut back in 1981. And politically, incentives failed to satisfy the public’s demand that work in welfare be seen as an obligation, not as a choice. More recently, advocates of “making work pay” argue that one can promote work by subsidizing wages or guaranteeing child or health care outside welfare. These supports will enable a single mother to make a life without cash aid, provided she works part-time. The benefits are popular because they are work connected. But nothing ensures that recipients now receiving cash will in fact choose to work instead. Some adherents also would time-limit aid to set a term to dependency. They are loathe, however, to have government explicitly tell people to work as a condition of aid.4 The conservative idea of enforcing work in return for support came to dominate reform because it was more popular and more practicable. Few welfare mothers worked, but if more did so, this would raise their incomes while also reducing dependency. Administrative work tests—where work effort is demanded as an eligibility condition for aid—avoided the fairness problem of incentives. They turned out to be more effective as well. Starting in the 1980s, a series of evaluations showed that mandatory work programs tied to AFDC had more power to raise the employment and earnings of recipients than any other strategy. Politically, work tests satisfied the voters’ demands for work without compromising their equal desire that government continue to aid families in need. Work enforcement emerged as a middle ground between the old policy of entitlement and the more extreme conservative proposal of simply eliminating welfare.5 Work requirements also offered more potential to change welfare than other conservative strategies. While fraud and abuse were indeed rampant when welfare expanded in the 1960s and 1970s, states moved speedily under federal pressure to clean up, and the rolls remained largely unchanged. Similarly, government sharply improved enforcement of child support, but this too had little effect on dependency. It proved too difficult to establish the paternity of absent fathers and to make them pay support regularly, in part because of their own employment problems. Putting welfare mothers to work, however, in many cases could take a family off cash welfare. The mother might still need child care or food stamps, but dependency could be reduced. However, deciding and enforcing the required work standards makes serious demands on government. The potential for political conflict or administrative breakdown is great. The dilemmas of traditional welfare are traded for institutional challenges that may prove just as difficult.

Development of Reform

•

21

THE LIBERAL PHASE OF REFORM Wisconsin dramatizes the shift from liberal reform to more conservative approaches. Traditionally, the state was known for generosity. Already by 1960, Wisconsin ranked sixth among the states in the aid it paid per recipient in AFDC. The state accepted an onus to “reallocate goods and services from the prosperous to the less fortunate members of the Wisconsin community.” The idea was accepted that “the needy of all ages—the handicapped, mentally ill, emotionally disturbed, single parents with young children, juvenile offenders, the unemployed, and a host of others—were to be taken care of, helped, and rehabilitated.”6 Yet less than 1 percent of Wisconsin citizens depended on AFDC in 1960, compared to 1.6 percent nationally.7 Following the doubling of the AFDC rolls in the late 1960s and early 1970s, most states used their control of benefits to curb dependency. They either cut benefits or failed to raise them enough to offset inflation, thus allowing them to decline in real terms. This limited the number of families that had incomes low enough to qualify for aid. In the early 1970s, however, Wisconsin sharply raised its benefits. In 1975, moreover, it decided to increase its payment standard in step with a national plan for a lowerincome budget set by the U.S. Bureau of Labor Statistics, which was indexed for inflation.8 These steps caused benefits to rise steadily through the early 1980s. The increases transformed Wisconsin from a state with average benefits into a leader. Table 2.1 shows the monthly amounts paid by Wisconsin and the median state to a needy three-person family (benefits vary by family size) at several points between July 1970 and January 2000. Benefits are shown in both nominal terms (current dollars) and in real (inflation-adjusted) terms. In January 1970, Wisconsin’s benefit precisely matched that of the median state. Already by July 1975, the state’s rank had soared to fourth, behind only Hawaii, Alaska, and Connecticut. Between July 1970 and January1987, when the median state’s benefits fell by a third in real terms, Wisconsin’s actually rose by 4 percent. Only California and Maine raised them by more. All other states reduced them.9 We usually think of welfare as paying benefits well below the federal poverty line. But by one calculation, Wisconsin’s cash benefit combined with in-kind benefits such as food stamps and Medicaid generated an income equal to 128 percent of the federal poverty line in 1986, a level second only to Minnesota.10 These changes partly explain why Wisconsin experienced growing welfare rolls, as shown in the bottom line of table 2.1. The surge extended to General Relief (GR), a local aid program without federal funding that

22

•

Chapter 2

TABLE 2.1 Nominal and Real Benefit Levels for Wisconsin and the Median State and Wisconsin’s Caseload, 1970–2000 (benefit figures in dollars) Benefits Nominal (current dollars) Wisconsin Median state Real (January 2000 dollars) Wisconsin Median state Wisconsin’s rank (50 states and D.C.) Wisconsin’s caseload (cases in 1,000s)

July 1970

July 1975

January 1987

January 1994

January 2000

184 184

342 235

544 354

517 366

628 421

796 796

1065 732

826 537

597 423

628 421

26.5

4

5

12

5

21.8

53.0

98.3

158.4

6.7

Sources: Benefit data from U.S. Congress, House, Committee on Ways and Means, Overview of Entitlement Programs: 1994 Green Book: Background Material and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, July 15, 1994), pp. 375–77; and U.S. Congress, House, Committee on Ways and Means, 2000 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, October 6, 2000), pp. 389–90. Note: Inflation adjustment uses consumer price index for all urban consumers, not seasonally adjusted, with 1982–84 = 100, downloaded from U.S. Bureau of Labor Statistics Web page, October 22, 2002. Caseload data from U.S. Department of Health and Human Services, Office of Human Services Policy, Aid to Families with Dependent Children: The Baseline (Washington, DC: U.S. Department of Health and Human Services, June 1998), p. 30 (annual averages used for July 1970 and July 1975); Wisconsin Department of Workforce Development (other times).

chiefly supported single working-aged men, mainly in Milwaukee. This program soared from 5,000 cases in an average month in 1980 to nearly 23,000 in 1985.11 Growing dependency meant growing fiscal strain. Between 1971–72 and 1984–85, the state’s AFDC spending in current dollars grew by an average of over 13 percent a year, and in some years by over 20 percent. By 1985–87, just before Tommy Thompson took office, the state was spending $1.6 billion, or 16 percent of all state spending, on AFDC and other welfare programs. Wisconsin had unusually high public spending per capita, and a third of the difference was due to welfare.12 To this point, growing AFDC rolls attracted only limited criticism, in part, because they seemed due to forces beyond the state’s control. The 1970s were a troubled decade for the economy in much of the country, and the recession of the early 1980s was the most severe since the Great Depression. It was especially sharp in Wisconsin, decimating the heavy

Development of Reform

•

23

industry, such as steel and automobiles, for which the state was known. Between 1979 and 1983, manufacturing employment fell 17 percent in Wisconsin, compared to 12 percent in the rest of the country, although it also rebounded strongly during the ensuing recovery.13

REFORM BEGINS Reform effort began seriously in 1979, when the legislature authorized a Welfare Reform Advisory Committee to look into the welfare problem. It was chaired by Robert Haveman, an economist at the University of Wisconsin–Madison who had formerly directed the Institute for Research on Poverty there. IRP was an important, federally funded center for antipoverty scholarship. The committee recommended ambitious changes in work incentives and child support enforcement. Its report was well received by the media and political leaders, including Lee Dreyfus, the Republican governor. But its emphasis was on expanding opportunity for the recipients, not on enforcing changes in lifestyle.14 One eventual result was the enactment, in 1984, of a state earned income credit, a wage subsidy for the working poor that supplemented a federal credit.15 Another result was pioneering reforms in child support enforcement. Following the ideas of Irwin Garfinkel, a colleague of Haveman’s at IRP, the state instituted a fixed standard for the amount of court child support awards, curbing the discretion of judges. It also mandated that judgments be deducted from absent parents’ paychecks automatically.16 The changes improved Wisconsin’s already strong performance in enforcing support. But the relentless advance of Wisconsin’s AFDC rolls continued. THE CRISIS OF 1985–86 Work-enforcing reform began in 1985–86. The state somehow woke up to its welfare problem more fully than before. One trigger was simply the steady rise in the rolls. They would eventually crest at just over 100,000 cases and 300,000 persons in April 1986. In a state population of under 5 million persons, numbers like this provoked alarm. The recession was long past. Now for the first time, government rather than society was clearly blamed for the welfare problem. Politicians, led by Joe Andrea and other Democratic legislators from the southeast corner of the state, charged that Wisconsin had become a welfare magnet. Its generous benefits, they claimed, were causing poor families to migrate to Kenosha or Milwaukee from less generous states—especially from Chi-

24

•

Chapter 2

cago. Indeed, the increase in the state’s poor population in the early 1980s was twice as large as could be explained by other causes.17 The issue blindsided Tony Earl, the Democratic governor. He called for research on the issue; studies left it unclear whether people really moved to Wisconsin to get higher welfare.18 But the liberal welfare consensus was shattered. When Earl next proposed the usual welfare benefit rise, legislators trimmed it. They told the Department of Health and Social Services, the state welfare department, to use the savings to develop a better welfare work program.19 This decision was a watershed. A consensus on a work strategy for reform had already gelled—prior to the election of Tommy Thompson as governor. The new program became the Work Experience and Job Training program. WEJT provided more intensive case management and education and training services than any previous welfare work program. Most controversially, it included a Community Work Experience Program (CWEP), or unpaid jobs to which clients would be assigned who persistently failed to find private positions.20 Enacted in April 1986, WEJT was implemented first in five counties in 1987 and later expanded to most of the state. It proved to be critical to Wisconsin’s exemplary ability to enforce work in welfare, first under AFDC and then W-2, as chapter 4 shows. When Thompson, who had previously been Republican leader in the Assembly, ran against Earl for governor in 1986, few expected him to win. Earl represented the liberal establishment that had lately dominated state politics. Democrats had controlled the Assembly and the governorship since 1971 (with the exception of Dreyfus’s governorship in 1979– 83) and the Senate since 1975. Earl saw himself as a problem-solving policymaker, but he was also a maladroit politician who got out of step with the voters on several counts. He was unwilling to face up clearly to the need for basic change in welfare, which the public viewed as out of control. Thompson capitalized on that sentiment, among other issues, to oust him.21

THOMPSON’S INITIAL REFORMS When Thompson took office in 1987, he played out a reform scenario much like what had occurred in 1985–86. Money was saved on cash aid, then invested in reform programs, but on a grander scale. Thompson’s most contentious initial proposal was for an actual cut in benefits, not just a limited increase. The budget for 1987–89 as passed by the legislature cut benefits by a token 1 percent, but by adroit use of his item-veto power, Thompson raised that to 6 percent.22 As Table 2.1 shows, that dropped

Development of Reform

•

25

the state’s benefit ranking to twelfth by 1994. But Thompson made no further changes in grants until 1997—when he raised them. By 2000, that again made the state a benefit leader—but only after reform had obliterated the rolls. Thompson also made no large cuts in the burgeoning General Relief program until 1996, when it was converted to a state block grant to counties, thus capping the state obligation.23 The benefit cut was to generate $23 million in savings. Thompson proposed to invest nearly all of that in improvements in the welfare system. Besides new money for support services, this chiefly meant new programs meant to demand greater personal responsibility from government’s dependents. First, Thompson expanded WETJ to much of the state. He also expanded the Community Work Experience Program in smaller counties independently of WEJT; in practice it operated similarly, offering a range of services and not just government jobs.24 Another innovation was Learnfare, an ambitious attempt to extend the logic of work requirements into areas besides employment. Welfare parents were now to be required, not just to work, but to keep their children in school. Teenagers aged 13–19 who were on welfare had to maintain steady attendance in school, on pain of reduction in their families’ grants. The idea was to make welfare adults more accountable for the behavior of their children. Enacted in 1987 and implemented in 1988, Learnfare encountered serious implementation problems, as detailed in chapter 8. Nevertheless, in 1993 it was extended to welfare children aged 6–12 in Brown, Fond du Lac, Kenosha, and Rock Counties, and in 1997 statewide. (See Figure 2.1 for the Wisconsin counties.) A third program, Children First, attempted to go beyond the earlier child support reforms. Those changes had put more pressure on absent fathers to support their families, but they could not assure that the fathers actually worked enough to do so. Children First required that nonpaying fathers who were remanded to the program either participate or pay up, on pain of going to jail. Enacted in 1987, the program was implemented in Fond du Lac and Racine in 1990, then expanded to other counties. As explained in chapter 8, it became a pioneering experiment in paternalistic social policy.

THE WAIVER EXPERIMENTS Another key dimension of Thompson’s approach to reform was what became known as waiver experiments. While AFDC allowed states great leeway in setting benefits and some aspects of eligibility, they could not normally flout national guidelines in other respects if they wanted to re-

26

Chapter 2

•

Superior Douglas

Bayfield Ashland

Iron Vilas

Burnett

Washburn

Sawyer

Florence Price

Polk

Oneida

Forest Marinette

Rusk

Barron

Lincoln

Langlade

Taylor St. Croix

Chippewa

Menominee

Dunn

Marathon

Eau Claire

Pierce Pepin

Clark

Oconto

Shawano

Door

Eau Clair

Kewaunee Portage

Wood

Buffalo

Waupaca

Outagamie

Jackson

Brown

Trempealeau Monroe

La Crosse

Juneau

Winnebago Calumet Manitowoc

Waushara

Adams

La Crosse

Green Bay

Marquette Green

Lake

Fond du Lac

Sheboygan

Vernon Sauk

Richland

Ozaukee

Dodge Washington

Crawford Grant

Columbia Dane

Iowa Lafayette

Madison Green

Rock

Jefferson

Milwaukee

Waukesha

Milwaukee Walworth

Racine Kenosha

Kenosha

Figure 2.1. Wisconsin and its counties.

tain federal funding. However, Washington sometimes waived its rules if a state claimed that an experiment would teach something about how to improve welfare.25 Typically, a waiver would be granted for a few years conditional on cost neutrality and an evaluation to test the program. Starting in the 1980s, many states obtained waivers to experiment in AFDC, chiefly to institute more demanding work programs than those normally allowed by federal policy. At the time when Thompson was elected, most governors still hoped to strengthen work tests at the federal level. That effort eventually led to the Family Support Act of 1988 (FSA), which expanded work programs. But the resistance to serious work tests was still strong among liberal community groups and Democratic members of Congress. To evade that impasse, the waiver process developed as an alternative route to reform. Governors saw the political advantage of launching waivers, and the Reagan, Bush, and Clinton Administrations all promoted this as a way of producing change. By 1991, around thirty projects were running in eighteen states.26

Development of Reform

•

27

Perhaps due to its initial commitment to liberal ideas of welfare, Wisconsin had not been among the early welfare experimenters. Tommy Thompson changed that. Just after taking office, in February 1987, he attended a meeting of the National Governors Association in Washington. The governors endorsed a massive new welfare work program, to be funded and run largely by the federal government. Thompson was the only governor to dissent. He believed that Wisconsin was already “way out front” on welfare reform, and to wait for change at the center would hold him back. Instead, he went to the U.S. Department of Health and Human Services (DHHS), the federal welfare agency, and asked for waivers to let him carry out his own experiments. By 1996, Wisconsin had received eleven waivers, the most of any state.27 Besides Learnfare, the first group of Thompson waiver experiments to be approved and implemented included: • 20-Hour Work Requirement: Since 1967, welfare parents had faced no work tests until their youngest children turned six, a rule that exempted two-thirds of the caseload. FSA lowered this age to three, but still exempted about half of mothers. This waiver, implemented in 1988, allowed Wisconsin to obligate mothers to work when their children were as young as one, provided child care was provided. Initially, they had to participate twenty hours a week; after FSA, up to forty hours. As explained in chapter 4, this waiver was crucial to Wisconsin’s ability to build up work requirements. • $30 and 1/6 Work Incentive: This was a revision of the standard AFDC work incentive in effect in 1987. That provision allowed a welfare mother who went to work to keep $30 a month and one-third of the remainder from her wages, before the rest was deducted from her grant. Essentially, starting in 1989, Wisconsin slightly reduced that incentive but stretched it from four months to a year, in hopes of making it more effective.28 • Medicaid Assistance Extension: In 1987, welfare recipients were normally allowed Medicaid coverage only on welfare, not off. Lest they hesitate to leave welfare for work, this waiver, beginning in 1989, assured recipients of Medicaid coverage for twelve months after they left AFDC for employment. In 1988, FSA would mandate such coverage nationwide.29 • 100-Hour Rule Demonstration: In an attempt to mitigate AFDC’s antimarriage effects, Congress had allowed coverage of two-parent needy families since 1961, but only if the principal earner (usually the father) worked less than 100 hours a month. This waiver deleted that restriction, starting in 1991, in hopes of promoting more two-parent families on welfare.

Alongside the waivers, Thompson negotiated an advantageous financial deal with Washington that was to prove vital in funding the Wisconsin reform. Cuts in welfare dependency could be anticipated as a result of the experiments. Both Wisconsin and Washington would save money, be-

28

•

Chapter 2

cause both shared in the funding of AFDC. After complex negotiations, the state was allowed to claim $148 million in savings over 1987–97, including $9 million in the first year. Wisconsin would be reimbursed from this pool for the cost of its experiments at a 50 percent rate. For every dollar it spent, that is, it could claim a dollar of federal funds.30 A renewed push to promote waivers occurred in the first Bush Administration, and Tommy Thompson responded with a new list of initiatives. Several of these tinkered further with the incentives surrounding AFDC, without directly denying aid or conditioning it on work. • Vehicle Asset Limit Demonstration: Starting in 1995, this waiver raised the value of an automobile one could own and still qualify for AFDC from $1,500 to $2,500, in hopes that more recipients would buy cars to get to work. • Special Resource Account: AFDC also forbade a family to have more than $1,000 in assets and qualify for aid. This waiver, also beginning in 1995, allowed a family to accumulate up to $10,000 in a “special resources account” for education or training purposes. • Two-Tier AFDC Benefit Pilot Program: To weaken the “welfare magnet” effect that might attract recipients to Wisconsin, this waiver paid migrants to the state for the first six months the same benefit they received in their former state. Usually this was less than Wisconsin’s benefit. The idea was controversial, because the U.S. Supreme Court had already disallowed residency requirements on eligibility for AFDC.31 Thompson had trouble selling the idea to DHHS and the legislature. Two-Tier finally began in 1994 in Kenosha, Milwaukee, Racine, and Rock counties.32

More ambitious was Parental and Family Responsibility (PFR), popularly known as “Bridefare.”33 Much of welfare’s growth was generated by unwed births to teen parents. PFR strengthened work incentives still further, beyond the level of the waiver above, while at the same time reducing the additional welfare that a teen mother could claim if she had additional children on the rolls. It also set aside the 100-hour rule and other restrictions on two-parent welfare. The changes aimed to motivate a teen mother to work rather than add to her family, and to cause her spouse to work and join the household. PFR began in 1994 in Juneau, Milwaukee, Oneida, and Rock Counties. In addition to incentives, PFR included paternalistic elements also seen in Learnfare and Children First. The counties running PFR were to provide the clients with intensive case management. The young parents were supposed to receive education on proper parenting, family-planning, and work behavior. They could be sanctioned if they did not cooperate with the training.

Development of Reform

•

29

THE RADICAL PHASE OF REFORM A third wave of the waiver campaign came in 1993, when the Clinton Administration took office with welfare reform as a leading goal. Like his Republican predecessors, Clinton pushed welfare experimentation, and Tommy Thompson responded with yet further ideas. They were still more severe than what had gone before. One of these was the AFDC Benefit Cap: A common suspicion was that some welfare mothers had additional children in order to gain higher welfare benefits, since grants rose with the size of the family. Under this program, which began statewide in 1996, mothers were denied any benefit increase for children born longer than ten months after they went on the rolls, although the children still received food stamps and Medicaid. The new rule did not apply to clients subject to PFR. Far more important was Work Not Welfare.34 WNW experimented with a form of welfare that was radically time-limited and work-conditioned, as would later be institutionalized in W-2. Work tests and time limits had come to dominate welfare reform thinking on the national scene. But as Tommy Thompson could claim, WNW was “the first program that absolutely required welfare recipients to find work.”35 That Washington approved a waiver for so radical a program shows which way the wind was blowing. WNW, which was implemented in Fond du Lac and Pierce Counties in 1995, built on earlier waiver programs. The $30 and 1/6 earnings disregard was stretched from one year to two, the 100-hour rule was set aside, and a benefit cap was imposed on children born on the rolls (the benefit cap waiver was not yet in effect). In addition, food stamps were cashed out (their cash value was added to the AFDC grant), and any child support received was paid directly to the family rather than to the welfare system. The family received a $50 disregard for unearned income. But above all, with certain exceptions, eligible families could stay on WNW for only 24 months in any 48-month period. Equally important, the family had to “earn” its benefits with employment activities even while on the rolls. Adult family members could be required to participate up to forty hours a week, the obligation varying with the number of adults in the family and the size of the grant. The benefit was docked $4.25—the minimum wage—for every hour below forty. The activities that qualified were strongly slanted toward actual work. Education and training programs were limited to those that could be completed within two years. However, to support working, the program also provided generous subsidized child and health care. Like PFR, WNW had paternalistic aspects. Families were assigned to case

30

•

Chapter 2

management teams that explained the program to them and implemented the services. WNW was radical for another reason too. For reasons explained in chapter 3, its legislation was amended to require the state to abolish and replace AFDC no later than 1999. The planning to do that thus went forward while all the later experiments were implemented.

THE DIVERSION PROGRAMS Starting in 1994, the state implemented other radical reforms that anticipated W-2. The earlier experiments—like the 100-Hour Demonstration— applied to a small share of clients statewide, or, like Work Not Welfare, affected a larger share only in selected counties. The new programs reached more widely. Two of them applied to Milwaukee, with its large and entrenched caseload, which had never before felt the full brunt of reform. The new programs involved what has become known as diversion. They no longer sought, like earlier experiments, to change the work incentives within welfare, or to require work of people already on welfare. Rather, they sought to keep people from going on the rolls at all. They did not do this in the traditional manner by denying aid or tightening up on the initial eligibility rules; in Wisconsin, “quality control” was never a mainstay of reform. Rather, they did it by pressing applicants to consider whether they had exhausted all possibilities for support before asking for aid. These other options included getting help from family, friends, or local charities, and, above all, looking for work immediately. The effect was to cut sharply the intake of new applicants to the rolls, even among people who qualified for aid on income grounds. Normally, when applicants approached local welfare agencies, they were immediately processed to determine their eligibility. Under Work First, they were first counseled against unnecessary dependency and invited to pursue other options, including immediate participation in JOBS. Work First was technically a modification of JOBS and required no waiver. After initial development in Grant, Jackson, and Kenosha Counties (see chapter 5), it was implemented in November 1994 in 18 counties, not including Milwaukee. By the end of 1995 it extended to 48 counties out of 72.36 Self-Sufficiency First (SSF) was a tougher version of the same thing. Under Work First, applicants for aid did not technically have to cooperate with diversion. They could refuse to enter work activities without penalty until their applications for aid had actually been approved. Under the SSF

Development of Reform

•

31

waiver, however, they had to attend work orientation sessions and put in 60 hours looking for work for 30 days prior to going on aid—or their applications would be denied. SSF’s companion was Pay for Performance. PFP toughened the sanction for noncooperation with work requirements. As in Work Not Welfare, recipients of aid had to “earn” their benefits by putting in twenty to forty hours of activities, on pain of reductions in their grants. These activities might include work, job search, or training. Previously, adults who refused to cooperate with work tests lost only their portion of the welfare grant for several months; aid continued for other family members. Under PFP, however, hours of assigned activities that a client missed without good cause were docked from the grant at a rate of $4.25 an hour, the minimum wage. If participation fell below a quarter of assigned hours, the grant was ended entirely.37 SSF and PFP were implemented in March 1996 for the entire state. They were thus the first reform programs to impact Milwaukee seriously. They produced little short of a revolution there. Tens of thousands of recipients suddenly faced serious demands to look for work and participate in work programs for the first time. Amid great confusion, huge numbers simply left the rolls, with the majority apparently taking jobs, as I elaborate in later chapters. In large part, W-2 merely institutionalized the changes that SSF, PFP, and earlier innovations had already established.

WAS POLICYMAKING EFFECTIVE? Were the reforms “effective welfare policies” in the sense used in Table 1.3? Did policymakers act wisely to achieve their ends given what they knew at the time? Overall, they clearly did. Their goals were chiefly to enforce work and, secondarily, to reduce the welfare rolls. One might have promoted those ends simply by cutting back aid, but cuts in eligibility or benefits played only a minor role in the Wisconsin story. The state was committed to generosity to the poor, so it had to find a way to raise work levels and limit dependency while keeping open the door to aid. That required it to address the problems that caused people to become poor or dependent. In the early stages, Thompson and his aides aimed at various targets, including keeping children in school, toughening child support payment, and promoting marriage. Eventually, however, they focused mostly on requiring clients to participate in work programs and to look for and accept available jobs. These

32

•

Chapter 2

were the aims, they found, that they could most readily advance. Those are the aims, above all, of Work Not Welfare and the diversion programs. By pursuing work, after a decade of reform, Wisconsin finally did transform its welfare system, and did so far in advance of the nation. Casual Policymaking The manner of the reform may appear less impressive. According to academic models of policymaking, decision makers should first define their problem, then conceive options for solving it, then forecast the likely results of choosing one or the other. They then choose the option that achieves the original goals most fully, or at least cost. By this standard, the analysis behind the early Wisconsin reforms can easily appear perfunctory. Tommy Thompson claims to have thought up Learnfare while driving across the state with an aide during the 1986 election campaign. The idea for the Special Resource Account came from the work of Michael Sherraden, an economist who emphasizes savings rather than work as the key to overcoming poverty. Eloise Anderson, a senior administrator for DHSS, conceived PFR and then sold it to the legislature. Work Not Welfare was launched in a hurry by Thompson; Gerald Whitburn, the secretary of DHSS; and other aides as a response to Clinton’s welfare reform initiative.38 The idea of diversion apparently occurred first to Jason Turner, a DHSS manager who headed welfare reform initiatives under Thompson. In 1993, during the planning of WNW, Turner wondered why those who were eligible for aid on an income basis should necessarily receive all the benefits to which they were entitled, if they had other options. He had heard of programs in Grant County and also in Oregon that attempted to limit eligibles to the aid they really needed, and he went to visit them. Out of this thinking emerged first Work First and then Self-Sufficiency First. Turner also invented Pay For Performance, based on an experimental program he had heard of in Chicago where recipients progressed toward employment by undertaking rising hours of activities over time.39 Planners consulted past research and evaluation on welfare and poverty selectively. In the 1980s, academic experts on poverty tended to doubt that enforcing work in welfare was desirable or possible.40 However, there was research showing that welfare recipients turn over rapidly in jobs, rather than never working at all. That suggested to the Wisconsin planners, not that work was impossible, but that it must be repeatedly enforced. They also took from evaluations of past welfare work programs

Development of Reform

•

33

the conclusion that training to increase skills was ineffective and reform should stress “work first.”41 However, the shape of reform was set more by policymakers’ own experiences and values than by any external input. In this, Wisconsin resembles other states. While state officials may know of prominent experiments elsewhere, their knowledge of such findings is often sketchy. Governors and senior agency leaders attach greatest weight to their own program history. This approach is most legitimate politically, and “what works” depends on implementation details that are particular to the state. Officials also take little direct input from their own electorate. Rather, they decide what to do by their own lights, then sell it to the legislature and the public.42 For similar reasons, Wisconsin officials also devoted only limited attention to evaluating their programs. They were required to evaluate the waiver programs as a condition of getting permission for them from Washington, but they placed little weight on these assessments. The state haggled over how to do the evaluations, in some cases persuading federal officials to accept less-than-experimental research designs.43 The studies themselves also ran into significant problems.44 As chapter 10 shows, those that were finished do not suggest that these experiments achieved much. This disinterest reflects the trust policymakers had in their own intuition. Jason Turner remarked that officials could tell 80 percent of what was effective or ineffective about a waiver program prior to an evaluation.45 That might be plausible in the setting he faced, where previous policies had manifestly failed and most of the new programs represented large departures from the status quo. If change were more incremental, evaluations would be more needed to gauge whether it achieved anything. Senior officials have shown more interest in evaluating W-2. That may be because the new structure is clearly well established, and there is thus more need to fine-tune it.46 Even at the outset, it was unlikely that most of the early Thompson innovations would produce much change. The experiments based on incentives did not change welfare radically. Other programs involved more forceful interventions, but they affected only handfuls of clients. The twenty-hour work requirement, for instance, affected few clients because most counties in 1988 lacked the resources to implement a work test for more mothers than they already served; this changed later, as chapter 4 shows. PFR involved an elaborate design, but five months after it was implemented in four counties in 1994, only 122 clients were enrolled in it. Similarly, Two-Tier, also beginning in 1994, enrolled all of 436 clients in four counties in its first three months.47

34

•

Chapter 2

The Limits of Analysis However, this pattern of policymaking must have been more effective than it seems, or the state would not have achieved the revolution it did. This was so, I think, for two main reasons. First, the orthodox model of policymaking assumes that goals can be specified independent of means. The reality is that goals and means are interdependent. Government only pursues goals that it has means to achieve. If programs fail, it shifts to other goals where it has programs that can succeed.48 In Wisconsin, as in the nation, welfare reform focused more closely on work over time because this proved to be the goal that could break the mold of traditional welfare. As Thomas Kaplan says of W-2, “the central themes of the new program—personal responsibility, work, and revamping the mission of the welfare offices—emerged only gradually and haltingly from a decade of experimentation.”49 A more serious problem is that the orthodox model assumes too much prior knowledge. The usual criticism is that it requires too much calculation. In practice, decisionmakers “satisfice,” or consider only limited options, because they cannot know enough to optimize.50 But this criticism still assumes that we at least know what options to investigate. That may be true for an old problem with a long policy history. Wisconsin in the mid–1980s, however, did not have a lot to go on. In the past, antipoverty programs had stressed work incentives or compensatory education and training for the low-skilled. This probably was why WEJT initially emphasized education and training, and why many of the early waiver programs relied on incentives to influence behavior. But the evaluation history of such programs was disappointing.51 Mandatory programs that required positive behavior as a condition of aid would eventually demonstrate more leverage over lifestyle, but the idea of enforcement as a social strategy was barely visible when the Wisconsin reform began. The first serious evidence that mandatory work programs could raise work levels and cut dependency on AFDC appeared only in 1984–86, from an evaluation in San Diego.52 And the San Diego results did not settle all questions. Could welfare work programs be voluntary as well as mandatory? Could they stress education and training for well-paying jobs, or must they demand going to work immediately in lower-skilled, lower-paid jobs? In the 1980s, most experts felt that clients and staffs would be motivated to succeed only if programs were voluntary. They also believed that government had to invest in clients’ skills to make them more self-sufficient in the long term, even if this meant they did not leave welfare quickly.53 Only in the later 1980s did evidence appear suggesting that mandatory programs were preferable.54 Only in the mid–1990s did it become clear

Development of Reform

•

35

that requiring work in available jobs achieved more than attempting to improve skills.55 The general progression in Wisconsin is from family programs to work programs, from incentives to enforcement, and within the work programs from an education and training focus to work first. That partly reflected evaluation evidence as it came in, but since Wisconsin was so far out front, it also reflected what officials themselves sensed was working in the field. If we remember what policymakers knew at the time, the decisionmaking process was more rational and less confused than it may appear. Viewed in context, it really was problem-solving of a high order. John Kingdon has suggested that policymaking in American politics often springs from unpredictable combinations of problems, policy ideas, and political opportunities. For government to act, there has to be challenge that everyone recognizes, a policy that might plausibly overcome that problem, and clear political support. The Wisconsin reform took shape from a strenuous effort to bring all these elements into alignment. There was a will to change, but the welfare problem had to be defined in a way that permitted change. The Thompson Administration and the legislature came to define it in terms of work because this was the meaning where agreement was broadest and where also effective programs were available.56 Policymaking As Education The orthodox view of policymaking also treats policy analysis as separable from politics. Serious theorists of analysis of course recognize that politics imposes limits on the choices officials can meaningfully discuss. There is no point in considering options to solve a problem that will be rejected by the legislature or the voters. But this still assumes that analyzing policy is something separate from selling it, that politics influences policymaking rather than the other way around. Actually, policy decisions can influence the political process, and programs on the books can prepare the way for future programs. Tommy Thompson realized that it was more important to begin a process of change than to know precisely where it was headed. By proposing one initiative after another, he got people thinking about change, and he kept his opponents off balance. He legitimized the idea that welfare, which had been sacrosanct, could be changed without the roof falling in. Thompson aimed to get the state sufficiently committed to reform so that there was no going back. To this end, the difficulties the early programs encountered were almost a virtue, because they compelled further policymaking.57 Each program Thompson pushed through thus had an educative effect. The early experiments can easily seem inconsequential, and they came on

36

•

Chapter 2

stream too late to get much credit for Wisconsin’s unusual caseload fall, which dates from 1986.58 But politically, they were essential. They changed the discourse surrounding welfare and thus prepared the way for the later and more radical programs, whose effect is undoubted. Policymaking is a process as much as a decision. The process of reform also prompted administrators and the recipients themselves to reassess welfare. Chapter 4 shows how officials changed JOBS from emphasizing services to clients to demanding that they work in return for aid. They made these changes, in part, because of the clear messages they received from the legislative arena about the direction of reform. As for the recipients, as chapter 9 will show, many of them began leaving aid as early as 1986–87, as soon as serious controversy about welfare broke out. They got a message about self-reliance long before serious reforms were even legislated, let alone implemented. Statecraft Wisconsin-style reveals a subtle mix of policymaking and politicking that can never be reduced to a theory. Politics set an agenda for policymaking, but the making of the reform and its favorable effects also reshaped the politics, in a synergistic cycle. The real secret of success was not that knowledge came ahead of action, or analysis ahead of politics, but that both politicians and administrators did their jobs so well.

CHAPTER 3

The Politics of Reform

THIS CHAPTER COVERS the extraordinary politics of the Wisconsin reform.1 I emphasize the overall patterns and events up through the decision to abolish AFDC in 1993. Deep political division has often impeded welfare reform, at the national level and in many urban states. Yet Wisconsin moved, in only seven years, from a liberal regime that embodied entitlement to a decision to end that system. What can explain this? One interpretation is conventionally partisan. The political winds were shifting to the Republicans. Tommy Thompson made effective use of the welfare issue to win the governorship and then keep his opponents off balance. This view, however, cannot explain why reform began prior to Thompson’s election as governor, or how he sold his early measures to a legislature that Democrats still controlled. I place more weight on a high-minded political tradition that drove both parties, though rivals for power, to focus on a common problem. That in turn drove them toward a common, work-based solution. They had the support of the bureaucracy. Meanwhile, groups outside the legislature that might have blocked change were surprisingly quiescent. This moderate and problem-solving style typifies what Daniel Elazar called a moralistic culture.

THE POLITICS OF WELFARE Welfare politics has not generally been edifying. Ever since AFDC became a national issue in the 1960s, the system proved difficult to change, despite its unpopularity. In Washington, the history is one of repeated Congressional stalemates or, at best, grudging compromises that satisfied neither party. A clearly radical reform came only in 1996, and it too was nonconsensual. In the 1960s and 1970s, liberal planners proposed to expand welfare to cover the two-parent poor and raise benefits. But their principal vehicles—President Nixon’s Family Assistance Plan and President Carter’s Program for Better Jobs and Income—were defeated in Congress. Various partisan and group divisions prevented action. Moderates of both parties would not accept an expansion of welfare without clear work tests, which liberals in turn rejected. Welfare came to be seen, in the words

38

•

Chapter 3

of Joseph Califano, an aide to both Presidents Johnson and Carter, as “the Middle East of domestic politics”—a subject on which agreement seemed impossible.2 With national politics tending right, the liberal approach to welfare became impolitic by the late 1970s, and the conservative approach took over. Since then, Congress has toughened up child support enforcement and fraud and abuse control. It has stiffened work tests bearing on the adult recipients, and it has devolved much more control of welfare to the states. But most of these measures rested on difficult compromises between the parties. Liberals and many Democrats have resisted change for fear that a public commitment to the poor would be lost. Conservatives and many Republicans, for their part, hungered to undo entitlement. No deep consensus joined the two sides. When radical change came with PRWORA in 1996, it was chiefly because Republicans had at last won a majority in Congress. PRWORA was a defiantly conservative measure that flouted several liberal principles— not only entitlement but open-ended aid and federal leadership in social policy. PRWORA also stressed familist goals, such as reversing unwed pregnancy and promoting marriage, that were anathema to the left. Clinton signed PRWORA with misgivings, chiefly because he had promised to end the old welfare. Advocate groups protested, and Clinton’s welfare planners resigned from the administration.3 PRWORA gained a majority in Congress, but nothing more. Many states have also faced political difficulties over welfare. Especially in the more liberal, urban states of the North, welfare sparked deep divisions between the parties and, often, tensions between state and local governments as well.4 TANF forced states to confront these differences because it demanded that they institute tougher worker tests and assume more responsibility for welfare policymaking than ever before. As I show in chapter 11, not all states responded well. In Wisconsin, however, we do not see stalemate, or grudging compromise, or the simple triumph of the right. Rather, after a liberal era dominated by Democrats, both parties collaborated in the reforms launched chiefly by Tommy Thompson. Despite infighting over details and timing, leaders on both sides of the aisle crafted a reform that was both radical and consensual. THE SETTING FOR REFORM Some might attribute this success to economic or social conditions that made it easier to reform welfare in Wisconsin than elsewhere. The state, for example, has recently enjoyed low unemployment. A tight labor

Politics of Reform

•

39

market clearly is one reason for the state’s dramatic fall in dependency. But many other states also had good conditions in the late 1980s or the 1990s, without enacting comparable changes or reducing their caseloads by as much. Nor was the state’s welfare caseload clearly less disadvantaged than the norm. Minority status is one indicator of disadvantage. In 1990, 31 percent of Wisconsin’s AFDC recipients were black and 6 percent Hispanic, somewhat lower than figures of 40 and 16 percent for the nation.5 On the other hand, the state’s unusual caseload decline made its cases more nonwhite than the average over time, since the more employable cases were the first to leave welfare. By 1994, 52 percent of Wisconsin’s caseload was nonwhite, compared to 59 percent nationally. By 1998, the figure was 83 percent for Wisconsin, compared to 64 percent nationally.6 And throughout the reform process, the nonwhite caseload was overwhelmingly concentrated in Milwaukee , making race more sensitive than if nonwhites had been more dispersed. In December 1993, the caseload was 60 percent black and 10 percent Hispanic in Milwaukee, while it was 69 percent white in the rest of the state.7 Outside Milwaukee, the caseload is mostly white, little distinct from the rest of the population. Rural welfare recipients tend to be short-term, the lower end of the working class.8 But in Milwaukee, reform had to confront one of the most segregated and depressed cities in America. Between 1970 and 1990, the city’s economy sharply declined and lost factory employment. The nonpoor fled the city as poor blacks moved in. Ghetto poverty sharply expanded, particularly in the 1980s.9 By 1990, 47 percent of all blacks and 65 percent of all poor blacks in Milwaukee lived in areas that were 40 percent or more poor, the highest figures for any city in the country.10 Yet despite this, the state eventually enforced the same severe work policies in the city as in the rest of the state. Turning to politics, one might suspect that the Wisconsin public is unusually hostile to welfare, but it is not.11 Polls from the late 1980s and early 1990s do find about 80 percent of Wisconsin residents concerned that the state’s generous benefits were drawing recipients from other states. Forty-one to 53 percent, a plurality, believed benefits were too high, while less than 10 percent called them too low. Large majorities thought that most recipients of aid did not really need it. Between 1988 and 1998, welfare along with other social issues ranked on average third among eight leading policy concerns, led only by taxes and crime. At the same time, however, around 90 percent of respondents endorsed the principle of helping people unable to support themselves. Table 3.1 shows support for certain leading proposals for reforming welfare. The reforms are numbered, ranked by percentage of all respondents in favor. Large majorities want to make absent fathers pay child

40

•

Chapter 3

TABLE 3.1 Support by Wisconsin Residents for Welfare Reform Measures (percent in favor)

Reform Proposal 1. Taking money out of the paychecks and taxing refunds of fathers who refuse to make child support payments. 2. Limiting new residents of the state to the benefit they would have received in their former state for six months. 3. Requiring recipients to hold a public or private job in order to receive benefits under W-2. 4. Requiring people who move to Wisconsin to wait a period before receiving welfare. 5. Requiring that welfare recipients lose some of their benefits if their children drop out of school or are absent more than a pre-set number of days per year. 6. Providing public service jobs with the government for welfare recipients who are able to work if they cannot find work with a private employer. 7. Requiring teenage single parents who have children out of wedlock to live with their parents or another responsible adult as a condition of receiving welfare benefits. 8. Limiting additional support that single teenage mothers receive from welfare if they give birth to a child while on welfare. 9. Allowing a family to supplement welfare payments with income from a minimum-wage job without decreasing the amount of their welfare payments. 10. Denying all welfare benefits to any woman under the age of twenty-one who has a child out of wedlock. Children whose mothers could not find any way to support them would be placed in a residential child care facility. 11. Mothers on welfare receiving higher benefits if they marry the father of their child. Abolition of AFDC by 1999 Passage of W-2 Passage of PRWORA

Percent Total

Percent Whites

Percent Blacks

95

96

93

93

NA

NA

87

NA

NA

87

NA

NA

85

86

72

84

84

88

74

74

75

68

69

54

57

NA

NA

32

33

10

23

22

27

73 87 66

NA 88 67

NA 80 56

Source: Gordon S. Black, “The Wisconsin Citizen Survey” (Milwaukee: Wisconsin Policy Research Institute), various editions between February 1988 and August 1998. Questions showing results by race come from August 1994, October 1996, and August 1998. Note: On subjects where there was more than one poll, I have chosen a representative result. The only large discrepancy occurred on “two-tier” (question 2). The reported result, in August 1994, was 93 percent. In January 1992, it was 75 percent. This might be because in the earlier question the benefit limit was open-ended, while in the later one the limit was only for six months.

Politics of Reform

•

41

support, force new migrants to wait six months for aid and then receive the same benefits as in their former state, and require that welfare adults work and keep their children in school as a condition of aid. Weaker majorities want to make teen parents live at home, limit additional aid for children born on the rolls, and strengthen work incentives. Only minorities favor the more conservative measures of denying aid entirely to unwed mothers under twenty-one or promoting marriage by single mothers. The bottom panel of the table shows that clear majorities supported the end of AFDC and the enactment of W-2 and PRWORA. Of the three laws, PRWORA garnered the least support, probably because of its punitive and antigovernment overtones. Like other Americans, Wisconsites opposed the “abuses” of welfare rather than the principle of aid. They demanded that adult recipients show responsibility, above all by working, but they showed leniency toward families in difficulties not of their own making.12 As I showed in chapter 12, the public’s will to change welfare was probably stronger in the Midwest than in much of the country. Yet the changes people wanted were not unusual. What stands out about Wisconsin is not what the public wanted out of welfare reform but the fact that the government gave it to them. In this state, more than anywhere else in the country, welfare really was changed to condition aid on serious work effort, thus driving most of the employable off the rolls. The reasons for that transformation must lie chiefly, not in society, but in the special properties of Wisconsin’s government.

WAS REFORM PARTISAN? One view of the politics of reform was that it sprang from combat among the parties.13 This view has some plausibility because of the close battle for party control in the state. Table 3.2 lists the governors of the state since the late 1960s. Republicans and Democrats have alternated. Since 1971, only Tommy Thompson held the office for two or more terms. He won his fourth term in 1998, then resigned in 2001 to become Secretary of the U.S. Department of Health and Human Services in the second Bush Administration. Table 3.3 shows the balance of party power in the state legislature since 1969. Democrats controlled the Assembly for a long period starting in 1971, but their edge narrowed over time until the Republicans took control by a thin margin in 1995. In the Senate, Democratic control lasted from 1975 until 1993, after which control shifted by a seat either way in several elections. In 2002, Democrats recovered the statehouse after sixteen years, but Republicans captured both houses of the legislature.

42

•

Chapter 3

TABLE 3.2 Governors of Wisconsin, 1965–2003 Governor Warren P. Knowles Patrick J. Lucey Martin J. Schreiber Lee S. Dreyfus Anthony S. Earl Tommy G. Thompson Scott McCallum Jim Doyle

Party

Term Began

Term Ended

Republican Democrat Democrat Republican Democrat Republican Republican Democrat

January 1965 January 1971 July 1977 January 1979 January 1983 January 1987 February 2001 January 2003

January 1971 July 1977 January 1979 January 1983 January 1987 February 2001 Jan 2003

Sources: State of Wisconsin, 2001–2002 Blue Book (Madison: Wisconsin Legislature, Joint Committee on Legislative Organization, 2001), p. 710; “The 2002 Elections: State by State,” New York Times, November 7, 2002, p. B14. Note: Knowles served three two-year terms. Following a constitutional amendment ratified in 1967, the term became four years. Schreiber, who had been lieutenant governor, became governor in 1977 when Lucey resigned to become U.S. Ambassador to Mexico. Lieutenant Governor McCallum became governor in February 2001 when Thompson resigned to become Secretary of the U.S. Department of Health and Human Services. All other governors were elected on the election day preceding the beginning of their terms.

There is no denying that welfare trends in Wisconsin partly reflected which party held power. In 1969, in response to riots in Milwaukee, Republicans controlling the legislature attempted to cut AFDC benefits by 17 percent. One motive was to save money, but Republican rhetoric questioned the very legitimacy of aid. Democratic legislators fought the cuts, and radical priests led welfare recipients from Milwaukee in marches on Madison, leading to violent demonstrations at the capitol. A compromise somewhat reduced the cut, but the episode galvanized churches, advocates, unions, and liberal politicians to defend welfare. In the 1970 and 1972 elections, Democrats recruited candidates and campaigned on a theme of Republican heartlessness. The electorate responded. Democrats seized control of the Assembly and the governorship in 1970 and added the Senate four years later. For the first time in state history, they controlled the entire state government.14 Democrats were chiefly responsible for the liberal welfare regime that the state built up in the later 1970s and early 1980s. But controversy returned due to the welfare migration issue in the mid– 1980s. Although conservative Democrats first raised the alarm, Republicans soon joined them. The issue embarrassed the Earl administration and helped elect Tommy Thompson governor in 1986. Earl could never satisfy his liberal political base—lower-income voters and advocate groups committed to welfare—while still responding to an enraged public.15 Even afterward, Democrats remained vulnerable on the issue.

Politics of Reform

•

43

TABLE 3.3 Party Control of the Wisconsin State Legislature, 1969–2003 (numbers of seats held by parties at indicated times) Assembly Year January 1969 January 1971 January 1973 January 1975 January 1977 January 1979 January 1981 January 1983 January 1985 January 1987 January 1989 January 1991 January 1993 April 1993 January 1995 June 1996 January 1997 April 1998 January 1999 January 2001 January 2003

Senate

Democrats

Republicans

Democrats

Republicans

48 67 62 63 66 60 59 59 52 54 56 58 52 52 48 48 47 47 44 43 41

52 33 37 36 33 39 39 40 47 45 43 41 47 47 51 51 52 52 55 56 58

10 12 15 18 23 21 19 17 19 19 20 19 15 16 16 17 17 16 17 18 15

23 20 18 13 10 10 14 14 14 11 13 14 15 17 17 16 16 17 16 15 18

Sources: State of Wisconsin, 2001–2002 Blue Book (Madison: Wisconsin Legislature, Joint Committee on Legislative Organization, 2001), p. 272; “Wisconsin Legislative Spotlight,” Wisconsin legislature Web page, November 23, 2002. Note: Vacant seats are not shown. The Assembly has 99 seats (prior to 1973, 100 seats), the Senate 33. Elections of April 1993, June 1996, and April 1998 were to fill vacancies in the Senate. Representatives in the Assembly serve two-year terms, Senators four-year terms.

Tommy Thompson was first elected to the Assembly in 1966 and became Republican leader of the Assembly from 1981. He had opposed Democratic spending programs for two decades, earning the sobriquet “Dr. No.” During his campaign for the statehouse, he promised to cut welfare benefits so that Wisconsin would no longer be a “welfare magnet.” Then money could be saved and taxes cut. Then the state would attract jobs and investment rather than welfare recipients, the reverse of what it was doing. This was the traditional Republican program—to cut back government in favor of the private sector.16 Following Thompson’s victory, Democratic leaders such as Tom Loftus announced that they had “irreconcilable differences” with Thompson

44

•

Chapter 3

over welfare. The governor, for his part, gloried in the fact that welfare was a “fantastic campaign issue” that made his opponents feel as if “their hearts had been cut out.”17 He depicted leading Democrats as unrepentant liberals who initially allowed him only “limited experiments.”18 They at first resisted approving Learnfare, the flagship of Thompson’s first reform agenda, but concurred so as not to be painted as opponents of change.19 Later the legislature resisted expanding the program to schoolchildren younger than teenagers, approving this only in several counties in 1993. It rejected Two-Tier twice before approving it in 1993. Especially in his early years in office, Thompson depended on Wisconsin’s strong governorship to get his way. He possessed a partial veto power that allowed him to pick and choose among the provisions of bills submitted to him by the legislature. He used 290 partial vetoes to reshape the 1987 budget bill, the vehicle for most of his initial programs. Other vetoes resurrected Personal and Family Responsibility in 1991 when the legislature passed it only in a weakened form. These actions angered Democrats, who felt that the governor had abused his office. They contested the 1987 vetoes, but lacked the votes to override.20 They did, however, push through a constitutional amendment in 1990 that somewhat restricted the partial veto power.21 Thompson then kept the initiative by launching one reform experiment after another. Democrats continually had to react to his agenda rather than advancing their own. Key members of Thompson’s team were James Klauser, a longtime aide whom he made Secretary of Administration, and Gerald Whitburn, who was Secretary of the Department of Health and Social Services from 1991 to 1995. According to Whitburn, the earlier reforms, successful or not, created such “momentum” for change that Democrats had no choice but to accept W-2.22 One of the motives for Work Not Welfare was to block Democrats from recovering the initiative on the heels of President Clinton’s promise to “end welfare as we know it.” WNW was announced in 1993, before Clinton’s plan, and presented as a test of his idea.23 Democrats compromised with Thompson in self-protective ways, and this undercut their opposition. When first elected in 1986, Thompson persuaded Tim Cullen, the influential Democratic majority leader of the Senate, to become his Secretary of Health and Social Services, the department in charge of welfare reform. In addition to helping promote reform, this removed Cullen as a potential gubernatorial rival. It also made Tom Loftus, Speaker of the Assembly, the presumptive challenger to Thompson. This forced Loftus to trim his liberal sails in order to face a statewide electorate. In 1990, he did run against Thompson, and was defeated soundly.

Politics of Reform

•

45

By this time, many Democrats found it easier to support Thompson’s schemes than to oppose them. Representative Margaret Krusick, a Democrat from Milwaukee, remarked in 1991 that “If lawmakers are responsive to their constituents, they will vote for welfare reform,” for not to do so would be to “risk . . . losing their seats in the Legislature.”24 A shift on Learnfare expansion was apparent. In Thompson’s early years, Democratic leaders had opposed the program, but in the 1992 elections, almost a third of candidates supporting it were Democrats.25 The ultimate reason for Democratic reverses, perhaps, was the strong public resolve to see welfare changed. This made it impossible for Democrats to assemble a coalition against Thompson the way they had done against the Republican cuts in the early 1970s. The defense of welfare drew its strongest support from liberal legislators from the Madison area, and from liberal advocate groups there and in Milwaukee. Neither element could deliver many popular votes. As for the recipients and minority groups reliant on welfare, their voting turnout was too low to give them leverage. It might seem inevitable that the Republicans eventually regained a majority in the legislature, in 1993–95, and went on to enact the radical W-2, overthrowing the liberal welfare system. Thompson’s defeat of Earl in 1986 certainly accelerated change. So did the Republicans’ recovery of a legislative majority, as this allowed them to enact the most radical phase of reform. The immediate reason for the enactment of Self-Sufficiency First and Pay for Performance, in 1995, was that Republicans had swept the 1994 elections, earning control of both legislative houses as well as the governorship for the first time since 1971. Democrats were cowed, as well, by GOP victories in other Midwestern states and in Congress in the same elections.26 Press accounts tend to laud Thompson as the maestro of reform, as if he and the Republicans had done it alone.27 The governor’s leadership and tactics were undeniably effective, and Republican legislators displayed impressive discipline in supporting him for over a decade. Due to both his personal and institutional powers, Wisconsin’s governor was by 1994 one of the most potent in the country.28 Thompson’s departure for Washington in 2001 left a vacuum in Madison that the state GOP has struggled to fill.29 One result of partisan combat currently is a serious budget deficit, which the state has closed more with smoke and mirrors than serious economies.30 But the partisan account of reform cannot explain two central facts: Serious change began prior to the election of 1986, when Democrats controlled the entire state government. And even after Thompson’s emergence as governor, he was able to get his programs approved by a stillDemocratic legislature. Even after Republicans took control, the parties

46

•

Chapter 3

were so evenly divided that reform could not advance seriously other than on a bipartisan basis. Verbal dueling between the two sides concealed the real collaboration between them. The very Democratic leaders whom the governor said opposed him agreed to join a bipartisan Governor’s Commission on Welfare Reform that he set up in early 1987. Chaired by DHSS Secretary Tim Cullen and including all four legislative party leaders, this body approved most of the initial planks in Thompson’s program, although not the benefit cut.31 Many, even most Democratic legislators ended up voting for Thompson’s proposals. In 1993, a number of Democrats signed up as cosponsors of Work Not Welfare, which time-limited welfare, and the great majority of representatives and senators from both parties voted for the plan to abolish AFDC, which was linked to WNW.32 The character of reform also shows the thumbprints of both parties. Work and other obligations were increasingly demanded of the recipients, but through government rather than by simple cutbacks in aid. From an early point, Wisconsin was headed toward the conservative yet big-government solution that would finally be embodied in W-2.

THE BUREAUCRACY Several things about the policymaking in Madison took decisions out of the simple partisan framework suggested above. One of these was initiative by the bureaucracy. Tony Earl hesitated to reform welfare himself, but he also named strong executives to his administration. They aimed to solve the state’s problems, without being deterred by political fears. Linda Reivitz, his secretary of Health and Social Services, saw a need to address welfare well before it became an election issue. She declared in the midst of the 1986 election campaign that “The dysfunctions in the system, the barriers to employment, are insane.” She was referring to weak work incentives and the 100-hour rule for two-parent welfare, among other problems.33 An important figure under her was Peter Tropman, a former Democratic state representative from Milwaukee who headed the Division of Policy and Budget at DHSS from 1977 to 1987, under administrations of both parties. This background gave him an unusually broad perspective on the welfare problem. He had originally favored liberal welfare. As a politician, he joined in the Democratic coalition to establish that system in the early 1970s. But he gradually came to believe that the system treated welfare families too generously compared to the working poor. AFDC, he told me, was simply “out of sync with the experiences of most Americans. What adult people do is work.” By the 1980s, as national politics

Politics of Reform

•

47

turned conservative, he realized that Tony Earl was in “very big trouble” over welfare.34 So in 1985–86, he wrote several wide-ranging papers designed to promote new thinking. One of these, on the “welfare mess,” used government figures to describe the welfare dilemma. AFDC was growing, and welfare mothers were increasingly disadvantaged and long-term. Neither they nor their absent spouses usually took responsibility for their families. And welfare as it was could not fix itself. Tropman concluded, contrary to liberal orthodoxy, that “welfare is not a part of the solution, but part of the problem.” Instead, welfare recipients should be held to the “same expectations” as other people, should move toward “financial independence,” and should have to “work,” as this was the best route to independence. Welfare should become, not an entitlement, but a “transitional” supplement to earnings and child support. On the other hand, government should provide all necessary services, such as child care, to enable recipients to work. At much the same time, Tropman drafted a plan for the legislature for what become the WEJT program; it aimed to expand participation in work activities through a combination of new obligations laid on the recipients and new state spending.35 Although nominally a Democrat, Tropman’s greatest loyalty was to the regime. His mission was to show that government could “solve problems,” even entrenched poverty. In another paper, he addressed the crisis of credibility that had hit government nationwide. The Reagan Administration symbolized a widespread sense that government was wasteful, ineffective, even detrimental to American values. But rather than react defensively, Tropman said, DHSS should improve its programs. If it realized savings from reduced AFDC, it should “reinvest” them in “economic development, day care and benefit increases.” Likewise, savings from better child support enforcement should go into “general relief work programs, or education and training programs.” Thus, social programs would not be eliminated but rather shift their focus from the nonworking to a working citizenry. Improving implementation also might entail the “privatization” of some functions through contracting, even if government remained in overall charge.36 All this, later reforms and W-2 would do. Tropman circulated these papers within the government and the university. He failed to persuade Tony Earl to run on his ideas in the 1986 campaign, but some of the same ideas were taken up by Tommy Thompson. After the change of administration in 1987, Tropman was kept on at DHSS for several months in order to help frame the detailed Thompson proposals.37 Career officials in DHSS also played an entrepreneurial role. Aides to Tropman, including Gail Propsom and Bob Wagner, helped him write his

48

•

Chapter 3

papers. Together, Wagner and Tropman wrote a paper proposing experimental programs in high-welfare counties to get the caseload down.38 These and other analysts from DHSS and the Department of Administration helped translate Tommy Thompson’s ideas into operating programs. They also staffed the Governor’s Commission on Welfare Reform that ratified most of Thompson’s early proposals.39 Thus, the plans Thompson sent to the legislature were, in part, the work of a bureaucracy that Democrats had helped to build.

REPUBLICAN MODERATION More important for promoting consensus, however, was the surprising moderation of the parties in the legislature. The traditional conservative stance toward welfare had been either to oppose aid to people of working age on principle, or to reject the extra spending and bureaucracy needed to reform welfare. But in the 1980s, Wisconsin conservatives in both parties tended to attack the abuses in welfare rather than the institution or its cost. Joseph Andrea tore the skin off the migration issue, but he told me that he wanted to “provide child care, provide a job, to get these people jump-started”; he never questioned the need for welfare in some form.40 To liberal defenders of aid, such as Tony Earl, that sort of concern, despite the passion it aroused, sounded less “ideological” and more “practical” than earlier conservative attacks on assistance.41 Reform largely side-stepped the sensitive family issues at the heart of welfare. Despite Bridefare and the Benefit Cap experiment, Thompson’s proposals paid only limited attention to growing unwed pregnancy, the chief force, besides nonwork, that made families poor. He made no serious attempt to stigmatize illegitimacy or to promote marriage, as Republicans in Washington were to do in the 1990s. This accorded with public opinion. As Table 3.1 shows, the least popular reform ideas in Wisconsin were those that involved penalizing unwed pregnancy or promoting marriage. Republicans also were sobered by their earlier defeats over welfare. In reforming the system, they wanted to appear “less punitive,” Tropman said, lest they ignite another advocate/liberal crusade.42 Tommy Thompson, despite his partisan past, turned moderate once he entered the statehouse. He did not seek to cut back public commitment to welfare in any basic way, the goal that Republicans in Congress attempted with PRWORA. He did not reform welfare mainly to save money. That again followed public opinion, which wanted restructuring rather than economy to be the chief goal of reform.43 Thompson did want to cut taxes and improve the business climate, but he also expected that welfare reform

Politics of Reform

•

49

would cost new money in the short-term.44 Learnfare raised costs by $9 million and the buildup of work programs by $30 million.45 The state did eventually save large sums on reform, but this was a windfall, due to the unexpected caseload decline. Thompson’s initial proposal to cut AFDC benefits might sound like a traditional conservative gambit to reduce spending and taxes. But he justified it as a way of funding bigger work programs. The point was not to cut back welfare as such but to shift resources from income maintenance toward employment. “Cut, but invest,” Thompson said, much like Tropman. Most Democrats could justify that to their consciences and constituents.46 The administration was to make no further changes in AFDC benefits for ten years. This showed its “commitment,” Whitburn said, to the humane purposes of reform.47 And in 1997 Thompson actually raised stipends paid to cash recipients on W-2 by 21 percent, a step upsetting to some conservatives.48 As table 2.1 shows, this again gave Wisconsin one of the most generous benefit levels in the country. Furthermore, Thompson spent lavishly on the child care, health care, and other services that Democrats demanded as the price of reform. Very seldom during the reform process did Republicans voice public concern about costs.49 Instead, it was Democrats who used claims of waste to criticize Thompson’s initiatives, especially Learnfare.50 The governor also compromised with Democrats on operational matters. He postponed the inauguration of Learnfare from 1987 to 1988 because of implementation problems in Milwaukee.51 Most of his welfare experiments were piloted outside the city, because of its impacted caseload. The assault on Milwaukee welfare, the final redoubt of the old system, occurred only toward the end of the reform process. All told, the Thompson program might appear conservative. But perceptive Democrats noted how distant it was from past Republican positions. Tony Earl remarked ironically to me that “The kind of proposal, W-2, that Thompson has now would have been rejected out of hand as too liberal and too expensive by the Republican legislative leaders of the late 1960s. He’s gone through a real transformation.” Thompson himself admitted that he had learned a lot, while John Norquist, the Democratic mayor of Milwaukee, credited him with “perhaps the best welfare reform program in the country.”52 DEMOCRATIC CRITICISM OF WELFARE Even more critically, most Democrats were not unwilling welfare reformers. Conservative Democrats like Joe Andrea were the first to attack welfare excesses. They came chiefly from the urban, southeastern corner of

50

•

Chapter 3

the state—Kenosha, Racine, and Milwaukee. These were the counties most exposed to migration from Illinois, and officials there were upset about it.53 Liberal Democrats came mostly from Dane County, around Madison, the state capital, an area hospitable to feminism, environmentalism, and other liberal causes. Lee Dreyfus, the Republican governor prior to Earl, once dubbed Madison “a city in Wisconsin surrounded by reality.”54 Andrea “made a career” of fighting the “Dane County liberals,” he proudly told me, leading to “slam-bang debates” in the Democratic legislative caucus.55 But even liberal Democrats questioned welfare in terms seldom heard outside Wisconsin. Tom Loftus was one of the Democrats’ most forceful leaders during an Assembly career that lasted from 1977 to 1991. He was Speaker of the Assembly from 1982 through Thompson’s early years as governor, and ran against him for governor in 1990. He defended generous welfare, but at the same time, he called AFDC “a hopeless program” that was “too broke to be fixed.”56 He opposed Thompson’s proposals to cut benefits, but he also demanded bluntly that liberal critics of Thompson explain how they would solve the personal and family problems of the poor.57 John Norquist was another liberal who, as an assistant majority leader in the Senate, reacted favorably to Thompson’s initial proposals. Later, as a highly visible mayor of Milwaukee, he ridiculed traditional welfare as “a failure” that “ought to be abolished.”58 His main objection to Thompson’s programs was not that they were too tough on recipients but that they offered the needy any cash aid at all. He would have preferred a system that offered only jobs.59 Pressed by a liberal journalist about what assistance he would give to an indigent teenage mother, Norquist retorted: Nothing. I wouldn’t give her any cash grant. She should stay home and not have the baby. Look what happened in the past when someone had an illegitimate baby. But now we are bribing young girls to have these babies. Individuals may indeed continue to act irresponsibly, but we will not design a system around them. I know it sounds hard, but there can be no cash benefits for underage mothers.

As for an adult single mother who will not or cannot work: She will be offered work. There will be day care and health care. Maybe not the greatest jobs in the world, but there are jobs, even if it’s emptying bedpans. But, I say, if she doesn’t want to work, if she chooses to separate herself from the system, that’s OK too. But not with taxpayer money. If you want to be homeless, OK, but not on government time. Yes, at some point we say, “Tough luck, buddy, you’re on your own.”60

Politics of Reform

•

51

The remarkable thing about these comments is their anger and authenticity. Norquist utterly refused to condescend, as most Democratic politicians do, to the perceived vulnerability of the recipients. He expected certain good behaviors from everyone, however disadvantaged. He reacted to irresponsible welfare adults as he would to a misbehaving member of his own family. This earnestness and innocence, we will see, reflect Wisconsin culture. Many other Democratic politicians and officials spoke out against welfare on their own, without waiting for their leaders and not simply to blunt Republican sallies. During Thompson’s early years as governor, Milwaukee struggled with growing General Relief. County executives, welfare administrators, and city council members from Milwaukee— most of them Democrats—complained of the expense and abuses of the program. They demanded that Madison take over more of the costs and allow more stringent work and residency requirements.61 Such complaints dissociated their party from any simple defense of welfare. The Work Ideal As these reactions suggest, Democrats regarded welfare as a failure because it looked like a bad deal compared to employment. Why not work, they asked? Employment, not single-parenthood, was the route toward inclusion in the society. In this they differed from liberals in other states and most intellectuals, who often view work as a threat to poor families. Yes, work is good, they say, but only if the effort is voluntary and cushioned with elaborate child care and other support services. Wisconsin liberals simply had higher expectations for the capability of the poor. Such sentiments drove Democrats to start reforming welfare well before Tommy Thompson became governor, when they still dominated Madison. Loftus immersed himself in the subject as soon as he entered the Assembly in 1977. He was an eager member of the 1979 Welfare Reform Advisory Committee, fastened on child support as the best alternative to AFDC, and became the chief sponsor of Irwin Garfinkel’s child support reforms in the early 1980s. When he ran against Thompson in 1990, he proposed a guaranteed form of child support as a substitute for AFDC.62 Already during the crisis of 1985–86, when the migration issue flared and the benefit increase was trimmed, the Democratic legislature set up two special committees to study the welfare problem.63 Already, both Earl and Loftus had the idea that a freeze and even a reduction in benefits could be justified in order to expand work program.64 Out of the ferment emerged WEJT, the real beginning of serious welfare reform in Wisconsin. Due to these precedents, when Thompson advanced his proposals, the Democratic leadership did not reject all change. Not only was Loftus an

52

•

Chapter 3

innovator, but, as mentioned above, his counterpart as floor leader in the Senate, Tim Cullen, became Thompson’s chief reform advocate and administrator. Cullen was vital to selling the Thompson program to his Democratic colleagues.65 His successor as Senate majority leader, Joe Strohl, invented the idea of “two-tier”—paying incoming recipients only the benefit they would have received in their previous state.66 These were not leaders afraid to grasp the political nettles of reform. Nor did Democratic innovation cease with Thompson’s election. In 1991, in response to the governor’s initiatives and recent increases in the rolls, Walter Kunicki, the Speaker of the Assembly, set up a Special Committee on Welfare Review to suggest further changes. It advised a crackdown on welfare fraud in Milwaukee and recommended milder versions of the differential benefits and restrictions on general assistance than Thompson had proposed.67 After the decision in 1993 to replace AFDC, Democrats set up yet another legislative panel to decide how to do that.68 Riemer and Norquist A key figure in shaping Democratic thinking was David Riemer, a lawyer with a long interest in social policy who served as Milwaukee Mayor Norquist’s Director of Administration. From that post, he promoted the idea of employment as a solution to poverty. He wrote a book arguing that the main cause of dependency lay, not in welfare itself, but in a labor market that did not pay the low-skilled enough to live on. The solution was to create jobs through government as necessary, and also to subsidize wages.69 Work requirements are generally seen as a conservative policy, but Riemer is one of several theorists on the left who want to enforce work for progressive ends. Traditional welfare leads only to poverty, segregation, and conservative politics, they argue. The way to reform welfare was, as Norquist suggested, to guarantee recipients jobs rather than income. Only this can realize equality and a common society.70 The most direct test of a job guarantee would come in the New Hope project, described in chapter 8, of which Riemer was the chief inspiration. As early as 1987, Riemer proposed to Tropman that AFDC be replaced with a job creation program.71 The idea struck a chord with Norquist and other Democratic politicians. Norquist endorsed replacing aid with government jobs starting in 1990, and he and Riemer advocated the idea in many speeches and hearings. On several occasions in 1992–93, Shirley Krug and Barbara Notestein, both Democratic representatives from Milwaukee, worked with Norquist and Riemer to draft concrete proposals for job-based aid. These plans, known variously as “Wisconsin Works” or the “Wisconsin Jobs Connection,” became competitors to Thompson’s

Politics of Reform

•

53

Work Not Welfare.72 The programs guaranteed jobs, but much as in W-2 the positions were time-limited, and cash aid was denied to “parents who refuse to work.”73 In 1994, the idea of replacing cash with jobs was again pressed by Chuck Chvala, another unsuccessful Democratic challenger to Thompson’s reelection.74 The most radical step Democrats took was to propose the total abolition of AFDC, Food Stamps, and General Relief in 1993. An amendment attached to the authorization for Work Not Welfare mandated that these programs expire at the end of 1998. The system replacing them was supposed to support the unemployable, guarantee jobs to the employable, and assure affordable child and health care. Democrats were driven to this step partly by their frustration at always reacting to Thompson. To propose abolition put him and his advisors on the defensive, for a moment. The leader of the move was Antonio Riley, a black Democratic representative from Milwaukee. He had been pushed into action on welfare by his constituents, hit on the idea of abolition, and sold it to his party, with the support of Norquist and Riemer.75 Convinced that abolition was an impractical charade, most of Thompson’s advisors told him to veto the provision. But, after confining the abolition to AFDC and eliminating the Democratic conditions, he accepted it, and it became the mandate for W-2. What is remarkable here is not only that Democrats turned against a signature program but that they trusted their opponents to collaborate in redesigning welfare. Mutual distrust helps to explain the bitterness over welfare in Washington and many urban states. Democrats may have assumed that they would have enough power in Madison to control the redesign. But by October 1993, when abolition passed, they had already lost control of the governorship and the Senate to the Republicans, and in the 1994 elections they would lose the Assembly as well. They apparently assumed that, whoever was in control, they would be able to work with Thompson and his party. That was partly because of the new moderation shown by the GOP, but also because their own thinking about reform had become so radical.

A MORALISTIC STYLE The civic-mindedness of Wisconsin’s politicians reflects the political habits of the state. In a famous analysis, Daniel Elazar identified three strands in American public culture. Where “individualism” dominates, for instance in the mid-Atlantic region and many urban states, politics is seen mainly as a way of distributing material advantages to individuals and interests. This approach tolerates a politics oriented to economic subsidy

54

•

Chapter 3

and payoffs to parties and their supporters. In “traditionalistic” states, chiefly in the South, government is given only a limited role, oriented to the maintenance of traditional values and racial hierarchy. Wisconsin represents the “moralistic” culture that dominates upper New England, the upper Midwest, and parts of the West. There government has a stronger orientation to the public interest. Politicians are expected to pitch their proposals as serving the society as a whole. Benefits for narrower interests are suspect. Government is also seen more ambitiously, as a positive instrument for achieving the public good.76 This civic and problem-solving ethos is what Wisconsin policymakers displayed over welfare reform. One interpretation of the Wisconsin consensus might be that policymakers simply disagree less than they do in other states. Welfare policymaking was in this sense bipartisan. That might reflect public opinion. Polls show that Wisconsin is above average among the states in the liberalism of both opinion and policy, but it is not conspicuously liberal or Democratic, nor is opinion strongly polarized.77 The state’s dutiful social ethos, as I show below, discourages flamboyant politics and urges the parties to work together. The attitude is “live and let live,” Tony Earl told me, not “kill or be killed.” As another leader remarked, “We do everything in Wisconsin by bipartisan committee.”78 At the same time, however, party elites in Wisconsin are more liberal than the national norm, and they appear to be among the most polarized in the country.79 The parties are also evenly matched and competitive as organizations. During 1989–94—the heart of the reform period—the parties in Wisconsin (and also Pennsylvania) shared power more evenly than those in any other states.80 Legislative control was so finely balanced that each party had an incentive to seize tactical advantages, and did so. The partisan edge swayed back and forth over the course of the reform process.81 Problem Solving The ground of the consensus lay not in any special propensity to agree but in the subtler fact that the welfare debate in Wisconsin was so directly and sincerely about policy. The desire to protect the status quo or one’s immediate political advantage was outweighed by the drive to find a defensible solution to an agreed-upon problem. In the 1970s, that problem had been seen as simply supporting the poor generously. By the 1980s, however, the problem had shifted to reducing dependency, yet without eliminating generous aid. That dual focus, over time, drove both parties toward work enforcement, as only this policy could square that circle. In short, a process of learning within government drove most participants

Politics of Reform

•

55

toward a work-based reform that was both thoroughgoing and largely agreed upon.82 Debate about policy suffused even the bureaucracy. In 1986, Gerald Berge, who managed income maintenance within DHSS, wrote an elaborate rebuttal to Tropman’s paper on the “welfare mess.” He argued that welfare was not clearly responsible for other social problems, as Tropman claimed. Instead, the blame lay with social inequality outside welfare. Rather than cut benefits, the government should build up child support as an alternative to welfare. But Berge shared with Tropman a belief in sincere policy argument. His statement, studded with statistics and references to outside authors, is the work of an expert and a true believer. It alludes only in passing to the political interest that the department or the Earl administration might have in fending off criticism.83 In Wisconsin, the parties orient more to issues than organization. They are coalitions of able self-starters who get into office on their own, chiefly because they are interested in policy problems. The pattern is the opposite of machine politics, where politicians are motivated mostly by office and its material perquisites. In Wisconsin, issues are taken seriously in their own right, not only as means to power.84 Tom Loftus, whom one associate called a “policy wonk,” announced that he was “after the next generation, not the next election.”85 Both parties engaged in serious inquiry, and did not impose preformed solutions. Trust The key to agreement was the fact that neither party was so cynical that protective reflexes overrode common solutions. In less well-governed states, and in Washington, Democrats may have their own misgivings about welfare, but they fear to broach serious change lest they unleash a right-wing backlash against the poor. They also shelter the many interests who gain something from traditional welfare—not only the recipients but the program staffs and community groups whose livelihood is serving the dependent. On the other side, Republicans seek mainly to cut welfare, not rebuild it, because they fear new costs and they distrust the intentions, or the competence, of the bureaucracy. In Wisconsin, however, the combatants focused on the problem more than on the political stakes. The Democratic legislators I interviewed mostly spoke with rueful candor about the need to reckon with serious dependency. They were not in denial. For example, Dismas Becker, a former Democratic majority leader in the Assembly, had been among the radical priests who led the demonstrations to defend welfare in 1969. Later he went into politics. But by the 1980s he found that the complaints

56

•

Chapter 3

about welfare migration were “legitimate.” It was “inevitable” that Democrats would respond.86 Barbara Notestein, another example, had come to the Assembly out of a career in welfare rights and community organizing. She sat for a heavily Democratic district in Milwaukee. In another state, she would have defended welfare. Yet she was able to hear the complaints from constituents about abuses in AFDC, and she concluded that “something had to be done.” She agreed with the Republicans that the best answer seemed to be to “move more people toward work.” Hence her own proposals to replace AFDC with guaranteed jobs. She thought W-2 was too severe, but recognized that it had substantial support in both parties. The new system might be improved, but entitlement was dead.87 The Role of Ideas A further feature of the moralistic style was a heavy emphasis on ideas. Among conservative intellectuals, the dominant approach to welfare reform probably is that of Charles Murray—simply to cut back welfare so that the poor are forced to function for lack of alternatives. They must work and marry or starve.88 That view has strongly influenced Republican reform positions in Washington. But in Wisconsin, leaders were drawn more to my own idea of “reciprocity” or “social contract.” In this view, the problem with welfare was not that it was generous but that it was permissive, failing to enforce key civilities, such as work, in return for aid. To require work also would address the leading economic problem of the poor, which was lack of steady employment.89 I held a visiting position at the University of Wisconsin in Madison in early 1987, just as Thompson launched his first reforms. I advocated work enforcement as a basis for reform in several meetings with legislators and administration officials, and in public talks.90 To my surprise, these encounters had influence.91 Homegrown thinkers like Tropman and Riemer, as well as many politicians, had already hit on reciprocity as a rubric for change. However, my ideas helped them tie what they were doing to poverty research and the national welfare debate.92 Policymakers had closer ties to the intellectual world than is usual in most states. In the framing of W-2, Gerald Whitburn remarked, planners studied not only conservative experts like myself and Charles Murray but the work of liberals such as David Ellwood, Christopher Jencks, Mickey Kaus, and William Julius Wilson.93 Officials in the bureaucracy were knowledgeable about academic research on welfare and poverty. They seemed to have one foot in academia. Indeed, some began their careers as administrators and then moved to research appointments at the University of Wisconsin—Madison.94

Politics of Reform

•

57

In the moralistic style, legislators worry more about satisfying their own ideas of the public interest than catering to the voters. At the same time, they remain clearly accountable to the people. Across the states, when the leaderships are polarized, as they are in Wisconsin, even small changes in the partisan balance can produce large shifts in policy. The public tends to react to excessive activism by government by electing more Republicans to the legislature.95 This is exactly the process we see at work in Wisconsin. In the 1970s, a popular shift toward the Democrats established the liberal welfare system. In the 1980s and 1990s, a shift the other way enfranchised the radical changes leading to W-2. Thus the public chastised leaders of each extreme, until they worked together to produce the welfare the people wanted. The whole episode is a triumph of policymaking and self-government.

OUTSIDE GROUPS In most urban states, several groups outside of government commonly deter or discourage welfare reform. Most minority leaders, advocate groups, and intellectuals are hostile to conservative approaches to reform. They cannot deliver many votes, but they often have enough moral and intellectual authority to intimidate politicians who might want to question entitlement. It is difficult to contemplate making recipients work or stay in school in the face of community leaders or commentators who call this immoral or ignorant, and who have the ear of the media. An important group right of center, on the other hand, is business. In many settings, corporate leaders resist the extra spending and bureaucracy that it takes to reform welfare. In Wisconsin, however, the liberal groups that might have opposed reform were quiescent, ambivalent, or ineffective, while business was surprisingly supportive. Blacks Black leaders are often presumed to oppose conservative reform, because so many blacks rely on aid. In the Wisconsin legislature, Thompson was challenged by Senators Gwendolynne Moore and Gary George, two black Democrats representing poor areas of Milwaukee. Moore characterized Thompson as a “successful demagogue” who “beat up on people of color” and whose reform amounted to a “new form of slavery.” She accepted a work requirement for recipients with school-age children, but she thought overcoming poverty required a whole series of new programs, and she was one of only four senators to oppose the abolition of AFDC.96

58

•

Chapter 3

George attacked Learnfare as “no different from South Africa.” To him, two-tiered benefits implied that “If you’re black, you don’t get fair treatment in America.” Welfare reform, he told me, occurred to “satisfy the blood-lust” of taxpayers.97 However, while Moore and George were respected, few other blacks or Democrats followed them. A more influential black was Antonio Riley, the representative who in 1993 led the abolition of AFDC. He sat for the Milwaukee district that Moore had occupied before she moved up to the Senate. He estimated that two-thirds of the families in his district drew assistance of one kind or another; indeed, he himself had once been on General Relief. Nonetheless, he declared that “Welfare is the jailer of our people.” It caused blacks to be “written off” by the rest of society. Like many white liberals, Riley assumed that work was better than welfare, even for needy one-parent families. Like Riemer, he wanted to redesign welfare around employment and to “make work pay,” if necessary by using government jobs. He became impatient with Republican “tinkering” and advocated a clean break with the past.98 More blacks followed him than Moore or George.99 Riley’s success reflects the fact that blacks in Wisconsin shared the same strong opposition to traditional welfare as other groups. Table 3.1, which shows support for various welfare reform measures, also breaks out white and black support for many of these ideas. On most of the items, black support was very similar to white. The only substantial difference occurred over the idea of denying benefits entirely to unwed young mothers, but here only a minority of either race was in favor. Black support for W-2 and PRWORA is almost as strong as white. Of course, blacks largely agree with whites about what to do about welfare everywhere in the country. What is special about Wisconsin is that a radical reform was not stymied by racial sensitivities among elites. In Milwaukee, aggressive criticism of welfare was acceptable, even in public, even from blacks. In a poll in 1992, a majority of residents, including blacks, supported abolition of the old system.100 Will Martin is a former aide to Tommy Thompson who became director of Employment Solutions, one of the private agencies that first ran W-2 in Milwaukee. Like Antonio Riley, he is an ambitious black professional with conspicuous ability and conservative values. He told me that under the old system, “We’ve enabled people to be dysfunctional.” “We’ve held people to a lesser standard.” So the “moral premise” of reform had to be that “it is simply unconscionable to leave somebody on welfare.”101 It is the same angry but innocent tone heard from John Norquist and other white politicians. Due to the weakness of black opposition, advocates defending entitlement on racial grounds did not become the voice of Milwaukee in the

Politics of Reform

•

59

welfare debate, as they have of other big cities. Instead, the city was spoken for mainly by its elected leaders, most of whom, as mentioned earlier, were moderate Democrats aggrieved by the old system. Blacks in Wisconsin simply did not force the rest of the polity to treat welfare as a racial issue. Community Groups Equally surprising, community organizations opposed to conservative reform had little clout. They had been very influential during the liberal era of welfare in the 1960s and 1970s. They spearheaded the resistance to the Republican cuts in 1969. In the 1980s and 1990s, pro-welfare groups were still active in Milwaukee and Madison. Everyone I spoke to in Wisconsin mentioned the Welfare Warriors. They took the classic welfare rights position in defense of entitlement. A single mother, in their view, discharged the obligation of the departed father in raising his child. Thus, “Every mother is a working mother,” deserving of public support, whether or not she works for pay.102 Another such group, called ACORN (Association of Community Organizations for Reform Now), staged demonstrations at welfare centers. But compared to the 1960s, advocacy groups drew little attention. Politicians heard them, but did not obey. What undercut the opposition was that few advocates would now defend entitlement. In the changed setting of the 1980s, even liberal groups realized that it was impolitic to resist work tests when the support for them was so strong. Further, most also thought that it was fair, at least in principle, that adult recipients should have to work in return for aid. Church groups found the issue of entitlement too sensitive to take a clear position on it. Thus, the Welfare Warriors’ demand for entitlement was rejected, even by other liberal organizations.103 Some of my respondents intimated that the Warriors may have damaged their cause more than they helped it. Most groups confined their opposition to details. Typically, they accepted the work demand in principle but wanted the terms of work—pay, support services, chances for better jobs, and so on—to be more generous than the Republicans did. As one example, Pat DeLessio, a prominent poverty lawyer, told me that “I generally support a principle that requires people to work,” provided that “it offers them a way to get training and get out of poverty.”104 Advocates also criticized the troubled implementation of reform in Milwaukee, which led to the improper sanctioning of many clients.105 The Thompson administration reached out to the groups sufficiently to make them feel heard. The Welfare Warriors were left isolated.106

60

•

Chapter 3

Academics A final source of outside opposition was liberal intellectuals, of whom the most important were poverty experts at the universities. Their views should have carried special weight in Wisconsin, where there is a tradition of the university advising the government. Expertise about poverty is especially strong at the University of Wisconsin, home of the Institute for Research on Poverty. IRP’s experts had enjoyed ready access to DHSS under governors of both parties. For two decades prior to Thompson, they had schooled the bureaucracy in an entitlement approach to welfare.107 IRP experts were not indifferent to the lifestyle issues in welfare. They had masterminded Wisconsin’s pathbreaking child support reforms in the 1980s. They also wanted to raise work levels among the recipients. However, they preferred to use voluntary measures such as improved work incentives, training, and child care. Most opposed enforcing work as a condition for aid.108 Under Thompson, IRP lost its entre´e to state government. Instead, the administration turned to other experts who were readier to discuss the functioning problems on welfare—myself, Charles Murray, and also David Ellwood and Mickey Kaus. As we will see, W-2 was designed by a task force that drew on various experts in and outside the state.109 In opposition to Thompson, IRP issued studies casting doubt on the feasibility of work tests.110 Another academic poverty center, the Employment and Training Institute (ETI) at the University of Wisconsin—Milwaukee, performed hostile evaluations of some of Thompson’s early programs.111 But as long as the caseload was falling—which it did almost continually from 1986—and as long as little hardship resulted, the criticism carried little weight. Business In most parts of the country, business has played little direct role in welfare reform. Employers are of course interested in hiring the new workers made available by reform, but businessmen have seldom led social policy development. That partly reflects the cultural gulf that separates social planners, including academics, from the corporate sector. It also reflects businessmen’s frequent disdain for government. Especially in urban states with expensive regimes, business tends to view the state bureaucracy as a rival and a parasite. To them, it seeks only to tax and regulate. It seems to serve special interests, such as organized labor, trades and professions licensed by government, and public employees themselves. These fears run strong in urban states with Elazar’s individualist culture, such as Mas-

Politics of Reform

•

61

sachusetts, New York, or Pennsylvania. In these settings, corporate leaders are suspicious of new social initiatives, even to reform welfare.112 In Wisconsin government, however, there is relatively little patronage or waste to oppose. Accordingly, businessmen adopt a more trusting style. They enter into the debate about the public interest that is at the heart of moralistic politics. While they generally prefer smaller government, they do so with less antigovernment animus than elsewhere. This perspective became the main platform of the postwar Republican party.113 So business did not oppose Thompson’s new programs as wasteful. It was open to the argument that welfare could be recast, as against simply cut. Business even helped develop new programs. In Milwaukee, as described in chapter 8, a local business group sponsored the New Hope Project, an ambitious work guarantee program with a liberal pedigree. In Madison, the local Chamber of Commerce took over several housing projects troubled by social problems. Using business and public funds, it renovated the apartments, lowered rents, and installed new management. It hired a staff person to help residents with employment problems. Crime fell, work levels rose, and morale improved. The project’s mission—“to do whatever is necessary to better the community”—is as broad as government’s.114 These episodes give a sense that the entire state participated in the struggle to reform welfare, not only the government. The players did not agree about everything, but they did agree that reform was critically important. They pursued a quest to replace traditional welfare, and it led toward work-based aid.

CONCLUSION Chapter 2 argued that the Wisconsin reform constituted effective policymaking given what leaders knew at the time. We now see one of the reasons for that success: a high-minded political process. By focusing sincerely on an agreed problem, politicians achieved close to the best reform they might have wished. They were not forced by compromise or stalemate to achieve much less. The key to Wisconsin’s achievement was that each party reassured the other about its innermost concern. Democrats abandoned entitlement, because aid would now be conditioned on work. But Republicans abandoned the ambition to use reform to downsize government. One side agreed to enforce work, while the other agreed to do it through government, rather than by denying aid and letting the marketplace do the enforcing. Benefits to the poor were actually increased. The result was a

62

•

Chapter 3

new welfare state of almost European generosity—but stiffly premised on employment. I do not mean that every aspect of welfare politicking in the state was high-minded; on issues other than welfare reform, Wisconsin politicians may well have been less so. And maneuvering for personal and short-term advantage did go on around the problem-solving. The state has recently been irresolute in addressing its budget deficit, in part because Scott McCallum, who succeeded Tommy Thompson as governor, was less masterful. Embarrassing episodes of political corruption—anathema to the moralistic style—surfaced in Milwaukee and Madison.115 Wisconsin is high-minded relative to most other states. In Washington, welfare is still such a partisan issue that, were Democrats to recover control of Capitol Hill, parts of PRWORA might well be undone. In Wisconsin, partisan agreement ran far deeper. Even a Democratic return to dominance in Madison would not shake the basics of the state’s new welfare regime. The early appearance of that consensus is the first reason why the state led the nation toward a new welfare order. It is connected, as well, to the second: the state’s strong administrators. The bureaucracy implemented the reform brilliantly, in part, because it received such clear directions from politicians about the direction of change. With the agenda clear, officials knew the problem they had to solve.

CHAPTER 4

Implementing Work Requirements

MOST DISCUSSIONS OF welfare reform concentrate on its inflammatory politics.1 That is where the drama seems to be. There was drama enough in the politics of the Wisconsin reform. However, implementing reform has been as difficult for government as legislating it. Political leaders love to announce showy new welfare programs, as these always grab public attention. Under TANF, state and local leaders say they are implementing work requirements vigorously. But unless the new demands can be executed down to the operating level, they remain symbolic.2 The implementation of work tests is demanding. Recipients must receive support while being referred to employment programs, where they must receive child care and other services in order to participate. Enough recipients must participate so that work becomes, not exceptional, but a routine part of welfare. Then the employable will seek work on their own, allowing welfare to concentrate on serving the most disadvantaged. These demands have often proven insuperable in the past. Meeting them is the chief challenge welfare managers face. I focus here on how Wisconsin built up work requirements in AFDC up through the mid–1990s. Though little noticed at the time, that process laid the groundwork for the radical Wisconsin Works. Indeed, due to these changes, most of the AFDC caseload had already confronted a serious work demand and left the rolls before W-2 was implemented in 1997. Without that, the new system might have been unmanageable, or unaffordable. Much of this early development occurred in the late 1980s and early 1990s. That may seem long ago, and yet the concrete things Wisconsin did then to make the work test real are what many other states—and some European countries—are struggling to do today. In the bureaucratic statecraft of welfare reform, as in the politics, Wisconsin was a pioneer. Little prior research describes this process. There are studies of the decisions states make to institute work requirements, but without hard evidence of how programs actually changed “on the ground.”3 Some recent studies describe administrative changes designed to raise work levels at the service level, but only in the mid–1990s.4 There are evaluations of work programs that discuss their implementation to a limited extent.5 There are also studies of implementation in more depth, but at one point in time.6 No one, to my knowledge, has attempted to track the implementation of work requirements in the detail I do here, or over time. That

64

•

Chapter 4

probably is because, to do it, one would need the unusual access that I had in Wisconsin. My sources include program data, state documents explaining work policies, and interviews with Wisconsin officials in Madison and in ten counties between 1987 and 2000 (chiefly in 1994–97).

THE IMPLEMENTATION PROBLEM Implementation first appeared as a serious difficulty in American social programs in the 1970s. Many of the innovative education and employment programs invented by the Great Society failed to achieve their goals, and one reason was that they were not carried out as designed. At first, the chief difficulty appeared to be that federal authorities could not stop state or local governments from using federal money for unauthorized purposes.7 Later, a more top-down view developed that programs could achieve their goals if top-level actors put enough political will and money behind them, if program design was improved, and if lower-level agencies were competent and motivated.8 Both state and local initiative figured in the Wisconsin reform. In this chapter I describe how the state government pushed and pulled the Wisconsin counties to institute the first serious work tests in AFDC. Madison’s role in state governance is quite strong, despite the tendency of moralistic states to favor decentralized administration.9 Although Wisconsin’s welfare system is run by local agencies, the state government controls it by virtue of heavily funding local benefits and administrative costs.10 In chapter 5, I shift to a bottom-up perspective. Notwithstanding strong state leadership, the counties differed widely in how they approached welfare reform, and some of them became bellwethers for state policy. Below, I describe early welfare work programs and the early history of these efforts in Wisconsin. I go on to the buildup of funding and reporting systems that transformed the programs after the mid–1980s, then to the policy changes that gave them rising authority over the caseload. I also discuss how state officials attuned the bureaucracy to their goals. All these measures made the work test unavoidable for welfare adults in Wisconsin by the mid–1990s. That drove the caseload down well in advance of W-2.

EARLY WELFARE WORK PROGRAMS The oldest national work program is the Employment Service (ES), now known as the Job Service (JS). Established in 1933, the ES is a federally funded labor exchange that helps voluntary job seekers find employment. Although the ES might serve welfare recipients, Congress in 1967 added the Work Incentive (WIN) program, which aimed specifically to place

Implementing Work Requirements

•

65

employable welfare adults in work or training. The control of WIN was vested jointly in the ES, which did the job placement, and in state and local welfare departments, which provided child care and other support services. Wisconsin excelled in running these and other work programs because of its tradition of strong administration (see chapter 12). Research in the 1970s found that the ES in Wisconsin exemplified the most efficient form of ES organization. It had personnel who approached their task analytically and managers who delegated heavily to the staff serving the clients, while also holding them accountable for results, including job placements. For similar reasons, Wisconsin also had an outstanding WIN organization.11 Yet Wisconsin’s work programs failed in the sense that they did not restrain dependency. As shown in chapter 2, the welfare rolls rose inexorably from the 1970s through the mid–1980s. This was chiefly because the state steadily raised benefits. Secondarily, until the mid–1980s, work programs in the state lacked a mandate truly to enforce work. Dependency was not yet widely controversial. That changed, as we saw, in the mid– 1980s. There were bureaucratic problems as well. As in other states, the Employment Service in Wisconsin proved less capable of placing welfare recipients in jobs than welfare officials wanted. WIN could obligate its clients to participate as a condition of aid. The ES itself, however, did not enforce work. Its routines had been formed around helping job seekers who came to it on a walk-in basis. So it gave lower priority to serving the mandatory clients that WIN referred to it. It was uncomfortable with the paternalistic tasks necessary to serve welfare adults—not only enforcing work but providing special support services. WIN also faced problems getting the employment and welfare sides of the program—often located in different places as well as run by different agencies—to work together. In response to these dissatisfactions, Congress, in the Omnibus Budget Reconciliation Act (OBRA) of 1981, allowed states to replace WIN with a “WIN Demonstration Program” wholly controlled by the welfare department. This Wisconsin did in 1983, when it inaugurated the Wisconsin Employment Opportunity Program (WEOP). Control was vested in the Department of Health and Social Services (DHSS), the state welfare department. But since DHSS still contracted with the ES for job search services, little changed. In addition, WIN’s funding was reduced by Reagan-era cuts. In part, WEOP was a response to this, as it concentrated mainly on job search and dispensed with much of the costlier education and training that WIN had provided. Wisconsin augmented WEOP with other work-oriented activities that had been permitted by OBRA, such as employment search and the Community Work Experience Program, in which recipients

66

•

Chapter 4

“worked off” their grants in unpaid government jobs. However, due to the cuts, by the mid–1980s, WEOP’s staff fell from 300 to 100 and the program could serve only 22 of Wisconsin’s 72 counties.12

THE BUILDUP BEGINS As described earlier, when the state seriously awoke to its welfare problem in 1985–86, the first new program it enacted was the Work Experience and Job Training program. WEJT was supposed to improve on WEOP, by providing, along with CWEP, more education and training, and also by reaching more of the caseload. Control of it was firmly vested in DHSS, demoting the ES to a contractor’s role in some counties. After Tommy Thompson won the governorship, DHSS seized on WEJT to expand participation in welfare work. The program was instituted first in five counties in 1987, then by 1989 in thirty-three counties, including all the largest ones. By 1990, either WEJT or a combination of CWEP and job search was running in every country.13 WEJT gave the state a running start on implementing the Job Opportunities and Basic Skills Training Program, the next welfare work structure to come down from Washington. JOBS, which resembled WEJT, was established by the Family Support Act of 1988. Where WIN had obliged mothers to participate when their youngest child turned six, JOBS lowered that age to three, allowing it to involve more of the caseload. Like WEJT, the new program also provided more funding for training, education, and child care. Wisconsin was the first state to lodge its JOBS plan with Washington, and it was one of fourteen states to implement the program in July 1989, the earliest possible date.14 The plan was approved, the state noted, “Because we had in place the required aspects of the program in a majority of counties.”15 By July 1990, JOBS was operating throughout Wisconsin. WEJT and other earlier programs were folded into it. Now the state could make a work obligation real on the ground. As one state official remarked to me, “The JOBS program was everywhere, and no one was exempt because JOBS did not exist.” BUILDING THE BUREAUCRACY To fully implement the new programs, the state committed itself to an ambitious expansion of the welfare work bureaucracy. The goal was to change work from a paper requirement to something that was built into actual welfare operations.

Implementing Work Requirements

•

67

Funding In WIN and JOBS, federal funds matched specified shares of state spending. These monies were insufficient to implement WEJT and JOBS in the manner Wisconsin wished. So extra funds were provided. The first five counties to start WEJT received $1.8 million from Madison to serve 3,500 clients. By 1990, WEJT cost $36 million to serve 50,000 clients.16 This went well beyond previous WIN and WEOP funding. In JOBS, the state gave counties $1,200, later $1,400, per case to run the program annually. Each county also received a child care grant of about 17 percent of its JOBS grant. The state was prepared to expand WEJT from 1989 on even if it spent more money than Washington would match with federal funds. It also doled out extra funds to meet special needs—such as meeting federal standards for a 40 percent participant rate for two-parent families in AFDC, to raise participation rates in Milwaukee, or to ensure that counties’ JOBS funding did not drop just because caseloads were now falling.17 These increases occurred even though the caseload declined from 1986 onward, except for a brief increase during the recession of 1991–92, and one might have expected funding to fall as well. State appropriations for AFDC work programs rocketed from $1 million in the mid–1980s to $13 million in 1987–88 and $34 million in 1988–89.18 State effort clearly outpaced matching funds from Washington, which grew much more slowly. The state’s federal JOBS authorization grew only moderately, from $24.5 million in 1990 to $28.2 million in 1995, in constant 1995 dollars.19 But the state’s outlays for welfare employment and training soared from about $15 million in 1988 to over $80 million in 1995, counting both state and federal funds, also in 1995 dollars. The rise was particularly steep during 1988–90, during the launching of WEJT and JOBS.20 Some of the new money came from the extra federal funding that Wisconsin obtained through the “waiver savings” deal that Thompson made with Washington (see chapter 2). But since this money matched state spending, Wisconsin had to commit its own money up front to create its welfare work network. Separate figures on personnel are unavailable, but state officials told me that much of the new money went simply to hire more staff for welfare work programs so that they could involve the caseload more fully. By 1994, state and local JOBS workers reported in interviews that case managers in the program typically had caseloads of around 150 clients each, the range running from 75–80 in small counties to 300 in Milwaukee, the largest city. This is below the typical numbers found in other WIN and JOBS programs I have studied, especially in urban settings.

68

•

Chapter 4

Rising staffing interacted with falling caseloads to permit much closer supervision of the recipients. Both developments reduced the number of cases that a given staff member was responsible for. So staff could now serve a higher proportion of the remaining cases. They could help and motivate relatively more of them to get jobs or dispense with welfare. The more they did this, the more the rolls fell, and their grip tightened further. Thus, as one official put it, the growing bureaucracy “ratcheted” the caseload down. Many other states, where limited funding was a constraint on JOBS, were less fortunate.21 Management Information Systems Wisconsin also made critical investments in computerized management information systems (MIS). Rationalizing these systems had started in the 1970s, when the state combined the paperwork used to apply for AFDC, Food Stamps, and Medicaid. It computerized much of case management operations for these programs using a Computer Reporting Network. CRN, which streamlined data collection and case decisions across the state, was the most advanced such system in the country at the time.22 CRN had the limitation, however, that it covered only the eligibility and grant payment side of welfare. As employment became more important, the state built up reporting systems in that area. At the time of WEOP, local work offices monitored client assignments using a Job Service Information System that was based on the ES’s Employment Security Automated Reporting System. ESARS recorded how many clients were in certain statuses within offices, but it did not allow managers to track individual cases as CRN did. And as the welfare department took control of work programs in most localities, the ES’s systems were no longer available. In 1989–90, welfare implemented its own improved Work Programs Reporting System (WPRS), which could track individual cases, in order to run JOBS. Finally in 1994, the state implemented the Client Assistance for Reemployment and Economic Support system. CARES automates the input of client data, determination of eligibility, and the calculation of benefits. It encompasses both the claims payment and work sides of welfare. This fact reflects the welfare department’s dominance of work programs. CARES also records clients’ activities within work programs, although not the wages or other features of private jobs they enter. The system is complicated, its implementation was troubled, and it had to be modified for W-2 in 1997. But it was still an improvement and a better system than many states had.23

Implementing Work Requirements

•

69

Administrative Costs One apparent downside to the bureaucratic buildup was a dramatic rise in administrative costs. Overall in welfare, Wisconsin appears efficient. If one looks at Medicaid (the health program for the poor) and Food Stamps as well as AFDC, and if one considers all the fiscal years between 1989 and 1993, the state spent $148.96 in federal matching funds to administer each case annually, compared to a national average of $159.30. In 1993, administration ate up only 4.8 percent of federal program costs in Wisconsin, compared to 5.8 percent for the nation as a whole.24 But state and local costs specifically to run AFDC exploded. In 1986, prior to the WEJT buildup and the Thompson reforms, Wisconsin spent only $10.56 on administration per month per AFDC case, the lowest figure in the country. By 1994, however, that amount had soared to $61.55, only the thirty-seventh best among states. Wisconsin’s proportional increase since 1986—almost six times—was far and away the largest in the country.25 Much of this cost went for the new personnel and bureaucratic apparatus, including reporting systems, needed to run Wisconsin’s elaborate welfare reform programs. Thus, from the time of WEJT, Wisconsin invested millions in stronger government to get control of welfare. In return for these expenses, however, the state promoted an equally dramatic drop in welfare dependency. As I detail in chapter 10, the resulting savings in grant payments and other benefits more than offset the new spending on bureaucracy and services. These gains were already apparent by the early 1990s. At least in financial terms, the gamble paid off.

RAISING PARTICIPATION Armed with these bureaucratic resources, the state could now change the design of work programs so that they promoted work more effectively. As mentioned in chapter 2, the Wisconsin reform began at a time when the best design of welfare work programs was disputed. But by the late 1980s, it was reasonably clear that the best programs were mandatory and enforced participation stringently. Wisconsin moved in that direction with rule changes that obligated more recipients to participate in WEJT or JOBS, and for more hours. Normal federal rules would have made this difficult. Under WIN, welfare mothers could not be obligated to work until their youngest child was six, exempting most of the caseload. FSA lowered this age to three, but even then only 41–44 percent of welfare adults were mandatory for

70

•

Chapter 4

the program.26 But one of the first waivers Tommy Thompson obtained allowed the state to loosen these rules. It could obligate mothers to participate when the youngest child was as young as one. Under this authority, the state in 1988 set the age at two, then in 1995 lowered it to one. A large majority of mothers were now mandatory for work. The waiver also allowed the states to require twenty hours of participation a week, or half-time. After FSA, which normally required only twenty hours, Wisconsin could demand forty hours. The state began to demand thirty-five hours of effort a week in Milwaukee in 1995, and in 1996 an expectation of as many as forty hours was instituted statewide. The fulltime standard was set as part of Pay for Performance, the 1996 reform that anticipated W-2 in expecting ongoing activities from welfare parents. Pressure was put on the contractors running JOBS and on the clients to meet these standards. Most of the contractors were county welfare departments. As of 1994, they were expected to serve 80 percent of the recipients referred to them by the income maintenance side of welfare. I discuss below the financial incentives that backed this up. Monitoring of client compliance also increased. Traditionally, recipients subject to work obligations had been allowed two failures to report to assignments before the state began proceedings to “sanction” them (reduce their grants). In 1995, that was cut to one. The penalty for noncompliance also rose. Traditionally, only the share of the cash grant attributable to the adult who failed to cooperate was cut; benefits continued for the rest of the family, including children. As part of PFP, however, the state changed to full-family sanctions. Benefits were now reduced by the minimum wage (then $4.25) for each hour by which a participant’s activities fell short of his or her assignment, and if hours fell below 25 percent of hours assigned, the entire grant was canceled. The Food Stamp benefit was then reduced as well.27 Under JOBS, states were supposed to achieve certain activity levels among recipients on pain of reductions in their federal funding. Seven percent of recipients that states judged “mandatory” for JOBS (that is, employable) were supposed to participate in 1990, rising to 20 percent by 1995. Wisconsin was more ambitious. In addition to changing the rules for clients, Madison pressed the counties to raise participation rates. Table 4.1 shows the percents of AFDC adults mandatory for JOBS and two measures of participation in JOBS, for both Wisconsin and the median state in the nation from 1991 through 1996. The first participation rate is the official JOBS measure—roughly, the share of mandatory adults meeting the twenty-hour standard. The second is the share of welfare adults, mandatory or not, who participated in JOBS to any extent. The figures show that Wisconsin did not classify an unusual percent of its welfare adults as mandatory for JOBS in the early 1990s, but the fig-

Implementing Work Requirements

•

71

TABLE 4.1 Rates of JOBS Mandatoriness and Participation among AFDC Adults in Wisconsin and the Median State, 1991–96 (figures in percent) Year

Wisconsin Percent of AFDC Adults JOBS Mandatory Participation for JOBS Rate

1991 1992 1993 1994 1995 1996

40.1 52.7 40.4 40.7 56.5 66.7

38.8 33.3 30.9 31.7 35.6 60.6

Median state Overall Participation Rate

24.7 20.9 14.5 14.2 23.6 37.0

Percent of AFDC Adults JOBS Mandatory Participation for JOBS Rate

39.9 39.3 43.6 41.6 42.1 42.6

14.8 17.7 17.9 20.8 27.6 33.0

Overall Participation Rate

12.7 12.4 12.5 13.1 15.4 19.4

Sources: U.S. Administration for Children and Families; U.S. Congress, House, Committee on Ways and Means, Overview of Entitlement Programs: 1994 Green Book: Background Material and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, July 15, 1994), 357–59; U.S. Congress, House, Committee on Ways and Means, 1998 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, May 19, 1998), 484–86. Note: Data are for federal fiscal years. The JOBS participation rate is the percent of mandatory adults participating in JOBS for required hours. The overall participation rate is the percent of all AFDC adults participating in JOBS at any level.

ures rose sharply to two-thirds by 1996, reflecting the above rule changes. In JOBS participation, the state was consistently well above the median state, especially at the beginning and end of the period, and even further above JOBS standards. In overall participation, Wisconsin dipped close to the median state in 1993–94, but again soared well above by 1995– 96. In JOBS participation in 1996, Wisconsin ranked eighth; among larger, more urban states, only Florida, Oregon, and Washington scored above it. In overall participation in 1996, Wisconsin was fourth, with only three small rural states ahead. Between 1994 and 1996, Wisconsin raised its overall participation by nearly 23 percentage points. This was by ten points the most in the nation.28

TOWARD WORK FIRST Other rule changes reoriented JOBS from an emphasis on education and training toward employment in available jobs. When the program began in the late 1980s, most experts believed that the best way to overcome dependency and poverty among poor families was to improve the skills of parents so that they could get better jobs. This view was reflected in

72

•

Chapter 4

JOBS, whose rules favored assignment to education for parents who had not completed high school. However, it became apparent by the early 1990s that JOBS was overemphasizing remediation. Many clients were in school, but few were going to work. States were impatient to raise participation and employment.29 As noted in chapter 2, evaluations of the 1990s finally confirmed that programs that stressed work first realized larger gains in employment and earnings than those stressing education and training. In Wisconsin, the state initially said that JOBS was to target “the most difficult to employ.” This meant that “Job Search is not the primary activity.” Rather, there should be “an emphasis on using education and training . . . as means to improve the employability of welfare recipients.” In 1990, the legislature endorsed the idea that the aim was “long-term training/education,” including postsecondary programs that clients might take more than two years to complete.30 But by the early 1990s, reflecting the national trend, DHSS pressed JOBS contractors to place more clients in work. According to state officials, in 1994, the contractors were expected to place 12 percent of oneparent cases and 40 percent of two-parent cases in “work components,” which included CWEP, Work Experience (another form of unpaid work), Work Supplementation (where welfare subsidizes a private-sector job), or on-the-job training (OJT). In 1995, these figures were raised to 18 and 50—later 60—percent, respectively. Restraints were placed on higher education, which had allowed some recipients to go to school for years without taking a job. In 1995, education programs were limited to two years in length except in unusual circumstances, and four-year programs were disallowed. In that year also, clients in education or training programs for less than twenty hours a week were expected to undertake work components as well, and in Milwaukee, 75 percent of all clients in education or training were expected to be in work components or working part-time. Finally in 1996, the inauguration of Pay For Performance required that recipients “earn” their grants through near full-time hours of work activities. PFP forced counties to abandon the remediation strategy, if they had not already done so. The shift in priorities is dramatized in table 4.2. It shows the proportions of JOBS participants whom Wisconsin and the country as a whole assigned to various activities, or components, over 1991–96. Wisconsin sharply cut the percent of clients involved in assessment. This was a sign of tightened administration, since assessment delays the assignment of participants to activities. The state never assigned many people to job training, but zeroed out even those. It sharply cut its use of remedial education and postsecondary education. Meanwhile, it built up job search. Work readiness is a motivational activity that declined in use, probably

Implementing Work Requirements

•

73

TABLE 4.2 Percentages of JOBS Clients Assigned to Various Components in Wisconsin and the Nation, 1991–96 Component Assessment Job training Remedial education Postsecondary education Job search Work readiness Work supplementation CWEP

WI US WI US WI US WI US WI US WI US WI US WI US

1991

1992

1993

1994

1995

1996

42.5 21.3 2.4 14.5 17.8 25.6 21.1 19.4 13.8 11.9 5.9 5.9 0.4 0.2 2.8 2.8

10.2 16.5 3.2 15.8 15.9 24.5 24.4 14.9 14.8 9.4 4.2 4.5 1.0 0.1 3.3 3.7

12.4 17.6 3.1 15.4 16.6 23.3 22.7 15.7 13.4 8.0 2.8 4.5 1.2 0.3 3.7 4.3

8.2 12.5 3.7 15.8 14.6 23.9 25.7 18.8 15.8 5.5 3.5 5.2 2.3 0.2 4.9 3.6

6.7 11.1 0.0 15.8 12.2 22.7 16.7 16.7 22.6 6.1 3.6 6.2 1.6 0.3 5.2 3.8

4.6 10.3 0.0 13.4 9.6 19.7 12.1 13.7 37.9 9.0 2.3 7.4 3.0 0.3 14.0 8.8

Sources: U.S. Administration for Children and Families; U.S. Congress, House, Committee on Ways and Means, 1996 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, November 4, 1996), 420–21; U.S. Congress, House, Committee on Ways and Means, 1998 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, May 19, 1998), 482–83. Note: Data are for federal fiscal years. Remedial education includes high school, GED, and ESL. Job training includes job skills and vocational training. Post-secondary education includes both assigned and self-initiated. CWEP means Community Work Experience Program and includes other unpaid work. OJT is omitted because it was very small and changed little.

in favor of job search. Work supplementation and CWEP grew alongside job search, although less strongly. In most cases, the shifts are sharper than seen for the country as a whole. According to a government study, the share of active JOBS participants that Wisconsin assigned to workoriented activities soared from 40 percent in 1994 to 88 percent in 1997.31

DIVERSION A final policy change is that the state tried to divert needy families from going on aid at all. The reasoning was that, although most families that sought welfare were needy enough to qualify, few had exhausted other

74

•

Chapter 4

possibilities for income. By seeking support from friends, family, or local charities, or by going to work, most could avoid welfare. Under Work First, developed in Grant and Kenosha Counties and instituted in most other counties in 1994–95, applicants for aid had to spend two weeks of looking for jobs while their papers were processed; many obtained jobs and never went on aid. In 1995, applicants for two-parent AFDC, and for one-parent AFDC in Milwaukee, faced eight weeks looking for work if they went on the rolls. In 1996, the requirement to look for work before one’s application was approved was extended under Self-Sufficiency First, the companion to Pay For Performance. Work First had applied only to applicants mandatory for JOBS, and the sanction for noncooperation, once benefits were approved, was only denial of the noncooperator’s share of the grant. Under SSF, all applicants had to spend thirty days looking for work prior to receiving aid, and the penalty for noncooperation was loss of the entire grant. Note that the door to aid was not simply closed. Rather, “resource specialists” worked with applicants to arrange some way to meet their needs other than going on welfare. DHSS figures show that, during 1995, 35 to 55 percent of applicants for AFDC in the counties running Work First were “diverted.” That is, they were potentially eligible for AFDC but were talked out of it by the resource specialist, failed to complete their applications, or got a job before going on the rolls. Of course, many applicants for aid always fail to follow through, and there are no statewide numbers prior to Work First to prove that the incidence increased. But local and state staff say it did. In Racine, which did compile figures, 75 to 80 percent of applications had traditionally been approved. Under Work First in September 1995, that rate fell to as low as 29 percent, depending on how many of the initial contacts one assumes would have filed papers and gained aid under the earlier rules.32 One apparent cost of diversion, however, was that participation in JOBS fell. In October 1993, state data shows that 86 percent of recipients referred to JOBS were enrolled in the program. Two years later, the rate was only 69 percent. The reason was not that administration had loosened; it was getting tighter, with fewer clients inactive.33 State officials say that one explanation was that the shift to the CARES computer system in 1994 lowered apparent participation for technical reasons. Another was that the change in the age-of-child rule earlier in 1995 sharply increased the pool of clients mandatory for JOBS, thus reducing the enrollment rate. But the major explanation was probably that the new work and diversion policies caused more recipients to leave welfare quickly. They would go on aid and be referred to JOBS, but leave welfare before they could actually enroll in the work program.34

Implementing Work Requirements

•

75

MOTIVATING THE BUREAUCRACY The Wisconsin welfare bureaucracy, both state and local, handled the complexities of the new work policies and the many Thompson experiments with impressive efficiency.35 But many of its personnel preferred the less demanding welfare that had prevailed prior to the reforms. While most officials recognized some need to change, many opposed the new work requirements, fearing that they would lead only to hardship.36 Tommy Thompson and his lieutenants recognized these doubts. In an interview, Gerald Whitburn, secretary of DHSS between 1991 and 1995, remarked on the need to “bring the bureaucracy along.” In part, the administration accomplished this through leadership. Thompson was careful to challenge the bureaucracy to reform welfare without attacking it. He defined the problem and told officials to solve it. He set “a clear direction,” my respondents said, then trusted administrators to follow through. “He turned us loose,” two of my informants told me. This had a “huge impact on staff,” another said, motivating them to accept and even lead change. At the same time, the state altered funding incentives to align bureaucratic self-interest with reform. Welfare benefit payment and welfare work programs were run in most localities by country governments, chiefly by their local welfare departments, although in a few cases by other public agencies. These agencies in turn often subcontracted to private organizations to provide services. In Milwaukee, the country government had declined to run JOBS, so the state had contracted it to the welfare department, the ES, and five nonprofit service organizations. In 1994, counties were required to “earn” 55 percent of their JOBS funding for two-parent cases by placing these clients in jobs or work components. They gained $600 per full-time job placement, $300 per parttime placement, and $200 per month for placements in work components. In 1995, this performance pool rose to cover 60 percent of two-parent JOBS monies and 20 percent of the much larger budget for one-parent AFDC. The payments per placement also rose. These incentives were the lever behind the plan requirements mentioned above. The state also imposed a novel system of bureaucratic competition on the agencies that ran JOBS in Milwaukee, which it viewed as underperformers. During the first half of 1994, six of the seven providers were reorganized into two teams, one led by the local welfare department, the other by Goodwill Industries, a large nonprofit agency. The budget was divided between them and they were told to compete to achieve job placements and other goals, with the results to affect funding in the next fiscal cycle. Clients were allocated randomly between the two teams, to ensure a fair test. Faced with this ultimatum, each team scrambled to call in

76

•

Chapter 4

unassigned clients and assign people to job search as they had never done before. Job entries soared. By the time the competition ended in June 1994, placements were 29 percent above, and full-time placements 38 percent above, the level for the same period in 1993.37 No other change could explain this. The providers disliked the competition. They preferred a cooperative system where everyone worked together to run JOBS. Accordingly, in 1995 the state replaced the teams with incentive payments similar to those set for other counties in 1994.38 Again, agencies scrambled to make placements, because now they earned money by doing so. Job entries again soared, to as much as twice the monthly level achieved in 1994. For all of 1995, Milwaukee County achieved 7,400 jobs, a 52-percent rise over 1994.39 Finally in December 1995, as detailed in chapter 7, counties were told that they had to reduce the AFDC caseload by 25 percent and achieve other goals between September 1995 and August 1996 in order to be assured of selection to administer W-2. By these pressures and suasions, Wisconsin’s leaders attuned the bureaucracy more fully to their goals. THE EFFECTS OF WORK REQUIREMENTS As discussed in chapters 9 and 10, WEJT or JOBS had little effect on the rolls through 1990; they may actually have slowed caseload fall. That is not surprising, given that in these years the programs stressed education and training over actual employment. That changed with the shift toward higher participation and work first. Now work programs drove the caseload down and work levels up. Comparisons of counties show that those that were tougher about enforcing work put more recipients in jobs and had more caseload decline (see chapters 5 and 9). Recipients left by several routes, all of them tied to JOBS. One was entering employment. Despite a falling caseload, JOBS achieved rising job entries—16,067 in 1994, 19,504 in 1995, and 23,280 in 1996. While Milwaukee had traditionally lagged the state, the increase in entries in the city now was even larger than elsewhere, as the demanding policies of 1995–96 took effect. After 1996, job placements fell, probably due to an accelerated caseload drop as W-2 was implemented.40 A second route was diversion. As applicants for aid were confronted up front with the demand to work or to make other arrangements for themselves, more did so. This depressed new entries to the rolls and, as departures continued, drove the caseload down. In the eighteen counties that instituted Work First in November 1994, the caseload had already fallen 40 percent since January 1987, a period of almost eight years. The

Implementing Work Requirements

•

77

remaining cases then fell another thirteen points just in the next fifteen months (still relative to January 1987), as more clients got jobs or avoided applying for AFDC under the new policies. Self-Sufficiency First and Pay For Performance accelerated the caseload decline statewide. Between March 1996, when they were implemented, and September 1997, when W-2 began, the number of people on the rolls plummeted from 175,541 to 88,575, or 50 percent.41 A third route was sanctions. Recipients can be removed from the rolls for failure to fulfill requirements linked to aid. In the early 1990s within Wisconsin, counties that raised their sanctioning levels typically had more caseload decline (see chapter 9). In a federal study comparing states in 1996, Wisconsin had the most benefit terminations—5,700—out of thirty-three states examined. This may suggest that many clients were in conflict with the program. Actually, in the vast majority of cases, clients simply failed to enroll in JOBS or to appear for assigned activities, and then their grants were reduced or terminated.42 By all these routes, the number of case closures came to exceed openings. In the mid-1980s, at the height of the caseload, around 20,000 cases went on and off the rolls in a single quarter. By late 1989, both entries and exits had dropped almost in half, suggesting that most of the shorttime cases had already left the rolls. Entries dropped again sharply in 1995, after the diversion programs began, and again in 1996 when SelfSufficiency First and Pay For Performance came on line, while exits rose. The difference between entries and exits grew to 2,000–7,000 a year, and it was this gap, year after year, that drove the rolls down.43

CONCLUSION The Wisconsin story confirms that commitment at the center matters for policy implementation. Governments can execute social programs more fully if they put the requisite political and financial resources behind them, and improve their design. Wisconsin achieved work enforcement when it decided to take that mission seriously, developed a bureaucracy equal to the task, and made its programs more demanding. This success is the more impressive because it occurred in just the kind of complex, human service program that has been most subject to implementation failures. Previous top-down successes typically involved regulatory programs—such as school busing or voting rights—where a strong legal mandate was enough to compel change at the local level.44 Here, government had to gear up to enroll large numbers of people and then equip and mobilize them to work. Stiffening the legal mandates was part of the story, but equally important was bureaucratic statecraft—the devel-

78

•

Chapter 4

opment of a strong, efficient, and engaged administrative structure. By funding and staffing its programs amply, and by motivating local agencies to get on board, Wisconsin overrode all obstacles. A paradox of the story is that government had to grow, in a sense, in order to shrink. It was only by building up the bureaucracy that Wisconsin could get better control of its caseload. Only then could more recipients be made active in work programs, which in turn moved them off the rolls or deterred them from entering. Only when the new work demands became effective at the operating level was employment truly built into the welfare mission. Only then did the culture of welfare change, for staff and recipients alike.

CHAPTER 5

Local Variations

MOST ACCOUNTS OF welfare reform in Wisconsin give the credit to statelevel officials, Governor Tommy Thompson chief among them.1 While their role is undoubted, the revolution in Wisconsin welfare also owed much to the local level. Kenosha County pioneered the idea of enforcing high levels of participation in work programs, Kenosha and Sheboygan developed programs focused on work first rather than education and training, and Grant County developed the policy of diversion, or finding alternatives for needy families other than welfare. State officials say that all these ideas influenced W-2.2 We shift here from chapter 4’s top-down perspective to bottom-up, but with a new twist. Scholars of implementation sometimes suggest that when localities reshape social programs funded from above, they must be doing things of which higher-level governments disapprove. To innovate is to deviate.3 Some counties did resist state efforts to enforce participation in work programs and work first, and Madison finally imposed its will on them. The lead counties, however, innovated in ways parallel to the state. Like the parties in the legislature, the governments focused on the same problem—rising dependency—and this drove their policies in the same direction: toward work enforcement. Indeed, they competed to find solutions. It was another reflection of the earnest, moralistic style. Elazar says that moralistic states prefer decentralized administration, but here strong leadership at both the state and local levels collaborated to generate a new policy.4 We can also use variation among the counties to break out of the limitations of a case study, as suggested in chapter 1. We can show that the more enterprising counties were also those that framed the most effective welfare policies, given what was known at the time. They favored mandatory participation and work first. We do not find more passive counties that did this, nor do we find reforming counties that took other directions. These policies, in turn, can be shown to have maximized program performance, as that was measured in JOBS. This demonstrates in microcosm the theme of this book, that good government is the key to successful welfare reform. My account is based chiefly on interviewing Wisconsin welfare officials at the state level and in ten counties in the mid–1990s.5 I included most

80

•

Chapter 5

of the largest counties in the state in terms of population (Dane, Kenosha, Milwaukee, Racine) and also those that state officials viewed as reform pioneers (Grant, Kenosha, Sheboygan). However, I also visited counties that were not high performers in the state’s eyes. These had either resisted work first (Douglas, Marathon, Milwaukee, Racine) or had only recently changed in that direction (Dane, Fond du Lac, Winnebago).6 In each county I (and in some cases a research assistant) spoke to a selection of managers and line staff in the welfare department and JOBS. We asked how the counties processed clients in JOBS, what approaches they thought worked best, and how they explained the rapid decline in the Wisconsin caseload. I also compared how counties said they handled their caseloads with how they actually did, as shown by program data. Finally, I constructed statistical analyses to show that the policies of the leading counties did indeed promote JOBS performance.

KENOSHA The most important of the pioneer counties was Kenosha, a mixed urban and rural county in the southeast corner of the state. During the reform period, Kenosha faced unemployment problems due to a declining industrial base. It also confronted a migration of low-income people from Chicago—the problem that had helped to trigger the Wisconsin reform in the mid–1980s. Leadership A key to the statewide reform was that politicians and administrators performed complementary roles. The governor and legislators defined the welfare problem and called for change. But they left administrators leeway to craft actual programs, and they trusted them to do so. The same thing happened in Kenosha. Welfare was given a mandate for reform by the local Public Welfare Board, most of whose members were county supervisors. Most of the members were also Democrats, but the board operated in a nonpartisan manner, a hallmark of Wisconsin’s political style. Board members called for a reform promoting employment, but did not decide details. Welfare managers delighted that they could “go to them with ideas” and get support, yet not be told what to do. The principal political figure was County Executive John Collins, by all accounts a forceful and effective leader. Collins, said one official, was “as liberal a Democrat as they come,” but he also treasured a “strong work ethic.” The Kenosha reform would above all vindicate that value.

Local Variations

•

81

The county also developed an enterprising group of welfare managers. Clark Earl became director of the welfare department in 1981. Finding Kenosha’s existing welfare work program to be ineffective, he was willing, he told me, to “knock heads” and “make everybody change,” and he did so. That style eventually made enemies. Earl came in conflict with Collins and left the county by 1993—but not before he had broken the mold of welfare in Kenosha. A key lieutenant was George Leutermann, who headed income maintenance under Earl. He kept that role after leaving county employment to work for Goodwill Industries, after Goodwill became chief contractor for the new Kenosha work program. Later still, he headed up Maximus, one of the private agencies chosen to run W-2 in Milwaukee. A third figure was Larry Jankowski, who joined Goodwill in 1989 to head up the Job Center, which was to run JOBS and other work programs in Kenosha. This ambitious group, seconded by other managers and staff, transformed welfare in the county. WEJT State officials agreed that existing work programs were ineffective. This was why they created WEJT in 1986, which led on to JOBS. The Kenosha managers applied to run one of the initial pilots for WEJT in order to get extra financing for their ideas, and they were selected.7 Working through a “management group” of directors from several local agencies, they crafted a model designed mainly to achieve early client involvement and maintain it. Recipients found mandatory for WEJT were to go directly from eligibility determination into motivational training, then into skill assessment, then into a series of training or work assignments, with scarcely a break. Constant monitoring of cases by staff was to produce continual “engagement,” thus countering the tendency of clients to drop out (see chapter 8). Participation was also demanding. Recipients were to achieve a “simulated work week” of as much as thirty-two hours of activity, to get them used to the demands of employment. The idea, Earl said, paralleled the thinking behind the Saturation Work Initiative Model (SWIM) in San Diego, an important federally funded experiment of the same era that tried to maximize participation.8 The initial implementation of the program in 1987 failed. WEJT was contracted to the local Private Industry Council (PIC), the agency that ran training programs under the federal Job Training Partnership Act (JTPA). The PIC, however, chose to serve only the 250 participants specified in its contract. This it did by “creaming,” or concentrating on the most employable cases, as JTPA traditionally did. The county had expected that it

82

•

Chapter 5

would serve the entire welfare caseload found mandatory for WEJT, or over 1,000 cases, as only this would change the nature of welfare. A team of academic consultants recommended that the county reassert control of the program, and it did so.9 Later in 1987, the management group recontracted WEJT to Goodwill and other nonprofit agencies. For several months, in Leutermann’s words, the Kenosha organization “went on a rampage.” Implementation of the program was slowed while staff audited every case to find out what their clients were really doing, or not doing. Out of this crucible, Kenosha developed the capacity to monitor its cases and follow up on dropouts as closely, perhaps, as any welfare department had ever done. This administrative mastery would be the foundation of everything the county achieved. Work First The model had still to be fine-tuned. Initially, WEJT everywhere in the state stressed putting recipients in education and training rather than right into jobs. The goal was to raise skills so that clients could get well-paid positions and thus get entirely off welfare and stay off. Later evaluations were to show that this approach achieved less than putting people to work immediately in the jobs they could already get (see chapter 2). Already in 1988–89, the Kenosha managers realized that many clients were failing to complete training assignments and go on to work. So they shifted away from “education for education’s sake,” Jankowski said, toward education aimed more directly at jobs. Then in late 1989, Governor Thompson approached them and asked them to pilot a work first program for the state, offering them extra money to do so. A working group of personnel from all levels sketched out a model that was similar to WEJT but emphasized immediate work more strongly. This meant an initial week of orientation and assessment followed by seven weeks of job search, after which cases not getting jobs were put in community service (unpaid) positions. The philosophy now was that “You do not need an education to get a job,” one supervisor told me. Rather, “To get a job, you need to work.” The county continued to refer many clients to education and training—but now they normally had to be working at least part-time first. That step away from the remediation strategy would eventually be followed statewide, then nationwide. Kenosha became one model for the Work First program, oriented to both job search and diversion, that the state implemented starting in 1994.10

Local Variations

•

83

Resources The Kenosha managers were adept at getting extra money from the state and Washington in order to implement their ideas. Whenever an experimental program came along, they applied to run it. “If it sneezed we piloted it,” Leutermann joked. “I still have calluses on my knees.” For the five WEJT pilots, the state allocated $1.8 million in new state money. To pilot Work First, Thompson offered the county $100,000—it bargained for and got $350,000. Later, after JOBS replaced WEJT, Kenosha, and also Jackson County, received funding sufficient to serve 70 percent of their caseloads, because they had achieved high participation, whereas other counties were funded at only 47 percent.11 Still further funds came for county experiments in Job Centers, case management, and client tracking. Kenosha used this money to hire new staff and build up the bureaucracy so that clients could be served and monitored with unprecedented intensity. The program used teams of staff to serve clients, a structure that called for more personnel than single case managers (see below). Most of this buildup occurred in the late 1980s, according to respondents. From 1990 on, WEJT and JOBS had fifty to fifty-seven staff annually, or about one for every fifty cases—an unusually high ratio for welfare work programs.12 Since caseloads fell virtually throughout the reform period, this alone permitted smaller caseloads, without any additional hiring. Effects of WEJT WEJT’s primary goal was to raise “engagement” in the program, and this it achieved. According to Kenosha data, by early 1989, almost two-thirds of the recipients eligible for WEJT were either participating in the program or employed at least part-time—rates apparently exceeding even those achieved in SWIM.13 Participation here meant any activities at all. Under JOBS, to be counted as participants, clients had to undertake at least twenty hours of activities a week, and complete three-quarters of them. On that basis, Kenosha WEJT in 1990 achieved participation rates as high as 55 percent.14 This was vastly above federal norms for JOBS, which in 1990 required only 7 percent of eligibles to participate, rising to 20 percent in 1995. Kenosha’s programs were not rigorously evaluated, but the raw numbers are encouraging. A county follow-up succeeded in interviewing 39 percent of all clients who had left the rolls between late 1989 and early 1991. Over 60 percent were employed, that rate was sustained even a year after leaving welfare, and the rate was slightly higher for clients who

84

•

Chapter 5

had been in WEJT than for those who had not.15 A state report found that Kenosha JOBS placed 30 percent of its clients in jobs in 1993, the highest of ten large county programs.16 The program probably did improve outcomes, over what otherwise would have occurred, because it closely approximated the form that has been found to be most effective in evaluations elsewhere—a focus on work tempered with some attention to good jobs.17 In 1992, the National Alliance of Business honored Kenosha as the best JOBS program in the country. Tensions with the State Despite their achievement, however, the Kenosha managers “weren’t easy to deal with,” recalled Gerald Whitburn, who was state welfare director in the early 1990s. He recalled that Collins used to call him up and bend his ear to get extra funding. He respected the Kenosha managers because they “walked the talk,” and showed conspicuous “professionalism,” but he also found them “tenacious” about their own ideas. The program, said another state manager, finally became a “wayward child.” The issue, curiously, was work first. Although Kenosha led the way toward a focus on actual employment, the county resisted the extreme of this approach represented by later state policy. Its model was really participation first, not work first. It demanded the immediate immersion of clients in the program, but they also received preparation before they were expected to go out and get a job. One could not expect recipients to go out “cold turkey” and land positions, Jankowski remarked. They needed first to be oriented and motivated. W-2’s policy of demanding employment even as a condition of aid, Earl said, was simply a way of “blowing people off welfare.” Kenosha also opposed the two-year time limits that W-2 placed on any one assignment. Even under work first, the county still placed many clients in training alongside work. The idea of combining work with training, Leutermann said, got “lost” in W-2.

GRANT Grant is a rural county in the southwest corner of the state. Here the local welfare director, Jon Angeli, had created a work program in his own image over a decade. His approach was “family-centered.” He believed that it was usually problems within a poor family, for instance spousal or child abuse, that threw it onto welfare and kept it there. The solution was intensive intervention by caseworkers. Angeli recruited welfare and JOBS personnel who embraced his approach, then sent them to the University

Local Variations

•

85

of Iowa to be trained as “family development” specialists at the National Resources Center on Family Based Services. Grant’s approach was closer to traditional social work than Kenosha’s. The idea was that past failures caused a family to feel powerless, then to react in dysfunctional ways, then to abandon self-sufficiency for welfare, in a descending spiral. Caregivers could reverse that spiral with attention and services that would restore confidence, improve functioning, and eventually get a family off aid. Work was the end of the process rather than the beginning. Only thus could one help what Angeli called the “sociologically unemployed.” His sizable staff of social workers and other aides allowed him to achieve caseloads of thirty to fifty, permitting close attention to each case. Thus, Grant’s job placement performance in 1993 was not unusual for the state, well below that of Kenosha or Sheboygan.18 At the same time, the approach promoted work indirectly. Welfare might lavish services on families, but it also opposed “enabling.” It asserted “empowerment,” or the idea that the poor must ultimately be responsible for solving their own problems. Families were to be seen as “capable and functional,” not as “incompetent.”19 From this premise, tough work expectations followed. JOBS in Grant County was severe about expecting clients to seek and accept jobs as soon as possible. In this it reflected the conservative ethos of the surrounding farming community. In rural areas, families face more social pressure to leave assistance than they do in cities.20 If the county did not excel in job entries in 1993, that might be because it had already driven the most employable cases off the rolls. Another implication was diversion. To the extent possible, Angeli believed, families in trouble were to look to their “natural support systems,” including the “extended family,” before they asked for public support. Grant, like Kenosha, was a pilot for work first, but its main innovation was not immediate job search. Rather, starting in 1990, the county demonstrated that one could keep most applicants for aid from going on the rolls if one confronted them about other options at the outset. The “family resource specialist” who did this was a talented young woman who had herself been on welfare. She found ways to get applicants into JOBS or work or to make other arrangements for them without putting them on the rolls. “A lot of people come in here expecting us to solve their problems,” she remarked, but they learned that the solutions lay “within” them. Not everyone agreed with diversion. Grant, like Kenosha, came into tension with the state, but for being too tough rather than not tough enough. Some other counties thought Angeli’s approach was virtually illegal. Grant was talking people out of assistance that they had a legal right to. That feeling reflected the traditional, entitlement view of welfare that the state reform would finally reject. Angeli and his team responded that

86

•

Chapter 5

they were merely serving the needy most effectively. Although the state funded the county to develop diversion, managers in Madison were cool about adopting it statewide, until Work First in 1994. Jean Rogers, head of income maintenance at the state level, recalled that Angeli “was pushing the envelope as far as he could push it.”

SHEBOYGAN Sheboygan is a small city with a mixed economy on the Lake Michigan shore north of Milwaukee. Much of its caseload consisted of Hmong refugees who migrated to the state from Southeast Asia following the Vietnam War. While most of these families were two-parent, they had serious difficulties adjusting and learning English, and thus many subsisted on welfare. Here, as in Kenosha, the county welfare department had become dissatisfied with the existing welfare work operation. JOBS and earlier programs had been run by the local branch of the Job Service, and they had served only a small part of the caseload. Sheboygan’s solution was not a home-grown program. Rather, the department contracted JOBS to Curtis and Associates, a private firm that had run employment programs in several states. Curtis created a hardernosed version of Kenosha. There was a similar stress on up-front job search, but the goal was more often full-time employment. Kenosha’s frequent combination of work with training tended to leave families on the rolls, receiving partial grants to supplement part-time wages. Sheboygan aimed to get families entirely off welfare, which full-time work would usually do. If a client had not found work after four to six weeks of job search, he or she would enter a community service job but continue to look for work. If clients were assigned to remediation, it was usually English-language training rather anything more advanced. The program kept clients active through the same close follow-up as Kenosha. By 1992, Sheboygan claimed to have achieved a participation rate by JOBS standards of 56 percent, like Kenosha’s vastly above-national norms. The county became known as a forceful placement operation. In the last half of 1992, by its own reckoning, 28 percent of the county’s cases recorded earned income, compared to a state average of 18 percent.21 In 1994, Sheboygan placed 40 percent of its more disadvantaged clients in jobs, a rate second in the state among twenty-seven midsized programs.22 Sheboygan also excelled at diverting applicants from aid, through temporary assistance or by referring them to local charities. This was one of the first counties to implement Work First in 1994. The result was a sharp fall in the county caseload. The state assessed Sheboygan’s Work First

Local Variations

•

87

operation so favorably that it later contracted with the county and Curtis to consult with other counties about how to reduce their own caseloads, as this was a requirement for the departments to be chosen to run W-2.

CHANGING COUNTIES Besides these pioneers, other counties supported welfare reform by showing that they could change toward higher participation and work first, once they faced pressure to do so. In some cases, change occurred because a county became a pilot for one of the Thompson experiments. Those programs, strongly shaped by the leading counties, helped spread new ideas around the state. Dane Dane County, which contains Madison, Wisconsin’s capital and secondlargest city, was one of the most liberal jurisdictions in the state. Here JOBS had strongly stressed education and training, rather than putting recipients to work or reducing the caseload. But by the 1990s, a failure to raise participation and generate jobs led to disquiet at the state level and, finally, in Dane itself. Local administrators candidly admitted shortcomings. Then JOBS went through changes similar to, although less far-reaching, than those Kenosha had undergone years before. On the advice of the same consultants that served Kenosha, in 1992 JOBS was taken away from the PIC, which had run it poorly. It was recontracted to other nonprofit agencies, which the county monitored more closely. In 1992–94, the county hired many more JOBS case managers, reduced staff caseloads, enforced participation, and reassessed all cases for employability. “We can now say,” it asserted, “that our case managers know their caseloads.” The length of time recipients could study before going to work was cut back. By all these changes, the program began to close the gap with state averages in terms of participation and job placements.23 In 1994, Mary Ann Cook came from the state level to take over as JOBS manager, promising further steps in the same direction. The impetus for the changes was in part political. “The county board went to Kenosha and fell in love,” Cook told me. Pressure also came from business. In a tight labor market, employers objected to JOBS’ heavy emphasis on education and training when jobs were going begging. But the largest pressure came from within. JOBS staff realized that their approach to the program, oriented to human capital, was outdated. They simply wanted to perform better.

88

•

Chapter 5

Winnebago Winnebago County contains Oshkosh, a small city in east central Wisconsin. As in Dane, the JOBS program here had originally favored education and training. There had been no overt pressure from political leaders or business to move toward employment, yet welfare and JOBS managers concluded that they had to change. They could see where state policy was going, and they feared that state funding for education and training might cease. So they joined Work First, effective February 1995. The philosophy was now to be “work first, career second,” local staff told me. A resource specialist had been installed, as in Grant, to divert applicants from welfare when they first applied for aid. Should they persevere, welfare intake workers were supposed to read them a statement about how JOBS in Winnebago had changed. The goal now was to make them “self-sufficient.” They were supposed to “benefit from the good economy and abundant well-paying jobs” in the area. Should they apply for or receive aid, they would immediately confront a newly-tough JOBS program that demanded near-immediate job search. The JOBS redesign had been done by a staff member imported from San Diego, the inventor of high-intensity welfare work programs including SWIM. While some officials had misgivings about the shift, others were enthusiastic. They felt they could at least reverse a longtime growth in dependency. Fond du Lac Fond du Lac is another small inland city with an economy of light manufacturing. Here too, the traditional approach to JOBS emphasized education and training. When I first interviewed in Fond du Lac in 1994, staff described remediation for the low-skilled as their “highest priority,” and they believed long-term education was “very successful.” They hesitated to sanction clients (i.e., reduce grants) for noncompliance with requirements. They deprecated the rising pressure to enforce participation in JOBS, or to reduce the caseload. All that just meant “money, money” for the agency, one case manager told me—not helping people. The program seemed slackly run compared to Kenosha or Sheboygan, with many clients allowed to remain uninvolved for long periods. But as in Kenosha and Dane, pressure to change had come from the county executive, other politicians, and business. Fond du Lac applied for and was accepted (along with Pierce County) to pilot Work Not Welfare, the radical Thompson experiment that required that families earn

Local Variations

•

89

their aid and limited their eligibility to two years at a stretch. To implement this program, one manager admitted, would mean “a shift in our paradigm.” WNW began in Fond du Lac in January 1995. By the time I returned in June of that year, diversion was in full swing, the JOBS program had been toughened up, educational assignments had been shortened, and the county was enforcing work as never before. Notably, the welfare department ran JOBS both before and after the change solely with its own personnel, without contractors. The shift had been wrenching, causing some loss of morale and the departure of some staff. They had been replaced by others more committed to the new work mission.

LAGGING COUNTIES A final group of counties resisted the drift toward enforced participation and work first. In these cases, the reason was not usually that JOBS was poorly run. Indeed, by the standards of other states, JOBS like earlier welfare work programs was run well almost everywhere in Wisconsin. Rather, these counties had different goals from the state’s. For them, work programs should support the choices recipients made, and should seek high-paying jobs, rather than using work to change the nature of welfare. One result was that they did not attempt to monitor their clients with the intensity seen in Kenosha or Sheboygan. Racine Racine is a large county, situated just north of Kenosha and feeling some of the same pressure over welfare migration. It was a willing pilot of several of the Thompson experiments. It was also a leader in child support enforcement, another frontier of antipoverty policy. In welfare work, however, Racine took a more moderate line than its neighbor to the south. It was less insistent that welfare adults participate in JOBS. When they did, the county demanded fewer hours from them, and it often allowed them to go into education and training rather than immediate work. Overall, it was less demanding than the three vanguard counties, although more so than Dane. The key difference from the pioneers, and also Dane, was that there was less internal pressure to change. Welfare officials, and the elected officials standing behind them, were more content with JOBS as it stood.24

90

•

Chapter 5

Marathon Marathon is a moderate-sized county in north-central Wisconsin containing the city of Wausau. Here the dispute with state policy was sharper than in Racine. When I interviewed in 1995, the JOBS manager told me that she disagreed with work first and remained committed to an education-and-training approach. JOBS here sent many clients to the local technical college. The county refused to participate in any of the Thompson experiments. Some staff did feel that the emphasis on education had been overdone, but at most they favored shorter training assignments more focused on jobs. The county adjusted to state directives no more than necessary, rather than taking them as cues for change, as in Dane or Winnebago. For example, JOBS placed some clients in work, but in part to reap the bonuses the state had begun paying for job entries, the better to spend the money on training. By state standards, JOBS was loosely run, allowing frequent nonparticipation. Far from rebuilding itself around the work mission, as in Kenosha, the welfare department in Wausau continued to give aid on the basis of need alone. It was largely detached from JOBS. Douglas Douglas County is in the north of the state and includes the city of Superior. Here again a relatively passive JOBS program felt little urgency about moving recipients into jobs. One reason might be that this locality was less exposed to welfare migration than the southern counties. Its economy was also less prosperous, with fewer jobs available. The difference from the leading counties was less that JOBS was wedded to remediation than that the agencies running the program had never fully internalized the work mission. The welfare department saw its role as giving clients all the benefits they were entitled to, without attempts to divert. The Job Service, which ran work activities, emphasized assigning clients to public jobs as much as finding positions with private employers. Services to help families, which Grant had made its centerpiece, were relegated to a group of nonprofit agencies, funded by foundations. Douglas had joined Work First, but that program had not reflected or produced the changes seen in Sheboygan or Winnebago. The resource specialist stationed at welfare intake was unable to stop eligibility clerks from processing applicants for benefits before alternatives could be discussed. Welfare officials told me that putting recipients to work was “not part of our job.”

Local Variations

•

91

Milwaukee Milwaukee County contains the state’s largest city and—by far—its largest welfare caseload. Here as in other lagging counties, local JOBS agencies were reluctant to enforce participation in the program, and they emphasized education and training heavily at the expense of job placements. More important, JOBS in Milwaukee was poorly run compared to elsewhere in the state, although probably well run by the standards of other big cities. The county welfare department had declined responsibility for the program, for fear it would prove costly. So JOBS was overseen by the state welfare department and contracted to several local agencies, public and private. It was also loosely run, respondents told me, with many clients sitting unassigned, without participation, for years. This was partly because staff had higher caseloads, reflecting lower funding.25 The Milwaukee agencies were also larger and thus less efficient than those found in smaller counties. Units had to communicate with each other on paper, whereas staff from smaller operations could do so face-to-face. George Leutermann summed up the situation in 1995, when he was working in the city: “Milwaukee hasn’t reduced the caseload because they haven’t had in place a system that could engage enough people to make an impact.” The state attempted to tighten things up, through the competitive funding arrangements and incentive payments seen in chapter 4. But the unresponsiveness of the existing system ultimately led to the contracting of W-2 to a set of entirely private organizations.

STRUCTURAL INNOVATIONS Besides policy changes that led toward work first and diversion, the leading counties and some others pioneered innovations in the organization of welfare work programs. Case Management One of these was in how staff in welfare work programs served the clients. Traditionally, particularly in large programs, there are separate personnel to determine initial eligibility for welfare and then to run the JOBS program, to which the employable recipients are referred. Some clients get lost in the referral, which was one reason for frequent no-shows for JOBS. One solution was to “co-locate” welfare and JOBS staff in the same of-

92

•

Chapter 5

fices, to facilitate referral. Another was to have a single case manager handle both benefits and JOBS. Research suggests that this “integrated” approach achieves higher participation and employment.26 This was the approach ultimately adopted in W-2. Kenosha developed a third alternative—assigning clients to “integrated service teams.” Income maintenance, JOBS, and child care personnel would meet with the client together. With face-to-face communication, staff could quickly address any problems that might be blocking the client’s participation. JOBS staff could also quickly tell welfare of any nonparticipation that might warrant a sanction. Under Work Not Welfare, Fond du Lac adopted this approach. It helps to account for the unusual ability of these programs to monitor their caseloads closely. On the other hand, it was costly, because more staff members were needed to serve a given caseload. Largely to minimize cost, Sheboygan relied mainly on single staff interacting with groups of clients in orientation and motivation sessions. Contracting Another innovation was the frequent contracting out of work programs to private agencies. While contracting was not new, it became more common as counties sought to change welfare. They often saw outside agencies as a way to do this without having to alter existing public bureaucracies. In Kenosha, WEJT and JOBS were contracted out because the county board viewed hiring under civil service as too costly. Also, Clark Earl preferred a “corporate culture” where operators would be more accountable for results and more “customer responsive.” The reasoning was similar to the ideas of “reinventing government” that influenced W-2 (see chapter 7). Racine had a structure much like Kenosha’s, while Sheboygan contracted its entire program to Curtis and Associates. Grant contracted with several agencies for a consortium of five local counties that ran JOBS together. In Milwaukee, the JOBS structure used both public and nonprofit bodies. Fond du Lac, on the other hand, rejected contracting, despite the rigors of refocusing its existing organization on work. Job Centers A third innovation was the idea of merging welfare work with other programs, such as the Job Service, JTPA (the PIC), or vocational rehabilitation. It seemed inefficient to run these programs separately, since they all involved determining eligibility for benefits, appraising clients for employability, and then referring them to training programs or to job search in

Local Variations

•

93

the private sector. Why not create a “one-stop shop,” funded by all the programs, that could serve all the clienteles together? This idea had been recommended by a state-level task force in 1985 and promoted by the state department of labor, overseer of the Job Service and JTPA. Grant was apparently the first county to create a Job Center, in 1985, but it was centered on welfare. In 1987, the state provided funds to develop Job Centers, with money for 1989–90 going to Kenosha and three other counties.27 With support from County Executive Collins, Kenosha used those and other monies to buy a failed shopping center and move all the participating agencies into it. The center opened in 1990 with half a dozen agencies, but grew to nearly twenty. Its rooms and offices are plush and modern, poles apart from the typical run-down welfare office. When people in Kenosha seek aid, they now go to this impressive facility, not to a separate welfare department. A large sign outside proudly proclaims the “Kenosha County Job Center.” The center symbolizes Kenosha’s commitment to welfare, but also to making the system more work-connected and mainstream. In the county’s words, the center “forges in the welfare recipient’s mind the connection between receiving economic assistance and preparing for work supported self-sufficiency.”28 Similar, though less extensive, Job Centers were created in Dane and Racine. Centers were also organized in Milwaukee, although they were overtaken by W-2. Meanwhile, the employment programs outside welfare developed their own Job Centers, a move later mandated by the federal Workforce Investment Act. Many of the W-2 Job Centers became co-located with these, and by 2001 the state was considering combining the two structures.29

HOW CLIENTS WERE ASSIGNED The description above is based largely on my interviews at the state and county levels, plus program documents. I also checked these impressions against how the counties actually assigned their clients, according to JOBS reporting data. Client assignments for October 1993, a typical month of that year, are shown in table 5.1. They mostly confirm the picture respondents gave of the counties. Compared to the state overall, the leading counties excelled in the proportion of clients referred whom they actually enrolled. Enrollment is a good measure of a program’s authority, its ability to get recipients to take the work test seriously. The lead counties also minimized people who were in unproductive holding statuses, or who were sanctioned for noncooperation. Interviews and past research suggested that demanding work programs actually reduce sanctions because they leave

94

•

Chapter 5

TABLE 5.1 Clients in Administrative Statuses in Research Counties, October 1993 (percent)

Holding Enrolled Statuses

Sanctioned

Job Search

Regular Jobs

Unpaid Jobs

PostRemedial secondary Education Education

Leading Counties Kenosha Grant Sheboygan

88 87 90

24 24 11

1 2 1

8 6 8

37 26 35

3 7 5

11 5 32

12 17 11

1 3 3

6 15 13

8 10 9

Changing Counties Dane Winnebago Fond du Lac

76 84 84

34 26 33

4 3 1

2 4 1

16 26 24

Lagging Counties Racine Marathon Douglas Milwaukee

78 91 95 86

16 29 52 32

8 4 3 8

3 3 3 4

18 19 23 19

0 0 3 1

8 23 3 4

15 9 12 5

State

86

30

5

4

22

1

7

9

Source: Calculated from county caseload data from the Department of Health and Social Services. Note: The base for all percentages is all recipients referred to the program. County figures that exceed the state’s are in bold. Clients in holding status include those on hold, scheduled to participate, unassigned, and those whom the program does not intend to serve or has asked to be exempted from JOBS. Sanctioned means JOBS has asked welfare to reduce a grant for noncooperation. Regular jobs include full- and parttime. Unpaid jobs include community work experience and work experience.

participants in less doubt, compared to more passive programs, about what is expected of them. They convey work demands to clients informally, not by “throwing the book” at them. High sanctioning is thus a mark of a weak program, not a strong one.30 Above all, the lead counties had high proportions of clients in job search and actual work; 40 percent in both Kenosha and Sheboygan were working, either in regular or community service jobs. One surprise is that the leaders also had fairly high proportions in education. That was partly because their high enrollments gave them more clients to assign to all activities. Kenosha assigned many to training, and Sheboygan to English language instruction, alongside work. The other counties were less tautly run and less demanding. Enrollment usually ran lower than in the leading counties, and holding statuses and sanctioning ran higher. While Marathon and Douglas had unexpectedly high enrollment, they also had more clients inactive than the leaders, espe-

Local Variations

•

95

TABLE 5.2 Clients in Administrative Statuses in Research Counties, October 1995 (percent)

Holding Enrolled Statuses

Sanctioned

Job Search

Regular Jobs

Unpaid Jobs

PostRemedial secondary Education Education

Leading Counties Kenosha Grant Sheboygan

79 91 81

38 24 10

4 2 1

9 9 12

39 33 44

3 5 9

4 1 14

7 0 6

6 5 11

5 6 10

8 9 10

Changing Counties Dane Winnebago Fond du Lac

81 68 81

26 14 29

2 2 10

5 3 2

24 38 34

Lagging Counties Racine Marathon Douglas Milwaukee

82 88 80 76

15 20 22 23

2 1 3 5

11 3 12 12

29 28 27 24

2 5 6 6

6 19 5 2

14 15 11 5

State

69

20

3

8

23

5

6

7

Source: See table 5.1.

cially Douglas. Job search and work activity were mostly lower than among the leaders, but, except in Marathon, education assignments were not conspicuously high. Where the lagging counties fell behind the leaders, above all, was in making their clients active. As one might expect, the counties that were changing by the time I interviewed them in 1994–95 fell between the leading and lagging groups in many of these respects. Table 5.2 shows similar data for October 1995, two years later. The impact of the state policies described in chapter 4 is apparent. The gaps between the groups of counties have narrowed. Now almost all the counties are enrolling above the state’s overall level. All except Kenosha and Grant show reduced assignment to holding statuses. Sanctioning in most of the changing and lagging counties has fallen—but a sharp rise occurred in Fond du Lac due to the unusual demands of Work Not Welfare. Almost all the counties show rises in job search and employment, with the increases generally greatest for the lagging group. The work rise extended to unpaid work assignments, or “workfare,” notably in Fond du Lac, again due to WNW. Less clearly, there is a decline in education assignments, particularly in the leading counties. Changes this sharp in only two years show a remarkable capacity for bureaucratic change.

96

•

Chapter 5

The shifts largely track shifts in the statewide figures, where we also see a fall in holding assignments and sanctioning, a rise in job search and work, and a fall in education. Enrollment also fell, for reasons discussed in chapter 4. Yet the counties still differ in characteristic ways. Sheboygan still stands out for exceptionally tight administration. Kenosha and Sheboygan are still relatively tough on regular employment, and Grant less so, while Dane and Milwaukee—the two most liberal jurisdictions—remain relatively soft. Marathon has raised work levels but not job search and continues to emphasize education. Thus, the initial implementation of work tests in Wisconsin was interactive. Madison was not fully in harmony with the counties, even the leaders, who also differed among themselves. And yet by a complex interplay, the two levels found their way toward policies of higher participation and work first. The lead counties set the pace, the state followed, and changes in state policy then drove all the counties in the same direction. Like ships in a convoy, they all turned in the direction the state wanted, yet within those bounds each continued to chart its own course. The variations also confirm the association of good government with effective welfare policies, the second proposition in table 1.3. The leading counties, those with the most energy and initiative, all headed toward the mandatory, work-focused type of program that was supported by both experience and research. I heard of no cases where such programs emerged without county initiative, nor did reforming counties develop training-focused programs. Some jurisdictions opposed work first, but I heard of none that had innovated at the level of a Kenosha or Grant. The leading counties were also responsible for the structural innovations, such as Job Centers, that benefited the entire state.

PERFORMANCE MEASURES But the bottom line is the third proposition in table 1.3: Did the new, more demanding programs actually lead to successful outcomes? Were the more enterprising counties also the most successful in concrete results? We can show at least that the demanding policies they favored also promoted performance in JOBS, as that was measured at the time. Wisconsin developed its work policies largely within the context of the JOBS program, created by the Family Support Act of 1988. FSA demanded that states raise the share of their employable recipients who participated actively in JOBS to 20 percent by 1995. The act also directed the U.S. Department of Health and Human Services to develop further performance measures for JOBS. The Department never did this, largely because its efforts were overtaken by the election of the Clinton Adminis-

Local Variations

•

97

tration in 1992 and the beginning of a new phase of welfare reform. Wisconsin, however, was able to transfer to JOBS a set of performance measures similar to those that had been used earlier in the Work Incentive program, the predecessor of JOBS.31 I used them in the following form: 1. Job entries: Percent of registrants entering jobs expected to last at least 30 days. “Registrants” here and below means all clients referred to JOBS, whether or not they enrolled. 2. Working registrants: Percent of registrants working full- or part-time while still on AFDC. 3. Cases closed: Percent of registrants whose JOBS cases are closed, often due to employment. This usually meant that they left welfare as well. 4. Retention at 30 days: Percent of job entrants contacted who were still employed after 30 days. 5. Retention at 180 days: Percent of job entrants contacted who were still employed at 180 days. 6. Job entry wage: The average hourly pay of registrants entering jobs.

The first three of these indicators measure job quantity, the last three job quality. Better jobs are those that pay more and that recipients keep longer.

COUNTY PERFORMANCE Table 5.3 shows how the ten research counties performed on these measures in 1993.32 Kenosha, Grant, and Sheboygan stand out for their high rates of job placement, clients working while on welfare, and cases closed, relative to the average county. They are even further above figures for the state as a whole, which are depressed by Milwaukee. Grant is below average on job entries because it focused more on diversion; Kenosha is below average on case closures because its clients often combined part-time work and education or training, and thus tended to stay on the rolls longer. In contrast, the lagging counties, other than Racine, performed well on job quality, due to their commitment to enhancing skills, but they made relatively few job entries. In most ratings, the changing counties were below average on both job quantity and quality, perhaps because in 1993 their changes had just begun. To some extent, the leading and lagging counties simply differed about goals. The first tried to use employment in a wide range of jobs to change the nature of welfare; the latter pursued well-paying jobs without seeking the same overall change. The good-job strategy, however, generated fewer jobs and thus helped fewer recipients than work first. And Kenosha’s strategy of combining work with training allowed it to perform well on

98

•

Chapter 5

TABLE 5.3 Performance of Research Counties on JOBS Performance Measures, 1993 Job Entries

Working Registrations

Cases Retention Closed 30 Days

Retention 180 Days

Job Entry Wage

68.8 48.8 64.7

5.62 5.49 5.40

46.9 45.1 54.3

6.15 5.70 5.33

Leading Counties Kenosha Grant Sheboygan

25.6 20.0 27.6

33.8 34.2 37.9

44.2 51.9 50.8

91.8 75.0 81.2

Changing Counties Dane Winnebago Fond du Lac

12.3 19.1 18.5

16.3 29.0 28.2

40.4 43.9 42.4

82.2 70.4 88.0

Lagging Counties Racine Marathon Douglas Milwaukee

9.3 18.7 13.5 11.7

20.0 24.2 22.9 14.9

46.4 40.5 45.9 35.0

82.4 89.2 96.1 88.9

53.6 62.9 74.0 57.1

5.36 5.67 5.85 5.69

Country average

22.0

30.9

48.5

84.3

55.8

5.60

State

16.2

22.8

41.8

85.6

56.4

5.75

Source: Wisconsin Department of Health and Social Services. Note: Figures in bold are above the average for all counties.

five out of six measures—to achieve, that is, both job quantity and job quality. Superb administration allowed the county, apparently, to square the circle. The results also support Kenosha in its quarrel with the state over work first. But was the high performance of the leading counties really due to their demanding policies? Already by 1994, when I began to interview in the state, state officials thought it was. At a first cut, the numbers bear them out. Figure 5.1 shows the relationship between the percent of clients enrolled in JOBS and the percent entering jobs in the seventy-two Wisconsin counties in 1993, with the ten research counties numbered. The three vanguard counties stand high on both proportions, with mainly smaller, more rural counties above them. Milwaukee and Racine, the largest of the lagging counties, bring up the rear. Figure 5.2 shows a similar pattern for the percent of clients assigned to job search versus the percent entering jobs.33 Of course, the superior performance of the leading counties might be due to something other than their policies. Perhaps their clients were more employable or more jobs were available than in the lagging counties. We

Local Variations

•

99

Figure 5.1. Percent of clients enrolled in JOBS versus percent entering jobs, Wisconsin counties, 1993. Numbers stand for research counties: (1) Dane, (2) Douglas, (3) Fond du Lac, (4) Grant, (5) Kenosha, (6) Marathon, (7) Milwaukee, (8) Racine, (9) Sheboygan, and (10) Winnebago. (Source: Wisconsin Department of Health and Social Services)

Figure 5.2. Percent of clients assigned to job search in JOBS versus percent entering jobs, Wisconsin counties, 1993. Numbers stand for research counties— see figure 5.1. (Source: Wisconsin Department of Health and Social Services)

100

•

Chapter 5

TABLE 5.4 Influences on the Percent of Clients Entering Employment, Wisconsin County JOBS programs, 1993 Compared to the average county, a county had A higher percentage of clients entering employment if it had more Clients enrolled in JOBS Clients in job readiness activities Population density (persons for square mile) A lower percentage of clients entering employment if it had more Clients who were black Growth in employment Source: Lawrence M. Mead, “Optimizing JOBS: Evaluation versus Administration,” Public Administration Review 57, no. 2 (March/April 1997): 113–23.

have to allow for these demographic and labor market factors before we can say that how the programs treated clients affected their outcomes. Accordingly, I measured variables of all these kinds for each county and did analyses that related all of them to each of the performance measures. When we control for the social and economic environment, does the assignment of clients still matter?34 Clearly it does, at least for the job quantity measures. Among all the measurable influences on job entries, table 5.4 shows those that turned out to be important. Counties had a higher proportion of clients entering jobs than the average if they had a higher proportion enrolled in JOBS. There is no substitute for simply getting the recipients into the program. A higher proportion of clients in job readiness—a motivational activity prior to job search—had a similar effect, as did a higher population density in the county. A county had lower-than-average job-entry performance, however, if it had a higher proportion of clients who were black, and if it had higher growth in employment. The latter finding is counterintuitive; it may mean that in a growing economy the most employable people already have jobs; welfare recipients are then less employable and less placeable. Policy factors are even more important for working registrants, or the share of JOBS registrants who hold jobs while still on welfare. Table 5.5 shows that a country tended to score higher on this measure if it had a higher percent of clients enrolled in JOBS, or in job readiness, or in job search. Meanwhile, the only environmental term to matter is the male share of the caseload. No labor market factor mattered. For case closures, results look somewhat different. To work their way off the rolls, welfare mothers need to earn above the welfare eligibility level, not just take any job. The employability of the clients thus matters more than it does for job entries. In table 5.6, accordingly, we find that a

Local Variations

•

101

TABLE 5.5 Influences on the Percent of Clients Working While Still on Welfare, Wisconsin County JOBS programs, 1993 Compared to the average county, a county had A higher percent of clients working if it had more clients who were Enrolled in JOBS In job readiness activities In job search Male Source: See table 5.4.

county had relatively more case closures if it had a higher proportion of clients under twenty-five; but fewer if it had a higher proportion who were disadvantaged, black, or Asian.35 The percent of high school graduates is a negative influence, not positive as one might expect. One reason may be that, in Wisconsin in 1993, high school graduates in JOBS were still likely to pursue further education rather than work and leave welfare. Yet the enrollment rate is again a positive influence on performance. The rate at which clients are sanctioned (that is, have their grants reduced for noncooperation) is negative. This confirms the idea, mentioned above, that frequent sanctioning is an ineffective way to convey work expectations. All of these analyses are quite strong, in the sense that they account for most of what differentiates one county’s performance from another’s on these three measures.36 My ability to explain the three job quality measures—retention at 30 days, retention at 180 days, and job entry wage—was weaker. These are the measures that the lagging counties focused on, at the expense of job entries and closing cases. The main influences on whether a county performed well in these terms seemed to be the level of education of its caseload, especially the share of clients who were high school graduates. There were some indications that policy still mattered, but if anything, higher participation and job search were associated with lower retention and job entry wages rather than higher. That makes sense, for as the proportion of clients who look for work rises, average employability falls and the quality of the job people can get also falls, causing retention also to decline. This was why the lagging counties feared a high-placement policy, and it was partly to uphold retention that Kenosha tried to combine work in available jobs with training for “better” positions. But no policy had much effect on these three outcomes. Overall, the results validate the state’s perception about how to promote work and minimize dependency. State officials believed that whether JOBS clients went to work, or left welfare through work, was something that program operators largely controlled, regardless of the employability

102

•

Chapter 5

TABLE 5.6 Influences on the Percent Clients with Cases Closed, Wisconsin County JOBS Programs, 1993 Compared to the average county, a country had A higher percent of cases closed if it had More clients enrolled in JOBS More clients aged under twenty-five A higher labor force participation rate A lower percent of cases closed if it had more clients who were Sanctioned for noncompliance Classified as disadvantaged High school graduates Black Asian Source: See table 5.4.

of clients or labor market conditions. The keys were to raise participation in JOBS and then require that more recipients actually look for work. While the more voluntary, training-oriented approach favored by the lagging counties achieved marginally higher-quality jobs, it did not produce enough of them to help much of the caseload. It did not change the nature of welfare so that employment was truly built into the welfare experience.

MILWAUKEE One may doubt that these analyses do justice to Milwaukee, Dane, and other more urban counties. Those programs, many feel, must inevitably perform worse than average because their caseloads are more disadvantaged. Milwaukee’s caseload decline lagged behind that of smaller counties and the state as a whole until late in the reform process. Today, welfare is down by 90 percent for the state as a whole, but less so in Milwaukee. In January 1987, the county comprised 39 percent of the total state AFDC caseload, but by January 2002, it accounted for 79 percent of the remaining cash aid cases in the state.37 As described in chapter 3, Milwaukee is a heavily black and Hispanic city with serious social problems. One might doubt the employability of recipients there, and also the availability of jobs. Between 1977 and 1987, the city lost 37,000 manufacturing production jobs, long the mainstay of low-skilled workers. These losses were more than offset by increases in retail and service employment, although the growth was greater in the suburbs than in the central city. In the region as a whole in October 1995, there were 32,529 job openings for 25,220 unemployed persons.38 In

Local Variations

•

103

1993, Milwaukee actually had lower unemployment and much higher wages than the average county. Demographic and labor markets factors are important, but they were already allowed for in the above analyses. If one controls for policy differences alongside the environment, it does not appear that there is much special about Milwaukee.39 The largest difference between the city and the leading counties was not in the employability of its clients or the labor market but in the emphasis of the local JOBS agencies on “good” jobs instead of on putting more recipients to work. And, as noted earlier, the program was loosely run, with participation effectively voluntary. These were the main reasons why only 12 percent of Milwaukee JOBS adults entered jobs in 1993, compared to 22 percent in the average county. Thus, the state’s insistence that Milwaukee adopt tougher policies appears sound, at least given its goals. While the county’s clients were more disadvantaged than average, it still could demand more of them, and then more would work and leave welfare. It eventually did just that, but only after the total reorganization described in chapter 7.

DO WE BELIEVE IT? These results suffice to establish that the more enterprising counties exceeded the others in terms of the performance outcomes prevalent in JOBS. They also suggest that to enforce participation and actual work in JOBS was the best way to achieve higher work levels on welfare, which was the chief aim of the reform from the state’s perspective. I say suggest because these results do not prove, properly speaking, that required participation or work first had any effect on the clients.40 They do not constitute an evaluation of JOBS, in the sense of proving that the program raised the employment or earnings of its clients above what they would have achieved otherwise. Rather, the results compare more and less successful programs within JOBS. They show that counties that enforced participation and required work generally outperformed programs that did not. We do not know that clients in JOBS did better than equivalent recipients who never entered the program. One may also question the statistical nature of the analysis. The goal was to allow for the features of the counties other than policy that might have affected their performance. While I controlled statistically for the employability of the clients and the labor market, one can do this only to the extent those forces are measurable. The largest doubt is about whether the clients in the various programs differed in unmeasured ways, what statisticians call selection bias. There is also doubt as to whether the policy

104

•

Chapter 5

variables capture all the significant dimensions of the programs being tested. There can be other statistical problems with such models as well.41 The usual way of avoiding these problems has been experimental evaluation. Here, a sample of people eligible for a program is drawn, and then the sample is randomly allocated between people who do and do not get the new treatment. Random assignment makes it highly improbable that the experimental group, which receives the new program, will differ in any important way—measured or unmeasured—from the control group, which does not. The difference in outcome between the two groups is then a valid measure of the program’s “impact”—how much it improves on what would have happened without the program. Evaluations of this type, chiefly done by the Manpower Demonstration Research Corporation during the 1980s and 1990s, were crucial for establishing the effectiveness of welfare employment programs. But experimental evaluations tell us little about how programs function and the policymaking behind them, which is my main focus here. In an experiment, the program tested usually remains a “black box.” Even if it shows impact, one is unsure what feature of the program produced that result. There is also doubt about how fully the program was implemented. Not everyone eligible for the new program actually enters it, but impacts are often proportional to the participation rate. Finally, experiments are usually run at local sites where it can be assumed that the experimental clients and control clients face the same labor market. How, then, does one generalize across evaluations done at many sites where conditions differ? In all these respects, statistical analyses such as those above have advantages. One sees the effect of individual policies within programs, such as enrollment or job search, so that the black box problem is reduced. The administrative measures also capture the degree of implementation in a program; indeed, implementation, as measured by the enrollment rate, becomes a key variable in explaining performance. And analyses can be constructed across multiple sites, provided one can measure program features, the demographics of clients, and the labor market at all of them. One reason to trust these results (and also the analyses of caseload decline in chapter 9) is that they are backed up by the interviewing I did of staff at the state level and in the ten research counties. What respondents said about “what works” anticipated most of the statistical results. Moreover, my statistical findings run largely parallel to those of the experimental welfare work evaluations just mentioned. These also suggest that welfare work programs perform better when they enforce participation and focus on actual employment rather than long-term training. Such programs have more power to raise working hours than they have to raise the quality of jobs.42

Local Variations

•

105

ADMINISTRATIVE EXCELLENCE Academic policy experts tend to think that successful programs result mainly from good policy analysis and money, but in these respects Wisconsin was not exceptional. Policy analysis played a minor role at the state level in the early stages of reform, and the same was true at the local level. Officials acted more on the basis of experience and trial-and-error than on deeply laid plans. And mistakes were made. With the apparent exception of Grant, every one of these counties had to redesign its work program at some point, either because it failed to perform, or contractors failed, or due to outside pressures. Managers “groped along.”43 As for money, its importance was clear but easy to exaggerate. Wisconsin social programs are generally well funded. In Kenosha extra resources clearly helped the county implement its innovations. But much of what all the counties achieved came through restructuring, rather than by expansion or the hiring of new staff. Much of the staff increase needed to enforce work was “funded” through the caseload decline. This shrank the caseloads of the existing staff, allowing them to “ratchet” the caseload down simply by giving attention to relatively more cases.44 What mattered more to the implementation, as in the legislation, of reform was governmental excellence. The keys to success at the state level were that elected leaders could agree about reform, and administrators could implement a complex series of changes. We see the same potent combination at the local level. Among the research counties, in every case but Milwaukee, elected leaders were able to agree on a basic policy, without serious divisions. And in every case but Milwaukee, serious operational problems did not impede programs from achieving that policy, or changing it as leaders deemed necessary. Localities also showed high-mindedness. When Tommy Thompson experimented with welfare, he depended on counties volunteering to pilot his programs. To do this brought political credit to a county, but it also required difficult changes in local expectations about aid and in welfare routines. Yet volunteer sites were never lacking. Because of its special requirements, WNW was the most demanding of the experiments. Yet no fewer than fifteen counties volunteered to pilot it.45 Most important, the leading counties were exemplars of strong organization. Researchers have described what effective welfare work programs look like. Managers must articulate a vision of a new policy and then recruit and organize staff who believe in it. That vision will be some version of the idea of social contract—serving clients in new ways, but also holding them to “high expectations.” Recipients will be helped, but they must also help themselves, above all by participating steadily. Managers must trust staff with discretion about how to achieve program goals, but

106

•

Chapter 5

also hold them accountable for key outcomes such as job placements. The staff in turn must hold clients responsible for staying involved and looking for work. Reliable information systems must report back on both staff and client compliance.46 The leading Wisconsin counties achieved all of this. The Kenosha managers and Jon Angeli in Grant created essentially new organizations to realize their ideas. Sheboygan purchased such an organization from Curtis. All these programs were suffused with the upbeat but demanding values of successful work programs. Kenosha’s watchwords were blazoned on its walls—“Expect Success,” “Believe in the Power of Work,” and “Unconditional Positive Regard.” The program aimed to be positive toward all its clients—and just for that reason to expect more from them. From a distance, such slogans can seem corny or empty. But by believing in them, the staff motivated both clients and themselves. The tough side of all these programs was that managers held everyone involved, from the clients on up, strictly accountable for results. The biggest difference among the counties was not in policy, but in how fully they claimed ownership of their programs. The low-performing counties were above all passive and detached. They went through the motions of a JOBS program imposed from above, without caring strongly about it. That effectively left it up to the recipients to decide whether to work, and little change resulted. In the leading counties, in contrast, managers were proactive. They strained to use welfare to generate positive social change. That energy, as much as money, drove them toward more demanding programs. The faith that effort was meaningful, that success lay within their grasp, was finally what motivated both staff and clients to achieve their goals. Perhaps even a program opposed to work first might have succeeded, had it been implemented with the same conviction. But the lagging counties showed nothing like the same conviction. The commitment to success overrode even failures and redirections along the way. State-level officials had shown fortitude through the long years of experimentation, prior to W-2. The leading and changing counties showed the same. They put their past behind them and turned, without looking back, to something better.47 When I asked Jean Rogers what distinguished the pathbreaking counties from the rest, she cited their willingness to experiment, but also their “tenacity” and “singleness of purpose.” I was reminded of Max Weber’s dictum that governing “takes both passion and perspective.” It is the “strong and slow boring of hard boards.”48

CHAPTER 6

The Emergence of W-2

THE LAST AND MOST dramatic stage of the Wisconsin reform was Wisconsin Works, or W-2. The earlier reforms had largely been incremental in character. Each new program built on earlier ones. Most instituted work requirements in some fuller way. While W-2 did that too, it also broke more decisively with AFDC. It had several features not found in the old welfare, or in TANF in any other state. It represented a vision of what a postreform aid system should be. Its enactment was somewhat more partisan than the earlier stages of reform, but consensus held. Somehow, Wisconsin achieved that rare thing—a major policy innovation cut virtually out of whole cloth. Below, I describe the unusual planning of W-2, the principals of the new system, and its enactment in 1995–96, including the politics. I then describe the new system in more detail, characterize what is unique about it, and assess it. W-2 was an impressive display of policymaking, but was it effective policymaking? Largely, I conclude that it was, although it was a bit too severe.

THE PLANNING OF W-2 The earlier stages of the Wisconsin reform, as described in chapter 2, had rested on limited policy analysis. Policymakers thought up and tried new programs, drawing on precedents in other states or the wisdom of the bureaucracy. That was enough as long as each reform was limited. The overall drift of reform was more important than any one program. But in 1993, the legislature called for the total replacement of AFDC. How to do that called for greater thought. The “Wisconsin Idea”—a close collaboration with the state university—had earlier been crucial to the state’s leadership of national social policy, as chapter 12 will show. In the 1980s, experts from the university’s Institute for Research on Poverty had conceived the state’s pioneering child support reforms. On welfare reform, however, such a partnership was impossible. As noted in chapter 3, most university academics opposed Tommy Thompson’s conservative brand of reform. They shared his desire to raise work levels on welfare, but they opposed enforcing work as a condition of aid. This reflected not only their greater liberalism but their

108

•

Chapter 6

belief that barriers in the society, such as lack of jobs or child care, largely accounted for low work levels and other difficulties among the poor. To the governor’s team, however, mere conditions never could explain social problems. Such associations were “totally useless” for policymaking, said Jason Turner, a senior welfare official, because IRP would “infer causality where there was none.”1 Rather, the administration blamed past failures on the permissive character of traditional welfare. So the “Wisconsin Idea” was maintained, but the administration turned for advice to outside sources. The instrument was the Hudson Institute, a moderate-to-conservative think tank headquartered in Indianapolis. Its president, Leslie Lenkowsky, was himself an expert on welfare. Hudson offered to fund a task force to think through the welfare redesign, and Thompson and DHSS Secretary Gerald Whitburn agreed. Hudson chose Andrew Bush, from the Republican Ways and Means Committee staff in Washington, to coordinate the project in Madison. Named in 1994, the task force was called the 99 Group, because AFDC was to be replaced by January 1999. It included Bush but was headed by Jason Turner, the manager of welfare reform for DHSS. Other members included Jean Rogers, chief of Economic Support (the welfare division of DHSS), and several of her aides—Shannon Christian, Laura Kaye, and John Wagner. Christian had come from the U.S. Department of Labor under Elizabeth Dole, while Wagner had previously worked for the Wisconsin Policy Research Institute, a conservative think tank. The team also included two assistants to James Klauser, a close aide to Thompson who headed the Department of Administration with responsibility for state budgeting. Thus, the group had some independence from day-to-day government, but it also had close ties to DHSS and the Thompson statehouse. All members contributed to the plan that emerged, but Turner and Rogers played the pivotal roles. Jason Turner had been fascinated by welfare since childhood. He had served in the Reagan Administration in the Department of Housing and Urban Development, then in the Bush Administration as head of the Office of Family Assistance, in charge of implementing JOBS. Since joining DHSS in 1993, he had helped Whitburn and Rogers plan and implement reform in Wisconsin.2 Rogers came out of an equally long career in Republican politics and appointive positions within Wisconsin. A close ally of Thompson and a forceful manager, she headed welfare for close to a decade, from 1991 to 2000. The group was deeply knowledgeable about welfare, but in a very different way from most academics. The IRP researchers were adept at describing the poor or welfare populations using data analysis, but had little direct knowledge of welfare as an institution. Except for IRP’s child support experts, they seldom visited welfare offices or talked to people who worked there. This detachment reflected the incentives of academia, in

Emergence of W-2

•

109

which methodological sophistication counts for more than low-tech investigation. In contrast, the 99 Group members read research but were more expert in the operations of welfare—both the politics and the administration of programs. They sought to recast welfare, not simply as policy—as the academics would have done—but as an institution. THE PRINCIPLES OF W-2 The new policy design, Turner told me, aimed “to change the nature of the system.”3 W-2 was not based directly on any of the earlier reform programs, not even Work Not Welfare, which is often seen as a pilot for it. Rather, the first step—rare in policy analysis—was a return to first principles. The following were the key benchmarks, to judge from interviews with leading group members and the plan that they produced. AN END TO ENTITLEMENT

The most critical decision was to overturn the principle of entitlement. Welfare, that is, would no longer be given simply on the basis of economic need, without behavioral conditions. If you were needy, you would no longer be guaranteed income, Andy Bush said. Rather, you would get “the means to support your family,” chiefly through working.4 W-2 also ended entitlement in the legal or budgetary sense, as PRWORA was to do at the federal level. Applicants no longer had a legal right to benefits simply because they met the eligibility rules. Poverty lawyers had construed welfare as a “property right” that could be withdrawn only with elaborate due process, Jason Turner noted, and that was “deadly” to any idea of conditioning aid on work.5 AN EMPLOYMENT FOCUS

The best way to help parents meet their obligations was to help them work. “W-2,” Bush wrote, “places the goal of consistent attachment to the workforce above all others.” The very name evokes the form testifying to earnings that workers submit with their federal tax returns. Work would become a virtual precondition of aid, not something expected of recipients only after they received benefits. The new plan sought to sustain and promote work throughout a family’s time on the rolls. Rather than give aid because people were unable to work, it would help parents “work through crises so they can remain self-supporting.” W-2 would be part of a “broader workforce assistance” system designed to do this for the population as a whole.6 Of course, work levels are difficult to measure directly, particularly for people not on welfare. In practice, the planners took caseload trends as

110

•

Chapter 6

their best marker of success or failure. On the other hand, the plan posed no challenge to the principle of aid. It was not antigovernment. Its leading goals did not include quality control (keeping the ineligible off the rolls) or saving money.7 Indeed, as shown below, W-2 was expected to cost more money than AFDC. The services it offered to families outside cash aid would actually expand government’s responsibility for the working poor. In the old system, only people on welfare received aid, while only people off welfare had to work. The latter were much more numerous than the former. The new system reversed that. Now assistance in various forms was to be offered to the entire low-income population—child care, health care, wage subsidies—but it was work connected. The main task of the welfare system was no longer aid but rather employment—moving the adults who were not working into jobs. Then they along with other low-income workers would qualify for the broader subsidy system.8 THE WORK STANDARD

The W-2 plan opens with a listing of principles, all of which follow from the basic decision to guarantee work rather than aid. The most important was the first—“For those who can work, only work should pay.” Getting aid should be as much like working as possible. This argued against allowing recipients to train for better jobs while on aid, or against giving them special work incentives to “make work pay” if they went to work, because these benefits were not normally available to workers off welfare. Wisconsin would be one of the few states to deny such subsidies.9 Rather, recipients were to be treated the same as “low-income families who work for a living.” More might be done for them only if more was done for all other low-paid workers.10 INDIVIDUALISM

W-2 is individualist in spirit. The focus is squarely on needy adults, and only they, strictly speaking, are participants in W-2. Adults have to have children to be eligible; the single needy are not covered. Yet aid is not directed to mothers as such; the plan treats fathers and mothers equally. Nor is it aimed at children, as if what their parents contributed were secondary, as AFDC had assumed. The plan mentions supporting the family and community, but these goals are secondary. W-2 covers two-parent as well as single-parent families and seeks to enforce child support, but it makes no wider attempt to enforce marriage or deter illegitimacy, as some conservatives favor. Instead, the plan focuses closely on helping needy parents—married or not—support their families, chiefly through working. THE LIGHT TOUCH

Another implication was that contact with cash aid should be minimized. AFDC had made the “critical mistake,” Bush wrote, of making

Emergence of W-2

•

111

families dependent on a wide range of benefits on an open-ended basis, simply because they fell below the need standard at a moment in time.11 Rather, families should be helped if possible without putting them on the rolls long-term. The planners had already seen, with Work First, that merely asking eligibles about how they might provide for themselves diverted many from applying for aid. W-2 routinized diversion, and it aimed at the “light touch.” Applicants would be given only the aid they really needed. Those in short-term need could receive “bridge loans” to tide them over, without going on the rolls.12 THE FOUR TIERS

If aid was unavoidable, it should be given in the way least likely to promote long-term dependency. It was assumed initially that everyone could and should work. Applicants were not to be assessed for employability or shunted into remediation at the outset. Everyone would look for work, and the preferred option was to obtain unsubsidized jobs, obviating any need for cash aid. Only those who could not do this were subsidized, at one of several levels.13 The higher tiers would always earn more than the lower, so that participants had an incentive to move up to private jobs if they could. Only the completely incapacitated would be referred to disability programs. THE FLAT GRANT

Cash aid would be paid to recipients only on the lower two tiers, where clients were assigned to community service jobs (CSJ) or remediation programs. A distinctive feature of the new plan was that grants varied by tier but not by family size. Participants in community service jobs received higher benefits than those not working. But on either level, they received the same grant regardless of the number of children in the family, whereas in AFDC larger families had received more. The rationale was not the conservative idea of a family cap—denying more benefits to children born on the rolls, to deter mothers from having them. Rather, it was the ideal of making the welfare experience as much like working as possible. Just as regular workers did not receive higher pay when they enlarged their families, so W-2 participants should not either. CHILD SUPPORT PASS THROUGH

For the same reason, W-2 participants were to receive any child support paid to the family, without any reductions in their grants. Under national AFDC policy until PRWORA, welfare mothers whose absent partners paid child support received only $50 of it per month, with the rest going to reimburse government for the costs of welfare. Under PRWORA, states were allowed to rescind even this, and most did so, so that a welfare family now gained nothing from child support. Absent parents then

112

•

Chapter 6

had a reason to avoid paying support through the formal system, or to help the family surreptitiously, lest all the money go to the government. Those incentives made enforcing paternity and the payment of support very difficult.14 With W-2, Wisconsin chose rather to pay all support to the family, even though revenue was lost to government. No other state has done this. The thinking was partly to change the incentives, so that more fathers and mothers would cooperate with support enforcement. Principally, however, the rationale was again the work standard. Under W-2, participants would already be “earning” their benefits through work or other mandatory effort. They should then receive child support without reductions, just as working families outside welfare did.15 SUPPORT SERVICES

W-2 levied severe and immediate demands on recipients to work for any aid their received, but at the same time it obliged government to support them in working as never before. The plan called for a large expansion of child care, health care, and transportation services. Enhanced subsidies would help families afford child care, to which they would contribute co-payments. The novelty was that all these services were to be provided, not only to current welfare recipients, but to the entire lowincome working population. The principle that welfare recipients and the working poor deserve the same treatment here justifies, not doing less for recipients, but doing more for other people.16 In the old welfare, people on aid could get some benefits not available otherwise, such as wage subsidies and guaranteed health care. Once they were on the rolls, they could keep some of these benefits even if their income rose above the welfare level, indeed even if they went to work and left welfare. This privileged them over comparable families who never went on welfare. W-2 largely abolishes this “dip-down.” The working poor can now get support services and wage subsidies as good or better than those available to cash aid recipients, but without going on cash aid. IMPLEMENTATION

Radical in its provisions, the W-2 plan was also radical in its administrative arrangements. The functions of eligibility determination and employment, previously separate in welfare, would be combined in a new position. The Financial and Employment Planner (FEP) would control both the eligibility and assignments of cases. And welfare agencies, like recipients, were to be denied any entitlement to welfare. Local welfare departments would have to compete with private entities for the right to run W-2. The agencies chosen, public or private, would be held account-

Emergence of W-2

•

113

able through performance contracting. Furthermore, W-2 agencies would usually operate out of job centers alongside other programs, so that all benefits and services would be readily available to clients. In short, welfare would be reinvented, and so would government.17 The planners did honor the traditional, economic goals of welfare. Participants who worked and took full advantage of the new system would emerge better off, they promised. Through W-2, a mother with three children could escape poverty on W-2 provided she worked full-time—even at the minimum wage—and claimed other benefits such as food stamps and the federal- and state-earned income tax credits (EITC).18 Yet the chief goals of the plan were civic rather than economic. The aim was to redefine social citizenship in ways that would promote social cohesion, rather than undercutting it as AFDC had done. To this end, work, even in unpaid jobs, was not to be rejected as punitive, nor to be guaranteed as an entitlement. It was rather an obligation that, when met, conferred belonging in the community. In “urban areas such as Milwaukee,” the plan noted, citizens worried about “the apparent disorder and lack of control over public space.” Recipients in community jobs could help improve this “public environment,” thus “creating in the minds of the residents a joining of public purpose with the provision of income support through work for [the] needy but worthy.”19 The very idea reflects the earnestness, and the innocence, of Wisconsin’s public culture. The plan’s innovations sprang from the core idea that W-2 was to assure people the opportunity to support their families, and only that. By the spring of 1995, the conception had gelled. With input from other group members, Turner drafted the plan and submitted it to Whitburn and Thompson. Both were surprised at how far-reaching it was, but they embraced it.20 In August, the governor announced it, and a detailed version went to the legislature.

THE ENACTMENT OF W-2 Remarkably, the legislature changed very little. The planners began the process of persuasion while the W-2 design was still in formation. To prevent the legislature crafting a reform on its own, Whitburn in May 1994 formed a “Bi-Partisan Legislative Working Group” on welfare reform. It included key legislators such as the Republican John Gard and the Democrats Antonio Riley and Gwendolynne Moore. Over the summer of 1994, this body held “listening sessions” in Madison and Oshkosh at which experts, group representatives, welfare administrators, recipients, and others spoke.21

114

•

Chapter 6

The plan leaked to the press in July 1995, prior to its announcement by Thompson.22 The leakers were probably opponents who hoped to stir up resistance to the radical features of the plan. But remarkably, little political damage resulted. The plan, Andy Bush remarked, “became this powerful idea” that nothing could derail.23 By the time the legislation passed in March 1996, members had changed only these aspects: BUREAUCRATIC COMPETITION

Welfare agencies and public service unions were alarmed by the opening of W-2 to bureaucratic competition. Loss of the program would threaten their revenues and jobs. In response, legislators created a “right of first selection” (RFS). This assured county governments and agencies of their choice (usually welfare departments) of becoming W-2 operators provided they met several conditions, to be specified by DHSS. On this basis, the vast majority of counties did qualify to run W-2. The administration framed the criteria to exclude Milwaukee, however, as this was where it believed the greatest operational problems lay.24 W-2 in Milwaukee would finally be contracted out to private agencies, as described further in chapter 7. CHILD CARE

The W-2 plan sharply expanded public subsidies for child care, but it also demanded that recipients of these subsidies share more of the cost than previously. The co-pays aimed to spread the subsidy over a wider population than the earlier beneficiaries, centered on traditional welfare. But this meant higher costs for some families who had earlier received free care. It also created incentives for recipients to prefer cheaper, less regulated forms of care. This affronted the views of social service groups that regulated care is better. In response, the legislature limited co-pays to 16 percent of beneficiaries’ gross income, still higher than many recipients had paid previously. The drawback was a sharper “notch” (increase in charges) when a family’s income finally rose above the W-2 ceiling. The approval of a new, less regulated category of subsidized care was also disputed, but passed without change. HEALTH CARE

As in child care, the W-2 plan would have extended health subsidies to a broader population than before, but at the price of cutting benefits or raising co-pays for some groups already receiving care, principally from Medicaid. Wisconsin would have needed a federal waiver to do this, and after protests by client groups and providers, Washington refused it in December 1996. But later, as explained below, further federal funding permitted creating a satisfactory health program after all.

Emergence of W-2

•

115

EMERGENCY ASSISTANCE

An emergency assistance program had been part of AFDC and would have died with it. The legislature chose to continue it, to provide temporary help particularly to homeless families. Its preservation strengthened the safety net available to those poor who might refuse to cooperate with the new W-2. The program was slightly liberalized in 1999.25 None of these changes altered the essence of W-2, although they did increase its cost by some $119 million over two years.26 The core of the plan—the rebuilding of aid around work opportunity rather than entitlement—went unchallenged. The process of amendment and adjustment to specific interests did not deny the outcome all coherence, as is common in American politics. Jason Turner had thought the 99 Group would get only half its plan through the legislature. Jean Rogers was more hopeful, and she was right. They got over 90 percent through. “For a concept to end up on the street,” Turner exclaimed, “is almost unheard of!”27 THE POLITICS OF PASSAGE How could it happen? As in chapter 3, a partisan interpretation conveys only part of the truth. The immediate reason was that Republicans won a majority in both houses in the 1994 elections, when Tommy Thompson was also reelected for a third term in the statehouse. Any welfare reform he proposed was sure to command Republican support in the legislature. With the GOP in control, fundamental challenges to the Thompson agenda became improbable. The new majority was to enact Self-Sufficiency First and Pay for Performance—really preludes to W-2—in 1995 and then W-2 itself in 1996. Had Democrats kept control of even one house in Madison, the pace of change would surely have slowed. W-2 would have been more moderate, providing more cash aid to the employable and allowing more education and training in place of work. It might have guaranteed jobs, as W-2 technically did not. Most Democrats probably did think that W-2 was too severe, and its enactment was less agreed on than most of the earlier reform programs. Yet a majority of both parties still accepted the essence of reform. Both had given Thompson and his planners a mandate to redesign AFDC. Majorities of both parties had also agreed about the essence of change—to abandon entitlement and rebuild aid around work, buttressed by new support services. Disagreements were largely confined to the exact severity of the work test and the generosity of the services.28 Nothing during the enactment shook this core agreement, although vociferous disputes emerged on some issues.

116

•

Chapter 6

As before, a few Democrats opposed the reform root and branch. To Senator Gwendolynne Moore, W-2 was “involuntary serfdom, a new form of slavery”; it would lead to “driving a deeper and deeper wedge between races.” 29 But other prominent Democrats endorsed the plan. Mayor Norquist pronounced it “excellent,” regretting only that it did not cover single people as well as families. The latter also was the concern of Representative Shirley Krug, another Milwaukee Democrat.30 Norquist and David Riemer kept their eye on their goal, which was to replace AFDC with a work-based system; achieving that mattered more to them than defeating Thompson. While they wanted CSJs to pay regular wages, they accepted the conservative tenet that aid should be given only for employment. Representative Antonio Riley characterized the bill as a good start that “needs a lot of work.” He hoped to accomplish that “in a bipartisan way.” He worried that the plan contained too little money for education and training, and he doubted that enough jobs would be available for all those expected to work.31 In the end, W-2 was too severe for Riley, and he voted against the final bill. But he still told me later that “Parts of W-2 are light years better than the old system.” “We cannot go back to AFDC as we knew it at the time that we voted to replace it,” he averred, “because more welfare recipients at that time were not working.”32 For Riley, as for most politicians of either party, raising work levels was indispensable, and any measure that failed to do this was not true reform. Opponents accepted that aid would now be conditioned on a serious work obligation; they simply wanted a less onerous test than Republicans. The administration conceded enough to win over most legislators. For instance, it added to W-2 an Employment Skills Advancement Program (ESAP), which gave small grants to workers for training, and it allowed 10,000 students already in vocational or technical training to complete their programs under the new system.33 Above all, the Republicans allowed the changes to agency selection and support service mentioned above, even though this raised costs. In short, they sweetened the pot.34 This was sufficient to obtain passage precisely because few legislators raised issues of principle. In the end, W-2 gained a large, bipartisan majority in both houses. The final vote in the Assembly was 73–25 with 22 out of 47 Democrats in favor. In the Senate, it was 27–26, with 10 out of 16 Democrats in favor. The consistent opponents, led by Senators Moore and Gary George, were isolated. Representative Barbara Notestein, an observant Democrat, admitted that much of the support for change came from her own party.35 And despite continuing friction over some points, Democrats never seriously questioned the basics of reform thereafter. When Ed Garvey, the

Emergence of W-2

•

117

next Democratic candidate for governor, ran against Tommy Thompson in 1998, he also questioned only details.36 One might call this simply political prudence. Most Democrats, admitted Assembly Minority Leader Walter Kunicki, were simply afraid to go on record against “welfare reform.”37 But, like the earlier stages of change, W-2 was a process to which both parties had contributed. Despite the jousting for advantage, both had entered sincerely into the struggle to redesign welfare. That quest had already separated most Democrats from rear-guard defense of the old system, even though Republicans got to design the new one. The enactment of W-2 was the triumph of an idea over the narrower forms of partisanship.

OUTSIDE GROUPS Outside government, the story was much the same. Resistance centered more on the details than the fundamentals of the plan. Opponents judged that passage could not be stopped, but opposition from community groups was also half-hearted. Demonstrations against passage did occur, but not on the scale seen during the confrontations over welfare in 1969. In Milwaukee, there were angry meetings and a march of 500 people led by Senator Moore. In Madison, twenty-five welfare mothers led by Representative Tammy Baldwin (Democrat of Madison) staged a demonstration at the Capitol.38 But officeholders did not respond as they had before. The chief opposition called itself the Policy Group on Welfare Reform. It included the Wisconsin Council on Children and Families, the Institute for Wisconsin’s Future, and other community groups. It asked that participants in community as well as regular jobs receive the minimum wage and EITC, that participation be optional for mothers until children were two years old, and that co-payments for health and child care not start until a family’s income reached the poverty line.39 But other voices supported the plan. Julia Taylor, the executive director of the YWCA called W-2 “a tremendous opportunity. It’s a once-in-a-lifetime possibility to make families self-sufficient and rebuild the central city.”40 The community groups were weakened by a lack of powerful allies. Not only were Mayor Norquist and David Riemer backing W-2, but they had involved a number of community groups and leaders in the New Hope Project, the business-financed experiment in work-based reform (see chapters 3 and 8), which in many ways paralleled W-2. One could hardly support a liberal program that based all aid on work and then oppose a reform with many similarities simply because it came from a Republican governor. The Policy Group failed to ally with the

118

•

Chapter 6

academic opponents of W-2, and this cost it intellectual and analytic credibility.41 The opposition was weakened, as well, because W-2 perpetuated the “welfare industry” often criticized by conservatives. Since the program demanded work rather than simply denying aid, it required that many services—child care, job placement, government jobs—be delivered to the caseload. Money was to be made doing this, and some of the community groups that might have opposed reform instead became contractors to W-2. These included the private organizations (one of them the YWCA) that ended up running the new system in Milwaukee. Some of the potential opposition was thus co-opted. Said Pat DeLessio, “You’ve got African-American agencies that bought into this program.” Said Gwendolynne Moore, “This is a plantation,” and “The would-be activists are being paid to be the overseers.”42 But aside from self-interest, most community organizations were no longer willing to defend the moral core of the old welfare system, which was entitlement, or giving aid to the needy without behavioral conditions. Many service providers found that a work requirement was constructive in tackling entrenched poverty, even if they wanted the new system to be more forthcoming in various ways.43 A feature of the moralistic political style is that participants in political debates feel driven to focus on the common problem, not defend personal stakes. Just as it was widely seen as improper for Republican legislators to try to save money on welfare in 1969, when the poor seemed to need greater help, so now it was seen as improper to defend traditional welfare rights when the problem was to reverse excessive dependency. To defend entitlement in this setting required unusual arguments. The most serious came from the Roman Catholic Church. The Wisconsin Catholic Conference wrote a statement on reform that struck many liberal themes, such as the responsibility of “social and economic conditions” for poverty and the fear that an end to traditional welfare would be “devastating to children and families.” The core of the Catholic position, however, was the theological conviction that human beings were “sacred” simply because they were “created in the image . . . of God.” This implied that “the poor, especially children, have a moral claim on the resources of the community,” that they should receive aid even if their parents are undeserving.44 In this spirit, Rembert Weakland, the Catholic Archbishop of Milwaukee, attacked W-2 as “a tragedy for the poor and a moral blemish on the earth’s most affluent society.” But Thompson, although a Catholic himself, retorted that Weakland, who was studying in New York, should “come back to Wisconsin and read his Bible.” Thompson and his allies were undeterred because they sincerely believed that what they

Emergence of W-2

•

119

were doing would do more for the poor than the existing system. Why, the governor’s press secretary asked, should Catholics defend a system that “hurts the people they claim to protect.” Representative John Gard, who led the reform effort in the Assembly, told me that he, too, was a Catholic, but he had to tell the bishops that “our patience is wearing thin.”45 On the other side of the political spectrum, business did not raise objections to W-2 despite the millions of new dollars that the new system was expected to cost (see below). If anything, business leaders were more relaxed about reform than the administration. Some suggested that W-2 could usefully be less work first, more training oriented, than planned. Milwaukee leaders successfully lobbied Thompson and the federal government to provide more funding for New Hope.46

THE ENACTED SYSTEM W-2 was cleared by the legislature on March 13 and signed by Governor Thompson on April 25, 1996. The new system radically changed what Wisconsin expected of its adult welfare recipients—and what it promised to them in return. It was simultaneously demanding and generous, and in launching it, the state took a heroic fiscal gamble. Provisions of W-2 These were the principal provisions of W-2 as enacted. Some details, as I note, were later changed:47 ELIGIBILITY

W-2 offers cash aid to custodial parents with children who are at least eighteen and have gross income at or below 115 percent of the federal poverty line. They may be married, cohabiting, or single. The family may not have assets exceeding $2,500, excluding vehicles worth up to $10,000 and one home. The custodial parent must cooperate in establishing child support, and must not have refused a job within 180 days or be in receipt of Supplemental Security Income (SSI). If the family is on W-2, the noncustodial parent is also eligible for some services. Subsidized child and health care is offered to individuals in families under 165 percent of poverty (later changed to 185 percent), and once they receive it they remain eligible up to 200 percent of poverty.48 Subsidized child care is limited to parents in high school or in work or work programs. Health coverage excludes most people eligible for employerbased health benefits.

120

•

Chapter 6

THE FOUR TIERS

Most participants in W-2 are immediately placed in one of four work assignments, depending on their employability as judged by the Financial and Employment Planner: • Unsubsidized employment: The preference is that individuals find a regular job. They may receive, in addition to that income, wage subsidies (federal and state EITC), food stamps, child and health care, and case management. The work requirement is the hours demanded by the regular job. • Trial jobs: Individuals with more limited skills may get a regular job paying at least the minimum wage, but W-2 subsidizes the employer up to $300 a month. These clients also get federal and state EITC, food stamps, child and health care, and case management. The work requirement is expected to be forty hours a week. • Community service job (CSJ): Individuals not yet able to obtain a regular job are placed in an unpaid position arranged by W-2, in return for a cash grant. They also receive food stamps and support services but not wage subsidies. They are expected to work up to thirty hours a week and to engage in education or training for up to an additional ten hours. • W-2 transition: Individuals unable to work regularly will undertake some required work alongside remediation activities in return for a cash grant. They also receive food stamps and support services but not wage subsidies. They may be required to work up to twenty-eight hours a week and engage in education and training for up to twelve hours a week. CASH GRANTS

The cash stipend for participants in CSJs was originally $555 a month, raised in 1997 to $673. For participants in W-2 Transition, the grant was originally $518 a month, raised in 1997 to $628. Grants do not vary with family size. In addition, participants searching for private jobs can receive job access loans to tide them over, with a maximum of $1,600 over a year. Loans are supposed to support employment and be repaid within one or two years using a combination of cash and community service. TIME LIMITS

Placement on any of the bottom three (subsidized) tiers is limited to twenty-four months, with extensions on a discretionary basis where clients show that they are trying to find a job but are unable to. In addition, trial jobs are limited to three months in length each, and community service jobs are limited to six months each, with some extensions possible. W-2’s lifetime limit is sixty months, as under federal TANF law.

Emergence of W-2

•

121

SANCTIONS

Cash grants are reduced by the minimum wage (originally $4.25 an hour, currently $5.15 an hour) for each hour of required activities that the participant fails to attend without good cause. In addition, participants may receive “strikes” for showing a pattern of nonparticipation or noncooperation; receiving three strikes disqualifies a client from further eligibility for that tier. OTHER W-2 PLACEMENTS

Some W-2 eligibles may receive benefits without immediate placement on a work tier. These include caretakers of a child less than twelve weeks old (when the work requirement begins), custodial parents under eighteen, and pregnant women without born children. Only the first group receives a cash grant, the others case management. For eighteen- or nineteen-year-olds without a high school diploma, the obligation is to finish school or get a certificate of general education development (GED). Applicants claiming incapacity remain in W-2 but are referred to the Department of Vocational Rehabilitation for possible admission to federal disability programs, either Supplemental Security Income or Disability Insurance. PLACEMENTS OUTSIDE W-2

Wisconsin did not transition all of AFDC to W-2. Since the system was aimed at employable adults with children, it excluded families in which the caretaker was not on the grant or not employable. These cases were transferred to two parallel programs outside W-2, run by the Department of Health and Family Services: • Kinship Care: Families where relatives of children have assumed responsibility for them, but are not legally responsible and are not themselves eligible for W-2, could receive $215 per child per month, subject to an assessment and background check. • Caretaker Supplement: Under legislation following W-2, caretakers receiving Supplemental Security Income as disabled, and hence not employable, could receive a supplement of $250 for the first child per month, and $150 for the second.

During the transition to W-2, some 11,000 cases transferred to these programs, which support around 12,000 cases today. This means that calculations that compare AFDC to W-2 somewhat overstate the decline in the state’s welfare rolls.49

122

•

Chapter 6

The Demands of W-2 The new system was severe because of its light touch, its denial of entitlement to aid, and above all its immediate and unyielding work test. In Wisconsin under AFDC, and in most states even under TANF, there was some grace period between going on aid and having to meet a work test. In W-2, one has to earn one’s wages or grant from the very first, with only loans to cover needs before the first check is cut. For parents who decline these terms, there is little recourse other than to scrape by without cash aid, or to turn their children over to the child welfare system.50 Activity demands are not only immediate but strong—as much as forty hours down to the lowest and least employable tier. The only concession to limited employability is that clients on the bottom two tiers do not have to work in regular jobs. Their obligation is more to become employable than to work in the usual sense, but their activities are still extensive and mandatory. The year after W-2’s enactment, a work requirement was extended to the second parent in two-parent cases claiming federally funded child care, in order to satisfy TANF requirements. The combined hours of the two parents were set at fifty-five a week.51 Notably, W-2 lacks work incentives in the usual sense. Many other states increased the amount welfare recipients could earn without large reductions from their grant, the better to “make work pay.” This was supposed to give recipients a leg up in moving into the workforce, by allowing them to draw income from both work and welfare during the transition.52 W-2 has no such incentives because participants are not usually permitted to combine wages with cash grants. They get only wages on the top tier, only cash grants on the bottom two tiers.53 In trial jobs, wages are subsidized, but, as events proved, very few participants ended up on this tier. So for most working clients, work incentives are limited to the federal and state EITCs. The reason for this might have been to avoid the disincentives to move up to higher incomes that work disregards can create as the subsidies are withdrawn at higher earnings.54 But the fundamental reason was the idea of taking regular low-wage employment as the benchmark for the new aid system: Workers on W-2 were not to get a better deal than workers off the rolls. W-2 also allowed little room for education or training. The goal was to move all participants into regular jobs, at whatever wage, without first studying or training for “better” positions, as JOBS had allowed. Again, the conviction was that this was more effective, and also that it was fairer to the working poor. The Employment Skills Advancement Program was highly restrictive. To qualify for grants, claimants normally had to be working forty hours a week, and to have worked for nine months in an

Emergence of W-2

•

123

unsubsidized job.55 Participants could of course enroll in training programs outside W-2, but like other low-skilled workers, they would do so using their own time and resources. The Generosity of W-2 Although the new system demanded that all employable adults work, it also supported work on a lavish scale. Support services were sharply extended and their funding increased to an extent never before seen in the Badger State. In principle, Wisconsin was already generous with child care. Subsidized care was available to AFDC recipients and to other families below 75 percent of the median income, or approximately $28,000 for a family of three, which was 222 percent of the federal poverty line. But the state had not actually funded enough care to meet all the demand.56 Pressure on the system had increased because of a court decision in 1993 mandating that child care funding be provided in JOBS as an entitlement—that is regardless of available funds.57 With W-2, earlier child care programs were consolidated and a more serious commitment was made. Whereas eligibility had earlier been tied in some way to welfare, it could now extend to 200 percent of poverty without any necessary tie to cash aid. That level took in a large share of the working poor and near poor. The goal was to meet all demand so defined. This was accomplished partly by raising co-pays, so that recipients did more to help pay for care, and by creating a new category of less regulated care, but also by massive increases in state funding. Whereas $90 million had been budgeted for child care in 1996–97, this amount soared to $158 million in 1997–98 and $180 million in 1998–99, or double the earlier figure.58 In health as in child care, Wisconsin had already been generous. The state’s Medicaid program was among the most expansive in eligibility and services in the nation, covering some groups above the welfare level. As mentioned above, the W-2 health program attempted to trim initial eligibility slightly and impose some co-pays in order to extend coverage to the broader working poor, but a federal waiver was denied. Then in 1997 more federal funding became available through the State Child Health Insurance Program (SCHIP). Wisconsin was then able to negotiate federal approval for BadgerCare, a new health program that provides the broad coverage originally planned for W-2, using Medicaid, SCHIP, and state monies. In the new system, which began in 1999, there are no charges to recipients below 150 percent of poverty, and co-pays above this are limited to 3 percent of family income.59 Despite the new

124

•

Chapter 6

commitments, the state expected to save money overall on health care. This was because W-2 would shift many Medicaid recipients off welfare and into private jobs, where they would be covered by employer health insurance. As for transportation, W-2 agencies are expected to provide this service out of their general W-2 funding. The state, however, also spends $2 million annually on special transportation projects to assist W-2 participants.60 Even the conservatives voting for W-2 thought that they had to guarantee such services in order to remove the excuse that welfare recipients could not work for lack of support. Observed Representative John Gard, point man for W-2 in the Assembly, “Individuals who seek assistance always have said if child care and health care are available, they’ll go to work. W-2 delivers on both of those.”61 It was another expression of the paradox that reducing dependency in some ways required larger government rather than smaller. The Cost of W-2 But could the state actually afford to do this? Tommy Thompson had warned his state that reforming welfare would cost more than leaving it unchanged.62 Table 6.1 shows the major expenditures anticipated for W-2. Over the first two years, from 1997 to 1999, it would cost $269 million a year just to administer the new system. The state would spend considerably larger sums on subsidized employment, child care, and health care, and it would spend almost $130 million on the parallel kinship care and SSI programs. All told, the cost was well over $1 billion a year. To meet it, the state could call on $372 million in federal funds, but it would have to defray the rest itself.63 Of course, the state was already spending large sums on welfare. During passage, the administration contended that W-2 need cost no more than this. In fact, the Legislative Fiscal Bureau, an arm of the legislature, estimated that it would cost $66 million a year more than the old system. The estimates assumed a falling caseload where lower costs for grant payments would offset the new investments in support services. With the rejection of the health care waiver, however, health costs were projected to rise, not fall as originally planned.64 But more significant, costs were highly uncertain. There was doubt about the future prices of child and health care. Above all, there was doubt about how many of the existing recipients would transition to the new system, and how many eligibles would claim the new benefits in future. The potential claimants included much of low-income Wisconsin.

Emergence of W-2

•

125

TABLE 6.1 Estimated Spending on W-2 during 1997–99 (millions)

State administration Local administration Subsidized employment Child care Health care Kinship care Children on SSI Other Total

1997–98

1998–99

Total

18.8 124.3 228.3 158.5 445.5 38.0 24.5 25.4

18.6 106.8 181.3 180.2 475.2 41.2 24.5 15.2

37.4 231.1 409.6 338.7 920.7 79.2 49.0 40.6

1,063.3

1,043.0

2,106.3

Source: Legislative Fiscal Bureau, Wisconsin Works: W-2, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1997), p. 62.

Perversely, policymakers felt they had to extend coverage precisely to get control of the welfare problem. In order to undo entitlement in some senses, they had to embrace it in another. They had to agree to pay whatever it took to assure all poor adults the chance to support their families. The state had made large new commitments essentially on faith.65 The risks were serious because the state was vulnerable to unexpected costs. Wisconsin had helped fund its welfare experiments with its special “waiver savings” deal with Washington. With this plus regular federal monies, the state had defrayed 64 percent of its record employment and training outlays in 1995. But with the enactment of PRWORA in 1996, this deal came to an end. Wisconsin would now get only its regular TANF block grant, and any spending above this would come entirely out of state pockets. On top of this, Governor Thompson had agreed in 1995, under legislative pressure, to shift over $1 billion in school costs from local property taxes to state funding, creating shortfalls in the rest of the budget. The collision between this promise and other needs created serious budgetary strain. By 2001, the state faced a budget deficit of $3.7 billion over two years.66 That the crunch was not worse was due to the dramatic decline in the caseload, which was far greater than anyone anticipated. Among other savings, this allowed Wisconsin to transfer some of its federal TANF money from cash grants to child care, where costs rose rapidly.67 The state more than won its gamble. It actually saved money on welfare reform (see chapter 10). But no one knew this in 1995–96 when the key commitments were made.

126

•

Chapter 6

LATER CHANGES Even with the concessions made during enactment, W-2 was less consensual than the earlier stages of reform. This led to ongoing efforts by Democrats to modify some aspects of the plan, although again without touching its fundamentals. Their efforts were emboldened when, in 1996 following the enactment of W-2, they regained control of the state senate. In response, the administration agreed to some minor liberalizations. The most important change was the benefit increase that Governor Thompson proposed, and the legislature accepted, in 1997, prior to the implementation of W-2. Grants on the lower two tiers of W-2 were raised to $673 (for CSJ positions) and $628 (for W-2 Transition). Some of Thompson’s aides regretted this step, but it was consistent with the administration’s big-government style of conservatism: Dependency was to be controlled through work requirements rather than low benefits. Thompson, however, vetoed a provision added by the legislature requiring that the state pay assistance to the second parent in two-parent cases once both parents were obligated to work under TANF rules.68 Child care was also liberalized. There was continuing unhappiness about the cost of care imposed on the recipients, despite the 16 percent cap set on how much of a family’s income could go to co-pays. An easing was discussed in the 1997 budget cycle. In 1999, Thompson himself proposed reducing the cap to 12 percent, and this was enacted. At the same time, the income ceiling for initial eligibility for child care was raised from 165 to 185 percent of poverty.69 The child care system is now called Wisconsin Shares. For some time, as noted in chapter 3, Milwaukee’s John Norquist, his advisor David Riemer, and several Democratic legislators had favored replacing welfare with guaranteed jobs. But they wanted these jobs to pay regular wages and benefits, rather than just welfare as W-2 provided, in part because the recipients would then qualify for EITC. Norquist relaunched this proposal just as the W-2 plan emerged, and Riemer argued for it in the national press.70 Once Democrats strengthened their hand in Madison, they tried to amend the rules for CSJs accordingly. This, however, would have given the positions much the same prerogatives as private-sector employment, violating the plan’s premise that regular jobs be favored over subsidized ones. The administration resisted. The compromise was, starting in 1999, to allow two of the Milwaukee W-2 agencies to offer wage-paying jobs, and to allow other W-2 agencies to offer them for up to fifteen hours a week.71 The most important pressures, however, arose over W-2’s severe workfirst emphasis. The program sharply restricted the ability of recipients to be supported while going to school rather than working, as they had done

Emergence of W-2

•

127

routinely in the 1970s and 1980s. ESAP, as mentioned, was highly restrictive (it was repealed in 2001 to save money). After W-2 was on the books, community groups and some Democratic politicians pressed to allow up to fifteen hours of education or training to count toward the thirty-hour work requirement required of recipients in CSJs. One argument for this was that the remaining recipients had more skill limitations than those who had left the rolls. Pressure also came from technical colleges that had suffered declining enrollments due to the work first policy. The legislature added the fifteen-hour provision to the 1997–99 budget, and Thompson vetoed it.72 However, in 1998 he named a special committee to consider the issue in more depth. It included several state legislators as well as representatives from community groups and technical colleges. Its report, in the end, recommended only minor changes to W-2. It stressed rather the chances that W-2 already offered to pursue higher skills alongside working. It reaffirmed the state’s decision no longer to allow remediation in place of work.73 The next year, the budget for 1999–2001 liberalized access to training under W-2 slightly. Under a new Workforce Attachment and Advancement Program, recipients could enroll in technical college courses for up to fifteen hours as part of their work requirement, but only if they also worked and their FEPs judged the activity necessary for them to get regular employment. Thompson vetoed more lenient language.74 Administration officials realized that W-2 appeared severe to many, and they were concerned to moderate that image.75 At the same time, they managed to defend the original architecture of the system. This was not just because Tommy Thompson was still governor and wielding his veto pen (he was reelected for a fourth term in 1998). It was also because support for the essence of the new system—work tests and support services in place of entitlement—was too broad and deep to be challenged. THE UNIQUENESS OF W-2 All states changed their welfare systems in some ways following PRWORA, but the degree varied from the minimal (merely to conform to the act) to the radical. By common consent, Wisconsin’s reform was the most thoroughgoing in the nation. Comparisons with other states highlight the features of W-2 found nowhere else—the immediate work test, the flat grant, and the 100-percent pass-through of child support.76 But more important than novelty was the sharp focus of the new system on promoting work among needy parents. In the words of one study, “Wisconsin provides the most dramatic example of how the traditional service and benefit delivery structure can be changed to transform welfare

128

•

Chapter 6

Figure 6.1. Scores of states on work requirements and work supports, 1998. (Source: Committee for Economic Development, Welfare Reform and Beyond: Making Work Work, New York: Committee for Economic Development, 2000, chap. 2)

into a work-based system.”77 Toward this end, the state was willing to take expensive steps that other states avoided. One of these was providing public jobs, so that recipients could be required to work whether or not private jobs were available. Outside Wisconsin, only New York City has done this on a large scale. Another step was support services that observers find unusually generous by national standards.78 W-2 is simultaneously very conservative and very liberal. In one analysis of state rules under TANF, Wisconsin ranked second to Virginia for the most restrictive welfare system in the country, because of its severe work requirements. But it was also by a considerable margin the most generous state in the country because of its lavish eligibility and benefit standards.79 A study by the Committee for Economic Development scored the states on the degree to which they required work or limited eligibility, through work tests or time limits, and on the degree to which they supported work, through work incentives and support services. There was little association between these two indexes, yet Wisconsin had close to the highest rating on both of them. Figure 6.1, which is based on these data, arrays the states on these two dimensions, with the more extreme values labeled. Wisconsin stands out. It is more exacting about work than other affluent, northern states, but it is also far more generous with work supports than most southern or western states.

Emergence of W-2

•

129

Generally, states that pay high benefits take as their goal raising work levels, while those that pay low benefits usually prefer to reduce dependency. Tennessee is a case of a low-benefit state that has focused on raising work levels.80 Wisconsin is a state that, notwithstanding its high benefits and supports, also seeks to minimize dependency. The implication is not, as in the low-benefit South, that the state has only a limited antipoverty agenda. It is rather that Wisconsin wants to help the low-income population off cash welfare rather than on. In a sense, W-2 is narrow because of its strong focus on work. In another sense, however, the system is broad because it encompasses more than welfare in the traditional sense. While W-2 as such provides cash, jobs, and support services, state policymakers understand the system to include or to imply such other programs as Kinship Care and Caretaker Supplement, the federal and state EITCs, and child support enforcement. Wisconsin Shares and BadgerCare, while providing support services for W-2, are designed to serve a wider population off cash welfare. Observers say that few if any other states have such an expansive vision of what welfare reform might mean.81 The W-2 universe may grow yet further in future. The state planned to use money from the Welfare-to-Work program, a federal grant for broad employment purposes, to serve several groups on the edges of the traditional welfare population—criminal offenders, noncustodial parents, and immigrants.82 Tommy Thompson spoke of diverting federal welfare money that is unneeded for W-2 to such other uses as postemployment case management, education, early childhood development, and housing, provided federal rules permit.83 All this points to the ambitious vision of government that lies behind W-2. The protean character of W-2 also reflects the earlier concordat underlying the entire Wisconsin reform. Democrats had boldly abandoned entitlement, while Republicans had boldly accepted new financial commitments. In return, both gained a radical new system that served their major goals and transcended their differences. This was statecraft sufficient to refound the welfare state. It opened up new vistas for constructive social policy nationwide.

W-2: AN ASSESSMENT In the terms of table 1.3, the emergence of W-2 was a display of governmental excellence. As in the earlier stages of the Wisconsin reform, we witness a policymaking process in which the search for solutions dominated partisan rivalry. The new system was deeply laid, innovative, and unusually coherent. But was this also effective welfare policy? That is, did

130

•

Chapter 6

W-2 as enacted embody the best that was known about how to reform welfare at the time, at least given employment goals? I would say mostly yes, but not completely. W-2 was shaped by the evaluation literature as it stood in the mid– 1990s. The 99 Group knew that welfare work programs that enforced participation and demanded work first had outperformed those that were voluntary or training oriented. They also were swayed by Wisconsin’s success in moving in those same directions prior to 1995. In addition, the diversion programs, starting with Work First in 1994, had appeared to convert would-be aid applicants into workers with great efficiency, by persuading them to avoid welfare entirely. Accordingly, W-2 was built firmly on the bases of diversion, mandatory participation, and work first. Applicants were confronted first with their responsibility to avoid welfare if possible, then simultaneously to participate and work—if they wanted to get any aid. In all these ways W-2 embodied the best thinking of its time. In crafting W-2’s radical administrative arrangements, the planners had less information to go on, but they guessed right. Research was about to appear showing that vesting the eligibility and employment functions in the same case manager, as W-2 did, raised participation and reduced dependency by more than assigning them to separate staff, as welfare had traditionally done. Yet few localities have unified case management in the same way.84 The emphasis on diversion, however, was arguably excessive even in 1995. When Jason Turner first told me of the plan’s severe and immediate work demand, my reaction was that welfare was being turned into the Employment Service. That is, Wisconsin would no longer offer aid to the needy. It would instead offer them help in getting a job. The effect, I thought, would be to deter most of the current eligibles from even seeking aid. It would drive most recipients off the rolls, and so it did. It was then unclear how many of them would, in fact, work. By refusing to give aid to people up front, the system might lose much of its leverage over their behavior. The system assumed that the more employable would avoid welfare while the less employable would still apply for it. But with such strong diversion, the danger was that those most in need of aid would be scared away entirely.85 This would lead to confusion during the implementation period (chapter 7). The system would need to reach out to the most troubled families (chapter 8). In the event, the hardship created by W-2 turned out to be limited and its overall effects positive, including stunningly high work levels (chapter 10). Thus the planners were finally vindicated. But the new regime took a risk that a system with a softer front end might have avoided. “On the message side,” Andy Bush admitted, W-2 was “almost too successful.”86

Emergence of W-2

•

131

W-2’s emphasis on work first seemed well justified at the time of design. But, as noted in chapter 5, Kenosha and other counties already found the policy too extreme. The Kenosha strategy was to get people into the program immediately and supervise them closely—what I would call paternalism—rather than put them to work at once. Said Larry Jankowsky, one of the Kenosha managers, applicants needed some time to “adjust” to the idea of working. Kenosha’s pre-employment preparation had provided that.87 The county also permitted many recipients to work only parttime while they pursued training or education for better jobs. Later in the 1990s, further national evaluations supported Kenosha’s position. The highest-performing programs seemed to be those that had, not strict work first, but a general work focus coupled with some elements of education and training.88 The most outstanding of these experiments was a JOBS program in Portland, Oregon. It emphasized placing clients in “good” jobs, not just in any jobs they could find. It was tough about enforcing participation, but clients were often assigned to short-term training in order to enhance their chances for well-paying positions. Portland’s ability to obtain many full-time jobs with benefits, even for disadvantaged clients, makes it a promising model for the postreform era, when welfare’s chief challenge will be to raise incomes rather than work levels.89 Such a program would have reduced dependency more slowly than W-2, but with less risk of hardship and perhaps with better earnings results. In its severity, then, W-2 probably overshot by several notches—if service to its own clients is the criterion. To be fair, however, the new system also was concerned with effects off the caseload. It aimed to reaffirm the work obligation for the entire low-income population. Portland’s welfare caseload was 70 percent white and more educated than the average. Wisconsin’s by 1995 were over half nonwhite and strongly disadvantaged, raising issues of integration.90 The state faced the entrenched poverty of Milwaukee, much of it due to low employment. To change the signals in such a city required a program that unambiguously affirmed the norm of work in available jobs. To do that was, perhaps, even more important than maximizing impacts for the caseload. In the end, W-2 would focus first on work enforcement and then on paternalism. In its first incarnation, it primarily drove caseloads down, forcing work levels up mainly off the caseload. But after the first shock, the program shifted toward designing customized, Portland-like regimes of services and requirements for the more troubled cases. I describe those stages in the next two chapters.

CHAPTER 7

Implementing W-2

W-2 drew attention chiefly for its policy innovations, but the way it was implemented was equally significant. Wisconsin aimed, not just to carry out the new system, but to do so in an ambitious way that upgraded the performance of the bureaucracy. Many states have found it challenging to implement serious welfareto-work programs. It is not enough, experience has shown, merely to spend money on new services and work incentives meant to move recipients into jobs; the program must also be mandatory, making clear to clients that participation and work effort is a condition of aid.1 Conversely, it is not enough to levy a tough work test and then stint on support services.2 One must also have a bureaucracy able to make the work requirement real “on the ground,” by assisting and overseeing the individual clients.3 In some basic sense, Wisconsin had already crossed these Rubicons. By the time of W-2, there already was a serious work test in the Badger State, backed up by money and bureaucracy. As table 7.1 shows, in the decade between the peak of the welfare rolls in 1986 and the passage of W-2 in 1996, the reform process, coupled with good economic conditions, had already driven the rolls down by 37 percent. The quality of the bureaucracy had emboldened the state. Gerald Whitburn, Secretary of the Department of Health and Social Services, remarked to me that the Thompson administration felt free to demand work exactly because it could count on capable local administrators to prevent hardship.4 But the administration was still dissatisfied. Work requirements had not been fully implemented in the biggest cities, particularly Milwaukee. And throughout the state, W-2 planners wanted greater accountability from the welfare system. To achieve it, they turned to “reinventing government.” Reinvention says that public agencies should have to compete with private or nonprofit organizations to deliver public benefits. To the extent possible, they should be held accountable using performance measures rather than administrative procedures. That means funding agencies on the basis of results, not costs or persons served. Competitive and market pressures largely replace accountability through the political process, as in traditional bureaucracy.5 Accordingly, the state chose the agencies to run W-2 competitively, and it oversaw them largely on the basis of performance rather than proce-

Implementing W-2

•

133

TABLE 7.1 Welfare Caseloads and Caseload Change in Wisconsin, 1986–2001 April 1986

December March September March December 1994 1996 1997 1998 1998

Statewide 100,335 73,714 Over April 1986 (%) −27 Milwaukee 38,263 36,865 Over April 1986 (%) −4 Balance of state 62,072 36,849 Over April 1986 (%) −41 Milwaukee as percent of state caseload 38 50

August 2002

62,888 −37 33,718 −12 29,170 −53

31,476 −69 21,889 −43 9,587 −85

13,342 −87 11,445 −70 1,897 −97

9,078 −91 7,670 −80 1,408 −98

9,849 −90 7,846 −79 2,003 −97

54

70

86

84

80

Source: Calculated from caseload data from Department of Workforce Development. Note: Cases include AFDC or W-2 cash cases and exclude Kinship Care and Caretaker Supplement.

dural rules. This process led to the ouster of the public welfare department from cash aid administration in Milwaukee and several other counties. It was the most radical privatization of welfare seen anywhere in the country. Some critics see in this a retreat from public commitment to the poor.6 At least in Wisconsin, that view is mistaken. Welfare remained a public enterprise, with public officials firmly in control. The combination of radical policy change with restructuring did, however, present a vast challenge to implementers, especially in Milwaukee. Mistakes were made on a large scale. The state took risks with the administration of W-2, as it had earlier done in its design. The outcome, however, was still fortunate. The state emerged with a new aid system that promoted work and integration for the poor, as AFDC had not. PRELUDES TO W-2 Until 1994, work enforcement in Wisconsin had rested mainly on the developments in the JOBS program surveyed in chapter 4. The state had gradually tightened up the requirements for AFDC adults to participate in JOBS and to look for work. In Milwaukee, however, these stipulations largely remained formalities. Large numbers of recipients were inducted into JOBS on the computer, but no actual activities were expected of many of them. According to officials I interviewed, thousands of cases sat in “holding pools” for years. As explained in chapter 5, the city as yet lacked the inclination and the capacity to activate individual clients. “Expectations are not immediate,” one official told me, because there was no “immediate connection” between behavior and consequences. Especially, “Sanctions have not been

134

•

Chapter 7

implemented.” As one W-2 staff member later told me, “There wasn’t any case management being done then”; hence, there was no serious intention to “put you in employment.” This began to change in 1994 when the state forced the agencies running JOBS to compete for job placements. In 1995, it instituted financial incentives rewarding job placements, and in December of that year, Work First, the first of the diversion programs, was implemented in Milwaukee. Now the case managers began to call in inactive cases and try to mobilize them to work, but their caseloads also grew to 300 or more—too large to deal with cases individually. The next, and far more radical, step was the inauguration of SelfSufficiency First and Pay For Performance—statewide—in March 1996. These, it will be recalled from chapter 2, were the radical diversion programs that demanded job search from applicants before they could go on aid, and then demanded regular activities from recipients. These were the first programs that, as one Milwaukee manager put it, required “personally touching every case every month.” The agencies had to call in recipients and verify what they were doing. They had to hire more case managers. They strained to create community service positions (CSJs) so that recipients who did not find jobs could still fulfill their work obligations. With their diversion and tough work tests, SSF and PFP anticipated much of W-2. “The revolution is PFP,” the same manager told me. “It’s all over after that. If we can do PFP, W-2 will be a cakewalk.” Some observers doubted how real the new demands were. Many recipients were apparently assigned to CSJs on paper without actually entering jobs, as also happened with W-2.7 Nevertheless, the new programs drove engagement up and caseloads down. As table 7.1 shows, the statewide caseload fell by a further thirty-two points from March 1996 through September 1997, when W-2 began. In one Milwaukee job center, participation in JOBS jumped 36 percent and job entries climbed from 1,000 to 2,500 between 1995 and 1996. Other people were diverted; in one county, 46 percent of the applicants deemed likely to qualify for aid between March and July 1996 failed to do so. Most of the sanctions imposed were due to SSF and PFP.8 But of 5,182 sanctions issued in Milwaukee through August 20, 1996, 44 percent were later found to be improper. Either the agency had the facts wrong or its requirements had been met. Officials changed procedures to minimize the mistakes. Yet despite the confusion and the sharp fall in the rolls, there was little sign of hardship. AFDC recipients proved far more capable of coping without welfare than anyone imagined. Many simply disappeared. Observers speculated that many of them were already working “off the books.”9 These patterns of declining rolls, climbing work

Implementing W-2

•

135

levels, administrative errors, yet few apparent negative effects would be repeated with W-2. The state urged agencies in Milwaukee to implement SSF and PFP in ways that anticipated W-2. Recipients should put in a “simulated workweek,” just as if they were employed. The state actually instituted W-2 in March 1997 in Pierce and Fond du Lac Counties, which had earlier been the pilots for Work Not Welfare.10 These steps, however, did not save the big city from later traumas.

THE W-2 PROCESS In a restructuring of state government decided in 1995, the Department of Health and Social Services, which had run welfare since 1967, was divided. The Division of Economic Support, with responsibility for AFDC and welfare reform, was transferred to the Department of Industry, Labor and Human Relations (DILHR), which was then renamed the Department of Workforce Development (DWD). The remainder of DHSS became the Department of Health and Family Services (DHFS), with responsibility for health, disability, and child care licensing, among other programs. DILHR had been Wisconsin’s labor department, in charge of labor programs outside welfare—chiefly the Employment Service, or Job Service, and training efforts under the Job Training Partnership Act. The transfer of welfare to the new DWD, which took effect in July 1996, signaled that welfare was now to be principally a work program rather than an income maintenance operation.11 But there was little need, at least in Wisconsin, to convince the welfare department to stress work. It was already doing so, and the shift to DWD brought little operational change. The same managers and staff, led by Jean Rogers, continued to run Economic Support. More important, welfare and its work programs at the local level continued to be run by the same agencies as before, chiefly county welfare departments. While these departments sometimes used the ES or JTPA as contractors for JOBS, the labor programs had no assured role, and the shift to the DWD did not give them one. Thus, there was no need in most places to run welfare work through welfare and labor agenices jointly, a structure that has troubled reform in other states (see chapters 4 and 11). The greater change with W-2 was within local agencies. The new administrative process diverts applicants from the rolls if possible and then, if they go on aid, enforces work. On approaching the agency, applicants first encounter a receptionist, who may refer them immediately to resources outside W-2, such as computerized listings of available jobs. If

136

•

Chapter 7

they want information about W-2, they next meet with a “resource specialist,” the role invented by Grant County as part of its diversion strategy. The RS gathers eligibility information and inquires whether the family could support itself through friends and relatives or local charities, such as food banks. People who go forward to apply for W-2 must comply with child support requirements and can be asked to look for work for up to 30 days before receiving aid.12 If applicants seek W-2 and appear eligible, they go on to the Financial and Employment Planner, the linchpin of the W-2 system. Using the information gathered earlier, the FEP determines final eligibility for W-2 and places the client on one of W-2’s four employment tiers. An assignment to unsubsidized employment or trial jobs means that the family receives no cash aid, only noncash benefits (child and health care), case management, and, possibly, job access loans to tide them over until the first paycheck. Only community service and transition placements receive cash. FEPs are given total discretion to make this assignment. Applicants might appeal the decision to the agency, but until recently, even if they won they received no back benefits.13 As mentioned earlier, joining the eligibility and employment functions promotes higher participation and lower dependency than leaving the roles separate. Fusion also eliminates a long-standing tension within the welfare bureaucracy. Often case workers who deal with employment have looked down on eligibility staff as mere clerks with lesser qualifications, poisoning relations between them. With the roles joined, that problem faded. The new structure underlined the fact that cash aid was no longer an entitlement. A core reason for the earlier separation of eligibility from employment operations was the view of federal administrators in the 1960s that access to income should rest on economic criteria alone, and not on the client’s receptivity to advice from the agency. As long as that separation held, applicants could claim a right to aid simply because they were needy, and the agency could demand work or other activity only afterward. In W-2, applicants have a right to ignore the resource specialist and apply for aid, but the right to cash itself is denied. They get cash only if the FEP judges they are not immediately employable, or if they first look for work and fail. Even then, they get aid subject to demanding activity conditions. FEPS were to have caseloads of no more than fifty-five so that they could work with families more intensively than JOBS had done. The RS and FEP refer applicants to a Support Services Planner (SSP) to arrange noncash benefits, particularly child care, food stamps, and Medicaid. Some applicants may need only these services and avoid W-2 itself. Federal rules required that eligibility for Food Stamps and Medicaid be

Implementing W-2

•

137

determined by public employees. In the majority of counties where the welfare department ran W-2, the FEP could do this. But in the few, including Milwaukee, where W-2 was privatized, the FEP determined eligibility for child care, but an SSP working for the welfare department determined eligibility for Food Stamps and Medicaid. The actual provision of these benefits was also through welfare, and that linkage would prove troublesome in the early months of W-2. Finally, W-2 would usually operate out of job centers serving social programs other than W-2 and its support services. These might include the Job Service, where job listings were available, or programs for the disabled, the elderly, or children. Thus, the new administrative setup fused the eligibility and employment sides of W-2, then linked cash aid as closely as possible to non–cash benefits, and then placed this entire structure inside facilities providing other resources. This reflected the idea that W-2 was not just a program but an idea that embraced much more than welfare in the narrow sense. It also reflected a sincere desire to make programs accessible. The Wisconsin reform limited dependency by insisting on work and not by making benefits otherwise hard to get.

BUREAUCRATIC COMPETITION The most disruptive aspect of the new process was competitive selection of the agencies to run W-2. The 99 Group had been drawn to contracting as a way of shaking up the welfare system. One of its models was America Works (AW), a proprietary job placement firm outside Wisconsin that had long claimed that it could put welfare recipients to work better and cheaper than the existing system. AW advocated not only work first but the payment of agencies on the basis of job entries achieved, rather than clients served. The idea stirred resistance and rebuttals from AW’s public and nonprofit competitors.14 The 99 Group doubted that some welfare departments were either willing or able to run W-2. So it made the competitive choice of contractors part of the W-2 plan.15 The earlier administration of welfare in Wisconsin had already involved contracting. By the early 1990s, DHSS’s initial contract for JOBS went in most counties to local welfare departments, but in some cases it went to employment agencies or other public programs, and in a few cases to nonprofit organizations. The employment agencies were usually the Private Industry Council (PIC), which administered JTPA training programs. In 1996, JOBS’ last year, these initial contractors included 38 welfare departments, 6 employment programs, and 22 private organizations, covering 72 counties. The private bodies included several consortia that served groups of rural counties. In Milwaukee, the state dealt with several

138

•

Chapter 7

agencies, as explained below. The county agencies, in turn, often subcontracted further at the local level.16 As noted in chapter 5, subcontracting was prominent in innovative counties such as Kenosha or Sheboygan. W-2 made this contracting process universal and explicit. The 99 Group had wanted DHSS to choose its local partners with a completely free hand. The legislature had imposed a “right of first selection.” RFS allowed counties and their agencies to win the right to run W-2 without competition provided they met certain criteria, to be defined by DHSS. For the first contract, covering 1997–99, the department required that counties reduce their caseloads by 25 percent between September 1995 and August 1996, with some adjustment for counties that had already had substantial caseload fall. Agencies had to have certain percentages of adult recipients employed or in work activities within JOBS, and they had to reduce AFDC spending compared to the previous year. They also received credit for having been pilots for various Thompson reform programs and for other reasons.17 While RFS applied to county governments, the law provided that a county that gained RFS also had to offer at least a subcontract to its existing JOBS agency. This ensured that local agencies that had performed well would not be ousted from the new system. The counties responded strongly to this challenge, as they had to the earlier JOBS performance incentives. DHSS had expected that less than half the old agencies would be chosen; but by August 1996, no fewer than 67 of the 72 counties had earned RFS for the initial contract to run W-2, a tribute to their resiliency.18 Milwaukee was much the largest of the five counties that missed. Of the four others, one was given a contract for W-2 on a discretionary basis. Seven other counties won RFS but decided not to exercise it; all these also were small, except for Waukesha, a suburban county outside Milwaukee. This confirmed the judgment of the 99 Group that not all counties even wanted to run W-2.19 Thus, eleven counties were opened to bureaucratic competition. Milwaukee was broken up into six regions (see further below), so a total of sixteen jurisdictions were put up for bids. In Milwaukee and several smaller counties, the winning agencies were proprietary or nonprofit organizations, which gave the outcome an antigovernment face. But in another case the winner was a PIC, in another a Job Center backed by the welfare department and the ES as well as a private firm. And in two counties that had been denied RFS, the public providers rallied and won entry to W-2 competitively. In the end, only nine counties were privatized. Of the initial RFS losers, only Milwaukee and midsized Walworth County were kept out of W-2, and they declined to bid competitively.20 The initial awards covered a start-up period running from March through August 1997 as well as the first full W-2 contracts running from the inauguration of the program in September 1997 through the end of

Implementing W-2

•

139

1999. For the second contracts, covering 2000–2001, the RFS criteria included several financial and administrative criteria but also performance measures such as maintaining full engagement of clients in the program and limiting recidivism to the rolls by those leaving welfare. For the third contract, covering 2002–2003, the RFS criteria were simply to satisfy the performance measures in effect for 2000–2001 (see below).21 These later rounds produced little change. A large majority of the counties continued to win RFS, and in the few counties opened for competitive bidding, little shift in agency responsibility resulted. At most, there might be a slight trend to further privatization over time. The most important change was that Employment Solutions, one of the five private agencies chosen to run W-2 in Milwaukee, encountered financial irregularities and withdrew from the program at the end of 2001. Its caseload was given to two of the other private agencies.22 The early RFS criteria represent the leading instance in the entire history of the Wisconsin reform where goals appeared to focus more on cutting caseloads and costs than on changing the nature of welfare. The standards probably did accelerate the caseload decline, and the contracting process did drive down the cost of welfare. One estimate was that the state saved $10 million on the contract for 1997–99 compared to what it would have spent by continuing to fund the established agencies. Another estimate was $28 million.23 Yet the greater effect of contracting was to make the administrative structure focus on results as never before.

PERFORMANCE CONTRACTING The other linchpin of the new accountability system was performance contracting. Agencies were not only chosen on a merit basis. They also had to operate under terms that tied their payoffs to results. At the same time, the state loosened the regulations it had previously placed on agency operations. As with bureaucratic competition, performance contracting had already begun under JOBS—the incentive payments systems the state instituted in the mid–1990s to shift the priorities of programs toward higher participation and work first (see chapter 4). For the initial W-2 contract, covering 1997–99, the state set only administrative standards.24 Agencies were to decide eligibility for W-2 quickly and accurately, reassess eligibility at specified intervals, refer cases properly for child support enforcement, and so on. DWD was concerned that the agencies avoid errors in case processing, a worry that proved to be well founded. Planners also felt that it was premature to define outcome measures before more was learned about what the new program could achieve.

140

•

Chapter 7

The incentives of the contract, however, made clear that the state’s initial priority was to make job entries quickly and reduce the rolls. The funding that agencies received were based on their existing caseloads but assumed that these would drop 30 percent over the life of the contract. Funding was fixed at the outset, while the contractors were left free to decide what proportions of clients to place on the four tiers of W-2’s employment ladder. This gave them a strong incentive to place as many clients as possible in private-sector jobs, since then they avoided the cost of creating the CSJs and remedial programs required on the bottom two tiers. There were incentives to underspend. The agencies faced no required spending levels, yet they could keep as profit up to 7 percent of any money they did not spend. They could also keep 10 percent of unspent money beyond this; the rest (90 percent) was to be returned, with half going back to DWD and the other half going to “reinvestment in the community for services to low-income persons.” Reinvestment programs were to serve the goals of TANF and be subject to DWD approval. Underspending proved to be substantial, due to the unexpected caseload decline. For the second contract, covering 2000–2001, in addition to activity measures, the state now defined a set of performance indicators akin to those used in earlier welfare employment programs and in TANF at the federal level. The criteria, and the “base level” expected on each were: • Full engagement: 80 percent of cases must be engaged in appropriate activities. • Basic educational activities: 80 percent of participants without high school education must be in educational activities. • Entered employments: 35 percent of participants must have obtained a job. • Wage rate: Those gaining jobs must have wages at least equal to those reported for 1998. • Job retention: 75 percent of participants entering jobs must still be employed after 30 days, and 50 percent after 180 days. • Health insurance: 30 percent of jobs obtained must provide health insurance.

Notably, these standards did not include driving the caseload down or saving money. Indeed, the legislature forbade rating agency performance on these bases starting in 2002.25 Agencies achieving above these levels received funding bonuses. In addition, the base level criteria (measured for the first year of the contract) were also the standards for RFS for the contract for 2002–2003. Thus, the criteria for bureaucratic competition and performance contracting finally converged. For the 2002–2003 contracts, the performance criteria were

Implementing W-2

•

141

much the same, but some administrative criteria were added. In addition, in the contracts for 2002–2003, DWD began to try to adjust the standards expected of an agency in light of its environment. Some norms were lowered, for example, if a county suffered abnormally high unemployment.26 Wisconsin officials and observers believe that their performance system makes the bureaucracy more accountable for results than in any other state.27 Thus the accountability system was fine-tuned over time. At the outset, the state knew only that much of the caseload could readily be moved off the rolls into jobs. Earlier reforms had already promoted this, and W-2 aimed at more of the same. But after the collapse of the caseload, DWD and the counties settled down to managing a much smaller caseload that also needed more help to work. The current performance system optimizes operations around that somewhat different mission. It is part of the shift toward paternalism discussed in chapter 8. PRIVATIZATION IN MILWAUKEE Privatization changed the administrative structure most radically in Milwaukee County, yet here was where privatization was already most important prior to W-2. The county welfare department had refused to be Madison’s chief contractor for JOBS. It feared that it would have to assume the cost of public employment programs if JOBS were curtailed. It also disapproved of the idea of enforcing work on welfare, which JOBS already embodied.28 Accordingly, the state welfare department (DHSS) made itself the agent for JOBS in Milwaukee and contracted the program to several local agencies. By the time I interviewed in 1994, these providers included the Job Service, which was a public agency, and several nonprofit organizations— Goodwill Industries, Opportunities Industrialization Center (OIC), the YWCA, the Consortium of Hispanic Agency Directors (CHAD), Lau Family, and Greater Access to Employment and Self-Sufficiency (GATES). CHAD served mainly Hispanics, while Lau Family catered to Asians. GATES, which arranged support services to recipients seeking training or jobs, was run jointly by welfare and several of the nonprofits, so the county welfare department still had some role. The Job Service performed intake to JOBS, but then few recipients were expected to participate actively. Those who were called in were offered their choice of agency. Most chose one of the nonprofits, which offered various types of education and training. Clients could also arrange remediation programs themselves and seek child care through GATES. The

142

•

Chapter 7

Job Service was the only agency that stressed job search, and it was also the only one to sanction clients actively for nonparticipation. Thus, JOBS in Milwaukee was passive and effectively voluntary. Some of its clients did get jobs, and the program claimed credit for this, but it seldom pushed people actively to work. Compared to other counties in JOBS, as shown in chapter 5, Milwaukee achieved well- paying jobs, but relatively few of them. It lacked the strong work message and the capacity to oversee clients found in higher-performing counties such as Kenosha. Chiefly for this reason, caseload decline in Milwaukee lagged well behind that of the state as a whole. As table 7.1 shows, compared to April 1986, cases on AFDC had fallen 27 percent statewide through the end of 1994, but the figure was only 4 percent in Milwaukee versus 41 percent in the rest of the state. Serious work enforcement then began in Milwaukee due to pressure from the state and the diversion programs. By September 1997, as W-2 began, the statewide caseload was down by two-thirds, compared to 43 percent in Milwaukee and 85 percent in other counties. The transition to W-2, which lasted until March 1998, saw a dramatic further fall in the Milwaukee rolls, but the county climbed to 86 percent of the remaining state caseload due to the near extinction of cash aid outside the city. Since then, the statewide caseload has stabilized at under 10 percent of its 1986 level, and the Milwaukee proportion has dropped slightly.29 Convinced that Milwaukee County could not run W-2, the state had denied it RFS. The county did not compete for a contract. Rather, under separate legislation, DWD agreed to contract with the county PIC to oversee W-2 operations in Milwaukee. The PIC has tracked how the W-2 agencies have handled client requests for time-limit extensions, among other matters. Milwaukee also obtained permission for its welfare department to continue to run the key support services under W-2—child care, Food Stamps, and Medicaid.30 For W-2, Milwaukee was broken up into six regions, in the belief that it was too large to be run as a single unit. The six regions were awarded as follows: • Employment Solutions, Inc. (ESI), an offshoot of Goodwill, gained the contracts for two regions, in black areas to the north and center of the city. As mentioned above, ESI withdrew from W-2 at the end of 2001. Its regions were then given to YW Works and UMOS. • Opportunities Industrialization Center (OIC) gained the contract for another largely black area of central Milwaukee. • Maximus, a national human services firm and one of two proprietary agencies among the five, gained the contract for the southwestern sector of the city.

Implementing W-2

•

143

• YW Works, a proprietary offshoot of the YWCA and two other nonprofits, gained the contract for the city’s northeastern sector. • United Migrant Opportunities Services (UMOS), a nonprofit that served migrant labor, gained the contract for the heavily Hispanic southeastern sector of the city.31

Since Milwaukee now comprised the great bulk of the statewide caseload, the state had staked the future of W-2 on these private providers, several of them new and untried. Furthermore, the agencies then contracted with other private organizations for various services, and sometimes with each other. Public agencies seemed very much sidelined. However, there was less actual change than meets the eye. Other than Maximus, all the agencies or their antecedents had participated in the earlier JOBS structure, and all of them hired managers and staff from these earlier organizations or the public agencies. And while Maximus was new, its Milwaukee operation was headed by George Leutermann, a creator of the Kenosha reforms that had earlier set an example for the state. All the new agencies zealously embraced the mission of W-2—refounding assistance on work. Maximus’s principal offices were emblazoned with the same upbeat slogans seen in Kenosha—“You are the best . . . Expect success,” and “At Maximus, our business is successful families.” All the players claimed to be motivated as much by service to the community as by the chance to make money; all claimed to spend money on services and administration beyond what was required under the initial contracts.32 All used some variant of the W-2 administrative process described above. At the same time, they had different styles. Employment Solutions stressed ready access for applicants and outreach to the community. Maximus and OIC sought to meet the demands of employers. Maximus created a “Max Academy” that offered a short, intensive course in job readiness. YW Works and UMOS were more service oriented, more attuned to helping clients overcome the problems that kept them from working.33 In the run-up to the inauguration of W-2 in September 1997, the new agencies hired frantically and designed new procedures. Many staff aiming to become FEPs came from positions, such as welfare eligibility worker, where their duties had been narrower. The state mandated that they be trained to exercise their new powers to assign clients, and also to enter data correctly into the CARES computer system, but in many cases the training was insufficient.34 The new staffs often failed to provide the individualized attention that the new system promised.35 Privatization by itself could not reverse the long history of administrative mediocrity in Milwaukee. Many managers and staffs simply did not

144

•

Chapter 7

know their jobs or care very much about them. The city’s agencies were probably better run than in other big cities where I have interviewed. But when W-2 began, they were well below the level achieved in the rest of the state. STRUGGLES TO IMPLEMENT The combination of these long-standing limitations with the radical changes imposed by W-2 led to monumental difficulties during the early months of the new program. These were easily the worst administrative problems in the entire history of the Wisconsin welfare reform. Some disruptions, indeed, were inevitable given the scale of the challenge. During the transition period, from September 1997 through March 1998, the state had committed itself to placing over 20,000 recipients in unsubsidized employment and creating CSJ positions for an even larger number— figures far beyond what had been achieved before.36 Low Take-up of W-2 Take-up of the new program fell far below expectations. This was partly because recipients on AFDC did not automatically transition to W-2. They had to come in and apply for the new program. DWD required that agencies show “due diligence” in contacting their cases. So a “scramble” ensued, Jean Rogers said, to call in recipients and orient them to W-2.37 Agencies strained to contact their caseloads using letters, television ads, even home visits. But many AFDC recipients never reacted to the changeover. According to a survey, 70 percent of former AFDC recipients in Milwaukee remembered receiving notice of the conversion, with smaller proportions attending orientations and meetings. But around 70 percent found the process confusing and difficult, and 28 percent said they took no steps to convert to W-2.38 Thus, when AFDC died in March 1998, many cases simply died with it. DWD could only hope those who did not come forward really did not need the aid.39 Those AFDC recipients who did respond were often deterred from seeking further aid by the tough work and participation demands of the new system. This meant that the caseload fall—already rapid—accelerated still further. Most of this reflected diversion. Even after March 1998 when W-2 was fully operational, around 7,000 cases were diverted from the new system in the rest of 1998.40 The sharpness of the decline unnerved administrators. In December 1997, Linda Stewart, Secretary of DWD, declared that it could not continue, and yet it did.41 As table 7.1

Implementing W-2

•

145

shows, statewide the number of cash cases fell by over half during the transition period, and most of this occurred in Milwaukee. By the time the dust settled, cash welfare had subsided to a residual operation throughout the state. As mentioned earlier, the decline in part reflected cases that were transferred to the new Kinship Care and Caretaker Supplement programs, but here too there were problems. Kinship Care outgrew its funding, and the state had to scramble to find more money for it.42 Denials of Aid This revolution brought a chorus of complaint. While the state thought it was earnestly transitioning AFDC to the new system, opponents, such as Senator Gwendolynne Moore, thought it was driving mothers off the rolls without any serious effort to help them work.43 Complaint focused most sharply on the propensity of FEPs to define as “job ready” hundreds of applicants who seemed employable but were currently jobless or working part-time. This placed them on W-2’s top tier, where they could receive only case management and support services, but not cash aid. Applicants could contest the decisions, but then they confronted an appeals process that was more informal and less predictable than the fair hearings procedure that had prevailed under AFDC. On appeal, most clients won their cases, making the agencies look incompetent.44 After an outcry, the agencies curbed unsubsidized placements. They cut the period during which they expected applicants to look for work to ten days. If people were still jobless at that point, they would commonly be assigned to a CSJ slot, where they could receive cash aid. CSJ became the “default assignment” for most cases transitioning to W-2.45 The agencies, however, had difficulty creating enough CSJ slots. As earlier under PFP, some ostensible assignments were not real. This meant that for some clients, entitlement persisted. Denials of Child Care Another acute problem was child care. The state had promised that subsidized care would be available to all those leaving welfare for work, but at first Milwaukee failed to deliver. This was chiefly because of the troubled interface between the new W-2 agencies and the county welfare department, which retained control over child care, food stamps, and health coverage. Although the department stationed staff at the W-2 agencies to deal with these benefits, it was slow to approve new child care providers and also to reimburse the providers for serving W-2 families. Some children were denied care.46 In October 1998, the state was forced

146

•

Chapter 7

to give Milwaukee extra funding to accelerate certifications and reimbursements.47 At first, the state never came close to spending the millions of dollars it had committed to child care as a condition of enacting W-2. It succeeded in wiping out waiting lists for care, but the buildup of subsidized slots fell far short of forecasts. Budget planning for 1997–99 assumed that 77,000 children would be in state-supported child care by December 1997, but in fact only 19,500 were, or fewer than had been served or on waiting lists a year earlier. The shortfall reflected the operational problems, as well as the unexpected caseload decline.48 More recently, however, child care costs have grown sharply, and the state is finally feeling the pressure of the commitments it has made. Sanctions In the early months of W-2, the activity and work demands of the new system were not well enforced. This was because the W-2 agencies were busy transitioning cases, and because they hesitated to sanction while clients were still adjusting to the great change. Eventually, the state grew alarmed that “engagement” would fall. Wisconsin might then fail to meet the severe federal participation norms set in PRWORA for two-parent cases in AFDC. The proportion of Milwaukee recipients fulfilling work or job-search obligations had fallen from 72 percent in October 1996 to 49 percent in November 1997. Accordingly, in December 1997, Jean Rogers wrote a memo to agency directors demanding that sanction levels increase, and they promised to comply.49 Sanctions then sharply rose. By the third quarter of 1998, more than a quarter of W-2 cash cases (those on the CSJ or transition tiers) were getting sanctioned every month, and by March 1999, the figure had risen to 42 percent. In a survey done in 2000, Milwaukee FEPs asserted that noncompliance was common among CSJ participants. Strong majorities of the FEPs believed that sanctioning was essential to control the problem.50 As cases disappeared, the role of sanctions finally declined. Thirtyone percent of cases were under sanction in October 1999, only 21 percent in December 2000.51 In 1998, Wisconsin had 5 and 18 percent of its caseload under full and partial sanction, compared to 1 and 5 percent for the average state. But the absolute number of such sanctions was small—477 and 1,899 cases, respectively.52 This was partly because Wisconsin is not one of the larger urban states, but also because its tough policies had already driven most families off the rolls. Paradoxically, more sanctions finally led to fewer.

Implementing W-2

•

147

TABLE 7.2 Assignment of W-2 Participants to Employment Tiers, 1998–2000 Planning Assumptions

Unsubsidized Trial jobs CSJ W-2 transition

Transition

New Cases

March 1998

December 1998

December 1999

December 2000

15 10 50 25

40 15 35 10

29 0 59 11

31 0 51 18

40 0 39 21

41 0 38 22

Source: Department of Workforce Development. Note: Some figures do not add due to rounding. CSJ category includes cases that receive cash aid without a work assignment due to having a child under twelve weeks of age.

The Weakening of Work Tests Another issue concerned the assignment of clients to the four tiers of the W-2 employment ladder. In budget planning, DWD assumed that participants would be distributed over the four levels as shown in table 7.2. While few of the cases transitioning from AFDC were expected to be immediately employable, most of the new intake after this, it was assumed, would obtain jobs in the private sector. By March 1998, at the end of the transition period, 29 percent of cases were assigned to unsubsidized employment. The number was depressed by the controversy over the assignment of the jobless to this tier. Later, unsubsidized work rose to projected levels. Less than 1 percent of cases, however, ever went into trial jobs, largely because employers feared the paperwork and were willing to hire without the subsidy.53 A higher proportion of cases than expected went onto the grant-paying tiers, CSJ and W-2 transition. CSJ mushroomed during the changeover from AFDC, then somewhat declined. This tier turned out to be somewhat less demanding about required activities and work than Pay For Performance had been, or than the state expected. Not only did some early CSJ assignments exist only on paper, but the requirements of the positions fell over time. By early 1998, W-2 had backed away from its forty-hour standard and accepted more like thirty hours. Many cases were placed in thirty-hour CSJ positions, perhaps with some education on the side.54 And the education became more formal, thus compromising the policy that W-2 would not allow regular instruction. In fact, the state allowed the Milwaukee agencies to provide CSJ participants with classroom training for more than ten hours a week provided it occurred at central CSJ work sites.55

148

•

Chapter 7

Meanwhile, W-2 Transition grew in relative size over time. Planners had assumed that this tier would receive only the most troubled cases. Here, as in CSJ, there was a tendency through 1998 and 1999 for fewer clients to be assigned to actual work or job search and for more to go into remediation. The nature of the training shifted away from specific job skills and toward general work readiness and “soft skills.” The bulk of transition participants were placed in substance abuse or other rehabilitation programs.56 As I note in chapter 8, the rising emphasis on remediation reflected the declining employability of clients as the caseload fell and, in response, the tendency of the program to become more paternalistic. It reflected as well the pressures mentioned earlier to allow more education and training under W-2. The Thompson administration resisted. David Riemer chastised the Milwaukee agencies for giving too many clients cash to pursue education and training, without work. This he called a “terrible mistake” that would take welfare back to the old days.57 The slippage, however, was not enough to affect the work-centered essence of W-2, simply because so few cases remained. The new system enforced work off the rolls more than it did on. Time Limits A final struggle concerned time limits. Under W-2, placement on any of the bottom three tiers is limited to two years. This meant that cases on CSJ or transition that were getting a cash grant could not remain in that status indefinitely. W-2 allows extensions of up to six months on a discretionary basis, to avoid hardship. Second and third extensions are possible. But DWD worried that to be lenient would lead to problems later. Large numbers might go “over the cliff” when TANF’s five-year limit came up in 2001.58 W-2 agencies were supposed to work intensively with participants from the eighteen-month point to be sure they understood the time limits and did everything they could to leave cash aid. That process became intense in 1999 as many cases approached the two-year mark. The agencies decided which cases deserved extensions, with final approval by DWD in Madison. The reviews consumed considerable bureaucratic effort. Few extensions were granted. Most recipients left their tier before two years; most often, they left welfare or switched to a different tier where they received a fresh 24-month clock. In Milwaukee, only 20 percent of CSJ cases reached the limit without an extension, only 1 percent with an extension; for transition cases, the figures were 15 and 23 percent. The local agencies usually asked for extensions only for cases that were cooperating and trying hard to work but faced difficult obstacles, particularly

Implementing W-2

•

149

medical problems. On the other hand, around 80 percent of their requests were approved at higher levels; the others were withdrawn or still pending. And most of the cases denied an extension moved to another tier, found a job, or qualified for disability coverage.59

PUBLIC REACTION W-2 was implemented, especially in Milwaukee, amid a storm of negative publicity. Despite the system’s generous aspects, it first appeared to many people to be simply a stringent work test that withdraw aid from families accustomed to receiving it. As before, there were protests against the plan. In November 1997, a Milwaukee group called Coalition for Change staged a rally of 150 people in Madison calling for changes. In December, 400 people gathered in Milwaukee to complain of difficulties they had had dealing with W-2. In April 1998, after the crackdown on sanctioning, activist groups led a protest march in Milwaukee. In July, following a convention by ACORN, a leading opponent of reform, a thousand attendees protested outside Governor Thompson’s Milwaukee office, their main demand being to unionize W-2 participants.60 Clients’ views of the new system were decidedly mixed. In the survey of former AFDC recipients mentioned above, 76 percent felt that W-2 workers wanted the participants to succeed and 71 percent said the new system motivated people to get jobs and get off welfare; 56 percent said that W-2 provided useful training and good job opportunities. However, 70 percent also felt that W-2 demanded too much of parents with young children, and 45 percent felt that government should give checks to needy families without requiring work—a return to entitlement. Only 43 percent felt the new program was better than AFDC—47 percent said it was not. The main suggestions for improvements were to allow more education and training and improve administration and support services.61 In a survey of W-2 participants done in 1998 and 1999, thin majorities said W-2 was generally on the right track and rejected the idea that the program was a waste of time or hostile to taking care of one’s family. But larger majorities rejected the idea that W-2 helped recipients get a job or a better job or made them feel better about themselves. In this study, and also in DWD’s own study of welfare leavers, W-2 caseworkers received better reviews, for fair and efficient treatment of clients, than did their program.62 The results for participants are probably more negative than for former AFDC recipients because the former were still on welfare, while many of the latter had left and felt better about their lives. That interpretation

150

•

Chapter 7

squares with a journalistic survey of “dozens of former welfare recipients in the Milwaukee area” reported in September 1998: Few [former recipients] say they miss the old system. Yet some have fallen through the cracks, confused or defeated by a system that at times has befuddled even the people administering it. Others entered the reform program reluctantly, but found success through support and hard work. And even families that improve their conditions through work tend to make progress only gradually. The new program has helped lead some former welfare recipients to jobs they can keep and advance in, jobs that afford them the income, flexibility and security they need to be financially independent.63

Over time, as described in the next chapter, W-2 became more supportive of the remaining, often troubled cases on cash aid. Several FEPs in Milwaukee suggested to me in interviews that much of the opposition to the new demands was transient. One reported that when SSF and PFP were instituted, clients showed “hostility” because they weren’t used to “doing anything” for their money, but later they had “calmed down.” Another FEP remarked that when W-2 came in, a “majority” of recipients reacted as if to say, “They can’t do this to us.” But this was because the clients were “taking the system for granted.” Once they saw that change was “inevitable,” they also saw the “merits” of it.64 For program staff, this interpretation might be self-serving, but it also squares with what we know about the psychology of the seriously poor. Most welfare adults are ashamed not to be working and supporting their families. Advocate groups may demand entitlement; recipients seldom do. Thus, most support work requirements in principle, and this makes them acquiescent when work actually is enforced.65 Over time, the opposition to W-2 subsided into the criticism of details that we saw in the legislative arena in chapter 6. In March 1998, Antonio Riley lauded W-2’s focus on work. His main desire was simply to allow more education and training so that participants could work their way up to better jobs.66 At a Milwaukee meeting in June 1998, community groups drew up a list of ways to improve W-2, for instance by speeding up child care subsidies. They left unquestioned the principles of the program.67 THE FATE OF THE AGENCIES As mentioned in chapter 6, the state took a huge gamble in launching W-2 when its costs were unpredictable and potentially huge. The startling caseload decline saved the state from red ink, but it also launched the W-2 agencies onto a roller coaster from boom to bust. They have in a

Implementing W-2

•

151

sense solved the welfare problem in Wisconsin, but they may also have put themselves out of business. Initially, welfare personnel had supported W-2 with some reluctance. Staff and some managers feared that if reform drove the caseload down further, their positions might be at risk. But true to the moralistic style, they did not defend this partial interest with conviction. State and county elected officials took the risk of job loss as a sign that welfare reform had succeeded, not that it had failed. They assured the established agencies a chance to compete for a role in W-2, through RFS. Once that was done, bureaucratic resistance faded. The ensuing collapse of the rolls and the financial problems I describe below did, however, weaken civil service morale.68 The initial W-2 contracts, which covered September 1997 through December 1999, were valued at $651 million. But, chiefly due to falling rolls, the agencies spent only 63 percent of this, or $414 million. Under the terms of the contracts, the state retained $84 million of the surplus, but the agencies nevertheless garnered $65 million in profits, although at first they earnestly spent much of this on operations. They also gained another $83 million that they were supposed to spend on community reinvestment initiatives. This they did, funding programs for work support, housing, transportation, education and training, mental health, pregnancy prevention, and other purposes loosely tied to welfare.69 These programs served the most expansive conception of W-2. But the extent of surpluses put strong pressures on the politicians to reduce windfalls in future. For the second W-2 contract, covering 2000– 2001, agency funding was cut to $369 million, or by 43 percent, and the amount of surplus the agencies could retain for either profit or community reinvestment was limited to 7 percent. Tommy Thompson had to veto attempts by the legislature to cut the level further.70 These decisions did not deter any of the contractors for 1997–99 from signing up for 2000– 2001, in part because they could use their surpluses on the first contract to offset the cuts in the second. But that cushion was gone by the third round, for 2002–2003, when the value of the contracts was cut further, to $297 million. Two small counties withdrew for financial reasons even though they had won right of first selection.71 The broader problem was that, as the cash welfare caseload fell, it could no longer support welfare agencies around the state. From the agencies’ fiscal viewpoint, enforcing work was all too successful. Most of their clients left the cash welfare rolls, yet their responsibility for the low-income population was wider than before, particularly if one defined W-2 broadly to include related programs. The remaining cash recipients demanded intensive help because of their disadvantages, but many others received case management, others just child or health care. Still others came into

152

•

Chapter 7

TABLE 7.3 Wisconsin Family Case Counts, April 1995 and April 2000

Case assistance (AFDC/W-2) Other cash cases Case management cases Family Food Stamp cases Family Medicaid cases Child care cases Total unduplicated case count

April 1995

April 2000

62,752 11,215 0 78,904 111,170 9,844 118,595

6,642 11,553 4,032 44,863 101,991 18,784 114,725

Percent change −89 +3 −43 −8 +91 −3

Source: Rebecca Swartz, “What Is a ‘Case’ in Post-Reform Wisconsin? Reconciling Caseload with Workload” (Madison: Hudson Institute, Welfare Policy Center, March 2001), table II-4. Note: Other cash cases include Kinship Care, SSI Caretaker Supplement, and cases where the caretaker is not legally responsible for the children. The 1995 child care figure is an estimate for 1996; no earlier figures are available. Total cases for 1995 excludes child care and includes child-only cases later placed under Kinship Care and Caretaker Supplement.

job centers expecting incidental aid with all manner of problems. Only some of these people were on caseloads that agencies earned money to administer. Table 7.3 shows how the composition of the “welfare” population (here limited to family cases) shifted between April 1995 and April 2000. There was a dramatic decline in cash cases, all in AFDC and W-2, with lesser declines in Food Stamps and Medicaid. But there was a near-doubling in cases receiving subsidized child care. The overall caseload actually changed very little, if it is defined to include all these clienteles. The fiscal shifts are shown in table 7.4. In 1996, under AFDC, three-quarters of Wisconsin’s welfare spending went to cash aid and much of the rest to child care. In 2000, by contrast, employment and child care were the major expenditures, well ahead of cash assistance and family programs. Again, overall spending changed little, but it shifted sharply away from the traditional welfare caseload. The years 1999–2001 saw massive increases in child care funding, yet only half of W-2 clients received these subsidies.72 For the public W-2 agencies, which are the majority, the decline of welfare might just mean relying more on child and health care funding and less on W-2. But for the private W-2 agencies that never controlled the noncash benefits, the problem is more serious. In order now for local welfare agencies to survive, they have to be funded in a way detached from specific caseloads. Their task might be to keep an entire low-income population working and otherwise functioning, regardless of the number

Implementing W-2

•

153

TABLE 7.4 Wisconsin Welfare Expenditures for 1996 and 2000

Total spending Percent allocated to Cash assistance Workforce development Child care Family stability Tax transfers

1996

2000

399,806,233

383,887,224

76 8 13 3 0

16 25 30 16 13

Source: Rebecca Swartz, “What Is a ‘Case’ in Post-Reform Wisconsin? Reconciling Caseload with Workload” (Madison: Hudson Institute, Welfare Policy Center, March 2001), figures II-6 and II-7. Note: Family stability means programs to reduce unwed pregnancy and improve health and nutrition. Tax transfers means TANF funds spent on the state earned income tax credit.

dependent on regular benefits such as W-2. That implies funding them on a capitation basis, in terms of population potentially served, rather than in terms of current caseloads. They might function as “employment maintenance organizations,” much like HMOs do in health care. David Riemer hit on this idea prior to W-2, and it may represent the future.73 Welfare officials were hoist on their own petard. They faced these problems only because of their success in driving down traditional family assistance. Reform had enforced work on the welfare caseload but extended benefits to a much larger employed population. The intense dependency of a narrow group was exchanged for the lesser dependency of a broader one. “Welfare” was no longer so distinct from the rest of society. Thus, providing aid was no longer so different from providing other public services such as police or fire protection.

RESEARCH AND EVALUATION The crafting of W-2 had been more deliberate than the earlier stages of the Wisconsin reform, and the arrangements for assessing W-2 were more thorough as well. If nothing else, the state needed some way to handle the outside research pressures that arose from the very prominence of the Wisconsin reform. By the early 1990s, welfare officials and scholars from around the country, and abroad, inundated Wisconsin with requests for data and access to programs, in order to understand what the state had achieved.

154

•

Chapter 7

Suspicion still separated Wisconsin leaders from the academic world, much of which continued to oppose the state’s work-enforcing approach to reform. Researchers from UW’s Institute for Research on Poverty (IRP) and the Employment and Training Institute, and from outside the state, still wanted to study W-2, but the Thompson administration hesitated to give them entre´e, lest they embarrass the government. In the long run, however, the academics’ estrangement from policymakers was damaging to both sides. Researchers needed access to have anything to study, and government needed systematic information about its reform. A further problem was that, even had welfare research been less politicized, academic inquiry about poverty or welfare was often of little use to government. Academics were often more interested in describing and analyzing social problems than in showing how government could affect them. They seldom addressed institutional questions of how best to run programs. Such results are not useful to policymakers or administrators seeking to improve policy. Why, then, should officials incur the cost to provide data to researchers or allow them to interview staff or clients? Tensions eased once the success of reform was a fait accompli. It became less plausible to say that social and economic conditions made dependency inevitable, as IRP researchers had suggested. There was more interest in evaluating the new programs and also in improving their operations. This gave greater prominence to those IRP scholars who did have an interest in the institutions. They made vital contributions to improving the management and administration of welfare work programs.74 A novel approach to the research interface problem was developed in 1996 by Jean Rogers, head of Economic Support at DWD, and Michael Wiseman, an economist and welfare expert from UW—Madison.75 Rogers realized the value of research, provided it was policy relevant. Wiseman was unusual among academics in supporting the Wisconsin reform and in emphasizing the importance of management and administration. Together, they created the Management and Evaluation Project. MEP’s mission was to promote research on W-2 that serves state needs. It was funded largely by the Hudson Institute and other foundations. In a way it continued the mission of the 99 Group. MEP was led by DWD senior administrators and research staff, with some academic representation.76 It ruled on outside requests for data and access. For approval, it expected that the research be methodologically sound and relevant to policy and that the data demands not overburden DWD. MEP also tried to anticipate the research policymakers would need, and it commissioned studies that were lacking. Some outside groups approached MEP seeking funding for projects. Others had funding but wanted their studies to be MEP-approved. MEP also oversaw evaluations of programs being done by DWD. To advise on methodologi-

Implementing W-2

•

155

cal issues, it had a technical advisory committee of experts both in and outside the state. MEP sponsored or approved close to thirty studies. Some of these were program evaluations, while others tried to assess the effects of W-2 by following cases that have left the rolls. One study targeted the implementation of W-2 in Milwaukee, and all the studies addressed administrative issues to the extent possible. I have drawn on this research heavily for this book. Outside observers agree that Wisconsin has unusually rich information on its reforms compared to other states, and that much of the credit goes to MEP.77 Funding for DWD research, however, was sharply cut in 2001 due to the state’s budget crunch, and MEP ended in 2002. In some of these projects, UW researchers are involved, so a partial rapprochement between the university and the administration has occurred. In the long run, the state’s tradition of collaboration between academia and government will probably reassert itself. This partly reflects the confidence administrators have that their reforms are, in fact, successful. In other states, notably New York, the estrangement between intellectuals and conservative policymakers is far greater, and government resists virtually all outside inquiry. There is no MEP in New York, or in any other state, to my knowledge.

BEYOND WELFARE W-2 is the most far-reaching redesign of welfare by any state. The final verdict on its success rests with its consequences, which chapter 10 will survey. But already in 1999, W-2 won an Innovations in Government Award from Harvard’s Kennedy School of Government. The program was cited for reducing the caseload but more for its structural innovations—especially the merging of the eligibility and employment roles in case management and the generous provision of support services.78 The larger importance of W-2 is that it proved that traditional welfare could virtually be abolished, even in a major city. Many have seen entitlement aid as a price that America has to pay to the poor who, due to various inequities, fail to enter mainstream society. Especially, it is a tribute that America pays to poor blacks to appease their anger and forestall urban riots as in the 1960s.79 In the view of others, opposition to welfare is also driven by race. Americans, including the media, exaggerate how “black” welfare is, and this motivates their aversion to it.80 W-2 refuted these arguments, at least for Wisconsin. The new plan was enacted by a largely consensual political process in which race played little role. It was then implemented even in heavily nonwhite Milwaukee, albeit with many problems. When one drives into Milwaukee today, one

156

•

Chapter 7

enters a city where traditional cash welfare has mostly disappeared. Tens of thousands of families no longer troop to Twelfth and Vliet Streets, the headquarters of the county welfare department, to get aid. Rather, much smaller numbers go to each of the local W-2 agencies, where they are mostly working in community jobs. One can imagine the fury, among some, that such a change might cause. But one can also imagine a brighter future for racial integration in Wisconsin, because less welfare and more work is a key to black advancement in America. Provided that other social policies, such as law enforcement and education also succeed—a big if— one can even imagine the dissolution of the ghetto in Milwaukee. That prospect is still distant, as chapter 10 will show. Yet at least the aid system now supports the patterns of life that promote integration, rather than hindering them. It now drives its participants toward the center of American life, which is the workplace, rather than supporting them outside. It has become like a doctor’s office, meant to help in need but with the goal of a quick return to functioning. As cash welfare fades and the W-2 agencies serve a much wider population in more diffuse ways, they have become cheerleaders for mainstream American life.

CHAPTER 8

Paternalism

THE MAIN LINE of the Wisconsin reform was the enactment and implementation of general work requirements. In large part, that process was impersonal. The demands were enforced across the board for the employable adults. Many people were simply driven off the rolls by the demands, or by diversion, and nobody paid them much heed. But viewed closer up, the process often involved a good deal of individualized attention. Case managers helped recipients arrange the child care and other services they needed to work. They also followed up on clients to be sure they fulfilled their obligation, which was to work or participate in assigned programs. This administrative style, which I call paternalism, seems to have developed spontaneously in Wisconsin. It has become the state’s main contribution to antipoverty policy, after its more noted policy innovations.1 Paternalism was a conspicuous aspect of Learnfare and Children First, two early Thompson experiments that were supposed to keep welfare children in school and enforce child support payment by absent fathers. Paternalism was also critical to the more successful JOBS programs and the New Hope project, an important liberal experiment in work-based reform in Milwaukee. And it emerged as essential to the evolution of W-2. After the early period of caseload decline, FEPs hand-tailored regimes of services and requirements for the more troubled cases that remained on the rolls. Paternalism appears to work in this state, with its gifts for bureaucracy and its conformist culture. But could it work more widely?

WHAT PATERNALISM MEANS In welfare and other antipoverty programs, the trend is toward social contract, the notion that the adult recipients of government assistance should fulfill certain behavioral requirements in return for support.2 Preeminently, that means working, but some states also require parents to keep their children in school or to get them vaccinated. Paternalism adds to this an administrative system where case managers oversee individual clients to make sure the social contract is realized. Traditionally, policy analysts have seen case managers as advocates for the poor who helped

158

•

Chapter 8

them get all the benefits to which they were entitled. In paternalism, case managers do this but they are authority figures as well as helpmates. They arrange services, but they also check up on their clients to be sure they fulfill their responsibilities. They provide a combination of help and hassle. In several policy areas, it appears that programs that combine assistance with oversight perform better than those that provide benefits more passively.3 Personalized direction apparently meets the needs of many poor adults. Research suggests that most poor people share the values of the larger society. They aspire to get through school, work, obey the law, and keep their children out of trouble. However, their actual lifestyle often falls far short of these values. School failure, unwed pregnancy, and crime are common, and few adults work consistently. Social barriers such as low wages or racial bias do not suffice to explain this.4 More plausibly, the poor learn about mainstream values, but they fail to internalize them in early life with enough force to govern actual conduct. One reason probably is weak parenting and disorganized family life. Many poor adults seem to appreciate paternalism precisely because it provides the consistent, personalized direction they have been lacking. If nothing else, case managers can explain to them social requirements that they often do not understand. That structure of attention makes it possible for them to achieve the orthodox values, such as work, in which they already believe. The defeatism of their lives is then relieved. In Milwaukee, for example, one welfare mother was put to work loading washing machines, what most would regard as an unrewarding task. But she remarked, “I was happy because it gave me something to do besides sitting at home every day.”5 Social programs need to tell recipients directly and personally what is expected of them.6 John Gardner is an employment trainer I know in Milwaukee who worked for the Congress for a Working America (CFWA), the left-leaning community group that spawned New Hope. He later led a successful movement to take over the school board in Milwaukee. Although a former AFDC recipient himself, he is hard-line about expecting work from the poor, on or off welfare.7 He tells the young men he trains, “I’ll do anything to help you. I’ll get you a job. But if you disappoint the employer—if you make me look bad—if you screw me over—you better watch out. I’m coming after you. I’m in your face. You’ll wish you’d never been born.”8 It sounds harsh to middle-class people, who are used to more solicitude from authority figures. But Gardner says his clients respond. His attitude tells them that what they do—or fail to do—is important. He at least shows them attention and concern. Many of them never had anybody come after them before. They often grew up in neglectful families where parents paid them little regard of any kind. At least case managers are

Paternalism

•

159

consistent. They operate by a set of rules. And by fulfilling those rules, the clients can find a security they have not known. Enforcing the rules in this consistent and personal way actually puts the clients in control.9 Society used to delegate the management of its dependents to professional or uniformed groups. Teachers upheld standards for children in the schools, social workers for recipients on welfare, policemen on the street. In recent decades, that authority has largely been withdrawn, at least in dealing with the poor. In the 1960s and 1970s, urban schools and welfare departments often seemed too severe toward the poor; since the 1980s, they have seemed too lenient. Through its elected representatives, society today is directly setting educational standards and work requirements, also stiffening child support enforcement and law enforcement. There is still discretion in how the rules are applied to individual cases, but caseworkers now are agents of social norms far more directly than before.

LEARNFARE An early arena for paternalism in Wisconsin was Learnfare, enacted in 1987. The program extended the idea of social contract from welfare to education. In poor families, many teenagers drop out of school, often due to pregnancy. Learnfare required that children on welfare aged thirteen to nineteen stay in school (or, if already dropouts, return there) until they graduated from high school, earned a GED, or reached age twenty, on pain of a cut in the family’s AFDC grant. In the first half of 1990, an average of 2,240 children were sanctioned monthly under the program, around three-quarters of them in Milwaukee.10 Other than his benefit cut, Learnfare was the most controversial of Tommy Thompson’s initial welfare proposals. Advocates attacked it as punitive and ineffective, as did Democrats in the legislature. As a result, the program was enacted only in a limited form, but Thompson restored its original parameters using his item veto.11 If the program survived, it was perhaps because the basic idea of enforcing school on the rolls was popular in Wisconsin’s earnest culture (see table 3.1). Learnfare also suffered serious implementation problems. The program presumed that the schools could tell which children were truant and communicate this to welfare, but doing this proved difficult, especially in the largest school systems, in Milwaukee and Dane Counties. Many teachers there also opposed the program initially, although they later became more accepting. Further, in 1990 a lawsuit halted the program in Milwaukee, after many sanctions there were shown to be in error. Reforms were instituted to verify that students were out of attendance before they could be penalized.12

160

•

Chapter 8

Case management had not been part of Learnfare’s design, but it was added to deal with the difficulties. Administrators learned that many clients did not understand Learnfare’s requirements, nor did the state always know why many failed to comply. “It’s not enough to sanction the child,” admitted DHSS Secretary Patricia Goodrich, “unless we know why they are absent in the first place.” That signaled a need for more outreach and support services.13 Federal authorities in 1990 mandated case management as a condition of granting the waiver for the expansion of the program (see below), and that demand was seconded by the settlement of the lawsuit. Families with teens in danger of a sanction would now be approached by case managers. They could either show that they were in compliance or cooperate with the worker to resolve the problems that prevented attendance, for instance a lack of child care.14 Following these changes, sanctioning in the program fell sharply.15 The state understood the effort in paternalist terms, as an exercise in tough love. Together, the threat of sanctions and the case manager were supposed to motivate teens to take school seriously. The sanction was the “catalyst,” case managers were taught, that would get families to work with them to resolve problems. “Strike a balance between caring and toughness”—help students, but also show them that “you mean what you say.”16 “The sanction alone doesn’t work,” one Learnfare manager told me. “They have to see a body.” However, the system lacked authority. Troubled teens did not have to accept case management; only 14 percent did so. Nor did the case managers control whether a teen was sanctioned. Perhaps for that reason, the new system achieved little.17 In Kenosha, however, an aggressive outreach effort with special funding achieved more. Here, a third of truant teens accepted case management, the highest figure of ten counties in a state evaluation. Among teens who accepted case management, their relationship with the case manager was the thing they liked most about the system.18 This sort of response validates the idea behind paternalism—that the poor are socially isolated as much as they are needy, and what they primarily need is a relationship. Learnfare had difficulty showing that it reduced truancy on welfare. Two evaluations failed to find clear effects.19 But in 1994, the program was expanded to cover welfare children aged six to twelve in several counties. The authority of the case worker was increased. Now failure to cooperate with the case manager became sanctionable.20 An evaluation of the new program for children aged ten to twelve did find small effects on school attendance.21 Results might have improved because the attendance problems of younger children are more tractable—or because strengthening case management did something to raise impacts. Interestingly, a simi-

Paternalism

•

161

lar program in Ohio recorded similar results. Even with case management, it achieved only small gains on school attendance or completion. But results improved in a program in Cleveland with more intensive outreach and assistance.22 Expanded Learnfare was instituted statewide in 1997 as part of W-2, but so few families now receive cash welfare that the practical importance of the program is small. The state is thrown back on earlier enforcement devices to combat truancy among the poor.

CHILDREN FIRST The other early reform initiative that developed case management was Children First. This program took on another difficult problem—nonpayment of child support by noncustodial parents, most of them fathers. A principal reason why families go on welfare is that fathers do not support them. It is difficult to establish paternity and make the fathers pay. They have little interest in paying through the official support system, as then most of the money goes to reimburse government for the cost of welfare, little to the family. Many low-income fathers also find it difficult to pay because of their own employment problems. It is not enough to “throw the book” at them by making the enforcement system more punitive, experience has shown; they also need help to work.23 A further enforcement problem is that a nonpaying father’s employment status is hard to verify. Is he working, and if not, could he work? Government lawyers can haul him into family court where judges tell him to get a job and pay up. But if, on further nonpayment, the father claims that he could not find a job, a court will hesitate to hold him in contempt. Under Children First, the father is told, not directly to get a job, but to report to a work program where supervision is enhanced. This is an obligation he cannot evade. If he is not jobless as he claims, the participation requirement will conflict with his unreported job and force him to pay up, or go to jail. If he does participate, he really is jobless and needs help to work. The program provides help, but also verifies that the father participates and looks for work in good faith. By providing new monitoring capacity, the program both assists fathers and enforces support more effectively.24 Counties running Children First are supposed to refer to it low-income noncustodial parents who owe child support judgments, are delinquent in paying them, are working less than thirty-two hours a week, and have no current means to pay. The parents must either pay their judgments for three consecutive months or participate in the program for sixteen weeks. Children First was authorized in 1987 and implemented initially in 1990

162

•

Chapter 8

in Fond du Lac and Racine Counties. It was later expanded, using federal waiver savings. Currently it serves thirty-nine counties including, since 1996, Milwaukee. Originally, it offered only work in community services jobs, but in 1993 it was broadened to offer training and other services.25 Wisconsin has long been a leader in child support enforcement. Chapter 2 summarized the innovations that the state pioneered in the 1980s. The two pilot counties were leaders even by Wisconsin standards. Fond du Lac has a computerized system for tracking cases to verify payment and prioritize cases for attention by enforcement lawyers. Racine has a special Child Support Department that brings together the prosecutors who seek judgments against noncustodial parents and the family court commissioners who render them. The county’s dedication to support enforcement is intense. It pays a staff of thirty-five to hold 350 support hearings a week.26 In Racine, one manager proudly told me, “We take control.” The early signs were that Children First had improved on earlier child support efforts. It had a strong “smoke-out” effect—most of its clients complied with the requirements by paying their judgments rather than by attending the program. The proportion of parents paying and the amount they paid seemed to increase.27 This, of course, could reflect other causes than the program, such as a tendency for more cooperative fathers to be selected to participate. An experimental evaluation was also done in Racine, a federal requirement since the program drew waiver savings money. Despite positive early findings, the final results failed to show clear impacts on child support payments or employment. Clients in Children First improved their employment, earnings, and child support payments, but so did clients in a control group subject to Racine’s normal child support system. The latter, of course, was a tough standard to exceed. The evaluation was also compromised; out of good intentions, the program served many clients in the control group.28 However, counties that entered the Children First program in its early years did raise their collections over the period 1983–95 by more than did counties who undertook the program later or not at all. This does not prove an impact for Children First per se, but it does suggest that the program was one of a complex of initiatives that caused counties running the program to outperform the others.29 If the program did not achieve more, one reason is simply that the child support problem, like the teen pregnancy/school attendance problem, is profoundly difficult. For obscure reasons, absent fathers are harder to motivate to work than the welfare mothers served by welfare work programs. What made Children First different from the earlier child support regime, above all, was case management. The program provided the personalized assistance and supervision that the earlier system lacked. The men get more attention than they would from an ordinary work program,

Paternalism

•

163

because the manager is dedicated to their needs. One of his roles, as in Learnfare, is simply to explain to the clients the obligations they face, which they often do not understand. At the same time, he holds the clients accountable for the promises they make to judges regarding work effort. To a father assigned to him, the case manager is a personal authority figure whom he cannot escape. Case managers “don’t mess around” about establishing contact with their clients, one state official told me. In most counties, they go to the court hearings where the men are remanded to them and immediately confront them with a “real person.” This makes it much more likely that the men will in fact show up at the program and enroll. There the staff work out with them an appropriate schedule of activities. Most programs stress job readiness and work search, similar to JOBS.30 Staff then monitor compliance. If clients disappear, they promptly refer them back to court, where they may be cited for contempt and jailed. The managers thus wield both the carrot and the stick. In addition, Racine offers extra services funded by private foundations. These include basic education and a “parental responsibility class,” in which fathers are confronted with their responsibility to support their children. They are asked what obituary their children would write about them if they were to die today.31 This group is the brainchild of Jerry Hamilton, a noted case manager who personifies Children First. Tommy Thompson lauds him as a father to fathers: “He is warm and caring, but tough as nails on the fathers when he has to be. When the fathers come in the door, he tells them, ‘You’re going to join the fathers’ club and learn how to be a dad.’ And they do.”32 The idea is to teach obligations and generate pressure to fulfill them from both staff and other clients. Doing this at a personal level is effective, staff say, in a way that just levying formal requirements is not. Said one Kenosha case worker, “There’s nothing more motivating than to have someone bugging you.” As with Learnfare, an outside evaluation confirms much of the Children First experience. The Parents Fair Share (PFS) Demonstration was run in seven cities between 1992 and 1996 to test a similar enforcement model. PFS had strong smoke-out effects and raised the proportion of parents paying at least some support, but it did not clearly raise fathers’ employment or earnings. The program worked best, however, at the sites where staff successfully combined their helping and enforcing roles, and where case managers and child support officials collaborated well, as they did in Fond du Lac and Racine.33 For all its potential, Children First is still a small program. Although implemented in most counties, it serves only a small proportion of the noncustodial fathers who are potentially eligible, due mainly to limited state funding. In 1999–2001, the Children First budget was only $1.14 million a year, enough to fund only 2,850 slots, although counties man-

164

•

Chapter 8

aged to serve as many as 4,958 clients in 2000 using other monies. But there are an estimated 50,000 noncustodial parents who owe arrearages to the state on account of welfare.34 The will to implement such a program more fully appears to exceed even Wisconsin’s capacity, at least to date.

WORK PROGRAMS The most important development of paternalism occurred in WEJT and JOBS. As earlier chapters have shown, the expansion and toughening of these work programs throughout the state gave Wisconsin an ability to enforce work in welfare ahead of most of the nation. It was the routines developed then that allowed the state to implement the much more severe work demands of SSF, PFP, and W-2. At the core of the early work program was enhanced oversight of individual clients. In the liberal era of Wisconsin welfare, there had been few work expectations, and the early years of WEJT and JOBS aimed more to place clients in education or training than in available jobs. Relations between caseworkers and clients were trusting but limited. Yet even in this period, welfare staff had an urge to promote the independence of their clients.35 The goal of JOBS, DHSS wrote in 1990, was to “achieve economic self-sufficiency” by establishing “expectations of both recipients and program on the basis of responsibility and personal dignity”36— in short, through social contract. It was already clear from earlier welfare work programs that simply to levy work requirements had little effect. What brought the demands home to recipients was to hear about them directly from staff. The best programs used intense interaction with staff to motivate people to go out and seek jobs. Later research also shows that effective programs set high expectations for clients and stress moving them into jobs quickly.37 Wisconsin administrators already sensed this. Thus, when work expectations strengthened in the 1990s, they turned to case management to promote work. The intention, at least, was not to punish but to mobilize. JOBS staff wanted to get their “hands on” more clients, two state officials told me, because that was “positive” and “human” and it would save the clients from becoming “depressed.” Leading Counties In 1987, the state established two pilot projects in case management intended to promote employment. The grants went to Kenosha and Grant, two of the leading counties in JOBS discussed in chapter 5.38 Kenosha used the funds and other monies to create the formidable administrative

Paternalism

•

165

machine behind WEJT, which carried over to JOBS. The county required mothers to participate in WEJT from the moment they requested aid, so that they never knew welfare without a tie to work. The aim was to “capture people immediately—the first day they apply,” recalled Larry Jankowski, one of the Kenosha managers. Applicants underwent “immediate immersion right away” and faced eight weeks of activities before they even “came up for air.” And unlike earlier work programs, WEJT was open-ended. It did not terminate after several weeks or months and let mothers who had not found jobs go back on aid without working. Rather, said George Leutermann, there was “no exit” short of going to work, leaving welfare, or moving out of the county. To Jankowski, the key was “setting expectations” but also “providing quality services” so people could act to meet the standards. This, he added, “forces them to do stuff that deep down inside they know they should do.” To execute this concept, as noted in chapter 5, Kenosha created teams of eligibility, employment, and service staff that served individual clients. WEJT case managers went “one on one” with clients to resolve problems that might be preventing work or participation, calling on other team members to help. The teams also facilitated subsequent monitoring of clients. With everyone located together in the Job Center, case managers who heard of a client dropping out of the program could easily inform income maintenance to initiate a sanction. Or if income maintenance heard that a client had left a job, employment personnel could quickly be notified and get that person back into job search. There was also regular checking of client attendance at activities run by outside contractors. Case managers called in clients for a review every three months, and they were themselves closely supervised by their own managers. This was the machinery behind Kenosha’s outstanding activity levels. The other leading counties found Kenosha’s team structure too expensive because of the staff levels required, but they found other ways to achieve a similar result. In Grant, Jon Angeli used his extra administrative resources for front-end services, to divert the employable from welfare, and for interventions in troubled families by specially trained caseworkers. He aimed to be “nurturant,” but also demanding.39 The style of his staff could be distinctly personal and confrontational, to judge from interviews. Several of them told me of their determination to “get in the face” of clients if necessary to force them to address the “real issues” that kept them on aid. This style, they suggested was like “tough parenting,” necessary to set “boundaries” and get people to face their responsibilities. Grant allowed considerable discretion to field workers, in contrast to Kenosha’s tightly scripted monitoring procedures.

166

•

Chapter 8

Lacking the same special funding, Sheboygan operated with fewer staff and higher caseloads than the other leaders. However, by using large group orientations, the program still conveyed a strong work message to its caseload. Sheboygan also worked hard on verifying client attendance at education and training assignments at the local technical college and elsewhere in the community. Case managers claimed to work their caseloads intensively to be sure everyone was active. “I know what they’re doing,” one of them boasted to me. “Making them active, tracking them, not letting them sit there in limbo,” she went on, was the key to “motivation” and, hence, to jobs. Neither WEJT nor JOBS in Wisconsin was rigorously evaluated. But close staff interaction with clients has been a feature of the most successful welfare work programs evaluated elsewhere in the country—in Riverside, California, and Portland, Oregon.40 Other Counties One of the reasons the other counties I visited did not perform at this level was that they were less paternalistic. They monitored clients less intensively, so they obtained less activity and employment from them. Case managers in Racine, for instance, claimed to follow caseloads closely and check up on attendance at activities. But they sounded less zealous than their counterparts in the leading counties, and the checking appeared less frequent. This reflected a lesser dissatisfaction about traditional welfare in Racine. This was in contrast to child support enforcement, where the county monitored its cases intensely. An acceptance of undemanding aid was still clearer in Milwaukee, but the city also faced serious administrative limitations. It lacked the staff to monitor caseloads closely. Mostly, case managers tracked clients on their computers, without the personal contact and involvement seen in the leading counties. The machine could tell them where clients were assigned, but not whether they were actually fulfilling their assignments. One manager remarked to me cynically, “We know where everybody is; they just aren’t doing anything.” I did encounter some staff who yearned for a tighter and more demanding operation. One claimed to run a job club (a jobseeking group) as if it were a boot camp. “Clients respect the fact that we don’t put up with any crap,” he asserted proudly, “from them or employers.” Another manager bemoaned that clients were “still looking for a personal touch.” They wanted a “face” to deal with, and the “computer” was no substitute. The first thing counties seeking to improve had to do was build up case management. Dane in 1993–94 tripled its case managers in order to cut

Paternalism

•

167

caseloads, permit closer oversight, and thus reduce inactive cases. Fond du Lac in 1994–95 went over to Kenosha-like teams of staff in order to make its radical change from an education focus to Work Not Welfare. Both counties became more directive by putting more staff in direct contact with clients. In Milwaukee, it was the later shift to SSF and PFP in 1996 that finally forced the county to deal with clients on an individualized basis. Its inability to do this with conviction, however, led to its ouster from W-2. The work programs transformed welfare in Wisconsin. Learnfare and Children First had far lesser effects. It is fair to conclude that paternalism works better in welfare work than in the other programs. That may be because a supervisory regime was implemented better in the work programs over a longer period. More likely, the work problem in welfare is simply more tractable than truancy or child support. Enforcing school or child support pits government against the deeply rooted problems of the poor family—unwed pregnancy and the troubled ties of low-income fathers and mothers. In contrast, enforcing work cuts with the grain of what poor adults themselves can and want to do.

NEW HOPE The most important liberal experiment in paternalism was the New Hope Project.41 This was not strictly a welfare reform but an antipoverty program aimed at a broad slice of low-income America. New Hope ran between 1994 and 1998 in two poor neighborhoods of Milwaukee, one heavily black and the other heavily Hispanic. The inspiration for the project came from David Riemer, a liberal lawyer with roots in the Congress for a Working America, New Hope’s parent. As seen in chapters 3 and 6, Riemer played a key role in justifying work-based welfare reform to the left in Wisconsin. He argued that the main cause of poverty was not the welfare system, as conservatives said, but the fact that low-skilled jobs did not pay enough to support families. The answer was for government to guarantee a minimum income and services to all regular workers. A similar goal of “making work pay” lay behind the Clinton Administration’s welfare reform proposals in Washington. The Republican Congress elected in 1994 accepted an expanded EITC but rejected the Clinton health care plan. It also rejected the administration’s welfare reform plan in favor of the more severe PRWORA. New Hope then became the principal trial for a liberal version of workbased welfare reform.

168

•

Chapter 8

The Program New Hope made its offer to people in its target neighborhoods who were eighteen or over, willing and able to work, and who had an income at or below 150 percent of the federal poverty line. That included many twoparent families and single people, not all of them strictly poor, as well as the much needier female-headed families covered by welfare. To all these adults, New Hope promised a job (either in the private sector or, if necessary, a community service job, or CSJ), a wage sufficient to bring the family to the poverty line, plus subsidized child and health care. The benefits were available for three years. But to get any of them, clients had to work at least thirty hours a week. New Hope shared with the main drift of welfare reform a determination to refound aid on employment and to deemphasize education and training in favor of actual work. That common theme dramatizes the convergence of left and right in Wisconsin about how to deal with poverty. New Hope was, however, more generous in its eligibility and benefits than W-2. It covered all poor adults rather than just parents with children, and the income limit for cash benefits was higher. New Hope’s CSJs paid the minimum wage and qualified the holder for federal and state wage subsidies, as W-2’s CSJs did not. Child and health care were more generous as well. Perhaps most important, New Hope avoided W-2’s stress on diversion; it made no attempt to deter clients claiming benefits provided they worked.42 The Outcome Experience threw in doubt New Hope’s premise that poor adults were blocked from supporting themselves by barriers outside themselves. Over 12,000 adults were estimated to be eligible for the program in its two target neighborhoods, yet the project had great difficulty recruiting 1,362 of them to participate. Many others were blocked by illness or family problems. Once the program began, take-up of the benefits was surprisingly low, belying the idea that the clients were strongly in need. In a given month, less than half the eligibles claimed any of the benefits, in part because they did not work the required thirty hours.43 New Hope did record some good effects, but not of an order that would validate an idea of serious economic barriers to employment. For participants, gains in employment and earnings (relative to the control group) came mostly among people who were not working thirty hours when the program began, even though they used the benefits less than those already working. Clients already working full-time actually cut back their work

Paternalism

•

169

hours slightly, a phenomenon seen in earlier income guarantee programs. There were no clear reductions in welfare receipt.44 The program assumed that work was not worthwhile for the lowskilled without special benefits, but the program made less difference to its clients than its planners assumed. In Milwaukee, work already paid for most of those who were able to seek and hold jobs. Not only were jobs plentiful, but most workers could already get benefits such as health or child care from existing social programs or their employers. W-2 was more realistic about the causes of poverty. These lie more in the internal problems of families and the defeatism of many poor adults than in the external opportunity structure. Accordingly, W-2 focused more on changing the lifestyles of the poor and preparing them for work than on altering payoffs in the labor market. The Institution New Hope, however, succeeded splendidly as an institution, and in this it had much to teach W-2. The origins of the program showed the same capacity for collective problem-solving that appeared in the enactment of the Thompson experiments and W-2. Although proposed by a community group that was well left of center, New Hope garnered support from corporations and government, even though they were much more conservative. The Greater Milwaukee Committee, a business body, lined up funding from local firms, while Washington and Madison kicked in some of the savings in welfare that they expected from the program. Foundations, both liberal and conservative, also contributed. Cooperation was possible because each side addressed the leading concerns of the other. CFWA conditioned New Hope’s benefits seriously on work, rather than giving them away, and it made sure that its project was well administered and evaluated. Business and government, for their part, reacted to the enterprise in a practical spirit, asking whether it would work rather than rejecting out of hand any expansion of government. This impressive capacity to collaborate across normal political divisions reflects the high-mindedness of Wisconsin political culture. Even more notable was the New Hope organization. With a staff of only twenty-five, it administered a complex benefit structure for hundreds of clients. It also developed pathbreaking procedures that turned the traditional logic of welfare upside down. In the past, welfare determined whether applicants for aid fell short of some standard of need, then paid them a grant to make up the difference. Aid was presumed to be the basis for family income; earnings were an afterthought. In New Hope, however, earnings were the basis of income, and supplements were added to that. At the end of each month, clients had to turn

170

•

Chapter 8

in wage stubs to indicate their earnings. The project then calculated what government subsidies they would receive, then added any New Hope supplement needed to reach the poverty line. It did this using a computerized system custom-designed for New Hope. Administrators also verified that clients had averaged the thirty hours a week needed to qualify for benefits. Work-oriented Kenosha developed a similar procedure under JOBS, and something like this is bound to become routine in the work-based welfare of the future. Case Management The core of the New Hope organization was its case managers, each responsible for about seventy-five eligibles. These “project reps” oriented clients to the program, arranged their benefits, then adjusted them each month in the manner just described. The combination of eligibility and work oversight roles was much as with FEPs in W-2. The project reps of course helped clients who came to them seeking benefits, but since takeup was low, they also sought out nonparticipating eligibles to get them involved. That meant locating them, persuading them to work, solving service and family problems that might get in the way, then helping them get a job to satisfy the thirty hours, if necessary through a CSJ. The staff were coaches and hand-holders, but they also believed in work first. The emphasis was on meeting expectations in order to get benefits. If paternalism means help and hassle, New Hope fused the two more seamlessly than any program has done. Clients responded very positively. High proportions said that the help and encouragement they got from the project reps was the best thing about New Hope. Many contrasted the skill and resourcefulness of their reps favorably to the treatment they had received from case managers in welfare. The largest psychological effect of the program came in the social support experienced by families with children. They received markedly more practical assistance, counseling, and emotional support than families not in the program.45 Over time, New Hope suggests, case managers with powers to both help and obligate the dependent might begin to reverse their discouragement. New Hope also reveals that, once a serious work test is in place, the whole logic of welfare administration can be reversed. The priority is no longer to keep undeserving people out of the program. The work test—if well enforced—already accomplishes that. The risks actually run the other way: Families in need of assistance may now avoid it because the demands linked to aid seem too forbidding. Welfare then has to go into business—to reach out to eligibles and persuade them to work in order to get the support offered to the working poor. The job of welfare administrators is no longer to fence an entitlement around with restrictions.

Paternalism

•

171

It is to sell a social contract to which both society and the individual contribute. The future of welfare probably lies in various income subsidies and other benefits tied to work. To make such a structure effective, New Hope suggests, a paternalistic staff is essential. Poor families are unlikely to fathom the benefits without help, nor are they likely to seize them without encouragement at a personal level. Many poor will probably have to remain administratively dependent on government even as their financial dependence is reduced by earnings.

WISCONSIN WORKS Despite its differences with New Hope, W-2 was driven toward a similar paternalism after its initial implementation and the fall in the caseload. As originally designed, the program tried to minimize dependency through diversion and the “light touch”—giving families only the aid they needed to maintain their independence. Yet case management was built into the system from the start. FEPs might try to avoid giving people cash aid, but they would also offer the participant “an ongoing relationship to the extent necessary,” planners said. Even after a client was working, they would “continue to be available to offer advice and assistance.” They were supposed to help people “work through crises,” Andy Bush wrote, such as breakdowns of child care, so that they could keep working.46 FEPS are “presumed to exercise moral authority for the good of their charges,” one observer wrote.47 “I prefer to think of us,” one Milwaukee FEP told me, “as coaches and mentors.” They could play these roles because they were supposed to have caseloads of no more than fifty-five cases each. The dramatic caseload collapse at the outset of W-2 left a more disadvantaged caseload, for many of whom work was not immediately possible. Perhaps a third of recipients, for example, are victims of childhood sexual abuse, a legacy that can explain erratic lives in adulthood. As such cases became relatively more important, FEPs’ tasks shifted from “processing former welfare recipients,” one journalist wrote, to “managing the individual workaday needs of fragile families.” Parallel changes have occurred in other states.48 W-2 responded to this situation, first of all, through outreach to the community. After the transition from AFDC, several of the Milwaukee agencies feared that they had overdone diversion. They wanted to be sure that families in need would still approach them. As early as 1998, they began to run advertisements in local media to publicize their services, and they spread the word through community organizations. They also sent

172

•

Chapter 8

workers out to visit families that had been sanctioned, to persuade them to return to the system.49 A second response was the softening of the work test described in chapter 7. FEPs allowed more clients to go onto the CSJ and transition tiers of W-2 than planners expected, while forcing fewer into the private sector. A third response was that FEPs began to design individualized regimes of benefits and services for the most troubled families. This meant varying combinations of work assignments and remediation programs, including counseling or substance abuse treatment, for clients on W-2’s bottom two tiers. The agencies faced pressure to solve families’ problems because of the two-year time limit on placements. The agencies also faced incentives to keep people working once placed, because their performance measures included job retention as well as job entries. Clients who left subsidized for unsubsidized jobs were to receive “follow-up case management” for at least six months.50 Far from diverting clients, W-2 progressively downplayed the “light touch,” and in 2002 it was abandoned.51 Like New Hope, W-2 entered into ever-closer “engagement” with its clients. Maximus developed a “doorbell” service to wake people up for their jobs in the morning. FEPs now “call participants at home to make sure they get up to prepare their children and themselves for school and work, drive them to appointments or to child care when other options are not available, and provide reassurance and encouragement through extended verbal communication.”52 They had the time to do this because the caseload decline cut the number of cases they had to manage far below expectations. In March 1999, for example, Maximus had on board 164 total staff to serve only 1,976 total cases, 1,170 of them receiving cash.53 Fortunately, with its strong case managers and four tiers, the original design of W-2 accommodated this shift toward paternalism. After visiting one of the Milwaukee agencies, one close observer remarked that “In some respects, if I were blindfolded I would have thought I was in Oregon.”54 He referred to the exemplary Portland JOBS program. It enforced participation stringently but was notably less given to diversion and work first than W-2 had been. Thus W-2 compensated for the severity of its work demands and its initial implementation. The long-term acceptability of the program probably depended on this.

BEYOND WELFARE REFORM Paternalism may well represent the future of welfare. Work-based welfare reform is rapidly deflating the welfare rolls around the country, as it has already done in Wisconsin, but there is no prospect that public assistance

Paternalism

•

173

will disappear. Poor families encounter their most serious problems at home, not in the workplace. Intimate difficulties stemming from family history are bound to keep some troubled families as “wards of the state.”55 And even for those who are working, some degree of ongoing support and assistance probably is essential for many to work steadily and gain better jobs over time. Again, an administrative dependence on government may well outlive financial dependence. That prospect is perhaps nowhere closer in America than in Grant County. There Jon Angeli and his staff helped create the work first and diversion strategies that transformed welfare at the local and then at the state level. Virtually all of Grant’s former welfare cases are now working, but Angeli does not think his task is over. Clients may no longer be on cash aid, yet they still need help to keep working and to advance. They have to address the more personal problems that may bar them from employment. They have to learn “basic life-skills,” Jon writes. There is, he says, a need for “individualized support services that modify self-defeating behaviors, develop healthy family-systems, and encourage positive peer groups within the community.” For welfare, that means there must be “a paradigm shift that transcends our previous focus on employment.”56 So like New Hope and other agencies in W-2, Angeli has gone into business. Realizing that his former clients would not automatically return to him, he set out to “entice them in new and different ways.” His welfare department and the local Job Center created an organization called Grant County Works. GCW markets job retention and development to former recipients in the area and their employers. It runs training sessions for low-income workers on such subjects as “dealing with difficult people,” “motivation in the workplace,” and “managing conflict.” According to Angeli, employers are more willing to hire low-skilled workers when they know these backup services are available.57 But now that W-2 has declined to only a handful of cases, Jon wonders how he can maintain the paternalistic structure necessary to do all this. The immediate issues are operational. The funding Grant gets as a W-2 contractor is no longer enough to support the caseworkers and the other apparatus Angeli needs to develop his mission to the working poor. As discussed in the last chapter, the meaning of “caseload” must be redefined in some way that recognizes welfare’s ongoing duties to help people who have left traditional aid. Another question is authority. Welfare’s ability to enforce work traditionally depended on its ability to condition aid on work requirements. How can it still promote work when people have left cash assistance and thus no longer have an obligation to the agency? Purely voluntary work programs have not, in the past, been as effective. One possibility is to

174

•

Chapter 8

implement work tests in Food Stamps more seriously, as this is a program on which many working families that have left cash aid continue to rely. Angeli to date has counted more on appealing to the self-interest of workers and their employers. A larger issue has to do with administrative quality. While paternalism helps to justify demanding programs, it also raises questions of arbitrariness. Those fears are least when staff enforce requirements that are clearly stated in law. They are greater the more discretion is involved, and some discretion is unavoidable. The more lower-level staff, rather than more accountable officials, make decisions, the more their ability and dedication matter. They must be competent, and they must act out of real concern, not just routine. The gruffness of a John Gardner would not get the response it does if he had not previously proven his dedication to his clients. These virtues are most likely to exist in places like Wisconsin where bureaucratic quality runs high. Could less-gifted localities do the same? The final issues surrounding paternalism are political. There may be strong policy arguments for supervisory programs. But can it finally be legitimate for programs to manage the dependent in this personalized way? Even if there were no discretion or abuse, paternalism inevitably reflects on the dignity of its clients.58 That cost is accepted for the greater good of overcoming poverty. Such a choice may be acceptable in a rural county like Grant, or in a state like Wisconsin, where the culture stresses duties over rights. It would be less legitimate in more rights-oriented parts of the country. More liberated states like New York or California may simply be unable to overcome poverty in ways consistent with their political values. I return to these issues in chapter 12.

CHAPTER 9

The Decline of Welfare

THIS IS THE FIRST of two chapters on the consequences of welfare reform in Wisconsin.1 We now address the bottom line in table 1.3: Were the state’s policies, although apparently effective, actually successful? This chapter focuses on the dramatic decline in the state’s caseload, from its height in 1986 through the early years of W-2. Did welfare reform produce the decline? Chapter 10 will take up the other effects of reform. The decline of welfare in Wisconsin, and the nation, has proven surprisingly hard to explain. It might seem obvious that welfare reform reduced dependency, since that was its leading goal, next only to raising work levels on the rolls. But reform got serious in most states only in the middle 1990s, when other conditions also favored a reduction in dependency. Economic conditions were the best in a generation. There was also a buildup of new government benefits designed to promote work. Economic studies differ in how much of the fall they attribute to welfare reform. In Wisconsin, I first asked local officials what they thought was driving welfare down. I then tested those ideas with quantitative analyses that showed how counties with more and less caseload fall differed, in policy and otherwise. In a similar way, I compared Wisconsin to other states. In both analyses, the jurisdictions with the greatest caseload decline were those that enforced work and child support best, although economic conditions and other factors also mattered. These results do not strictly prove that welfare reform drove the rolls down, but they are consistent with that view. Wisconsin stands out from other states for its willingness to combine high welfare benefits with stringent demands for work and child support. That combination virtually abolished traditional family welfare, replacing it with the new regime we have already seen.

RESEARCH ON CASELOAD CHANGE In Wisconsin, the AFDC caseload fell by around 90 percent from its crest in 1986 through the implementation of W-2 in 1997–98 (see Table 7.1). The decline was almost continuous, but for a slight increase during 1991– 92, the years of recession. The national caseload of AFDC (later TANF)

176

•

Chapter 9

underwent two dramatic shifts in the same period. After changing little since the mid–1970s, the number of families on the program jumped 32 percent between 1989 and 1994. But cases then plummeted by 58 percent between 1994 and 2001, a decline without precedent.2 Figure 1.1 portrays the trends. The changes have attracted much research attention, but finding clear causes has been difficult. Some studies simply fail to account for the changes well. One reason is that welfare levels shifted far more sharply than the background social and economic conditions that scholars usually rely on to explain them. The growth of 1989–94 was probably caused in part by rising unwed pregnancy and the recession of the early 1990s, but much more must be involved.3 As for the decline since 1994, field observers suggest that a synergy occurred. Work tests drove recipients off the rolls into the arms of waiting employers.4 In research by economists, the relative importance of reform and other factors is unclear. Some studies say welfare reform was the largest cause, but with important roles for low unemployment or the Earned Income Tax Credit. Reform becomes more dominant the closer one comes to the present.5 The EITC, which was raised in 1990 and 1993, is a subsidy paid by the federal tax system to workers with low earnings. It currently boosts their wages by as much as 40 percent, worth up to $4,000.6 Other authors, to the contrary, rate the economy as more important than welfare reform, even much more so.7 Linked to the welfare decline has been an employment rise. Poor adults work at much lower levels than the better-off, but as welfare rolls have fallen, their work levels have soared. In 1993, only 44 percent of poor female heads of family with children worked at all. That rate soared to 64 percent by 1999, before dropping to 58 percent in 2001 in the recession.8 That gain might have come from stiffer work demands in welfare, but one economic study attributes it mainly to the growing EITC.9 The tax credit did give low-skilled people who were not working more reason to do so, although the subsidy also allows those already working to cut back their effort without losing income. Work incentives in welfare had not previously shown much influence on work levels.10 One reason for the uncertainty, I suspect, is the difficulty of measuring welfare reform. The economic studies gauge it by whether a state had a federal waiver of varying kinds (e.g., to toughen work tests, impose time limits, etc.). The waivers were mostly granted by Washington in the 1990s, prior to PRWORA, to states wishing to experiment with welfare. As noted in chapter 2, Wisconsin had the largest number of these permissions. But waivers indicate only very crudely how far new requirements are enforced “on the ground,” as recipients experience them.

Decline of Welfare

•

177

A second problem is that if one measures reform using waivers, one cannot tell what specific policies this means. Other national data bases do not help. Statistical research on welfare is based mostly on government surveys or other data bases gathered for other purposes. Researchers themselves seldom have any direct contact with welfare or its recipients. Economists typically assume that if benefits such as AFDC or EITC change behavior, they do so through altering incentives. Recipients are assumed to react to payoffs so as to maximize their income, subject to constraints. But as argued in the last chapter, the poor often do not optimize in this sense. To get them to “do the right thing,” social programs increasingly use administrative suasions. And whether a family is subject to work or child support requirements is not indicated in the standard data bases. One can tell whether individuals are receiving welfare, but not whether they also face pressure to work or conform to child support requirements. To indicate whether a state had a waiver for welfare reform is a poor substitute for directly measuring the degree of work or child support enforcement. Thus, the economic studies may underplay the role of welfare reform in driving caseloads down.

THE WISCONSIN CASELOAD DECLINE Wisconsin’s caseload decline has aroused much the same uncertainty as the national welfare fall. Since the state’s decline outpaced that of virtually all other states, one might presume that the chief cause must be welfare reform. The state has been far more distinctive in its welfare policies than in its social or economic conditions. But it is difficult to connect the decline specifically to the state’s actions, at least in the early years of reform. The caseload starts to fall in 1986, well before any of Tommy Thompson’s reforms were even legislated, let alone implemented, so how could the decline be due to them? The only policy change that clearly reduced the rolls was the benefit cut of 1987, which simply made fewer people eligible for aid.11 According to one economic study, Wisconsin had the largest decline in dependency per capita of any state between 1987 and 1996, but 53 percent of it was due to economic conditions and only 11 percent to welfare reform.12 Economic conditions, however, were not notably better in Wisconsin than in other states. Perhaps unemployment in Wisconsin was only 4.4 percent in 1989, but eighteen states had lower rates. And while unemployment rose in Wisconsin only 0.7 points during the recession years of 1990–92, thirteen states were more fortunate.13 Most of these other states

178

•

Chapter 9

had much smaller caseload declines. Again, the economy cannot be the whole story.14 My approach to explaining the decline was to begin with direct inquiry and build up to statistical analyses. I and a research assistant interviewed welfare officials at the state level and in five counties in 1995.15 Local people who serve recipients have the most direct sense of why they are leaving the rolls. We asked officials what was causing the caseload to fall. I then tested these theories with several quantitative analyses. In the statistical work, I measured welfare reform using actual program data, rather than waivers. This reveals more about the actual policies driving change. The result, I think, is a more true-to-life understanding of what was going on. The drawback is that I can only compare jurisdictions within the state; I cannot directly explain the overall caseload’s decline over time, as the economic studies do.16 Nevertheless, my results strongly suggest that welfare reform was the principal driver behind the caseload fall. This research centers on the mid–1990s, prior to the more dramatic caseload decline caused by the diversion programs and W-2. But this earlier period is the one that arouses the most dispute. If welfare reform had strong effects then, we can be sure it did later, when the state adopted more drastic policies. Wisconsin in that period was pursuing the milder sort of work enforcement that many states are pursuing today, to judge from what we know of the implementation of TANF.

WHAT OFFICIALS SAID The reasons our respondents gave for the caseload decline suggested that welfare reform was uppermost. Table 9.1 breaks out responses by DHSS staff from those from the counties. The leading reason—cited by well over half of respondents—was the buildup of welfare work programs. Chapter 4 detailed how WEJT and JOBS were strengthened to make the work test real in much of Wisconsin, well before W-2. Administrators claimed they could see the results. Every time we “ratchet up the manditoriness,” one state official remarked, it “dents the caseload.” Some respondents asserted that the inauguration of WEJT in 1987 brought work programs to a critical mass, triggering the caseload decline. The programs in themselves probably did not do this at first, due to their heavy initial emphasis on education and training. The state realized by 1989 that WEJT had not cut the cost of welfare.17 Nevertheless, WEJT and JOBS did establish procedures to involve the caseload. That gave the state leverage when it later moved to demand higher participation and work first.

Decline of Welfare

•

179

TABLE 9.1 Reasons Given for Wisconsin Welfare Caseload Fall by Officials Interviewed, May–June 1995 (percent)

Work programs The economy Expectations Support services Child support Benefit cuts Administrative progress Demographic change Other Don’t know

DHSS Staff

County Staff

Overall

57 52 67 24 29 24 19 14 5 14

59 51 21 26 5 0 3 3 0 15

58 52 37 25 13 8 8 7 2 15

Source: Interviews of 21 DHSS and 39 county officials, May–June 1995. Note: “Don’t know” includes respondents who did not speak to the issue. Percentages add to more than 100 because respondents could mention multiple reasons.

Second in importance in explaining the decline, to our respondents, was economic conditions. There was a clear-cut “labor shortage,” the same state official remarked, making it easy to find jobs; job readiness, he joked, now demanded no more than “casting a shadow and having a pulse.” The effects were perhaps most apparent in Milwaukee. One might have doubted the ability of the local economy, with its recent loss of factory jobs, to absorb the legions of unskilled adults shifted off welfare by work enforcement. But the boom of the 1990s, drew even very marginal workers into the workforce. By the late 1990s, welfare recipients in Milwaukee were gaining jobs that paid an average of $7.39 an hour and averaged thirty-five hours a week; 72 percent of these jobs offered health insurance. This despite the fact that only 57 percent of the hires were high school graduates and only 68 percent had any recent work experience.18 The third factor is barely suggested in the economic studies—the changed expectations created by the reform process itself. Many of our respondents claimed that Tommy Thompson’s experiments and the controversies about them had depleted the rolls by giving a message that “work” and “responsibility” were now to be expected of recipients. The caseload was falling, Jean Rogers said in 1995, simply because “we keep coming at them” with one innovation after another. The “endless drumbeat” that welfare was changing, Gerald Whitburn asserted, had done more to motivate people to leave the rolls and work even than the JOBS program. Tommy Thompson’s “leadership” had given “a very clear and consistent message,” another state official said, and that “makes a tremen-

180

•

Chapter 9

dous difference.”19 While state officials might be expected to tout the governor’s accomplishments, we heard the same from some local welfare people, some of whom criticized the conservative direction of change. Thompson’s early experiments probably did little, in themselves, to reduce the caseload. Learnfare failed to show clear effects and was not even aimed at cutting AFDC in the short run. Several of the waiver programs improved work incentives, but research and experience do not suggest that this would raise work levels by much. Most of the experiments affected only selected counties. Yet our respondents repeatedly asserted that all these programs had driven the rolls down by changing the moral climate around aid. Learnfare especially had done this. Recipients got a “message of responsibility,” one respondent said. It was the “combination of the program and the rhetoric,” another said, that motivated many to leave the rolls.20 A quarter of officials mentioned the support services provided as part of reform—especially child care but also training and transportation—as a cause of caseload decline. If one combines this factor with the program buildup and the change in expectations, then welfare reform in all these guises becomes the dominant cause of welfare decline in Wisconsin, at least in the eyes of these officials. In addition, our respondents often spoke of less visible bureaucratic improvements that had helped to cut the rolls. The state had implemented CARES, its new state-of-the-art welfare computer system. CARES calculated benefits automatically, thus reducing errors. Teams of state and local officials had reduced errors and fraud in some regions. The picture one gets is of a problem-solving bureaucracy relentlessly grinding down its challenges. With every advance, the state grew more able to help its dependents and to enforce obligations on them. For the poor in Wisconsin, there is less and less escape from the supportive but controlling hand of government.21 Child support is rarely mentioned in academic studies of caseload change,22 and yet our respondents often asserted that it was one reason for the decline in Wisconsin. The child support reforms mentioned in earlier chapters had begun to bear fruit. With a uniform payment scale and automatic withholding, officials could update child support orders more quickly and thus collect more support. Higher and more reliable support payments, in turn, allowed more single mothers to make a life off public aid, especially if they worked at least part-time. In 1997–99, in fact, child support was the most important subsidiary income source for families leaving welfare due to employment.23 Equally significant was what our respondents did not mention. While nearly a quarter of state-level staff cited the Thompson benefit cut as reducing the rolls, no local official did. They simply had not observed it to

Decline of Welfare

•

181

TABLE 9.2 Numbers of JOBS Participants and AFDC Cases, Wisconsin, 1994–96 JOBS Participants

AFDC Cases

Year

Number

Percent Change Over Prior Year

Number

Percent Change over Prior Year

1994 1995 1996

29,862 32,913 39,787

10.2 20.9

76,217 70,604 55,501

−7.4 −21.4

Source: Department of Health and Social Services and Department of Workforce Development. Note: Years are calendar years. Figures for participants and cases are monthly averages for the years.

cause any dramatic fall in the rolls. A few respondents credited demographic changes, such as a fall in unwed pregnancy. As for the EITC, the addition of a state EITC to the federal credit made wage subsidies unusually generous in Wisconsin. And yet no official cited this as a reason why people were leaving welfare.24 To them, work programs and child support were far more salient. TRENDS OVER TIME Thus, our respondents suggested that rising work enforcement was the leading cause of caseload decline, although other factors mattered. Overall state caseload trends support this idea. Table 9.2 shows how Wisconsin’s AFDC rolls and its enrollment in JOBS shifted in opposite directions in the mid–1990s, when the work reforms began to bite. As participation in JOBS rises, so the welfare rolls contract just as sharply, and by very similar percentages. As noted in chapter 4, families left welfare through job placements, but also diversion and sanctions. A more telling question is how trends differed among the counties. If welfare reform was the main force behind the caseload fall, then we should expect to see more decline in counties with more demanding work policies. Largely, that is what we find. Table 9.3 shows caseload fall in the ten research counties discussed in chapter 5. The period covered is from January 1987, the beginning of the Thompson Administration, through the end of 1995, just before the onset of Self-Sufficiency First and Pay For Performance, which tended to flatten out differences among the counties. The left-hand column of figures shows that, over the whole period, caseload decline was generally greater in counties that were reform lead-

182

•

Chapter 9 TABLE 9.3 Change in Caseload in Research Counties, 1987–95 (percent, with January 1987 as base) County

Jan. 1987– Dec. 1995

Jan. 1987– Oct. 1994

Oct. 1994– Dec. 1995

Leading Counties Kenosha* Grant* Sheboygan*

−35 −69 −62

−26 −63 −42

−9 −6 −20

Changing Counties Dane Winnebago* Fond du Lac*

−16 −57 −71

−4 −40 −43

−12 −17 −27

Lagging Counties Racine Marathon Douglas* Milwaukee

−39 −39 −45 −11

−28 −30 −40 −4

−11 −9 −5 −6

County average State

−54 −33

−44 −24

−11 −9

Source: Calculated from county caseload data from the Department of Health and Social Services. Note: Starred counties were pilots for Work First or Work Not Welfare prior to December 1995. Leading, changing, and lagging counties are defined as in chapter 5.

ers, or were early pilots for Work Not Welfare or Work First, than in the other counties. The principal exception is Kenosha, which refused to push work first and closing cases as strongly as the state wanted. As the middle column shows, none of these counties, with the exception of Grant, had reduced its caseload by as much as the average of all counties through October 1994, just before Work First began. Most of my research counties were large, while the average reflects many small counties that found it easier to reduce their rolls. After the new programs came on stream, however, the right-hand column shows that Fond du Lac, Sheboygan, and Winnebago considerably exceeded the county average through the end of 1995. As the principal WNW site, Fond du Lac has a particularly sharp fall. Grant lagged, probably because of its unusual earlier decline, as did Kenosha and Douglas, probably because of their resistance to work first. Dane and the lagging counties trailed the leaders. For Dane or Milwaukee, the largest and also the most liberal jurisdictions,

Decline of Welfare

•

183

serious work enforcement did not begin until the start of Self-Sufficiency First and Pay For Performance in March 1996. One indication that policy, and not something else, explains the differences is that caseload decline in the Work First and Work Not Welfare counties accelerated as the new programs began. The two groups of counties had reduced their caseloads by 40 and 47 percent, respectively, from January 1987 through late 1994, a period of almost eight years. Work First began in November 1994 and WNW in January 1995. The Work First counties then reduced their caseloads a further thirteen points, and the WNW counties a further twenty-four points, just through the end of 1995.25 No other change could account for this.

A COUNTY-BASED ANALYSIS These analyses do not explicitly take into account other causes of the decline mentioned by our respondents, such as the economy. Counties may differ, not only in their welfare policies, but in how employable their clients are and, above all, in how favorable economic conditions are. So I constructed statistical models across all the counties, parallel to the models of JOBS performance seen in chapter 5. These relate the extent of caseload change in a county over 1986–94 to the county’s JOBS policies as well as the demographics of its clients and the local labor market.26 Based on the interviews and prior research, I expected that the various factors would affect caseloads as follows: • The JOBS Program: The higher the percentages of JOBS clients in program activities, and the lower the share in inactive statuses, the more people should enter work and the faster welfare should fall, if other conditions are equal. • Child Support Enforcement: The more child support collected in a county, the greater should be its caseload decline. • Population Trends: The more population is growing in a county, the more its welfare rolls should grow, or the slower they should fall, all else being equal. • Demographics: If the welfare caseload is more disadvantaged or less employable, or unwed pregnancy is higher, welfare should grow faster or fall slower. • Labor Market: If unemployment is lower or other economic conditions are more favorable, welfare should fall faster.

This was the general model. I tested several measures of each of these factors. Table 9.4 summarizes those that had a clear influence on caseloads during 1986–94.27 My hypotheses were largely confirmed. Control-

184

•

Chapter 9

TABLE 9.4 Determinants of AFDC Caseload Change, Wisconsin Counties, 1986–94 Compared to the average county, a county had A larger caseload decline if it had more Clients in active status, 1991 Increase in clients in active status, 1991–93 Rise in child support collections as percent of AFDC benefits, 1989–94 Increase in number of persons employed, 1986–89 A smaller caseload decline if it had more Clients in inactive status, 1991 Increase in clients in inactive status, 1991–93 Welfare adults aged 21–25, December 1986 Welfare adults Asian, December 1986 Welfare adults black, December 1986 Increase in persons employed, 1990–94 Increase in the unemployment rate, 1986–89 Increase in the unemployment rate, 1990–94 Source: Lawrence M. Mead, “The Decline of Welfare in Wisconsin,” Journal of Public Administration Research and Theory 9, no. 4 (October 1999): 608–10.

ling for other differences, counties had a larger caseload decline if they had high and growing proportions of welfare recipients active in the JOBS program, and if they had rising child support enforcement, all compared to the average county. Conversely, counties with high or rising proportions of clients in inactive status tended to have smaller caseload decline.28 Several caseload features also mattered; counties with high proportions of clients who were Asian or black had slower caseload decline, because these groups tended to be more disadvantaged and less employable than the norm. And the economy mattered. Rising employment during 1986–99 speeded caseload decline, while rising unemployment rates in 1986–89 and 1990–94 retarded the decline. It is unclear why having more adults aged 21–25 and rising employment during 1990–94 slowed caseload decline. The overall model was unusually powerful: Over 70 percent of the variation in caseload change across the counties is accounted for by these factors. The policy and demographic terms—not the economy—dominate. The JOBS and child support variables by themselves account for 48 percent of the variation, the demographic terms 46 percent, but the economic terms only 7 percent. Again, the results explain variation in caseload change within the state, not the overall caseload decline. But they show more policy detail than the economic studies. And they are at least consistent with the view expressed by our respondents that welfare reform was the main force behind change.

Decline of Welfare

•

185

I also did analyses for subperiods. A model for 1986–89 showed that counties that had run WEJT in 1989 had more caseload fall than others in this period, all else being equal. This confirms the idea that WEJT gave the state a running start toward JOBS and the state’s later work enforcement programs. A model for 1990–94 showed that a county’s caseloads were reduced, relative to the average, not only by activity in JOBS, but also by a rising level of sanctioning, or reducing clients’ benefits for noncooperation with JOBS. This may reflect efforts by the state in this period to get the counties to enforce the work tests more stringently.29 Some might doubt that results like this apply to Milwaukee, due to its disadvantaged caseload and doubtful labor market. But here as in the JOBS performance models in chapter 5, Milwaukee did not stand out as special.30 Apparently, what was most distinctive about the city was not its demographics or its economy but its performance in work and child support programs. Milwaukee simply did not enforce work or child support as well as most other counties. This view was confirmed by what happened after 1994, when the state demanded change from the city and obtained it. The more severe work policies of 1996–98 drove the caseload down in Milwaukee almost as fully as in the rest of the state. The environment proved to be no bar. COMPARISONS TO OTHER STATES A final question about the caseload decline is why Wisconsin was so different from most other states. Not only did it have virtually the greatest fall of any state, but it had vastly the largest decline among northern states with high benefits and big caseloads. To explore the reasons, I constructed statistical analyses of caseload change across the states that were much like those using the Wisconsin counties above. Again, these models explain differences in caseload change among jurisdictions. They do not, like the economic studies mentioned earlier, explain the overall trend in the national caseload. But again, the payoff is that we can measure state policies using actual program data and thus learn more about the forces that were probably driving change. I predicted that states would have less caseload rise, or more decline, if they had: • High or rising participation levels in welfare work programs • High or rising assignment of clients in such programs to job search or actual work, rather than education or training • High or rising enforcement of child support • Low or falling welfare benefit levels, as these features reduce eligibility for aid

186

•

Chapter 9

• High levels of governmental quality • Caseload demographics tending to reduce dependency—less unwed pregnancy or greater employability • Favorable economic conditions, particularly low unemployment

These expectations resemble those I had about the county-based analyses above, but now I can add benefit levels, because they vary across the states. I can also investigate the role of governmental quality more rigorously. Among the Wisconsin counties, I had to gauge governmental quality judgmentally, but across the states there are measures of various characteristics of government and politics. Findings from these variables are part of the argument linking governmental quality and successful welfare reform that I make in chapter 11. As before, I tested various measures for each of the above dimensions and assembled those that clearly affected caseloads. I constructed two models, one for the period 1989–94 when caseloads grew in most states, and the other for 1994–98, when caseloads fell in every state. In the first period, Wisconsin was one of only four states to record any decline, the others being Arkansas, Louisiana, and Mississippi. In the second period, Wisconsin had the third-greatest decline of any state, exceeded only by Idaho and Wyoming. No other urban, high-benefit state approaches this record. What explains it? The Period of Growth, 1989–94 The results for 1989–94 are summarized in table 9.5. States that had more of their welfare adults active in JOBS in 1991, and which increased the share of JOBS clients assigned to job search over 1991–94, had less caseload increase, compared to the average. A state also had less increase if it collected child support in a higher percentage of AFDC cases. Cutting the other way, states had more caseload growth if they paid higher benefits, which qualifies more needy families for aid, and also if they added eligibility for two-parent families (AFDC-UP) after 1989. This coverage, previously optional, was mandated by the Family Support Act of 1988. Twenty-two states, mainly in the South, had to add it. The government quality term that mattered in this period had to do with “reinventing government.” States that implemented such administrative reforms (e.g., performance measurement, decentralized decision making, a focus on consumer satisfaction) more fully had less caseload increase.31 This is our first indication that governmental performance has a palpable influence on welfare outcomes, even controlling for the policies that have the most immediate effect on clients.

Decline of Welfare

•

187

TABLE 9.5 Determinants of AFDC Caseload Change, American States, 1989–94 Compared to the average state, a state had A smaller caseload increase or more decline if it had More clients active in JOBS, fiscal 1991 More increase in JOBS clients in job search, 1991–94 More recipients receiving child support, fiscal 1989 Better implementation of reinvention of government More state population already on AFDC in 1989 A higher average AFDC case size, fiscal 1989 A larger caseload increase or less decline if it had A higher maximum AFDC grant, January 1989 Added two-parent (AFDC-UP) coverage after fiscal 1989 More growth in population, 1989–94 More increase in unemployment, 1989–94 Source: Lawrence M. Mead, “Caseload Change: An Exploratory Study,” Journal of Policy Analysis and Management 19, no. 3 (Summer 2000): 465–72. Note: The District of Columbia is omitted due to missing data.

The significant demographic influences were those affecting the demand for aid rather than the employability of the caseload. A state had less caseload increase if it had more of its population already on AFDC in 1989, but more increase if its population rose. Welfare growth was also lower in states that had larger welfare cases in 1989. Size of case is a proxy for unwed pregnancy. That is, states that had larger welfare cases in 1989 typically had older welfare mothers with more children, hence less unwed pregnancy among young mothers and thus less welfare increase in the ensuing years. Other studies agree that rising unwed pregnancy helped drive the welfare boom.32 Finally, as in other studies, caseload increase was larger if unemployment rose. Like the county-based analysis, the model is extremely strong, accounting for over three-quarters of the variation in caseload change across the states. The Period of Decline, 1994–98 The results for 1994–98, when caseloads were falling, are shown in table 9.6. The forces differentiating states were somewhat different. As in the earlier period, states had lower caseloads if they had more JOBS activity and better child support enforcement. They also had sharply lower caseloads if they had stronger sanctions, meaning grant reductions for failure to cooperate with JOBS. That meant canceling the grant entirely, either after a delay or immediately, rather than reducing it partially. Sanctions were not measurable in the earlier period. Higher caseloads resulted if, within JOBS, they assigned many clients to postsecondary education; this

188

•

Chapter 9

TABLE 9.6 Determinants of AFDC Caseload Change, American States, 1994–98 Compared to the average state, a state had A larger caseload decline if it had More increase in clients active in JOBS, fiscal 1994–96 Delayed full-family sanction Immediate full-family sanction More recipients receiving child support, fiscal 1994 A higher average AFDC case size, fiscal 1994 A smaller caseload decline if it had More JOBS clients assigned to postsecondary education, fiscal 1994 A higher maximum AFDC grant, January 1994 A higher rank for legislative effectiveness (CCSL) An individualistic political culture (Elazar) Source: Lawrence M. Mead, “Welfare Caseload Change: An Alternative Approach,” Policy Studies Journal 31, no. 2 (2003): 163–85. Note: Alaska and the District of Columbia are omitted due to missing data. Legislative effectiveness term is reversed so that higher values mean better rather than worse.

corroborates evaluations showing that work programs that favor education perform worse than those pushing work first. As before, states also tended to have higher caseloads if they had higher benefits. Two terms expressing government quality are now important. A state’s legislative effectiveness, as measured by the Citizen’s Conference on State Legislatures, reduced caseload decline, perhaps the reverse of expectations.33 The reason may be that welfare reform has generally been led by governors and welfare executives.34 Stronger legislatures may indicate weaker governors. The other governmental variable indicates whether a state is individualistic in political culture in Elazar’s sense, introduced in chapter 3. According to Elazar, individualistic states are more willing to accept diverse individual preferences than states that are moralistic or traditionalistic. That could mean higher dependency, as this result suggests.35 The only social or economic term that mattered was the same case-size term as in 1989. A search for other demographic influences failed. The decline in caseloads, like the rise, tended to hit all caseloads, regardless of makeup. Despite the weakness of these terms, the model still accounts for over 60 percent of the variation in caseload decline across the states.

WHY WISCONSIN WAS DIFFERENT We can use these models to investigate why Wisconsin was unusually able to restrain its caseload in the first period and drive it down in the second. Since the state’s decline was reasonably well predicted by both models,

Decline of Welfare

•

189

TABLE 9.7 Values of Wisconsin and Other States on Model for 1989–94

Factors promoting lower caseloads Percent of welfare adults active in JOBS, fiscal 1991 Change in percent of JOBS participants in job search, fiscal 1991–94 Percent of AFDC cases getting child support, fiscal 1989 Score for implementing reinvention of government Percent of state population on AFDC in 1989 Average AFDC case size, fiscal 1989 (persons) Factors promoting higher caseloads Maximum AFDC grant (family of 3) in January 1989 (dollars) State added two-parent coverage after fiscal 1989 Percent change in population, 1989–94 Change in unemployment rate, 1989–94 Caseload change (percent) Predicted (average 33.6) Actual (average 33.6)

LA

MS

CA

NY

IL

WI

–

–

–

–

–

+

+

–

+

–

+

+

–

–

–

–

–

+

+

–

–

+

+

+

+

+

+

+

+

+

a

+

+

a

+

+

–

–

+

+

–

+

+

+

–

–

–

−

–

–

+

–

–

–

–

–

+

+

–

–

32.5 28.1

4.9 12.7

−8.1 −7.6

8.9 −7.7 53.6 −10.3 −11.2 49.7

Source: See table 9.5. Note: Excludes the District of Columbia due to missing data. The letter “a” means that a state’s value on a variable equaled the average for all states, “+” that it was above the average, and “–” that it was below the average.

the reason must be that Wisconsin differed from other states in the factors that influence caseload change in these analyses. Table 9.7 lists the same variables from the 1989–94 analysis as in table 9.5. The table shows how the values on these terms for Wisconsin and several other states compared to the average value for all states used in the analysis. States that are above average on the factors promoting lower caseloads will tend to have a smaller caseload increase, or an actual decline, compared to the average increase, which was 34 percent. Conversely, states that are above average on the factors promoting higher caseloads will tend to have a larger increase. At the bottom I give the caseload change predicted for each state by the model, then its actual

190

•

Chapter 9

change. All these states are quite well predicted except Lousiana, where a small increase was expected rather than a decline. Louisiana and Mississippi were the only states that reduced their caseloads by more than Wisconsin. The model suggests that they achieved this mainly because of favorable social and economic conditions, not programming. They had a high initial dependency rate, low grant levels, low population growth, and less growth in unemployment than the average state. This allowed them to control caseloads even though they were below average in activity levels in JOBS and child support enforcement, and they had to add two-parent welfare coverage, all which tended to increase caseloads. The experience of the big urban states was very different. California and New York had high benefits and unemployment growth well above average. California also faced high population growth. These expansive forces were not offset by strong enforcement in work or child support, so a large caseload increase became unavoidable. Illinois did somewhat better, mainly due to lower population and unemployment growth. Wisconsin offered benefits almost on the level of California or New York. Its growth in population and in unemployment were slightly below par for the nation, but they were above the other states in the table except for California. Nevertheless, Wisconsin managed to stymie caseload growth with enforcement. It was easily the highest of these states in JOBS activity and child support. It also managed to shift JOBS marginally toward a job search in a period when most states shifted away. Table 9.8 shows the same results for the model for 1994–98. Now Mississippi achieved a greater-than-average caseload decline in part through a tough sanction policy, but it continued to display weak child support enforcement and also had high JOBS assignment to postsecondary education. California and New York still show high benefits and weak enforcement, so they recorded caseload declines well below average. Michigan and Oregon are moralistic states in Elazar’s terms that, as we will see, bear some resemblance to Wisconsin. They combine high benefits with some enforcement effort, and they avoid assigning many JOBS clients to higher education. But they use moderate sanctions. On balance, they had caseload declines slightly greater than average. Wisconsin, however, combines high benefits with Mississippi’s strong sanctions. It again has much the highest work and child support levels in the table, and thus records the greatest caseload decline in the nation, except for tiny Idaho and Wyoming. The Wisconsin decline is the more remarkable because it comes on top of the state’s unusual performance in the earlier period. In 1994–98, most states’ declines largely reversed the sharp increase they had had during 1989–94. That meant shifting off welfare many cases that had only come on the rolls recently, many of

Decline of Welfare

•

191

TABLE 9.8 Values of Wisconsin and Other States on Model for 1994–98

Factors promoting lower caseloads Change in percent of welfare adults active in JOBS, fiscal 1994–96 State has delayed full-family sanction State has immediate full-family sanction Percent of AFDC cases getting child support, fiscal 1994 Average AFDC case size, fiscal 1994 (persons) Factors promoting higher caseloads Percent of JOBS clients assigned postsecondary education, fiscal 1994 Maximum AFDC grant in January 1994 State’s rank in legislative effectiveness (CCSL) State is individualistic (Elazar) Caseload change (percent) Predicted (average −43.8) Actual (average −43.8)

MS

CA

NY

MI

OR

WI

+

–

–

–

+

+

– +

– –

– –

+ –

+ –

– +

–

–

–

+

+

+

a

+

a

+

–

+

+

–

–

–

–

–

–

+

+

+

+

+

–

+

+

+

–

+

–

–

+

–

–

–

−59.8 −26.7 −17.0 −47.4 −49.3 −74.2 −62.2 −21.5 −27.1 −46.0 −58.3 −79.8

Source: See table 9.6. Note: Excludes Alaska and the District of Columbia due to missing data. State’s rank in legislative effectiveness is inverted so that high ranks are better. The letter “a” means that a state’s value on a variable equaled the average for all states, “+” that it was above the average, and “–” that it was below the average.

them not highly dependent. Wisconsin, however, had reduced its caseload by 8 percent, and this was on top of an earlier decline that began in 1986. Thus, by 1994, the cream was already off AFDC in Wisconsin—and yet the state drove its rolls down by a further 80 percent in the ensuing four years. This reflected the forceful diversion and work policies that led up to W-2, and then W-2 itself. Wisconsin truly obliterated traditional welfare in a manner not seen anywhere else in America.

INSTITUTIONS MATTER In much of the research on welfare, caseload change results from impersonal social and economic forces that no one can control, such as unwed pregnancy or joblessness, or from benefits like EITC that descend from

192

•

Chapter 9

Washington and are invariant across the states. Governmental policies that might affect welfare appear secondary, in part because they are difficult to measure. But if we directly investigate localities in a hands-on manner, as I have tried to do in this study, then caseload change appears less inevitable. All depends on how government responds to the pressures toward dependency. Tables 9.7 and 9.8 already anticipate a theme of chapter 11: States respond differently depending on their political culture. The South limits dependency largely by curbing access to aid. Most urban states respond hesitantly, reflecting political divisions and administrative limitations; they thus suffer more increase in the rolls, or less decline, than the average. Good government states like Wisconsin, however, attempt to govern dependency rather than either forbidding or accepting it. Welfare is a river that sweeps toward them, but they seek to channel it. They address the social forces behind it with a combination of help and hassle. They promote work and education among the poor, and they enforce work and child support. To do that is a test for government. This was the true test that Wisconsin passed. The near extinction of cash welfare was a triumph, but an institutional triumph stands behind it. Wisconsin enforced work and drove the rolls down through both politics and administration. As shown earlier, politics set an agenda for change, and then policymaking reshaped politics, with the early experiments promoting further change, culminating in W-2. Meanwhile, administrators implemented repeated reforms. The combination of politics and administration changed the reality facing recipients. Most of them reacted by leaving the rolls for jobs. Only a minority was directly forced to do this. Many more responded to the changes in expectations and to the opportunities offered by a rich economy. In the end, both policy and public opinion affirmed and enforced the new welfare order.

CHAPTER 10

The Effects of Reform

THE LAST CHAPTER assumed that the rapid fall in Wisconsin’s welfare caseload was a desirable thing. But that fall caused or accompanied many other effects on the lives of the poor. Only if these were good can we finally say that welfare reform succeeded. There are few evaluations to go on, but other indicators show that life improved for most people directly affected by reform. Poverty is not vanquished in Wisconsin, but trends are more positive than they were prior to reform. These results depended in part on good economic conditions, yet the cooler economy of the last couple of years has not changed things much. Evaluations tend to focus on economic effects, because these are the most measurable—gains (or losses) in employment or income, or cuts in dependency. Non-economic effects, however, matter just as much. The chief danger posed by entrenched poverty, especially in Milwaukee, was not so much destitution as the isolation of the poor from mainstream society. The Wisconsin reformers aimed to bring the recipients and their families closer to the mainstream through a change in lifestyle—higher work levels. Also, evaluations focus on a program’s effects on its clients, but we should care about effects more broadly. The Wisconsin reform meant to change the opportunities and expectations surrounding the low-income population as a whole, not only the population on welfare. Through its new requirements and benefits, W-2 aimed to channel more effort by lowincome parents toward working steadily to support their families, and less toward subsisting on welfare and other social programs.

THE FALL IN CASELOADS Chapters 7 and 9 dramatized the decline in the cash welfare caseload, meaning AFDC and W-2. Other government programs for the workingaged also fell, although less radically. One of these was General Relief, the expensive local aid program for groups outside federal welfare. Following the reform of 1996, the cash caseload fell to around 1,500 cases and almost all the money now went for medical expenses.1 Table 10.1 shows how the caseloads for Food Stamps and Medicaid changed alongside AFDC/W-2 from 1985 through 2001. These in-kind

194

•

Chapter 10 TABLE 10.1 Decline of Wisconsin Caseloads, 1985–2000 (numbers of cases) Year

AFDC/W-2

Food Stamps

Medicaid

1985 1990 1995 1996 1997 1998 1999 2000 2001

96,625 79,274 70,604 54,954 34,892 11,336 7,844 6,649 7,641

129,959 97,675 115,763 100,256 83,523 73,529 72,134 78,688 91,540

148,812 134,366 164,105 158,258 148,821 145,951 149,153 167,979 185,440

Source: Department of Workforce Development. Note: AFDC/W-2 includes only cash cases and excludes Kinship Care and Caretaker Supplement. Medicaid figures cover only family cases related to cash welfare, not all family cases and not elderly or disabled. Medicaid figures for 1999–2001 include BadgerCare and Family Care, a longterm care program being piloted in nine counties.

programs had been offered to welfare recipients virtually as a package, so when families left cash aid they often left the other programs too. State policymakers never attempted o drive down in-kind aid the way they did cash assistance; through BadgerCare, indeed, they largely reversed the fall in health coverage. Food Stamps, however, fell largely in step with cash aid, although both have turned up in the recent recession. A disconcerting effect of welfare reform was to take families away from the special antipoverty programs that had surrounded them. Many recipients, when they left cash aid, left subsidized child care as well. The state’s new child care system claimed to serve the entire working poor, yet 93 percent of the claimants in Milwaukee in 1998 were still on welfare; only 7 percent were off.2 Another program affected was Head Start, the federal early education program for poor children. It had routinely recruited its clientele from AFDC, but as the number of welfare families fell, it had difficulty filling its rolls.3 The gain, however, was that as families left welfare for work, they moved closer to mainstream society. Now if they drew income from government, it would more likely be from middle-class social insurance programs such as Medicare or Unemployment Insurance.

FISCAL EFFECTS The changes in caseloads also transformed the finances of welfare. Welfare’s administrative costs increased sharply in Wisconsin, due to its

Effects of Reform

•

195

elaborate reform programs. The state generally ran welfare efficiently, but, as detailed in chapter 4, the onset of welfare reform brought an apparent deterioration. Between 1986 and 1994, administrative spending per case soared by almost six times. The state changed from a leader in administrative costs to a laggard.4 According to the Legislative Fiscal Bureau, the cost of running welfare programs grew from $112 million in 1986–87 to $193 million in 1993– 94 even though grant costs fell 18 percent in that period as caseloads fell. Critics of reform such as Senator Gwendolynne Moore seized on such figures to attack Tommy Thompson. “We’re paying more to get less under this administration,” she declared. The governor might claim to be saving money, “But in fact, the money going to poor families is dropping, while more and more is wasted by the bureaucrats.”5 Overall, however, the state clearly saved money on reform. Serious work enforcement had far greater diversion effects than anyone expected. The dramatic caseload more than offset the extra costs of Wisconsin’s elaborate welfare reform system. In 1997, Thompson noted with amazement that the caseload had fallen by 60 percent during his decade in office, thus cutting welfare spending by $317 million a year.6 W-2 had been expected to cost more than AFDC, but again the state was saved by the ongoing caseload decline. Table 10.2 shows Wisconsin’s welfare spending in three budget years: 1986–87, at the start of the Thompson Administration, 1993–94, which marked the end of the moderate reform period; and 1999–2000, after the implementation of W-2. Figures are given for cash benefits including AFDC/W-2, other programs (which include work programs, health, and child care), and administration. For many programs, the federal government reimbursed half or more of the state’s costs, so separate totals are given for state and federal spending. “Other” spending includes revenue from nonprogram sources such as child support collections. I also calculate the percentage change in spending between 1986–87 and 1999– 2000.7 The top panel of the table gives spending for all means-tested programs. It shows that although cash benefits fell sharply with the caseload fall, this saving was outweighed by large increases in the other categories, for a 78 percent rise in welfare costs overall. Allowing for inflation, the rise is still 18 percent. However, these figures cover several programs that are tangential to welfare reform, particularly Supplemental Security Income, an aid program for the elderly and disabled that grew by half in the period, and above all Medicaid, the huge and costly health program for the poor, whose costs more than doubled. While Medicaid and its adjunct BadgerCare were an integral part of the reform system, most Medicaid

196

•

Chapter 10

TABLE 10.2 Trends in Wisconsin welfare spending, 1986–2000 Percent Change 1986–2000

1986–87

1993–94

1999–2000

725,139,000 1,164,702,800 112,232,800 861,700,300 1,082,518,800 57,855,500 2,002,074,600 3,011,163,138

592,582,900 2,442,319,400 192,757,300 1,276,611,500 1,875,760,900 75,287,200 3,227,659,600 3,772,327,158

283,431,200 3,080,595,299 194,221,980 1,245,221,198 2,263,172,281 49,855,000 3,558,248,479 3,558,248,479

−61 164 73 45 109 −14 78 18

AFDC/W-2 and work Cash benefits 636,324,900 Other programs 11,499,600 Administration 92,608,800 State 291,437,700 Federal 391,190,500 Other 57,805,100 Total 740,433,300 In 2000 dollars 1,113,627,564

461,121,500 47,217,300 151,798,500 238,978,500 353,206,200 67,952,600 660,137,300 771,535,469

151,165,800 302,372,740 149,758,480 115,756,152 449,923,468 37,617,400 603,297,020 603,297,020

−76 2,529 62 −60 15 −35 −19 −46

All programs Cash benefits Other programs Administration State Federal Other Total In 2000 dollars

Source: Legislative Fiscal Bureau and Department of Workforce Development. Note: Upper panel includes all means-tested programs. Lower panel excludes Medicaid and programs unrelated to cash welfare reform. “Other programs” includes health, work programs, child care, and other social services.

spending goes to the elderly and disabled, not welfare families, so it reasonable to exclude it here.8 Omitting these programs leads to quite a different accounting. The bottom panel of the table is limited to AFDC, other cash benefits for the working-aged, child support collections, programs and services related to welfare employment, and associated administrative costs. Without SSI, cash benefit spending now declines even more rapidly. The rise in other programs is startlingly steep, due to huge increases for welfare work programs and child care, but the base is sharply reduced by the exclusion of Medicaid. Administrative costs grew strongly through 1993–94 but then leveled off. In nominal dollars in the lower panel, the total fell rapidly between 1986–87 and 1993–94, then more slowly through 1999–2000. The state spent much less on cash support for nonworking poor, and much more on noncash support for working families, both on and off welfare, and still came out ahead. Over the whole period, from 1986 to 2000, Wisconsin actually cut nominal spending by 19 percent, or about $137 million.

Effects of Reform

•

197

Savings flowed mostly to the state government while federal spending rose. Allowing for inflation, the overall economy approached half. The state currently has a budget crisis, but without the welfare savings the crunch would apparently have been worse than it is.9 The Wisconsin reform was not centrally about money. The state was willing to spend new resources to achieve its goals, and expected to do so. But by driving most families off the rolls, the state could be more generous to those remaining than would have been possible otherwise.

EVALUATING REFORM Earlier chapters have shown that, except in the drafting of W-2, the Wisconsin reform did not rest on any deep policy analysis. Policymakers believed they could figure out what worked using pilot programs and their own experience, with little formal evaluation. Most of the assessments the state did do were required by Washington, as a condition of approving waiver programs. Some of these studies were completed but showed no effects, or doubtful ones. An early evaluation of WEJT in Rock County found that it produced lower earnings and higher dependency compared to earlier policy, probably because of its focus on education rather than work. Neither Learnfare nor Children First showed clearcut impacts, although Children First may have had some. The 100-Hour Rule Demonstration and the Two-Tier AFDC Benefit Pilot Program also lacked clear effects.10 In retrospect, weak findings were hardly surprising. Experience has made clear that just putting recipients in training or changing benefit incentives has little effect on their behavior. Administrative demands to work in return for aid have larger effects. Once the diversion programs and more radical work tests came on stream starting in 1994, effects on caseloads and work became much clearer. Administrators could perceive them without benefit or research, so evaluations seemed even less justified than before. Once allowed to by PRWORA, the state terminated most of the remaining waiver studies; they had been overtaken by events.11 The only part of W-2 to receive an experimental evaluation was the 100-percent pass-through of child support collections to recipient families. The idea behind the policy was that if government did not take child support to reimburse itself for supporting welfare families, absent parents would more likely pay support, and would more often work in legal jobs, instead of working off the books to escape paying support. The evaluation found that mothers subject to the new policy gained perhaps $123–$142 a year in child support compared to the control group. More fathers paid

198

•

Chapter 10

support, and more also worked in legal jobs.12 These effects, though small, did confirm the theory behind the policy. The early programs seem to have achieved little in themselves, but they were part of a larger reform process that clearly did have effects. That process—the stiff new work tests coupled with support benefits—was too pervasive to evaluate rigorously. With everything changing, there was no way to shield a control group from the new policies, to provide baseline against which to measure change. The only plausible assessment was to see how social conditions changed over the period of reform, although these changes cannot be attributed rigorously to the new policies.13

EMPLOYMENT The principal goal of reform was to move welfare adults into jobs. With little question, Wisconsin achieved this, both on and off the rolls. Within welfare, the state attained work activity from an unusually high proportion of its recipients. Table 4.1 showed that participation rates in JOBS were well above the norm for the nation, even if at that time activity still reached only a minority of cases. TANF’s more severe rules redefined “work activities” to cover mainly actual work, with less room for education or training. Nationally, only 28–38 percent of cases were active by this standard in fiscal years 1997–2000. Wisconsin, however, achieved rates of 53–80 percent in these years, results that stood second or third among the states.14 Such numbers reflect the simple fact that in Wisconsin by this time, work had become a virtual precondition for aid. One purpose of stiff work requirements was to drive employable people off the rolls into jobs, and this the state also achieved. During the transition to W-2, state officials and other observers marveled at how quickly recipients abandoned welfare for jobs, even people who might have seemed unemployable. Families shifted wholesale from aid to work as their main support. The change was quickest in rural areas, slower in Milwaukee. Even after the transition to W-2, rapid departures for work continued. A majority of women who were on W-2’s two lower (cashpaying) tiers in 1998 were working in 1999.15 In several surveys by the Department of Workforce Development of recipients who had left welfare in 1998, near the beginning of W-2, 49– 54 percent of respondents said that employment was their main reason for departing the rolls. Other reasons, such as ineligibility or a disinclination to be on aid, came far behind. DWD also analyzed administrative data on case closures in the first two years of W-2; 64 percent of the cases had closed due to employment.16 Specifically in Milwaukee, work was the largest reason for recipients leaving welfare during the transition to

Effects of Reform

•

199

W-2, and the share of cases closing over two years due to employment was 62 percent.17 Once off the rolls, former recipients continued to work at high levels. In DWD’s leavers surveys, 58–62 percent of leavers said they were working at the time of the interview, and 82–83 percent had worked at some time since leaving the rolls. In Milwaukee, 72 percent of those who left the rolls during the W-2 transition were working in regular jobs in the month before they were surveyed. Administrative data show that 81–84 percent of leavers worked at some time in the year after leaving welfare. Working hours were also higher than one might expect—around half of working leavers, including those in Milwaukee, claimed to be employed at least forty hours a week.18 Dependency on cash aid fell as low as 7,000 cases, or around 20,000 if one adds Kinship Care and Caretaker Supplement. Much larger numbers continued to receive the in-kind programs, as table 10.1 shows. In DWD surveys, around half of leavers were still on Food Stamps and 71–77 percent were on Medicaid or BadgerCare when they were interviewed 6–22 months after leaving cash aid, and the figures ran somewhat higher in Milwaukee. But gradually, majorities left all the aid programs. After three years off the rolls, only 36 percent of leavers were still on Food Stamps and only 49 percent were still on Medicaid, though many more than this remained eligible.19 In surveys prior to welfare reform, fewer welfare mothers admitted to working than were in fact in jobs, often “off the books.” Perhaps when they left welfare they did not “go to work” but simply admitted to working already and dropped aid. However, among mothers who were on AFDC in 1990, 57 percent recorded earnings in the Unemployment Insurance system at some time in that year, rising to 65 percent in 1998. The share of those workers who were employed in all four quarters of the year rose much more sharply—from 37 percent in 1990 to 71 percent in 1998.20 Such numbers make it highly likely that actual as well as apparent work levels rose. EARNINGS AND INCOME DWD’s surveys found leavers with jobs earning wages of $7.42–$7.95 an hour on average, or well above the minimum wage. And over time, earnings rose. Among women who entered W-2 in its first year, average wages were $7.28 in 1998, rising to $8.07 in 1999. Average yearly earnings in the two years rose from $4,350 to $6,000 (38 percent), and from $5,600 to $7,650 (37 percent), considering only those who worked. The gains reflect both higher wages and more working hours. This is stronger in-

200

•

Chapter 10

come growth than previous research would have predicted. Over the longer term, earnings growth is even more dramatic. Working welfare recipients earned on average only $4,865 in 1990, but the figure rose to $12,787 by 1998—in constant 1998 dollars—or 163 percent.21 Of course, as earnings rose, many families lost eligibility for meanstested benefits, or simply left them. So overall family income did not rise as fast as earnings—and might even fall. According to administrative data, the personal income of welfare mothers who entered W-2 in its first year, including leavers and non-leavers, fell slightly, from $10,150 to $10,090 between 1998 and 1999. This calculation, however, considers only recorded earnings, W-2, the cash value of food stamps, and child support. In a survey, where mothers might consider other sources, their family incomes rose on average from $12,082 to $14,779 in the same years.22 Were they better off? Among those who left welfare in early 1998, 48 percent said that they had more money than when they were on aid, but 50 percent disagreed, and 68 percent said they were “just barely making it from day by day.” But at the same time, 68 percent disagreed that life had been better on welfare, and 59 percent thought they would not have to go on welfare again. In later DWD surveys, 46 percent of leavers said they were more satisfied with their lives than they had been on welfare; 38 percent were unchanged, only 16 percent less satisfied. Sixty percent felt better about themselves, 34 percent the same, only 7 percent worse. Specifically in Milwaukee, 39 percent of cases that went through the W-2 transition felt they were better off afterward, 32 percent the same, 29 percent worse; 36 percent said their standard of living was good or very good, 48 percent fair, only 16 percent poor.23 The picture is one of modest improvement for most families, although some unquestionably lost income. The gains in lifestyle are larger than the gains in income. Families might have reaped little in money terms from reform, or even lost ground, yet most felt they were better off overall. That suggests that simply to be earning their way, and in most cases off cash welfare, was worth a good deal to people. Whether or not they gained income, they had gained ground in the struggle to enter mainstream society.

POVERTY A further assessment is whether reform changed poverty rates. The federal poverty measure, used here, sets an income line for families that is supposed to cover their bare necessities. The thresholds are indexed for

Effects of Reform

•

201

inflation and are higher the larger the family. Welfare recipients usually begin in poverty, because welfare benefits in all states pay well below the poverty line. The leading criticism of W-2 was that, for most families, it did not overcome poverty. The Legislative Audit Bureau, an arm of the state legislature, tracked cases that left W-2 in early 1998. It found that, of those that filed tax returns for 1999, only 34 percent claimed incomes above the poverty line, 47 percent if one assumes that families also received federal and state EITC. Academic studies based on administrative data also found high poverty rates. Of mothers tracked in the childsupport evaluation, on and off the rolls, 77 percent were poor in 1998, falling to 67 percent in 1999. A study of two cohorts of leavers found 59 and 69 percent still poor in the year after leaving. Typically, the mother’s own earnings accounted for the bulk of family income, but added to this were some other resources from cash aid, food stamps, the EITC, and perhaps some earnings from spouses or other family members.24 However, it made all the difference whether a leaver was working or not. Wisconsin women who had been in welfare in 1990 and worked all year in 1998 earned an average of $16,389 in that year, or well above the poverty line for a mother and two children. Adding in tax credits, the total rises to almost $21,000. In Milwaukee, former AFDC clients who were working in regular jobs were typically off cash aid and above the poverty line. Those still in W-2, SSI, or Kinship Care, however, were still somewhat below poverty, while those neither on aid nor working were still worse off.25 Living just on W-2 cash benefits, a mother is well under the poverty line for a three-person family, but if she works full-time—even at the minimum wage—and collects EITC and food stamps, her income reaches 117 percent of poverty. A follow-up study in Dane County corroborates that estimate: A year after transitioning to W-2, respondents’ average income (counting cash sources plus food stamps) was at or slightly above the three-person poverty line, and higher for leavers who were working.26 A DWD study claimed that recipients could realize over $30,000 a year if they worked full-year, full-time at the wages they claimed and also accessed all remaining benefits.27 To blame W-2 for high poverty rates among nonworking leavers means taking their nonemployment as given. If the program overcomes poverty only with work, and not without, that is consistent with its original design. It never promised to support families above poverty by itself, but only to help working parents do so. That promise it does appear to fulfill.

202

•

Chapter 10

FAMILIES AND CHILDREN Since the original purpose of family welfare was to protect children, welfare reform must at least “do no harm” to them. The greatest direct impact reform is likely to have on children, besides changing family income, is altering access to health care. Based on parental report, 11 percent of the welfare children in the Wisconsin child support evaluation in 1998–99 were in only fair or poor health, twice the level for poor children in national data. However, only 17 percent of the children were without health coverage for part of 1999, up only 1 point over 1998, compared to figures for the nation of 25 and 23 percent. Wisconsin appears to have limited the loss of coverage despite its abrupt caseload fall. One indicator of this is that 98 percent of the children in the child support study were enrolled in Medicaid during 1998, compared to 55 percent for poor children nationally. And the state heavily promoted BadgerCare, which began in 1999.28 If the health of Wisconsin’s poor children is nevertheless subpar, that is probably for nonpolicy reasons, particularly the depressed condition of inner-city Milwaukee. DWD’s leaver surveys found some evidence that the condition of children had improved in other respects. Large minorities of respondents said their children’s health and their grades and behavior in school had recently improved, although as many or more reported no change. This feedback is consistent with results from evaluations of welfare work programs outside Wisconsin. These suggest that young children with parents in work programs tend to make cognitive gains and do better in school, compared to children from the control group. There are also some negative effects on behavior, mainly for older children. All these effects are small.29 Parents and children seem to gain most from welfare reform in noneconomic ways when the program combines work requirements with generous support benefits, as occurred, for instance, in the Minnesota Family Investment Program.30 In cash terms, W-2 is less generous than MFIP, but it does offer solid health and child care support, plus the child support pass-through. Thus, the program probably does generate some favorable effects on families, at least for those who maximize income by working.

HARDSHIP Even more alarming than damage to families would be signs of acute suffering due to welfare reform. Reform has put new strains on poor families in Wisconsin. In most counties, pressures on recipients to work devel-

Effects of Reform

•

203

oped gradually during the JOBS era. In Milwaukee, the sudden onset of serious work demands in 1996–97 plunged the city into turmoil. Some mothers, unable to cope, turned their children over to grandmothers on Kinship Care, one reason that program grew rapidly. Some families simply disappeared. When surveyed, 43–47 percent of W-2 leavers said they felt more stress and worry since quitting the rolls, only 17–24 percent less, with 30–40 percent unchanged.31 Despite this, however, signs of conspicuous hardship have been difficult to find. When surveyed, 16–20 percent of welfare mothers in the child support evaluation said that they had sometimes lacked enough to eat, had had utilities turned off for failure to pay the bill, or had had to move in with others due to inability to pay the rent during 1999. Half claimed to have lost phone service. Over half of the mothers had turned to friends, relatives, or charity for resources. All these figures declined slightly from 1998. The levels are somewhat higher than for the low-income population in general, but they resemble welfare conditions in other states.32 It is thus difficult to blame the problems specifically on Wisconsin’s welfare reform, despite its unusual severity. In its leaver surveys, DWD found that, in the two years prior to being interviewed, 71 percent of former recipients had gotten behind on paying utility bills, 55 percent had fallen behind on rent or house payments, and 26 percent had moved in with someone to share expenses, among other problems. However, such hardships occurred when people were both on and off the rolls; at most, problems ran somewhat higher after departure. And, as with poverty rates, outcomes turned on employment, with working leavers less likely to struggle than the nonworking.33 Child welfare outcomes are a key indicator of hardship. The chance that adults subject to welfare reform might harm their children appears high, because 39 percent of W-2 applicants in Milwaukee in 1999 had been investigated for child abuse or neglect in the preceding decade. Outof-home placements of children by child welfare authorities also soared 86 percent in Milwaukee during the 1990s. Some observers saw this as a crisis that might be blamed on reform.34 However, allegations of abuse, and the number of these charges that were substantiated, actually fell by more than half in Wisconsin between 1994 and 1998, the period when serious welfare reform hit the state. This could be due to reform—higher parental employment tends to reduce child welfare involvement, although higher parental stress increases it. Most likely, the growth in foster care placements is due, not to welfare reform, but rather to the inefficiencies of child welfare administration in Milwaukee. That problem led to lawsuits and eventually a state takeover of child welfare functions in the county.35

204

•

Chapter 10

A still more prominent assertion was that welfare reform produced an upsurge in homelessness in Milwaukee. Starting in 1996, shelters and food pantries complained of rising demand in the city. Advocates and community groups commonly blamed this on the recent rise in work demands and sanctioning in welfare.36 Nine percent of leavers said they had stayed at a homeless shelter within two years, but again the level was nearly the same on welfare as off.37 It appears that homelessness did increase, but from a low base. Joe Volk, chairman of the Milwaukee Shelter Task Force, estimated that the number of homeless families had risen by a quarter during the winter of 1996–97, compared to the previous year. That meant that forty-one more families were homeless on a given night. This was minuscule compared to the thousands of cases that had recently left the rolls.38 State officials say most shelter residents are not welfare leavers. One careful study that interviewed women in a Milwaukee shelter found that only 8 percent of them were homeless for reasons immediately tied to welfare sanctions, 15 percent if one counted earlier sanctions. In part, greater use of shelters and pantries might reflect the new diversion policies, the fact that welfare was referring applicants to local charities before immediately putting them on welfare, rather than the problems of leavers. It might also reflect a low level of subsidized housing in the state and changes in subsidized housing rules that made it easier to evict tenants. And it might reflect the fact that more shelters were built in Milwaukee, causing more people to come forward.39 If there was not more hardship, the reason might be that leaving cash welfare often makes less difference to the finances of families than people think. We ordinarily assume that families on welfare have no other resources, but this is seldom the case. A national study of benefit termination through 1996 in several states, including Wisconsin, found that 75 percent of sanctioned families had other income or benefits from other government programs. In Wisconsin, where most of the sanctions were due to SSF or PFP, 47 percent of those leaving the rolls also had private income from wages, pensions, or child support, and 32 percent had earnings.40 The main concern focuses on people who had left welfare without going to work. However, DWD’s surveys revealed that 15–18 percent of these people were married to a working spouse, and 60–65 percent of the rest had support from some other cash benefit program, such as Social Security, SSI, or Unemployment Insurance. Although welfare mothers claim to be without support, many actually live with spouses. When work requirements bite, some can simply rely on that support more openly and still not work.41 If all else failed, many leavers simply went back on cash aid. Among cases terminated in Wisconsin through 1996, 38 percent did

Effects of Reform

•

205

this, and later under W-2, 26 percent of leavers did the same. One might view this as an indictment of the program, but it was also reassuring.42 A fair criticism of W-2, made in earlier chapters, is that its initial design was too severe, and its implementation exposed the caseload to unnecessary risk. Some details were not completely thought through. Possibly, for example, families that shifted to Caretaker Supplement suffered because their stipend was less than they had received on AFDC. The system had to pay more attention to hard cases that might fall through the cracks, and it has done so.43 But if the state rolled the dice with W-2, it also won its gamble. The overall effects of reform on the society were clearly favorable. Hardship did occur, but it was seldom severe or lasting. It was the dog that didn’t bark. COMPARING WISCONSIN TO OTHER STATES A further way to assess the results of welfare reform is to compare Wisconsin to other states. Most also have seen large caseload declines. The national caseload decline appears to be driven by good economic conditions and new federal benefits, not only by changes in welfare. In other states, as in Wisconsin, many people seem to be leaving welfare simply due to the controversy about reform, even before tougher work tests are enacted or implemented. If so, then the Badger State’s unusual statecraft may be less critical to successful outcomes than I have implied. A recent summary of leavers studies nationwide found that 55–64 percent of leavers were working when surveyed, and 63–91 percent of them had worked at some time since leaving welfare. In most states, over 40 percent attributed their leaving to employment. Their average wages ran from $5.50 to $8.80 an hour. A majority remained poor, up to half suffered difficulties such as paying bills, and 18–35 percent returned to welfare at some point. All these results may seem comparable to Wisconsin’s.44 But on inspection, the state was distinctive in several ways. A cost of the Wisconsin reform was to produce unusual income losses as recipients left the rolls, because this was the only state to combine high benefits with a severe work test. The state drove the vast majority of the employable off the rolls, and these recipients had a lot of cash aid to lose. The lavish in-kind benefits the state continued to provide did not at first compensate.45 Meanwhile, the income to be had from working was not much different than in less generous states. The benefit, however, was probably to reduce poverty to an unusual extent among the low-income population as a whole. The impact on leavers was sharp, but the severe work tests and the support benefits also drove many people into jobs who otherwise would have gone on aid or

206

•

Chapter 10

TABLE 10.3 Trends in Poverty Rates for the United States and Wisconsin, 1989–98

All persons (percent) United States Wisconsin Wisconsin’s rank among states Persons under 18 (percent) United States Wisconsin Wisconsin’s rank among states Persons under 5 (percent) United States Wisconsin Wisconsin’s rank among states

1989

1993

1995

1996

1997

1999

13.1 10.7 14.5

15.1 10.9 8

13.8 8.9 4

13.7 8.7 2

13.3 9.2 4

11.9 8.4 6

18.3 14.9 22

22.7 15.9 11.5

20.8 13.9 11

20.5 12.2 4

19.9 14.3 7

17.1 10.9 6.5

20.1 17.7 22

26.5 19.4 9

24.2 16.7 13.5

23.4 14.9 6

21.7 15.6 7

18.9 12.8 9

Sources: Small-area estimates downloaded from Web page of U.S. Department of Commerce, Bureau of the Census, on December 1, 2002. Figures for 1989 derive from the 1990 Census, other figures from model estimations.

not worked. This in the long run raised incomes and reduced poverty rates. Poverty among children (under age eighteen) rose in Wisconsin between 1979 and 1993, then fell through 1998. That roughly tracks the change in welfare from entitlement to demanding work tests. Of course, it also tracks economic conditions; poverty trends in the nation were similar, although poverty levels in most states ran higher.46 Better evidence for the efficacy of reform is that Wisconsin gained ground on the rest of the country. Table 10.3, based on census estimates, shows that, nationally, overall poverty rose slightly between 1989 and 1993, then fell slightly. The trends for children under eighteen and under five are similar, except that levels run higher and the rise and fall are steeper.47 In all three cases, Wisconsin’s levels run lower and it gains in relative ranking among the states. After 1989, Wisconsin cut overall poverty so sharply that by 1996, only New Hampshire was less poor. In child poverty, the improvement is even steeper. After 1996, Wisconsin’s ranking worsens somewhat. This may reflect the initial severity of W-2 and its confused implementation. As for conditions affecting children other than poverty—rates of low birth weights, infant mortality, teen deaths, teen births, dropouts from school and work, female-headed families—the trends are less clear. Between 1990 and 1999, some of these indicators improve in Wisconsin, and some worsen; so does the state’s ranking on the indicators relative to other states. Wisconsin, however, remained well above average on almost all of them.48 The lack of clear effects might be because most of these

Effects of Reform

•

207

conditions have only uncertain ties to the changes in dependency and work levels that reform directly produced. Or it may be because, in the first instance, the Wisconsin reform was aimed much more at adults than children.

WHERE WISCONSIN EXCELS While other states had large caseload declines, table 1.1 shows that Wisconsin’s was almost the greatest in the nation. Especially, it outpaced decline in most other big-city states by a wide margin. It virtually ended the chance of living on cash welfare without work for the able-bodied, a watershed that probably no other urban state has crossed. This transformed welfare finances, and it apparently produced exceptional work levels among low-income families. The fiscal transformation of welfare in Wisconsin was extreme. Under TANF, the federal government gives fixed block grants to the states. When caseloads fell, states reaped a windfall. Nationally, the federal AFDC/ TANF dollars spent per family rose from $6,242 in 1995 to $8,330 in 1999, or 33 percent. In Wisconsin, however, the rise was far greater— from $7,238 to $15,975, or 121 percent, a figure exceeded only by Idaho, Kansas, Oregon, and Wyoming.49 As shown in chapter 7, funds were reallocated wholesale from cash welfare to support benefits. Compared to several representative states— California, Georgia, and Missouri— Wisconsin’s shifts were extreme. Between 1995 and 1999, the state cut overall social services spending by 8 percent and spending on cash assistance by 77 percent. Meanwhile, it raised spending on child care and child development by 168 percent. Wisconsin’s unspent federal TANF funds in 1999 were an enormous 155 percent of its single-year TANF grant. All these figures were easily the largest among the four states. For employment services, Wisconsin’s increase— 20 percent—was much the smallest. This was probably because the state had made much of its investment in work programs prior to 1995 and also because its extreme caseload decline left it with few clients to serve.50 Wisconsin’s other conspicuous achievement is unusually high work levels. Compared to elsewhere, the state has typical rates of earnings, poverty, and return to welfare among leavers—but higher work levels. In Wisconsin, over 80 percent of leavers were employed after leaving welfare, whereas in other states the rate is closer to two-thirds.51 This advantage seems due to the state’s intense effort to promote work, and it is the leading achievement of the Wisconsin reform. The best fix on comparative work levels comes from the Urban Institute’s National Survey of America’s Families (NSAF), a study of social

208

•

Chapter 10

TABLE 10.4 Work and Poverty Levels in Wisconsin and the Nation, 1997 and 1999 Wisconsin

United States

Wisconsin’s Rank (of 13)

1997

1999

1997

1999

1997

1999

72 92

73 91

62 89

62 88

1 1

1 2

77 96 8 12

80 94 7 10

63 94 13 21

67 94 11 18

1 3 1 1

1 9 2 1

30

25

45

39

1

1

38

36

46

43

1

2

19

21

23

23

2

3.5

Adults aged 25–54 working fullor part-time Below 200% of poverty Above 200% of poverty Single parents aged 25–54 working full- or part-time Below 200% of poverty Above 200% of poverty Adult poverty rate Child poverty rate (all families) Child poverty rate (single parent families) Adults below 200% of poverty with difficulty affording food in previous year Adults below 200% of poverty with problems paying mortgage, rent, or utilities in previous year

Source: Urban Institute, Snapshots of America’s Families II: A View of the Nation and 13 States from the National Survey of America’s Families (Washington, DC: Urban Institute, October 24, 2000). Note: Figures other than ranks are percents. Poverty figures apply to the previous year (1996 and 1998).

conditions surrounding low-income families in 1997 and 1999. NSAF generates findings on the nation and also on thirteen individual states, including Wisconsin.52 Table 10.4 summarizes the results for work and poverty levels. In Wisconsin, work rates for working-aged adults who are low-income (below 200 percent of poverty) are 10–11 points higher than the national average. For low-income single parents the gap is even larger—13–14 points. In both 1997 and 1999, these figures were the highest of the thirteen states. One might suspect that high work levels are simply characteristic of Wisconsin society, but figures for working-aged adults and single parents over 200 percent of poverty show a much smaller difference from national norms. Something about the state drives specifically low-income adults and single parents to work unusually hard. That something can only be welfare reform, both its severe work demands and its generous work supports. The lower half of the table shows that these work levels, along

Effects of Reform

•

209

TABLE 10.5 Comparing Wisconsin and Milwaukee to Michigan and Minnesota, 1997 Wisconsin Michigan

Minnesota

Adults not in school and employed full- or part-time Low-income adults aged 25–54 Single parents 69 69 Married mothers 57 57 Low-income adults aged 18–29 Single parents 67 61 Married mothers 54 50 Child poverty rate One-parent families 35 35 All families 14 13 Children in lower-income families With private health coverage 48 47 With public health coverage 40 41 Uninsured 12 13

Statewide

Milwaukee

United States

78 57

72 49

63 45

69 60

58 42

55 37

29 11

46 24

44 21

58 27 15

40 46 14

40 39 21

Source: Michael Wiseman, “In the Midst of Reform: Wisconsin in 1997,” Focus 20, no. 3 (Fall 1999): 15–22.

with Wisconsin’s benefit system, cut adult and child poverty and material hardship well below the rates typical of the country. Wisconsin also has unusually high rates of health insurance coverage for both low-income adults and low-income children. Much of the difference is due to unusually high rates of employer-based health coverage, reflecting in part the state’s strong emphasis on moving welfare recipients into private employment. The payoff came in lower-than-average rates of poor health among low-income adults and children. Again, these differences seem to reflect social policy rather than the nature of the state; the state’s advantage in health coverage or poor health among people above 200 percent of poverty is far smaller than its margin at lower incomes.53 A still more rigorous test is to compare Wisconsin to Michigan and Minnesota, its immediate Midwestern neighbors. These are states with many similarities to Wisconsin in economic and social conditions. As chapter 11 will discuss, they are also, like Wisconsin, states with goodgovernment traditions that we would expect to perform well in social policy. But as table 10.5 shows, the Badger State generally excels them in promoting employment and reducing poverty, at least in 1997. The NSAF survey was enlarged in Wisconsin to permit breaking out separate results for Milwaukee as well as the state as a whole, and these results also are shown here.

210

•

Chapter 10

Statewide, Wisconsin has conspicuously higher employment levels than the other states. For low-income adults aged 25–54 in Wisconsin, the difference between the work rate for single mothers and married mothers is sharply greater than in the other states. For ages eighteen to twentynine, work rates for both single and married mothers are higher in Wisconsin. Poverty rates in Wisconsin are also lower, especially for one-parent families. In health coverage, the other states have a slight edge, but again Wisconsin excels in obtaining private health coverage for families. In all these respects, Milwaukee lags behind the rest of Wisconsin. Nevertheless, the city is close to national averages for poverty rates while excelling those norms for work levels and public health coverage. IN THE FACE OF DISADVANTAGE Still more remarkably, Wisconsin achieved these results in the face of a more and more disadvantaged caseload. At the height of its welfare rolls in 1986, the state’s welfare adults were 51 percent high school graduates, 56 percent white, 29 percent black, and 5 percent Hispanic. Nineteen percent were men because two-parent cases comprised 27 percent of the caseload, well above national norms.54 All these figures made the caseload more employable than in most states. But as the state drove the rolls down, the more employable cases left soonest, and the demographics of the remaining cases were transformed. By October 1999, only 36 percent of the adults on cash aid were high school graduates, 12 percent white, and 2 percent male, while 69 percent were black and 7 percent Hispanic.55 As noted in table 7.1, Milwaukee’s share of the caseload soared from under half to over 80 percent. Changes like this occurred in other states too as their caseloads fell, but the Wisconsin shifts were much more extreme than the average. Wisconsin had driven the cream of its caseload off the rolls long before the national vogue for serious welfare reform began in the mid–1990s. But despite this, in leaver studies from the late 1990s the state attains economic outcomes on a par with those in other states, along with unusual work levels. This it seems to have achieved because of the rare force of its work tests coupled with very generous support services. The reformed system, indeed, was designed precisely to drive the more employable cases off the rolls early, so that both administrative attention and benefits could be concentrated on the harder-to-serve. In its later, more paternalistic form, W-2 became a vast engine devoted to preparing even the least employable for jobs. In so doing, Wisconsin pressed the limits of the labor market as probably no other urban state has done. As noted in chapter 9, a hot economy

Effects of Reform

•

211

rescued the state from the fear that jobs would be lacking for the massive numbers that reform demanded should work. Even low-skilled recipients in Milwaukee obtained jobs at respectable wages, and often with benefits. This was partly because the tight labor market left employers no choice, but also because the new policies motivated even the least employable to seek work. The cost was that many of the new hires were less job-ready than elsewhere. Compared to recipients hired in Chicago, Cleveland, or Los Angeles, those hired in Milwaukee displayed the most attitude problems, the highest absenteeism, and the lowest job retention. This was a sign of success, not failure. It reflected the fact that so many welfare recipients were hired in Milwaukee compared to the other cities.56 Those workers were available, in part, because Wisconsin insisted that they work, and enabled them to do so. This capacity to move even the least employable toward employable is, perhaps, the innermost achievement of the Wisconsin reform. That result holds more potential to uplift the inner city in the long term than any other. We see again that the state took risks, but its gamble was vindicated. The focus on work served economic goals, but above all the hope of social integration. A CHANGE IN DIRECTION If one focuses on the lifestyle of the poor, welfare reform may seem to change little. It alters the source of income for poor families, but seldom the fact that incomes are low, at least in the short run. It has still less effect on the non-economic problems of the poor, the cause of much of their distress. The great fact about poor areas, some would say, is not that work levels are low but that most families are female headed and children often succumb to crime, early pregnancy, and school failure. Welfare reform has little immediate effect on these dilemmas. Indeed, it dramatizes them. As welfare mothers struggle to work, their chief impediments are less likely to be insufficient jobs or child care than family and personal problems.57 The most serious shortcoming of welfare reform, as of welfare itself, is failure to address the problems of low-income men. The fact that lowincome fathers typically abandon their families is the first reason why many women go on welfare. Single parenthood damages the prospects of children as well. Government has not learned how to prevent this the way it has learned to promote work among the mothers. Programs like Children First that enforce child support on absent fathers have so far succeeded much less than enforcing work within welfare. At best, reform within welfare puts indirect pressure on the fathers, since now the mothers

212

•

Chapter 10

need help to work if they stay on welfare, or a partner they can depend on if they leave the rolls. But despite this, welfare reform and rising work levels do change the direction of the inner city. Nothing so positive has happened there since the civil rights reforms of the 1960s. The share of poor adults engaged in advancing themselves clearly has increased, and this has raised morale as well as incomes. Interviews with welfare mothers outside Wisconsin in 1997–99 suggest that they view going to work more favorably then they did prior to TANF. They think they can get more income, plus greater self-respect and also respect from their children, who prefer them to work, although they remain concerned about child care costs and the logistics of working.58 That life of greater demands, but greater respect, seems to be what a lot of former welfare families are now living in Wisconsin. Since W-2, city bus drivers in Milwaukee say they notice more people from poor areas traveling to jobs early in the morning. Despite the effort, there is “a new infusion of energy and activity,” observers say, and people feel more upbeat. “When W-2 first started, there were a lot of complaints,” one driver said. “But now that many people have gotten into it, their morale and self-esteem is boosted, and you can tell they feel good.” The structure imposed on the mothers has spread to the children, who now have to be in child care or school at an early hour. They hang around home less, and neighborhoods are calmer. Far from resenting this, children seem happier, pleased their parents are now working, and more respectful of them. Said a child care director, “We hear them say, ‘My mom’s at work,’ or ‘Grandma’s picking me up ’cause Mom’s working late,’ and you can hear the pride in their voices.”59 Another result of work is that families with earnings become connected to the characteristic concerns of the middle class. They can now imagine, maybe for the first time, going on vacation, buying cars, even homes, and moving to the suburbs. They also have to pay taxes, if only payroll levies (incomes are often still below income tax levels). Income tax–preparation firms like H&R Block have set up shop in once desolate areas of Milwaukee. Welfare leavers who are now working go to them to claim the state and federal EITC, which can raise incomes by half. But one must have earnings to qualify. To this extent, working really is a passport to a new life.60 Some observers even think that the changes are improving family life, the troubled realm that social policy has strained to reach. When denied an income without work, poor mothers have more reason to come to terms with fathers, to get support or at least child care from them. Nationally, the share of welfare mothers who lived with a partner jumped from 7 to 14 percent between 1997 and 1999. Social workers in Milwaukee

Effects of Reform

•

213

say that, faced with the new work demands, more single mothers have married their partners, or at least live with them, in order to have another adult to rely on. The changes are especially notable for low-income blacks. The new families are often conflicted, not always a gain for children. It is still clear, however, that on average children do better with two parents rather than one. The family is at least growing stronger, not weaker.61 Along with the recent decline in crime and drug addiction, the inner city may well have turned a corner. It is true that former recipients usually still have low incomes even if they work. The unaided job market probably cannot solve this, even in the long-term. Public benefits in some form will have to join with earnings to assure the poor a decent life. Fortunately, expanded EITC and health and child care, from both state and federal governments, have already begun to raise low-wage incomes on a large scale. Politically, the working poor are much more appealing than the nonworking.62 Raising work levels may be only half the solution to poverty, but it is much the tougher half. Having crossed that Rubicon, social policy now faces a brighter prospect than it has had for forty years. The war on poverty is far from won in Wisconsin, yet the state’s welfare reform must be accounted the nation’s leading victory over working-aged poverty since it became a national issue in the 1960s. America has had other victories over need. The working poor have gained hugely from economic growth, as the elderly and disabled have from more generous Social Security. But until the recent era of welfare reform, no change in social policy had shaken the core of the poverty problem, which was low work levels and other social problems among poor working-aged families. Like nothing else, work enforcement does reach to lifestyle, even if it is far from a total cure. When work requirements are combined with generous work supports and intense administration, the effect is enhanced. That is the combination that Wisconsin, more than any other state, achieved under W-2. It was a victory of money, but above all an institutional triumph, built out of political will and determined implementation. The solution to the welfare problem was not to junk welfare but rather to rebuild the welfare state. At bottom, poverty is a breakdown of order. Government’s task is to restore order. That was a mission made for Wisconsin’s masterful regime.

CHAPTER 11

Welfare Reform and Good Government

I HAVE ARGUED that Wisconsin’s governmental excellence empowered it to reform welfare successfully. As table 1.3 claims, the state’s high-minded and ambitious measures led to effective welfare policies, which in turn led to successful outcomes. That connection seems evident in the Wisconsin case, but to prove it more rigorously I must escape the limits of a single case. I did that in chapter 5 by comparing counties within the state. I do it here by comparing Wisconsin to other states.1 I define more closely what governmental excellence means, and predict it using Elazar’s theory of state political culture. I also define more closely what successful welfare reform means. I test the link between culture and success in three different ways. All suggest that these well-governed states are indeed the leaders of welfare reform. To relate successful welfare reform to general governmental capacity is new. As discussed in chapter 1, most of the policy research on welfare or poverty is about the social or economic conditions surrounding these problems, and ignores public institutions. There is some research on the implementation of welfare employment programs that does address institutions, but it makes no connection between outcomes and the general capacities of the regime.2 In addition, state politics scholars study which features of states affect their policies. But in the welfare area, the main focus has been on explaining the formal policies states choose. Benefits tend to run higher, it has been found, in states that are more liberal and Democratic, that are richer or more developed, or that have stronger party competition. But showing which states are generous does not establish which can perform the more complex task of reforming welfare. Today, that chiefly means, not changing benefits, but attaching work and other conduct requirements to them. Other studies explore those choices, showing that states with conservative politics or high welfare rolls were the likeliest to experiment with welfare reform, and that states with large black populations have set the stiffest work conditions.3 My focus here, however, is more on the difficulty states have making and implementing these decisions.

Welfare Reform and Good Government

•

215

THE MEANING OF GOOD GOVERNMENT What, exactly, was the excellence that Wisconsin showed? What precisely does “good government” mean? It could mean having institutions that look capable in some general sense. One national study from 1971 rated state legislatures in terms of whether they were well staffed, clearly accountable, well informed, independent of the executive or lobby groups, and representative; Wisconsin’s legislature ranked fifth in the nation. In 1994, the state was one of nine found by the National Conference of State Legislatures to have a fully professional legislature—full-time, strongly staffed, and well paid. This feature is in turn linked to more generous welfare policies.4 But again, making welfare generous is different from reforming it. State administrative quality is tougher to define, measure, or explain. At most, the research suggests, quality tends to run higher in richer and more urban states. Studies of public management suggest that quality varies widely around the country. In these studies, Wisconsin does not rank particularly high.5 That might mean that our impressions of its excellence are in error. More likely, research on institutions of this static kind fails to capture the qualities that Wisconsin showed. Those had to do, less with the state’s institutional assets, than with the way it used them to make and implement policy. What stands out is the ability of the state’s officials to set ambitious goals and achieve them. Somehow, they were able to agree on radical policies and then implement them strongly. The closest that past research comes to tapping this enterprising quality is in the literature on state innovation, or the ability to concoct new policies and programs. In a 1969 analysis, Jack Walker rated Wisconsin tenth among the states in overall innovation. He wrote that “Wisconsin, because of its leadership during the Progressive period and its early adoption of the direct primary, the legislative reference bureau, and workmen’s compensation, gained a reputation as a pioneering state that it has never lost.” A later analysis by Virginia Gray rated the state fifth in overall innovation, but first in the area of welfare. A still later analysis ranked Wisconsin eighth or tenth before 1930 but only thirty-fourth during 1930–70.6 But innovation is only one trait of good government. Political and administrative mastery would appear more central. What makes government “good” is not that it does new things but that it governs well. This means two things—to make effective policy and to implement it. Of course, policy is in part made during the implementation process. Effective policy is not necessarily novel; rather, it is prudent in light

216

•

Chapter 11

of a state’s circumstances, and it is broadly acceptable. To implement policy well means to carry out the agreed measures effectively and efficiently.

POLITICAL CULTURE How states use their institutions to govern is probably best captured by the idea of political culture. I use that term here to mean the long-term beliefs that people in a state have about what politics and government are like, and how they ought to be. In short, the attitudes that pervade how the institutions actually operate. Political culture clearly differs among the states.7 The best-known theory of state political culture is the scheme developed by Daniel Elazar. I used it in chapter 3 to characterize the highminded tone of welfare reform politics in Wisconsin. To Elazar, Wisconsin was one of several “moralistic” states where leaders were expected to explain their actions in terms of a public interest and sell them to an undifferentiated public, rather than serve more specific interests. Government was seen as an instrument for building the community and advancing its welfare. Also, public administration in such states was strong, although with a preference for decentralization. In contrast, the “individualistic” culture saw politics as a marketplace akin to the economy. Individuals and groups entered the political arena principally to serve their own ends, not for more exalted purposes. Government was well developed, but it served specific interests rather than pursuing broader social visions. Finally, in the “traditionalistic” culture, government was less open and more elitist than in the other cultures. Its purpose was restricted to preserving traditional values and limiting social change. In these states, parties counted for less than in the other cultures, and the bureaucracy tended to be underdeveloped and distrusted.8 All three cultures were found throughout the United States, yet each state was dominated by one of them. Table 11.1 lists the states Elazar placed under each heading, plus subgroupings that he identified within each culture. The scheme is shown as of 1984, the last revision Elazar made. This date is just before serious welfare reform began in Wisconsin. In general, the moralistic states are found in upper New England, the upper Midwest, and parts of the West and Northwest. The heartland of individualism is the urban states in the Mid-Atlantic region and lower Midwest, while the traditionalistic states lie in the South and Southwest. Culture shapes the values that states serve through government. Elazar’s moralism is similar to the earnest but intolerant “good government” style often associated with anti-machine, “reform” politics in American

Welfare Reform and Good Government

•

217

TABLE 11.1 Assignment of States to Elazar’s Political Cultures Moralistic (17)

Individualistic (17)

Traditionalistic (16)

Moralistic Colorado Maine Michigan Minnesota North Dakota Oregon Utah Vermont Wisconsin

Individualistic/moralistic Connecticut Illinois Massachusetts Nebraska New York Ohio Rhode Island Wyoming

Traditionalistic/individualistic Alabama Arkansas Florida Georgia Kentucky Louisiana New Mexico Oklahoma Texas West Virginia

Moralistic/individualistic California Idaho Iowa Kansas Montana New Hampshire South Dakota Washington

Individualistic Alaska Indiana New Jersey Nevada Pennsylvania Individualistic/traditionalistic Delaware Hawaii Maryland Missouri

Traditionalistic Mississippi South Carolina Tennessee Virginia Traditionalistic/moralistic Arizona North Carolina

Source: Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984), pp. 134–37. Note: States included in the case study analysis are shown in bold.

cities. Most states commonly described as “good government” fall within Elazar’s moralistic group. In my usage, governmental excellence connotes the manner of Wisconsin’s policymaking—its problem-solving political style and skilled administration. Legislators, led by the Thompson administration, framed policy in a sensible and ambitious way, given their goals. State and local officials then implemented it creatively. These are precisely the qualities highlighted in Elazar’s definition. They seemed to explain why Wisconsin achieved more in welfare reform than other states of my acquaintance. In this chapter I test more rigorously whether this apparent tie between good government in this sense and successful reform extends beyond Wisconsin. Elazar’s scheme is a widely respected one that has withstood criticism. He derived his classification in a nonrigorous way, based chiefly on his own observation of the politics and the historical development of each state (as discussed further in chapter 12). But more recent scholars using

218

•

Chapter 11

more systematic methods have produced parallel results. The scheme does capture an important and enduring dimension of interstate differences.9 The Elazar cultures are distinct from other political differences among the states, for example in political ideology or attachment to the political parties. As a group, the individualistic states tend to be the most liberal, and the traditionalistic states the most conservative, yet Elazar’s cultures tap a different dimension of belief.10 At the same time, the cultural types are associated with many other features of states politics and government, much in the way Elazar posited. Moralistic states tend to show high political participation, competitive parties, strong merit systems, and liberal and innovative programming, while traditionalistic states show less of all these things. Individualistic states tend to fall in the middle in these respects while showing strong parties and more centralized administration than the moralistic group.11 One would expect citizens in moralistic states to be more participatory and trusting and more favorable toward government intervention than people in individualistic states, who in turn would show these attitudes more than people in traditionalistic states. The evidence for this is mixed.12 The cultures appear to capture the distinctive beliefs of political elites and activists more than those of voters. In moralistic states, for example, leaders are more attuned to liberal policies and also more opposed to corruption than in the other cultures.13 However, this focus on elites is an asset here, because in state welfare policy the views of leaders have mattered more than those of voters.14 One criticism is that the Elazar scheme may express little more than the long-recognized contrast between northern and southern states. It is true that all the moralistic states are northern, while virtually all the traditionalistic states lie within the Old Confederacy. Critics also suggest that Elazar captures little more than the presence or absence of ethnic and racial pluralism. The more moralistic states tend to be socially more homogeneous than the less moralistic.15 While this is true, it is the specifically political attitudes defined by Elazar that most directly sway politics and policy. I argue in chapter 13 that the cultures largely predate today’s pluralism and are not a response to it. There is no suggestion in Elazar that one culture fosters greater capacity to govern than another. All three styles, he suggests, contribute something valuable to American politics. Traditionalism celebrated timeless moral values, individualism the values of compromise and freedom, moralism the quest for the good society.16 Other scholars, however, suggest that governments with traits that Elazar would call moralistic do tend to govern better than regimes lacking these traits, at least if the criteria emphasize “good government.” Robert Putnam showed that regional governments in Italy were markedly more effective and responsive in

Welfare Reform and Good Government

•

219

the more civic, northern regions of the country than they were in the less civic South.17 An analysis paralleling Putnam’s showed that civic culture was similarly related to government performance among the American states. Here civic culture meant qualities of civic engagement, beliefs in political equality, trust in others, and cooperative social structures, and was clearly related to culture in Elazar’s sense. Government performance included qualities of policy liberalism, innovation, and administrative effectiveness.18 This analysis, however, is highly aggregated. What it implies concretely about governmental style and capacity is unclear. How the three cultures govern is better revealed by the way they handle a specific issue, such as welfare. It is unclear in advance which culture will handle welfare reform best. The priorities of all three styles could be assets in that task. In part, reform involves pursuing new conceptions of the public interest, as the moralistic culture emphasizes; in part it requires compromising intense political differences, a strength of the individualistic style; in part it involves affirming traditional norms such as the work ethic, a strength of traditionalism. Nevertheless, I expected that the moralistic states would fare best because of the meaning that reform currently has. The clearest demand TANF makes of states is to raise work activity on the rolls. Thus reform chiefly means crafting and implementing complex regimes of requirements and support services that will move the dependent into jobs. That endeavor should come easiest to the moralistic states, due to their problem-solving approach to legislation and especially their strong public administration. In comparison, the individualistic culture should display a more interest-centered and conflictive political style and a less enterprising bureaucracy, in welfare as in other matters. Among the traditionalistic states, we should find less interest in—and capacity for—reforming welfare. THE LEADERS OF REFORM My first test of that hypothesis is to see which sort of state led in the development of reform. The idea of welfare work programs as a way of reforming welfare developed chiefly through experiments that states ran during the 1980s and early 1990s under waivers of normal federal rules granted by Washington. This route enabled states seeking change to sidestep the partisan stalemate that paralyzed welfare reform at the federal level until PRWORA in 1996.19 By the time PRWORA passed, forty-five states had held a total of 105 waivers. Necessarily, states of all three Elazar cultures participated, at

220

•

Chapter 11

least in that they ran experimental programs. However, two moralistic states—California and Wisconsin—had much the highest number of waivers: 10 and 11, respectively. And as a group, the moralistic states easily led. The three groups of states are about equal in number. Yet the moralistic group held 48 waivers, compared to 30 for the individualistic states and 26 for the traditionalistic states.20 More important, experiments in the moralistic states tended to be the most successful and had the most influence on the national debate. A required AFDC work program in San Diego in the mid–1980s was the first to show sizable impacts on the employment and earnings of welfare adults. San Diego followed up with the Saturation Work Initiative Model, a special, federally funded effort to maximize participation; it showed even larger impacts. Then, in the early 1990s, Riverside County achieved still larger effects by combining high participation and a work first policy.21 Such results emboldened welfare reformers in Washington to stiffen work requirements nationally, first in the Family Support Act of 1988 and later in PRWORA.22 California is a moralistic state. The pioneering programs succeeded, in part, because they enjoyed strong local political support. More important, they demonstrated masterful organization, a feature of the moralistic style. San Diego had developed a talented group of employment administrators, due to its involvement in previous federal training programs. In Riverside, a strong welfare director built a program committed to his vision of reform. He rallied his staff around a work first conception while also holding them accountable for results.23 These counties showed the same enterprising administrative style that we saw in Kenosha or Grant County in Wisconsin. Another noted program of the early 1990s was the Single Parent Employment Demonstration Program (SPED) in Kearns, Utah, near Salt Lake City. SPED sought to raise participation rates by broadening the range of permissible activities and through intense staff oversight of clients. Around three-quarters of clients participated in education and training and two-thirds in actual work activities—unusually high levels. Similar programs developed in Iowa and Colorado. These states as well as Michigan and Vermont changed policies to move welfare in these same directions statewide.24 All these states fall within Elazar’s moralistic group. The most extraordinary local program of the middle 1990s was probably JOBS in Portland, Oregon. One of eleven projects in the National Evaluation of Welfare-to-Work Strategies, Portland easily eclipsed the other sites. Local welfare directors created the program through an ambitious collaboration with other work programs and the community colleges. The agencies “owned” the program together, setting and enforcing

Welfare Reform and Good Government

•

221

performance expectations on each other. Engaged staffs motivated their clients to get and keep good jobs to a degree that few programs have achieved. The state backed the effort by setting clear expectations for programs and clients while leaving administrators free to innovate, just as elected officials did in Wisconsin.25 Oregon is another moralistic state. The moralistic states did not do everything right. In Wisconsin, as noted in chapter 5, some counties resisted work first, and Milwaukee had serious administrative problems. In California, Los Angeles and Alameda did not approach the caliber of San Diego and Riverside. They came under pressure to abandon training-oriented programs for work first.26 But the most noted programs did have a moralistic culture behind them. Something about that style drove officials to take on welfare as a problem they could solve. Equally significant, when welfare work experiments fell short due to government’s limitations, this typically occurred in nonmoralistic states. In Chicago, for instance, an attempt to impose a serious work search requirement on AFDC failed to achieve any clear impact on employment or earnings. That probably was because the welfare bureaucracy, facing a large urban caseload, simply could not get control of it in the way Kenosha, San Diego, or Riverside did. In Florida, an early JOBS-based experiment recorded only small impacts, in part due to insufficient child care funding.27 In Massachusetts, the Employment and Training Choices program, or E.T., was one of the most touted welfare work experiments of the 1980s, but its influence was fleeting. This was partly because E.T. was voluntary and training oriented, and later developments made clear that mandatory programs with a heavier work focus performed better. Just as important, although E.T.’s top management was skillful, the program failed to rebuild the welfare bureaucracy around work.28 As in other individualistic states, government in Massachusetts was large but not ambitious, meant to satisfy interests more than to solve problems. The system did not, as in the moralistic states, produce administrators with the enterprise to take the lead in reform—nor politicians willing to support them.

NEW YORK The Empire State serves best to dramatize how the individualistic states differed from the moralistic.29 The difference between New York and Wisconsin is not mainly in wealth or in generosity to the poor. Both are affluent states that pay high welfare benefits. Rather, the contrast is in their governments’ capacity to deal with the stresses and strains of welfare reform. In Wisconsin, welfare was substantially reconstructed around work

222

•

Chapter 11

by 1996, then further transformed with the inauguration of W-2 in 1997. In New York, the individualist tendency to defend established interests defeated fundamental change. The state was pushed into reform by Washington. Where Wisconsin held eleven waivers prior to PRWORA, New York held two. While New York’s welfare rolls have fallen substantially as elsewhere in the country, reform in New York is still sharply contested, even seven years after PRWORA. This chiefly reflects political paralysis and serious administrative problems. New York suffers far deeper political divisions than Wisconsin. Much more than in the Badger State, political conflict centers on how much government should do, rather than how best to do it. New York City Democrats insist on unusually costly and expensive programs, provoking bitter resistance from upstate Republicans and business. In welfare, most Democrats and liberals cling to entitlement, the idea that the poor deserve public support simply on the basis of need and without behavioral expectations. On the other side, Republicans and conservatives aim mainly to cut spending in order to solve the state’s serious budget problems. Neither side shows the magnanimity or creativity displayed in Wisconsin. The division paralyzes action because a Democratic majority is entrenched in the Assembly, a Republican one in the Senate, thus allowing either side to veto change.30 In Wisconsin, due to much broader agreement on reform, shifts in partisan control of the legislature matter much less. The impasse means that New York cannot set its own course. In response to both the Family Support Act of 1988 and PRWORA in 1996, the state was able to do little more than comply with new federal requirements. Contentious questions about how tough to be about work requirements were delegated to the counties because they could not be resolved in Albany. While change has occurred upstate, it is fiercely disputed in New York City, and there is no statewide consensus about direction.31 Rudolph Giuliani, the Republican mayor of New York City from 1994 to 2002, aggressively sought to reduce dependency, and he brought in Jason Turner as welfare director, from Wisconsin, to accomplish this. Turner’s attempts to divert people from TANF and put many recipients in government jobs drove the rolls down, but they earned nothing like the acquiescence seen in Milwaukee. Opposition from community groups, poverty lawyers, and local Democratic politicians was intense, and press coverage has been far more hostile than in Wisconsin. The Giuliani administration rebuffed all requests to research its program, so fearful was it of academic criticism. On the other side, outnumbered conservatives supported the mayor with equal stridency.32 The state also faces serious administrative problems, especially in New York City. The bureaucracy is large and well funded by national standards, but, as in Massachusetts, it is reactive rather than enterprising.

Welfare Reform and Good Government

•

223

Administrators do not take the same initiatives seen in Wisconsin, in part because political support may be lacking. They await direction from elected leaders, but the latter are often divided or cautious, unwilling to advocate change. The capability of the bureaucracy is substantial but short of the impressive level seen in Wisconsin. Public service is seen as a secure career, not as a calling. Agencies also are more resistant to managerial improvements and accountability. Whereas Madison rebuilt welfare administration with only minor resistance from organized labor, New York’s unions put up tougher resistance.33 One long-standing problem has been an inability to build job search requirements into welfare. Traditionally, New York counted upon the Employment Service to provide this aspect of work programs. In the 1980s, after disappointment with ES performance, the welfare department took control of work programs. But lacking the ES’s performance measures for job entries and other outcomes, it could never deliver more than the activity levels demanded by JOBS. In New York City, Turner tried to build up performance management in welfare. “Jobstat,” an elaborate system of indicators, now allows top officials to track what the local offices are achieving, and to hold them accountable. Turner also reorganized offices as “job centers” and brought in private contractors to run job search. But the changes were slowed by lawsuits. Due to these institutional problems, New York has not approached the thorough integration of the assistance and work missions seen in Wisconsin’s W-2, or in Riverside or Portland. Michael Bloomberg, Giuliani’s successor as mayor, has already moderated his policies. States with generous benefits like New York often do not feel free to reform welfare, because work demands might put in doubt the support that they have promised to poor families.34 In New York, an obligation to aid the needy is written into the state constitution. This sense that welfare is a benefit promised to a part of the community is typical of the individualistic style, whereas in a moralistic state like Wisconsin, the image of welfare is more instrumental and ambitious. The program is supposed to solve a social problem. If it fails to do so, it may be changed and improved, even at some risk to the recipients.

A PROCESS ASSESSMENT My second test of a connection between government capacity and welfare reform is to see how states of the different cultures implemented TANF, the most recent and radical phase of reform. This is a good test because states confronted TANF at the same time, under similar conditions, and with similar goals. Virtually all states set about changing welfare seriously

224

•

Chapter 11

in the mid–1990s, in a time of prosperity. And virtually all of them chose the goal of moving the adult recipients more rapidly into employment, although their emphasis on work first differed. Even Wisconsin, which started to reform far earlier than most states, carried out its most radical changes in the mid–1990s. Because of this unusual parallelism, differences in outcomes among the states should reflect differences in political style more surely than we could normally assume with state policymaking. The Meaning of Success This analysis is oriented to process, to how the states dealt with reform, rather than to their policies. Using prior research and my experience in several states,35 I defined successful implementation in terms of the following six criteria: Political performance: Policymaking: Did the state government frame its own policy and act sensibly to achieve it, in light of experience? Or was it pushed into reform by Washington, or did it make mistakes that could have been foreseen? Policymaking means taking a real decision about AFDC and setting a direction of one’s own, not simply passing enabling legislation. Consensus: Did the state enact reform with enough agreement that the key changes were largely accepted in and outside government? This implies avoiding divisive opposition both in the legislature and among outside lobby groups. Without consensus, policy can be hostage to the next election, and it is more difficult to motivate the bureaucracy to change. Resources: Was the state willing and able to fund the programs implied by its goals, including additional bureaucracy, child care, and other support services? The focus here is on funding adequacy rather than the detailed capacities of the bureaucracy. Funding need not be generous but only sufficient to support the policies a state has chosen.

Administrative performance: Commitment: Were senior administrators implementing reform committed to the task? Or were they rushed into it by politicians? Other commitment issues include morale among staff, union cooperation with reorganization in the bureaucracy, and the degree of cooperation between the state and local government. Coordination: Were the bureaucratic units or agencies charged with implementing welfare reform able to work together smoothly? Or were there serious divisions, conflicts, or communication problems among them? Indicators here include divisions within the welfare department or between it and other agencies.

Welfare Reform and Good Government

•

225

Capability: Did the bureaucracy have the expertise and facilities needed to implement reform, such as the capacity for case management and adequate reporting systems? I distinguish these specific capacities from more general resource adequacy, considered under political performance.

The three political and the three administrative criteria run roughly parallel. Policymaking and commitment express the political direction of reform, the goal that policymakers or administrators choose. Consensus and coordination express the degree of unity that officials succeed in building around their decisions, either in the political arena or during the administrative process. Resources and capability express at two levels whether a program “has the horses” to get the job done. As before, I use political culture to predict government capacity. I expect that the moralistic states will tend to perform the best by the six criteria, followed by the individualistic states and then the traditionalistic states. Evidence My evidence on whether states met these standards comes from two national projects that tracked the implementation of TANF in about two dozen states during the late 1990s. The Assessing the New Federalism (ANF) study at the Urban Institute, in Washington, D.C., covers thirteen states. The State Capacity Study (SCS) at the Rockefeller Institute of Government, at the State University of New York, covers twenty-one states.36 The ANF study focuses more on the formal policy decisions of states, SCS more on the political and administrative process, but they both say much about how well governments functioned. The two together cover twentythree states. I have added a twenty-fourth—Oregon—using other sources. This is a large sample—almost half the states. Table 11.1 shows that it is broadly representative. It includes several states from all three of Elazar’s cultures, and at least one from each of his subcultures. This evidence has obvious limitation for rating state performance. The studies are not all by the same authors. The evidence is second hand. The authors did not explicitly assess states’ performance as I do here. Rather, they described the reform process in the states, and I infer performance from this.37 The picture given of the local level discusses only some localities, which may not be representative. The number of documents on each state varies from one to six.38 To minimize uncertainty, I assess only whether a state met a criterion or not; I make no judgment about degrees. I assume that success is more likely to go unmentioned than are problems. So I assume that a state did meet a criterion unless the documents indicate otherwise. Also, each criterion mentions several specific challenges; a state failed a criterion if it encountered trouble with any one of these. Thus,

226

•

Chapter 11

TABLE 11.2 Association of Culture with Performance in Welfare Reform Overall Score 6 5 4 3 2 1 Number of states

Moralistic

Individualistic

Traditionalistic

MN OR UT WI KS MI WA CA CO

OH MO NJ NY RI MA

TN

9

6

Number of States

AZ NC WV AL MS TX FL GA

4 5 5 5 3 2

9

24

χ2 = 25.6, p = 0.004

for example, a state failed policymaking if it made no serious policy of its own in response to TANF or was pushed into reform by Washington or made hasty decisions that had to be reversed quickly. On the other hand, I have multiple documents on most states, and this gives reason for confidence. So does the fact that the ANF and SCS projects overlapped substantially. Eleven states were covered by both, including most of the larger states. The accounts from different documents and studies generally agreed closely. In no case did one source claim institutional success while another spoke of failure. At most, problems mentioned in one document go unmentioned in another. In these cases, I judged a state to have had the indicated problem.

PROCESS RESULTS Table 11.2 summarizes where states from the three cultures fell in overall performance, with one point given for each of the six criteria satisfied. There is a strong link between the three cultures and performance. Generally, as expected, the moralistic states score high, the traditionalistic states low, and the individualistic states in the middle. However, there are exceptions and considerable variation within the groups, for reasons explained below. In addition, the political and administration dimensions of performance, as I have defined them, were strongly linked. States high on one dimension tended to be high on the other.39 Table 11.3 gives more detail about how the sample states scored on the six performance criteria. They generally performed better politically than administratively.

TABLE 11.3 Performance of Case Study States in Implementing Welfare Reform Political performance3 State

Number of Documents1 Culture2

Policy

Consensus

Resources

Administrative performance3 Commitment Coordination Capability

Total

Minnesota Oregon Utah Wisconsin

4 2 2 6

M M M M

1 1 1 1

1 1 1 1

1 1 1 1

1 1 1 1

1 1 1 1

1 1 1 1

6 6 6 6

Kansas Michigan Ohio Tennessee Washington

2 3 2 2 4

MI M IM T MI

1 1 1 1 1

1 1 1 1 1

1 1 1 1 1

1 1 1 1 1

0 0 0 1 1

1 1 1 0 0

5 5 5 5 5

California Missouri New Jersey New York Rhode Island

3 1 4 4 2

MI IT I IM IM

1 0 1 0 1

0 1 1 0 1

1 1 1 1 1

1 1 0 1 1

0 0 0 1 0

1 1 1 1 0

4 4 4 4 4

Arizona Colorado Massachusetts North Carolina West Virginia

2 3 4 1 3

TM M IM TM TI

1 1 1 0 1

0 1 0 1 1

1 0 1 0 0

0 0 1 1 1

0 0 0 0 0

1 1 0 1 0

3 3 3 3 3

Alabama Mississippi Texas

2 2 4

TI T TI

0 1 0

0 0 1

0 0 0

1 1 0

0 0 0

1 0 1

2 2 2

Florida Georgia

4 2

TI TI

0 0

1 0

0 1

0 0

0 0

0 0

1 1

1

Indicates number of documents reviewed to rate the state. Indicates state’s political culture. “M” indicates moralistic, “I” individualistic, “T” traditionalistic. See Table 11.1. 3 In state ratings, “1” indicates that a state satisfied a criterion, “0” that it did not. 2

228

•

Chapter 11

Political Performance POLICYMAKING

Most states made a significant decision about reform, and most decided how to recast welfare sensibly in light of their own goals and experience. Typically, states were dissatisfied with the training-oriented work programs of the JOBS era and shifted toward work first, as TANF also required. The distinction of the moralistic states was largely that they had most often run experiments in welfare reform and thus had the most systematic evidence to go on. Ohio and Colorado delegated decisions heavily to their counties, but within a framework set from the center. A few states did not seriously make reform policy. New York was so deeply divided that it took no significant decisions about TANF but rather delegated the chief questions regarding work requirements to its counties.40 Alabama and Missouri were pushed into reform by federal action and appeared to have little welfare policy of their own. In several southern states (Florida, Georgia, North Carolina), policymaking appeared to be casual and highly personalized, with the governor or legislators offering reform plans with, apparently, little inquiry or evidence behind them.41 Texas was incoherent; the state claimed to pursue work first but based its policy on a waiver program focused far more on remediation. Disarray partly stemmed from a fragmented government where multiple policymakers acted with limited coordination. CONSENSUS

On this standard, states had more trouble. States in the leading two groups in table 11.3 were able to form a majority behind reform that stretched behind the legislature. They came to terms with advocates who questioned reform. New Jersey and Rhode Island achieved this through elaborate consultative processes. In Ohio, there was serious dissent, but prior to TANF. In the moralistic states, consensus came easier because critics accepted that reform was inevitable and did not oppose it frontally, the same pattern seen in Wisconsin. The other states had more difficulty. In Arizona, a proposal to privatize reform in Phoenix led to acute controversy. In Alabama, efforts to consult with community groups became too contentious to support action. With its legislature unable to act, Alabama implemented reform chiefly through administrative channels. In both New York and California, legislative debate on reform was intensely partisan. In New York City, and also in Georgia, Massachusetts, and Mississippi, advocates or civil rights groups rejected reform decisions as racist or illegitimate.

Welfare Reform and Good Government

•

229

RESOURCES

Most states devoted sufficient funds to reform to afford the programs and services that their decisions required. A key issue here was whether they would or could fund sufficient child care to meet the demand for subsidized care, either from welfare mothers going to work or from the larger low-income population. Some states (Colorado, Florida, Texas) were unable to do this due to fixed limits on state spending or revenues that had been entrenched in law or state constitutions. These curbs expressed antigovernment attitudes that overshadowed reform. In other cases (Alabama, Mississippi, North Carolina, West Virginia), the documents suggest, states were simply unwilling or unable to “pay the piper” for political reasons. Administrative Performance COMMITMENT

In the top two groups of states, administrators accepted the goals as their own, at least at a senior level, and significant resistance lower down was absent. Indeed, in some cases, as in Wisconsin, local officials had taken the initiative in reform, with support from elected leaders. In Florida and Georgia, however, officialdom was dragged into reform from above and showed little commitment to it. In Arizona, the welfare agency had been committed to a remediation approach to welfare and resisted the shift toward work first. In Texas, welfare reform was a lower priority to administrators than rebuilding nonwelfare employment programs and other initiatives. In Colorado and New Jersey, local agencies had a history of defiance toward the state capital, and this prevented them from embracing reform plans coming from the center. COORDINATION

The challenge of avoiding bureaucratic divisions emerged as the most serious institutional problem faced by reform, at least in this sample. One important division, within the welfare agency, was between the income maintenance clerks that determined eligibility for benefits and the case workers who supervised clients in their search for work, either directly or through referrals to contractors. The other important divide was between welfare as a whole and nonwelfare employment agencies that it might ask to serve welfare clients. The latter included the Employment Service, the federally funded job placement agency, and training programs under the federal Job Training Partnership Act.

230

•

Chapter 11

Wisconsin’s solution to harmonizing the elements was the most radical. Internal to welfare, the state unified the eligibility and case worker roles in the person of the Financial and Employment Planner. Externally, in most localities welfare kept control of the welfare work mission, using the outside employment agencies only as contractors. New York and Tennessee also kept welfare in control of local implementation. New York and Wisconsin transferred some or all of welfare administration to the labor department at the state level, but this was largely symbolic. In Minnesota, Oregon, and Washington, all the agencies worked collaboratively to implement reform, apparently with few frictions. In Utah, welfare and employment programs had been merged in a single agency. All the other states encountered significant problems. In some instances (Arizona, California, Georgia, Kansas, West Virginia), there were troublesome divisions between the income maintenance side of welfare and the employment side. In California, for example, Greater Avenues for Independence (GAIN) had been run almost apart from the eligibility operation, and this obstructed the implementation of work first under TANF. More often, the problem was the interface between welfare as a whole and the nonwelfare work agencies. As recounted in chapter 4, the ES had been ousted from an assured role in welfare work programs during the 1980s because it had seemed to serve recipients poorly. Typically, the experimental programs that developed work enforcement in the 1980s were welfare-controlled. With TANF, however, Congress toughened work standards but abolished JOBS and all restraints on how states might organize welfare work programs. Soon after, the Workforce Investment Act (WIA) of 1998 required that states place the ES, JTPA, and other nonwelfare work programs under a single umbrella, governed through local “workforce investment boards.” The boards were supposed to create “one stop shops” serving the clienteles of all their programs. Many states now turned the welfare work mission over to this new structure. With welfare rolls much reduced, it seemed inefficient to many states to continue a separate work program just for welfare. The change gave back to the voluntary employment agencies an assured role in welfare employment, and ignored their earlier failure. In the majority of states in this sample (Alabama, California, Colorado, Florida, Georgia, Kansas, Massachusetts, Michigan, Mississippi, Missouri, New Jersey, North Carolina, Ohio, Rhode Island, Texas), this attempt to merge welfare work with WIA created serious confusion. The problems included unclear procedures to refer clients to WIA, to serve them there, and to report back to welfare on the results. The attempt to use WIA was the one blemish on the performance of Kansas and Michigan, two of the moralistic states. In extreme cases, such as Florida and Texas, the state’s determination to turn welfare work over to WIA

Welfare Reform and Good Government

•

231

paralyzed welfare reform, at least in the short run. In the longer term, one may doubt whether the voluntary work agencies will show any more receptivity to the special requirements of enforcing work than they did before. CAPABILITY

In the top group of states, administrators appeared to have the talents and facilities they needed to cope with welfare reform. Welfare staffs took on the employment role, often unifying it with eligibility determination, and they had adequate information systems to run the new work programs. At the other extreme, in the bottom group the bureaucracy was overwhelmed by the demands of reform. In Mississippi, the problems led to a failed attempt to turn over welfare to the private sector. The states that had difficulty unifying case management included Massachusetts, Rhode Island, Tennessee, Washington, and West Virginia. Those where management information systems were inadequate included Florida, Georgia, and Tennessee. In the background of these struggles is a longstanding dilemma in the southern states—an inability to draw enough capable people into public service. That reflects the low pay and prestige of public service in the traditionalistic culture. Some may suspect that the strong association between the Elazar cultures and successful welfare reform is deceptive. Perhaps the moralistic states excelled because of advantages they had other than culture. These might include less divisive political patterns, greater funding for welfare reform, or a less divided society. Especially, northern states that are often moralistic have usually also been rich. Due to their high benefit levels under AFDC, they also received more federal funding per recipient from the TANF block grant than did low-benefit southern states. The northern states also have relatively small minority populations, hence lower dependency on AFDC, so they may have been more willing and able to transform welfare. I tried to explain the states’ performance in implementing TANF using a wide range of factors like these. They included measures of state funding for welfare reform and many features of states’ politics and society, including racial composition. (Some of these terms are used later in this chapter.) I considered only variables that showed a strong association with performance on their own. But controlling for politics, economics, race, or institutional features separately did not much reduce the power of the Elazar cultures in explaining performance. Only if one controlled for all of these factors at once was the importance of culture clearly weakened. Even then, moralistic states still had a clear edge in implementing welfare reform.

232

•

Chapter 11

Furthermore, it is artificial to control for factors such as funding or administrative quality, as if they had nothing to do with culture. As the case studies suggest, the forces that help or hinder successful implementation tend to run together. The moralistic states excel in part because they have developed capable and well-funded administrations over generations. Culture pervades how each group of states does things, and that shapes their ability to implement complex social policies such as welfare reform.

A POLICY ASSESSMENT My third assessment of the role of government capacity is a statistical analysis that relates the Elazar cultures to welfare policies that influence the welfare rolls.42 The focus is not on which states avoided problems in reforming welfare, as above, but on what policies they chose. The leading goal of reform in most states was to raise work levels among welfare adults, either on or off the rolls. Measures of success would be performance indicators such as the share of clients entering jobs or their wages or retention, but no such measures have been developed.43 And measures that center on the recipients would fail to capture the important effects that reform may have outside the caseload. Accordingly, I use caseload change as a proxy for changes in work levels. Clearly caseload decline has been the chief driver behind higher work levels and the other outcomes of reform (see chapters 9 and 10). To identify the policies that affect caseloads, I use the cross-state caseload change models presented in chapter 9. The first of these explained variations in caseload change across the states during 1989–94, the second during 1994–98. In most states, the first was a period of caseload increase, the second a period of decline. These models assembled all significant explainers of caseload change in these periods, including both policy and environmental variables. The first model also contained one indicator of governmental quality (implementing reinvention of government measures), and the second contained two (legislative effectiveness and Elazar’s individualism). In the real world, these governmental influences should affect caseloads, not directly, but via the policy variables. So for present purposes I excluded them from the models. After some other changes, the remaining policy variables from the two models are those shown in table 11.4.44 I then asked how these variables related to the features of a state’s regime. My measures included many attributes of politics and government derived mostly from state politics research. The political variables (Table 11.5) include measures of public opinion (partisanship, ideology) and po-

Welfare Reform and Good Government

•

233

TABLE 11.4 Policy Terms from Structural Models of Caseload Change 1989–94 Maximum AFDC grant (family of 3) in January 1989 (dollars) State added Unemployed Parent coverage after fiscal 1989 Percent of welfare adults active in JOBS, fiscal 1991 Change in percent of JOBS participants in job search, fiscal 1991–94 1994–98 Maximum AFDC grant (family of 3) in January 1994 (dollars) Change in percent of welfare adults active in JOBS, fiscal 1994–96 Percent of JOBS clients assigned postsecondary education, fiscal 1994 State has delayed full-family sanction State has immediate full-family sanction Percent of AFDC cases getting child support, fiscal 1994 Source: See tables 9.5 and 9.6.

litical culture. The latter terms include the Elazar cultures but also parallel measures developed by other scholars.45 The measures of institutional quality (table 11.6) indicate indicators for the modernization of government and administrative quality, meaning the organization and resources of the bureaucracy. For sources, see the appendix to this chapter.

POLICY RESULTS I then ran correlations between the policy terms and the government quality variables. Based on limited indications from past research, I expected the political opinion variables to associate positively with caseloads; the scales of these terms run in a liberal direction, so higher levels on these scales should mean higher acceptance of dependency. So should higher levels of Democratic party control or party competition. I expected the modernization and administrative quality terms to associate negatively, because capable government—running programs tightly and efficiently— appears to be one of the forces restraining dependency. The political culture variables are more complicated. I assume that states with more diverse or minority populations will favor higher welfare. But, given the recent leadership of moralistic states in welfare reform, this culture must be seen as opposed to dependency. For these states, nonworking welfare affronts the public interest. Traditionalistic states, mainly in the South, should also oppose welfare, but in the name of inherited values.46 The individualistic culture probably promotes dependency, as this mindset accepts various claimant groups getting what they can out

234

•

Chapter 11

TABLE 11.5 Government Quality Explainers: Political Variables Variable Name Political opinion partisan ideology policlib govid dgovid citid dcitid ranney1 ranney2 Political culture diverse ethnic minority econdev elazarm moral1 moral2 civic soccap elazari indiv1 indiv2 elazart trad1 trad2 sharkan south deep

Variable Label

Expected Tie to Caseload Change

State partisanship (Erikson, Wright, and McIver, 1993) State ideology (Erikson, Wright, and McIver, 1993) Policy liberalism (Erikson, Wright, and McIver, 1993) Government ideology (Berry et al., 1998) Change in government ideology (Berry et al., 1998) Citizen ideology (Berry et al., 1998) Change in citizen ideology (Berry et al., 1998) Ranney party control index (Bibby and Holbrook, 1996) Ranney party competition index (Bibby and Holbrook, 1996)

+ + + + + + + + +

Social diversity index (Sullivan, 1973) Ethnic diversity index (Hero and Tolbert, 1996) Minority diversity index (Hero and Tolbert, 1996) Economic development rank (Citizens Conference on State Legislatures, 1971) State is moralistic (Elazar, 1984) Moralism index (Johnson, 1976) Moralism index (Morgan and Watson, 1991) Civic culture score (Rice and Sumberg, 1997) Putnam social capital index (Putnam, n.d.) State is individualistic (Elazar, 1984) Individualism index (Johnson, 1976) Individualism index (Morgan and Watson, 1991) State is traditionalistic (Elazar, 1984) Traditionalism index (Johnson, 1976) Traditionalism index (Morgan and Watson, 1991) Political culture score (Sharkansky, 1969) State is in South State is in Deep South

+ + + + – – – – + + + – – – ? – –

Note: For the 1989–94 time period, govid and citid are measured in 1989, and dgovid and dcitid measure change over 1989–94. For the 1994–98 time period, govid and citid are measured in 1994; dgovid measures changes over 1994–96, and dcitid measures change over 1994–97, the last years available. Ranney measures cover years 1989–94. For ease of presentation, variables expressed as ranks are reversed so that higher values mean better rather than worse. This also reverses their expected association with caseload change. South includes AL, AR, DE, DC, FL, GA, KY, LA, MD, MS, NC, OK, SC, TN, TX, VA, and WV. Deep South includes AL, AR, FL, GA, LA, MS, NC, SC, and VA.

Welfare Reform and Good Government

•

235

TABLE 11.6 Government Quality Explainers: Institutional Variables Variable Name Modernization walker grayave graywel grumm legeffec poldevel account represnt

Variable Label Innovation score (Walker, 1969) Overall average innovation rank (Gray, 1973) Average welfare innovation rank (Gray, 1973) Legislative professionalism score (Grumm, 1970) Legislative effectiveness rank (Citizens Conference on State Legislatures, 1971) Political development rank (Citizens Conference on State Legislatures, 1971) Government accountability score (Bowman and Kearney, 1988) Representation score (Bowman and Kearney, 1988)

Administrative quality exec Executive centralization score (Bowman and Kearney, 1988) avgrade Overall public management grade (Barrett and Greene, 1999) managav Average of five public management grades (Barrett and Greene, 1999) results Managing for results grade (Barrett and Greene, 1999) human Human resources management grade (Barrett and Greene, 1999) infotech Information technology grade (Barrett and Greene, 1999) reinvent Reinvention of government score (Brudney, Hebert, and Wright, 1999) reinvimp Reinvention implementation score (Brudney, Hebert, and Wright, 1999) statemps Number of state government employees per 10,000 population (Morgan, Morgan, and Quitno, 1994) locemps Number of local government employees per 10,000 population (Morgan, Morgan, and Quitno, 1994) staff Staff/spending score (Bowman and Kearney, 1988)

Expected Tie to Caseload Change – – – – – – – – – – – – – – – – – – –

Note: For ease of presentation, variables expressed as ranks are reversed so that higher values mean better rather than worse. This also reverses their expected association with caseload change. Avgrade is the overall grade given to the states by Barrett and Greene (1999); managav is the average of their five grades in five areas of public management.

236

•

Chapter 11

TABLE 11.7 Correlations of Government Quality Measures with Caseloads and Policy Explainers from Structural Model for 1989–94: Political Variables

Variable Name Political opinion partisan ideology policlib govid dgovid citid dcitid ranney1 ranney2 Political culture diverse ethnic minority econdev elazarm moral1 moral2 civic soccap elazari indiv1 indiv2 elazart trad1 trad2 sharkan south deep

Percent Change in AFDC AFDC Grant 1989–94 1989

State Added UP

Percent Adults Active in JOBS 1991

Change In Percent of JOBS in Job Search

– + – – .24 – + – +

–.36 .59 .78 .35 –.31 .53 – –.26 .35

+ –.64 –.71 –.35 + –.54 – – –.31

–.29 – + – – + –.36 – –

+ – + + – + + + +

+ + + + – – – – – + + + + + + + – –

.69 .46 –.40 .49 .38 .36 .24 .67 .57 .30 .58 .55 –.68 –.67 –.65 –.77 –.62 –.50

–.44 –.38 + –.44 – – – –.45 –.28 –.38 –.46 –.53 .51 .42 .47 .38 .31 .32

+ + –.36 – + .34 .37 .29 .42 + + – – – –.25 –.31 – –

– – – – + + + – – + + – – – + – – –

Note: Correlations shown are significant at p = .10 or better. For other values, only signs are shown.

of government. The fact that individualism was a positive influence on caseloads in the original model for 1994–98 supports this. Tables 11.7 through 11.10 show the correlations between the governmental quality measures and the policy variables from table 11.4. For context, I also show correlations between the policy terms and caseload change, although the governmental influences are assumed to affect caseloads via policy. Correlations that were large enough to be sure of the connection are shown, otherwise only the signs.47

Welfare Reform and Good Government

•

237

TABLE 11.8 Correlations of Government Quality Measures with Caseloads and Policy Explainers from Structural Model for 1989–94: Institutional Variables

Variable Name

Percent Change in AFDC AFDC Grant 1989–94 1989

Modernization walker grayave graywel grumm legeffec poldevel account represnt Administrative quality exec avgrade managav results human infotech reinvent reinvimp statemps locemps staff

State Added UP

Percent Adults Active in JOBS 1991

Change In Percent of JOBS in Job Search

– + – – – – – .36

.69 .66 .53 .27 .46 .68 .40 .44

–.46 –.51 –.38 –.41 – –.53 – –

+ – + – + + + +

+ + + + + + – –.26

+ –.24 –.24 – – – + – + –.24 –

.30 – – – – – .26 – + – .26

– – –.24 –.25 – – –.26 – + + –.25

+ + + – – + + + + .27 –.25

+ – – – – + – –.29 .24 – –

Note: Correlations shown are significant at p = .10 or better. For other values, only signs are shown.

1989–94 Table 11.7 shows that, during 1989–94, caseload change itself had remarkably few ties to any of the government explainers. This confirms that the sharp jump in the caseload in this period was not a political act. No one willed it. The increase was driven largely by higher unwed pregnancy in the 1980s and higher unemployment in the 1990s, and it hit states of all politics and traditions. Turning to the policy terms, however, higher AFDC grants are clearly linked to more liberal state politics, by most of the measures.48 They are also linked positively to social and ethnic diversity and to moralistic or individualistic political cultures, but negatively to minority diversity and traditionalism (both of which probably denote the South). Grants are also tied positively to all the modernization measures and to a few of the administrative terms. In short, generous welfare is a feature of liberal and well-developed states. Turning next to the addition of two-parent

238

•

Chapter 11

welfare coverage; this policy was forced on conservative states by the Family Support Act of 1988, so the signs are largely the reverse of those for benefit levels. The percent of welfare adults active in JOBS begins to show the distinctive influence of culture. Political liberalism and moralistic political culture no longer cut in the same direction. Enforcing work is a conservative measure but nevertheless favored by states that are moralistic or civic. And although there are few significant ties to the institutional variables, most of the signs are positive. This suggests the penchant for well-run work programs that we see in an Oregon or Wisconsin. For change in the percent of JOBS clients assigned to job search, there are few clear findings, but the pattern of signs again suggests a link to moralism and institutional development. Building up work programs is the policy of states that seek to restrain welfare, not by cutting aid, but by promoting good behavior. That is precisely the animus of the moralistic culture. 1994–98 Tables 11.9 and 11.10 show parallel results for 1994–98. Far more associations appear between caseload change and governmental antecedents than in the earlier period. The decline of welfare after 1994 was far more shaped by politics than the earlier rise. The fall is least in the more liberal and individualistic states, and greatest in traditionalistic areas including the South. That confirms our expectations about these cultures. For moralistic states the influence is less clear. Grant levels show much the same associations as in the earlier results. Although few significant associations appear with change in the percent of adults active in JOBS, the pattern of signs again suggests that this is a conservative policy that, nevertheless, is favored by moralistic states. There is also some tie with modernization and administrative quality. For the percent of JOBS clients assigned to higher education, these linkages are less clear. That may be because enrolling recipients in JOBS yet assigning them to college is politically ambiguous; it enforces activity aimed at self-reliance, but in a form that postpones an actual obligation to work. The two sanction variables differ only in degree, but politically they emerge as virtual opposites. In the context of 1994–98, when welfare politics everywhere was turning conservative, enforcing only a delayed full-family sanction was a liberal policy. It was also favored by individualistic states and—to judge from the signs—moralistic ones, but not by traditionalistic states. Strong sanctions—an immediate full-family sanction—on the other hand was a conservative policy that was opposed in individualistic and, probably, moralistic areas but supported by tradition-

Welfare Reform and Good Government

•

239

TABLE 11.9 Correlations of Government Quality Measures with Caseloads and Policy Explainers from Structural Model for 1994–98: Political Variables Change in Percent Change in AFDC Variable name

Political opinion partisan ideology policlib govid dgovid citid dcitid ranney1 ranney2 Political culture diverse ethnic minority econdev elazarm moral1 moral2 civic soccap elazari indiv1 indiv2 elazart trad1 trad2 sharkan south deep

1994–98

Percent of

Percent JOBS

Percent of

AFDC Grant Adults Active Assigned Higher Moderate

Strong

AFDC with

1994

in JOBS

Education 1994 Sanctions Sanctions Child Support

+ .47 .39 .29 + .40 + + +

–.36 .59 .75 + + .48 + –.24 .29

– + + – – – – – +

– + – – + + + – –

– + + – + + + – –

– –.38 –.36 – – –.41 – + –

–.53 –.28 – – – – + –.37 +

.48 .38 + .35 – – – + + .38 .42 .41 –.28 –.28 –.32 – –.24 –.27

.70 .48 –.43 .46 .32 .38 .25 .64 .60 .33 .60 .52 –.65 –.70 –.63 –.78 –.63 –.49

– – – + + + + + + – – – – – – – – –

+ – – – + – – + .26 – + + – – – – –.24 –

.28 + + + + + + + + + .37 .31 – –.30 –.29 – – –

–.49 –.24 + – – – – – – – –.50 –.39 .34 .47 .45 .34 .31 .41

– –.28 –.51 – .55 .42 .43 .34 .41 – – – –.36 – – –.43 –.29 +

Note: Correlations shown are significant at p = .10 or better. For other values, only signs are shown.

alistic states and the South. The signs in table 11.10 suggest that moderate sanctions may be linked positively to modernization and negatively to administrative quality, whereas with strong sanctions the associations lean the other way. The child support variable returns, more strongly, to the pattern seen with the level of JOBS activity in 1989–94. Getting child support paid on more welfare cases is clearly a conservative policy. But it is favored by the

240

•

Chapter 11

TABLE 11.10 Correlations of Government Quality Measures with Caseloads and Policy Explainers from Structural Model for 1994–98: Institutional Variables Change in Percent Change in AFDC Variable Name

1994–98

Modernization walker .33 grayave + graywel + grumm .30 legeffec .30 poldevel .26 account + represnt + Administrative quality exec + avgrade – managav – results + human –.31 infotech – reinvent + reinvimp – statemps + locemps –.30 staff +

Percent of

Percent JOBS

Percent of

AFDC Grant Adults Active Assigned Higher Moderate 1994

in JOBS

Strong

AFDC with

Education 1994 Sanctions Sanctions Child Support

.67 .65 .50 + .45 .66 .38 .43

– + + – + + + –

– + + –.25 + – – .29

+ + + + + + – +

–.34 –.33 – – – –.35 – –.24

– + + –.32 – + + +

.31 – – – –.25 – + – .34 – +

+ + + + + .28 + .34 – + –

+ – – – + – – –.28 .27 – –.27

+ – – – – – + – – – –

+ + + – + + – + – .27 –

+ + + – + + – – + .27 –.33

Note: Correlations shown are significant at p = .10 or better. For other values, only signs are shown.

moralistic states and not in the South. Again, we see the effort by the moralistic states to combine generosity with demands for good behavior. Child support also shows some positive association with most of the institutional terms. What Policy Is Best? Summarizing, the individualistic states have high benefits and have adopted only moderate sanctions, a combination that reflects their ambivalence about welfare reform. The traditionalistic states have both low benefits and strong sanctions; they restrain dependence simply by keeping people off the rolls. One side maintains entitlement, while the other denies any broad responsibility for the poor. The moralistic states, however, combine high benefits with strong work and child support enforcement. That

Welfare Reform and Good Government

•

241

is, they affirm a strong government role but deny entitlement. They seek to help the poor while also demanding that the poor help themselves.49 In this third analysis, to say which culture promotes successful welfare reform, one must be willing to prefer one of these paths to the others. Is it best to reform welfare incrementally by compromise, simply to restrict aid, or to try to enforce better behavior within welfare? Evaluation results as well as the Wisconsin experience argue for the last option. The combination of support with demands to function comes nearest to meeting the needs of the seriously poor. Since that is the bent of the moralistic states, I again find them to be the leaders of welfare reform.

CONCLUSION The findings suggest that Wisconsin was not alone in the mastery it showed about reform. Several other states approached the task with a similar high-mindedness and a comparable administrative prowess. One might write a book about any of these states, and the story would be similarly compelling. Admittedly, culture is not all. California and Colorado are two moralistic states that, despite manifest administrative talent, were pulled down by their internal divisions, at both the policymaking and administrative levels. Unlike Wisconsin, California could not achieve statewide what its pioneering counties had done. Conversely, Tennessee, although a traditionalistic state, scored unexpectedly high, kept out of the top group only by the limited capability of its bureaucracy. Much still depends on how leaders use the resources they are given. Admittedly as well, the moralistic states are favored by the way welfare reform is defined. Current national policy demands the creation and management of complex welfare work programs, a task in which the moralistic style excels. If, instead, reform meant the expansion of entitlement welfare, as was proposed in the 1960s and 1970s, then individualistic states with their benefit-oriented ideas of welfare would be the leaders. Similarly, if reform meant simply cutting or terminating aid, then southern, traditionalistic states might become the exemplars. The moralistic states might not excel at every kind of policy. There may well be issues that they handle less well than some other culture. As mentioned above, Wisconsin currently has budget problems that have not elicited the same high-mindedness from its leaders as welfare. In these areas, the cultures might be less distinctive. Something about the welfare issue, however, drives the three groups of states to behave in divergent ways. It is a problem that is simultaneously intensely divisive, deeply threatening to traditional values, and profoundly difficult to solve. So it

242

•

Chapter 11

is especially prone to draw from the individualistic states a temporizing response, from the traditional states a repressive response, and from the moralistic states elaborate attempts to govern the society. All of these responses are understandable, given the traditions they reflect. But the moralistic response is nevertheless the most admirable. This is the approach that achieves the most for the poor, the evidence suggests, and that is also the most legitimate to the public. The other responses are those that recent decades have taught the nation to question. After the burgeoning of dependency and related social problems in the nation’s cities, to give aid without behavioral conditions cannot be defended. But after civil rights and welfare rights, and given today’s affluence, the idea of doing little or nothing for the poor is equally repugnant. Large majorities of Americans, of every race and class, reject entitlement welfare because of its permissiveness, but still think that needy children and families deserve aid.50 Thus, on both policy and political grounds, all states are driven to embrace the difficult goal of combining aid with requirements. That is what most already claim to be doing under TANF. To that end, the moralistic states seem most able to form a public will and then to execute it with conviction. At bottom, reform requires deep belief in the public enterprise, and this is just what they have. Statecraft is their me´tier.

APPENDIX TO CHAPTER 11

Sources for Government Quality Variables Barrett, Katherine, and Richard Greene, “Grading the States: A Management Report Card,” Governing 12, no. 5 (February 1999): 17–90. Berry, William D., Evan J. Ringquist, Richard C. Fording, and Russell L. Hanson, “Measuring Citizen and Government Ideology in the American States, 1960–93,” American Journal of Political Science 42, no. 1 (January 1998): 327–48. Data used come from an update from the State Politics and Policy Data Archive at Florida State University, url: www.pubadm.fsu.edu/archives. This data covers 1960–96 for government ideology and 1960–97 for government ideology. Bibby, John F., and Thomas M. Holbrook. 1996. “Parties and Elections.” in Politics in the American States: A Comparative Analysis, 6th ed., ed. Virginia Gray and Herbert Jacob (Washington, DC: Congressional Quarterly, 1996), pp. 78–121. Bowman, Ann O’M., and Richard C. Kearney, “Dimensions of State Government Capability,” Western Political Quarterly 41, no. 2 (June 1988): 341–62. Brudney, Jeffrey L., F. Ted Hebert, and Deil S. Wright, “Reinventing Government in the American States: Measuring and Explaining Administrative Reform,” Public Administration Review 59, no. 1 (January/February 1999): 19–30. Citizen’s Conference on State Legislatures, State Legislatures: An Evaluation of Their Effectiveness (New York: Praeger, 1971). Elazar, Daniel J., American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984). Erikson, Robert S., Gerald C. Wright, and John P. McIver, Statehouse Democracy: Public Opinion and Policy in the American States (Cambridge, England: Cambridge University Press, 1993). Gray, Virginia, “Innovation in the States: A Diffusion Study,” American Political Science Review 67, no. 4 (December 1973): 1174–85. Grumm, John G., “Structural Determinants of Legislative Output,” in Legislatures in Developmental Perspective, ed. Allan Kornberg and Lloyd D. Musolf (Durham, NC: Duke University Press, 1970), chap. 12 (pp. 429–59).

244

•

Chapter 11

Hero, Rodney E., and Caroline J. Tolbert, “A Racial/Ethnic Diversity Interpretation of Politics and Policy in the States of the U.S.,” American Journal of Political Science 40, no. 3 (August 1996): 851–71. Johnson, Charles A., “Political Culture in the American States: Elazar’s Formulation Examined,” American Journal of Political Science 20 (August 1976): 491–509. Morgan, David R., and Sheilah Watson, “Political Culture, Political Systems Characteristics, and Public Policies among the American States,” Publius 21 (Spring 1991): 31–48. Morgan, Kathleen O’Leary, Scott Morgan, and Neal Quitno, eds., State Rankings 1994: A Statistical View of the Fifty United States (Lawrence, KS: Morgan Quitno Corp., 1994). Putnam, Robert, Data on social capital index downloaded from Bowling Along web page (http://www.bowlingalone.com/data.php3), June 14, 2002. Rice, Tom W., and Alexander F. Sumberg, “Civic Culture and Government Performance in the American States,” Publius 27, no. 1 (Winter 1997): 99–114. Sharkansky, Ira, “The Utility of Elazar’s Political Culture: A Research Note,” Polity 2, no. 1 (Fall 1969): 66–83. Sullivan, John L., “Political Correlates of Social, Economic, and Religious Diversity in the American States,” Journal of Politics 35 (February 1973): 70–94. Walker, Jack L., “The Diffusion of Innovations among the American States,” American Political Science Review 63, no. 3 (September 1969): 980–99.

CHAPTER 12

Origins of Excellence

THE QUESTION inevitably posed by the last chapter is: What produces a Wisconsin? If well-governed states do best at welfare reform, how does a state become well governed? It is not a question usually asked about social policy. I need to explain, not only the origins of Wisconsin’s regime, but why this state has exceeded other good-government states in prominence. In so doing, I at last document the connection asserted in the top line of table 1.3: Something about Wisconsin’s past gave it unusual governmental capacity, which it then used to transform welfare. That something, I propose, was a rare combination of society and polity. Wisconsin is conservative in its social attitudes, but at the same time it has a masterful regime born of a singularly fortunate history. This is a state that remembers only success in its public endeavors. That confidence is not the least of its resources.

A CONSERVATIVE SOCIETY The first question is where the will to reform welfare comes from. As shown in chapter 3, people in Wisconsin did not want anything unusual out of welfare reform. They criticized abuses of the old system and wanted to enforce work. But they probably did want change more intensely than Americans in many urban states. The initial reason was that the Midwest has more conservative social attitudes than much of the rest of the country. Uncivil behavior of all kinds is less tolerated here than it would be on the East or West Coasts. Table 12.1 contrasts opinion in the average state, in Wisconsin, and in regions of the country on several issues that reveal social attitudes. The figures in the table are arbitrary index numbers based on interview questions. On most of the issues, a higher figure means a more liberal or biggovernment response. However, higher religiosity or support for capital punishment would usually be seen as more conservative. The data show that Wisconsin is a bit more liberal than the average state on tolerance and race, but on other issues a bit more conservative. It is also typical of the Midwest.1 The contrasts of regions, however, show that the Midwest is considerably more conservative than the Northeast on most of these issues. Indeed, on several of them it is almost as conservative as the South.

246

•

Chapter 12

TABLE 12.1 Opinion on Social Issues for 44 States, Wisconsin, and Regions State Average Wisconsin Northeast

Issue Tolerance for unorthodox opinions Acceptance of racial integration More pro-choice on abortion Religiosity Acceptance of homosexuality Acceptance of women’s rights Welfare spending Capital punishment for murder

Midwest

South Mountain

Pacific

.60

.63

.66

.60*

.52*

.63

.72*

.73

.74

.79

.72*

.66*

.79

.84*

3.80 .55

3.62 .58

4.15 .52

3.72* .57*

3.56* .57*

3.74* .55

4.37* .45*

.19

.16

.28

.17*

.14*

.17*

.30

1.43 .35

1.40 .26

1.53 .35

1.43* .34

1.32* .36

.73

.68

.71

.72

.71

1.52 .33

1.58 .35

.85*

.72

Source: Paul Brace, Kellie Sims-Butler, Kevin Arceneaux, and Martin Johnson, “Public Opinion in the American States: New Perspectives Using National Survey Data,” American Journal of Political Science 46, no. 1 (January 2002): 173–89. Note: Data are from General Social Surveys pooled over 1974–98. Results cover 44 states. HI, ID, ME, NE, NV, and NM are omitted. Northeast includes CT, MA, NH, NJ, NY, PA, RI, and VT. South includes AL, AR, DE, FL, GA, KY, LA, MD, MS, NC, OK, SC, TN, TX, VA, and WV. Midwest includes IL, IN, IA, KS, MI, MN, MO, ND, OH, SD, and WI. Mountain includes AZ, CO, MT, UT, and WY. Pacific includes AK, CA, OR, and WA. * Indicates that the statistic is different from the corresponding figure for the Northeast at a significance level of p = .10 or better.

The Mountain states are a little less conservative than the Midwest, while the Pacific is the most liberal region. These results confirm subjective impressions that the nation’s interior is more attached to traditional values than the coasts. The immigrants who settled Wisconsin—heavily Yankee, German, Polish, and Scandinavian—were particularly attached to bourgeois values. In my experience, Wisconsinites simply refuse to accept the extent of social problems they associate with welfare. That attitude penetrates staff in the welfare system and even liberal advocate groups, who then hesitate to defend entitlement. In New York, welfare officials I have interviewed are cynical and resigned about welfare and poverty. Hard experience has taught them to disbelieve in any solutions. But in Wisconsin, people in similar positions expressed anger rather than cynicism. It is the same innocent outrage shown by Wisconsin politicians of both parties during deliberations on welfare reform.2

Origins of Excellence

•

247

Importantly, these attitudes extend even to urban areas. After cities grew large in America in the nineteenth century, urban reformers tried to stamp out vice by outlawing drink and prostitution. They failed, and eventually most Americans accepted that cities are inevitably disorderly.3 That acquiescence, however, apparently never reached Milwaukee. The city might be big-government, even “socialist” in its politics, my respondents said, but it is conservative in a cultural sense. “Its German-Polish roots are still dominant,” Senator Gary George told me. “It’s not like New York.”4 Those sentiments made welfare reform popular even among urban voters. Nor do the races differ over culture. Most black families in Wisconsin migrated from somewhere else in recent decades, but they have imbibed the local ethos. Blacks were no less dependent on welfare in Wisconsin than in other states, but their political response differed. In big cities elsewhere in America, the recipients tend to be spoken for by community groups that defend entitlement. But in Wisconsin, blacks in poor areas reached out directly to their representatives—and opposed traditional welfare. Antonio Riley, the pivotal black leader of reform, told me that welfare reform was not initially an interest of his. He took it on only after black constituents, rich and poor, told him that AFDC was ruining their community. “We can’t keep doing what we’re doing,” they asserted. Similarly, Joe Strohl, a leading Democratic senator, who is white, recounted that older blacks had joined whites in his Racine district in criticizing younger blacks who were moving in from Chicago and going on welfare. “It’s okay for me now to talk about welfare,” Strohl concluded, “because my black constituents say the same thing as others.”5 Some observers think that the debate about welfare and poverty is caught up in the “culture war” between those who defend traditional social mores—such as the two-parent family—and those who tolerate divergent lifestyles.6 There is truth to that on the national scene, and in some urban states. In Wisconsin, however, a culture war is hard to detect. This empowered government to seek radical change.

POLITICAL CULTURE Wisconsin also acted more effectively to vindicate these attitudes than most other states. It displayed an unusual capacity to govern—to decide the contentious questions in welfare and then to implement its policies. Where does this capacity come from? It has not been a common question in political science. Scholars typically ask why American government is smaller than one might expect. Why did it never collectivize the economy or enact a com-

248

•

Chapter 12

prehensive welfare state, as several European countries did? American government somehow lacks the same social democratic tradition. This reflects the absence of socialist labor movement, weak political parties, and the dispersion of power by the Madisonian political system, which obstructs the enactment of far-reaching legislation.7 Another factor is America’s fluid class system and its individualist beliefs, which make the inequalities of capitalism more acceptable than they are to most Europeans. American political leaders typically seek to expand opportunity for ordinary citizens rather than to expropriate the rich.8 But the main question here is not why government is large or small but why it is able, or unable, to govern in the most basic sense. In general, American government is capable. Americans complain regularly about the incompetence of those who rule them. Yet their governments at all levels make and enforce rules that would be beyond the capacity of most other regimes. The main reason is that the country was founded chiefly by the British. Centuries before any other large country, England achieved the rule of law and government by consent. Government showed little corruption by the standards of the time. It was also accountable to a political class larger than itself, although politics was not fully democratized until the twentieth century.9 Britain bequeathed these institutions to its colonies. The Americans rebelled against George III, but not against these priceless governmental traditions. From the outset, political wealth would distinguish the new nation even more than its affluence and power. Indeed, it was government’s ability to knit together a vast continent with an open society that made the nation’s ascent to world leadership inevitable. For no other country has combined strong government with a capitalist economy on such a scale. The British inheritance chiefly explains why the potential for good government exists everywhere in America, as it may not in Putnam’s less civic Italy. The variations in capacity within the country, however, are still significant. That is one point of the Elazar political cultures. Elazar himself traced those variations largely to the groups that first settled various parts of the country. The moralistic states were founded by Puritans from Britain and their Yankee heirs as well as Scandinavians and other northern Europeans. These groups brought with them demanding ideas about the morality and purposes of government. They first settled in upper New England, and then their descendants moved on to the upper Midwest and Northwest. The individualistic states were settled chiefly by other British groups and people from the interior German states. Immigrants from Ireland and southern Europe also entered these regions and southern New England

Origins of Excellence

•

249

in the nineteenth century. These states came to be pluralistic and capitalist but without the unifying purposes of the moralistic region. The traditionalistic culture, finally, was founded by British aristocrats with landed traditions. In Virginia and points south, they created a plantation economy based on social hierarchy and, eventually, on slavery. Here the main purpose of government was the defense of hierarchy.10 Elazar attached special weight to the religious background of the initial founding groups, and later research confirms this. In New England, the Puritans strove to create a more exalted community than they had known in the old country. Politics was participatory, but deferred to moral leaders. New England town meetings asserted wide responsibility for local conditions and problems, and were willing to tax accordingly. Through unremitting effort, Puritans believed, the community could approach an ideal of ordered liberty. This politics lives on in the moralistic states’ exacting standards for public behavior, their abhorrence of partial interest and corruption. Above all, it explains their insistence that government pursue some vision of the collective good. In the mid-Atlantic states, the Quakers and other religious groups also pursued exacting goals for themselves, but their strong commitment to tolerance denied government a comparable authority over the community. That heritage plus a more pluralistic pattern of settlement shaped the individualistic culture, with its attachment to freedom and economic advancement. The traditionalistic culture, in contrast, was founded by cavaliers inspired by an Anglican, class-based vision of society. Here, government was to defend inherited values, but it had no mandate to strive in any ongoing way for a better society.11 In its formation, Wisconsin is a classic moralistic state. Its earlier settlers were Yankees from New England and New York, and immigrants from Britain, Ireland, Germany, and Scandinavia.12 These groups had generally known government to be competent and honest, if not democratic. Wisconsinites thus derived high moral expectations for their politics.13 As one welfare manager remarked to me innocently, “There is an attitude in Wisconsin of wanting to do things well.” I have not heard officials talk this way in other states. In individualistic states like New York, New Jersey, or Pennsylvania, attitudes are much more disabused. Wisconsin also escaped much of the suspicion of big government that is a staple of American politics. Early citizens of the state, much like Canadians, valued the help that government gave them in conquering the frontier. Whereas Americans assert rights to “life, liberty, and the pursuit of happiness” against government, Canadians believe that goals of “peace, order, and good government” are to be achieved through government. This also became the attitude in Wisconsin. Later German immigrants to

250

•

Chapter 12

the state came from collectivist, even socialist traditions, often Catholic in inspiration, so they too were pro-government. Wisconsin was also far enough north that it escaped significant migration from the South, unlike other parts of the Midwest. It thus avoided the limited view of government that pervades the southern tradition.14

PROGRESSIVISM The political virtues of the state were further intensified at the end of the nineteenth century by Progressivism. This was a national movement to clean up politics and control the burgeoning industrial economy. Characteristically, Progressives were educated or professional people who opposed the machine politics of the big cities. They also worried that largescale capitalism would deprive individuals of the traditional American chance to get ahead on their own. After prolonged agitation, Progressives achieved important state-level political reforms such as the secret ballot, the selection of political candidates by primaries, and the replacement of patronage administration by civil service. At the national level, Presidents Theodore Roosevelt and Woodrow Wilson presided over the enactment of the Federal Reserve system and a number of agencies to curb the abuses of big business. Academics commonly ask how Progressivism, like the rest of the American tradition, affected the scale of government. They usually find the movement too individualist and too ambivalent about the state to truly achieve its ends. The New Deal was more truly collectivist, although it too operated more to rationalize capitalism than to transform it.15 But more relevant here is the positive effects Progressivism had on government’s capacity to rule, especially in Wisconsin. Progressivism’s chief leader was Robert M. La Follette, a Wisconsin Republican who put himself at the head of the movement, if he did not invent it. With its backing, he was elected governor of the state in 1900 and U.S. Senator in 1906. He became a leading Progressive spokesman on the national scene, even founding his own Progressive Party and running for president on its ticket in 1924.16 Within Wisconsin, Progressivism seems to have fortified both dimensions of effective government—the capacities to decide contentious issues and then to implement programs. Until the 1890s, politics in the state displayed many of the divisions common throughout the country. Politics was open but stalemated. The parties were seen as beholden to narrow interests—business and various ethnic groups—and were not widely trusted. There were rumblings from below but a sense that any government response to popular grievances was likely to benefit mainly the well-

Origins of Excellence

•

251

connected. “Mugwump” political reformers were already active, but they were too elitist to command a wide following. Cynicism and division thus stymied a forceful governmental response to the urgent urban and economic problems of the day.17 Then the economic panic of 1893 generated widespread distress. Somehow reformers came together and found a broader and stronger voice. Issues were redefined not as struggles between interests but as crusades pitting an undifferentiated public against all the interests together. Instead of an endless argument between partial views, a clear public good was seen, at least on leading issues, that government could and should pursue. This was the movement that La Follette came to lead.18 Its first victories came over political reform and the establishment of municipal utilities in Milwaukee, but it went on to much more: To insure and enlarge the functions of representative government, the progressives of Wisconsin, before the opening of the first World War, put into operation a direct primary, an initiative and referendum, a corrupt-practices act, a law governing campaign expenditures, a civil service law, and an antilobbying law. To protect the citizen against organized economic power, the state set up transportation, industrial, and public utility commissions to fix rates and guarantee services. It also had a tax commission, a state income tax, an inheritance tax, regulatory life insurance and banking laws, an act to assist co-operatives, an unfair-trade-practices act, new weights and measures regulations, and laws governing stock and bucket shop operations. To promote public safety and welfare it had child and family-labor laws, industrial safety laws, public health and pure food laws, a workmen’s compensation act, a conservation law, and an excellent system of education and extension work.19

The state’s political reforms made it one of the most open polities in the country. Wisconsin had nominations by direct primary from 1906 and an accurate apportionment of legislative seats by population from 1954, well before the Supreme Court mandated this step nationwide. As noted in chapter 3, politicians in Wisconsin are first of all individualists who care about issues, only secondarily partisans.20 In response to popular pressures, this system produced the cavalcade of programs just mentioned. Later decades brought further social and economic initiatives. Welfare reform is only the latest expression of a hallowed reforming tradition. That so open a system could achieve so much taught hopeful lessons about democracy. Elected leaders could actually be the servants of the people. The public grew confident that government would respond to its concerns: As John Fenton wrote: Again and again this same pattern of protest-leadership-action repeated itself in Wisconsin’s history. Let the voters be disturbed about the party bosses, and

252

•

Chapter 12

the successful reform candidate succeeded in obtaining a primary election law; let them turn their eyes to hills denuded of their forest cover and to polluted rivers, and their candidate gave them an effective conservation program; when the workers bemoaned their insecure lot, their candidates procured for them the nation’s first unemployment compensation and workmen’s compensation laws.21

To obtain such a government, voters accepted higher taxes than in most states. By 1970, only five states had a higher tax burden per capita; by 1990, twelve states.22 The political capacity to respond depended vitally on the character of the parties. Intense ethnic and class diversity was not lacking in Wisconsin. The parties expressed those fissures, as elsewhere in the country. Typically, Democrats stood for the ethnic groups of the cities, Republicans for the well-off and business. One side typically favored more government, the other less. In the individualistic states, the two sides became more cynical and entrenched during the New Deal years, as battles over programs and subsidies intensified. That pattern has continued since. In Wisconsin, there was less cynicism. Republicans dominated state politics going back to the nineteenth century. Their reign was challenged by the Progressive party during the 1930s and 1940s, and this led to a change in tone. The Progressives disbanded after 1946, their remnants joining the two other parties, but their practical, antipartisan style lived on. Both Democrats and Republicans came to think of themselves more as social reformers than as partisans. While each party favored some interests over others, both felt a compulsion to address evident public problems. Both had to justify what they did to a broad audience. That impulse forced their actual policies together over welfare reform and many other issues, despite their disagreements, thus providing a stable mandate for change.23 When Tommy Thompson claimed to be a Progressive, it was not a pose. A reforming style was part of his partisan past. The same style came far less easily to the GOP in Albany or Washington. Alongside politics, Progressivism also built up public administration. Traditionally, administrators had been chosen by elected political leaders under the “spoils” system. The Progressives demanded that the state be something above and apart from politics. Officials were now to be appointed by nonpolitical criteria, including examination. Wisconsin’s “merit” system became one of the most far-reaching in the nation. Today, appointments in the state involve very little patronage.24 Because of its many programs and its willingness to tax, Wisconsin built up a substantial bureaucracy during the Progressive era and after. By 1992, Wisconsin paid its employees an average salary that was eighth-highest in the country.25 The opportunities drew capable people into public service. These

Origins of Excellence

•

253

were the gifted officials who implemented, and in part created, welfare reform in Wisconsin. The success of this bureaucracy over time, like the problem-solving political process, taught a positive view of government. As Robert Nesbit writes, “A legacy of the progressive tradition in Wisconsin is the belief that a highly professionalized bureaucracy, backed by enlightened legislation adequate for the task, will save us from either rapacity or folly.” Wisconsin’s history taught its citizens to have “a level of expectation in terms of honesty, competence, humane motivation, and service that is a cut or two above the average among the states.”26 In 1971, the Citizen’s Conference on State Legislatures developed an index of overall political development based on many background indicators. Wisconsin ranked second in the nation, behind Minnesota; it was followed by Michigan with California and Washington tied for fourth. While the association of political development with Elazar’s cultures was far from perfect, these top five states were all moralistic. On an index of economic development, however, the ranking was quite different. Here the top five states were California, New York, New Jersey, Connecticut, and Illinois. All but California were individualistic. Wisconsin ranked nineteenth.27 The contrast perhaps suggests the different priorities of the two cultures—the one toward collective goals, the other toward the individual pursuit of economic opportunity.

NATIONAL LEADERSHIP The Progressive movement, however, was a national one that strengthened government in many states. We still have to explain why Wisconsin, among all the moralistic group, became the most noted exemplar of welfare reform. The reason probably is the Badger State’s long history of national leadership in social policy. In 1911, Wisconsin established the first state workmen’s compensation system and, in 1931, the first state unemployment insurance system. It also, in 1911, established the first state income tax, the better to fund its many programs. During the Depression, Governor Philip La Follette, son of Robert La Follette, developed a public works program for the jobless called—in anticipation of W-2—Wisconsin Works. All these programs became models for other states and the nation.28 In this period, two scholars rate Wisconsin as the most innovative state in welfare and social policy in the country. It was also a leader in how fully it covered its eligibles and the benefits it paid in early social programs.29 While Wisconsin’s prowess may not extend to all areas of government, in social policy it is conspicuous. Chapter 2 noted the state’s leading role

254

•

Chapter 12

in child support policy in the 1980s. It was the first state to institute a state-earned income tax credit, on top of the federal EITC. Wisconsin has also been a leader in health policy; its innovations include not only BadgerCare but an unusually generous and well-run Medicaid program and Family Care, a pilot long-term care program.30 And in 1990 in Milwaukee, the state undertook the first publicly funded experiment with vouchers in the public schools. In education, as in welfare reform, local black leaders supported radical change out of the conviction that the existing institutions had failed. Advocates of choice now rank the state third in the nation in educational freedom, behind Arizona and Minnesota.31 Some scholars believe that states are deterred from enacting costly social programs out of a fear that generosity will attract poor people while driving businesses away to states with lower taxes. Wisconsin has been remarkably undeterred by such fears. It insisted on inaugurating unemployment insurance even in the depths of the Depression.32 Fear of welfare migration and a desire to cut taxes were motivations behind Tommy Thompson’s initial welfare reform proposals. But neither impulse figures much in the later history of welfare reform in the state. Rather, Wisconsin proceeded in an inner-directed way, taking few cues from outside. Leaders made large financial commitments to reform and reaped large windfalls without paralyzing concern for the competitiveness or fiscal health of the state. In the moralistic style, they simply assumed that they could and should “do the right thing.” Leaders and citizens in Wisconsin simply assume that the state should get out front on national problems, as it has done in the past.33 This confidence was itself one reason for the state’s wide influence. Outsiders saw Wisconsin reforming welfare by its own lights and succeeding, until they were driven to follow. When Thompson made his initial deal over waiver savings with Washington in 1987, he broke with the National Governor’s Association, which preferred to reform welfare through national legislation. By pursuing waiver experiments, Wisconsin outpaced other states in changing welfare, as did Michigan. By 1995, NGA was split over reform and the main state-level influence on PRWORA came from Thompson and Michigan Governor John Engler.34 Other states, notably West Virginia, were influenced by Wisconsin when they crafted their own reforms.35 Given his formative role, it was not surprising that Thompson finally left his governorship in 2001 to become Secretary of the U.S. Department of Health and Human Services—the federal welfare department—in the second Bush Administration. Although not native to Wisconsin, the previous secretary, Donna Shalala, had also come from the state, where she had been chancellor of the University of Wisconsin at Madison.

Origins of Excellence

•

255

THE WISCONSIN IDEA Shalala was far from the first University of Wisconsin figure to go to Washington. A distinctive feature of Wisconsin statecraft is the “Wisconsin idea”—a notion that government and the state university should collaborate in policymaking. It is another expression of the high-minded moralistic style. From its founding in 1848, the University of Wisconsin had aimed to serve the state as well as academic goals. In this it resembled other state universities, but in Wisconsin the tie between university and government was unusually close. John Bascom, who became university president in 1874, was one of Robert La Follette’s teachers. He inculcated in him, La Follette recalled, “sound citizenship” and “a proper attitude to public affairs.” The university, he believed, should produce graduates who were devoted to “equal rights and the public welfare.”36 As if to symbolize the university’s mentoring role, atop Bascom hill at the university, a seated statue of Abraham Lincoln gazes down upon the state capitol, only a mile away. This vision was realized, above all, in social policy. Under La Follette and after, many university faculty entered leading posts in government. A series of Wisconsin economists became master craftsmen of state and federal social programs. The first of them, John R. Commons, helped Robert La Follette create the state’s pioneering workmen’s compensation system. One of his students, Edwin Witte, was Executive Secretary of Franklin Roosevelt’s Committee on Economic Security, which drafted the Social Security Act. That Act extended nationwide the unemployment insurance system that Wisconsin had pioneered. Two other Wisconsinites who worked with Witte became major federal administrators. Arthur Altmeyer headed the Social Security Administration, and Wilbur Cohen became Secretary of the U.S. Department of Health, Education, and Welfare. One of Witte’s students, Robert Lampman, in turn, was an early architect of national antipoverty policy. While on leave from the university at the Council of Economic Advisors in the Kennedy Administration, Lampman did research showing the extent of the poverty problem and arguing for a national response. That challenge was taken up in 1964 by Lyndon Johnson’s War on Poverty.37 Poverty was an economic problem, not a moral one, Lampman wrote. It was caused by social conditions outside the poor, such as joblessness, that government could and should change.38 Upon returning to Wisconsin, Lampman helped establish the Institute for Research on Poverty at UW-Madison in 1966. IRP embodied Lampman’s approach to poverty, centered on the opportunity structure rather than conduct. For decades the Institute received a special federal grant

256

•

Chapter 12

to fund its research. It played a leading role in experiments in income maintenance that became the main basis of proposals by the Nixon and Carter Administrations to extend and reform welfare. IRP experts also developed Wisconsin’s innovative child support policies in the 1980s.39 As noted earlier, most IRP experts did not support the conservative drift of welfare reform in Wisconsin after 1986, so they were excluded from the crafting of Tommy Thompson’s reforms or W-2. But if the Wisconsin Idea died in its traditional form, it reappeared in the reliance that Thompson placed on other experts from outside the state. IRP still offered considerable expertise in program evaluation. Eventually, the Management and Evaluation Project drew upon that capacity, and this reestablished a tie between the IRP and the state.

A ROUSSEAUIAN COMMUNITY Most Americans see society as a collection of individuals getting ahead on their own. The stress is on individual rights, asserted if necessary against the community. Americans do believe in some duties that everyone should fulfill, including obeying the law and working for a living; this explains the fierce popularity of enforcing work in welfare.40 But beyond this, individualism reigns. People act vigorously to advance their own interests. Liberty lies in the “silence” of the laws, Hobbes wrote. The social rules are mere “hedges,” meant to keep people out of trouble with each other as they struggle to get ahead.41 The image is of hurrying pedestrians warily avoiding each other as they dash ahead on the streets of Manhattan. In Wisconsin’s more dutiful culture, however, social expectations go further than this. The society expects that good behavior in public will be instinctive and pervasive. Hence the self-restraint that characterizes people in the state and elsewhere in the Midwest, at least compared to more liberated regions. Individual urges retreat behind a mask of good manners. The social ideal is no longer one of competitive striving, but of easy, cheerful interaction among citizens. At the University of Wisconsin in Madison, people sit out on the terrace at the student union, overlooking Lake Mendota. Most are strangers to each other, yet such is the trust between them that they seem as relaxed as they would be in their own living rooms. Equally important, Wisconsinites think government can promote such a society. Much of the country probably would not say this. Most Americans think, following John Locke, that government exists to preserve rights to life, liberty, and property that exist in advance of government.

Origins of Excellence

•

257

The regime is legitimate only so far as it acts to advance these rights. That comes to mean that it acts chiefly to expand freedom. It may do this in a negative or positive sense—by leaving individuals alone or by helping them get ahead. But it cannot normally seek to shape individual conduct or character in more pervasive ways. This is the ethos that especially pervades Elazar’s individualistic culture. In Wisconsin, however, a tutelary role for government is more accepted. This reflects government’s past successes as well as the state’s conservative attitudes. The image of government sometimes seems closer to JeanJacques Rousseau than Locke. Rousseau is best known for the radical democratic regime imagined in The Social Contract; where the “general will” is supposed to rule, there is no representation or minority rights, and people can be “forced to be free.” But equally important are the social conditions Rousseau set for his system. A true democracy, he believed, could arise only in an uncorrupted, rural society that had formed its citizens around a common morality. Then they would spontaneously seek the common good through an “inward assent of their wills,” without coercion. One role of government was to nurture that common culture.42 In such a community, to achieve a collective end comes first, and the choices of individuals are subordinate. People define themselves, not apart from the community, but in terms of the commitments they have made to others. That republican ideal was prominent in early American political thinking. It was eclipsed by rights-oriented thinking in the nineteenth century and since. But the rights perspective ignores the civilities required to achieve a desirable society. Welfare reform is a latter-day attempt to recover the idea that citizenship must entail some minimum of virtue.43 Nationally, neither left nor right can usually imagine social policy as a moral teacher. To the left, still wedded to rights, radical welfare reform as in PRWORA represents the tyranny of the majority over the poor and dependent.44 To the right, government is a corrupter rather than the teacher of virtue. It was precisely the permissiveness of traditional welfare and other programs, conservatives believe, that promoted nonwork, nonmarriage, and the other ills of the inner city.45 In Wisconsin, however, welfare in its reformed shape is seen as a facilitator of community. Tommy Thompson, for example, struck a communitarian note when he signed the legislation creating W-2 in 1996. By implementing the new system, he declared, [W]e can change our state forever to one where those who are able to work do so, and where those who are not are given the incentives and supports they need to enable them to do so. We will be a state where all citizens are educated and trained to work and expected to do so, where communities work together

258

•

Chapter 12

to provide temporary help to those who need it, and where the government of the state acts to enable persons to work, instead of simply providing cash to individuals who are not working.46

Thus a diligent but demanding government pursues a more perfect society. The rhetoric still echoes that “city on a hill” that the Puritans, the founders of the moralistic style, aimed to create in America. State administrators did not shy away from the tutelary purposes of welfare. They saw “the expectations they have of themselves,” one of them told me, and they thus wanted the poor to meet the same “expectations.” In rural areas like Grant County, the welfare agency seeks to “educate the community that an investment in the low-income working population significantly contributes to . . . Grant County’s societal and economical future.”47 On the other hand, recipients were told to meet the community’s expectations. One Learnfare case manager said to her clients, “I want you to be a good citizen,” she told me. Compliance was not a matter of choice. The idea that work should be voluntary for recipients was “foreign” to him, said Jon Angeli, the welfare director. “Why would we want to do that to anyone?” he declared. For only by cooperating could recipients earn “the dignity of working for their AFDC.” We usually think of welfare as standing outside respectable society. It is a safety net into which the poor and dysfunctional fall. Angeli, however, was a local notable, known to everyone in his town. He and his staff saw themselves, not as compensating for social failure, but as leading the society toward its goals. Rather than undercutting social values as conservatives say, they defined and expressed those values. The goal was to “sell the community” on helping the hard cases, said one JOBS case manager, because in that way JOBS was also “facilitating the community.” “You create an environment in your staff,” Angeli remarked, “and that gets conveyed to your community.” While welfare reform was new, the attitudes behind it were timeless. “Those of us who are out here always felt like this,” he said.48 Grant’s success, however, depended heavily on its rural setting. In that simple environment, few differences divided the community except that between the poor and the nonpoor. The needy were not shielded from social demands to function by separate neighborhoods, one reason dependency runs lower in rural than urban areas. It had become “more and more difficult to stay on welfare for long periods of time” due to these “pressures” as well as the various welfare reform programs, one state administrator said. These were just the kind of suasions Rousseau counted on to keep his ideal polity civic. In the cities, the struggle for an inclusive society was much more difficult. Angeli complained that “People in Milwaukee don’t want to hold

Origins of Excellence

•

259

people responsible” the way Grant did. But it was harder to levy expectations effectively in the cities because of their scale and complexity. Above all, the nonwhite character of most of the poor in Milwaukee separated them from the mainstream in ways going beyond dependency or nonwork. Milwaukee had struggled with race and poverty for decades, and it was not notably more successful than other big cities. Conflicts over housing discrimination, school integration, and crime seemed endless.49 Nevertheless, Wisconsin’s earnest culture still changed the tone of these struggles. Government was at least responsive. It came up with repeated plans to address the problems, even if there were no easy answers. The belief that public solutions could be found was unquenchable. This reflected the state’s fortunate past: “Since the era of the early progressives at the turn of the century, the people of Wisconsin had come to believe in the power of legislation and in the idea that good laws enforced by honest and vigorous public servants could resolve virtually any problem involving discrimination and the exploitation of one group by another.”50 In similar battles in New York, in contrast, the mood was much less hopeful. Above all, there was much less trust in the fairness and capacity of government itself.51 Wisconsin’s social problems are not unusual, but its political culture is, and in that lay the society’s best hope. The vision behind W-2, stated by Tommy Thompson, was of a community of mutual aid in which all citizens claimed membership by working. Through welfare reform, government struck a telling blow to realize that hope, even if it remains distant. The job centers scattered around Milwaukee were new, but they defended the same ancient values as in Grant County. The combination of unyielding work tests and lavish support services did at least raise work levels sharply, while reducing need. The lifestyle of working had at least started to penetrate poor neighborhoods, even in Milwaukee. Heartened by its past, government continued to chip away hopefully at the social problem.

CONCLUSION Most states that have large governments also suffer serious social divisions. Government is rich enough to do much, but its ability to address problems is limited by political differences. This is the pattern typical of the individualistic states, such as Illinois, Massachusetts, and New York. In the traditionalistic states, society is more cohesive, but government also attempts to do less. This is the pattern in the South. One tradition is able to give out welfare without conditions, the other to deny it. Neither finds easy the governing of the dependent that current welfare reform requires.

260

•

Chapter 12

In Wisconsin, in contrast, society is cohesive, yet government is ambitious. Both the society and government are strong. The state was willing to address the welfare problem, even in Milwaukee, and it had the governmental capacity to do so. It willingly took on the paternalist mission of changing lifestyles, not simply aiding the poor. And because of its reforming history, Wisconsin also outpaced other well-governed states. This unique combination explains why Wisconsin led the nation toward the transformation of welfare. Wisconsin’s greatest resource, in the end, was its indomitable self-confidence. The state believed in the poor themselves, refusing to condescend to them as other generous states have done. They were able, it asserted, to join a community of citizens defined by their capabilities rather than their needs. Above all, Wisconsin had confidence in its regime. This is a government that, at least in its collective memory, has never been defeated. So it proceeded calmly to recast welfare, even in the big city. In the Biblical phrase, Wisconsin’s welfare reformers were “wise as serpents and innocent as doves.”52 They drew on deep knowledge and experience about welfare. But they also showed a childlike faith that they could, like previous Wisconsin reformers, overcome the problems of their day. Faced with the inner city, some might call that confidence naive. But without it, who would have dared the revolution? In statecraft, innocence is a kind of power.

CHAPTER 13

Implications

IN THIS FINAL CHAPTER, my principal question is what the Wisconsin story implies for the nation. This gifted state succeeded in transforming welfare so as to enforce work without ending aid to the needy. But if that achievement rested on special conditions, as the last two chapters suggest, it may not mean much for less fortunate localities, or the country. What does Wisconsin really say to more troubled urban states with large caseloads, such as California, Massachusetts, or New York? In part, Wisconsin’s contribution is to good policy. In W-2, we see an advanced conception of what a work-based welfare system might mean. Wisconsin also suggests the power of a paternalist approach to poverty to reconcile humanity with civility. Some other developments in social policy outside welfare show the same. More important, however, Wisconsin shows that governmental quality matters. Only a government that is competent, in both political and administrative ways, can carry out the complex measures that today’s poverty seems to require. To overcome poverty, therefore, we must worry about governmental quality as much as the substance of policy. That message is relevant to states outside the current good government group. The capacity of American government does appear to be growing. This bodes well for solutions to both race and poverty problems. In some ways, Washington has promoted stronger government in social policy, and it can still do so. Alongside institutional challenges, welfare reform still faces political criticism. The nation as a whole has not attained the consensus behind reform seen in Wisconsin. The effort to condition aid on good behavior remains deeply controversial. Nevertheless, enforcing and facilitating order through strong institutions remains the nation’s best approach to poverty.

A WAY FORWARD IN SOCIAL POLICY Wisconsin broke the mold of the old welfare debate, and that is its first contribution to a better social policy. Traditionally, the struggle was between doing more for the poor and doing less. The Wisconsin reform, however, was antigovernment in one sense but pro-government in an-

262

•

Chapter 13

other. By enforcing work, the state reduced costs and the number of recipients. It also privatized parts of social administration. At the same time, however, it provided still-generous benefits, added new support services, and accepted a wider responsibility for the working poor, on and off welfare. The Wisconsin story refutes the conviction of the traditional left that the solution to poverty is simply to give more resources to the needy. Merely to express humanitarianism through government is not a solution to the current social problem. On the other hand, Wisconsin also refutes the credo of the traditional right that welfare per se is an evil. The answer to poverty cannot be to yield to frustration and simply do less. Events proved that helping people, even generously, was not counterproductive as long as those aided also helped themselves. Once poor adults went to work, government could profitably do much more to aid them than before. And doing so is politically popular. All states have had this experience to some extent. Wisconsin, however, squeezed more out of reform than most other states. Its caseload decline and work increase were unusually steep. That was probably because it both demanded and facilitated work with unusual force. A third element was the intensive administration that tied the reformed aid system together. Close oversight of the recipients was critical to motivating them to get jobs and also to claim all the benefits the system offered. Some other instances of paternalism outside welfare reform dramatize the nature and potential of that approach. Criminal justice has seen a trend to use the police less as crime fighters and more as enforcers of order. Rather than pursue offenders after crimes occur, officers patrol neighborhoods and develop ties with residents in an attempt to deter offenses in advance. They assiduously prosecute minor crimes to forestall major ones. In short, they apply authority. They also seek consent for these actions, both through these local contacts and through the political process, just as welfare reform had a political basis in Wisconsin.1 In corrections, it has not been found that incarcerating lawbreakers changes their behavior after they are released. Attempts to resocialize delinquent youth through such devices as “boot camps” have failed. The regimentation is too brief, and after it, youth no longer face a controlling structure. More effective appears to be supervising youth or drug offenders once they are on probation or parole. What helps them go straight is the knowledge that the parole officer is constantly checking up on them. They typically appreciate the attention.2 A notable success for this approach occurred in Boston in the 1990s. There homicides due to gang violence had exploded; 155 youth died between 1990 and 1994. Then law enforcement agencies at all levels of government formed a collaboration known as Operation Cease Fire. The

Implications

•

263

project had many facets, including interdiction of the illegal weapons used in shootings. But the core of it was a preemptive stance by the agencies against violent behavior. Teams of officers visited poor neighborhoods to warn gangs that further offences would draw time in federal prison. Parole officers went into the field with the police to be sure that their charges obeyed curfews. Homicides of persons under seventeen dropped to zero between 1995 and 1997.3 Some offenders were locked up, but success came mainly from deterrence. Officials expressed and enforced the community norm against violence. They restored an expectation that had been lacking, and this changed behavior. The ability to do this was a substantial administrative achievement. But since 1999, homicides have rebounded somewhat. One reason appears to be that the oversight of potential offenders has declined in intensity and focus.4 The authority to enforce ultimately came from politics. Cease-Fire gained critical support from the Ten-Point Coalition, an alliance of local black clergy. Like black leaders in Milwaukee, desperation drove them to accept radical change. Cooperation with a largely white power structure was preferable, they decided, to the continuing destruction of their community. They agreed to support the crackdown in return for some input into how individual offenders were handled. Like the broad coalition behind reform in Wisconsin, this preempted resistance to the new policy.5 A further example of paternalism in some localities is public school reform. In education, as in welfare, experience has proven that merely to spend money achieves little. Just as higher welfare benefits do not cause poor adults to overcome dependency, so investing more in the schools has little tie to how much students learn. Instead, effective schools involve a subtle mix of standard-setting and support by teachers and outside authorities. Higher and clearer norms are set for students, and are then backed up by close monitoring and extra attention for slow learners. Such schools typically have strong principals and motivated teachers who maintain order and focus energies on learning.6 The impetus for schools to perform could come from forcing the public school system to compete for students, by allowing parents to send their children to charter or private schools using public money, but this is too controversial to have widespread appeal. Instead, most states are trying to improve education in the same manner as in welfare and criminal justice—through enforcement. Since the 1980s, most states have set stiffer performance standards for teachers and students. Many, as well, have spent new money on instruction, special help for students, and school facilities. The states that have made the most progress seem to be those that do both of these things well, such as North Carolina and Texas.7

264

•

Chapter 13

GOVERNMENTAL PRECONDITIONS The way forward in social policy, then, seems to entail this complex process of governance—setting improved standards and then investing in the various benefits and supports needed to achieve them. But to do these things well demands good government. That is another implication of Wisconsin. Elected leaders must decide contentious questions of social standards, and then administrators must execute complex programs. Not just Wisconsin but the moralistic states as a group seem best able to meet these challenges. That is because in that culture decisions justified by a public interest have unusual authority, and because public administration is strong. Thus, to improve social policy in America also requires improving the general quality of government. The very idea raises what might be called the Robert Putnam problem. In Making Democracy Work, Putnam and his associates showed that the northern regions of Italy were markedly better governed than the southern ones. This, they said, was because northern political culture was more civic. Citizens and their leaders in the north were more egalitarian, honest, and trusting of government and each other, compared to those in the south. These attitudes empowered government to be more responsive, effective, and uncorrupt.8 What repels about the analysis is its implicit determinism. Putnam suggested that the fate of the Italian regions had been settled deep in the Middle Ages. The towns of the Italian north had developed a much more civic style than the south, and from this difference a wide gap in public capacity followed. Putnam also said that the less civic regions could improve their capacity over time, but that progress would inevitably be slow. Does my analysis suggest the same fate for the less-well-governed parts of America? Yes and no. Differences in governmental quality among the states are indeed important and enduring. They do set some limits on what government can achieve in the short-term. But at the same time, America generally is better governed than Italy, and the differences in capacity among regions are probably smaller.9 The potential to be well governed appears to be widespread. Much depends on how leaders use the institutions they are given. American political culture may also be more malleable. Italy had no national government between the fall of Rome and unification in the nineteenth century, so its better-governed regions could not impose their institutions on the others. In America, in contrast, a national government with strong civic values was established in 1787 and has ruled with rising authority ever since. From the outset, this federal regime put pressure on the American traditions that were less civic. Over time, moralism has

Implications

•

265

strengthened, and this makes solutions to poverty more feasible. The effects are seen nationally, not only in Elazar’s moralistic states. Even at the founding, the existence of slavery was widely seen as incompatible with ideals of equal rights. The issue was obfuscated at the time, then later resolved in a civic direction, first in the Civil War and a century later in the civil rights revolution of the 1960s. Daniel Elazar judged that the overthrow of slavery and Jim Crow permanently weakened the traditionalist strain in the national makeup. The Civil War, he wrote, “made possible the ultimate ascendancy of the moralistic and individualistic cultures on the national plane.” Southern politics has since become more individualistic and less oligarchic.10 Later, the individualist culture came under pressure. This style promoted the politics of ethnic groups and economic interests that dominated the most populous states during the nineteenth century. But regimes based on horse trading and corrupt urban machines could not be reconciled with national ideals of honest and impartial government. Protests against the abuses led finally to the Progressive reforms to clean up elections and curb partisan influence in administration. The individualist style, which originally distrusted governmental power, also had to accept the welfare state as necessary to serve wider opportunity for ordinary Americans. That forced it to take on complex tasks of governmental management that were more native to the moralistic style.11 These events strengthened the moralistic streak in the culture at the expense of the others, and the trend continued over the twentieth century. Challenges to machine politics persisted after World War II, leading to the overthrow of Tammany Hall in New York and comparable changes in other cities. Clubhouse politics ceded to a more earnest style where candidates made appeals about issues directly to the public. Parallel changes hit national politics at the end of the 1960s. Demonstrations surrounding the1968 Democratic national convention led to reforms by both parties to open up the presidential nomination process. The selection of national candidates has since been dominated by direct primaries, similar to those the Progressives had imposed on state and local elections decades before. The national parties as organizations have become bystanders to elections dominated by self-starting candidates—much as in Wisconsin. In Washington, think tanks and other advisory groups that argue issues on the merits have obtained rising influence since the 1970s. The national press exerts pressure on leaders to approach issues in these terms rather than strike deals. Today, experts often define the alternatives among which public officials choose. Indeed, many appointees to executive office are experts themselves. At the same time, traditional pork-barrel politics retreated as fiscal pressures forced cutbacks in budget subsidies, and deregulation exposed protected industries such as airlines to competition.

266

•

Chapter 13

Interest groups faced opposition from public-interest lobbies. The shifts partly reflected the growth of a more educated, middle-class electorate that rejected the older, politicking style. The new, more open environment demanded that policies be readily justifiable in terms of a public interest.12 One can argue that the new, more earnest style weakens government by moralizing decisions, so that it is tougher to make practical compromises among opposed interests.13 That might be true regarding routine decisions about budgets and priorities, but government’s capacity to take on the largest issues and resolve them effectively is probably strengthened. As chapter 11 suggests, poverty or welfare reform is such an issue. Alongside these political changes, state and local government underwent a renaissance. Traditionally, most localities had not been known for strong bureaucracy. State governments were fragmented, and governors often had little authority over them because of the separate election of other state executives.14 But since the 1960s, many states have given governors more power, reduced plural executives, and reorganized the bureaucracy. Both governors and legislatures have stronger staffs. These steps plus the expansion of agencies gave localities much more capacity to implement programs than they had a generation or two ago.15 For all these reasons, the political cultures of the states are far from static. Trends toward greater moralism are palpable across the country. That is also because the state cultures are themselves complex. As noted in Table 11.1, Elazar assigned each state a dominant culture, but he also thought most states exhibited a secondary tendency as well. While seventeen states were dominantly moralistic, another ten states displayed moralistic tendencies. In the second and third edition of American Federalism, in 1972 and 1984, Elazar added a moralistic subtheme to the individualistic cultures of Massachusetts, New York, and Rhode Island, and to the traditionalistic culture of North Carolina. Only traditionalistic New Mexico lost a moralistic subtheme.16 These changes reflect the strengthening of moralism in American political culture over time. THE PLACE OF RACE A focus on change over time permits a fuller answer to the idea, considered briefly in chapter 11, that the Elazar cultures are just proxies for ethnic and racial differences among the states. The moralistic states as a group are socially the most homogeneous, and the traditionalistic states, with their large black minorities, are the least; the individualistic states fall in between.17 The implication may be that the cultures actually are negative responses to race. Perhaps moralism is the culture of a white, homogeneous society that will not accept nonwhites, while the other cul-

Implications

•

267

tures represent an effort to control or limit black claims. I would argue, rather, that the cultures predate the race problem, and the tendency of moralism to strengthen will favor solutions to poverty and race. At the national level, race definitely shaped social policy. The new social programs created by the Social Security Act of 1935 either excluded most blacks or allowed states the discretion to exclude them. This was to assure the South that the black labor that it depended on in agriculture would have no alternative to work. Great Society programs focused much more on improving the skills or education of the poor than on assuring them jobs or income, as a European welfare state might have done. This reflected, in part, widespread doubts as to whether the black poor were “deserving” enough to merit new social guarantees.18 The skills focus also reflected the manifest problems that poor blacks had in getting through school and staying employed. On the other hand, the poverty problem has become more distinct from race over time. The welfare poor are racially more diverse today than they were in the 1960s, due to rising numbers of Hispanic and Asian immigrants. The welfare boom of the 1960s was sharpest among blacks; by the time it was over, in 1973, 46 percent of AFDC cases were black, only 38 percent white, and 13 percent Hispanic; virtually none were Asian. By 1998, however, only 37 percent were black while 36 percent were white, 20 percent Hispanic, and 5 percent Asian.19 The officials seeking to reform welfare are equally diverse. This is one reason why it is plausible to discuss welfare reform in nonracial terms, as Wisconsin did. At the state level too, as chapter 11 showed, culture clearly has shaped responses to welfare reform. But it is incorrect to view these styles as responses to race as such. As Elazar showed, the three cultures were formed chiefly by the initial settlement of the American states, well before race became an important issue in politics. David Fischer has argued that race became an important question even in the South only in the later eighteenth century, when large-scale importation of slaves began, well after all the regional cultures had gelled. Indeed, the cavalier culture of Virginia created the southern racial caste system rather than the other way around. Southern whites imported Africans in order to re-create the rural proletariat that their aristocratic social vision required, a role that white immigrants had refused to play.20 If race were fundamental to culture, then culture would change when demography does. In cases of heavy migration in the American states, that sometimes happened. The originally Puritan culture of Massachusetts was “disrupted,” Elazar says, by the heavy immigration of Irish and other groups in the nineteenth century, producing the mostly individualistic culture that we see today.21 Heavy immigration to urban states around the turn of the twentieth century deepened the pluralist tendencies of those

268

•

Chapter 13

areas, fomenting corruption and machine politics. More typically, however, immigration to a state is slower. Then newcomers tend to take on the political style that the founding groups already established. When people from other regions move to a moralistic state, they become more moralistic themselves. This effect is particularly strong for people from individualistic regions.22 In Wisconsin, in-migration of blacks and Hispanics after World War II did not change the state’s moralistic culture. Rather, the new minorities appear to have taken on that style themselves. Black leaders displayed a more earnest and self-effacing style in welfare politics than seen in some other states. Elazar does remark that the migration of blacks to Detroit brought a traditionalistic element into Michigan,23 but that state still reformed welfare in a decidedly moralistic manner. Both states responded to the racial challenge by reasserting their accustomed style, not by changing it. More generally, what is distinctive about the political identity of states is not clearly explicable from their demographics.24 In this century, the race or welfare problems did not shape culture. Rather, each culture reacted to the problems in a characteristic way. When national welfare policy began in the 1930s, most blacks still lived in the South, so the repressive reaction of the traditionalistic culture was the first that black welfare claimants encountered. Then after the war, when many blacks moved north to the big cities, they encountered the more individualist urban political style. Blacks were now treated as a voting block, to be subsidized like others through public programs. This was the politics that acquiesced in the massive expansion of entitlement welfare in the 1960s and 1970s. Finally, as the black migration reached further north, it encountered the pro-government but less tolerant style of moralistic states like Michigan, Oregon, and Wisconsin. These are the states that neither reject nor accept welfare, but rather try to use it to promote a more civic society. Integrating racial minorities is certainly the greatest domestic challenge American government has faced. Nicholas Lemann has called race “the chief generator of doubt” about the national enterprise. As blacks moved north, they brought with them some of the South’s “tragic sense” about how little government could do to solve social problems.25 But much of the story in welfare reform is that some states have acquiesced in that pessimism, and some have not. Wisconsin policymaking displays earnestness and optimism even in dealing with heavily nonwhite Milwaukee. In welfare, the states that effectively enforce work uphold the hope that the poor of all races can finally join the society as equals. They defend the vision, going back to the Puritans, that effective government can be the instrument of positive social change. As the national culture evolves toward moralism, the Wisconsin response to poverty should become more common, and prospects for

Implications

•

269

progress will improve. The main pressure the other way probably comes from immigration. Discussion of this issue usually focuses on the effects on the nation’s racial makeup, but the effects on politics are more important. The state political cultures maintain their character only if immigration is gradual. Heavy inflows can make government less civic. That is hardly surprising, as immigrants almost always come from countries less well governed than our own. While immigration can continue, some reduction of the current intake—especially from Latin America—would be prudent.

NATIONAL POLICY What might the nation do to accelerate current trends toward a more civic political culture? In the past, Washington has actively promoted better administration in connection with welfare. This shows that others perceive my theme: good government matters. Administrative Requirements The founders of national social programs realized that government quality would affect the achievement of their ends. The Social Security Act was first enacted, in 1935, in an era when public administration was more prestigious than today. So the Act mandated that states’ blueprints for family welfare (then called Aid to Dependent Children) had to provide for the “efficient operation” of the plan.26 Federal oversight of welfare was initially vested in the Social Security Board, although it was later shifted to other agencies. Then in 1939, Social Security Chairman Arthur Altmeyer—from Wisconsin—complained to Congress that the agency could not secure from states the capable administration it needed. It had been forced to withhold funds from Illinois, Ohio, and Oklahoma. The reason: state civil service standards were inadequate. Accordingly, Congress amended the law to demand that states also institute and maintain “personnel standards on a merit basis.” In 1946, in addition, Congress also raised funding for state administrative costs.27 In the 1960s, federal officials used these and other provisions to try to induce states to pay higher grants in AFDC and to determine eligibility and benefits more uniformly. They wanted to eliminate the “localism” and “moralism” that caused states often to deny aid to the unwed mothers who flooded onto the rolls in that era. That goal reflected the philosophy of entitlement that shaped national welfare policy until the 1980s. To those ends, Washington tried to force states like Massachusetts to elevate administration standards, chiefly by raising the educational standards re-

270

•

Chapter 13

quired of new welfare officials. In this they expressed the rationalistic, instrumental view of government characteristic of the moralistic style.28 After the 1960s, federal attention shifted to other concerns. By the late 1980s, federal and state politicians abandoned entitlement, and welfare reform thinking shifted toward work enforcement, the sea change that Wisconsin came to lead. Finally, with PRWORA, the very language about “efficient operation” and “merit” systems was deleted from welfare law. The Republican majority in Congress deemed these provisions too intrusive now that states were given increased control over welfare. Management Information Systems But in other ways, Washington continued to pressure the states toward improved public administration. It did this through generous funding of state administrative costs and through the development of management information systems. From 1973 onward, the federal government funded local MIS development generously in several programs—AFDC, Medicaid, Food Stamps, JOBS, and child support enforcement. Washington guided the design of state systems, and it reimbursed as much as 90 percent of state costs. The goal was to prevent ineligibility and fraud in the payment of benefits and also to give program managers the information they needed to manage. By 1989, thirty-seven states had some sort of federally funded information system in AFDC. In the 1990s, this structure encountered difficulty with sharing of data across programs. The vogue for performance measurement meant that managers now needed data on the results achieved by clients, such as employment, whereas established information systems dealt mainly with eligibility and benefits. Those systems funneled data to the top of bureaucratic hierarchies, while the trend was to decentralize management responsibility to lower levels. Current systems do not support case management well at the operating level. PRWORA demanded more reporting of data on clients, while it also ended dedicated federal funding for welfare MIS.29 Nevertheless, MIS development did force states to think more systematically about what programs were supposed to achieve, as against merely about their costs or benefits. That was a step toward a more moralistic style. Performance Standards Performance measurement moved states in this same direction, even if it complicated information systems. “Reinventing government” said that public management should focus on missions rather than bureaucratic

Implications

•

271

rules, on outcomes rather than inputs.30 The idea entered welfare reform in a major way with the Family Support Act of 1988. FSA set targets for participation rates in the new JOBS program. By 1995, states were supposed to have 20 percent of their mandatory recipients (those they judged employable) involved in the program, on pain of cuts in their federal grants. While that level now seems low, the new targets were more clearly defined and enforced than anything in earlier law. States—especially the large urban states—struggled to meet them. Under PRWORA, work targets were again raised. Now in principle, all adult recipients were obligated to work, not only those states judged to be employable, and half of them were supposed to be active in “work activities” by 2002. The standards had the weakness that states could deduct from their activity targets any percent by which their AFDC caseloads had declined after 1995. Since the caseload decline was unexpectedly great, the work targets were reduced to levels nearly all states could meet. But despite this, states became steadily more demanding about work. This again forced them to think of welfare as a means to an end— employment—rather than an end in itself.31 Again, they shifted toward a more moralistic conception of government. When PRWORA is reauthorized, in 2003 or 2004, work standards will certainly be raised further. This will probably force urban states like New York, which have clung to entitlement, at last to abandon it. They will be forced to approach welfare as a project, not a payoff, and that again is to adopt a more earnest and goal-oriented style. Publicity At every stage, welfare reform has drawn steadily more publicity. Under FSA and, even more, under PRWORA, the advance of work tests in welfare was often covered in the general media, not only in specialist publications. Recurrent attention was given to the positive evaluations of work programs and, since 1994, to the precipitous decline in the rolls. The great question was the effects of change on the poor, both on and off welfare. The glare of attention hit states that traditionally had ignored social policy, not only leaders like Wisconsin or Michigan. With the nation watching, these more backward states had to devote unwonted attention to welfare. Legislators and administrators labored over reform for years. In so doing, they had to upgrade their performance. Many states participated in evaluations of their early waiver programs, not only those that had been innovators. Public and expert reaction to these studies forced state governments to think through what they were doing with welfare more fully than before. Florida, for example, is a state that has had considerable difficulty with welfare reform. Yet the state

272

•

Chapter 13

mounted two of the more important evaluations of welfare work experiments in the 1990s. The first of these was of a JOBS program that showed slim results, in part due to inadequate child care funding. The other was one of the nation’s first experiments in time-limited welfare. It demonstrated that a program with tough work and participation standards might drive all its cases off the rolls well in advance of the time limit.32 The explication of these results gave Florida a kind of exposure at conferences and in the media that was new.33 A second southern state that was ill prepared for reform was South Carolina. Here the bureaucratic infrastructure for serious work enforcement barely existed. Nevertheless, the state mounted one of the first and best of the many state leaver surveys that tracked the progress of recipients after they left the rolls.34 The southern states thus started to influence the national welfare debate. And they in turn were influenced by it. They were driven toward a more rationalistic, problem-solving style. The combination of federal standards and ceaseless publicity will probably upgrade the performance of the more backward states in social policy over time. While few will become leaders, the gap between them and exemplars like Wisconsin should narrow. Thus, ironically, the process of welfare reform may itself help to generate the governmental capacity that it needs. THE PROBLEM OF CONSENT Besides meeting these institutional demands, the other major challenge faced by welfare reform is ongoing dissent from the very idea of conditioning aid on good behavior. That principle is popular with voters, and recent work-based welfare programs have unquestionably succeeded. Yet many leaders, in and out of politics, continue to question the enforcement of work as a condition of aid. The federal government, and many urban states, have not achieved the concordat about reform seen in Wisconsin. Like a bad conscience, the resistance may deny the new policies full acceptance and thus undermine their achievements. The Right Some conservatives would still prefer to abolish aid for those of working age rather than condition it on employment. One argument memorably made by Charles Murray, is that welfare creates perverse incentives for the poor not to work or marry. Another contention, made by cultural conservatives, is that welfare promotes a broader moral breakdown. The right sees illegitimacy rather than nonwork as the great evil in American

Implications

•

273

life. If welfare is kept, they want to use it to promote abstinence and marriage, perhaps by turning the aid function over to conservative community groups, including churches.35 Such arguments influence the national Republican party. They helped to shape PRWORA, and the reauthorization of the act may provide more funding for programs to promote marriage. On this view, the Achilles’ heel of the Wisconsin reform was that it concentrated on employment, downplaying the more divisive family issue. “The essential defect of the Kenoshas,” Marvin Olasky wrote, “is that no bureaucracy, and no amount of money, can buy the reformation of morals that is desperately needed.” A renaissance of the poor can come only from people who believe in “Biblical concepts of challenge, personal involvement, and spiritual commitment.”36 Unfortunately, such views are currently impolitic and impractical. Most Americans are dismayed by the marriage problem rather than angered. They find the decay of the family distressing, but they are unready to enforce marriage in the same way as work.37 Nor has government developed programs that show clear impact on marriage or unwed pregnancy. Illegitimacy rates have fallen in the last decade, but this seems principally due to the rhetoric of personal responsibility surrounding welfare reform.38 To the extent we can promote marriage, in other words, it is through work. Before a marriage agenda can be taken more seriously, conservatives must sell it politically and develop policies that can credibly advance it. Conservatives sometimes suggest that simply by intoning timeless American values, they can cause the society to revert to the rigid ideas of personal probity that it held a century ago. That ignores the many Americans who have genuine difficulty living by those norms. For them, some effort to enforce and facilitate virtue is unavoidable. The best way back to a more wholesome society is through government, not around it. The Left Weightier opposition to welfare reform comes from the left. Liberal politicians and community groups strenuously opposed the stiffening of work tests in the last decade. Many people in the universities, the national media, foundations, and other nonprofit bodies feel the same, although they are less demonstrative. This opinion-making elite cares deeply about poverty. To them, the very existence of need in America is an indictment of our society, as is the apparent severity of welfare reform. These forces obtained nothing like the hearing in Washington that they did in Wisconsin. They were simply steamrollered by the Republican majority that took over Congress in 1994. PRWORA thus has much less

274

•

Chapter 13

bipartisan support behind it than W-2 enjoys.39 The resentments have already complicated the reauthorization of TANF. A third of House Democrats endorsed a plan that would undo much of the work requirements of 1996, and activist groups have demonstrated against Bush Administration proposals to stiffen work standards. In the Senate, Democrats demand easier work standards and more child care money than the administration wants.40 The disputes have delayed the legislation. Most academic poverty experts differ from government about how confining the social conditions faced by poor adults are, and how much they can progress on their own. Like the IRP experts who opposed Tommy Thompson, academics tend to be pessimistic about the environment and the poor themselves. They judge that steady work is impractical for most poor adults, whom they see as overwhelmed by an indifferent society. Most policymakers are more optimistic; they think that some opportunities exist for the low-skilled and that dependent adults can at least work steadily in some job.41 The actual outcome of welfare reform vindicated government against the critics. Although poverty is correlated with myriad social barriers, these did not prevent a 60 percent reduction of the welfare rolls when work was enforced. Part of that success was indeed due to an easing of constraints—a good economy and new benefits. But the new demands to work also caused vastly more nonworking adults to take jobs than any expert anticipated. In the policy arena, it was an embarrassing defeat for academia, comparable to the failure of the Kremlinologists to predict the fall of the Soviet Union. Justice Even if opponents agreed that work is possible, however, they would still question reform, due to deep political differences. Critics say that reform has so far done much more to depress dependency than to overcome poverty, and that is true. More families are working, but most are not much better off than before, and many remain poor. Nor has reform stemmed long-term trends toward greater inequality in America. These voices want poverty reduction, not only higher work levels, to become an explicit goal of welfare in the renewal of PRWORA.42 Yet reform ultimately does serve the cause of justice. In America, only the “deserving” can claim greater equality, and for the employable that requires a work history. The public may reject traditional welfare, but helping the working poor is popular. Although welfare reform was the doing mainly of conservatives, its political effects favor liberals. The more the rolls have declined since 1994, the more social politics has shifted to the left.43 As in Wisconsin, enforcing work has sharply reduced the welfare

Implications

•

275

rolls, but the welfare state has expanded in other senses. Already, the nation is spending more on child care and the EITC than it ever spent on AFDC. Some “civic” liberals have supported work-enforcing reform precisely to justify an expansion of government. Rather than feel defeated by the success of reform, liberals should seize this occasion to press their own agenda. Many problems of the inner city remain unsolved. New programs could be justified, for example, targeted at housing or at moving the poor to the suburbs.44 The working poor are much easier to help than the nonworking, so programs aimed at them could well succeed. Due to welfare reform, the United States is potentially closer to social democracy today than it was a decade ago. Entitlement If this logic fails to persuade many on the left, the reason ultimately is that work-enforcing welfare overthrows entitlement. However generous a Wisconsin may be to the poor, the benefits it now gives to families are mostly conditioned on employment. That violates the ideal that many still see in European welfare states, where aid is given as a right without any stipulations about good behavior.45 The defenders of entitlement today speak less of justice in the old socialist sense than they do of the family. Just as parents succor children in need, so government should support needy families without thought of reciprocity. The family member is one whose bond with others is intimate and unconditional. The ideal behind welfare reform, however, is the very different ideal of citizenship. The citizen is one who claims the same rights and obligations as others, in a public order outside the family. In that world, citizens may still make claims for support, but only if they first fulfill the common obligations. In the family, aid is given based on vulnerability, whereas in the polity it is given based on desert, on having first earned the respect of others.46 To the intellectual left, the family ideal serves a larger vision that the purpose of public life is an ever-expanding acceptance. Just as government has extended civil liberties, then political rights, then social rights to its citizens, so social policy should also free people from the trammels of private life. All the wounds and embarrassments of which people are ashamed should finally become known and accepted. Just as society has learned to extend recognition to women, to blacks, and—increasingly— to gays, so it should accept people who are different in whatever way. The dysfunctional, even the criminal are to be understood in the way they see themselves, and to understand is to forgive. All are to become persons worthy of what Ronald Dworkin calls “equal concern and respect.”47 Public life must finally embrace the full weight of the human condition.

276

•

Chapter 13

The ultimate goal here is not freedom or equality in any traditional sense, for that would be too impersonal. It is rather of a universal solicitude, expressed in emotional and not economic terms.48 The vision is alluring, yet it is a recipe for social breakdown. Unjudging solicitude may work in the family because some reciprocity is enforced among the members informally, due to close personal relationships. But citizens in the wider society have no such ties, so demands for reciprocity must be explicit. If social policy becomes all-accepting, society cannot defend obligations bearing on individuals. People can always find reasons for slighting social norms. Personal responsibility then dissolves and dependency grows, threatening equal citizenship. Government cannot foster intimate authenticity, but only the less authentic good behavior that characterized most people in Wisconsin. On all these deep grounds, dissent from welfare reform is likely to endure. This will deny current policies full legitimacy, and thus weaken them. One can only hope that those measures will in the end persuade the critics by their very success. The welfare state cannot be a family, yet it can promote stronger families on and off welfare. Families where adults have to work and children have to learn in school will, on average, be stronger than otherwise. Families will more likely have competent parents, and have two of them. They will be more capable of solicitude. There, in private life, is ultimately where solicitude must be found.

THE INSTITUTIONAL FRONTIER Social solutions finally depend on statecraft. This is the innermost meaning of Wisconsin. The Badger State demonstrates in miniature that governance—that complex combination of demands and supports—offers the nation’s best hope against poverty. We usually think of social programs as built out of economic things—incentives, budgets, benefits. But governmental capacity is so important that policymakers should focus even more on politics and administration. It is the institutions, more than any specific policies, that in fact achieve and embody the public interest.49 The nation’s greatest resource against poverty is not its wealth but the civility that has undergirded our politics since before the Founding. Inherited from the British, we scarcely perceive these virtues. Our culture is so self-critical that we easily imagine ourselves ill governed, rather than what we are—one of the world’s strongest polities. The authority of American government is unassailable. The chief threat to it does not come from principled opposition within our borders, nor terrorism without, but rather from the secession of the poor from mainstream institutions—the schools, the workplace, and the criminal law.

Implications

•

277

In crafting a response, we look to the innermost strength of the society, which is its politics. The long arm of that ancient virtue reaches out of the past to empower government against the social problem. In exemplars like Wisconsin and Michigan, it reaches out from the Midwest to impose, via national policy, a more demanding stance on the cities of the coasts, which are less civic. It also reaches out from ordinary citizens in every region, what Samuel Beer called the “great hinterland of common belief.”50 Elite culture is too freedom-oriented to restore order, but the voters finally insist that civility be upheld. Most of them are willing to pay the price, in bureaucracy as well as benefits, that this entails. Ultimately, the politicians speak for them, and welfare is being rebuilt around work. By drawing on civic values, the welfare state has returned to its origins—as a state. A century ago, the theorists of Progressivism traced the nation’s reluctance to regulate capitalism to its libertarian past. Jefferson had trumped Hamilton. The country was too much in love with freedom. It refused to recognize that to expand opportunity in some forms required public controls on the economy. So government regulation of the private marketplace began.51 In the New Deal and Great Society eras, however, this activity was less visible than the growth of national benefit programs, including welfare. Those programs developed an ethos of entitlement. In the case of welfare, that fomented dependency and other social problems. So now, through work and other conduct requirements, we also regulate recipients. It is a return to an earlier tradition that tried to use government to promote good citizenship.52 Work enforcement is far from a total solution to poverty, but it has promoted integration as nothing else in antipoverty policy has done. Around employment, institutions could be built that demand something meaningful of the poor, and then reward them for it. The key was that the work goal was achievable by many, and also esteemed by the society. By enforcing and supporting it, society helped to integrate the marginal. Efforts to enforce child support or learning in the schools are less advanced, but the goal is the same. When more fathers support their children, and when the children get through school, they achieve a belonging that reaches beyond the economic. The danger is that the will to enforce could falter. Flushed with its recent successes, welfare might revert to a spirit of entitlement. This would be a loss for the country, and the regime. That way lies a caste society, or a deeply divided one. The way forward is toward citizenship.

Notes

CHAPTER 1 INTRODUCTION 1. Comparing AFDC at its height to W-2, the drop is around 90 percent, but this somewhat overstates the state’s decline. When W-2 was implemented in 1997–98, the state diverted some cases from AFDC to other cash aid programs. See chapters 6 and 10. For lack of data, the tables and figures in this chapter consider only AFDC and W-2. 2. Lawrence Mead, “Welfare Reform: Meaning and Effects,” Policy Currents 11, no. 2 (Summer 2001): 7–13. 3. The DeParle series began with “Cutting Welfare Rolls but Raising Questions,” New York Times, May 7, 1997, pp. A1, B12, and ended with “Bold Effort Leaves Much Unchanged for the Poor,” New York Times, December 30, 1999, pp. A1, A20–21. Examples of local coverage include Lynda McDonnell, “Wisconsin Welfare Experiment Aims to Work,” St. Paul Pioneer Press (St. Paul, MN), February 5, 1995, and Gardiner Harris, “A New Welfare Command: Get a Job, Any Job,” Courier-Journal (Louisville, KY), August 11, 1996. 4. Institute for Wisconsin’s Future, Overview of Current Research on Wisconsin Works (W-2), vol. 2 (Milwaukee: Institute for Wisconsin’s Future, July 1999), lists twenty-five organizations with studies under way in the state, and this does not include myriad other studies by individual scholars. 5. Donald P. Green and Ian Shapiro, Pathologies of Rational Choice Theory: A Critique of Applications in Political Science (New Haven: Yale University Press, 1994). 6. H. George Frederickson, “The Repositioning of American Public Administration,” PS 32, no. 4 (December 1999): 701–11. 7. Typical compendia of this style of poverty research include Phoebe H. Cottingham and David T. Ellwood, eds., Welfare Policy for the 1990s (Cambridge, MA: Harvard University Press, 1989); Christopher Jencks and Paul E. Peterson, eds., The Urban Underclass (Washington, DC: Brookings, 1991); and Sheldon H. Danziger and Robert Haveman, eds., Understanding Poverty (New York: Russell Sage Foundation, and Cambridge, MA: Harvard University Press, 2001). For a fuller critique, see Lawrence M. Mead, “Comment on Mark R. Rank and Thomas A. Hirschl, ‘Rags or Riches,’ ” Social Science Quarterly 82, no. 4 (December 2001): 670–75. 8. Lawrence M. Mead, The New Politics of Poverty: The Nonworking Poor in America (New York: Basic Books, 1992), chaps. 4–7. 9. Richard P. Nathan and Thomas L. Gais, Implementing the Personal Responsibility Act of 1996: A First Look (Albany: State University of New York, Rockefeller Institute of Government, October 1999); Thomas Gais, “Welfare Reform Findings in Brief” (Albany: State University of New York, Rockefeller Insititute of Government, March 1, 2002).

280

•

Notes to Chapter 1

10. Theodore R. Marmor, The Politics of Medicare (Chicago: Aldine, 1975); Theda Skocpol, Protecting Soldiers and Mothers: The Political Origins of Social Policy in the United States (Cambridge, MA: Harvard University Press, 1992); Jane J. Mansbridge, Why We Lost the ERA (Chicago: University of Chicago Press, 1986). 11. John W. Kingdon, Agendas, Alternatives, and Public Policy, 2nd ed. (New York: HarperCollins, 1995); Frank R. Baumgartner and Bryan D. Jones, Agendas and Instability in American Politics (Chicago: University of Chicago Press, 1993). 12. Jack L. Walker, “The Diffusion of Innovations Among the American States,” American Political Science Review 63, no. 3 (September 1969): 980–99; Virginia Gray, “Innovation in the States: A Diffusion Study,” American Political Science Review 67, no. 4 (December 1973): 1174–85. 13. Jeffrey L. Pressman and Aaron Wildavsky, Implementation, 3rd ed. (Berkeley: University of California Press, 1984); Daniel A. Mazmanian and Paul A. Sabatier, Implementation and Public Policy (Glenview, IL: Scotts, Foresman, 1983); Malcolm Goggin, Ann O’M. Bowman, James P. Lester, and Laurence J. O’Toole, Jr., Implementation Theory and Practice: Toward a Third Generation (New York: HarperCollins, 1990). 14. Kim Quaile Hill, “In Search of Policy Theory,” Policy Currents 7, no. 1 (April 1997): 1–9; Paul A. Sabatier, ed., Theories of the Policy Process (Boulder, CO: Westview, 1999). 15. Lawrence M. Mead, “Policy Studies and Political Science,” Policy Studies Review 5, no. 2 (November 1985): 319–35; idem, “Public Policy: Vision, Potential, Limits,” Policy Currents 5, February 1995: 1–4. 16. Lawrence M. Mead, Beyond Entitlement: The Social Obligations of Citizenship (New York: Free Press, 1986); idem, New Politics of Poverty. 17. Martin Gilens, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy (Chicago: University of Chicago Press, 1999); Fay Lomax Cook and Edith J. Barrett, Support for the American Welfare State: The Views of Congress and the Public (New York: Columbia University Press, 1992); Steven M. Teles, Whose Welfare? AFDC and Elite Politics (Lawrence, KS: University Press of Kansas, 1996); R. Kent Weaver, Ending Welfare As We Know It (Washington, DC: Brookings Institution, 2000); Gary Bryner, Politics and Public Morality: The Great American Welfare Reform Debate (New York: Norton, 1998). 18. Robert D. Behn, Leadership Counts: Lessons for Public Managers from the Massachusetts Welfare, Training, and Employment Program (Cambridge, MA: Harvard University Press, 1991); Donald F. Norris, and Lyke Thompson, eds., The Politics of Welfare Reform, (Thousand Oaks, CA: Sage, 1995); M. Anne Hill and Thomas J. Main, Is Welfare Working? The Massachusetts Reforms Three Years Later (Boston: Pioneer Institute, 1998); Sarah F. Liebschutz, ed., Managing Welfare Reform in Five States: The Challenge of Devolution (Albany: Rockefeller Institute Press, 2000); and Carol S. Weissert, ed., Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States (Albany, NY: Rockefeller Institute Press, 2000). 19. Michael Wiseman, “State Strategies for Welfare Reform: The Wisconsin Story,” Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 515–

Notes to Chapter 1

•

281

46; Paul E. Peterson and Mark C. Rom, Welfare Magnets: A New Case for a National Standard (Washington, DC: Brookings, 1990), chap. 2; Thomas J. Corbett, “Welfare Reform in Wisconsin: The Rhetoric and the Reality,” in The Politics of Welfare Reform, ed. Norris and Thompson, chap. 2; Thomas Kaplan, “Wisconsin’s W-2 Program: Welfare As We Might Come to Know It?” in Learning from Leaders, ed. Weissert, chap. 4; idem, “Wisconsin Works,” in Managing Welfare Reform in Five States, ed. Liebschutz, chap. 6; Cyril Ignatius Kendrick, “The Roots of Welfare Reform: The Social Forces Underlying the Wisconsin Learnfare Program,” Ph.D. diss. (Blacksburg, VA: Virginia Polytechnic Institute, Department of Sociology, March 1993). 20. I checked my findings about the politics of reform against those of Michael Heaney, a graduate student at the University of Chicago, who interviewed more widely. His conclusions are similar. 21. I sat on the Steering and National Technical Advisory Committees of the Management and Evaluation Project, a foundation-funded body that oversaw research on the implementation and effects of Wisconsin Works. See chapter 7. 22. Mead, Beyond Entitlement, chaps. 3–4; Mead, New Politics of Poverty, chaps. 3–6. 23. Gilens, Why Americans Hate Welfare, chaps. 2, 8; Thomas Brock, David Butler, and David Long, “Unpaid Work Experience for Welfare Recipients: Findings and Lessons from MDRC Research” (New York: Manpower Demonstration Research Corporation, September 1993). 24. My approach is indebted to Gary King, Robert D. Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton: Princeton University Press, 1994). 25. The model oversimplifies in that each factor is linked directly only to the next one in the causal chain. I do think these were the major relationships, but one may well posit that governmental excellence also affects successful welfare reform, or that Wisconsin’s background might affect both welfare reform and its outcomes. 26. In two early papers completed in 1995–96: Lawrence M. Mead, “The New Paternalism in Action: Welfare Reform in Wisconsin” (Milwaukee: Wisconsin Policy Research Institute, January 1995); idem, “The Decline of Welfare in Wisconsin” (Milwaukee: Wisconsin Policy Research Institute, March 1996). 27. Teles, Whose Welfare; Daniel P. Moynihan, The Politics of a Guaranteed Income: The Nixon Administration and the Family Assistance Plan (New York: Random House, 1973); Norris and Thompson, ed., Politics of Welfare Reform. 28. John J. Mitchell , Mark L. Chadwin, and Demetra S. Nightingale, Implementing Welfare-Employment Programs: An Institutional Analysis of the Work Incentive (WIN) Program, R&D Monograph 78, U.S. Department of Labor (Washington, DC: U.S. Government Printing Office, 1980); Jan L. Hagen and Irene Lurie, Implementing JOBS: Progress and Promise (Albany: State University of New York at Albany, Rockefeller Institute of Government, August 1994); Marcia K. Meyers, Bonnie Glaser, and Karin Mac Donald, “On the Front Lines of Welfare Delivery: Are Workers Implementing Policy Reforms?” Journal of Policy Analysis and Management 17, no. 1 (Winter 1998): 1–22.

282

•

Notes to Chapter 2

29. Lawrence M. Mead, “Expectations and Welfare Work: WIN in New York State,” Polity 18, no. 2 (Winter 1985): 224–52; idem, “The Potential for Work Enforcement: A Study of WIN,” Journal of Policy Analysis and Management 7, no. 2 (Winter 1988): 264–88. 30. Richard P. Nathan, “Setting Research Priorities in the Era of Devolution,” in W-2 Research Assessment and Direction Conference, Wingspread, Racine, WI, December 10, 1999, ed. Shannon Christian and Rebecca Swartz (Indianapolis: Hudson Institute, June 2000), p. 45. 31. These other states include California, Illinois, Maryland, Massachusetts, Missouri, New Jersey, New York, Pennsylvania, Rhode Island, South Carolina, and West Virginia. CHAPTER 2 THE DEVELOPMENT OF REFORM 1. Michael C. Barth, George J. Carcagno, and John L. Palmer, Toward an Effective Income Support System: Problems, Prospects, and Choices (Madison, WI: University of Wisconsin, Institute for Research on Poverty, 1974); David T. Ellwood, Poor Support: Poverty in the American Family (New York: Basic Books, 1988). 2. Martin Anderson, Welfare: The Political Economy of Welfare Reform in the United States (Stanford, Calif.: Hoover Institution Press, 1978); Charles Murray, Losing Ground: American Social Policy, 1950–1980 (New York: Basic Books, 1984); Marvin Olasky, The Tragedy of American Compassion (Wheaton, IL: Crossway Books, 1992). 3. Robert Moffitt, “Incentive Effects of the U.S. Welfare System: A Review,” Journal of Economic Literature 30, no. 1 (March 1992): 15–19. 4. Ellwood, Poor Support, chap. 5. 5. Lawrence M. Mead, “Welfare Employment,” in The New Paternalism: Supervisory Approaches to Poverty, ed. Lawrence M. Mead (Washington, DC: Brookings, 1997), chap. 2. 6. John H. Fenton, Midwest Politics (New York: Holt, Rinehart, and Winston, 1966), pp. 71–72; William F. Thompson, The History of Wisconsin: Continuity and Change, 1940–1965 (Madison: State Historical Society of Wisconsin, 1988), p. 701. But to judge from table 2.1, Wisconsin’s generosity to AFDC declined by 1970, only to rise thereafter. 7. Wisconsin Expenditure Commission, “Welfare Spending” (Madison: Wisconsin Expenditure Commission, August 18, 1986), p. 3. 8. Ibid., p. 3; Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1991), p. 11. 9. Calculated from table 2.1 and its sources. 10. Wisconsin Expenditure Commission, “Welfare Spending,” pp. 9–10. 11. Ibid., p. 8. 12. Ibid., pp. 1–8. 13. Michael Wiseman, “State Strategies for Welfare Reform: The Wisconsin Story,” Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 518.

Notes to Chapter 2

•

283

14. Thomas J. Corbett, “Welfare Reform in Wisconsin: The Rhetoric and the Reality,” in The Politics of Welfare Reform, ed. Donald F. Norris and Lyke Thompson (Thousand Oaks, CA: Sage, 1995), pp. 31–32. 15. This state credit was first enacted in 1984, repealed in 1986, then reenacted in 1989, according to Johannes M. Bos, Aletha C. Huston, Robert C. Granger, Greg J. Duncan, Thomas W. Brock, and Vonnie C. McLoyd, New Hope for People with Low Incomes: Two-Year Results of a Program to Reduce Poverty and Reform Welfare (New York: Manpower Demonstration Research Corporation, August 1999), pp. 13, 13 n. 22. According to V. Joseph Hotz and John Karl Scholz, “The Earned Income Tax Credit” (Cambridge: National Bureau of Economic Research, January 2001), pp. 10–11, among the fourteen states and the District of Columbia that offer a state EITC, Wisconsin is unusually generous. 16. Corbett, “Welfare Reform in Wisconsin,” pp. 23, 33–34; Irwin Garfinkel, Assuring Child Support: An Extension of Social Security (New York: Russell Sage Foundation, 1992), chaps. 3, 5. 17. Paul E. Peterson and Mark C. Rom, Welfare Magnets: A New Case for a National Standard (Washington, DC: Brookings, 1990), chap. 3. According to Corbett, “Welfare Reform in Wisconsin,” p. 34 n. 12, the migration issue first surfaced in the early 1970s. 18. Peterson and Rom, Welfare Magnets, pp. 26–32, 42–44; Wisconsin Expenditure Commission, Report of the Welfare Magnet Study Committee (Madison: Wisconsin Expenditure Commission, December 1986). Later studies were also equivocal. For a review, see David E. Umhoefer and Celeste Williams, “Data, Emotions Cloud ‘Welfare Magnet’ Issue,” Milwaukee Journal, January 5, 1992, part A, p. 1. 19. Legislative Fiscal Bureau, 1985–87 Wisconsin State Budget (Madison: Legislative Fiscal Bureau, October 1985), vol. 1, p. 387. 20. John Pawasarat and Lois M. Quinn, “Wisconsin Welfare Employment Experiments: An Evaluation of the WEJT and CWEP Programs” (Milwaukee: University of Wisconsin–Milwaukee, Employment and Training Institute, September 1993), pp. 8–19. 21. Peterson and Rom, Welfare Magnets, pp. 32–38; Corbett, “Welfare Reform in Wisconsin,” p. 36. 22. For details of how he did this, see Wiseman, “State Strategies,” p. 523. 23. Under this reform, counties except Milwaukee receive a limited allocation from the state for GR cash benefits, provided they assure health benefits for the recipients. In Milwaukee, the state funds only GR health benefits. Both state and county funding for this purpose are channeled through the Medicaid program in order to claim a federal match. See Legislative Fiscal Bureau, General Assistance Block Grant Program, Informational Paper #46 (Madison: Legislative Fiscal Bureau, January 2001). 24. Pawasarat and Quinn, “Wisconsin Welfare Employment Experiments,” pp. 22–35, 85–93. 25. The authority to do this appears in section 1115 of the Social Security Act, in which the law for AFDC/TANF occupies Title IVA.

284

•

Notes to Chapter 2

26. Steven M. Teles, Whose Welfare? AFDC and Elite Politics (Lawrence, KS: University Press of Kansas, 1996), chap. 7; Michael Wiseman, “Welfare Reform in the States: The Bush Legacy,” Focus 15, no. 1 (Spring 1993): 18–36. 27. Wiseman, “State Strategies,” pp. 519–20; waiver data from the U.S. Administration for Children and Families. 28. Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1995), p. 13. 29. Related waivers extended Medicaid coverage to pregnant women and preschool children up to 185 percent of the poverty line (known as Healthy Start) and required that Medicaid recipients enroll in health maintenance organizations. These were early steps toward BadgerCare, the low-income health program that later became part of W-2 (see chapter 6). 30. Wiseman, “State Strategies,” pp. 524–25, 524 n. 6. 31. Shapiro v. Thompson, 394 U.S. 618 (1969). 32. John W. Kole, “Thompson’s dissent tries to make a point among colleagues,” Milwaukee Journal, March 1, 1987, Part J, p. 3; Legislative Fiscal Bureau, Welfare Reform Initiatives, Informational Paper #47 (Madison: Legislative Fiscal Bureau, January 1995), pp. 17–19. 33. The following discussion draws on Legislative Fiscal Bureau, Welfare Reform Initiatives, pp. 13–17, and other budget documents. 34. The following discussion draws on Legislative Fiscal Bureau, Welfare Reform Initiatives, pp. 19–29, and other budget documents. 35. Governor Tommy Thompson and Dr. William Bennett, “The Good News about Welfare Reform: Wisconsin’s Success Story” (Washington, DC: Heritage Foundation, March 6, 1997), p. 6. 36. Legislative Fiscal Bureau, Aid to Families with Dependent Children, January 1991, p. 29; State of Wisconsin, Wisconsin Welfare Reform (Madison: State of Wisconsin, April 1996), p. 11. 37. State of Wisconsin, Wisconsin Welfare Reform, pp. 19–20; Legislative Fiscal Bureau, “1995 Assembly Bill 21: AFDC Benefit Cap, Self-Sufficiency First and Pay-After-Performance” (Madison: Legislative Fiscal Bureau, January 25, 1995), pp. 3–5. 38. Tommy G. Thompson, Power to the People: An American State at Work (New York: HarperCollins, 1996), pp. 42, 57–60, 64–67; Jason Turner, remarks during “Welfare and Employment: Policy Making in Michigan and Wisconsin— A Roundtable,” annual conference of the Association for Public Policy Analysis and Management (APPAM), Pittsburgh, PA, November 1, 1996. Another account of Learnfare’s origin is that the idea came from a book that suggested that prison inmates be required to learn to read; Rich Chandler, a Thompson aide, spotted it and sold it to Thompson; see Cyril Ignatius Kendrick, “The Roots of Welfare Reform: The Social Forces Underlying the Wisconsin Learnfare Program,” Ph.D. diss. (Blacksburg, VA: Virginia Polytechnic Institute, Department of Sociology, March 1993), pp. 176–77. For the savings idea, see Michael Sherraden, Assets and the Poor: A New American Welfare Policy (Armonk, NY: M. E. Sharpe, 1991). 39. Thompson, Power to the People, pp. 39–40, 71–72; Jason Turner, remarks during APPAM; interview with Jason Turner, Madison, WI, November 14, 1996. The Chicago Program was Project Match; see Toby Herr, Robert Halpern, and

Notes to Chapter 2

•

285

Aime´e Conrad, Changing What Counts: Rethinking the Journey Out of Welfare (Evanston, IL: Northwestern University, Center for Urban Affairs and Policy Research, April 1991). 40. Laurence E. Lynn, Jr., ed., “Symposium: The Craft of Public Management,” Journal of Policy Analysis and Management 8, no. 2 (Spring 1989): 284– 306. 41. According to Turner, remarks at APPAM conference, the turnover research was by LaDonna A. Pavetti. It is summarized in Pavetti, “The Dynamics of Welfare and Work: Exploring the Process by Which Young Women Work Their Way Off Welfare” (Cambridge: Harvard University, Kennedy School of Government, October 1992). The evaluation report was on New Chance, an intensive but voluntary remediation program for poor single mothers. Its results, by most accounts, were disappointing. See Janet C. Quint, Johannes M. Bos, and Denise F. Polit, New Chance: Final Report on a Comprehensive Program for Young Mothers in Poverty and Their Children (New York: Manpower Demonstration Research Corporation, October 1997). 42. Thomas L. Gais and Richard P. Nathan, “Learning, Emulation, and Adaptation in the American States: The Case of Welfare Reform” (Albany: State University of New York, Rockefeller Institute of Government, November 6, 1999); David Greenberg, Marvin Mandell, and Matthew Onstott, “The Dissemination and Utilization of Welfare-to-Work Experiments in State Policymaking,” Journal of Policy Analysis and Management 19, no. 3 (Summer 2000): 367–82; Greg M. Shaw, “The Role of Public Input in State Welfare Policymaking,” Policy Studies Journal 28, no. 4 (2000): 707–20. 43. This occurred with Learnfare (see Kendrick, “Roots of Welfare Reform,” pp. 225–30, 233), and Work Not Welfare (see Thompson, Power to the People, pp. 68–70). 44. In the experimental assessment of Children First in Racine, the program mistakenly served many members of the control group, thus making its impact indeterminate (see chapter 8). In the 100-Hour Rule Demonstration, the experimental and control groups were incorrectly defined, putting the validity of the results in question (see Wiseman, “State Strategies,” p. 527). With Self-Sufficiency First, random assignment of clients to the program and the control group was compromised due to computing and procedural errors (personal communication from Thomas Kaplan, June 10, 2002). 45. Turner, remarks at APPAM conference. 46. This was the task of the Management and Evaluation Project, described further in chapter 7. 47. Wiseman, “State Strategies,” pp. 526, 528; Legislative Fiscal Bureau, Welfare Reform Initiatives, pp. 17, 19. 48. Aaron Wildavsky, Speaking Truth to Power: The Art and Craft of Policy Analysis (Boston: Little, Brown, 1979). 49. Thomas Kaplan, “Wisconsin’s W-2 Program: Welfare As We Might Come to Know It?” in Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States, ed. Carol S. Reassert (Albany: Rockefeller Institute Press, 2000), p. 87.

286

•

Notes to Chapter 3

50. Charles E. Lindblom., “The Science of ’Muddling Through,’ ” Public Administration Review 19, no. 2 (Spring 1959): 79–88; Herbert A. Simon, Administrative Behavior: A Study of Decision-Making Processes in Administrative Organization, 3rd ed. (New York: Free Press, 1976). 51. Lawrence M. Mead, The New Politics of Poverty: The Nonworking Poor in America (New York: Basic Books, 1992), chap. 8; Henry J. Aaron, Politics and the Professors: The Great Society in Perspective (Washington, DC: Brookings, 1978). 52. Barbara Goldman, Judith Gueron, Joseph Ball, and Marilyn Price, Preliminary Findings from the San Diego Job Search and Work Experience Demonstration (New York: Manpower Demonstration Research Corporation, February 1984); Barbara Goldman, Daniel Friedlander, and David Long, Final Report on the San Diego Job Search and Work Experience Demonstration (New York: Manpower Demonstration Research Corporation, February 1986). 53. Lynn, ed., “ Craft of Public Management”; Judith M. Gueron and Edward Pauly, with Cameran M. Lougy, From Welfare to Work (New York: Russell Sage Foundation, 1991). 54. Gayle Hamilton and Daniel Friedlander, Final Report on the Saturation Work Initiative Model in San Diego (New York: Manpower Demonstration Research Corporation, November 1989). 55. James Riccio, Daniel Friedlander, and Stephen Freedman, GAIN: Benefits, Costs, and Three-Year Impacts of a Welfare-to-Work Program (New York: Manpower Demonstration Research Corporation, September 1994; Stephen Freedman and Daniel Friedlander, The JOBS Evaluation: Early Findings on Program Impacts in Three Sites (Washington, DC: U.S. Department of Health and Human Services and U.S. Department of Education, September 1995). 56. John W. Kingdon, Agendas, Alternatives, and Public Policies, 2nd ed. (New York: HarperCollins, 1995). 57. Albert O. Hirschman, “The Principle of the Hiding Hand,” The Public Interest, no. 6 (Winter 1967): 10–23. 58. Wiseman, “State Strategies,” p. 528. CHAPTER 3 THE POLITICS OF REFORM 1. This chapter is largely based on Lawrence M. Mead, “The Politics of Welfare Reform in Wisconsin,” Polity 32, no. 4 (Summer 2000): 533–59. I acknowledge valuable comments on earlier drafts from Robin Brady, Michael Heaney, and Ted Lascher. 2. Lawrence M. Mead, Beyond Entitlement: The Social Obligations of Citizenship (New York: Free Press, 1986), chap. 5; Joseph A. Califano, Jr., Governing America: An Insider’s Report from the White House and Cabinet (New York: Simon and Schuster, 1981), p. 321. 3. R. Kent Weaver, Ending Welfare As We Know It (Washington, DC: Brookings Institution, 2000). 4. Donald F. Norris and Lyke Thompson, eds., The Politics of Welfare Reform (Thousand Oaks, CA: Sage, 1995).

Notes to Chapter 3

•

287

5. U.S. Congress, House, Committee on Ways and Means, Overview of Entitlement Programs: 1990 Green Book: Background Material and Data on Programs Within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, June 5, 1990), pp. 588–90. 6. U.S. Congress, House, Committee on Ways and Means, 2000 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, October 6, 2000), pp. 439–40. 7. Department of Health and Social Services, JOBS Annual Report (Madison: Department of Health and Social Services, December 1994), p. 46. 8. Mark R. Rank and Thomas A. Hirschl, “A Rural-Urban Comparison of Welfare Exits: The Importance of Population Density,” Rural Sociology 53, no. 2 (1988): 190–206, and Mark Robert Rank, Living on the Edge: The Realities of Welfare in America (New York: Columbia University Press, 1994). 9. Julie Boatright Wilson, “Milwaukee: Industrial Metropolis on the Lake” (Cambridge, MA: Harvard University, Kennedy School of Government, April 1995); Paul A. Jargowsky, Poverty and Place: Ghettos, Barrios, and the American City (New York: Russell Sage, 1996), pp. 49–57. 10. Jargowsky, Poverty and Place, pp. 81–82. 11. The following discussion is based on Gordon S. Black, “The Wisconsin Citizen Survey” (Milwaukee and Thiensville, WI: Wisconsin Policy Research Institute). I have used eight issues of this survey, running between January 1988, the earliest, and July 1998. 12. Martin Gilens, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy (Chicago: University of Chicago Press, 1999), chaps. 2, 8. 13. This section is indebted to David Dodenhoff, a research assistant who researched state journalism covering the Wisconsin reform for me. 14. Thomas Kaplan, “Wisconsin Works,” in Managing Welfare Reform in Five States: The Challenge of Devolution, ed. Sarah F. Liebschutz (Albany: Rockefeller Institute Press, 2000), pp. 107–108; interviews with Tony Earl and Peter Tropman, both in Madison, WI, June 12, 1997. I extended the Tropman interview by phone on May 18, 2000. Tropman was a Democratic state representative from Milwaukee, 1973–77, then administrator of the Division of Policy and Budget at the Department of Health and Social Services, 1977–87. 15. Interview with Linda Reivitz, Madison, WI, June 13, 1997. Reivitz was Secretary of the Department of Health and Social Services, in charge of welfare, under Governor Earl. 16. Paul E. Peterson and Mark C. Rom, Welfare Magnets: A New Case for a National Standard (Washington, DC: Brookings, 1990), chap. 2. 17. “Speaker, Governor Split on Welfare Reform,” Milwaukee Sentinel, January 8, 1988, part 1, p. 11; Jason DeParle, “Getting Opal Caples to Work,” New York Times Magazine, August 24, 1997, p. 47. 18. Tommy G. Thompson, Power to the People: An American State at Work (New York: HarperCollins, 1996), pp. 54, 75. 19. Steve Schultze, “New Learnfare Compromise Further Limits Scope of Plan,” Milwaukee Journal, November 13, 1987, part B, p. 2.

288

•

Notes to Chapter 3

20. Cyril Ignatius Kendrick, “The Roots of Welfare Reform: The Social Forces Underlying the Wisconsin Learnfare Program,” Ph.D. diss. (Blacksburg, VA: Virginia Polytechnic Institute, Department of Sociology, March 1993), pp. 219, 223; “Thompson’s vetoes stand as both parties claim victories,” Milwaukee Journal, October 1, 1987, part A, p. 17. 21. State of Wisconsin, 2001–2002 Blue Book (Madison: Wisconsin Legislature, 2001), pp. 208, 239. 22. Interview with Gerald Whitburn, Tomahawk, WI, March 15, 1998; interview with Rep. John Gard, Madison, WI, March 17, 1998. Gard chaired the committee on welfare reform that pushed W-2 through the Assembly. 23. Thompson, Power to the People, pp. 64–67; Thomas Kaplan, “Wisconsin’s W-2 Program: Welfare As We Might Come to Know It?” in Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States, ed. Carol S. Weissert (Albany: Rockefeller Institute Press, March 2000), p. 87. 24. Amy Rinard, “Some Democrats May Embrace Thompson Welfare Reform Plans,” Milwaukee Sentinel, December 23, 1991, part 1, p. 12. 25. Steve Schultze, “Democrats Leery of Expanded Learnfare,” Milwaukee Journal, September 13, 1989, part B, p. 10; Steven Walters and Neil H. Shively, “Expansion of Learnfare Unlikely,” Milwaukee Sentinel, June 6, 1990, part 1, p. 4; Eldon Knoche, “Survey Finds Political World Unrealistic,” Milwaukee Sentinel, October 28, 1992, Part 1, p. 10. 26. Gard interview. 27. “Getting Tough on the Poor,” Newsweek, October 14, 1990, p. 33; Dave Daley, “State Cuts Bold Path on Welfare,” Milwaukee Journal, July 17, 1994, pp. A1, A17; Norman Atkins, “Tommy Thompson: Governor Get-a-Job,” New York Times Magazine, January 15, 1995, pp. 22–25; “The Man Who Should Be King,” The Economist, August 7, 1999, p. 25. 28. Thad Beyle, “Governors: The Middlemen and Women in Our Political System,” in Politics in the American States: A Comparative Analysis, 6th ed., ed. Virginia Gray and Herbert Jacob (Washington, DC: Congressional Quarterly, 1996), pp. 221–38. 29. Richard Berke, “As Elections for Governor Loom, GOP Leaders Worry,” New York Times, April 29, 2002, p. A16. 30. James K. Conant, “Wisconsin’s Budget Deficit: Size, Causes, Remedies, and Consequences” (Fairfax, VA: George Mason University, Department of Political Science, 2002). 31. Kendrick, “Roots of Welfare Reform,” pp. 194–98; “Governor’s Commission Backs Welfare Changes,” Milwaukee Sentinel, April 2, 1987, part 1, p. 11. 32. Amy Rinard, “Welfare Reform Backed by Majority,” Milwaukee Sentinel, September 10, 1993, part 1, p. 9; Stephen Walters and Amy Rinard, “Assembly OKs Plan to End Welfare in ’99,” Milwaukee Sentinel, October 27, 1993, part 1, p. 1; Amy Rinard and Stephen Walters, “Work Not Welfare Goes to Governor,” Milwaukee Sentinel, October 29, 1993, part 1, p. 1. 33. Richard Bradee, “Reivitz Calls Some Welfare Rules ‘Insane,’ ” Milwaukee Sentinel, June 4, 1986, part 1, p. 1. 34. Tropman interview.

Notes to Chapter 3

•

289

35. Peter Tropman, “From the Welfare Mess to Wisconsin Works” (Madison: Department of Health and Human Services, 1986); idem, “Employment Programs for AFDC Recipients—A Vision” (Madison: Department of Health and Human Services, November 19, 1985). 36. Peter Tropman, “Human Services at the Crossroads” (Madison: Department of Health and Social Services, March 11, 1986). 37. Tropman interview; Kendrick, “Roots of Welfare Reform,” pp. 156–75, 187–90. 38. Bob Wagner and Peter Tropman, “Wisconsin Works” (Madison: Department of Health and Social Services, 1986). 39. Kendrick, “Roots of Welfare Reform,” pp. 178–80, 195, 264–90. 40. Interview with former Sen. Joseph Andrea, Kenosha, WI, June 11, 1997. Andrea sat for Kenosha in the Assembly and Senate from 1977 to 1997. 41. Earl interview. 42. Tropman interview. 43. Eighty-one percent of respondents supported these priorities in a poll in January 1992. See Gordon S. Black, “The Wisconsin Citizen Survey: January 1992” (Milwaukee: Wisconsin Policy Research Institute, January 1992), p. 30. 44. Kendrick, “Roots of Welfare Reform,” pp. 199–200; DeParle, “Getting Opal Caples to Work,” p. 54. 45. Thomas J. Corbett, “Welfare Reform in Wisconsin: The Rhetoric and the Reality,” in The Politics of Welfare Reform, ed. Donald F. Norris and Lyke Thompson (Thousand Oaks, CA: Sage, 1995), pp. 22, 27–28. 46. Thompson, Power to the People, pp. 43–44, 49. 47. Whitburn interview. 48. Joel Dresang, “Budget Plan Boosts W-2 Subsidies,” Milwaukee Journal Sentinel, February 13, 1997, part A, p. 15; Richard P. Jones, “Budget Breezes through Senate,” Milwaukee Journal Sentinel, September 26, 1997, part A, p. 1. 49. One occasion arose in 1993, when John Gard and other GOP legislators expressed fear that the proposed replacement of AFDC with a work program would only expand the cost of aid. But later, Gard accepted the higher cost of W-2. See Dave Daley, “‘Ending’ Welfare Will Cost More, GOP Legislators Say,” Milwaukee Journal, December 2, 1993, part B, p. 1; Daniel Bice, “W-2 Costs to Exceed Current Program,” Milwaukee Journal Sentinel, November 28, 1995, part B, p. 1. 50. Dave Daley, “Learnfare Costs $10.2 Million, Moore says,” Milwaukee Journal, June 2, 1994, part B, p. 5; Meg Jones, “Democrats to Ask Governor to Shut Down Learnfare,” Milwaukee Journal Sentinel, June 3, 1996, part B, p. 1. 51. Kendrick, “Roots of Welfare Reform,” pp. 230–34. 52. Earl interview; DeParle, “Getting Opal Caples To Work,” p. 54. 53. Examples include John W. Wahner and Jerome R. Stepaniak, who served as senior county welfare administrators in Milwaukee in the 1980s. They wrote “Welfare In-Migration: A Four-County Report” (Milwaukee: Wisconsin Policy Research Institute, December 1988) to document what they perceived to be largescale welfare migration to the southeast corner of the state. 54. Becker interview. 55. Andrea interview.

290

•

Notes to Chapter 3

56. Craig Gilbert, “Workfare Isn’t the Final Word,” Milwaukee Journal, May 18, 1987, pp. 1A, 11A; “Program Called Mean-Spirited,” Milwaukee Sentinel, April 18, 1990, part 2, p. 6. 57. At a meeting of the Thompson Welfare Reform Commission on April 1, 1987, I heard Loftus press liberal economists from the Institute for Research on Poverty at the University of Wisconsin to explain how anything they proposed would overcome unwed pregnancy among the poor. 58. Eldon Knoche, “Thompson Would Create Jobs for Welfare Recipients,” Milwaukee Sentinel, December 8, 1986, part 1, p. 1; Gretchen Schuldt, “Welfare Overhaul, Health Plans Linked,” Milwaukee Sentinel, July 28, 1993, part 1, p. 9. 59. Mary Beth Murphy, “Norquist Says Thompson’s Welfare Plan Falls Short,” Milwaukee Sentinel, September 16, 1993, part 1, p. 7; “Mayor, 9 to 5 Urge Clinton to Block W-2,” Milwaukee Journal Sentinel, May 22, 1996, part A, p. 12. 60. Marc Cooper, “Overthrowing the Welfare State,” Village Voice, May 31, 1994, pp. 35–36. 61. Doris Hajewski, “Enforce Welfare Wait, County Panel Says,” Milwaukee Journal, February 13, 1986, part 1, p. 1; Daniel P. Hanley, Jr., “Half-Year Limit on General Relief Sought,” Milwaukee Journal, February 18, 1992, part B, p. 4; Gretchen Schuldt, “Panel Urges End to Welfare Grants,” Milwaukee Sentinel, October 18, 1994, part 1, p. 5. 62. Tom Loftus, The Art of Legislative Politics (Washington, DC: Congressional Quarterly, 1994), chap. 8; Craig Gilbert, “Loftus: The Man and His Stance on Key Issues,” Milwaukee Journal, July 15, 1990, part A, p. 14. 63. These were a Select Committee on Work Incentives in the Assembly and a Legislative Council Study Committee, the latter a joint body drawing members from both legislative houses and the public. 64. Charles E. Friederich, “Earl May Try to Freeze Welfare,” Milwaukee Journal, September 28, 1986, part A, p. 1. 65. Kendrick, “Roots of Welfare Reform,” pp. 186–87, 198–89, 216. 66. “Strohl Pushes AFDC Limit for Families Moving to State,” Milwaukee Journal, July 31, 1986, part A, p. 14; Charles E. Friederich, “Welfare Plan Aims to Curb Migration,” Milwaukee Journal, January 5, 1989, part A, p. 1. 67. Dave Daley, “Committee to Study Welfare Alternatives,” Milwaukee Journal, November 1, 1991, part A, p. 10; Amy Rinard, “Panel Backs Welfare Reform to Aid in Arrests,” Milwaukee Sentinel, March 6, 1992, part 1, p. 4; “Welfare Compromise Gets Support from Bipartisan Legislative Panel,” Milwaukee Journal, March 12, 1992, part A, p. 15. 68. Dave Daley, “Democrats Ask for Welfare Reform that Avoids Bashing of Recipients,” Milwaukee Journal, June 22, 1994, part A, p. 12. 69. David R. Riemer, The Prisoners of Welfare: Liberating America’s Poor from Unemployment and Low Wages (New York: Praeger, 1988). 70. Philip Harvey, Securing the Right to Employment: Social Welfare Policy and the Unemployed in the United States (Princeton: Princeton University Press, 1989); Mickey Kaus, The End of Equality (New York: Basic Books, 1992). 71. Kendrick, “Roots of Welfare Reform,” p. 171. The discussion here and below is informed by David Riemer, “Chronology of Events Leading to Ending

Notes to Chapter 3

•

291

AFDC in Wisconsin and Requiring Creation of a Work-Based System” (Milwaukee: Department of Administration, July 18, 1997). 72. Amy Rinard, “Legislator Cites Welfare Migration,” Milwaukee Sentinel, February 27, 1992, part 1, p. 13; Amy Rinard, “Democrats Dust Off Welfare Plan,” Milwaukee Sentinel, May 29, 1993, part 1, p. 10; Dave Daley, “Democrats Tout Own Welfare Plan over Governor’s,” Milwaukee Journal, October 17, 1993, part B, p. 5. 73. Assembly Democratic Caucus, “Reforming Welfare in Wisconsin” (Madison: Assembly Democratic Caucus, n.d.). I received this pamphlet from Notestein. 74. Steve Schultze, “Chvala Offers a Wider Plan to Scrap Welfare,” Milwaukee Journal, August 22, 1994, part A, p. 2 75. Interview with Rep. Antonio Riley, Milwaukee, WI, March 16, 1998; Jason DeParle, “Getting Opal Caples To Work,” p. 35. 76. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984), chap. 5. Elazar’s typology plays an important role in chapters 11–13. 77. Virginia Gray, “The Socioeconomic and Political Context of States,” in Politics in the American States., ed. Gray and Jacob, pp. 28–29; Robert S. Erikson, Gerald C. Wright, and John P. McIver, Statehouse Democracy: Public Opinion and Policy in the American States (Cambridge, England: Cambridge University Press, 1993), chap. 2. 78. Earl interview; interview with Joseph Strohl, Madison, WI, June 12, 1997. Strohl was a state senator for Racine from 1978 to 1990 and was Democratic majority leader of the Senate, 1986–90. 79. Erikson, Wright, and McIver, Statehouse Democracy, chap. 5; Tbomas R. Dye, “Party and Policy in the States,” Journal of Politics 46, no. 4 (November 1984): 1097–16. 80. Bibby and Holbrook, “Parties and Elections,” in Politics in the American States., ed. Gray and Jacob, pp. 103–107. 81. Steven Walters and Neil H. Shively, “Expansion of Learnfare Unlikely,” Milwaukee Sentinel, June 6, 1990, Part 1, p. 4; 2/6/91, Jo Sandin, “Workfare’s Impact Challenged,” Milwaukee Journal, February 6, 1991, part A, p. 1; “State General Relief Rolls up 20% since ’89, Review Shows,” Milwaukee Journal, December 5, 1991, part A, p. 22; Kenneth Lamke, “Ament Urges 2-Tier Benefit System,” Milwaukee Sentinel, December 9, 1991, part 1, p. 4. 82. Hugh Heclo, Modern Social Politics in Britain and Sweden: From Relief to Income Maintenance (New Haven: Yale University Press, 1974). 83. Gerald A. Berge, “ ‘From the Welfare Mess to Wisconsin Works’—An Alternative” (Madison: Department of Health and Social Services, August 4, 1986); Kendrick, “Roots of Welfare Reform,” pp. 168–70, 171–73. 84. Leon D. Epstein, Politics in Wisconsin (Madison: University of Wisconsin Press, 1958), chaps. 2, 5; David R. Mayhew, Placing Parties in American Politics: Organization, Electoral Setting, and Governmental Activity in the Twentieth Century (Princeton: Princeton University Press, 1986), part 1; Peterson and Rom, Welfare Magnets, pp. 38–46. 85. Strohl interview; “Unimpassioned Look at Aiding Poor Vowed,” Milwaukee Sentinel, September 19, 1986, part 2, p. 10.

292

•

Notes to Chapter 3

86. Becker interview. 87. Interview with Rep. Barbara Notestein, March 17, 1998, in Madison, WI. 88. Charles Murray, Losing Ground: American Social Policy, 1950–1980 (New York: Basic Books, 1984). 89. Mead, Beyond Entitlement. 90. According to personal records, in early 1987 I met individually with several officials at DHSS (Secretary Tim Cullen, Al Fish, Jim Meier, Peter Tropman, and several of his analysts), and Betty Jo Nelson, minority leader of the Assembly. I also participated in two meetings of the Governor’s Welfare Reform Commission, where I and other experts conferred with Nelson and Tom Loftus, speaker of the Assembly. I interacted extensively with Irwin Garfinkel, Tom Corbett, Bernie Stumbras, Michael Wiseman, and other poverty experts and researchers at the Institute for Research on Poverty at the University. I also spoke to groups of DHSS staff and legislators alongside Garfinkel, gave a public lecture on welfare reform alongside Wiseman, and spoke to the Republican caucus in the Assembly. 91. Corbett, “Welfare Reform in Wisconsin,” p. 34; Kendrick, “Roots of Welfare Reform,” pp. 69–72, 190–91, 300–301. 92. In a study of the politics of welfare reform more thorough than mine, Michael Heaney, a graduate of the University of Chicago, has ranked 106 individuals and groups who influenced Wisconsin outcomes during 1993–99 (personal communication, August 22, 2000). The governor’s office ranked first. At thirty, I and Michael Wiseman, another welfare expert, were the highest-ranked academics. 93. Whitburn mentioned Ellwood, Kaus, and Wilson to me in our interview. He mentioned Jencks to Michael Heaney (personal communication from Heaney, June 2, 2000). 94. Examples would include Thomas Corbett, Thomas Kaplan, and Ingrid Rothe, all of whom in different ways were sources for this research. 95. Erikson et al., Statehouse Democracy, chaps. 5–7. 96. Norman Atkins, “Tommy Thompson,” p. 24; Steven Walters, “Senate OKs Aid Cutoff Experiment,” Milwaukee Sentinel, October 22, 1993, part 1, p. 1; Amy Rinard, “‘W-2’ Welfare Reform Proposal Called ‘Return to Slavery,’ ” Milwaukee Journal Sentinel, August 13, 1995, part B, p. 5; Cooper, “Overthrowing the Welfare State,” p. 37. 97. Steven Walters, “Panel Rejects Learnfare Expansion,” Milwaukee Sentinel, May 22, 1991, part 1, p. 1; Dave Daley, “Assembly Approves 2-Tier Welfare Checks,” Milwaukee Journal, March 24, 1992, part A, p. 1; interview with Sen. Gary George, Madison, WI, March 19, 1998. 98. DeParle, “Getting Opal Caples to Work,” p. 34; Cooper, “Overthrowing the Welfare State,” p. 33; Antonio Riley, “From a Central City Legislator: Why I Said No,” Milwaukee Journal, February 27, 1994, part J, p. 1; Riley interview. 99. According to my interview with Riley, five out of six black representatives in the Assembly voted for his proposal to abolish AFDC, while both black senators (George and Moore) opposed it. 100. David E. Umhoefer, “Criticism of Welfare Runs Deep,” Milwaukee Journal, February 12, 1992, part A, p. 1. 101. Interview with Will Martin, Milwaukee, WI, December 4, 1995; DeParle, “Getting Opal Caples to Work,” p. 36.

Notes to Chapter 3

•

293

102. Interview with Pat Gowens and other Welfare Warriors in Milwaukee on March 16, 1998. 103. Interview with Marcus White, Interfaith Conference of Greater Milwaukee, in Milwaukee, WI, March 16, 1998. The Interfaith Conference represents various churches and synagogues in Milwaukee. By one account I heard, the Warriors were voted down 14–1 on this issue at a meeting of community groups sometime in 1994–95. 104. Interview with Pat DeLessio, a staff attorney for Legal Action of Wisconsin, Milwaukee, WI, March 16, 1998. 105. DeLessio interview; White interview; Amy Rinard, “Welfare Reform Changes Sought,” Milwaukee Journal Sentinel, December 21, 1995, part B, p. 5; Fran Bauer, “Legislature Told to Go Slow on W-2,” Milwaukee Journal Sentinel, December 29, 1995, part B, p. 1. 106. White interview. 107. Kendrick, “Roots of Welfare Reform,” p. 224. 108. The most forthright opponents were the economists Robert Haveman and Sheldon Danziger, who both served as directors of IRP. Irwin Garfinkel, the child support expert and also a former director, was more receptive to enforcement, and Haveman became so in later years. Thomas Corbett, Thomas Kaplan, and Michael Wiseman were less visible IRP figures who helped to create, implement, and evaluate WEJT and later welfare work initiatives in Wisconsin and beyond. I have drawn heavily on their publications in this book. As an academic institution, of course, IRP takes no collective position on political issues. 109. Thompson, Power to the People, pp. 64–65. One cost to IRP was the killing of the Child Support Assurance System (CSAS) conceived by Irwin Garfinkel. Under CSAS, single mothers with child support orders would have been assured some minimum of child support whether or not the father paid support. Wisconsin was poised to test the new system when Tommy Thompson was elected governor in 1986. Despite lobbying by IRP, with my support, CSAS died. See Loftus, Art of Legislative Politics, pp. 125–27. Later, a related program, the Child Assistance Program, was tested in New York. 110. E.g., Maria Cancian and Dan Meyer, “A Profile of the AFDC Caseload in Wisconsin: Implications for a Work-Based Welfare Reform Strategy” (Madison: University of Wisconsin, Institute for Research on Poverty, September 1995). 111. John Pawasarat and Lois M. Quinn, “Wisconsin Welfare Employment Experiments: An Evaluation of the WEJT and CWEP Programs” (Milwaukee: University of Wisconsin–Milwaukee, Employment and Training Institute, September 1993). 112. Edward C. Banfield and James Q. Wilson, City Politics (Cambridge, MA: Harvard University Press, 1963), chap. 18; Mark D. Brewer and Jeffrey M. Stonecash, “The Economy, Taxes, and Policy Constraints in New York,” in Governing New York State, 4th ed., ed. Jeffrey M. Stonecash (Albany: State University of New York Press, 2001), chap. 11. 113. John H. Fenton, Midwest Politics (New York: Holt, Rinehart, and Winston, 1966), p. 62. 114. Telephone interviews with Robert W. Brennan, President, Greater Madison Chamber of Commerce, May 19, 2000; Amy Overby, Madison Community

294

•

Notes to Chapter 4

Foundation, May 22, 2000; and Bob Koch, manager of Future Madison, May 31, 2000; Department of Health and Social Services, Wisconsin Works: 1999 Plan, Draft (Madison: Department of Health and Social Services, April 1995), p. 56. 115. Berke, “As Elections for Governor Loom”; Jodi Wilgoren, “In Wisconsin, Scandal, Outrage and Deficit Churn Up a Storm of Political Change,” New York Times, May 19, 2002, p. 18; David E. Rosenbaum, “Scandals Begin to Tarnish Wisconsin’s Political Luster,” New York Times, July 10, 2002, p. A12. CHAPTER 4 IMPLEMENTING WORK REQUIREMENTS 1. This chapter is based principally on Lawrence M. Mead, “Implementing Work Requirements in Wisconsin,” Journal of Public Policy 21, no. 3 (June 2001): 239–64. I acknowledge comments from reviewers. 2. Richard P. Nathan and Thomas L. Gais, Implementing the Personal Responsibility Act of 1996: A First Look (Albany: State University of New York, Rockefeller Institute of Government, 1999); Michael Wiseman, “How Workfare Really Works,” The Public Interest no. 89 (Fall 1987): 36–47. 3. Jan L. Hagen and Irene Lurie, Implementing JOBS: Progress and Promise (Albany: State University of New York at Albany, Rockefeller Institute of Government, 1994); Donald F. Norris and Lyke Thompson, eds. The Politics of Welfare Reform (Thousand Oaks, CA: Sage, 1995); Nathan and Gais, Implementing the Personal Responsibility Act; Carol S. Weissert ed., Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States (Albany, NY: Rockefeller Institute Press, 2000). 4. LaDonna A. Pavetti, Pamela Holcomb, and Amy-Ellen Duke, Increasing Participation in Work and Work-Related Activities: Lessons from Five State Welfare Reform Demonstration Projects, Final Report, Volume II: Site Visit Summaries (Washington, DC: Urban Institute, 1995); U.S. General Accounting Office, Welfare Reform: Three States’ Approaches Show Promise of Increasing Work Participation (Washington, DC: U.S. General Accounting Office, 1997); U.S. General Accounting Office, Welfare Reform: States Are Restructuring Programs to Reduce Welfare Dependence (Washington, DC: U.S. General Accounting Office, 1998). 5. Judith M. Gueron and Edward Pauly, with Cameran M. Lougy, From Welfare to Work (New York: Russell Sage Foundation, 1991). 6. Lawrence M. Mead, “Optimizing JOBS: Evaluation versus Administration” Public Administration Review 57, no. 2 (1997): 113–23; Marcia K. Meyers, Bonnie Glaser, and Karin Mac Donald, “On the Front Lines of Welfare Delivery: Are Workers Implementing Policy Reforms?” Journal of Policy Analysis and Management 17 no. 1(1998): 1–22. 7. Richard R. Nelson, The Moon and the Ghetto (New York: Norton, 1977); Jeffrey L.Pressman and Aaron Wildavsky, Implementation, 3rd ed. (Berkeley: University of California Press, 1984). 8. Daniel A. Mazmanian and Paul A. Sabatier, Implementation and Public Policy (Glenview, IL: Scott, Foresman, 1983).

Notes to Chapter 4

•

295

9. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984), pp. 117–18, 202, 221; Russell L. Hanson, “Intergovernmental Relations,” in Virginia Gray and Herbert Jacob, Politics in the American States: A Comparative Analysis, 6th ed. (Washington, DC: Congressional Quarterly, 1996), pp. 67–69. 10. Sheila R. Zedlewski, Pamela A. Holcomb, and Amy-Ellen Duke, Cash Assistance in Transition: The Story of Thirteen States (Washington, DC: Urban Institute, December 1998), pp. 22–23. 11. Mark Lincoln Chadwin, John J. Mitchell, Erwin C. Hargrove, and Lawrence M. Mead, The Employment Service: An Institutional Analysis, U.S. Department of Labor, R&D Monograph 51 (Washington, DC: U.S. Government Printing Office, 1977); John J. Mitchell, Mark L. Chadwin, and Demetra S. Nightingale, Implementing Welfare-Employment Programs: An Institutional Analysis of the Work Incentive (WIN) Program, U.S. Department of Labor, R&D Monograph 78 (Washington, DC: U.S. Government Printing Office, 1980); Lynn Schmitt, “Employment and Training Programs in Wisconsin” (Madison: Department of Workforce Development, 1996). 12. Thomas J. Corbett, “Welfare Reform in Wisconsin: The Rhetoric and the Reality,” in Politics of Welfare Reform, ed. Norris and Thompson, p. 34. 13. Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1991), p. 22. 14. Michael Wiseman, “State Strategies for Welfare Reform: The Wisconsin Story,”Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 16; data from the U.S. Administration for Children and Families. 15. Department of Health and Social Services, “JOBS—Job Opportunities and Basic SkillsTraining Program“ (Madison: Department of Health and Social Services, 1990), p. 2. 16. Legislative Fiscal Bureau, Aid to Families with Dependent Children, pp. 22–26. 17. Legislative Fiscal Bureau, “1989–91 Budget: Health and Social Services— Public Assistance—Federal Welfare Reform, AFDC Employment and Training Programs” (Madison: Legislative Fiscal Bureau, 1989), p. 3; Department of Health and Social Services, “JOBS—Job Opportunities and Basic Skills Training Program,” p. 11; Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1995), pp. 21, 24. 18. John Pawasarat and Lois M. Quinn, “Wisconsin Welfare Employment Experiments: An Evaluation of the WEJT and CWEP Programs” (Milwaukee: University of Wisconsin–Milwaukee, Employment and Training Institute, September 1993), pp. 48–51. 19. Calculated from JOBS data in U.S. Congress, House, Committee on Ways and Means, Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, 1991–96). 20. Wiseman, “State Strategies for Welfare Reform,” pp. 529–30. 21. Hagen and Lurie, Implementing JOBS. 22. Corbett, “Welfare Reform in Wisconsin,” p. 21.

296

•

Notes to Chapter 4

23. Thomas Kaplan, “Wisconsin’s W-2 Program: Welfare as We Might Come to Know It?” in Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States, ed. Carol S. Weissert (Albany: Rockefeller Institute Press, 2000), pp. 104–105. 24. U.S. Department of Health and Human Services, Welfare Administrative Costs (Washington, DC: U. S Department of Health and Human Services, 1995), tables 1 and 8. 25. U.S. Administration for Children and Families, Time Trends, Fiscal Years 1985–1994 (Washington, DC: U.S. Department of Health and Family Services, 1995), p. 7. 26. Nationally, the percents all welfare adults made mandatory for JOBS fell in this range, 1992–96. Data from the U.S. Administration for Children and Families. 27. The Food Stamp benefit was reduced by $4.25 per missing hour, although not below $10 a month, due to federal rules. 28. Based on calculations from the sources for table 4.1. 29. Irene Lurie, “A Lesson from the JOBS Program: Reforming Welfare Must Be Both Dazzling and Dull,” Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 580–84. 30. Department of Health and Social Services, “JOBS—Job Opportunities and Basic Skills Training Program,” pp. 17–20. As one sign of the times, when I taught at the University of Wisconsin—Madison in 1987, I had a Ph.D. student who was on welfare. She had been allowed to turn her welfare grant, in effect, into a graduate fellowship. 31. U.S. General Accounting Office, States Are Restructuring Programs, pp. 27–35. 32. Data from the Department of Health and Social Services. 33. See tables 5.1 and 5.2 in chapter 5. 34. Note that these data show enrollment in a single month. In table 4.1, numbers run lower because JOBS participation is measured over a year, and it required activity beyond enrollment. 35. In the JOBS period, the best evidence for the bureaucracy’s competence was simply the high participation rates achieved. During my fieldwork, the Wisconsin work programs seemed to be better run than similar programs I had encountered in other states, especially New York. In chapter 11, I compare Wisconsin to other states more rigorously during the TANF period. 36. Cyril Ignatius Kendrick, “The Roots of Welfare Reform: The Social Forces Underlying the Wisconsin Learnfare Program,“ Ph.D. diss. (Blacksburg, VA: Virginia Polytechnic Institute, Department of Sociology 1993), pp. 168–73, 201– 202, 223–24. 37. Department of Health and Social Services, JOBS Annual Report (Madison: Department of Health and Social Services, 1994), p. 28. The competitive system was the invention of Jason Turner, who managed welfare reform initiatives for the Department of Health and Social Services. 38. Milwaukee’s providers had to make job placements at least equal to 60 percent of the level in the previous year, a threshold missing in the incentive system elsewhere in the state. Above 60 percent, they earned $600 for a full-time

Notes to Chapter 5

•

297

placement, $300 for part-time, plus $200 per client per month placed in a work component. 39. Data from the Private Industry Council (PIC) of Milwaukee County. The PIC had a coordinating role in welfare reform in Milwaukee. See chapter 7. 40. Data from the Department of Workforce Development. 41. Data from the Department of Health and Social Services and the U.S. Administration for Children and Families. 42. U.S. General Accounting Office, Welfare Reform: States’ Early Experiences With Benefit Termination (Washington, DC: U.S. General Accounting Office, 1997a), pp. 20–32. 43. Data from the Department of Workforce Development. 44. Martha Derthick, New Towns In-Town: Why a Federal Program Failed (Washington, DC: Urban Institute, 1972); Pressman and Wildavsky, Implementation; Mazmanian and Sabatier, Implementation and Public Policy. CHAPTER 5 LOCAL VARIATIONS 1. This chapter is largely based on Lawrence M. Mead, “Welfare Reform in Wisconsin: The Local Role,” Administration and Society 33, no. 5 (November 2001): 523–54; and idem, “Optimizing JOBS: Evaluation versus Administration,” Public Administration Review 57, no. 2 (March/April 1997): 113–23. 2. Interviews with Jason Turner on November 14, 1996, and October 10, 2000, and with Jean Rogers on October 10, 2000. Within the state welfare department, Turner was the chief manager of welfare reform from 1993 to 1997, and Jean Rogers headed the income maintenance division from 1991 to 2000. I also heard similar opinions from many other officials during my interviews. 3. Jeffrey L. Pressman and Aaron Wildavsky, Implementation, 3rd ed. (Berkeley: University of California Press, 1984); and Eugene Bardach, The Implementation Game: What Happens after a Bill Becomes a Law (Cambridge, MA: MIT Press, 1977). 4. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper and Row, 1984), pp. 117–18, 202, 221. 5. I interviewed in Dane, Fond du Lac, Kenosha, Milwaukee, Racine, and Sheboygan in July 1994, and in Douglas, Grant, Marathon, Winnebago, and again Fond du Lac in June 1995. I made later visits to Kenosha in 1997 and to Milwaukee in 1995 and 1999. In 1994 and 1995 and on several later occasions, I also interviewed state officials. I also collected program documents from counties. The Kenosha and Dane accounts also rely on Thomas Kaplan, Thomas Corbett, and Elizabeth Boehnen, “First Report Form” (Albany: State University of New York, Rockefeller Institute of Government, July 1, 1997), written for the State Capacity Study at the Rockefeller Institute of Government at the State University of New York. There may have been other leading or changing counties, but I did not hear of them. 6. I depended initially on state officials to characterize which counties were high and low performers, but their judgments proved accurate. It is better to select cases for a study like this on the basis of variation in the independent term (here

298

•

Notes to Chapter 5

program features) rather than the dependent term (here program performance). But it makes little difference provided a wide range of variation in both terms is explored, as I achieved. See Gary King, Robert D. Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton: Princeton University Press, 1994), pp. 128–40. 7. The other WEJT pilots were in Douglas, Jackson, Racine, and Rock Counties. The program was later expanded to much of the state, as recounted in chapter 4. 8. Gayle Hamilton and Daniel Friedlander, Final Report on the Saturation Work Initiative Model in San Diego (New York: Manpower Demonstration Research Corporation, 1989). SWIM’s interim results were published in 1988. According to Jon Angeli, the 32-hour simulated work week originated in Grant County (personal communication, August 10, 2001). 9. Thomas Corbett, Lawrence Mead, Bernie Stumbras, and Michael Wiseman, A Management Assessment Report on the Pilot Work Experience and Job Training Program in Kenosha County (Madison: University of Wisconsin—Madison, Institute for Research on Poverty, December 1987). 10. Work first was also piloted in Jackson County, where the emphasis was also on employment before training, and in Grant County, where the stress was on diversion. 11. Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, 1991), pp. 22–23; idem, Aid to Families with Dependent Children, Informational paper #45 (Madison: Legislative Fiscal Bureau, 1995), p. 21. 12. This calculation is based on staff data from annual Kenosha JOBS reports and county caseload data from the state Department of Health and Social Services. 13. “Welfare Reform on the Front Lines: The La Follette-Kenosha County Project,” La Follette Policy Report 2, no. 2 (1990): 6. 14. Michael Wiseman, “Sample Family Support Act Job Opportunities and Basic Skills Training (JOBS) Participation Data (Revised)” (Madison: University of Wisconsin—Madison, La Follette Institute of Public Affairs, November 24, 1991). 15. “Evaluation Report: KCDSS Post Participant Tracking Project” (Milwaukee: Milwaukee Employment and Training, 1991). A contact rate of 39 percent may appear low, but it was 39 percent of all leavers, not of a sample of leavers as in most recent leavers studies. 16. Department of Health and Social Services, JOBS Annual Report (Madison: Department of Health and Social Services, 1994), p. 6. 17. This is the approach developed by the Portland JOBS program, the most successful site in the National Evaluation of Welfare-to-Work Strategies (NEWWS). Kenosha WEJT also resembled Portland JOBS in that it was the creation of cross-agency collaboration. On NEWWS, see Gayle Hamilton, Stephen Freedman, Lisa Gennetian, Charles Michalopoulos, Johanna Walter, Diana Adams-Ciardullo, Anna Gassman-Pines, Sharon McGroder, Martha Zaslow, Surjeet Ahluwalia, and Jennifer Brooks, with Electra Small and Bryan Ricchetti, National Evaluation of Welfare-to-Work Strategies: How Effective Are Different Welfare-to-Work Approaches? Five-Year Adult and Child Impacts for Eleven Pro-

Notes to Chapter 5

•

299

grams (New York: MDRC, November 2001). On Portland JOBS, see Susan Scrivener, Gayle Hamilton, Mary Farrell, Stephen Freedman, Daniel Friedlander, Marisa Mitchell, Jodi Nudelman, and Christine Schwartz, National Evaluation of Welfare-to-Work Strategies: Implementation, Participation Patterns, Costs, and Two-Year Impacts of the Portland (Oregon) Welfare-to-Work Program (Washington, DC: U.S. Department of Health and Human Services and U.S. Department of Education, May 1998). 18. Department of Health and Social Services, JOBS Annual Report, p. 6. 19. Arlene D. Siss, “Grant County Department of Social Services Child Day Care Synopsis” (Lancaster, WI: Grant County Department of Social Services, n.d.). 20. Mark R. Rank and Thomas A. Hirschl, “A Rural-Urban Comparison of Welfare Exits: The Importance of Population Density,” Rural Sociology 53, no. 2 (1988): 190–206. 21. “Sheboygan County JOBS Program Self-Evaluation of 1992” (Sheboygan, WI: Curtis and Associates, n.d.), pp. 3, 9. 22. Department of Health and Social Services, JOBS Annual Report, p. 9. “Disadvantaged,” following JOBS rules, included recipients who had been on AFDC for thirty-six out of the last sixty months, single parents under twenty-four who lacked a high school diploma or much work experience in the last year, or adults in a family due to become ineligible for AFDC within two years due to the youngest child turning eighteen. 23. “Dane County JOBS Program Self-Evaluation for Program Year 1993” (Madison: Dane County Department of Human Services, n.d.). 24. “Racine County Human Services Department 1993 JOBS Program SelfEvaluation” (Racine: Racine County Human Services Department, n.d.). 25. Milwaukee in 1993 was funded at a level that assumed that only 35 percent of its caseload was mandatory for JOBS, as against higher levels for other counties. See Legislative Fiscal Bureau, Aid to Families with Dependent Children, January 1995, p. 21. 26. Thomas Brock and Kristen Harknett, “A Comparison of Two Welfare-towork Case Management Models,” Social Service Review 72, no. 4 ( Fall 1998): 493–520. 27. Legislative Fiscal Bureau, State Job Training Programs (Madison: Legislative Fiscal Bureau, 1991), p. 13. 28. “Welcome to the Kenosha County Job Center” (Kenosha, WI: Kenosha County Job Center, 1994), p. 3. In addition to these innovations, Kenosha developed a customized reporting system, the Job Center Information System, that was better at following cases through assignments than the state system at the time. Designed by Michael Wiseman of the University of Wisconsin—Madison, it was displaced by the CARES system in 1994. 29. Jennifer Ehrle, Kristin Seefeldt, Kathleen Snyder, and Pat McMahon, Recent Changes in Wisconsin Welfare and Work, Child Care, and Child Welfare Systems (Washington, DC: Urban Institute, September 2001), pp. 10–11. The merger might, however, pose problems for welfare reform, as discussed in chapter 11.

300

•

Notes to Chapter 5

30. Lawrence M. Mead, “Expectations and Welfare Work: WIN in New York State,” Polity 18, no. 2 (Winter 1985): 236–42. Kenosha, however, threatened many more clients with sanctioning and then brought them into compliance with “reconciliation” consultations, avoiding an actual sanction. In October 1993, Kenosha was above the state norm in assignments to reconciliation, and in October 1995 both Kenosha and Sheboygan were. 31. John J. Mitchell, Mark L. Chadwin, and Demetra S. Nightingale, Implementing Welfare-Employment Programs: An Institutional Analysis of the Work Incentive (WIN) Program, R&D Monograph 78, U.S. Department of Labor (Washington, DC: U.S. Government Printing Office, 1980), pp. 273–96. 32. Due to a change in computer systems in 1994, 1993 is the last year for which this analysis is possible. That year also catches the JOBS program at its height, just before the more radical, diversion-oriented reforms that the state implemented from 1994 on. 33. The correlations between the variables was .52 in figure 5.1, .39 in figure 5.2. Both figures are significant at .05. 34. In the regression models following, the administrative terms use program data. The demographic control terms were calculated from a special breakdown of the JOBS caseload by county in December 1993 done by the state welfare department at my request. Labor market terms come from the U.S. Bureau of Labor Statistics. Each performance measure was regressed on several measures of the three types. In tables 5.4–5.6, I assemble the explainers of the first three measures that proved significant at .10 or better. For further explanation of and detail on this approach, see Mead, “Optimizing JOBS,” and idem, “Performance Analysis,” in Policy into Action: Implementation Research and Welfare Reform, ed. Mary Clare Lennon and Thomas Corbett (Washington, DC: Urban Institute Press, 2003). 35. As above, “disadvantaged” was defined according to JOBS rules. 36. R2 for the three models was .54, .61, and .59. 37. Calculations from data from the Department of Health and Social Services and the Department of Workforce Development. 38. Mark Alan Hughes, “Learning from the ‘Milwaukee Challenge,’ ” Journal of Policy Analysis and Management, vol. 15, no. 4 (Fall 1996): 567; Thomas Brock, Fred Doolittle, Veronica Fellerath, and Michael Wiseman, Creating New Hope: Implementation of a Program to Reduce Poverty and Reform Welfare (New York: Manpower Demonstration Research Corporation, October 1997), pp. 35–47. 39. The Milwaukee case was well predicted in all the above analyses. Omitting it from the regressions did not change the results much, nor was a dummy variable for the county significant in any of them. See Mead, “Optimizing JOBS,” p. 115. 40. For a fuller discussion of the issues below, see Mead, “Performance Analysis,” and idem, “Are Welfare Employment Programs Effective?” in Social Programs That Work, ed. Jonathan Crane (New York: Russell Sage, 1998), chap. 11. 41. Such as multicollinearity (not a problem here) or the endogeneity (nonindependence) among the explainers. 42. Judith M. Gueron and Edward Pauly, with Cameran M. Lougy, From Welfare to Work (New York: Russell Sage Foundation, 1991), p. 148. A recent meta-

Notes to Chapter 6

•

301

analysis based on many MDRC studies found, similar to my results, that the largest influence on a program’s earnings impacts was how much emphasis it gave to moving clients quickly into jobs. See Howard S. Bloom, Carolyn J. Hill, and James Riccio, “Linking Program Implementation and Impacts: Lessons from a Pooled Sample of Welfare-to-Work Experiments,” Journal of Policy Analysis and Management 22, no. 4 (Fall 2003). Admittedly, with training programs under the Job Training Partnership Act, there is only a limited association between programs that perform well by official measures and those that evaluated well in the national JTPA evaluation. See James Heckman, Carolyn Heinrich, and Jeffrey Smith, “Assessing the Performance of Performance Standards in Public Bureaucracies,” American Economic Review 87, no. 2 (May 1997): 389–95; Burt S. Barnow, “Exploring the Relationship between Performance Management and Program Impact: A Case Study of the Job Training Partnership Act,” Journal of Policy Analysis and Management 19, no. 1 (Winter 2000): 118–41. This is probably because JTPA is a voluntary program, so that the enforcement methods that generate effects in welfare work programs cannot be used. 43. Robert D. Behn, “Management by Groping Along,” Journal of Policy Analysis and Management 7, no. 4 (Fall 1988): 643–63. 44. As explained in Mead, “Optimizing JOBS,” p. 115, administrative funding levels relative to caseloads had hardly any influence on county performance in JOBS. 45. Tommy G. Thompson, Power to the People: An American State at Work (New York: HarperCollins, 1996), p. 70. 46. Eugene Bardach, Improving the Productivity of JOBS Programs (New York: Manpower Demonstration Research Corporation, 1993); Richard D. Behn, Leadership Counts: Lessons for Public Managers from the Massachusetts Welfare, Training, and Employment Program (Cambridge, MA: Harvard University Press, 1991); Lawrence M. Mead, Beyond Entitlement: The Social Obligations of Citizenship (New York: Free Press, 1986), chap. 7. 47. When I first visited Fond du Lac in 1994, I had Bardach, Improving the Productivity of JOBS Programs, with me. I expected to mention it to the managers, who had just agreed to join Work Not Welfare and expected wrenching changes. But when I first met with them, there was the Bardach pamphlet on the table. They had already read and discussed it. 48. “Politics as a Vocation,” in Max Weber, From Max Weber: Essays in Sociology, ed. and trans. H. H. Gerth and C. Wright Mills (New York: Oxford University Press, 1958), p. 128. CHAPTER 6 THE EMERGENCE OF W-2 1. Interview with Jason Turner, New York, NY, February 27, 2001. But as explained in chapters 3 and 7, several less visible IRP researchers made important contributions to the administrative side of welfare reform. 2. Jason DeParle, “Getting Opal Caples to Work,” New York Times Magazine, August 24, 1997, p. 37.

302

•

Notes to Chapter 6

3. Turner interview, February 27, 2001; Jason Turner, remarks on “Welfare and Employment: Policy Making in Michigan and Wisconsin—A Roundtable,” conference of the Association for Public Policy Analysis and Management (APPAM), Pittsburgh, PA, November 1, 1996; David Dodenhoff, “Privatizing Welfare in Wisconsin” (Thiensville, WI: Wisconsin Policy Research Institute, January 1998), p. 7. 4. Interview with Andrew Bush, New York, NY, April 9, 2001. 5. Turner interview, February 27, 2001. 6. Andrew Bush, “Replacing Welfare in Wisconsin,” Public Welfare 54, no. 2 (Spring 1996) 18–19, 21. 7. Thomas Kaplan, Thomas Corbett, and Elisabeth Boehnen, “First Report Form” (Albany: State University of New York, Rockefeller Institute of Government, July 1, 1997), pp. 12, 62. 8. I owe this insight to Michael Wiseman. 9. Jennifer Ehrle, Kristin Seefeldt, Kathleen Snyder, and Pat McMahon, Recent Changes in Wisconsin Welfare and Work, Child Care, and Child Welfare Systems (Washington, DC: Urban Institute, September 2001), p. 19. 10. Department of Health and Social Services, Wisconsin Works: 1999 Plan, Draft (Madison: Department of Health and Social Services, April 1995), pp. 1–6. 11. Bush, “Replacing Welfare in Wisconsin,” p. 19. 12. Department of Health and Social Services, Wisconsin Works, pp. 5–6, 14, 27, 32–33. 13. Ibid., pp. 7–12. 14. Maureen R. Waller and Robert Plotnick, “Effective Child Support Policy for Low-Income Families: Evidence from Street Level Research,” Journal of Policy Analysis and Management 20, no. 1 (Winter 2001): 89–110. 15. Turner interview, February 27, 2001. Any child support received was still included in income in determining eligibility for W-2, until 1999 when an amendment excluded it starting in 2000. 16. Department of Health and Social Services, Wisconsin Works, p. 72. 17. Ibid., chaps. 3, 7. 18. Ibid., pp. 21–23. 19. Ibid., pp. 46–47. 20. Bush interview. 21. I was one of several experts to speak to the group at a session in Madison on July 12, 1994. 22. Amy Rinard, “Private Firms May Run New Welfare System,” Milwaukee Journal-Sentinel, July 1, 1995, part A, p. 1. 23. Bush interview. 24. Dodenhoff, “Privatizing Welfare,” pp. 10–15. 25. Legislative Fiscal Bureau, Wisconsin Works: W-2, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1997), p. 59; idem, Wisconsin Works (W-2) and Other Economic Support Programs, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 2001), p. 30.

Notes to Chapter 6

•

303

26. Amy Rinard, “W-2 Changes Would Cost $119 Million,” Milwaukee Journal Sentinel, February 28, 1996, part B, p. 1. 27. Turner, remarks at APPAM conference; Turner interview, February 27, 2001; Jean Rogers, personal communication, June 5, 2002. 28. Kaplan, “Wisconsin’s W-2 Program,” p. 93 29. Amy Rinard, “ ‘W-2’ Welfare Reform Proposal Called ‘Return to Slavery,’ ” Milwaukee Journal Sentinel, August 13, 1995, part B, p. 5; Daniel Bice and Steven Walters, “Democrats Unhappy with Lack of Solutions,” Milwaukee Journal Sentinel, January 31, 1996, part A, p. 8. 30. Amy Rinard, “Welfare Plan Puts All to Work,” Milwaukee Journal-Sentinel, August 4, 1995, part A, p. 1. 31. Amy Rinard, “W-2 Welfare Bill Remains a Work in Progress,” Milwaukee Journal Sentinel, November 20, 1995, part B, p. 5; idem, “Foes Say W-2 Will Force Students to Quit School,” Milwaukee Journal Sentinel, January 3, 1996, part A, p. 5; Bice and Walters, “Democrats Unhappy.” 32. Interview with Representative Antonio Riley, Milwaukee, WI, March 16, 1998; Joel Dresang, “ ‘We cannot go back,’ W-2 Advocate Riley Says He’ll Push to Refine Work-Based Program,” Milwaukee Journal Sentinel, March 31, 1998, p.12. 33. Amy Rinard, “400 Pack Capitol Hearing on W-2,” Milwaukee Journal Sentinel, January 5, 1996, part A, p. 1; Amy Rinard, “Job Training Plan Added to W-2 Bill,” Milwaukee Journal Sentinel, February 20, 1996, part A, p. 1. 34. Interview with Representative John Gard, Madison, WI, March 17, 1998. 35. Amy Rinard, “W-2 Rolls through Assembly,” Milwaukee Journal Sentinel, March 8, 1996, part A, p. 1; Amy Rinard, “Senate Passes W-2 Plan,” Milwaukee Journal Sentinel, March 14, 1996, part A, p. 1; Interview with Rep. Barbara Notestein, Madison, WI, March 17, 1998. 36. Steven Walters, “ELECTION 98, Garvey Describes How He Would Improve W-2,” Milwaukee Journal Sentinel, October 16, 1998, p. 2. 37. Amy Rinard, “Big Margin Predicted for W-2 Passage,” Milwaukee Journal Sentinel, January 22, 1996, part A, p. 1. 38. Alan Borsuk, “Public Sounds Off on Welfare Plan,” Milwaukee Journal Sentinel, September 30, 1995, part B, p. 3; Amy Rinard, “Foes Say Welfare Reform Plan Will Trap Poor Moms in Poverty,” Milwaukee Journal Sentinel, October 3, 1995, part B, p. 7; Meg Kissinger, “Marchers Rake Thompson Welfare Plan,” Milwaukee Journal Sentinel, November 12, 1995, part B, p. 5; Rinard, “400 Pack Capitol Hearing on W-2”; Meg Kissinger and Jesse Garza, “Black Legislators Attack State’s Welfare Reforms,” Milwaukee Journal Sentinel, March 24, 1996, Part B, p. 3. 39. Amy Rinard, “Welfare Reform Changes Sought,” Milwaukee Journal Sentinel December 21, 1995, part B, p. 5. 40. Jim Stingl, “W-2’s Toughest Test Awaits: Milwaukee,” Milwaukee Journal Sentinel, May 18, 1997, p. 6. 41. Michael Heaney, personal communication, June 2, 2000. Heaney is a graduate student at the University of Chicago who has closely researched the politics of welfare reform in Wisconsin.

304

•

Notes to Chapter 6

42. Bruce Murphy, “The Myth of Welfare Reform,” Milwaukee Magazine, May 1994, p. 118; interview with Pat DeLessio, Milwaukee, WI, March 16, 1998; interview with Senator Gwendolynne Moore, Madison, WI, March 17, 1998. 43. Michael Massing, “The End of Welfare?” New York Review of Books, October 7, 1999, p. 23. 44. Wisconsin Catholic Conference, “Reforming Welfare by Valuing Families,” 3rd draft (Milwaukee: Wisconsin Catholic Conference, January 31, 1995); Steven Walters, “Catholic Group Wants W-2 Changes,” Milwaukee Journal Sentinel, May 15, 1996, part B, p. 5. 45. Frank Aukofer, “Weakland Opposes W-2 Program,” Milwaukee Journal Sentinel, July 5, 1996, part A, p. 1; Richard P. Jones, “Governor Demands Apology from Weakland,” Milwaukee Journal Sentinel, July 9, 1996, part A, p. 1. 46. Kaplan et al., “First Report Form,” p. 32. 47. The chief sources for the following include: Legislative Fiscal Bureau, Wisconsin Works, January 1997; idem, Wisconsin Works, January 2001; and Kaplan, “Wisconsin’s W-2 Program.” I also consulted materials from the Department of Workforce Development web page: http://www: dwd.state.wi.us. 48. To this minor extent, there still is a “dip-down” in W-2. That is, eligibility for support services can be maintained at incomes above the ceiling for initial eligibility. 49. Kaplan, “Wisconsin’s W-2 Program,” p. 95. Data on caseloads is from the Department of Health and Family Services. 50. Department of Health and Social Services, Wisconsin Works, pp. 85–89. 51. Legislative Fiscal Bureau, 1997–99 Wisconsin State Budget: Comparative Summary of Budget Provisions Enacted As 1997 Act 27 (Madison: Legislative Fiscal Bureau, December 1997), vol. 2, pp. 1260–62. 52. Gregory Acs, Norma Coe, Keith Watson, and Robert I. Lerman, Does Work Pay? An Analysis of the Work Incentives under TANF (Washington, DC: Urban Institute, July 1998). 53. A partial exception was that amendments of 1997 and 1999 allowed recipients to combine an unsubsidized part-time job with a prorated CSJ position. See Legislative Fiscal Bureau, Wisconsin Works, January 2001, pp. 9–10. 54. With W-2, the cash benefit generates little disincentive to work, although if one considers other benefits the disincentives are more serious. See Barbara L. Wolfe, “Incentives, Challenges, and Dilemmas of TANF: A Case Study,” Journal of Policy Analysis and Management 21, no. 4 (Fall 2002): 577–86. 55. Legislative Fiscal Bureau, Wisconsin Works, January 1997, p. 51 56. Department of Health and Social Services, Wisconsin Works, p. 72 57. Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1995), p. 24. 58. Legislative Fiscal Bureau, Wisconsin Works, January 1997, pp. 34–35. 59. Ibid., pp. 42–49; Brian K. Bruen and Joshua M. Weiner, Recent Changes in Health Policy for Low-Income People in Wisconsin (Washington, DC: Urban Institute, March 2002), pp. 7–14. 60. Legislative Fiscal Bureau, Wisconsin Works, January 2001, p. 16; Kaplan, “Wisconsin’s W-2 Program,” p. 82.

Notes to Chapter 6

•

305

61. Joel Dresang, “Cash Welfare Caseload Drops 61%,” Milwaukee Journal Sentinel, May 8, 1998, p. 1. 62. DeParle, “Getting Opal Caples to Work,” pp. 54, 59. 63. These federal funds included $318.2 million from the TANF block grant and $53.3 million from the child care and development block grant. See Legislative Fiscal Bureau, Wisconsin Works, January 1997, p. 72. 64. Legislative Fiscal Bureau, Wisconsin Works, January 1997, pp. 23, 61–72. 65. Michael Wiseman, “State Strategies for Welfare Reform: The Wisconsin Story,” Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 537– 38, 541. 66. Wiseman, “State Strategies,” pp. 538–39; James K. Conant, “Wisconsin’s Budget Deficit: Size, Causes, Remedies, and Consequences” (Fairfax, VA: George Mason University, Department of Political Science, 2002); Andrew Reschovsky, “Wisconsin’s Structural Deficit: Our Fiscal Future at the Crossroads,” La Follette Policy Report 13, no. 1 (Spring 2002): 5–7, 21–22. 67. Ehrle et al., Recent Changes in Wisconsin, pp. 13–14. 68. Legislative Fiscal Bureau, 1997–99 Wisconsin State Budget, vol. 2, pp. 1257–58, 1260–62. 69. Bob Lang, “Copayments and Income Eligibility Limit for W-2 Child Care” (Madison: Legislative Fiscal Bureau, June 4, 1997); Legislative Fiscal Bureau, 1999–2001 Wisconsin State Budget: Comparative Summary of Budget Provisions Enacted as 1999 Acts 9 and 10 (Madison: Legislative Fiscal Bureau, January 2000), vol. 2, pp. 1505–1508. 70. Mike Nichols, “Jobs Plan Would Replace AFDC,” Milwaukee Journal-Sentinel, May 28, 1995, part B, p. 3; David R. Riemer, “Welfare Recipients Need Wages, Not Workfare,” New York Times, December 30, 1996, p. A15. 71. Kaplan, “Wisconsin’s W-2 Program,” pp. 90–91. 72. Joel Dresang, Milwaukee Journal Sentinel, “Demonstrators Protest W-2 Training System,” November 7, 1997, p. 3; Richard P. Jones, “Senate Urged to Expand W-2 Learning,” Milwaukee Journal Sentinel, January 29, 1998, p. 1; Mike Johnson and Margo Huston, “W-2 Work or Else: In W-2 and Forced Out of School, ” Milwaukee Journal Sentinel, February 8, 1998, p. 1; Geeta SharmaJensen, “W-2 Work or Else: Businesses Told to Back Training Aid, ” Milwaukee Journal Sentinel, March 12, 1998, Business p. 1; Steven Walters, “Governor’s Vetoes Remold State Budget,” Milwaukee Journal Sentinel, October 12, 1997, part A, p. 1. 73. Governor’s Wisconsin Works (W-2) Education and Training Committee, Step Up: Building a Workforce for the Future (Madison: Department of Workforce Development, June 1998). 74. Legislative Fiscal Bureau, Wisconsin Works, January 2001, pp. 8, 31; Steve Schultze, “Welfare Advocates Criticize Thompson for Vetoing W-2 Changes in Budget,” Milwaukee Journal Sentinel, October 30, 1999, p. 1. 75. I heard remarks to this effect from J. Jean Rogers, Administrator of Economic Support, and Linda Stewart, Secretary of the Department of Workforce Development, at meetings of the Management and Evaluation Project, in Madison, WI, on July 9, 1998, May 16, 2000, and June 27, 2000.

306

•

Notes to Chapter 6

76. L. Jerome Gallagher, Megan Gallagher, Kevin Perese, Susan Schreiber, and Keith Watson, One Year after Welfare Reform: A Description of State Temporary Assistance for Needy Families (TANF) Decisions As of October 1997 (Washington, DC: Urban Institute, June 1998), pp. 33, 39–40, 42; Sheila R. Zedlewski, Pamela A. Holcomb, and Amy-Ellen Duke, Cash Assistance in Transition: The Story of Thirteen States (Washington, DC: Urban Institute, December 1998), p. 46; LaDonna A. Pavetti, “Welfare Policy in Transition: Redefining the Social Contract for Poor Citizens with Children and for Immigrants,” Understanding Poverty, ed. Sheldon H. Danziger and Robert H. Haveman (New York: Russell Sage Foundation; Cambridge, MA: Harvard University Press, 2001), p. 246, says that Utah also has a flat grant, but it was associated with a benefit cut, whereas in Wisconsin benefits were raised. 77. Pamela A. Holcomb, LaDonna Pavetti, Caroline Ratcliffe, and Susan Riedinger, Building an Employment-Focused Welfare System: Work First and Other Work-Oriented Strategies in Five States (Washington, DC: U.S. Department of Health and Human Services, June 1998), p. 77. 78. Douglas J. Besharov and Peter Germanis, “Welfare Reform—Four Years Later,” The Public Interest, no. 140 (Summer 2000): 21–22; Shannon Christian and Rebecca Swartz, eds., W-2 Research Assessment and Direction Conference, Wingspread, Racine, WI, December 10, 1999 (Indianapolis: Hudson Institute, June 2000), p. 33. 79. Heather McCallum, “Welfare As We Know It Now . . . State Approaches to TANF,” paper delivered at the conference of the American Political Science Association, Atlanta, GA, September 2, 1999. 80. Richard P. Nathan and Thomas L. Gais, “Is Devolution Working? Federal and State Roles in Welfare,” Brookings Review 19, no. 3 (Summer 2001): 26–28. 81. Demetra Smith Nightingale and Kelly S. Mikelson, “An Overview of Research Related to Wisconsin Works (W-2)” (Washington, DC: Urban Institute, February 4, 20000); Sheila R. Zedlewski, Pamela A. Holcomb, and Amy-Ellen Duke, Cash Assistance in Transition: The Story of 13 States (Washington, DC: Urban Institute, December 1998), p. 46. 82. Kaplan, “Wisconsin’s W-2 Program,” pp. 98–99; Ehrle, Recent Changes in Wisconsin, p. 12. 83. Remarks at a meeting of the Subcommittee on Business, Welfare Reform, and the Low-Wage Labor Market, Committee for Economic Development, Milwaukee, WI, March 23, 1998, and at a conference on “Next Steps in Welfare Reform,” Manhattan Institute for Policy Research, Washington, DC, April 14, 1999. 84. Thomas Brock and Kristen Harknett, “A Comparison of Two Welfare-toWork Case Management Models,” Social Service Review 72, no. 4 (December 1998): 493–520; Pamela A. Holcomb and Karin Martinson, Implementing Welfare Reform across the Nation (Washington, DC: Urban Institute, August 2002), p. 5. The first paper reporting the Brock and Harknett findings appeared in March 1997. 85. Wiseman, “State Strategies,” pp. 536–37. 86. Bush interview. 87. Interview with Larry Jankowsky, Kenosha, WI, June 11, 1997.

Notes to Chapter 7

•

307

88. Dan Bloom and Charles Michalopoulos, How Welfare and Work Policies Affect Employment and Income: A Synthesis of Research (New York: Manpower Demonstration Research Corporation, May 2001). The recent studies summarized here were from the National Evaluation of Welfare to Work Strategies. 89. Susan Scrivener, Gayle Hamilton, Mary Farrell, Stephen Freedman, Daniel Friedlander, Marisa Mitchell, Jodi Nudelman, and Christine Schwartz, National Evaluation of Welfare-to-Work Strategies: Implementation, Participation Patterns, Costs, and Two-Year Impacts of the Portland (Oregon) Welfare-to-Work Program (Washington, DC: U.S. Department of Health and Human Services and U.S. Department of Education, May 1998). David Greenberg, Karl Ashworth, Andreas Cebulla, and Robert Walker, “When Welfare-to-Work Programs Work Well: Explaining Why Riverside and Portland Shine So Brightly” (Baltimore: University of Maryland at Baltimore County, Department of Economics, May 2002), argue that some part of Portland’s superiority probably reflects relatively advantaged clients. 90. Scrivener et al., Portland (Oregon) Welfare-to-Work Program, pp. 9–12; U.S. Congress, House, Committee on Ways and Means, 2000 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, October 6, 2000), pp. 439–40. CHAPTER 7 IMPLEMENTATING W-2 1. The Family Independence Program in Washington State was a well-funded program that lacked clear impacts, probably because it was voluntary. See Sharon K. Long, Demetra S. Nightingale, and Douglas A. Wissoker, The Evaluation of the Washington State Family Independence Program: Final Report (Washington, DC: Urban Institute, December 1993). 2. Florida’s Project Independence showed only small impacts, probably because child care had to be cut back due to limited funding. See James J. Kemple, Daniel Friedlander, and Veronica Fellerath, Florida’s Project Independence: Benefits, Costs, and Two-Year Impacts of Florida’s JOBS Program (New York: Manpower Demonstration Research Corporation, April 1995). 3. The WIN Demonstration Program in Chicago demanded work search of a large urban caseload but achieved only small impacts, probably because of weak administration. See Daniel Friedlander, Stephen Freedman, Gayle Hamilton, and Janet Quint, Final Report on Job Search and Work Experience in Cook County (New York: Manpower Demonstration Research Corporation, November 1987). 4. Interview with Gerald Whitburn, Tomahawk, WI, March 15, 1998. 5. David E. Osborne and Ted Gaebler, Reinventing Government: How the Entrepreneurial Spirit Is Transforming the Public Sector (Reading, MA: AddisonWesley, 1992). 6. Michael B. Katz, The Price of Citizenship: Redefining the American Welfare State (New York: Metropolitan Books, 2001), chap. 6.

308

•

Notes to Chapter 7

7. Jason DeParle, personal communication, January 4, 2001. I have also heard such comments from Michael Wiseman, a close observer of Wisconsin administrative problems. 8. U.S. General Accounting Office, Welfare Reform: States’ Early Experiences with Benefit Termination (Washington, DC: U.S. General Accounting Office, May 1997), pp. 20–21, 26–28, 48–49, 51–52. The job center and county mentioned in the GAO report were not identified. 9. General Accounting Office, Benefit Termination, pp. 51–52; Jason DeParle, “Cutting Welfare Rolls But Raising Questions,” New York Times, May 7, 1997, pp. A1, B12; Adam Cohen, “The Great American Welfare Lab,” Time, April 21, 1997, pp. 74–76. 10. Legislative Fiscal Bureau, Wisconsin Works: W-2, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1997), pp. 90–93; Jason Turner and Steve Perales, “W2 Implementation in Milwaukee County” (Madison: Department of Health and Social Services, February 20, 1996). 11. Tommy G. Thompson, Power to the People: An American State at Work (New York: HarperCollins, 1996), pp. 195–96. 12. Legislative Fiscal Bureau, Wisconsin Works (W-2) and Other Economic Support Programs, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 2001), p. 4. 13. Suzanne Lynn, Complaint Resolution in the Context of Welfare Reform: How W-2 Settles Disputes (New York: Manpower Demonstration Research Corporation, November 2001); Thomas Kaplan, personal communication, June 10, 2002. 14. Jan Rosenberg and Sol Stern, “America Works: A Venture to End Dependency,” City Journal 3, no. 3 (Summer 1993): 66–71; Eric Pooley, “Nice Work If You Get It,” New York, January 10, 1994, p. 10; Esther B. Fein, “For Job-Finding Concern, a Troubled Past,” New York Times, March 1, 1994, pp. A1, B4; Fred Siegel, “Poverty Empire Strikes Back,” New York Post, March 11, 1994. 15. David Dodenhoff, “Privatizing Welfare in Wisconsin” (Thiensville, WI: Wisconsin Policy Research Institute, January 1998), pp. 7–9. In addition to W-2, the contractors were to run the Food Stamp Employment and Training program. FSET was supposed to put to work recipients of food stamps who were judged to be employable. Traditionally FSET was less important than JOBS or W-2, in part because work tests for cash aid take precedence; FSET served largely single men. 16. Department of Health and Social Services, “JOBS—Job Opportunities and Basic Skills Training Program” (Madison: Department of Health and Social Services, 1990), p. 7; Dodenhoff, “Privatizing Welfare,” p. 5; personnel communications from Ginevra Ewers, Department of Workforce Development, June 8, 2001. Contracts were also made with Indian tribes, but they are ignored here. 17. Jean Rogers, “Selection Criteria for Wisconsin Works (W-2)” (Madison: Department of Health and Social Services, December 27, 1995). 18. Jean Rogers, remarks at meeting of the Management and Evaluation Project, Madison, WI, July 9, 1998. 19. Dodenhoff, “Privatizing Welfare,” p. 12.

Notes to Chapter 7

•

309

20. Ibid., pp. 12–17. 21. Jean Rogers, “Right of First Selection for the Next Wisconsin Works Contracts (Starting January 2000)” (Madison: Department of Workforce Development, November 13, 1998); and Jean Rogers, ”Updated Information on Right of First Selection for the Next W-2 Contracts“ (Madison: Department of Workforce Development, January 26, 1999). 22. “Sixty-two Agencies Qualify for Early W-2 Service Contract Renewal” (Madison: Department of Workforce Development, March 9, 1999); personal communications from Jude Morse, an official of DWD involved in contracting, June 1, 2001, September 27, 2002, and October 1, 2002; “W-2 Agencies Selected for Next Contract Period” (Madison: Department of Workforce Development, September 19, 2001); “Employment Solutions Inc. Will Transition Out of Running W-2 Program” (Madison: Department for Workforce Development, June 7, 2001). 23. Dodenhoff, “Privatizing Welfare,” p. 23. Joel Dresang, “Study Touts Use of Private Agencies,” Milwaukee Journal Sentinel, January 29, 1998, p. 1. 24. The following relies on Thomas Kaplan, Thomas Corbett, and Elizabeth Boehnen, “First Report Form” (Albany: State University of New York, Rockefeller Institute of Government, July 1, 1997), pp. 12, 29, 42–45, 48; and Dodenhoff, “Privatizing Welfare,” p. 19. 25. Jennifer Noyes, remarks and briefing materials at a meeting of the Management and Evaluation Project, Madison, WI, June 27, 2000; Department of Workforce Development, “Performance Standards for the 2001–2002 W-2 and Related Programs Contract” (Madison: Department of Workforce Development, n.d.); Legislative Fiscal Bureau, Wisconsin Works, January 2001, p. 41. 26. Noyes, remarks and briefing materials at MEP; Department of Workforce Development, “Performance Standards for the 2002–2003 W-2 and Related Programs Contract” (Madison: Department of Workforce Development, n.d.). 27. I heard this opinion from Department of Workforce Development officials and Thomas Corbett, a well-informed welfare expert at the University of Wisconsin—Madison, at a meeting of the Management and Evaluation Project, Madison, WI, June 29, 2001. Thomas L. Gais, Richard P. Nathan, Irene Lurie, and Thomas Kaplan, “Implementation of the Personal Responsibility Act of 1996,” in The New World of Welfare: An Agenda for Reauthorization and Beyond, eds. Rebecca M. Blank and Ron Haskins (Washington, DC: Brookings, 2001), pp. 60– 61, find that states have not yet created incentives to tie the funding of agencies to the achievement of the goals of welfare reform. This is just what Wisconsin had done. 28. I heard this account of the county’s motivation from several respondents in my interviews in Milwaukee and elsewhere. 29. As noted earlier, these figures ignore Kinship Care and Caretaker Supplement and thus somewhat overstate the actual decline of cash welfare. 30. Legislative Audit Bureau, Wisconsin Works (W-2) Program: An Evaluation, Report 01–7 (Madison: Legislative Audit Bureau, April 2001), pp. 70–75; Dodenhoff, “Privatizing Welfare,” pp. 12–15.

310

•

Notes to Chapter 7

31. In the same period, child welfare administration was decentralized and privatized in much the same way. See Kristin S. Seefeldt, Laura K. Kaye, Christopher Botsko, Pamela A. Holcomb, Kimura Flores, Carla Herbig, and Karen C. Tumlin, Income Support and Social Services for Low-Income People in Wisconsin (Washington, DC: Urban Institute, December 1998), pp. 44–49. 32. Dodenhoff, “Privatizing Welfare,” pp. 21–23. 33. These impressions are derived from my field interviewing in Milwaukee in 1999. 34. Thomas Kaplan, “Wisconsin’s W-2 Program: Welfare As We Might Come to Know It?” in Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States, ed. Carol S. Weissert (Albany: Rockefeller Institute Press, 2000), pp. 105–107. 35. Jason DeParle, “Wisconsin Welfare Experiment: Easy to Say, Not So Easy to Do,” New York Times, October 18, 1998, pp. 1, 24; idem, “For Caseworker, Helping Is a Frustrating Struggle,” New York Times, December 10, 1999, pp. A1, A26. 36. Michael Wiseman, “State Strategies for Welfare Reform: The Wisconsin Story,” Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 536. 37. Jean Rogers, remarks at MEP, July 9, 1998. 38. Rebecca Swartz, Jacqueline Kauff, Lucia Nixon, Tom Fraker, Jan Hein, and Susan Mitchell, Converting to Wisconsin Works: Where Did Families Go When AFDC Ended in Milwaukee (Indianapolis: Hudson Institute, 1999), pp. 12–13. 39. Department of Workforce Development, “W-2 Enters New, Stable Phase” (Madison: Department of Workforce Development, March 26, 1998). 40. Data from the Department of Workforce Development. 41. Joel Dresang, “W-2 Work or Else: Welfare Rolls Fell Faster in October,” Milwaukee Journal Sentinel, December 16, 1997, p. 1. 42. Mary Beth Murphy, “High Demand May Lead to Waiting List for Kinship Care,” Milwaukee Journal Sentinel, February 11, 1998, p. 1; idem, “Extra Money Found for Kinship Care,” Milwaukee Journal Sentinel, March 3, 1998, p. 4; idem, “Surplus Funds Urged to Bail Out Program,” Milwaukee Journal Sentinel, September 1, 1998, p. 1. 43. Interview with Senator Gwendolynne Moore, Madison, WI, March 17, 1998. 44. Institute for Wisconsin’s Future, “Transitions to W-2: The First Six Months of Welfare Replacement” (Milwaukee: Institute for Wisconsin’s Future, June 1998), pp. 4–7; Lynn, Complaint Resolution. 45. Thomas Kaplan, “Wisconsin Works,” in Managing Welfare Reform in Five States: The Challenge of Devolution, ed. Sarah F. Liebschutz (Albany: Rockefeller Institute Press, 2000), p. 112. 46. DeParle, “Wisconsin’s Welfare Experiment”; Margo Huston, “W-2 Work or Else: County Falls Far Behind on W-2 Child-Care Aid,” Milwaukee Journal Sentinel, November 25, 1997, p. 1; idem, “Audit Rips County’s Day-Care Program,” Milwaukee Journal Sentinel, June 2, 1998, p. 1. 47. Kaplan, “Wisconsin Works,” pp. 110–11, 113–14. 48. Kaplan, “Wisconsin’s W-2 Program,” pp. 99–100.

Notes to Chapter 7

•

311

49. Margo Huston, “State Wants Tougher W-2 Enforcement,” Milwaukee Journal Sentinel, January 22, 1998, p. B1; idem, “W-2 Agencies Promise to Toughen Up,” Milwaukee Journal Sentinel, February 3, 1998, pp. 1A, 6A. 50. Kaplan, “Wisconsin Works,” pp. 115; Thomas Kaplan and Thomas Corbett, with Victoria Mayer, “The Implementation of W-2,” in Maria Cancian and Dan R. Meyer, W-2 Child Support Demonstration Evaluation Phase 1: Final Report (Madison: University of Wisconsin—Madison, Institute for Research on Poverty, April, 2001), vol. 2, tables II.1.5 and II.1.6. 51. Legislative Audit Bureau, Wisconsin Works, p. 54. 52. U.S. General Accounting Office, Welfare Reform: State Sanction Policies and Number of Families Affected (Washington, DC: U.S. General Accounting Office, March 2000), pp. 28–29, 52–55. 53. Legislative Audit Bureau, Wisconsin Works, pp. 62–64. 54. Noyes remarks at MEP, June 27, 2000. 55. Kaplan, “Wisconsin’s W-2 Program,” p. 91. 56. Susan Gooden, Fred Doolittle, and Ben Glispie, Matching Applicants with Services: Initial Assessments in the Milwaukee County W-2 Program (New York: Manpower Demonstration Research Corporation, November 2001), chaps. 1, 3. 57. Margo Huston and Joel Dresang, “W-2 Work or Else: W-2 Cash for Training Called ‘A Step Back,’ ” Milwaukee Journal-Sentinel, September 17, 1998, p. 1. 58. Jean Rogers, remarks and briefing materials at meeting of the Management and Evaluation Project, Racine, WI, December 9, 1999. 59. Kelly S. Mikelson, “Wisconsin Works: Meeting the Needs of Harder to Serve Participants” (Washington, DC: Urban Institute, March 2001); Susan Gooden and Fred Doolittle, Exceptions to the Rule: The Implementation of 24Month Time-Limit Extensions in W-2 (New York: Manpower Demonstration Research Corporation, December 2001). 60. Crocker Stephenson, “Protesters Voice the Problems of W-2 at Capitol,” Milwaukee Journal Sentinel, November 5, 1997, p. 5; Joel Dresang, “Hundreds Hear Women Tell of W-2 Trouble,” Milwaukee Journal Sentinel, December 11, 1997, p. 3; Margo Huston, “Marchers Call for Moratorium on W-2 Sanctions,” Milwaukee Journal Sentinel, April 5, 1998, p.3; Joel Dresang, “Activists March for W-2 Participants,” Milwaukee Journal Sentinel, July 28, 1998, p. 1. 61. Swartz et al., Converting to Wisconsin Works, pp. 23–24. 62. Kaplan and Corbett, “Implementation of W-2,” table II.1.7; Department of Workforce Development, Survey of Those Leaving AFDC or W-2, January to March 1998: Preliminary Report (Madison: Department of Workforce Development, January 13, 1999), p. 11. 63. Joel Dresang and Margo Huston, “W-2 Work or Else: With Welfare Ended, Will Families Be Secure?” Milwaukee Journal Sentinel, September 25, 1998, p. 15. 64. These quotes come from interviews conducted in Milwaukee, on March 19, 1999. 65. Lawrence M. Mead, “Welfare Employment,” in The New Paternalism: Supervisory Approaches to Poverty, ed. Lawrence M. Mead (Washington, DC: Brookings, 1997), pp. 64–67. See chapter 8.

312

•

Notes to Chapter 7

66. Joel Dresang, “ ‘We cannot go back,’ W-2 advocate Riley says he’ll push to refine work-based program,” Milwaukee Journal Sentinel, March 31, 1998, p.12. 67. Margo Huston, “W-2 Work or Else: Community Leaders Call for Fining W-2 Agencies That Are Slow with Child Care Subsidies,” Milwaukee Journal Sentinel, June 19, 1998, p. 3. 68. Jean Rogers, personal communication, August 29 and September 10, 2001; Thomas Corbett, personal communication, June 20, 2002. 69. Legislative Fiscal Bureau, Wisconsin Works, January 2001, pp. 38–41; Noyes, remarks and materials at MEP. These amounts including Indian tribes. 70. Legislative Fiscal Bureau, Wisconsin Works, January 2001, pp. 34–36, 38– 41; Kaplan, “Wisconsin Works,” p. 116. Again, the amounts include Indian tribes. 71. Department of Workforce Development, “Wisconsin Works (W-2) and Related Programs Contract for the period January 1, 2002, through December 31, 2003, Appendix B” (Madison: Department of Workforce Development, n.d.); Jude Morse, personal communication, October 1, 2002. However, for both 2001– 2002 and 2002–2003, counties received small additional appropriations for community investment. 72. Legislative Fiscal Bureau, Wisconsin Works, January 2001, pp. 44–45, 54, 61. 73. Rebecca Swartz, “What Is a ‘Case’ in Post-Reform Wisconsin? Reconciling Caseload with Workload” (Madison: Hudson Institute, Welfare Policy Center, March 2001); David R. Riemer, “Replacing Welfare with Work: The Case for an Employment Maintenance Model,” Focus 16, no. 2 (Winter 1994–1995): 23–30. 74. Key figures were Thomas Corbett and Thomas Kaplan, both of whom had earlier careers in state welfare administration, and Michael Wiseman, an economist at the La Follette Institute of Public Affairs. In this study, I have relied on their research heavily. 75. The following discussion draws on my own experience and on Shannon Christian and Rebecca Swartz, eds., W-2 Research Assessment and Direction Conference, Wingspread, Racine, WI, December 10, 1999 (Indianapolis: Hudson Institute, June 2000), pp. 19–26. 76. The MEP steering committee was chaired first by Jean Rogers, later by DWD Secretary Linda Stewart, and Jennifer Noyes, who succeeded Rogers as head of Economic Support. Michael Wiseman was vice chair and I was the other academic member. The other members were mostly senior research and statistics staff from DWD. 77. Christian and Swartz, eds., W-2 Research Assessment and Direction Conference, pp. 33, 47. 78. 2000 Taubman Center Annual Report (Cambridge, MA: Harvard University, Kennedy School of Government, Taubman Center for State and Local Government, 2000), p. 28. 79. Frances Fox Piven and Richard A. Cloward, Regulating the Poor: The Functions of Public Welfare, updated ed. (New York: Vintage, 1993). 80. Martin Gilens, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy (Chicago: University of Chicago Press, 1999).

Notes to Chapter 8

•

313

CHAPTER 8 PATERNALISM 1. This chapter is largely based on Lawrence M. Mead, “The New Paternalism in Action: Welfare Reform in Wisconsin” (Milwaukee: Wisconsin Policy Research Institute, January 1995), pp. 36–41, and idem, “The Rise of Paternalism” and “Welfare Employment,” both in The New Paternalism: Supervisory Approaches to Poverty. ed. Lawrence M. Mead (Washington, DC: Brookings, 1997), chaps. 1–2. Interviews in the counties occurred in 1994–99. For details, see chapter 5, note 5. 2. Lawrence M. Mead, Beyond Entitlement: The Social Obligations of Citizenship (New York: Free Press, 1986). 3. The individual chapter authors in Mead, ed., New Paternalism, show this in the areas of welfare employment, teen pregnancy prevention, child support, homelessness, drug addiction, and education. 4. Lawrence M. Mead, The New Politics of Poverty: The Nonworking Poor in America (New York: Basic Books, 1992), chaps. 4–7. 5. Jason DeParle, “Cutting Welfare Rolls but Raising Questions,” New York Times, May 7, 1997, p. B12. 6. For two analyses by psychiatrists consistent with this view, see George E. Vaillant, “Poverty and Paternalism: A Psychiatric Viewpoint,” and Miles F. Shore, “Psychological Factors in Poverty,” both in New Paternalism, ed. Mead, chaps. 10–11. For the liberal view that external constraint is the problem, see Thomas J. Kane, “Giving Back Control: Long-Term Poverty and Motivation,” Social Service Review 61, no. 3 (September 1987): 405–19. 7. Marc Cooper, “Overthrowing the Welfare State,” Village Voice, May 31, 1994, p. 36. 8. I believe this conversation occurred in Milwaukee sometime in 1994 or 1995, but I cannot pinpoint the date. 9. For a similar conception of helping the poor as involving “challenge,” see Marvin Olasky, The Tragedy of American Compassion (Wheaton, IL: Crossway Books, 1992), chaps. 1, 6, 12–13. 10. Legislative Fiscal Bureau, Learnfare, Informational Paper #56 (Madison: Legislative Fiscal Bureau, January 1991), p. 1. 11. Craig Gilbert, “Stop Learnfare, Groups Urge US,” Milwaukee Journal, April 4, 1990, Part A, p. 1; George Gerharz, “Wisconsin’s Learnfare: A Bust,” New York Times, January 29, 1990, p. A23; Cyril Ignatius Kendrick, “The Roots of Welfare Reform: The Social Forces Underlying the Wisconsin Learnfare Program,” Ph.D. diss. (Blacksburg, VA: Virginia Polytechnic Institute, Department of Sociology, March 1993), pp. 213–23. 12. Marcus E. Ethridge and Stephen L. Percy, “A New Kind of Public Policy Encounters Disappointing Results: Implementing Learnfare in Wisconsin,” Public Administration Review 53, no. 4 (July/August 1993): 340–47. 13. Isabel Wilkerson, “Wisconsin Makes Truancy Costly by Tying Welfare to Attendance,” New York Times, December 11, 1989, p. B12. 14. Legislative Audit Bureau, An Evaluation of the Learnfare Program: Final Report (Madison: Legislative Audit Bureau, April 1997), pp. 57–59; idem, “An

314

•

Notes to Chapter 8

Evaluation of Learnfare Case Management Services” (Madison: Legislative Audit Bureau, February 9, 1994), pp. 19–23. 15. Associated Press “Better Record Keeping Reduces Milwaukee County Sanctions,” Milwaukee Journal, February 12, 1991, part B, p. 3; idem, “Learnfare Penalties Decreasing,” Milwaukee Sentinel, June 3, 1991, part 1, p. 10. 16. Department of Health and Social Services, Learnfare Case Management Manual (Madison: Department of Health and Social Services, May 1994), pp. 1, 7. 17. Legislative Audit Bureau, “Evaluation of Learnfare Case Management Services,” pp. 25–43. 18. Ibid., pp. 25–26, 28; Thomas A. Van Eck, “Kenosha County Learnfare Enhanced Services Program: Evaluation Summary Report” (Kenosha: Kenosha County Department of Social Services, December 1993). 19. John Pawasarat, Lois M. Quinn, and Frank Stetzer, “Evaluation of the Impact of Wisconsin’s Learnfare Experiment on the School Attendance of Teenagers Receiving Aid to Families with Dependent Children” (Milwaukee: University of Wisconsin—Milwaukee, Employment and Training Institute, February 1992); Legislative Audit Bureau, Evaluation of the Learnfare Program. 20. Legislative Fiscal Bureau, Welfare Reform Initiatives, Informational Paper #47 (Madison: Legislative Fiscal Bureau, January 1995), pp. 8–9. 21. Legislative Audit Bureau, An Evaluation of Effects of Learnfare Expansion (Madison: Legislative Audit Bureau, April 1997). 22. Johannes M. Bos and Veronica Fellerath, LEAP: Final Report on Ohio’s Welfare Initiative to Improve School Attendance Among Teenage Parents (New York: Manpower Demonstration Research Corporation, August 1997); David Long, Robert G. Wood, and Hilary Kopp, LEAP: The Educational Effects of LEAP and Enhanced Services in Cleveland (New York: Manpower Demonstration Research Corporation, October 1994). 23. Ronald B. Mincy and Elaine J. Sorenson, “Deadbeats and Turnips in Child Support Reform,” Journal of Policy Analysis and Management 17, no. 1 (Winter 1998): 44–51; Maureen B. Waller and Robert Plotnick, “Effective Child Support Policy for Low-Income Families: Evidence from Street Level Research,” Journal of Policy Analysis and Management 20, no. 1 (Winter 2001): 89–110. 24. Dan Bloom and Kay Sherwood, Matching Opportunities to Obligations: Lessons for Child Support Reform from the Parents’ Fair Share Pilot Phase (New York: Manpower Demonstration Research Corporation, April 1994), chap. 1; Fred Doolittle and Suzanne Lynn, Working with Low-Income Cases: Lessons for the Child Support Enforcement System from Parents’ Fair Share (New York: Manpower Demonstration Research Corporation, May 1998), pp. 1–2, 4–6, 37. 25. Legislative Fiscal Bureau, 1993–95 Wisconsin State Budget: Comparative Summary of Senate Bill 44, Enacted As 1993 Act 16 (Madison: Legislative Fiscal Bureau, November 1993), pp. 459–60; personal communication, Department of Workforce Development, August 16, 2002. 26. Joe Klein, “ ‘Make the Daddies Pay,’ ” Newsweek, June 21, 1993, p. 33. 27. Office of Policy and Budget, “Children First: Community Work Experience Program for Non-Custodial Parents” (Madison: Department of Health and Social

Notes to Chapter 8

•

315

Services, May 1991), pp. 10–11, 13–17; John Wagner, “The Children First Program: A Report to the Administrator” (Madison: Department of Health and Social Services, Division of Economic Support, June 1993), pp. 5–8; Sandra Cleveland, “Process Evaluation of Wisconsin’s Children First Program” (Madison: Department of Health and Social Services, December 22, 1995), pp. 22–27. 28. Sandra Cleveland, “Summary of 1994 Children First Data” (Madison: Department of Workforce Development, July 17, 1996), pp. 7–8; Ron Blasco, Children First Program: Final Evaluation Report (Madison: Department of Workforce Development, November 2000), pp. 11–20. 29. Blasco, Children First, pp. 21–36. 30. Cleveland, “Process Evaluation,” pp. 17–18. 31. Office of Policy and Budget, “Children First,” pp. 8, 21. 32. Tommy G. Thompson, Power to the People: An American State at Work (New York: HarperCollins, 1996), p. 56. 33. Fred Doolittle, Virginia Knox, Cynthia Miller, and Sharon Rowser, Building Opportunities, Enforcing Obligations: Implementation and Interim Impacts of Parents’ Fair Share (New York: Manpower Demonstration Research Corporation, December 1998); John M. Martinez, and Cynthia Miller, Working and Earning: The Impact of Parents’ Fair Share on Low-Income Fathers’ Employment. (New York: Manpower Demonstration Research Corporation, October 2000). 34. Jay Hein, “Non-custodial Parents: What’s Next for Wisconsin?” (Madison: Hudson Institute, June 29, 2001), pp. 7, 13—14, 23. 35. Joel F. Handler and Ellen Jane Hollingsworth, The “Deserving Poor”: A Study of Welfare Administration (New York: Academic Press, 1971). 36. Department of Health and Social Services, “JOBS—Job Opportunities and Basic Skills Training Program” (Madison: Department of Health and Social Services, 1990), p. 2. 37. Mark Lincoln Chadwin, John J. Mitchell, and Demetra Smith Nightingale, “Reforming Welfare: Lessons from the WIN Experience,” Public Administration Review 41, no. 3 (May/June 1981): 375–76; Eugene Bardach, “Implementing a Paternalist Welfare-to-Work Program,” in The New Paternalism, ed. Mead, chap. 8; Howard S. Bloom, Carolyn J. Hill, and James Riccio, “Linking Program Implementation and Impacts: Lessons from a Pooled Sample of Welfare-to-Work Experiments,” Journal of Policy Analysis and Management 22, no. 4 (Fall 2003). 38. Legislative Fiscal Bureau, Aid to Families with Dependent Children, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 1991), p. 29. 39. Grant County officials prefer to call their style “parental” instead of paternalistic. By this they mean a combination of the “maternal” (nurturance) and the “paternal” (accountability). Personal communication from Deb Hughes, August 13, 2001. 40. James Riccio, Daniel Friedlander, and Stephen Freedman, GAIN: Benefits, Costs, and Three-Year Impacts of a Welfare-to-Work Program (New York: Manpower Demonstration Research Corporation, September 1994); Susan Scrivener, Gayle Hamilton, Mary Farrell, Stephen Freedman, Daniel Friedlander, Marisa Mitchell, Jodi Nudelman, and Christine Schwartz, National Evaluation of Welfare-to-Work Strategies: Implementation, Participation Patterns, Costs, and Two-

316

•

Notes to Chapter 8

Year Impacts of the Portland (Oregon) Welfare-to-Work Program (Washington, DC: U.S. Department of Health and Human Services and U.S. Department of Education, May 1998). 41. The following is based on Lawrence M. Mead, “The Twilight of Liberal Welfare Reform,” The Public Interest, no. 139 (Spring 2000), pp. 22–34. 42. Thomas Brock, Fred Doolittle, Veronica Fellerath, and Michael Wiseman, Creating New Hope: Implementation of a Program to Reduce Poverty and Reform Welfare (New York: Manpower Demonstration Research Corporation, October 1997), pp. 5–8; Thomas Kaplan and Ingrid Rothe, “New Hope and W-2: Common Challenges, Different Responses” (Madison: University of Wisconsin— Madison, Institute for Research on Poverty, May 1999). 43. Brock et al., Creating New Hope, chaps. 4–5; Johannes M. Bos, Aletha C. Huston, Robert C. Granger, Greg J. Duncan, Thomas W. Brock, and Vonnie C. McLoyd, New Hope for People with Low Incomes: Two-Year Results of a Program to Reduce Poverty and Reform Welfare (New York: Manpower Demonstration Research Corporation, August 1999), chap. 3. 44. Bos et al., New Hope, chaps. 4–7. 45. Ibid., pp. 48–54, 96–97. 46. Department of Health and Social Services, Wisconsin Works: 1999 Plan, Draft (Madison: Department of Health and Social Services, April 1995), pp. 13– 14; Andrew Bush, “Replacing Welfare in Wisconsin,” Public Welfare 54, no. 2 (Spring 1996): 19. 47. Thomas Kaplan, “Wisconsin’s W-2 Program: Welfare As We Might Come to Know It?” in Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States, ed. Carol S. Weissert (Albany: Rockefeller Institute Press, March 2000), p. 101. 48. Jason DeParle, “Early Sex Abuse Hinders Many Women on Welfare,” New York Times, November 28, 1999, pp. 1, 28; Joel Dresang, “W-2: Work in Progress: Third of three parts,” Milwaukee Journal Sentinel, December 14, 1999, p. 1; Pamela A. Holcomb and Karin Martinson, Implementing Welfare Reform across the Nation (Washington, DC: Urban Institute, August 2002). 49. Jean Rogers, remarks at meetings of the Management and Evaluation Project, Madison, WI, July 9, 1998, and Racine, WI, December 9, 1999. 50. Legislative Fiscal Bureau, Wisconsin Works (W-2) and Other Economic Support Programs, Informational Paper #45 (Madison: Legislative Fiscal Bureau, January 2001), p. 6. 51. Mary C. Rowin, “Incorporating the Informed Choice Philosophy into W-2” (Madison: Department of Workforce Development, July 8, 2002). 52. Kaplan and Rothe, “New Hope and W-2,” p. 11. 53. Interview with George Leutermann at Maximus, Milwaukee, WI, March 18, 1999. Admittedly, these staff also administered other programs such as community reinvestment and Children First. 54. Tom Corbett, personal communication, April 4, 2001. 55. Neil Gilbert, “Why the New Workfare Won’t Work,” Commentary, May 1994, pp. 47–49.

Notes to Chapter 9

•

317

56. Grant County Department of Social Services, “The Secret to Welfare Reform” (Lancaster, WI: Grant County Department of Social Services, April 20, 2001), pp. 3, 10. 57. Ibid., pp. 5–8. 58. Mead, “Rise of Paternalism,” pp. 24–27. CHAPTER 9 THE DECLINE OF WELFARE 1. This chapter is substantially based on Lawrence M. Mead, “The Decline of Welfare in Wisconsin” (Milwaukee: Wisconsin Policy Research Institute, March 1996); idem, “The Decline of Welfare in Wisconsin,” Journal of Public Administration Research and Theory 9, no. 4 (October 1999): 597–622; idem, “Caseload Change: An Exploratory Study,” Journal of Policy Analysis and Management 19, no. 3 (Summer 2000): 465–72; idem, “Welfare Caseload Change: An Alternative Approach,” Policy Studies Jorunal 31, no. 2 (2003): 163–85. Interested readers should consult these papers for data sources and answers to technical questions. 2. Caseload data from the U.S. Administration for Children and Families web page, last updated May 23, 2002. 3. Rebecca M. Blank, “What Causes Public Assistance Caseloads to Grow?” (Evanston, IL: Institute for Policy Research, Northwestern University, September 1997); Geoffrey Wallace and Rebecca M. Blank, “What Goes Up Must Come Down? Explaining Recent Changes in Public Assistance Caseloads,” in Economic Conditions and Welfare Reform, ed. Sheldon H. Danziger (Kalamazoo, MI: Upjohn Institute for Employment Research, 1999), pp. 49–89; Congressional Budget Office, “Forecasting AFDC Caseloads, with an Emphasis on Economic Factors” (Washington, DC: Congressional Budget Office, July 1993). 4. Jason DeParle, “A Sharp Decrease in Welfare Cases Is Gathering Speed,” New York Times, February 2, 1997, pp. 1, 18; idem, “Success and Frustration, As Welfare Rules Change,” New York Times, December 30, 1997, p. A1, A16. 5. John H. Bishop, “Is Welfare Reform Succeeding?” (Ithaca: Cornell University, School of Industrial and Labor Relations, August 14, 1998); Council of Economic Advisors, “The Effects of Welfare Policy and the Economic Expansion on Welfare Caseloads, An Update” (Washington, DC: Executive Office of the President, August 3, 1999); David T. Ellwood, “The Impact of the Earned Income Tax Credit and Social Policy Reforms On Work, Marriage, and Living Arrangements” (Cambridge, MA: Harvard University, Kennedy School of Government, November 1999); June E. O’Neill and M. Anne Hill, “Gaining Ground? Measuring the Impact of Welfare Reform on Welfare and Work” (New York: Manhattan Institute, Center for Civic Innovation, July 2001). 6. U.S. Congress, House, Committee on Ways and Means, 2000 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, October 6, 2000), pp. 808–12. 7. Council of Economic Advisors, “Explaining the Decline in Welfare Receipt, 1993–1996” (Washington, DC: Executive Office of the President, May 9, 1997);

318

•

Notes to Chapter 9

James P. Ziliak, David N. Figlio, Elizabeth E. Davis, and Laura S. Connolly, “Accounting for the Decline in AFDC Caseloads: Welfare Reform or Economic Growth” (Madison: University of Wisconsin, Institute for Research on Poverty, November 1997). 8. Calculated from data from the March 1994 Current Population Survey, table 19; the March 2000 Current Population Survey, table 17; and the March 2002 Annual Demographic Survey, table 17. 9. Bruce D. Meyer and Dan T. Rosenbaum, “Welfare, the Earned Income Tax Credit, and the Labor Supply of Single Mothers” (Cambridge, MA: National Bureau of Economic Research, September 1999). 10. Robert Moffitt, “Incentive Effects of the U.S. Welfare System: A Review,” Journal of Economic Literature 30, no. 1 (March 1992): 13, 17–18. Studies of EITC suggest that it does raise work levels, but to my eye not by enough so that it could account for the large recent rise in work levels. See V. Joseph Hotz and John Karl Scholz, “The Earned Income Tax Credit” (Cambridge: National Bureau of Economic Research, January 2001), pp. 39–49. 11. Michael Wiseman, “State Strategies for Welfare Reform: The Wisconsin Story,” Journal of Policy Analysis and Management 15, no. 4 (Fall 1996): 528– 29, 542. 12. Ziliak et al., “Accounting for the Decline in AFDC Caseloads,” pp. 21–27. 13. Calculated from unemployment data in U.S. Department of Commerce, Bureau of the Census, Statistical Abstract of the United States 1992 (Washington, DC: U.S. Government Printing Office, 1992), p. 402, and U.S. Department of Commerce, Bureau of the Census, Statistical Abstract of the United States 1994 (Washington, DC: U.S. Government Printing Office, September 1994), p. 419. 14. My own time series analysis of the Wisconsin caseload decline also found unemployment to be the main detectable cause, but there were anomalies and much that was unexplained. See Mead, “Decline of Welfare,” Journal of Public Administration Research and Theory, pp. 604–606. 15. These counties were Douglas, Fond du Lac, Grant, Marathon, and Winnebago. During these interviews, I also asked about the policy issues discussed in other chapters. In other interviewing in the state, respondents sometimes offered opinions about the caseload decline, but I exclude those responses here. Sixty respondents were interviewed, 21 at DHSS in Madison and 39 in the counties. The vast majority were public officials, but a few worked for community groups that were contractors to welfare or JOBS. In one case I interviewed two officials together, and did not record who said what; I treat them as a single respondent. 16. This because my models are simple cross-sections. I explain change over time only in the sense that some terms incorporate time in the way they are calculated. The economic studies are mostly pooled or panel studies with an explicit time dimension. 17. Bob Lang, “1989–91 Budget: Health and Social Services—Public Assistance—Federal Welfare Reform, AFDC Employment and Training Programs” (Madison: Legislative Fiscal Bureau, May 3, 1989), p. 3.

Notes to Chapter 9

•

319

18. Harry J. Holzer and Michael A. Stoll, “What Happens When Welfare Recipients Are Hired?” (Washington, DC: Georgetown University, Georgetown Public Policy Institute, May 2000). 19. This is consistent with the finding of Wallace and Blank, “What Goes Up Must Come Down,” p. 63, that for a state to have a Republican governor depresses caseloads. 20. Stephen H. Bell, “Why Are Welfare Caseloads Falling?” (Washington, DC: Urban Institute, March 2001), pp. 30–32, notes that waivers have anticipatory effects on caseloads in several of the economic studies. 21. One theory might be that Wisconsin drove its rolls down simply by making it tougher for people to establish eligibility for welfare. However, no administrator we interviewed in the state mentioned such a policy, and no Thompson initiative has been aimed at toughening eligibility determination other than by establishing conduct requirements. The state outperformed most states in the accuracy of its grant payment in both 1986 and 1994. According to federal statistics; it denied aid improperly in only 708 cases in 1994. See U.S. Congress, House, Committee on Ways and Means, Overview of Entitlement Programs: 1990 Green Book: Background Material and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, June 5, 1990), pp. 1237–40; and U.S. Congress, House, Committee on Ways and Means, 1998 Green Book: Background Material, and Data on Programs Within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, May 19, 1998), pp. 465–72. 22. A recent exception is Chien-Chung Huang, James Kunz, and Irwin Garfinkel, “The Effect of Child Support on Welfare Exits and Re-Entries,” Journal of Policy Analysis and Management 21, no. 4 (Fall 2002): 557–76. As do I, the authors find an important child support effect on caseloads. 23. Department of Workforce Development, Welfare Closure Study: A Study of All AFDC and W2 Closures (September 1997 through September 1999) (Madison: Department of Workforce Development, March 2000), p. 12. 24. There was, however, one mention of increased disregards for dependent care. 25. Caseload data from the Department of Health and Social Services. 26. Results are limited to 1994 because this was the last year that the JOBS reporting categories had a consistent meaning across the state. After that, Work First and other more initiatives create variations not captured by the reporting data. JOBS first reported data in 1991. Due to a computer change in 1994, I use JOBS data for 1991–93. 27. There was enough theoretical guidance to specify the likely dimensions of the model (those given just above), but not the specific measures. So I tried various measures for each dimension and assembled those that were significant at .10 or better. Because of this fitting procedure, only the general model is verified, not this specific result. In another run of data, the significant variables might well differ, but the same general dimensions would tend to emerge. 28. “Active” refers to a series of assignments that make definite demands on recipients to participate and to work, look for work, or undergo education or

320

•

Notes to Chapter 10

training, while “inactive” means several administrative statuses that exempt clients from such demands. 29. Mead, “Decline of Welfare,” Journal of Public Administration Research and Theory, pp. 610–12. 30. Despite its large size relative to other counties, Milwaukee did not appear as an outlier in any of the regressions. Nor was a dummy for Milwaukee significant in any of them, nor did the omission of this case change the results much. See Mead, “Decline of Welfare,” pp. 611–12. 31. This measure comes from Jeffrey L. Brudney, F. Ted Hebert, and Deil S. Wright, “Reinventing Government in the American States: Measuring and Explaining Administrative Reform,” Public Administration Review 59, no. 1 (January/February 1999): 19–30. 32. Blank, “What Causes Public Assistance Caseloads to Grow,” pp. 13–21; Congressional Budget Office, “Forecasting AFDC Caseloads.” 33. Citizen’s Conference on State Legislatures, State Legislatures: An Evaluation of Their Effectiveness (New York: Praeger, 1971), table 3. Here and below, I reverse variables expressed as ranks so that higher, rather than lower, is better. 34. Thomas L. Gais, “Concluding Comments: Welfare Reform and Governance,” in Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States, ed. Carol S. Weissert, (Albany: Rockefeller Institute Press, 2000), pp. 174–78. 35. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984), chap. 5. If substituted for individualism, the measure for whether a state was moralistic neared significance (p = .115) and was negative, as expected. CHAPTER 10 THE EFFECTS OF REFORM 1. Data from the Department of Health and Family Services. Data limitations prevent saying more about the caseload trend. 2. John Pawasarat and Lois M. Quinn, “Impact of Welfare Reform on Child Care Subsidies in Milwaukee County, 1996–1999” (Milwaukee: University of Wisconsin–Milwaukee, Employment and Training Institute, October 1999). 3. University of Wisconsin (Milwaukee) Center for Economic Development, “Support Utilization Among Head Start Parents in Wisconsin” (Milwaukee: University of Wisconsin Center for Economic Development, November 1999). 4. U.S. Department of Health and Human Services, Office of Inspector General, Welfare Administrative Costs (Washington, DC: U.S. Department of Health and Human Services, February 1995), table 8; U.S. Department of Health and Human Services, Administration for Children and Families, Time Trends, Fiscal Years 1985–1994, p. 7. 5. Richard Jones, “Welfare Mothers Deprived of Advocate,” Milwaukee Journal, July 10, 1994, p. B5; Bruce Murphy, “The Myth of Welfare Reform,” Milwaukee Magazine, May 1994, p. 118; State Senator Gwendolynne S. Moore, press release, July 11, 1994.

Notes to Chapter 10

•

321

6. Governor Tommy Thompson and Dr. William Bennett, “The Good News About Welfare Reform: Wisconsin’s Success Story” (Washington, DC: Heritage Foundation, March 6, 1997), pp. 7–9. 7. Although this is the most thorough accounting I know of Wisconsin’s welfare spending, it is still incomplete. Figures are confined to programs that were within the Department of Health and Social Services in 1994. Federal project aids, block grant aids, program aids, and local assistance had to be excluded from the upper panel of table 10.2 due to missing data. However, these programs do not figure in the lower panel. 8. Separate budget figures on family cases under Medicaid are unavailable. 9. Neither James K. Conant, “Wisconsin’s Budget Deficit: Size, Causes, Remedies, and Consequences” (Fairfax, VA: George Mason University, Department of Political Science, 2002), nor Andrew Reschovsky, “Wisconsin’s Structural Deficit: Our Fiscal Future at the Crossroads,” La Follette Policy Report 13, no. 1 (Spring 2002): 5–7, 21–22, who analyze the crisis, mentions welfare reform as a cause. 10. John Pawasarat and Lois M. Quinn, “Wisconsin Welfare Employment Experiments: An Evaluation of the WEJT and CWEP Programs” (Milwaukee: University of Wisconsin–Milwaukee, Employment and Training Institute, September 1993), chap. 11; Wisconsin Department of Workforce Development, “100Hour Rule Final Report” (Madison: Department of Workforce Development, September 27, 1996); Maximus, Evaluation of the AFDC Two-Tier Benefit Demonstration Project: Final Report (Boston: February 1998). 11. According to Maria Cancian, Thomas Kaplan, and Ingrid Rothe, with Hwa-Ok Park and Dan Ross, “Wisconsin’s Self-Sufficiency First/Pay for Performance Program: Results and Lessons from a Social Experiment” (Madison: Institute for Research on Poverty, 1999), the evaluation of PFP showed a significant effect on the chance that recipients would leave welfare and remain off. 12. Maria Cancian and Dan R. Meyer, W-2 Child Support Demonstration Evaluation Phase 1: Final Report (Madison: University of Wisconsin—Madison, Institute for Research on Poverty, April 2001), vol. 1, pp. 34–49, 62–69. This evaluation was required by Washington because the child support pass-through still required a federal waiver. 13. Burt S. Barnow, Thomas Kaplan, and Robert A. Moffitt, eds., Evaluating Comprehensive State Welfare Reform: The Wisconsin Works Program (Albany: Rockefeller Institute of Government, 2000). 14. Data from the U.S. Administration for Children and Families. 15. Joel Dresang, “W-2 Work or Else: Surprising 30% of W-2 participants working unsubsidized,” Milwaukee Journal Sentinel, May 22, 1998, p. 1; Jean Matheson, “W-2 Work or Else: Survey Indicates Program is Successful,” Milwaukee Journal Sentinel, February 23, 1999, Waukesha edition, p. 3; Maria Cancian and Daniel R. Meyer with Chi-Fang Wu, “Program Participation of Mothers on W-2,” in Cancian and Meyer, W-2 Child Support Demonstration Evaluation, vol. 2, chap. 4; Maria Cancian and Robert Haveman, “Patterns of Labor Market Performance among Low-Income Wisconsin Single Mothers,” in Cancian and Meyer, W-2 Child Support Demonstration Evaluation, vol. 2, chap. 5, p. 5. 16. Department of Workforce Development, Survey of Those Leaving AFDC or W-2: January to March 1998: Preliminary Report (Madison: Department of

322

•

Notes to Chapter 10

Workforce Development, January 13, 1999), pp. 5–6; idem, Wisconsin Works Leavers Survey: Those Who Left W-2 Cash Assistance, April 1998 through December 1998: Final Report (Madison: Department of Workforce Development, November 2001), pp. 147–48; idem, Welfare Closure Study: A Study of all AFDC and W2 Closures (September 1997 through September 1999) (Madison: Department of Workforce Development, March 2000), pp. 8–10. The first leavers survey document covers departures in early 1998, while the second summarizes three surveys covering departures later in 1998. 17. Rebecca Swartz, Jacqueline Kauff, Lucia Nixon, Tom Fraker, Jan Hein, and Susan Mitchell, Converting to Wisconsin Works: Where Did Families Go When AFDC Ended in Milwaukee (Indianapolis: Hudson Institute, 1999), pp. 14, 20; Department of Workforce Development, Welfare Closure Study, p. 16. 18. Department of Workforce Development, Survey of Those Leaving AFDC or W-2, p. 7; idem, Wisconsin Works Leavers Survey, pp. 28–29, 42–43; Swartz et al., Converting to Wisconsin Works, pp. 25–26; Maria Cancian, Robert Haveman, Daniel R. Meyer, and Barbara Wolfe, “Before and After TANF: The Economic Well-Being of Women Leaving Welfare” (Madison: University of Wisconsin–Madison, Institute for Research on Poverty, February 2002), p. 10. The Swartz study surveyed not leavers but a sample of cases that went through the transition to W-2, only some of whom left welfare. The proportion working in regular jobs was 72 percent of the leavers from this sample, but 41 percent of the entire sample; 69 percent of the entire sample worked either inside or outside W-2. “Leavers” here excludes people who shifted from AFDC or W-2 to SSI or Kinship Care. 19. Department of Workforce Development, Survey of Those Leaving AFDC or W-2, p. 14; idem, Wisconsin Works Leavers Survey, pp. 78–80; Swartz et al., Converting to Wisconsin Works, p. 17; Maria Cancian, Robert Haveman, Thomas Kaplan, Daniel Meyer, Ingrid Rothe, and Barbara Wolfe, with Sandra Barone, “The Take-Up of Medicaid and Food Stamps by Welfare Leavers: The Case of Wisconsin” (Madison: University of Wisconsin—Madison, Institute for Research on Poverty, June 2001), pp. 28–31. 20. Sammis B. White and Lori A. Geddes, “Economic Lessons for Welfare Mothers” (Thiensville, WI: Wisconsin Policy Research Institute, February 2001), pp. 5–7. Some of these gains may represent a “welfare rebound” effect—the normal tendency of many people who go on aid to recover economically over time. 21. Department of Workforce Development, Survey of Those Leaving AFDC or W-2, p. 7; idem, Wisconsin Works Leavers Survey, p. 44; Cancian and Haveman, “Patterns of Labor Market Performance,” p. 5; White and Geddes, “Economic Lessons,” pp. 7–8. 22. Maria Cancian and Daniel R. Meyer, “Mothers’ Income and Economic Well-Being,” in Cancian and Meyer, W-2 Child Support Demonstration Evaluation, vol. 2, chap. 6, pp. 1–11. 23. Department of Workforce Development, Survey of Those Leaving AFDC or W-2, p. 12; idem, Wisconsin Works Leavers Survey, pp. 117, 119; Swartz et al., Converting to Wisconsin Works, pp. 34–35. 24. Legislative Audit Bureau, Wisconsin Works (W-2) Program: An Evaluation (Madison: Legislative Audit Bureau, April 2001), pp. 43–45; Cancian and

Notes to Chapter 10

•

323

Meyer, “Mothers’ Income and Economic Well-Being,” p. 23; Cancian et al., “Before and After TANF,” pp. 6–7, 24. 25. White and Geddes, “Economic Lessons for Welfare Mothers,” pp. 7–8; Swartz et al., Converting to Wisconsin Works, p. 29. 26. U.S. Congress, 2000 Green Book, pp. 394–96; Irving Piliavin, Mark Courtney, and Amy Dworsky, “Early Transfers from AFDC to W-2: The Experiences of 100 Dane County Families” (Madison: University of Wisconsin—Madison, Institute for Research on Poverty, February 2000), p. 21. 27. Department of Workforce Development, Welfare Closure Study, pp. 13– 14. This calculation treated child and health care costs as income, which some will find unrealistic. 28. Cancian and Meyer, W-2 Child Support Demonstration Evaluation, vol. 1, pp. 84–87; Arthur Reynolds and Barbara Wolfe, “Child Well-Being among W-2 Families,” in Cancian and Meyer, W-2 Child Support Demonstration Evaluation, vol. 2, chap. 9, pp. 2–8, 36–37. 29. Department of Workforce Development, Wisconsin Works Leavers Survey, pp. 129–30; Gayle Hamilton, with Stephen Freedman and Sharon M. McGruder, National Evaluation of Welfare-to-Work Strategies: Do Mandatory Welfare-toWork Programs Affect the Well-Being of Children? A Synthesis of Child Research Conducted As Part of the National Evaluation of Welfare-to-Work Strategies (Washington, DC: U.S. Department of Health and Human Services and U.S. Department of Education, June 2000). The New Hope Project, in Milwaukee, also recorded positive effects on children in school. See Johannes M. Bos, Aletha C. Huston, Robert C. Granger, Greg J. Duncan, Thomas W. Brock, and Vonnie C. McLoyd, New Hope for People with Low Incomes: Two-Year Results of a Program to Reduce Poverty and Reform Welfare (New York: Manpower Demonstration Research Corporation, August 1999), chap. 7. 30. Virginia Knox, Cynthia Miller, and Lisa A. Gennetian, Reforming Welfare and Rewarding Work: A Summary of the Final Report on the Minnesota Family Investment Program (New York: Manpower Demonstration Research Corporation, 2000); Pamela A. Morris, Aletha C. Huston, Greg J. Duncan, Danielle A. Crosby, and Johannes Bos, How Welfare and Work Policies Affect Children: A Synthesis of Research (New York: Manpower Demonstration Research Corporation, March 2001). 31. Jason DeParle, “Cutting Welfare Rolls but Raising Questions,” New York Times, May 7, 1997, pp. A1, B12; idem, “As Welfare Rolls Shrink, Load on Relatives Grows,” New York Times, February 21, 1999, pp. 1, 24; Department of Workforce Development, Wisconsin Works Leavers Survey, pp. 115, 118. 32. Cancian and Meyer, “Mothers’ Income and Economic Well-Being,” pp. 11–23. 33. Department of Workforce Development, Wisconsin Works Leavers Survey, pp. 100–13. The interviews were done 12–22 months after the respondents left welfare in 1998. 34. Mark E. Courtney, Irving Piliavin, and Peter Power, “Involvement of TANF Applicants with Child Protective Services” (Madison: University of Wisconsin— Madison, Institute for Research on Poverty, July 2001), p. 3; May Beth Murphy,

324

•

Notes to Chapter 10

“Child Abuse Reports Reach High in County,” Milwaukee Journal Sentinel, January 26, 1998, p. 3; Nina Bernstein, “Foster Care System Wary of Welfare Cuts,” New York Times, November 19, 1995, pp. 1, 26; Frances Fox Piven, “Thompson’s Easy Ride, The Nation, February 26, 2001, pp. 4–5. 35. U.S. Congress, 2000 Green Book, pp. 708–709; Jennifer Ehrle, Kristin Seefeldt, Kathleen Snyder, and Pat McMahon, Recent Changes in Wisconsin Welfare and Work, Child Care, and Child Welfare Systems (Washington, DC: Urban Institute, September 2001), pp. 15–16; Courtney et al., “Involvement of TANF Applicants with Child Protective Services,” pp. 18–24. 36. Meg Kissinger and Georgia Pabst, “Hunger, Homelessness Rise Sharply,” Milwaukee Journal Sentinel, November 26, 1996, part A, p. 1; Edward Walsh, “Homelessness and Hunger Rise in Milwaukee: Is Welfare Reform to Blame?” Washington Post, December 27, 1996. Many later articles with these allegations appear in the Milwaukee Journal Sentinel throughout 1997–99. 37. Department of Workforce Development, Wisconsin Works Leavers Survey, p. 107. 38. De Parle, “Cutting Welfare Rolls,” p. B12. 39. Kristin S. Seefeldt, Laura K. Kaye, Christopher Botsko, Pamela A. Holcomb, Kimura Flores, Carla Herbig, and Karen C. Tumlin, Income Support and Social Services for Low-Income People in Wisconsin (Washington, DC: Urban Institute, December 1998), p. 50; David Dodenhoff and Susan Vergeront, Homeless Families in Milwaukee after Welfare Reform (Milwaukee: Center for SelfSufficiency, 1999); Thomas Kaplan and Daniel R. Meyer, “Toward a Basic Impact Evaluation of the Wisconsin Works Program,” in Evaluating Comprehensive State Welfare Reform, ed. Barnow, Kaplan, and Moffitt, pp. 153–54. 40. U.S. General Accounting Office, Welfare Reform: States’ Early Experiences With Benefit Termination (Washington, DC: U.S. General Accounting Office, May 1997), chap. 3. 41. Department of Workforce Development, Survey of Those Leaving AFDC or W-2, p. 10; idem, Survey of Those Leaving W-2 Assistance Who Did Not Return, April through June, 1998: Preliminary Report (Madison: Department of Workforce Development, May 2000), pp. 15–16; Douglas J. Besharov and Peter Germanis, “Welfare Reform—Four Years Later,” The Public Interest, no. 140 (Summer 2000): 27–32. 42. General Accounting Office, States’ Early Experiences With Benefit Termination, pp. 33–34; Legislative Audit Bureau, Wisconsin Works, pp. 46–52. 43. Swartz et al., Converting to Wisconsin Works, pp. 37–38; Amy L. Sherman, “The Lessons of W-2,” The Public Interest, no. 140 (Summer 2000): 44–48. 44. Christine Devere, “Welfare Reform Research: What Do We Know About Those Who Leave Welfare?” (Washington, DC: Congressional Research Service, March 13, 2001). Gregory Acs and Pamela Loprest, with Tracy Roberts, “Final Synthesis Report of Findings from ASPE ‘Leavers’ Grants” (Washington, DC: Urban Institute, November 27, 2001), report similar results. 45. Jason DeParle, “Wisconsin Welfare Overhaul Justifies Hope and Some Fear,” New York Times, January 15, 1999, pp. A1, A21; Sarah Brauner and Pamela Loprest, “Where Are They Now? What States’ Studies of People Who Left Welfare Tell Us” (Washington, DC: Urban Institute, May 1999), pp. 6–7.

Notes to Chapter 10

•

325

46. National Center for Children in Poverty, “Child Poverty in the States: Levels and Trends from 1970 to 1998” (New York: Columbia University, Mailman School of Public Policy, 2000). 47. The Census shows a Wisconsin child poverty rate of 16 percent in 1993, whereas the National Center for Children in Poverty has 19 percent. The difference probably reflects different estimating procedures using the same CPS data. 48. Annie E. Casey Foundation, Kids Count Date Book 2002: State Profiles of Child Well-Being (Baltimore: Annie E. Casey Foundation, 2002), pp. 46–56, 157. In 1999, Wisconsin ranked seventeenth or better on all of these indicators, except for child death rate, where it was thirtieth. Its overall ranking was seventh in 1990, eleventh in 1999. 49. U.S. Congress, 2000 Green Book, pp. 411, 416–17. The shift is somewhat exaggerated because non-TANF caseloads such as Wisconsin’s Kinship Care are excluded. 50. Deborah A. Ellwood and Donald J. Boyd, “Changes in State Spending on Social Services since the Implementation of Welfare Reform: A Preliminary Report” (Albany: Rockefeller Institute of Government, February 2000). 51. Cancian et al., “Before and After TANF,” p. 30. Wisconsin is toward the high end of the ever-worked range mentioned in summaries of leavers studies— 63–91 percent in Devere, “Welfare Reform Research,” and 57–90 percent in Acs and Loprest, “Final Synthesis Report.” 52. The other states were Alabama, California, Colorado, Florida, Massachusetts, Michigan, Minnesota, Mississippi, New Jersey, New York, Texas, and Washington. 53. Urban Institute, Snapshots of America’s Families II: A View of the Nation and 13 States from the National Survey of America’s Families (Washington, DC: Urban Institute, October 24, 2000). The children’s result may appear to conflict with the result, reported above, that poor children were unusually sick in Wisconsin. The reason probably is that the latter figure was for children under poverty and on welfare, where this one is for all children under 200 percent of poverty. 54. Calculated from caseload data from the Department of Health and Social Services. 55. Department of Workforce Development, “Welfare Closure Study,” pp. 14–15. 56. Harry J. Holzer and Michael A. Stoll, “What Happens When Welfare Recipients Are Hired?” (Washington, DC: Georgetown University, Georgetown Public Policy Institute, May 2000). 57. Jason DeParle, “Early Sex Abuse Hinders Many Women on Welfare,” New York Times, November 28, 1999, pp. 1, 28; idem, “Bold Effort Leaves Much Unchanged for the Poor,” New York Times, December 30, 1999, pp. A1, A20–21. 58. Ellen K. Scott, Kathryn Edin, Andrew S. London, and Joan Maya Mazelis, “My Children Come First: Welfare-Reliant Women’s Post-TANF Views of WorkFamily Trade-offs and Marriage” (Kent, OH: Kent State University, Department of Sociology, n.d.). The weighting of the benefits of working relative to the costs seems more positive in these interviews than in similar ethnography in Kathryn Edin and Laura Lein, Making Ends Meet: How Single Mothers Survive Welfare

326

•

Notes to Chapter 11

and Low-Wage Work (New York: Russell Sage Foundation, 1997), where the research dates from the early 1990s. 59. Michele Derus, “Making the Bus Means Making a Living on W-2,” Milwaukee Journal Sentinel, June 28, 1998, pp. 1, 12. 60. Jason DeParle, “On a Once Forlorn Avenue, Tax Preparers Now Flourish,” New York Times, March 21, 1999, pp. 1, 20. 61. Sheila R. Zedlewski and Donald W. Alderson, “Before and After Reform: How Have Families Changed?” (Washington, DC: Urban Institute, April 2001); Blaine Harden, “2-Parent Families Rise After Change in Welfare Laws,” New York Times, August 12, 2001, pp. 1, 30. 62. Jason DeParle, “Better Work Than Welfare—But What If There’s Neither?” New York Times Magazine, December 18, 1994, pp. 58, 74. CHAPTER 11 WELFARE REFORM AND GOOD GOVERNMENT 1. This chapter draws on Lawrence M. Mead, “Welfare Reform: The Institutional Dimension,” Focus 22, no. 1 (Special Issue 2002): 39–46; idem, “Welfare Caseload Change: An Alternative Approach,” Policy Studies Journal 31, no. 2 (2003) 163–85; and “State Governmental Capacity and Welfare Reform,” paper presented at the annual conference of the Association for Public Policy Analysis and Management, Washington, DC, November 7, 2003. 2. Eugene Bardach, Improving the Productivity of JOBS Programs (New York: Manpower Demonstration Research Corporation, December 1993); Robert D. Behn, Leadership Counts: Lessons for Public Managers from the Massachusetts Welfare, Training, and Employment Program (Cambridge, MA: Harvard University Press, 1991); Donald F. Norris and Lyke Thompson, eds., The Politics of Welfare Reform (Thousand Oaks, CA: Sage, 1995). 3. On benefit levels or spending, see Richard E. Dawson and James A. Robinson, “Inter-Party Competition, Economic Variables, and Welfare Policies in the American States,” Journal of Politics 25, no. 2 (May 1963): 265–89; Charles F. Cnudde and Donald J. McCone, “Party Competition and Welfare Policies in the American States,” American Political Science Review 63, no. 3 (September 1969): 858–66; Lawrence E. Gary, “Policy Decisions in the Aid to Families with Dependent Children Program: A Comparative State Analysis,” Journal of Politics 35, no. 4 (November1973): 886–923; Edward T. Jennings, Jr., “Competition, Constituencies, and Welfare Policies in American States,” American Political Science Review 73, no. 2 (June 1979): 414–29; Jeff Stonecash and Susan W. Hayes, “The Sources of Public Policy: Welfare Policy in the American States,” Policy Studies Journal 9, no. 5 (Spring 1981): 681–98; Russell L. Hanson, “The ‘Content’ of Welfare Policy: The States and Aid to Families with Dependent Children,” Journal of Politics 45, no. 3 (August 1983): 771–85; Robert D. Plotnick and Richard F. Winters, “A Politico-Economic Theory of Income Redistribution,” American Political Science Review 79 (no. 2, June 1985): 458–73; Jack Tweedie, “Resources Rather than Needs: A State-Centered Model of Welfare Policymaking,” American Journal of Political Science 38, no. 3 (August 1994): 651–72; and Charles Barrilleaux, Thomas Holbrook, and Laura Langer, “Electoral Competi-

Notes to Chapter 11

•

327

tion, Legislative Balance, and American State Welfare Policy,” American Journal of Political Science 46, no. 2 (April 2002): 415–27. On welfare experimentation and TANF decisions, see Robert C. Lieberman and Greg M. Shaw, “Looking Inward, Looking Outward: The Politics of State Welfare Innovation under Devolution,” Political Research Quarterly 53, no. 2 (June 2000): 215–40; Yvonne Zylan and Sarah A. Soule, “Ending Welfare As We Know It (Again): Welfare State Retrenchment, 1989–1995,” Social Forces 79, no. 2 (December 2000): 623–52; and Joe Soss, Sanford P. Schram, Thomas P. Vartanian, and Erin O’Brien, “Setting the Terms of Relief: Explaining State Policy Choices in the Devolution Revolution,” American Journal of Political Science 45, no. 2 (April 2001): 378–95. 4. Citizens’ Conference on State Legislatures, State Legislatures: An Evaluation of Their Effectiveness (New York: Praeger, 1971); Samuel C. Patterson, “Legislative Politics in the States,” in Politics in the American States: A Comparative Analysis, 6th ed., ed. Virginia Gray and Herbert Jacob (Washington, DC: Congressional Quarterly, 1996), pp. 175–76; John G. Grumm, “Structural Determinants of Legislative Output,” in Legislatures in Developmental Perspective, ed. Allan Kornberg and Lloyd D. Musolf (Durham, NC: Duke University Press, 1970), chap. 12. 5. Lee Sigelman, “The Quality of Administration: An Exploration in the American States,” Administration and Society 8, no. 1 (May 1976): 107–44; Charles Barrileaux, Richard Feiock, and Robert E. Crew, Jr., “Measuring and Comparing American State Administrative Characteristics,” State and Local Government Review 24 (Winter 1992): 12–18; Katherine Barrett and Richard Greene, “Grading the States: A Management Report Card,” Governing 12, no. 5 (February 1999): 17–90; Katherine Barrett and Richard Greene, with Michele Mariani, “Grading the States 2001: A Management Report Card,” Governing 14, no. 5 (February 2001): 20–108; Ann O’M. Bowman and Richard C. Kearney, “Dimensions of State Government Capability,” Western Political Quarterly 41 (June 1988): 341–62. 6. Jack L. Walker, “The Diffusion of Innovations Among the American States,” American Political Science Review 63, no. 3 (September 1969): 980–99; Virginia Gray, “Innovation in the States: A Diffusion Study,” American Political Science Review 67, no. 4 (December 1973): 1174–85; Robert L. Savage, “Policy Innovativeness As a Trait of the American States,” Journal of Politics 4, no. 1 (February 1978): 212–24. 7. Samuel C. Patterson, “The Political Culture of the American States,” Journal of Politics 30, no. 1 (February 1968): 187–209; Harry Eckstein, “A Culturalist Theory of Political Change,” American Political Science Review 82, no. 3 (September 1988): 789–804. 8. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984), chap. 5. 9. Charles A. Johnson, “Political Culture in the American States: Elazar’s Formulation Examined,” American Journal of Political Science 20 (August 1976): 491–99; David R. Morgan and Sheilah Watson, “Political Culture, Political Systems Characteristics, and Public Policies among the American States,” Publius 21 (Spring 1991): 31–43; Joel Lieske, “Regional Subcultures in the United States,”

328

•

Notes to Chapter 11

Journal of Politics 55, no. 4 (November 1993): 888–913. I prefer Elazar’s schemes to these later formulations because it is better known, corresponds well with my field observations of various states, and is statistically powerful. 10. Robert S. Erikson, Gerald C. Wright, and John P. McIver, Statehouse Democracy: Public Opinion and Policy in the American States (Cambridge, England: Cambridge University Press, 1993), pp. 62–72, 150–76, 246. 11. Johnson, “Political Culture in the American States,” pp. 499–507; Morgan and Watson, “Political Culture,” pp. 36–47; Ira Sharkansky, “The Utility of Elazar’s Political Culture: A Research Note,” Polity 2, no. 1 (Fall 1969): 66–83; Jody L. Fitzpatrick and Rodney E. Hero, “Political Culture and Political Characteristics of the American States: A Consideration of Some Old and New Questions,” Western Political Quarterly 41, no. 1 (March 1988): 145–53. 12. Richard A. Joslyn, “Manifestations of Elazar’s Political Subcultures: State Public Opinion and the Content of Political Campaign Advertising,” Publius 10, no. 2 (Spring 1980): 37–58, generally confirms the expectation, while Peter F. Nardulli, “Political Cultures in the American States: An Empirical Examination of Elazar’s Formulation,” American Politics Quarterly 18, no. 3 (July 1990): 287– 315, and Timothy D. Schiltz and R. Lee Rainey, “The Geographic Distribution of Elazar’s Political Subcultures among the Mass Population,” Western Political Quarterly 31, no. 3 (September 1978): 410–15, question it. For a rebuttal of Schiltz and Rainey, see Robert L. Savage, “Looking for Political Subcultures: A Critique of the Rummage Sale Approach,” Western Political Quarterly 34, no. 2 (June 1981): 331–36. 13. John Kincaid, “Political Cultures of the American Compound Republic,” Publius 10, no. 2 (Spring 1980): 6–11, 13–14; Susan Welch and John G. Peters, “State Political Culture and the Attitudes of State Senators toward Social, Economic Welfare, and Corruption Issues,” Publius 10, no. 2 (Spring 1980): 59–67. 14. Tweedie, “Resources Rather than Needs”; Eric M. Uslaner and Ronald E. Weber, “The ‘Politics’ of Redistribution: Toward a Model of the Policy-Making Process in the American States,” American Politics Quarterly 3, no. 2 (April 1975): 130–70. 15. Rodney E. Hero and Caroline J. Tolbert, “A Racial/Ethnic Diversity Interpretation of Politics and Policy in the States of the U.S.,” American Journal of Political Science 40, no. 3 (August 1996): 851–71. 16. Elazar, American Federalism, pp. 141–42. 17. Robert D. Putnam, with Robert Leonardi and Raffaella Y. Nanetti, Making Democracy Work: Civic Traditions in Modern Italy (Princeton: Princeton University Press, 1993). 18. Tom W. Rice and Alexander F. Sumberg, “Civic Culture and Government Performance in the American States,” Publius 27, no. 1 (Winter 1997): 99–114. Because the performance measure stresses the liberalism of policy, New York, California, and Massachusetts rank highest on it, and Wisconsin only nineteenth. 19. Steven M. Teles, Whose Welfare? AFDC and Elite Politics (Lawrence, KS: University Press of Kansas, 1996), chap. 7. 20. Data from the U.S. Administration for Children and Families. I regressed a state’s number of waivers on dummy variables for whether it was moralistic

Notes to Chapter 11

•

329

or individualistic (traditionalistic was the reference category). Moralism had a coefficient of 1.2, significant at .096; individualism was nonsignificant. 21. Barbara Goldman, Daniel Friedlander, and David Long, Final Report on the San Diego Job Search and Work Experience Demonstration (New York: Manpower Demonstration Research Corporation, February 1986); Gayle Hamilton and Daniel Friedlander, Final Report on the Saturation Work Initiative Model in San Diego (New York: Manpower Demonstration Research Corporation, November 1989); James Riccio, Daniel Friedlander, and Stephen Freedman, GAIN: Benefits, Costs, and Three-Year Impacts of a Welfare-to-Work Program (New York: Manpower Demonstration Research Corporation, September 1994). 22. Michael Wiseman, ed., “Research and Policy: A Symposium on the Family Support Act of 1988,” Journal of Policy Analysis and Management 10, no. 4 (Fall 1991): 588–666. 23. Gayle Hamilton, Interim Report on the Saturation Work Initiative Model in San Diego (New York: Manpower Demonstration Research Corporation, August 1988), pp. 15–17, 83–84; Riccio et al., GAIN, pp. 46–55; Bardach, Improving the Productivity of JOBS Programs. 24. LaDonna A. Pavetti, Pamela Holcomb, and Amy-Ellen Duke, Increasing Participation in Work and Work-Related Activities: Lessons from Five State Welfare Reform Demonstration Projects, Final Report (Washington, DC: Urban Institute, September 1995); Gary C. Bryner, “Welfare Reform in Utah” (Albany: Rockefeller Institute of Government, 1997). 25. Susan Scrivener, Gayle Hamilton, Mary Farrell, Stephen Freedman, Daniel Friedlander, Marisa Mitchell, Jodi Nudelman, and Christine Schwartz, National Evaluation of Welfare-to-Work Strategies: Implementation, Participation Patterns, Costs, and Two-Year Impacts of the Portland (Oregon) Welfare-to-Work Program (Washington, DC: U.S. Department of Health and Human Services and U.S. Department of Education, May 1998), chap. 2; Eugene Bardach, Getting Agencies to Work Together: The Practice and Theory of Managerial Craftsmanship (Washington, DC: Brookings, 1998), chap. 7; Gayle Hamilton, Stephen Freedman, Lisa Gennetian, Charles Michalopoulos, Johanna Walter, Diana Adams-Ciardullo, Anna Gassman-Pines, Sharon McGroder, Martha Zaslow, Surjeet Ahluwalia, and Jennifer Brooks, with Electra Small and Bryan Ricchetti, National Evaluation of Welfare-to-Work Strategies: How Effective Are Different Welfare-to-Work Approaches? Five-Year Adult and Child Impacts for Eleven Programs (New York: MDRC, November 2001). Chapter 6 mentioned Portland as a model that was somewhat less work first than W-2, and perhaps preferable. 26. Los Angeles shifted toward work first in the mid–1990’s and achieved improved results. See Stephen Freedman, Jean Tansey Knab, Lisa A. Gennetian, and David Navarro, The Los Angeles Jobs-First GAIN Evaluation: Final Report on a Work First Program in a Major Urban Center (New York: Manpower Demonstration Research Corporation, June 2000). 27. Daniel Friedlander, Stephen Freedman, Gayle Hamilton, and Janet Quint, Final Report on Job Search and Work Experience in Cook County (New York: Manpower Demonstration Research Corporation, November 1987); James J. Kemple, Daniel Friedlander, and Veronica Fellerath, Florida’s Project Indepen-

330

•

Notes to Chapter 11

dence: Benefits, Costs, and Two-Year Impacts of Florida’s JOBS Program (New York: Manpower Demonstration Research Corporation, April 1995). 28. Behn, Leadership Counts; M. Anne Hill and Thomas J. Main, Is Welfare Working? The Massachusetts Reforms Three Years Later (Boston: Pioneer Institute for Public Policy Research, 1998), chap. 4; Melissa Buis, “Taking States Seriously: Welfare Reform Implementation in Massachusetts and Oregon,” Ph.D. diss. (Brandeis University, 2001), chap. 4. E.T. claimed positive effects, but its evaluation was nonexperimental, hence less authoritative than the MDRC studies. See Demetra Smith Nightingale, Douglas A. Wissoker, Lynn C. Burbridge, D. Lee Bawden, and Neal Jeffries, Evaluation of the Massachusetts Employment and Training (ET) Program (Washington, DC: Urban Institute Press, 1991). 29. This section is based largely on Lawrence M. Mead, “Kicking New York’s Dependency Habit,” City Journal 1, no. 4 (Summer 1991): 41–49; Sarah F. Liebschutz, “Welfare Reform in New York: A Mixed Laboratory for Change,” Managing Welfare Reform in Five States: The Challenge of Devolution. ed. Sarah F. Liebschutz (Albany: Rockefeller Institute Press, 2000), chap. 4; Harvey Catchen, “Ending Welfare As We Know It,” in Governing New York State, 4th ed., ed. Jeffrey M. Stonecash (Albany: State University of New York Press, 2001), chap. 15; and Jay Hein, “New York Reformed: How New York Was Transformed from U.S. Welfare Capital to Reform’s Epicentre,” in Douglas J. Besharov, Peter Germanis, Jay Hein, Donald K. Jonas, and Amy L. Sherman, Ending Dependency: Lessons from Welfare Reform in the USA (London: Civitas, May 2001), pp. 82– 101. I also draw on my experience consulting with welfare officials about reform in New York and Wisconsin. 30. Stonecash, ed., Governing New York State. For example, inability to agree on how to implement TANF cost the state $1 million a day in federal welfare funding in late 1997. See David Firestone, “New York Losing Millions for Lack of Welfare Plan,” New York Times, October 9, 1996. 31. Harvey Catchen, “Ending Welfare As We Know It,” in Governing New York State, ed. Stonecash, chap. 15. In 1995, New York did require counties to impose workfare in general assistance, a program of local aid. Some counties, particularly New York City, later applied the same policy to AFDC, but this was not widely agreed upon. See Susan A. Riedinger, Laudan Y. Aron, Pamela J. Loprest, and Carolyn T. O’Brien, Income Support and Social Services for LowIncome People in New York (Washington, DC: Urban Institute, February 1999), p. 20. 32. See, instance, Heather Mac Donald, “Welfare Reform Discoveries,” City Journal 7, no. 1 (Winter 1997): 14–31. City Journal is published by the Manhattan Institute, a prominent conservative think tank in the city. 33. New York City’s Human Resources Administration was able to unify case management only in 2000. This it did by merging the old eligibility and employment workers after strenuous negotiations with the affected unions. Personal communication from Mark Hoover, who was deputy commissioner of HRA under Jason Turner, January 10, 2003.

Notes to Chapter 11

•

331

34. Sheila R. Zedlewski, Pamela A. Holcomb, and Amy-Ellen Duke, Cash Assistance in Transition: The Story of Thirteen States (Washington, DC: Urban Institute, December 1998), pp. 6, 40, 44–45. 35. Besides New York and Wisconsin, states where I have research or consulting experience include California, Illinois, Maryland, Massachusetts, Missouri, New Jersey, Pennsylvania, Rhode Island, South Carolina, and West Virginia. 36. The UI sample includes Alabama, California, Colorado, Florida, Massachusetts, Michigan, Minnesota, Mississippi, New Jersey, New York, Texas, Washington, and Wisconsin. The RIG sample includes Arizona, California, Colorado, Florida, Georgia, Kansas, Michigan, Minnesota, Mississippi, Missouri, New Jersey, New York, North Carolina, Ohio, Rhode Island, Tennessee, Texas, Utah, Washington, West Virginia, and Wisconsin. UI used research teams that visit states from Washington. RIG used political scientists who are local to each state. 37. The ratings I gave were less subjective than they may seem. I first read the cases in 2001, then reassessed the ratings in May 2002 and again in May 2003. Each assessment involved six ratings of twenty-four states, or 144 decisions. At the first rerating, I changed three decisions, at the second, two. Effects on the quantitative results reported below were negligible. I also asked Alan Weil and Thomas Gais, the leaders of the ANF and SCS projects, to review my ratings, and they did not differ with them. This does not imply that they would necessarily make the same ratings if they reviewed the same documents in the same detail that I did, only that they did not find my judgments unreasonable based on their knowledge. I emphasize that the criteria I use and the conclusions drawn from them are entirely my own; they should not be attributed to Gais or Weil, or to their projects, funders, or institutes. 38. The ANF documents are on the Urban Institute web page. Some of the RIG studies appeared in Liebschutz, ed., Managing Welfare Reform in Five States, and Carol S. Weissert, ed., Learning from Leaders: Welfare Reform Politics and Policy in Five Midwestern States (Albany, NY: Rockefeller Institute Press, 2000). For a complete listing, see Mead, “State Government Capacity.” 39. The relationship of cultural type with political performance alone was X2 = 12.1, p = .061; with administrative performance alone, X2 = 10.0, p = .124. The ordinal correlation between the two dimensions was Gamma = .54, p < .014; Tau b = .39, p < .019. 40. Compared to Ohio or Colorado, one lacks the sense that New York took any important decisions about AFDC at the state level. Rather, it agreed to let counties decide. There were state-level policies, but they dated from the past and did not embody fresh decisions. 41. This pattern recalls the factional style that V. O, Key, Jr., attributed to southern politics years ago in Southern Politics in State and Nation (New York: Vintage Books, 1949). One might object that the policymaking behind the Wisconsin reform was also highly casual and personalized, with little policy analysis behind it, as noted in chapter 2. This, however, was true mostly during the earliest phase of reform, when there was little research or experience to go on. The later

332

•

Notes to Chapter 11

evolution of JOBS (see chapters 4–5) was closely based on experience, and the planning of W-2 (chapter 6) was unusually systematic. By the mid–1990s, much more evaluation information existed about “what works” in welfare reform, and a casual, personal style is no longer so understandable. 42. This analysis is indebted to Dennis O. Grady, “State Government Administrative Capacity and Welfare Reform Implementation,” paper presented at the annual meeting of the American Political Science Association, Atlanta, GA, September 2–5, 1999. 43. TANF does have measures of the job entry rate, job retention by recipients who have gone to work, and their wage rate, which it uses to award “high performance” bonuses to some states. Wisconsin won high performance bonuses in 1999 and 2000. But these measures are available only on states that apply for the bonuses, which omits several of them. 44. In the model for 1989–94, when I withdrew the reinvention of government term, one case had to be dropped to avoid heteroscedasticity. This made the variables for child support enforcement and the percent of population on AFDC in 1989 nonsignificant, so they too were dropped. For 1994–98, when I withdrew legislative effectiveness and individualism, I had to omit a different case to avoid heteroscedasticity, and case size was omitted as nonsignificant. 45. To conserve degrees of freedom, I tested indicators only for Elazar’s three principal cultures, not for the subcultures he defined within them. I also preferred to represent the cultures using dummy variables rather than the single scale developed by Sharkansky in “The Utility of Elazar’s Political Culture.” The cultures are nominal categories that cannot validly be arrayed on a single scale. In this I follow Johnson, “Political Culture in the American States,” p. 499; Kincaid, “Political Cultures,” p. 7; and Elazar himself; see Daniel J. Elazar, “Steps in the Study of American Political Culture,” Publius 10, no. 2 (Spring 1980): 225–26, 229–30. 46. Since Sharkansky’s version of Elazar links moralism and traditionalism, its expected tie to caseload change is unclear. In my results it operates largely as a measure of traditionalism. 47. These are Pearson correlation coefficients, which vary from .00 to 1.00. A positive sign indicates that the variables vary in the same direction; when one increases, so does the other. A negative sign means that they vary inversely; more of one means less of the other. 48. Partisanship is negative because many southern states are Democratic yet conservative. Democratic party control (ranney1) is also negative, probably for similar reasons. 49. For a parallel analysis covering a wider range of programs, see Marcia K. Meyer, Janet C. Gornick, and Laura R. Peck, “Packaging Support for Low-Income Families: Policy Variation across the United States,” Journal of Policy Analysis and Management 20, no. 3 (Summer 2001): 457–83. 50. Martin Gilens, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy (Chicago: University of Chicago Press, 1999), chaps. 2, 8. This and other studies of the public opinion surrounding poverty are based on national samples. It is possible that attitudes may differ importantly among states

Notes to Chapter 12

•

333

of the different Elazar cultures. But as mentioned above, the major differences on welfare appear rather to be among elites. CHAPTER 12 ORIGINS OF EXCELLENCE 1. Wisconsin is below the Midwest in support for welfare, but this result does not appear valid. Wisconsin, a high-benefit state, scores forty-third among the states on this indicator. Other high-benefit states (California, New York) also score unexpectedly low. The correlation between the welfare opinion indicator and benefit levels is actually negative. This may be because, as Brace et al. admit (pp. 181–82), the welfare indicator may not be stable. 2. I base this comparison on interviewing of social welfare personnel in New York, Wisconsin, and other states. 3. Paul Boyer, Urban Masses and Moral Order in America, 1820–1920 (Cambridge, MA: Harvard University Press, 1978). 4. Interview with former Rep. Dismas Becker, Madison, WI, June 13, 1997; interview with Sen. Gary George, Madison, WI, March 19, 1998. 5. Interview with Rep. Antonio Riley, Milwaukee, WI, March 16, 1998; interview with former Sen. Joseph Strohl, Madison, WI, June 12, 1997. 6. Dennis Coyle and Aaron Wildavsky, “Requisites of Radical Reform: Income Maintenance Versus Tax Preferences,” Journal of Policy Analysis and Management 7, no. 1 (Fall 1987): 1–16; Myron Magnet, The Dream and the Nightmare: The Sixties’ Legacy to the Underclass (New York: William Morrow, 1993). 7. Charles Noble, Welfare As We Knew It: A Political History of the American Welfare State, New York: Oxford University Press, 1997; David Brian Robinson and Dennis R. Judd, The Development of American Public Policy: The Structure of Policy Restraint (Glenview, IL: Scott, Foresman, 1989). 8. Seymour Martin Lipset, “Radicalism or Reformism: The Sources of Working-Class Politics,” American Political Science Review 77, no. 1 (March 1983): 1– 18; Louis Hartz, The Liberal Tradition in America: An Interpretation of American Political Thought since the Revolution (New York: Harcourt, Brace & World, 1955); Richard Hofstadter, The American Political Tradition and the Men Who Made It (New York: Knopf, 1948). 9. Thomas Ertman, Birth of the Leviathan: Building States and Regimes in Medieval and Early Modern Europe (Cambridge, England: Cambridge University Press, 1997). 10. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper & Row, 1984), pp. 122–31. 11. David Hackett Fischer, Albion’s Seed: Four British Folkways in America (New York: Oxford University Press, 1989). 12. State of Wisconsin, 2001–2002 Blue Book (Madison: Wisconsin Legislature, 2001), pp. 684–85. 13. Leon D. Epstein, Politics in Wisconsin (Madison: University of Wisconsin Press, 1958), pp. 17–18, 30; Robert C. Nesbit, Wisconsin: A History (Madison: University of Wisconsin Press, 1973), p. 150.

334

•

Notes to Chapter 12

14. Nesbit, Wisconsin, p. 180; John H. Fenton, Midwest Politics (New York: Holt, Rinehart, and Winston, 1966), pp. 6–7, 47, 67–68. 15. Richard Hofstadter, The Age of Reform: From Bryan to F. D. R. (New York: Knopf, 1955); Grant McConnell, Private Power and American Democracy (New York: Vintage, 1970). 16. Nesbit, Wisconsin, chaps. 25–26. 17. Ibid., chaps. 15–24; David P. Thelan, The New Citizenship: Origins of Progressivism in Wisconsin, 1885–1900 (Columbia, MO: University of Missouri Press, 1972), chaps. 1–3. 18. Thelan, New Citizenship, chaps. 4–13. 19. Russel B. Nye, Midwestern Progressive Politics: A Historical Study of Its Origins and Development, 1870–1951 (East Lansing, MI: Michigan State College Press, 1951), p. 217. 20. Epstein, Politics in Wisconsin, pp. 23, 27, 77–160. 21. Fenton, Midwest Politics, p. 68. 22. Richard F. Winters, “The Politics of Taxing and Spending,” in Politics in the American States: A Comparative Analysis, 6th ed., ed. Virginia Gray and Herbert Jacob (Washington, DC: Congressional Quarterly, 1996), pp. 322–24. 23. Epstein, Politics in Wisconsin, chaps. 3, 5–8; William F. Thompson, The History of Wisconsin: Continuity and Change, 1940–1965 (Madison: State Historical Society of Wisconsin, 1988), p. 448. 24. Wilder Crane and A. Clark Hagensick, Wisconsin Government and Politics (Madison: University of Wisconsin—Extension, Institute of Governmental Affairs, 1976), p. 8.2; Epstein, Politics in Wisconsin, p. 30. 25. Richard C. Elling, “Bureaucracy: Maligned Yet Essential,” in Politics in the American States, ed. Gray and Jacob, p. 299. 26. Nesbit, Wisconsin, pp. 508, 532. 27. Citizen’s Conference on State Legislatures, State Legislatures: An Evaluation of Their Effectiveness (New York: Praeger, 1971). 28. Joel Dresang and Dave Umhoefer, “Pioneers, Then and Now, at the Social Frontier,” Milwaukee Journal, August 29, 1999, pp. 45, 55. 29. Virginia Gray, “Innovation in the States: A Diffusion Study,” American Political Science Review 67, no. 4 (December 1973); Christopher Howard, “Social Policies in the American States, 1900–1935: A Statistical Analysis” (Williamsburg, VA: College of William and Mary, Department of Government, August 2000). 30. Dresang and Umhoefer, “Pioneers,” p. 55; Brian K. Bruen and Joshua M. Weiner, Recent Changes in Health Policy for Low-Income People in Wisconsin (Washington, DC: Urban Institute, March 2002). 31. Jay P. Greene, Paul E. Peterson, and Jiangtao Du, “School Choice in Milwaukee: A Randomized Experiment,” in Learning from School Choice, ed. Paul E. Peterson and Bryan C. Hassel (Washington, DC: Brookings, 1998), chap. 13; Jay P. Greene and Chester E. Finn, Jr., “Grading States on Their Education Freedom,” Wall Street Journal, September 19, 2000, p. A26. 32. Robertson and Judd, Development of American Public Policy, pp. 374–75.

Notes to Chapter 12

•

335

33. Nesbit, Wisconsin, p. 532. 34. Timothy Conlan, From New Federalism to Devolution: Twenty-Five Years of Intergovernmental Reform (Washington, DC: Brookings, 1998), pp. 278–81. 35. L. Christopher Plein and David G. Williams, “First Report Form: West Virginia” (Albany: Rockefeller Institute of Government, 1997), pp. 10–11. 36. Nye, Midwestern Progressive Politics, chap. 4. 37. Nesbit, Wisconsin, p. 426–27; Irwin Garfinkel, Assuring Child Support: An Extension of Social Security (New York: Russell Sage Foundation, 1992), p. ix. 38. Executive Office of the President, Annual Report of the Council of Economic Advisors (Washington, DC: U.S. Government Printing Office, 1964), chap. 2. 39. Institute for Research on Poverty: A History (Madison: University of Wisconsin—Madison, Institute for Research on Poverty, 1995–96). In 1996, most of the grant was shifted to a consortium of the University of Chicago and Northwestern University. In 2002, IRP lost its remaining funds as a national poverty research center but received a grant for regional research on poverty. The two review panels, on which I served, felt that IRP’s research too heavily stressed statistical analysis of national data bases. It had also given insufficient attention to the political and administrative factors in antipoverty policy. 40. Lawrence M. Mead, Beyond Entitlement: The Social Obligations of Citizenship (New York: Free Press, 1986), chaps. 10–11. 41. Thomas Hobbes, Leviathan (Oxford: Blackwell, 1946), pp. 138–39, 143, 227. 42. Jean-Jacques Rousseau, Political Writings, ed. and trans. Frederick Watkins (Edinburgh: Thomas Nelson, 1953), p. 169. 43. Michael J. Sandel, Democracy’s Discontent: America in Search of a Public Philosophy (Cambridge, MA: Harvard University Press, 1996). 44. Demetrios James Caraley, “Ending Welfare As We Know It: A Reform Still in Progress,” Political Science Quarterly 116, no. 4 (Winter 2001–2002): 557–58. 45. Mead, Beyond Entitlement; Magnet, Dream and the Nightmare. 46. Letter to members of the Assembly, April 25, 1996. 47. Grant County Department of Social Services, The Secret of Welfare Reform (Lancaster, WI: Grant County Department of Social Services, April 20, 2001), p. 7. 48. Interview with Jon Angeli, Director, Grant County Department of Social Services, Lancaster, WI, June 2, 1995. 49. Thompson, Wisconsin, chap. 8. 50. Ibid., p. 336. 51. Jonathan Rieder, Canarsie: The Jews and Italians of Brooklyn against Liberalism (Cambridge, MA: Harvard University Press, 1985); Jim Sleeper, The Closest of Strangers: Liberalism and the Politics of Race in New York (New York: Norton, 1990). 52. Matthew 10:16.

336

•

Notes to Chapter 13

CHAPTER 13 IMPLICATIONS 1. James Q. Wilson and George L. Kelling, Jr., “Broken Windows: The Police and Neighborhood Safety,” The Atlantic , March 1982, pp. 29–38; George L. Kelling and William J. Bratton, “Taking Back the Streets,” City Journal 4, no. 3 (Summer 1994): 38–46. 2. Wesley G. Skogan, “Crime and Punishment,” in Politics in the American States: A Comparative Analysis, 6th ed., ed. Virginia Gray and Herbert Jacob (Washington, DC: CQ Press, 1996), pp. 391–92; Mark A.R. Kleiman, “Coerced Abstinence: A Neopaternalist Drug Policy Initiative,” in The New Paternalism: Supervisory Approaches to Poverty, ed. Lawrence M. Mead (Washington, DC: Brookings, 1997), chap. 6. 3. “How to Start a Cease-Fire: Learning from Boston,” Time, July 21, 1997, p. 28; Ford Foundation, Innovation in American Government, 1997 (New York: Ford Foundation, 1997), p. 31. 4. Fox Butterfield, “Killing of Girl, 10, and Increase in Homicides Challenge Boston’s Crime-Fighting Model,” New York Times, July 14, 2002, p. 14. 5. Christopher Winship and Jenny Berrien, “Boston Cops and Black Churches,” The Public Interest, no. 136 (Summer 1999): 52–68. 6. Chester E. Finn, Jr., “Paternalism Goes to School,” in New Paternalism, ed. Mead, pp. 227–31. 7. Diane Ravitch and Joseph P. Viteritti, eds., New Schools for a New Century: The Redesign of Urban Education (New Haven: Yale University Press, 1997); David Grissmer and Ann Flanagan, “Exploring Rapid Achievement Gains in North Carolina and Texas” (Washington, DC: National Educational Goals Panel, November 1998). 8. Robert D. Putnam, with Robert Leonardi and Raffaella Y. Nanetti, Making Democracy Work: Civic Traditions in Modern Italy (Princeton: Princeton University Press, 1993). For a parallel analysis centered on the less civic Italian south, see Edward C. Banfield, The Moral Basis of a Backward Society (Glencoe, IL: Free Press, 1958). 9. Tom W. Rice and Alexander F. Sumberg, “Civic Culture and Government Performance in the American States,” Publius 27, no. 1 (Winter 1997): 109–10. 10. Daniel J. Elazar, American Federalism: A View from the States, 3rd ed. (New York: Harper and Row, 1984), pp. 133–37, 192–93. 11. Elazar, American Federalism, 3rd ed., pp. 132–33. 12. Hugh Heclo, “Issue Networks and the Executive Establishment,” in The New American Political System, ed. Anthony King (Washington, DC: American Enterprise Institute, 1978), chap. 3; John W. Kingdon, Agendas, Alternatives, and Public Policies, 2nd ed. (New York: HarperCollins, 1995); Marc K. Landy and Martin A. Levin, eds., The New Politics of Public Policy (Baltimore: Johns Hopkins University Press, 1995); James Q. Wilson, Political Organizations (Princeton: Princeton University Press, 1995). 13. Edward C. Banfield and James Q. Wilson, City Politics (Cambridge, MA: Harvard University Press, 1963).

Notes to Chapter 13

•

337

14. James Q. Wilson, “Corruption: The Shame of the States,” The Public Interest, no. 2 (Winter 1966): 28–38. 15. Elazar, American Federalism, 3rd ed., pp. 188–89. 16. Daniel J. Elazar, American Federalism: A View from the States (New York: Crowell, 1966), pp. 96–111; Daniel J. Elazar, American Federalism: A View from the States, 2nd ed. (New York: Crowell, 1972), pp. 116–19; Elazar, American Federalism, 3rd ed., pp. 134–37. 17. Rodney E. Hero and Caroline J. Tolbert, “A Racial/Ethnic Diversity Interpretation of Politics and Policy in the States of the U.S.,” American Journal of Political Science 40, no. 3 (August 1996): 851–71. 18. Margaret Weir, Ann Shola Orloff, and Theda Skocpol, eds., The Politics of Social Policy in the United States (Princeton: Princeton University Press, 1988). 19. U.S. Congress, House, Committee on Ways and Means, Overview of Entitlement Programs: 1990 Green Book: Background Material and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, June 5, 1990), p. 580; U.S. Congress, House, Committee on Ways and Means, 2000 Green Book: Background Material, and Data on Programs within the Jurisdiction of the Committee on Ways and Means (Washington, DC: U.S. Government Printing Office, October 6, 2000), p. 439. 20. David Hackett Fischer, Albion’s Seed: Four British Folkways in America (New York: Oxford University Press, 1989), pp. 388, 811–12. 21. Elazar, American Federalism, 3rd ed., pp. 131, 134, 150–52. 22. Russell L. Hanson, “The Political Acculturation of Migrants in the American States,” Western Political Quarterly 45, no. 2 (June 1992): 355–83. 23. Elazar, American Federalism, 3rd ed., p. 132. 24. Robert S. Erikson, Gerald C. Wright, and John P. McIver, Statehouse Democracy: Public Opinion and Policy in the American States (Cambridge, England: Cambridge University Press, 1993), chap. 3. 25. Nicholas Lemann, The Promised Land: The Great Black Migration and How It Changed America (New York: Knopf, 1991), pp. 6–7, 149, 200. 26. Social Security Act (1935), sec. 402(a)(5). 27. Martha Derthick, The Influence of Federal Grants: Public Assistance in Massachusetts (Cambridge, MA: Harvard University Press, 1970), pp. 98–101. 28. Derthick, Influence of Federal Grants. 29. Terrence Maxwell, “Information Federalism: History of Welfare Information Systems” (Albany: Rockefeller Institute of Government, 1999); U.S. General Accounting Office, Welfare Reform: Improving State Automated Systems Requires Coordinated Federal Effort (Washington, DC: U.S. General Accounting Office, April 2000). Washington, however, continues to pay 80–90 percent of the cost of MIS improvements in state child support enforcement programs. 30. David E. Osborne and Ted Gaebler, Reinventing Government: How the Entrepreneurial Spirit Is Transforming the Public Sector (Reading, MA.: AddisonWesley, 1992), chaps. 4–5. 31. Thomas L. Gais, “Welfare Reform Findings in Brief” (Albany: State University of New York, Rockefeller Institute of Government, March 1, 2002).

338

•

Notes to Chapter 13

32. James J. Kemple, Daniel Friedlander, and Veronica Fellerath, Florida’s Project Independence: Benefits, Costs, and Two-Year Impacts of Florida’s JOBS Program (New York: Manpower Demonstration Research Corporation, April 1995); Dan Bloom, James J. Kemple, Pamela Morris, Susan Scrivener, Nandita Verma, and Richard Hendra, The Family Transition Program: Final Report on Florida’s Initial Time-Limited Welfare Program (New York: Manpower Demonstration Research Corporation, December 2000). 33. At several conferences I attended, Don Winstead, who was Welfare Reform Administrator for the Florida Department of Children and Families for much of this period, was a skillful exponent of Florida policies. In 2002, he was named Deputy Assistant Secretary of the U.S. Department of Health and Human Services, in charge of planning for income maintenance. 34. South Carolina Department of Social Services, Division of Program Quality Assurance, “Survey of Former Family Independence Program Clients: Cases Closed during April through September, 1998” (Columbia: South Carolina Department of Social Services, June 2000). 35. Charles Murray, Losing Ground: American Social Policy, 1950–1980 (New York: Basic Books, 1984); idem, “The Coming White Underclass,” Wall Street Journal, October 29, 1993, p. A14; Heather Mac Donald, “The Real Welfare Problem Is Illegitimacy,” City Journal 8, no. 1 (Winter 1998): 35–47. 36. Marvin Olasky, “Welfare States,” National Review, October 23, 1995, pp. 44–48. 37. Arland Thornton, “Attitudes, Values, and Norms Related to Nonmarital Fertility,” in U.S. Department of Health and Human Services, Public Health Service, Report to Congress on Out-of-Wedlock Childbearing (Washington, DC: U.S. Government Printing Office, September 1995), pp. 201–15. 38. Isabel V. Sawhill, “The Perils of Early Motherhood,” The Public Interest, no. 146 (Winter 2002): 80–83. 39. Lawrence M. Mead, “The Politics of Conservative Welfare Reform,” in The New World of Welfare, ed. Rebecca M. Blank and Ron Haskins (Washington, DC: Brookings Institution, 2001), pp. 212–17. 40. “Will Tinkering Bring the Building Down?” The Economist, March 2, 2002, pp. 29–30; Robin Toner, “Rallies in Capital Protest Bush Welfare Proposals,” New York Times, March 6, 2002, p. A17; Robert Pear, “50 Senators Ask Daschle for Debate on Renewing Welfare Law,” New York Times, September 12, 2002, p. A12. 41. Lawrence M. Mead, The New Politics of Poverty: The Nonworking Poor in America (New York: Basic Books, 1992), chaps. 1–7. 42. Robin Toner, “Helping the Poor in the Post-Welfare Era,” New York Times, March 3, 2002, pp. WK1, WK7. 43. Mead, New Politics of Poverty, chaps. 10–11. 44. Jason DeParle, “The Silence of the Liberals,” Washington Monthly 31, no. 4 (April 1999): 12–17. 45. Katherine McFate, Roger Lawson, and William Julius Wilson, eds., Poverty, Inequality, and the Future of Social Policy: Western States in the New World Order (New York: Russell Sage, 1995).

Notes to Chapter 13

•

339

46. Lawrence M. Mead, Beyond Entitlement: The Social Obligations of Citizenship (New York: Free Press, 1986); George Lakoff, Moral Politics: What Conservatives Know That Liberals Don’t (Chicago: University of Chicago Press, 1996). 47. Ronald Dworkin, Taking Rights Seriously (Cambridge, MA: Harvard University Press, 1977). 48. This is the cultural model that Lakoff, Moral Politics, calls Nurturant Parent. It is also similar to the worldly tolerance that emanates from the New York Times Magazine, with its constant articles focusing on cutting-edge developments in politics and culture. 49. Samuel P. Huntington, “Political Development and Political Decay,” World Politics 17 (April 1965): 405–17. 50. Samuel H. Beer, “In Search of a New Public Philosophy,” in New American Political System, ed. King, p. 44. 51. Herbert Croly, The Promise of American Life, ed. Arthur M. Schlesinger, Jr. (Cambridge, MA: Harvard University Press, 1965). 52. Michael J. Sandel, Democracy’s Discontent: America in Search of a Public Philosophy (Cambridge, MA: Harvard University Press, 1996), pp. 324– 28, passim.

Index

ACORN (Association of Community Organizations for Reform Now), 59, 149 administrative requirements of national policy, 269–70 advisory groups, 265 AFDC (Aid to Families with Dependent Children), 18–20, 279n.1; abolition of, 37, 46, 53, 58, 144, 292n.99; administrative costs of, 69, 194–95; assets, and eligibility for, 28; automobiles, and eligibility for, 28; caseload declines in, 76–77, 175–76, 181; child care under, 123; child support under, 111; criticism of, 18–19; enrollment growth, 21–23; funding for, 67, 270; support for, 40–41; vs. W-2, 107, 110–11; waiver experiments, 25–28, 176–77, 219–20, 319n.20; Wisconsin vs. national enrollment levels, 21; work requirements of, 20 (see also work requirements, implementing). See also TANF AFDC Benefit Cap, 29, 48 AFDC-UP (Aid to Families with Dependent Children—Unemployment Parent), 186 Aid to Families with Dependent Children. See AFDC Aid to Families with Dependent Children—Unemployment Parent (AFDCUP), 186 Alabama, 228–29 Altmeyer, Arthur, 255, 269 American government. See government America Works (AW), 137 Anderson, Eloise, 32 Andrea, Joseph, 23, 48, 49–50, 289n.40 ANF (Assessing the New Federalism; Urban Institute), 225–26, 331n.36 Angeli, Jon, 84–86, 106, 165, 173–74, 258–59, 298n.8 Anglican vision, 249 Arizona, 228, 229, 230 Assessing the New Federalism (ANF; Urban Institute), 225–26, 331n.36 Association of Community Organizations for Reform Now (ACORN), 59, 149

automobiles, 22–23, 28 AW (America Works), 137 BadgerCare, 123, 194, 202, 254 Baldwin, Tammy, 117 Bardach, Eugene: Improving the Productivity of JOBS Programs, 301n.46 Bascom, John, 255 Becker, Dismas, 55–56 Beer, Samuel, 277 Bell, Stephen H., 319n.20 Berge, Gerald, 55 Bi-Partisan Legislative Working Group, 113–14 black box problem, 104 blacks: migration of, 268; welfare reform supported by, 57–59, 247, 292n.99 Bloomberg, Michael, 223 boot camps for delinquent youth, 262 Bos, Johannes M., 283n.15 Boston, 262–63 Bridefare. See PFR Britain, 248 Brock, Thomas W., 283n.15 Bush, Andrew, 108, 109, 110–11, 114, 130 Bush (George H. W.) Administration, 26 business, 60–61, 119 Califano, Joseph, 37–38 California: caseload declines in, 190; coordination in, 230; economic development of, 253; liberal culture of, 174, 328n.18; moralistic culture of, 220, 241; political development of, 253; political divisions in, 228, 241; welfare spending in, 207 Canada, 249 capability, 225, 231–32 capitalism, 248, 250 CARES (Client Assistance for Reemployment and Economic Support), 68, 74, 143, 180 Caretaker Supplement, 121, 145, 205 Carter Administration, 37, 256 Catholics, 118–19 CFWA (Congress for a Working America), 158, 167

342

•

Index

CHAD (Consortium of Hispanic Agency Directors), 141 Chamber of Commerce (Madison), 61 Chandler, Rich, 284n.38 Chicago, 211, 221 Child Assistance Program (New York), 293n.109 child care: under AFDC, 123; costs of, 275; W-2 denials of, 145–46; and W-2’s emergence, 114, 119, 123, 125–26 children: abuse/neglect of, 203; health of, 209, 325n.53; W-2’s effects on, 202, 206–7, 211–13, 325n.48, 323n.29 Children First: assessment of, 285n.44; creation/expansion of, 25; evaluation of, 197; paternalism of, 28, 157, 161–64, 167; success of, 211 child support, 23, 161; under AFDC, 111; and caseload declines, 180, 183; evaluation of, 197–98, 202, 321n.12; in moralistic states, 239–40; under PRWORA, 111–12; and W-2’s emergence, 111–12, 302n.15; Wisconsin leadership in, 253– 54. See also Children First Child Support Assurance System (CSAS), 293n.109 Christian, Shannon, 108 Chvala, Chuck, 53 Citizen’s Conference on State Legislatures, 188, 253 City Journal (New York City), 330n.32 civility, 276–77 civil rights, 265 Civil War, 265 class diversity, 252 Cleveland, 161, 211 Client Assistance for Reemployment and Economic Support. See CARES Clinton Administration: PRWORA signed by, 38; on waivers, 26; welfare reform goals of, 29–30, 32, 44, 96–97, 167 Coalition for Change, 149 Cohen, Wilbur, 255 Collins, John, 80, 84 Colorado, 220, 229, 241, 331n.40 commitment, 224–25, 229 Committee for Economic Development, 128 Committee on Economic Security, 255 Commons, John R., 255 Community Service Job. See CSJ

Community Work Experience Program. See CWEP Computer Reporting Network (CRN), 68 Congress for a Working America (CFWA), 158, 167 Connecticut, 253 consensus, 224–25, 228 consent, 272–76, 339n.48 conservatives vs. liberals, 19–20, 245–47, 262, 272–74 Consortium of Hispanic Agency Directors (CHAD), 141 Cook, Mary Ann, 87 coordination, 224–25, 229–31 Corbett, Thomas J., 283n.17, 292n.90, 293n.108, 309n.27, 312n.74 counties, welfare reform by, 79–106, 297– 98n.6; administrative excellence in, 105– 6; analysis results, assessing, 103–4, 301n.42; and case management, 91–92; and client assignments, 93–96; and contracting to private agencies, 92; Dane County, 87, 93, 96; Douglas County, 90, 94–95; Fond du Lac County, 88–89, 92, 95, 135, 301n.46; Grant County, 79, 84–86, 90, 92–93, 95–97, 105–6, 136, 298n.8; Jackson County, 83; and job centers, 92–93; Kenosha County, 79, 80–84, 92–98, 101, 105–6, 131, 298nn.15 and 17, 299n.28, 300n.30; Marathon County, 90, 94–96; Milwaukee County, 91, 92–93, 96–98, 102–3, 105, 299n.25; performance of counties, 96–102, 300nn.32–36; Pierce County, 135; Racine County, 89, 92–93, 97–98; Sheboygan County, 79, 86–87, 92, 94, 96–97, 106, 300n.30; statistical results, assessing, 300n.41; Winnebago County, 88. See also individual counties criminal justice, 262 CRN (Computer Reporting Network), 68 CSAS (Child Support Assurance System), 293n.109 CSJ (Community Service Job), 126–27, 134, 145, 147 Cullen, Tim, 44, 46, 51–52, 292n.90 Curtis and Associates, 86–87, 92, 106 CWEP (Community Work Experience Program), 24–25, 65–66, 73 Dane County: caseload declines in, 182– 83; income in, 201; paternalism in, 166– 67; welfare reform in, 87, 93, 96

Index Danziger, Sheldon, 293n.108 data, sharing of, 270 decline of welfare. See welfare, decline of DeLessio, Pat, 59, 118 delinquent youth, 262 Democratic national convention (1968), 265 Democrats, 252. See also politics of reform demographics, 183–84, 187, 210, 267 DeParle, Jason, 5, 279n.3 Department of Health and Family Services (DHFS), 135 Department of Health and Social Services. See DHSS Department of Industry, Labor, and Human Relations (DILHR), 135 Department of Workforce Development. See DWD Depression, 253, 254 Detroit, 268 development of reform. See welfare reform DHFS (Department of Health and Family Services), 135 DHHS (U.S. Department of Health and Human Services), 96; and ES, 65, 66; and the IRP, 60; waivers granted by, 27– 28; and WEJT, 66; welfare reform initiated by, 46–47, 292n.90 DHSS (Department of Health and Social Services), 135, 138, 141, 164 DILHR (Department of Industry, Labor, and Human Relations), 135 Douglas County, 90, 94–95, 182 Dreyfus, Lee, 23, 24, 50 drinking, outlawing of, 247 DWD (Department of Workforce Development), 135, 139–42, 144, 147–48, 154– 55, 198–200, 203 Dworkin, Ronald, 275 Earl, Clark, 81, 84, 92 Earl, Tony, 24, 46–47, 48, 49, 51, 54 earned income tax credit. See EITC earnings/income, W-2’s effects on, 199– 200, 211, 212, 213 economic panic (1893), 251 education, paternalism in, 263 effective welfare policy, definition of, 13– 14, 281n.25 effects of reform, 1, 193–213. See also W2, effects of

•

343

EITC (earned income tax credit), 23, 113, 176, 181, 191–92, 212, 283n.15, 318n.10 Elazar, Daniel, 12, 13–14, 79, 188, 216, 248–49, 265, 267. See also political/public culture Ellwood, David, 56, 60 emergence of W-2. See W-2, emergence of emergency assistance, 115 employment, W-2’s effects on, 198–99, 209–10, 322nn.18 and 20 Employment and Training Choices (Massachusetts), 221, 330n.28 Employment and Training Institute (ETI; University of Wisconsin—Milwaukee), 60, 154 Employment Security Automated Reporting System (ESARS), 68 Employment Service (ES), 64–65, 66, 223, 229–30. See also Job Service Employment Skills Advancement Program (ESAP), 116, 122–23, 127 Employment Solutions, Inc. (ESI), 139, 142, 143 England, 248 Engler, John, 254 entitlement, 13; and PRWORA, 109, 270; and W-2’s emergence, 109, 118, 122, 125; vs. work ideal/tests, 37–38, 51–53, 56, 59, 61, 241–42, 269–70, 275–76, 339n.48 equal rights, 265 ESAP (Employment Skills Advancement Program), 116, 122–23, 127 ESARS (Employment Security Automated Reporting System), 68 ESI (Employment Solutions, Inc.), 139, 142, 143 ethnic diversity, 252 ETI (Employment and Training Institute; University of Wisconsin—Milwaukee), 60, 154 European welfare states, 247–48, 275 experimental evaluations, 104 experts, 265 families/children, W-2’s effects on, 202, 211–13, 323n.29 Family Assistance Plan, 37 Family Care, 254 Family Independence Program (Washington State), 307n.1

344

•

Index

Family Support Act. See FSA fathers, 211. See also Children First; child support Fenton, John, 251–52 FEP (Financial and Employment Planner), 112–13, 136–37, 145, 171–72, 230 fiscal effects of W-2, 194–97, 207, 321n.7 Fischer, David, 267 Fish, Al, 292n.90 Florida, 221, 228–29, 231, 271–72, 338n.33 Fond du Lac County: caseload declines in, 182; Children First in, 161–62; paternalism in, 167; welfare reform in, 88–89, 92, 95, 135, 301n.46 Food Stamp Employment and Training (FSET), 308n.15 food stamps, 29, 53, 69, 70, 113, 136–37, 193–95, 296n.27 FSA (Family Support Act; 1988), 26, 66, 69–70, 96, 186, 220, 271 FSET (Food Stamp Employment and Training), 308n.15 GAIN (Greater Avenues for Independence; California), 230 gangs, 262–63 Gard, John, 113, 124, 289n.49 Gardner, John, 158, 174 Garfinkel, Irwin, 23, 51, 292n.90, 293nn.108 and 109 Garvey, Ed, 116–17 GATES (Greater Access to Employment and Self-Sufficiency), 141 GCW (Grant County Works), 173 General Relief (GR), 21–22, 25, 51, 53, 193, 283n.23 George, Gary, 57–58, 116, 247, 292n.99 George III, king of England, 248 Georgia, 207, 228, 229, 230–31 German immigrants, 249–50 Gilens, Martin, 332–33n.50 Giuliani, Rudolph, 222 good government, 2, 214–42; capability, 225, 231–32; commitment, 224–25, 229; consensus, 224–25, 228; coordination, 224–25, 229–31; evidence for, 225–26, 331nn.36 and 37; leaders of reform, 219–21, 328n.20, 329nn.25 and 26, 330n.28; at the local level, 105–6; meaning of, 215–16; moralistic approach to, 13–14 (see also moralistic

states); in New York, 221–23, 330nn.30, 31, and 33; policy assessment, 232–33, 332nn.43–45; policymaking, 224–25, 228, 331–32nn.40 and 41; policy results, 233–41, 332nn.46–48; and political culture, 216–19, 328n.9, 328n.18 (see also individualistic states; moralistic states; traditionalistic states); process assessment, 223–26, 331nn.36 and 37; process results, 226–32, 331–32nn.39–41; resources, 224–25, 229; state political cultures as a measure of, 12; and W-2’s success, 13–16, 281n.25 Goodrich, Patricia, 160 Goodwill Industries, 75, 81, 141 government: capability/size of, 247–48, 261, 272; excellence of, definition of, 13–14, 281n.25; suspicion of, 249–50. See also good government Governor’s Commission on Welfare Reform (Wisconsin), 46, 48, 292n.90 GR. See General Relief Grand Old Party (GOP). See Republicans Granger, Robert C., 283n.15 Grant County: caseload declines in, 182; paternalism in, 165, 173–74, 315n.39; rural setting of, 258; WEJT in, 165, 315n.39; welfare reform in, 79, 84–86, 90, 92–93, 95–97, 105–6, 136, 298n.8; work goals/responsibility in, 258–59 Grant County Works (GCW), 173 Gray, Virginia, 215 Greater Access to Employment and SelfSufficiency (GATES), 141 Greater Avenues for Independence (GAIN; California), 230 Greater Milwaukee Committee, 169 Great Society, 64, 267, 277 Greenberg, David, 307n.89 Hamilton, Alexander, 277 hardship, 202–5, 323n.33 Haveman, Robert, 23, 293n.108 Head Start, 194 health care, 114, 119, 123–24, 202 health insurance coverage, 209 Healthy Start, 284n.29 Heaney, Michael, 292n.92, 303n.41 Hispanics, migration of, 268 Hmong refugees, 86 Hobbes, Thomas, 256 homelessness, 204

Index homicides, 262–63 Hoover, Mark, 330n.33 Hotz, V. Joseph, 283n.15 H&R Block, 212 Hudson Institute, 108 Human Resources Administration (New York City), 330n.33 100-Hour Rule Demonstration, 27, 29, 30, 197, 285n.44 Idaho, 190, 207 illegitimacy. See unwed pregnancies Illinois, 190, 253 immigration, 267–68, 269 implementing W-2. See W-2, implementing implications of W-2. See W-2, implications of Improving the Productivity of JOBS Programs (Bardach), 301n.46 income. See earnings/income, W-2’s effects on Indian tribes, 308n.16, 312nn.69 and 70 individualism, 256–57, 265 individualistic states, 53–54, 218–23, 320n.35, 328–29n.20; AFDC grants to, 237; caseload change in, 238; dependency supported by, 233, 236; Elazar on, 188; Massachusetts, 60–61, 221, 266; New York, 60–61, 221–23, 249, 266, 331n.40; Pennsylvania, 60–61, 249; performance of, 226; political differences in, 259; sanctions in, 238, 240; settlers in, 248–49; on welfare reform, generally, 241–42 Institute for Research on Poverty. See IRP Institute for Wisconsin’s Future, 117 integration, racial, 268 interest groups, 266 Interfaith Conference of Greater Milwaukee (1998), 293n.103 Iowa, 220 IRP (Institute for Research on Poverty; University of Wisconsin—Madison), 23, 60, 107–9, 154, 255–56, 290n.57, 292n.90, 293nn.108 and 109, 335n.39 Italy, 218–19, 248, 264 Jackson County, 83 Jankowski, Larry, 81, 82, 84, 165 Jefferson, Thomas, 277 Jencks, Christopher, 56 Jim Crow, 265

•

345

Job Center, 81, 93, 299n.28 JOBS (Job Opportunities and Basic Skills Training Program), 11, 296n.26; bureaucratic competitive/incentive approach, 75–76, 296–97nn.37–39; and caseload declines, 181, 183–85, 187–88, 319n.26; case management in, 91–92; client assignments by, 93–96; by county, performance of (see counties, welfare reform by); creation of, 66, 96; disadvantaged, definition of, 299n.22; educational emphasis of, 71–72, 296n.30; enrollment decline in, 74, 296n.34; funding for, 67– 68, 75; goals of, 164; job entries raised by, 76; participation rates for, 198, 271; paternalism of, 164–65; in Portland, 131, 172, 220–21, 298n.17, 307n.89; services emphasized by, 36; work-first emphasis of, 71–73; work requirements of, 36, 66, 69, 70–71, 133 Job Service (JS), 64, 92–93, 141 Jobstat (New York City), 223 Job Training Partnership Act. See JTPA Johnson, Lyndon, 255 JTPA (Job Training Partnership Act), 81, 92–93, 229, 301n.42 justice, 274–75 Kansas, 207, 230 Kaplan, Thomas, 34, 293n.108, 312n.74 Kaus, Mickey, 56, 60 Kaye, Laura, 108 Kenosha County: caseload declines in, 182; Learnfare in, 160; as a model for state welfare reform, 12, 84; paternalism in, 164–65; WEJT in, 164–65; welfare reform in, 79, 80–84, 92–98, 101, 105–6, 131, 298nn.15 and 17, 299n.28, 300n.30 Key, V. O., Jr., 331n.41 Kingdon, John, 35 Kinship Care, 121, 145, 203 Klauser, James, 44, 108 Krug, Shirley, 52, 116 Krusick, Margaret, 45 Kunicki, Walter, 52, 117 labor market/unemployment, 183–84, 210–11; in Milwaukee, 102–3, 185, 210–11; shortages, 179; W-2’s effects on, 198–99, 322nn.18 and 20 La Follette, Philip, 253

346

•

Index

La Follette, Robert M., 250, 251, 255 Lakoff, George, 339n.48 Lampman, Robert, 255 Latin America, 269 Lau Family, 141 law enforcement, 262–63 leadership, national, 253–54 Learnfare: assessment/evaluation of, 197, 285n.43; costs of, 49; creation/expansion of, 25, 27, 44, 49; effects on caseloads, 180; origins of, 32, 284n.38; paternalism of, 28, 157, 159–61, 167; support for, 45 left vs. right. See conservatives vs. liberals Legislative Audit Bureau, 201 Legislative Council Study Committee (Wisconsin), 51, 290n.63 Legislative Fiscal Bureau, 124, 195 Lemann, Nicholas, 268 Lenkowsky, Leslie, 108 Leutermann, George, 81, 82, 83, 84, 91, 165 liberal states, 174, 237 liberals vs. conservatives. See conservatives vs. liberals liberty, 256 lobbies, 266 Locke, John, 256 Loftus, Tom, 43–44, 50, 51–52, 55, 290n.57, 292n.90 Los Angeles, 211, 329n.26 Louisiana, 189–90 machine politics, 265, 267–68 Madison, 62, 87 Management and Evaluation Project. See MEP management information systems (MIS), 68, 270, 337n.29 Manhattan Institute (New York City), 330n.32 Manpower Demonstration Research Corporation, 10, 104 manufacturing employment, 23 Marathon County, 90, 94–96 marriage, 212–13, 272–73 Martin, Will, 58 Massachusetts: administrative standards in, 269–70; capability of, 231; individualistic culture of, 60–61, 221, 266; liberal culture of, 328n.18; reform opposed in, 228

Maximus, 81, 142, 143, 172 McCallum, Scott, 62 Medicaid: administrative costs of, 69, 194– 96; caseload declines for, 193–94; eligibility for, 136–37; generosity of, 123–24; in Wisconsin, 254 Medicaid Assistance Extension, 27, 284n.29 Meier, Jim, 292n.90 MEP (Management and Evaluation Project), 154–55, 256, 281n.21, 285n.46, 312n.76 merit system, 252, 269–70 Michigan, 190, 209, 220, 253 Midwestern conservatism, 245–46 Milwaukee, City of: blacks’ support of reform in, 58–59; conservatism of, 247; demographics of, 39, 102, 155–56, 259; economy of, 39; health insurance coverage in, 210; homelessness in, 204; JOBS in, 75–76, 103, 296–97n.38, 300n.39; labor market/unemployment in, 102–3, 185, 210–11; Learnfare in, 159; PFP/ SSF’s effects on, 31; political corruption in, 62; poverty in, 131, 193, 210, 259; reform delays in, 49; riots of 1969, 42; SST’s effects on, 31; welfare decline in, 3; Work First in, 134 Milwaukee County: caseload declines in, 142, 182–83, 185, 309n.29, 320n.30; paternalism in, 166; privatization in, during W-2 implementation, 138–39, 141– 44, 309n.29, 310n.31; welfare reform in, 91, 92–93, 96–98, 102–3, 105, 299n.25 Milwaukee Shelter Task Force, 204 Minnesota, 209, 230, 253 Minnesota Family Investment Program, 202 MIS (management information systems), 68, 270, 337n.29 Mississippi, 190, 228–29, 231 Missouri, 207, 228 Moore, Gwendolynne, 57–58, 116–18, 145, 195, 292n.99 morale/respect, 212 moralistic culture, national, 268–69 moralistic states, 64, 188, 216–23, 320n.35, 328–29n.20; AFDC grants to, 237; California, 220; capability in, 231– 32; caseload change in, 238; child support in, 239–40; consensus in, 228; de-

Index pendency opposed by, 233; Elazar on, 79, 216; Michigan, 190; Oregon, 190; performance of, 226; policymaking in, 228, 331–32nn.40 and 41; and race, 266–67; reform’s favoring of, 241; sanctions in, 238–39; settlers in, 248–49; trend toward, 266; Wisconsin, 12–14, 54, 79, 216, 220, 221–23, 249; work requirements in, 238, 240–41 Mountain states, conservatism of, 246 Mugwumps, 251 Murray, Charles, 56, 60, 272 Nathan, Dick, 15 National Alliance of Business, 84 National Conference of State Legislatures, 215 National Evaluation of Welfare-to-Work Strategies (NEWWS), 220–21, 298n.17 National Governor’s Association, 254 national leadership, 253–54 national policy, 269–72, 277, 337n.29 National Resources Center on Family Based Services (University of Iowa), 84– 85 National Survey of America’s Families (NSAF), 207–8, 209 Nelson, Betty Jo, 292n.90 New Chance, 285n.41 New Deal, 250, 252, 277 New Hampshire, 206 New Hope Project (Wisconsin), 52, 61, 117, 119, 157, 167–71, 323n.29 New Jersey, 228, 229, 249, 253 New Mexico, 266 NEWWS (National Evaluation of Welfareto-Work Strategies), 220–21, 298n.17 New York City, 222–23, 228, 330nn.31 and 33 New York State: bureaucracy in, 222–23; caseload declines in, 190; coordination in, 230; cynicism about welfare/poverty in, 246, 259; economic development of, 253; good government in, 221–23, 330nn.30, 31, and 33; individualistic culture of, 60–61, 221–23, 249, 266, 331n.40; liberal culture of, 174, 328n.18; political divisions in, 222, 228, 330nn.30 and 31 New York Times Magazine, 339n.48 99 Group, 108–9, 130, 137, 138 Nixon Administration, 37, 256

•

347

Norquist, John, 49–53, 58, 116, 117, 126 North Carolina, 228–29, 263, 266 Northwestern University, 335n.39 Notestein, Barbara, 52, 56, 116 Noyes, Jennifer, 312n.76 NSAF (National Survey of America’s Families), 207–8, 209 Nurturant Parent, 339n.48 OBRA (Omnibus Budget Reconciliation Act; 1981), 65 Ohio, 228, 331n.40 OIC (Opportunities Industrialization Center), 141, 142 Olasky, Marvin, 273 Omnibus Budget Reconciliation Act (OBRA; 1981), 65 One-Hundred-Hour Rule Demonstration. See 100-Hour Rule Demonstration. Operation Cease Fire (Boston), 262–63 Opportunities Industrialization Center (OIC), 141, 142 Oregon, 190, 207, 220–21, 230 Pacific states, liberalism of, 246 Parental and Family Responsibility Demonstration Project. See PFR Parents Fair Share (PFS), 163 paternalism, 1–2, 157–74; of Children First, 157, 161–64, 167; in Dane County, 166–67; in Fond du Lac County, 167; in Grant County, 165, 173–74, 315n.39; in Kenosha County, 164–65; of Learnfare, 157, 159–61, 167; in localities, 262–63; meaning of, 157– 59; in Milwaukee County, 166; of the New Hope Project, 157, 167–71; in Racine County, 166; in Sheboygan County, 166; of W-2, 157, 171–72; of work programs, 164–67 Pavetti, LaDonna A., 285n.41 Pay for Performance. See PFP Pennsylvania, 60–61, 249 performance standards, 270–71 Personal Responsibility and Work Opportunity Reconciliation Act (1996). See PRWORA PFP (Pay for Performance): caseload declines due to, 77; creation of, 115, 134; evaluation of, 321n.11; origins of, 32, 45; scope of, 31; work requirements of, 31, 70, 72

348

•

Index

PFR (Parental and Family Responsibility Demonstration Project; Bridefare), 28, 32, 33, 44, 48 PFS (Parents Fair Share), 163 PIC (Private Industry Council), 81, 137, 297n.39 Pierce County, 135 pluralism, 267–68 police, 262 Policy Group on Welfare Reform, 117–18 policymaking: analysis limits of, 34–35; as education, 35–36; goals vs. means in, 34; in good government, 224–25, 228, 331–32nn.40 and 41; as problemsolving, 35; in reform’s early stages, 105; satisficing in, 34 political/public culture, 53–54, 60–61, 216–19, 328n.9; on American government, 247–50; and good government, 216–19, 328nn. 9 and 18; malleability of, 264–65; secondary tendencies in, 266. See also individualistic states; moralistic states; traditionalistic states political science, 9–10 politics of reform, 37–62; academics’ opposition to reform, 60, 107–8, 293nn.108 and 109; bipartisan support for W-2, 115–17; blacks’ support of reform, 57– 58, 247, 292n.99; bureaucracy’s role in, 37, 46–48, 56, 62; business’s support of reform, 60–61; community groups’ support of reform, 59, 293n.103; conservatives vs. liberals, 37–38, 48, 57; Democratic criticism of welfare, 49–53, 116, 289n.53; entitlement vs. work ideal/ tests, 37–38, 51–53, 56, 59, 61, 241–42, 269–70, 275–76, 339n.48; ideas’ role in, 56–57; influential figures in, 292nn.90 and 92; moralistic culture of, 37, 53–57; outside groups’ support of reform, 37, 57–61, 293nn.103, 108, and 109; partisan view of, 37, 41–46, 62; problem-solving approach to, 54–55; Republican moderation in, 48–49, 53, 61; setting for reform, 38–41; trust displayed in, 55– 56; welfare politics, 37–38 population trends, 183 pork-barrel politics, 265 Portland, 131, 166, 172, 220–21, 298n.17, 307n.89, 329n.25 poverty: and need to aid the poor, 242, 332–33n.50; paternalism toward the

poor, 158–59; and race, 267; recovery from, 1; studies of, 9; unwed pregnancies as a cause of, 48, 290n.57; W-2’s effects on, 193, 200–201, 211, 213, 274, 323n.27; War on Poverty, 255; in Wisconsin vs. other states, 205–6, 208–9, 210, 325nn.47 and 53. See also New Hope Project Private Industry Council (PIC), 81, 137, 297n.39 problem-solving approach to reform, 54– 55 Program for Better Jobs and Income, 37 Progressivism, 250–53, 265, 277 Project Independence (Florida), 307n.2 Project Match (Chicago), 32, 284n.39 Propsom, Gail, 47–48 prostitution, 247 PRWORA (Personal Responsibility and Work Opportunity Reconciliation Act; 1996): blacks’ support of, 58; child support under, 111–12; conservatism of, 38; and entitlement, 109, 270; goals of, 274; and marriage, 273; passage of, 2, 38; radical nature of, 2; support for, 40–41, 273–74; work requirements of, 220, 271 public administration, 252–53 publicity, 271–72 Public Welfare Board (Kenosha County), 80 Puritans, 249, 258, 267, 268 Putnam, Robert, 218–19, 248, 264 Quakers, 249 race, 39, 155, 266–69 Racine County: Children First in, 161–62, 163; paternalism in, 166; welfare reform in, 89, 92–93, 97–98 Reagan Administration, 26 recession (1980s), 22–23 reciprocity, 56, 276 Reivitz, Linda, 46, 287n.15 religion, 249 Republicans: vs. Democrats, 252; election wins of 1994, 45, 115, 273; reforming style of, 252. See also politics of reform resources, 224–25, 229 resources specialists, 135–36 respect/morale, 212 RFS (right of first selection), 138–39, 140 Rhode Island, 228, 231, 266

Index Riemer, David, 52–53, 56, 116–17, 126, 148, 153, 167 right of first selection (RFS), 138–39, 140 rights, 257, 275 right vs. left. See conservatives vs. liberals Riley, Antonio, 53, 58, 116, 150, 247, 292n.99 Riverside County (California), 166, 220 Rogers, Jean, 86, 106, 108, 115, 135, 144, 146, 297n.2; on caseload declines, 179; MEP created by, 154, 312n.76 Roman Catholic Church, 118–19 Roosevelt, Franklin, 255 Roosevelt, Theodore, 250 Rousseau, Jean-Jacques: The Social Contract, 257 Rousseauian community, 257–59 sanctions, 121, 238–39, 240 San Diego, 34, 220 Saturation Work Initiative Model (SWIM; San Diego), 81, 83, 220 SCHIP (State Child Health Insurance Program), 123 school choice, 254 SCS (State Capacity Study; Rockefeller Institute of Government), 225–26, 331n.36 Select Committee on Work Incentives (Wisconsin), 51, 290n.63 Self-Sufficiency First. See SSF settlers, 246, 248–50, 267 Shalala, Donna, 254–55 Sharkansky, Ira, 332nn.45 and 46 Sheboygan County: caseload declines in, 182; paternalism in, 166; WEJT in, 166; welfare reform in, 79, 86–87, 92, 94, 96–97, 106, 300n.30 Sherraden, Michael, 32 Single Parent Employment Demonstration Program (SPED; Utah), 220 slavery, 265, 267 social contract, 56 The Social Contract (Rousseau), 257 social democracy, 275 socialist labor movement, 248 Social Security Act (1935), 255, 267, 269 Social Security Board, 269 sources of research, 10–13 South Carolina, 272 Southern politics, 245, 331n.41 Special Resource Account, 28, 32

•

349

SPED (Single Parent Employment Demonstration Program; Utah), 220 spoils system, 252 SSF (Self-Sufficiency First): assessment of, 285n.44; caseload declines due to, 77; creation of, 115, 134; origins of, 32, 45; scope of, 31; work requirements of, 30– 31, 74 SSI (Supplemental Security Income), 195 SSP (Support Services Planner), 136 State Capacity Study (SCS; Rockefeller Institute of Government), 225–26, 331n.36 State Child Health Insurance Program (SCHIP), 123 state-earned income tax credit, 254 state income tax, 253 steel industry, 22–23 Stepaniak, Jerome R., 289n.53 Stewart, Linda, 144, 312n.76 Strohl, Joe, 52, 247, 291n.78 Stumbras, Bernie, 292n.90 success, meaning of, 224–25, 232 successful welfare reform, definition of, 13–14, 281n.25 Supplemental Security Income (SSI), 195 Support Services Planner (SSP), 136 SWIM (Saturation Work Initiative Model; San Diego), 81, 83, 220 Tammany Hall (New York), 265 TANF (Temporary Assistance for Needy Families; formerly AFDC), 332n.43; caseload reductions, Wisconsin vs. national, 3–5, 207, 325n.49; creation of, 2; implementing, 63; time limit for, 148; vs. W-2, 107; work requirements for, 3–4, 9–10, 198, 230, 242 (see also work requirements, implementing) taxes, 252, 253 Taylor, Julia, 117 teenage parents, births to, 28 Temporary Assistance for Needy Families. See TANF Tennessee, 129, 231, 241 Ten-Point Coalition (Boston), 263 Texas, 228–29, 263 think tanks, 265 $30 and 1/6 Work Incentive, 27, 29 Thompson, Tommy: as Assembly leader, 43; on the bureaucracy, 75; on caseload declines, 195; election of, 24, 45, 115;

350

•

Index

Thompson, Tommy (cont.) as governor, 41–42, 43–44; initial welfare reforms of, 24–25, 283n.23; on Learnfare, 32; moderation of, 48; Progressivism of, 252; as Secretary of Health and Human Services, 41, 254; vetoes used by, 44, 127; on Weakland, 118; as a welfare reform leader, 6, 12, 25–28, 35, 43–46, 105, 108, 257–58 (see also W-2); on WNW, 29 traditionalistic states, 54, 188, 216, 218– 20, 233, 237–41, 249, 259, 332n.46 Tropman, Peter, 46–47, 55, 56, 287n.14, 292n.90 truancy, 160. See also Learnfare trust, 55–56 Turner, Jason, 32–33, 108–9, 113, 115, 222–23, 296n.37, 297n.2 20-Hour Work Requirement, 27, 33 Two-Tier AFDC Benefit Pilot Program, 28, 33, 44, 197 UMOS (United Migrant Opportunities Services), 143 unemployment. See employment, W-2’s effects on; labor market/unemployment unemployment insurance, 253, 254, 255 United Migrant Opportunities Services (UMOS), 143 University of Chicago, 335n.39 University of Wisconsin, 255–56 unwed pregnancies, 48, 187, 273, 290n.57 urban reform, 247 U.S. Department of Health and Human Services. See DHHS Utah, 220, 230 Vehicle Asset Limit Demonstration, 28 Vermont, 220 Virginia, 128, 267 Volk, Joe, 204 W-2: vs. AFDC, 107, 110–11; assessment of, 33, 285n.46 (see also MEP); award won by, 155; BadgerCare, 284n.29; blacks’ support of, 58; caseload reductions, Wisconsin vs. national, 3–5, 7, 279n.1 (see also welfare, decline of); case management in, 92; costs of, 110, 115, 124–25, 195, 289n.49, 305n.63; effects on former recipients, 3, 5; family eli-

gibility for, 110; individualism of, 110; influence/studies of, 5–6, 279n.4; innovativeness of, 3, 7, 107, 127–29, 155, 306n.76; paternalism of, 157, 171–72; political origins of, 52–53; protests against, 149; publicity surrounding, 5–6, 279n.3; sources for research on, 10–12; success/meaning of, 13; success of, connecting government to , 13–16, 281n.25 (see also good government); support for, 40–41; vs. TANF, 107; time limits on assignments under, 84; vision behind, 259; work requirements for, 3–4, 9–10 (see also work requirements, implementing). See also W-2, effects of; W-2, emergence of; W-2, implementing; W-2, implications of W-2, effects of, 17, 193–213; achieved despite disadvantage, 210–11; caseload declines, 193–94; on earnings/income, 199–200, 211, 212, 213; on employment, 198–99, 209–10, 322nn.18 and 20; evaluating reform, 197–98, 321nn.11 and 12, 325–26n.58; on families/children, 202, 211–13, 323n.29; fiscal, 194–97, 207, 321n.7; on hardship, 202–5, 323n.33; on marriage, 212–13; on morale/respect, 212; on poverty, 193, 200–201, 211, 213, 274, 323n.27; Wisconsin’s strengths, 207–10, 325nn.49, 51, and 53; Wisconsin vs. other states, 205–7, 325nn.47 and 48 W-2, emergence of, 45–46, 107–31; and bureaucratic competition, 114; and cash grants, 111, 120, 126; and child care, 114, 119, 123, 125–26; and child support, 111–12, 302n.15; costs of W-2, 124–25, 305n.63; demands/severity of W-2, 122–23, 131, 205, 304nn.53 and 54; effectiveness assessment, 129–31; eligibility for, 110, 119, 123, 128, 304n.48; and emergency assistance, 115; enactment of, 113–15; and entitlement, 109, 118, 122, 125; the four tiers, 111, 120, 122; generosity of W-2, 123–24, 128; and health care, 114, 119, 123–24; and implementation, 112–13; liberalization of W-2, 126; and the light touch, 110–11, 122, 171–72; and outside groups, 117–19; placements outside of, 121; planning, 107–9; politics of pas-

Index sage, 115–17; principles, 109–13, 302n.15; provisions, 119–21, 304n.48; and sanctions, 121; and time limits, 120, 128; uniqueness of W-2, 127–29, 306n.76; work-first emphasis, 126–27, 130–31; work goals/requirements, 109– 10, 122–23, 128, 130 W-2, implementing, 132–56; agencies’ fate, 133, 150–53, 312nn.69–71; aid denials, 145, 319n.21; bureaucratic competition by agencies, 132, 134, 137–39, 308n.15; and caseload declines, 132, 134–35 (see also welfare, decline of); child-care denials, 145–46; FEP, role of, 112–13, 136– 37, 145, 171–72, 230; low take-up of W-2, 144–45; and performance contracting, 139–41, 309n.27; preludes to W-2, 133–35; privatization in Milwaukee County, 138–39, 141–44, 309n.29, 310n.31; public reaction, 149–50; research/evaluation, 153–55, 312nn.74 and 76; sanctions, 133–34, 146–47; time limits, 148–49; the W-2 process, 135–37; work tests, weakening of, 147–48 W-2, implications of, 17, 261–77; and consent, 272–76, 339n.48; and the institutional frontier, 276–77; for national policy, 269–72, 277, 337n.29; and race, 266–69; social policy, direction forward in, 261–63; social policy, governmental preconditions for, 264–66 W-2, politics of, 113–19 Wagner, Bob, 47–48 Wagner, John, 108 Wahner, John W., 289n.53 Walker, Jack, 215 Walworth County, 138 War on Poverty, 255 Washington State, 230–31, 253 Waukesha County, 138 Weakland, Rembert, 118 Weber, Max, 106 WEJT (Work Experience and Job Training Program): caseload declines triggered by, 178, 185; creation/expansion of, 24–25, 66, 81, 83, 298n.7; education/training emphasized by, 34, 82; evaluation of, 197; funding for, 67; in Kenosha County, 81–84, 92, 298nn.15 and 17; origins of, 47, 51; participation goals of, 83; pater-

•

351

nalism of, 164–67, 315n.39; work requirements of, 69 welfare: definition of, 1; dependency on, 1, 3, 65; effective policies of, definition of, 13–14, 281n.25; entitlement view of, 13; institutional approach to studying, 8–10; opinion indicator on, 333n.1; opposition to, 13; protests of, 117; and race, 155; racial makeup of caseloads, 39; reform of (see PRWORA; W-2; welfare reform). See also TANF; welfare, decline of welfare, decline of, 2–3, 16–17, 175–92; county-based analysis of, 183–85, 319– 20nn.26–28, 320n.30; institutional effects on, 191–92; officials on, 178–81, 319nn.20, 21, and 24; research on caseload change, 175–77, 318n.10; trends over time, 181–83; W-2’s effects on, 193–94; Wisconsin caseload decline, 177–78, 318nn.15 and 16; Wisconsin’s uniqueness in, 188–91; in Wisconsin vs. other states, 177, 185–88, 320nn.33 and 35 welfare rebound, 322n.20 welfare reform, 1, 18–36; administrative demands of, 15; altruism of, 7–8; analysis limits for, 34–35; beginning of, 23, 283n.15; costs of, 48–49, 289n.49; by county (see counties, welfare reform by); diversion programs, 30–31, 32 (see also Work First); effectiveness of, 31–36; goals of, 18–20, 48–49, 289n.43; influential figures in, 292nn.90 and 92; liberal phase of, 21–23; at the local level (see counties, welfare reform by); paternalism of (see paternalism); policymaking as education, 35–36; policymaking casualness, 32–33, 285nn.43 and 44; politics of (see politics of reform); publicity for, 271–72; public support for, 39–41, 45, 289n.43; radical phase of, 29–30; Thompson’s initial reforms, 24–25, 283n.23; turnover research on, 32–33, 285n.41; waiver experiments, 25–28, 33, 176–77, 197, 219–20, 254, 284n.29, 319n.20; welfare migration issue, 23–24, 28, 42–43, 48, 50, 283n.17, 289n.53; Wisconsin history of, 14–15 (see also PRWORA; W-2); work-enforcing reform instituted, 23–24; work goals/requirements of, 3–4, 9–10, 13, 18, 19–20,

352

•

Index

welfare reform (cont.) 31–33, 34–35 (see also work requirements, implementing) Welfare Reform Advisory Committee (Wisconsin), 23 welfare states, European, 247–48, 275 Welfare-to-Work, 129 Welfare Warriors, 59, 293n.103 WEOP (Wisconsin Employment Opportunities Program), 65–66, 67 West Virginia, 229, 230–31, 254 Whitburn, Gerald, 32, 44, 49, 56, 75, 84, 108; Bi-Partisan Legislative Working Group formed by, 113; on caseload declines, 179; on work requirements vs. hardship, 132 WIA (Workforce Investment Act; 1998), 93, 230–31 Wilson, William Julius, 56 Wilson, Woodrow, 250 WIN (Work Incentive Program), 64–65, 66–67, 69–70 WIN Demonstration Program (Chicago), 307n.3 Winnebago County, 88, 182 Winstead, Don, 338n.33 Wisconsin, 245–60; anger about welfare in, 246; background of, 13–14, 245–60, 281n.25; benefit raises/cuts in, 21–22, 24–25, 49; benefit terminations in, 77; budget deficit in, 62; bureaucracy in, 252–53; conservativism/liberalism of, 233n.1, 245–47, 257; coordination in, 230; counties, map of, 26; Democratic vs. Republican control of, 41–43; governors of, 41–42; immigrant settlers in, 246, 249–50; liberal culture of, 328n.18; moralistic culture of, 12–14, 54, 79, 216, 220, 221–23, 249; national leadership of, 253–54, 260; political corruption in, 62; political development of, 253; Progressivism in, 250–53; reforming tradition in, 250–51, 260; Rousseauian community in, 257–59; support of residents for welfare reform, 39–41; taxes in, 252; the “Wisconsin idea,” 255–56, 335n.39 Wisconsin Catholic Conference, 118 Wisconsin Council on Children and Families, 117 Wisconsin Employment Opportunities Program (WEOP), 65–66, 67

Wisconsin Policy Research Institute, 108 Wisconsin Shares, 126 Wisconsin Works. See W-2 Wiseman, Michael, 154, 292nn.90 and 92, 293n.108, 299n.28, 312nn.74 and 76 Witte, Edwin, 255 WNW (Work Not Welfare): and abolition of AFDC/Food Stamps/General Relief, 53; assessment of, 285n.43; creation/ expansion of, 105; Democratic support for, 46; effects on caseloads, 183; in Fond du Lac County, 88–89, 92, 95, 301n.46; origins of, 32; paternalism of, 29–30; political motives for, 44, 53; scope of, 30; work goals/requirements of, 29, 32; as a W-2 pilot, 109 women, earnings/employment of, 198, 201, 210 work ethic, 256 Work Experience and Job Training Program. See WEJT Work First: caseload declines due to, 76– 77; creation/expansion of, 30, 74, 82, 298n.10; in Douglas County, 90; effects on caseloads, 183; in Grant County, 298n.10; in Jackson County, 298n.10; in Kenosha County, 82; in Milwaukee, 134; origins of, 32; in Sheboygan County, 86–87; in Winnebago County, 88 Workforce Attachment and Advancement Program, 127 Workforce Investment Act (WIA; 1998), 93, 230–31 Work Incentive Program. See WIN work levels: for single vs. married mothers, 210; in Wisconsin vs. other states, 207– 9, 325n.51 workmen’s compensation, 253, 255 Work Not Welfare. See WNW Work Programs Reporting System (WPRS), 68 work requirements, implementing, 63–78; administrative costs of, 69; bureaucracy for, buildup of, 66–69, 77–78; bureaucracy for, motivation/competence of, 75– 76, 77–78, 296n.35, 296–97nn.37–39; and diversion from welfare, 73–74, 76– 77 (see also Work First); early programs, 64–66; effects of work requirements, 76–77; funding for, 67–68; management information systems for, 68; paternalism

Index of (see paternalism); the problem of implementation, 64; raising participation, 69–71, 296nn.26 and 27; studies of, 63– 64; the work-first concept, 71–73. See also JOBS work tests, 26, 27, 29

WPRS (Work Programs Reporting System), 68 Wyoming, 190, 207 YWCA, 118, 141 YW Works, 143

•

353