The Economic Status of the Aged [Reprint 2020 ed.] 9780520349810


159 34 60MB

English Pages 320 Year 2020

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

The Economic Status of the Aged [Reprint 2020 ed.]
 9780520349810

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

The Economic Status of the Aged

A Publication of the Institute of Industrial Relations University of California

m Economic Status of the Aged P E T E R O. S T E I N E R A N D R O B E R T

DORFMAN

Berkeley and Los Angeles, 1957 UNIVERSITY OF CALIFORNIA

PRESS

U N I V E R S I T Y OF C A L I F O R N I A P R E S S B E R K E L E Y A N D LOS A N G E L E S CALIFORNIA CAMBRIDGE UNIVERSITY PRESS LONDON, E N G L A N D ©

1 9 5 7 , BY

T H E R E G E N T S OF T H E U N I V E R S I T Y OF C A L I F O R N I A L I B R A R Y OF CONGRESS C A T A L O G C A R D N U M B E R :

57-62IO

P R I N T E D I N T H E U N I T E D S T A T E S OF A M E R I C A B Y T H E U N I V E R S I T Y OF C A L I F O R N I A P R I N T I N G D E P A R T M E N T

To Our Parents

Foreword

In 1950, the Institute of Industrial Relations of the University of California received a sizable grant from the Rockefeller Foundation to conduct a five-year interdisciplinary study of the problem of aging in an industrial society. The plans for the project were formulated under the leadership of Chancellor Clark Kerr, who was then Director of the Institute, and his Associate Director, the late Lloyd H. Fisher. The separate studies which eventually emerged as independent subdivisions of the project dealt with the economic status of the aged, the politics of the aged, the relationship of physiological and psychological age to chronological age, the social and psychological aspects of aging and retirement, employer and union policies toward the older worker, and retirement policy under Social Security legislation. The responsibility for guiding the project in its final stages has fallen chiefly to Dr. Margaret S. Gordon, now Associate Director of the Institute. The growing concern over the problem of aging stems in large part from recognition of the fact that, in modern industrial societies, including the United States, the proportion of aged persons in the population has been rising for many decades. In 1880, as Steiner and Dorfman point out, less than four per cent of the American population was 65 years old or older. By 1950, this proportion had increased to more than eight per cent and, in the predominant opinion of population experts, would go on increasing throughout the second half of the century, vii

viii

|

Foreword

Meanwhile, as a result of profound social and economic changes, the proportion of the aged population in the labor force was declining. The responsibility for providing for the economic needs of older persons was gradually being shifted from the aged themselves and their families to society as a whole. There was growing anxiety that, as the population continued to "age," the provision of adequate income for retired persons would become an increasingly heavy burden for the economy. Out of the rising awareness of these developments came a heightened interest in the possibility of expanding employment opportunities for those who had passed the conventional retirement age of 65. If the trend toward earlier withdrawal from the labor force could be reversed, it was argued, not only would a larger proportion of aged persons provide for their own economic needs and at the same time contribute to the total output of the economy, but they would also achieve a more satisfactory social and psychological adjustment in a job-oriented society. These views strongly influenced the development of plans for the Institute's project. The authors of the present volume set out to determine precisely what the labor-force and income status of the population aged 65 or older actually was. Finding the available statistical data inadequate, they arranged with the Bureau of the Census to conduct a special nationwide follow-up survey of persons 65 or older, in connection with the Current Population Survey of April, 1952. The result is one of the most valuable sets of data available on the economic problems of old age. The Steiner-Dorfman findings challenge several of the widely accepted views relating to the aged problem. One is the notion that the population will go on "aging" throughout the second half of the present century. Another is the prevailing impression that the essential core of the aged problem is the elderly man who has been forced into premature retirement through "mere age." Still another is the optimistic expectation that large numbers of older persons can be kept in the labor force if only the "artificial barriers" to their employment can be removed. On the basis of their findings, the authors argue cogently that a basic solution to the problem of economic distress in old age will be found only if the main focus of attention is shifted from the aged themselves to the earlier years of working life.

Foreword

| ix

A preliminary version of some of the authors' findings was published earlier in the American Economic Review.1 These articles attracted a great deal of interest and generated considerable controversy. It is recognized that many will disagree with the conclusions of Professors Steiner and Dorfman. In view of the comprehensive nature of their evidence and the care with which it has been analyzed, however, their findings are entitled to serious consideration by all who have an interest in the problems of the aged. Although the authors conclude that the most numerous and needy victims of the aging process in our society are the elderly widows and incapacitated older men whose plight cannot be cured merely by expanding employment opportunities, they do not deprecate the advantages of a flexible retirement policy for those workers who are capable and desirous of continuing in the labor force after some particular age has been reached. Even if only a small proportion of elderly workers will be affected by a flexible policy, the benefits to them and to the economy will be well worth achieving. Yet, as the authors urge, our main efforts must be directed elsewhere. We must arm our young and mature workers against the inroads of the aging process by providing them with adequate pension and insurance programs, with a clear understanding of their need to augment social provisions by individual saving, and, above all, with a flexible training and outlook that will enhance their ability to adapt to the inevitable changes in themselves and the world around them. Both of the authors of this volume were affiliated with the research staff of the Institute of Industrial Relations at the time the study was conducted. Peter O. Steiner is a member of the Economics Department of the University of California, Berkeley, and Robert Dorfman is now a member of the Economics Department at Harvard University. Arthur M. Ross, Director, Institute of Industrial Relations University of California

1 Robert Dorfman, "The Labor Force Status of Persons A g e d Sixty-five and Over," and Peter O. Steiner, "The Size, Nature, and Adequacy of the Resources of the A g e d , " American Economic Review, X L I V (May 1954), pp. 634-660.

Acknowledgments

This study was undertaken while the authors were associated with the Institute of Industrial Relations, University of California, Berkeley, and was materially assisted by a grant from the Rockefeller Foundation. In addition to institutional assistance, we have received many helpful suggestions during the last six years from numerous individuals, none of whom is responsible either for the views expressed or for any errors that may remain in the completed work. We are deeply grateful to the members of the Bureau of the Census, many of them unknown to us, whose extraordinary cooperation in the planning, enumeration, and tabulation of the Follow-up Survey is responsible for much of the value of the study. In particular, our thanks to Howard Brunsman, Henry Shryock, Gertrude Bancroft, and Herman Miller, whose patience and help exceeded reasonable bounds. Mrs. Edna Wentworth and other members of the staff of the Bureau of Old Age and Survivors Insurance made suggestions that improved the design of the survey and facilitated comparability of our data with that from related surveys. A number of graduate students at the University of California have made contributions to the study as Graduate Research Assistants of the Institute of Industrial Relations. Bennett Berger, Bernard Hall, Sidney Klein, Roy Ockert, Mrs. Judith Riggs, Joseph xi

xii

|

Acknowledgments

Rosenbaum, and Bruce Taft each spent a year or more on the project, and we appreciate their contributions. Alice Lundin carried out the painstaking task of putting the appendix tables into final form. Mrs. Ruth R. Steiner drafted the figures. Professor Melvin W. Reder of Stanford University made many contributions to the design of the study as well as giving us the benefit of a critical reading of the manuscript. Dr. Margaret S. Gordon read the entire manuscript at least twice, and her comments and suggestions have led to many improvements; more recently, as Acting Director of the Institute, she has assisted in many other ways. Professors W. L. Crum, Emily Huntington, and Harvey Leibenstein, each read major parts of the manuscript and provided helpful comments. Mr. John Gildersleeve, of the University of California Press, made manful efforts to improve the readability of the manuscript, and difficult passages that remain are due to our stubborn refusal to accept some of his suggestions. To all of the above, and others, we are grateful. But our greatest debt is to Lloyd H. Fisher. His premature death robbed us of his friendship and his help. The stimulus we have received from his extraordinary mind and personality will never leave us, and our hope is that the manuscript would have satisfied him. Peter O. Steiner Robert Dorfman September 5 , 1 9 5 6 Cambridge, Massachusetts

Contents

I

INTRODUCTION

1

Outline of the report, 5 II

THE TREND OF THE OLDER POPULATION

7

The number of the aged, 7; The proportion of the aged, 11; The age composition of the aged, 12 III

SOME CHARACTERISTICS OF THE AGED

14

Sex and marital status, 15; Living arrangement (family status), 20; Location and community type, 23; Education, 30; Nativity and color, 31; Socioeconomic status (longest occupation), 34; Age distribution, 35 IV

THE LABOR-FORCE STATUS OF OLDER MEN

38

A general view, 38; Occupation, health, and laborforce participation, 44; Reasons for retirement, 49; Occupational distribution and changes, 51; Relationship of marital status to labor-force status, 55; Extent of part-time employment, 58 V

THE LABOR-FORCE STATUS OF OLDER WOMEN

60

Status of women with work experience, 65 VI

THE RECEIPTS OF THE AGED

Needs of the aged, 68; The level and distribution of receipts, 72; Dependency status of the aged, 81; Economic status of the aged, 82 xiii

67

xiv VII

VIII

|

Contents

SUPPORT STATUS OF AGED UNITS

86

SOURCES OF RECEIPTS

93

Basic data, 93; Differences in economic position, 95; Are these differences significant?, 107; The role of particular sources, 111; Long term expectations, 121 IX

THE SIZE AND ADEQUACY OF THE ASSETS OF THE AGED

12g

The asset position of the aged, 130; Liquid asset holdings, 131; Sources of vulnerability: death of a spouse, 134; Life insurance, 134; Liquid assets, 138; Sources of vulnerability: illness, 142; Conclusion, 145 X

CONCLUSION

146

Appendix A :

SOURCE AND RELIABILITY OF DATA

155

Appendix B:

PART I, FIFTY YEAR PROJECTION OF POPULATION

168

PART N, EFFECT OF ERROR IN FORECAST MORTALITY

Appendix C : Appendix D :

RATES

186

DEVELOPMENT OF BUDGETS

192

COMPARISON OF THE FOLLOW-UP SURVEY OF THE AGED WITH THE SURVEY OF NATIONAL OLD AGE AND

Appendix E:

SURVIVORS' INSURANCE BENEFICIARIES

207

BASIC DATA OF THE FOLLOW-UP SURVEY

220

Index

295 TEXT TABLES

2.1. Population aged 65 and over, United States, 1930-2000, by sex

8

2.2. Trend of age specific death rates, 1900-1950, United States death registration area

8

2.3. Death rates for men and women, 1940 and 1950

11

2.4. Age composition of the aged, United States, 1930-2000

13

3.1. Number of persons aged 65 and over, 1880-1975

15

3.2. Marital status of persons aged 65 and over, by sex, April, 1952

17

3.3. Marital status of persons aged 65 and over, 1890-1950

18

xiv VII

VIII

|

Contents

SUPPORT STATUS OF AGED UNITS

86

SOURCES OF RECEIPTS

93

Basic data, 93; Differences in economic position, 95; Are these differences significant?, 107; The role of particular sources, 111; Long term expectations, 121 IX

THE SIZE AND ADEQUACY OF THE ASSETS OF THE AGED

12g

The asset position of the aged, 130; Liquid asset holdings, 131; Sources of vulnerability: death of a spouse, 134; Life insurance, 134; Liquid assets, 138; Sources of vulnerability: illness, 142; Conclusion, 145 X

CONCLUSION

146

Appendix A :

SOURCE AND RELIABILITY OF DATA

155

Appendix B:

PART I, FIFTY YEAR PROJECTION OF POPULATION

168

PART N, EFFECT OF ERROR IN FORECAST MORTALITY

Appendix C : Appendix D :

RATES

186

DEVELOPMENT OF BUDGETS

192

COMPARISON OF THE FOLLOW-UP SURVEY OF THE AGED WITH THE SURVEY OF NATIONAL OLD AGE AND

Appendix E:

SURVIVORS' INSURANCE BENEFICIARIES

207

BASIC DATA OF THE FOLLOW-UP SURVEY

220

Index

295 TEXT TABLES

2.1. Population aged 65 and over, United States, 1930-2000, by sex

8

2.2. Trend of age specific death rates, 1900-1950, United States death registration area

8

2.3. Death rates for men and women, 1940 and 1950

11

2.4. Age composition of the aged, United States, 1930-2000

13

3.1. Number of persons aged 65 and over, 1880-1975

15

3.2. Marital status of persons aged 65 and over, by sex, April, 1952

17

3.3. Marital status of persons aged 65 and over, 1890-1950

18

Contents

[

xv

3.4. Sex and marital status of individuals 65 and over, April, 1952

18

3.5. Median income of aged economic units by sex and marital status, 1952

20

3.6. Living arrangements of aged economic units, April, 1952

22

3.7. Aged and rate of population growth, 1940-1950, by census regions

23

3.8. Per cent of population 65 and over, 1950, and related economic data, by states

24

3.9. Aged persons by community type, 1950

28

3.10. Residence of aged economic units in sample, April, 1952

29

3.11. Level of school completed, persons 25 years and older and persons 65 years and older, 1950

30

3.12. Nativity of older persons and age of foreign-born persons, 1910-1950

31

3.13. Aged economic units: nativity and years in the United States, April, 1952

32

3.14. Socioeconomic status (longest occupation) of aged economic units with work experience, April, 1952

35

3.15. Age distribution and marital status of the aged, 1950

36

4.1. Older men by labor-force status, April, 1952

39

4.2. Distribution of men 65 and over by total unit income and relation of income to labor-force availability, April, 1952

40

4.3. Labor-force status, males 65 and over, by longest occupation group, April, 1952

41

4.4. Occupational distribution of males 65 and over who have ever worked and of employed males 14 and older, April, 1952

42

4.5. Labor-force status, males 65 and over, by type of community, April, 1952

42

4.6. Distribution of sample cases by longest occupation and laborforce status, April, 1952

45

4.7. Labor-force status and reasons for retirement;, men 65 and over, April, 1952

48

xvi

|

Contents

4.8. Labor-force status of older men retired under compulsory systems, April, 1952

49

4.9. Reasons for retirement for males 65 and over, by longest occupation group, April, 1952

50

4.10. Comparison of longest occupation with present or last occupation, men 65 and over, April, 1952

54

4.11. Labor-force status by marital status and longest occupation, men 65 and over, April, 1952

56

4.12. Labor-force participation of white males, 65 and over, by marital status, age group, and community type, 1950

57

4.13. Extent of labor-force participation, men 65 and over, 1951

58

4.14. Interest in full- or part-time work, men 65 and over not in the labor force, April, 1952

59

5.1. Labor-force status, women 65 and over, April, 1952

61

5.2. Labor-force status, women 65 and over, by community type, April, 1952

61

5.3 Older men and women by amount of work in 1951

62

5.4. Per cent in labor force, women and men 65 and over with work experience, by longest occupation, April, 1952

62

5.5. Women 65 and over who have worked since age 50, reasons for retirement, by marital status, April, 1952

63

5.6. Occupational distribution of women 65 and over with work experience, and of employed women 14 and over, April, 1952

63

5.7. Retirement status of women 65 and over who have worked since age 50, by longest occupation and marital status, April, 1952

64

5.8. Labor-force status of women 65 and over with work experience who are not in the labor force, April, 1952

66

6.1. Sample budget figures for selected subgroups of aged population

71

6.2. Aged with total receipts below budget levels, 1951

80

6.3. Units with receipts below lower subsistence budget

80

6.4. Economic status

84

Contents

I xvii

7.1. Support status

87

7.2. Support status: summary

88

7.3. Support status of the independent

88

7.4. Independent units above and below budget levels, by support status

90

7.5. Sources of support of the not self-supporting

91

8.1. Number of sources of receipts, 1951

95

8.2. Frequency of receipt of specified sources of receipts, 1951

96

8.3. Aggregate value of receipts, 1951

97

8.4. Size distribution of selected sources of receipts, 1951

98

8.5. Average (mean) total receipts of those with specified principal major sources of receipts, 1951

99

8.6. Form of occurrence of principal sources of receipts, 1951

100

8.7. Unrelated females with and without earnings, 1951

101

8.8. Sources of receipts: summary

104

8.9. Unrelated males with and without earnings, 1951

105

8.10. Couples with and without earnings, 1951

105

8.11. Principal major source of receipts for units classified according to socioeconomic status

108

8.12. Principal major source of receipts for units classified according to family status

111

8.13. Selected data on sources of receipts: classified by source, 1951

112

8.14. Impact of dissaving: where dissaving is principal major source of receipts, 1951

116

8.15. Impact of dissaving: where dissaving is a secondary source, 1951

"7

8.16. Dissaving as a secondary source: relative impact of dissaving and other secondary sources

118

9.1. Net value of total assets, April, 1952

130

9.2. Home ownership, April, 1952

130

9.3. Frequency of holdings of liquid and nonliquid assets

132

xviii

|

Contents

9.4. Holdings of liquid assets

132

9.5. Size of liquid asset holdings

133

9.6. Frequency and size of life insurance holdings of male members of couples, April, 1952

134

9.7. Life insurance and income of couples

136

9.8. Life insurance and income of couples: consolidated income classes

137

9.9. Comparison of life insurance and liquid asset holdings of couples

140

9.10. Combined holding of life insurance and liquid assets of couples

140

9.11. Medical service and medical payments of aged units, 1951

143

9.12. Annual medical payments of aged units, 1951

143

9.13. Noncash medical services received by aged units, 1951

144

9.14. Medical debts of aged units, April, 1952

145

TEXT FIGURES

1.1. Per cent of the population of the United States 65 years old and older, 1880-2000

2

2.1. Number of men and women 65 and over, United States, 19302000

10

3.1. Growth of the aged population of the United States, 18801950, with projections to 1975

16

3.2 Scatter diagram: per cent of population 65 and over and per capita income (dollars) by states, 1950

26

3.3. Scatter diagram: per cent of population 65 and over and per cent increase in total income, by states, 1940-1950

27

4.1. Occupational mobility: inflow and outflow rates of ten occupational groups, men 65 and over

52

6.1. Comparison of total receipts and total incomes, aged economic units, 1951

73

xviii

|

Contents

9.4. Holdings of liquid assets

132

9.5. Size of liquid asset holdings

133

9.6. Frequency and size of life insurance holdings of male members of couples, April, 1952

134

9.7. Life insurance and income of couples

136

9.8. Life insurance and income of couples: consolidated income classes

137

9.9. Comparison of life insurance and liquid asset holdings of couples

140

9.10. Combined holding of life insurance and liquid assets of couples

140

9.11. Medical service and medical payments of aged units, 1951

143

9.12. Annual medical payments of aged units, 1951

143

9.13. Noncash medical services received by aged units, 1951

144

9.14. Medical debts of aged units, April, 1952

145

TEXT FIGURES

1.1. Per cent of the population of the United States 65 years old and older, 1880-2000

2

2.1. Number of men and women 65 and over, United States, 19302000

10

3.1. Growth of the aged population of the United States, 18801950, with projections to 1975

16

3.2 Scatter diagram: per cent of population 65 and over and per capita income (dollars) by states, 1950

26

3.3. Scatter diagram: per cent of population 65 and over and per cent increase in total income, by states, 1940-1950

27

4.1. Occupational mobility: inflow and outflow rates of ten occupational groups, men 65 and over

52

6.1. Comparison of total receipts and total incomes, aged economic units, 1951

73

Contents

|

xix

6.2. Distribution of total receipts, by type of residence, couples, 1951

74

6.3. Distribution of total receipts, by type of residence, unrelated males, 1951

75

6.4. Distribution of total receipts, by type of residence, unrelated females, 1951

76

6.5. Distribution of total receipts, by family status, couples, 1951

77

6.6^ Distribution of total receipts, by family status, unrelated males, 1951

78

6.7. Distribution of total receipts, by family status, unrelated females, 1951

79

8.1-. Size distribution of receipts, by source of receipts, 1951

102

9.1. Face value of life insurance, by income, couples, 1951

139

APPENDIX TABLES

A.i. Standard error of estimated numbers, April, 1952, Follow-Up survey

157

A.2. Illustrative computation of per cent distribution

159

A.3.- Standard error of estimated percentages, April, 1952, FollowUp survey

162

A.4. Standard error of selected medians, April, 1952, Follow-Up survey

163

B.i. Survival relationships, women, 65-69

170

B.2. Fertility weights used in projecting births

172

B.3. Adjustments for increasing cohorts in early age groups

175

B.4. Projection equations for middle age groups

176

B.5. Projection of the male population of the United States, by age, 1950-2000: assumption A, higher fertility

180

B.6. Projection of the female population of the United States, by age, 1950-2000: assumption A, higher fertility

181

B.7. Projection of the total population of the United States, by age, 1950-2000: assumption A, higher fertility

182

Contents

|

xix

6.2. Distribution of total receipts, by type of residence, couples, 1951

74

6.3. Distribution of total receipts, by type of residence, unrelated males, 1951

75

6.4. Distribution of total receipts, by type of residence, unrelated females, 1951

76

6.5. Distribution of total receipts, by family status, couples, 1951

77

6.6^ Distribution of total receipts, by family status, unrelated males, 1951

78

6.7. Distribution of total receipts, by family status, unrelated females, 1951

79

8.1-. Size distribution of receipts, by source of receipts, 1951

102

9.1. Face value of life insurance, by income, couples, 1951

139

APPENDIX TABLES

A.i. Standard error of estimated numbers, April, 1952, Follow-Up survey

157

A.2. Illustrative computation of per cent distribution

159

A.3.- Standard error of estimated percentages, April, 1952, FollowUp survey

162

A.4. Standard error of selected medians, April, 1952, Follow-Up survey

163

B.i. Survival relationships, women, 65-69

170

B.2. Fertility weights used in projecting births

172

B.3. Adjustments for increasing cohorts in early age groups

175

B.4. Projection equations for middle age groups

176

B.5. Projection of the male population of the United States, by age, 1950-2000: assumption A, higher fertility

180

B.6. Projection of the female population of the United States, by age, 1950-2000: assumption A, higher fertility

181

B.7. Projection of the total population of the United States, by age, 1950-2000: assumption A, higher fertility

182

xx

|

Contents

B.8. Projection of the male population of the United States, by age, 1950-2000: assumption D, lower fertility

183

B.g. Projection of the female population of the United States, by age, 1950-2000: assumption D, lower fertility

184

B.10. Projection of the total population of the United States, by age, 1950-2000: assumption D, lower fertility

185

B.11. Probability of surviving until 2000

190

C . i . Index of budget ratios

201

C.2. Total "standard" budgets

204

C.3. Cash equivalent "standard" budgets

205

C.4. "Subsistence" budgets

206

D.i. Units of analysis in two surveys

209

D.2. Income distributions in two surveys: selected class intervals

214

D.3. Relationship between income and liquid asset holdings of the aged, 1946

215

D.4. Comparison of net worth (Beneficiary survey) with net value of total assets (Follow-Up survey), 1951

216

D.5. Home ownership: Beneficiary and Follow-Up surveys

218

(For list of tables presenting the basic data of the Follow-Up survey, see page 221.)

APPENDIX FIGURES

B.i. Survival relationships, women 65-69

171

B.2. Relation of women of childbearing age to children under five, high fertility assumption

173

D.i. Comparison of the size distribution of income in two surveys of the aged, couples, 1951

211

D.2. Comparison of the size distribution of income in two surveys of the aged, unrelated males, 1951

212

D.3. Comparison of the size distribution of income in two surveys of the aged, unrelated females, 1951

213

xx

|

Contents

B.8. Projection of the male population of the United States, by age, 1950-2000: assumption D, lower fertility

183

B.g. Projection of the female population of the United States, by age, 1950-2000: assumption D, lower fertility

184

B.10. Projection of the total population of the United States, by age, 1950-2000: assumption D, lower fertility

185

B.11. Probability of surviving until 2000

190

C . i . Index of budget ratios

201

C.2. Total "standard" budgets

204

C.3. Cash equivalent "standard" budgets

205

C.4. "Subsistence" budgets

206

D.i. Units of analysis in two surveys

209

D.2. Income distributions in two surveys: selected class intervals

214

D.3. Relationship between income and liquid asset holdings of the aged, 1946

215

D.4. Comparison of net worth (Beneficiary survey) with net value of total assets (Follow-Up survey), 1951

216

D.5. Home ownership: Beneficiary and Follow-Up surveys

218

(For list of tables presenting the basic data of the Follow-Up survey, see page 221.)

APPENDIX FIGURES

B.i. Survival relationships, women 65-69

171

B.2. Relation of women of childbearing age to children under five, high fertility assumption

173

D.i. Comparison of the size distribution of income in two surveys of the aged, couples, 1951

211

D.2. Comparison of the size distribution of income in two surveys of the aged, unrelated males, 1951

212

D.3. Comparison of the size distribution of income in two surveys of the aged, unrelated females, 1951

213

Chapter I

Introduction

The fable of the Hydra may be taken as a parable of social progress. He who fought the Hydra found that each time he cut off a head two new ones grew in its place. In the same way, each time we solve a problem in the struggle for social and economic progress the solution brings in train new, unexpected, and often tragic difficulties. The "problem of an aging population" is just such a dismal sideeffect of progress in other areas. It has its origins in three major developments. First is progress in medicine and public health, which has now reached the point where more than two-thirds of persons who survive infancy will attain the age of 65, and many will live well beyond. Second is the development of highly mechanized methods of production, which entail large, bureaucratized firms and technically adept work forces. Third is the cityward drift of population and the urbanization of our culture. The first of these developments has resulted in a large population of persons in the seventh and later decades of life. Because of the second, our economy has little place for those with the abilities and limitations of older workers. Because of the third, retired older persons find themselves in increasingly unfriendly surroundings. Out of these forces has arisen a new aspect of poverty in the midst of plenty—the plight of the older person in a modern, mechanized, urbanized society. The purpose of this monograph is to throw some light on this predicament.

1

Introduction

| 3

The problem has arisen suddenly, as social problems go. In 1880 less than 4 per cent of the American population was 65 or older. By 1950 this proportion had doubled. Figure 1.1 traces the growth of the older age groups as a proportion of the entire population. Until 1920 the rise was gradual. Then came twenty years of rapid increase followed by a barely perceptible slackening in the decade 1940-1950. It has been generally expected that the proportion of aged persons in the population would continue to increase during the second half of the present century. But the sustained high birth rates of the past fifteen years or so have meant that the youngest sector of the population has been growing rapidly. If high birth rates continue, the result may well be that the proportion of aged may cease to increase appreciably. The number of aged persons, however, will continue to rise rapidly, because of the increase in life expectancy. The fact that a group of social and economic problems now affects fourteen million people (the estimate for 1955) would command attention even if there were no other unusual aspects to the situation. A new and large social class—the retired class—is being created; a class with its own peculiar needs, interests, and demands, whose impact on the whole society increases as the class grows. Recognition of a new problem creates a need for new information. To be sure, the economic surveys conducted by the Bureau of the Census, the Bureau of Labor Statistics, the Survey Research Center, and state and local agencies, almost always include age among the variables studied and therefore contain much information about the aged. But the rest of the variables are, in generalpurpose surveys, so selected and classified as to be appropriate for understanding the needs and conditions of people in the active years of life. For this reason, these surveys are not well adapted to a study of the needs and conditions of older persons. A survey of their economic status must be specially designed in four principal respects. First, in economic surveys it is usual to use either the family or the individual as the basic economic unit. Neither is an appropriate unit for many older persons. An old couple frequently constitutes a subfamily living within a larger family composed of themselves, their adult children, and perhaps grandchildren. Thus the family is too broad an economic unit to use in the study of older persons.

4 | Introduction Yet, at the same time, the individual is too narrow a unit, for the economic resources of an aged couple must be considered jointly if an adequate picture of the economic status of either member is to be obtained. A newly defined economic unit, the couple, is required. Second, it is usual to rely on occupation as an indicator of socioeconomic status. For age groups in which at least the males are virtually all in the labor force this device is usually satisfactory. But a large percentage of older persons are not in the labor force, and many who are in the labor force are no longer in the occupational group to which they belonged during their years of peak activity. Thus it is not occupation but "normal occupation," in some sense, which must be used to indicate the socioeconomic status of older persons. In the present study, longest occupation is so used. Third, although income may be a sufficiently good indicator of economic status for most of the population, when applied to the aged it leaves many questions unanswered. For older persons the appropriate measure of economic resources includes the total amount available for consumption, including not only income, but also use of accumulated assets and any nonmonetary support or contributions received. Finally, the economic status of aged persons is so critically related to their health that no satisfactory analysis of their economic status can be made without information about how the state of their health affects earning capacity and expenses. These requirements made it necessary to conduct a nationwide survey of the noninstitutional aged population, specially designed to throw light on their economic status and problems, in order to provide some of the facts requisite for intelligent discussion of the problem of the aged and for formulation of a constructive social policy. In response to this need for data the Institute of Industrial Relations of the University of California contracted with the Bureau of the Census to conduct a nationwide survey of the financial status of households that included one or more persons aged 65 or over. Briefly stated: during April, 1952, there were special follow-up interviews with 60 per cent of those households in the Current Population Sample of the Census Bureau that contained members at least 65 years old. That survey, which provided the factual foun-

Introduction

| 5

elation of the present report, will be referred to throughout as the Follow-Up Survey: it included approximately 3,600 men and women aged 65 and over.1 Outline of the Report In chapter ii, by way of indicating the future magnitude of the problem of the older population, some projections of the American population by age and sex extending to the year 2000 are offered. The projections as far as 1975 are those of the Census Bureau. The extension to 2000 is presented without any pretense to refinement or accuracy, characteristics which are, in any event, unattainable. Chapter iii is devoted to general background material on the characteristics of the older members of the population, drawing heavily on Census Bureau reports. This material delineates the context in which the survey and the characteristics disclosed by it are embedded. It was, of course, impossible to analyze the available data in all conceivable ways, and it was therefore necessary to decide early in the study which lines of analysis were to be exploited thoroughly and which would have to be pursued less extensively or not at all. The most important of these decisions, and their justifications, are also discussed in this chapter. The presentation of the substantive findings of the survey begins in chapter iv and v, which discuss the labor-force status of men and women 65 and over, with special attention to the characteristics of those not in the labor force. These chapters are intended to help answer two key questions: Why are they not in die labor force? What are the possibilities of returning them to work? Chapters vi to ix concern the financial condition of the older members of the population. The main questions are: How large are their incomes? From what source are these incomes derived? What determines the size of an older person's income? What are their expenses and how great are they? If expenses exceed income how is the deficit financed? Do older persons have substantial financial resources to draw upon for ordinary living expenses or for the emergencies of life? 1 This survey was designed by the authors, was conducted by regular census enumerators, and was financed by a grant from the Rockefeller Foundation. Appendix A presents detailed information about the survey, and many of the basic data are presented in Appendix E. Text tables, unless otherwise indicated, are based upon the Follow-Up Survey and are presented without indication of source.

6

| Introduction

The report concludes, as a wise convention dictates that it must, with a recapitulation and interpretation of the findings. In particular we have tried, in the final chapter, to suggest the appropriate frame of reference for future research and ameliorative efforts. As this outline indicates, we do not attempt to present either a history or a theory of the economic problems of aging on the one hand, or a program for dealing with them, on the other. This report can best be described as a picture of the predicament of the aged in the United States in the year 1 9 5 2 with some discussion of the nature of their problems and some suggestions for solutions.

Chapter II

The Trend of the Older Population

The Number of the Aged The censuses of population taken in 1930, 1940, and 1950 provide convenient benchmarks for looking ahead. In the two decades 1930-1950, the number of persons 65 and over practically doubled, increasing from 6,630,000 to 12,270,000 (see table 2.1). The outlook for the future cannot be stated with as much certainty as the course of the recent past. Nevertheless, the Bureau of the Census has prepared forecasts of plausible trends of population growth up to 1975 and we have extended these forecasts to the year 2000.1 So far as these population forecasts relate to the older population, the persons whose numbers are being estimated have already been born and we believe that the estimates should be accorded a fairly high degree of confidence. The only important demographic factor affecting the number of older persons in this country during the next half century will be the death rate, and the forecasts presented depend heavily on the assumptions made concerning the future trend in mortality among older and middle-aged persons. Although death rates have proved to be relatively stable, as demographic factors go, they are by no means constant over time. As table 2.2 shows, progress in the arts of medicine and public health has had dramatic results throughout 1 U. S. Bureau of the Census, Current Population Reports, Series P-25, No. 78 (August 2i, 1953). The Census Bureau forecasts and our extensions of them are discussed more fully in Appendix B.

7

8

I Trend of the Older Population

T A B L E 2.1 P O P U L A T I O N A G E D 65 AND O V E R , U N I T E D S T A T E S , 1930-2000, B Y S E X ( T h o u s a n d s of persons) Year

Men

Total

Women

3>326 4,406

3.3°8 4,614

6,634

I940 95°

5.798

6,472

12,270

7.079 8,090

8,622

Ii,701

!97°

i3.29r

22,66o

1990

9.369 10,468

IS.42S

25.893

10,489

15.644

26,133

'93° J

9,020

18,885

IO>795

SOURCE: 1930-1950: 1950 Census of Population, Vol. II, Part I (Washington: U . S. Government Printing Office, 1953), Table 39, p. 93. 1960-1970: U. S. Bureau of the Census, Current Population Reports, Series P-25, No. 78 (August 21, 1953). 1980-2000: Computed. See Appendix B.

T A B L E 2.2 TREND or A G E SPECIFIC DEATH R A T E S , 1900-19jo, UNITED STATES D E A T H REGISTRATION A R E A Deaths per 1,000 exposed Age group 1900

1940

1950

All ages

17.2

10.8

9.6

Under 1

162.4 19.8

54-9 2.9

5"r4

3-9

i .0

15-24 25-34

5-9 8.2

3 >

35-44

10.2

45-54

15.0

55-64

27.2

22.3

8-5 19.i

65-74

56.4

48.0

40.7

!23-3 260.9

112.0

93-3 202.0

1-4

75-84 O v e r 84

2.0

5-2 10.6

235-7

Average Per cent per cent improvement improvement, 1940-1950 per decade, 1900-1950 II

II

33-0

27

40

1-4 0.6

41

52

31 26

40

!-3 1.8 3-6

26

35 42

!9 II

31 20

7 6

14

5

17

5

14

15

SOURCE: Death rates from U. S. National Office of Vital Statistics, Vital Statistics of the United States, 105O, Vol. I (Washington: U . S. Government Printing Office, 1954), p. 190. Rates of change computed.

the twentieth century. It is a commonplace that progress has been most marked in combating the diseases of infancy and childhood, and the trends recorded in the last two columns of the table accord with this. The main purpose in presenting the table, however, is to call attention to the very substantial rate of progress in dealing with the diseases of maturity and old age. Some of the major medi-

Trend of the Older Population

| 9

cal gains in the decade of the 1940s—improvements in the art of surgery, in the use of antibiotics, in the diagnosis and treatment of tuberculosis—have important applications to the diseases of the upper age brackets, and these developments were not fully disseminated by the end of the decade. The last column of the table shows that progress in the treatment of the diseases of the middle and end of life was accelerating in the middle of the century; there seems to be every reason to expect this trend to continue. Nevertheless, the forecasts presented in this chapter are based upon very conservative assumptions about future declines in the mortality rate. They assume, in general, that mortality will continue to decline until i960 at the rate displayed from 1939 to 1947 and will thereafter remain constant. For the extreme high-age groups—whites 80 years old and over and nonwhites 65 years old and over—it was assumed that mortality rates would remain at the level of the i94o's.a In making these assumptions we follow the precedent of the Census Bureau, although we feel that medical progress in the next half century is likely to be much greater than is here postulated. Even at the conservative rates of mortality assumed, the number of older persons will approximately double in the course of the next fifty years. Figure 2.1 gives a general impression of the course of the anticipated growth. The number of older women will climb by about 2,000,000 per decade until close to the end of the century. The pace of increase will slacken in the 1990's, when children born in the 1930's reach the old-age category, and will pause at a plateau until early in the twenty-first century, when the impact of the increased fertility of the later 1940's will be felt. This, of course, will occur beyond the period covered by our projections. The population of older men is expected to increase only about half as fast as the population of older women—at about 1,000,000 per decade—with the same slackening in the final decade of the century. In the middle-age groups in 1950 the mortality rate for men was more than 50 per cent higher than the mortality rate for women; even in the oldest age group a man was more than 10 per cent more likely to die in the course of a year than a woman (see table 2.3). s For a fully detailed statement of the mortality assumptions see U. S. Bureau of the Census, Current Population Reports, Series P-25, No. 4 3 (August 10, 1950). See also Appendix B, Part II.

O O O CM 0 01 0>

o co 0)

o

J

o (0 0)

o m 0)

o siOí

o to 0>

J

I

I

I

I

I

I

L

Trend of the Older Population |

11

By the end of the century, according to these estimates, there will be some 10,500,000 men and some 15,600,000 women over 65 in the country; about twice as many older people as at present.

The Proportion of the Aged The proportion of older persons increased strikingly during the years from 1930 to 1950. The future of this proportion is necessarily less predictable than that of the number of the aged, because it TABLE 2.3 DEATH R A T E S FOR M E N AND WOMEN, 1 9 4 0 AND 1 9 5 0

(Deaths per thousand) 19+0

1950 Age Men

Women

Men

Women

II. I

8.2

I2.0

9-5

Under i 1-4 5-14 i5~ 2 4 25-34 35-44 45-54 55-64 65-74 75-84 85 and over

37-3 i-5

47-7

0-1 00 «

All ages

6 l .9

3 •1

2.7

0.7

i-3 °-5

1.2

0.9

i-7

0.9

2.2

1-4

2.7

1.8

4-3

2.9

2-3 3-4 5-9

10.7

6.4

12.5

8.6

24.1

14.1

26.2

18.I

4-5

49.0

33°

54.2

41.9

I°4-3

84.0

121.3

103.7

216.4

191.9

246.4

227.6

SOURCE: See table 2.2.

depends on a projection of the entire population, which in turn, depends upon assumptions about the future of fertility. Fertility has proved in recent decades to be an unstable and unpredictable component of population change and, at the present time, appears to be reversing its long-term trend. It is unwise to place much confidence in any prediction of its future behavior, and we have therefore made two projections of the entire population, one based on the highest, the other on the lowest, fertility rate assumed in the Census Bureau projections. Even these two projections cannot be regarded as bracketing the range of possibilities. The higher anticipates merely a continuation of the fertility rates of the postwar period; the lower assumes a rather prompt return to the levels of the early 1940's. If economic and political conditions favorable to

12

| Trend of the Older Population

high fertility persist, the high fecundity rates of recent years may well continue; they may even rise if progress is made in combating the physical and emotional causes of sterility and low fecundity. At present there is no evidence to support the belief that fertility rates are likely to decline. Thus is appears that the projection that assumes higher fertility (see figure 1.1) presents the more plausible outlook in the light of currently available data. The indication, clearly, is for an immediate slackening in the recent pace of increase in the proportion of the population over 65. The current ratio of 8.2 per cent may be expected to rise to a peak of about 9 per cent around 1975 and there to stabilize. The sharp decline after 1980 is a reflection of the abnormally small number of births during the 1930s and should not be projected beyond the years for which calculations have been made. In summary, then, our expectation is that the increase in the proportion of older persons has about run its course, since the substantial increase in their number will be almost entirely offset by the growth of the population as a whole.8

The Age Composition of the Aged Longevity is on the increase. Not only do larger numbers and proportions of persons reach 65 than formerly but they survive for longer periods after reaching that age. The expectation of life for men who have attained 65 has increased from 11.8 years in 1930 to 13.0 years in 1950. For women of 65 the twenty-year increase was from 12.8 years in 1930 to 15.3 years in 1950/ A woman who reached 65 in 1930 could expect, on the average, to live to be almost 78; a woman who reached 65 twenty years later could expect to live to be slightly more than 80. When these trends are projected into the future, (see table 2.4) the percentage of older persons in the 65-69 age bracket declines gradually and substantially for both sexes, and the proportion over 7 5 increases by about 50 per cent. Thus in the future we shall have an increasingly old population of persons 65 and over. 8 This conclusion is affected by the difficulty of forecasting changes in mortality rates. For further discussions see Appendix B, part II. * These figures are for white persons. See Statistical Abstract of the United States, 1954 (Washington: Government Printing Office, 1955), Table 72, p. 73.

Trend of the Older Population

|

13

TABLE

2.4

A G E COMPOSITION OF THE A G E D , U N I T E D S T A T E S , 1 9 3 0 - 2 0 0 0

(Percentages) Age

1930

1940

1950

I960

1970

1980

1990

2000

43 29

42 28

39 28

38

39 28

38

34

28

27

28

28

30

33

34

34

35

39

MEN

65-69

42

70-74

30 28

7 5 AND OVER

WOMEN

65-69 7°-74 7 5 AND OVER SOURCE: See t a b l e 2 . 1 .

41 29

42 28

40

36 27

35 28

34 27

33 27

28

28





32

36

37

38

40

45

26

Chapter III

Some Characteristics of the Aged

From the sociological point of view older persons can be considered to constitute a large, disadvantaged minority, whose characteristics are constantly changing. In part the character of this minority is determined by the aging process which its members have undergone, in part by the characteristics of the general population some decades previously—the child is father to the man. The aged population also exhibits the consequences of differential survival rates within the general population, displays secondary efects of its inherent peculiarities, and, finally, reflects the attitudes of the general population toward its members. Thus one would expect the older population to differ from the general population in many respects other than age—and it does. Among the characteristics that appear relevant to the problems of older persons are: age distribution within the thirty-odd years of older life; the ratio between the sexes; marital status; family status; housing arrangements; places of residence by geographic region and size of community; color; nativity; educational background; occupational background. In our analysis the attention given to these different classifications has been very uneven, for two reasons. First, not all of the data yielded by the Follow-Up Survey permitted breakdowns sufficiently fine for meaningful analysis. Second, our study indicated only a minor relationship between some classifications and the two measures of economic status we have adopted, namely, income and labor-force participation. 14

Some Characteristics of the Aged

| 15

The classifications which we could analyze and which appeared to be fruitful for explanatory purposes were: sex, marital status, family status, residence, and occupational background. In the present chapter the distribution of older persons by these characterTABLE

3.1

NUMBER OF PERSONS AGED 65 AND OVER, 1 8 8 0 - 1 9 7 5 (Millions of persons) Year

Total

Male

Female

Males per 100 females

CENSUS DATA 1880 1890 1900 1910 1920

IOI

1-7 2.4

•9 1.2

•9 1.2

3 •1 4.0

1.6

1-5 2.0

102

2.0 2-5

2-4

IOI

3-3

3-3 4.6

IOO

6.5

90

7-5 8.6

86 78 73

'93°

4-9 6.6

1940

9.0

195°

12.3

4-4 5.8

104 IOI

96

PROJECTIONS

1955 I960

14.0

1965

1

7 1 7.6

97°

7-3 18.9

8.1

9-7 10.8

1975

20.7

8-7

12.0

I

IS-7

6-5

82 7Î

SOURCE: IQSO Census, Vol. II, Part I, p. 93. Current Population Reports, P-25, No. 87, tables I, 2, and 3.

istics, as revealed by the Follow-Up Survey, is discussed, as are the distributions by the other characteristics listed above, employing, where necessary, data from other sources.1 Sex and Marital Status Older women outnumber older men, because they live longer. (See table 2.3.) The rate at which the numbers of older men and women are moving apart is shown in detail in table 3.1 and figure 3.x. At the beginning of the century the expectation of life at birth was 1 T w o especially useful sources of information on the characteristics of older persons are U . S. Bureau of the Census, XJ. S. Census of the Population: 1950, Vol. II, Part 1 (Washington: Government Printing Office, 1953), and the various series of the Current Population Reports (Washington: U . S. Bureau of t h e Census). F o r brevity these will b e cited as 1950 Census, Vol. II, Part 1 and C P R Series, respectively.

MILLIONS

r

/ / / / / / / / / / /

18

16

14

/ / /

12

10

8

6

4

2

1880

J 1900

I 1920

I 1940

I I960

L 1980

Fig. 3.1. Growth of the aged population of the United States, 1880-1950, with projections to 1975. (Source: Current Population Reports P-25, No. 78, tables 1 , 2 , 3.)

Some Characteristics of the Aged

| 17

48 years for white males and 51 years for white females: in 1950 the expectation was 66 years for males and 71 for females.* Our projections for the next half century anticipate a striking increase in the disproportion between the numbers of older men and older women in the course of the next fifty years. The marital status of the older population reflects the difference in longevity, the hazards of the aging process, and the tendency of men to marry women younger than they. The marital distribuTABLE 3.2 M A R I T A L STATUS OF PERSONS A G E D 65 AND O V E R , BY S E X , A P R I L , 1 9 5 2

(Thousands of persons) Total

Male

Female

Status Number

All Single Married Spouse present Spouse absent Widowed Divorced

12,276

P e r cent

Number

99-9

S.7J6

P e r cent

IOO.0

Number

6,520

Per cent

IOO.O

988

8.0

476

512

Si.6

8.3

7-9

6,34°

3.938

68.4

2,402

36.8

6,104

49-7 !-9 39-3 i .0

3,802

66.1

2,302

136 1,260

2-4 21 . 9

35-3 1-5 54-7

82

I.4

236 4,826 122

100 3.566

40

0.6

SOURCE: Current Population Reports, P - 2 0 , N o . 44, t a b l e 1. NOTE: Some estimates of population totals taken from the Current Population Reports differ slightly from estimates derived from the Follow-up S u r v e y of Older Persons. These discrepancies result from the w e i g h t i n g procedure used b y the Census B u r e a u , which prepared both estimates, and h a v e not been suppressed. See Appendix A .

tion of older persons in April, 1952, is shown in table 3.2. Only 35 per cent of older women were living with their husbands. The comparable figure for women from 20 to 65 at the same date was 74 per cent. Of the older men, 66 per cent were living with their wives on the survey date, as compared with 78 per cent for men aged 20 to 65. This general situation is of long standing, as table 3.3 indicates. In the past sixty years the percentage of older men who were married has fallen slightly while the comparable ratio for women has varied only negligibly. In all likelihood, for the immediate future we can expect about two-thirds of older men and about one-third of older women to be married. 2 Expectations of life are taken from U. S. Bureau of the Census, Historical Statistics of the United States, 1789-1345 (Washington: Government Printing Office, 1 949)> P- 45, and U. S. Bureau of the Census, Statistical Abstract of the United States: 1954 (Washington: Government Printing Office, 1954), p. 72.

i8

| Some Characteristics of the Aged

In studying the economic condition of older persons the essential distinction with respect to marital status is simply the presence or absence of a spouse in the household. Estimates of the number of older persons in April, 1952, classified in accordance with this TABLE 3.3 M A R I T A L STATUS OF PERSONS A G E D 65 AND O V E R ,

1890-1950

(Thousands of persons) Male

Female

Year Total

Married

P e r cent married

Total

P e r cent married

Married

1890

1,234

870

70.s

1,184

418

35-4

I9OO

1,044

67. i

34-2

i,3°4

65.6

1,525 1,964

521

I9IO

1,555 1,986

687

35-°

I920

2,483

1,607

64.7

2,45°

830

33-9

I93O

2,117

6 3 7

3,308

2,8lO

63.8

4,613

1,147 1,584

34-7

I94O

3,326 4,406

I95O

5,733

3,767

65.7

6,522

2,328

35-7

SOURCE: IQSO Census,

34-3

Vol. I I , P a r t I, pp. 179-180.

TABLE 3.4 S E X AND M A R I T A L STATUS OF INDIVIDUALS 6 5 AND O V E R , A P R I L ,

1952

(Thousands of persons) M a r i t a l status

Male (46.1 per cent)

Female (53.9 per cent)

Total

Spouse absent or no spouse Spouse present Spouse 65 or over Spouse under 65

I,8LO

4,230

6,040

3,763

2,289

6,052

2,136

2,136

4,272

1,627

153

1,780

Total

5,573

6,519

12,092

distinction, are given in table 3.4. The category "spouse absent" groups all persons who have no spouse living in the same household—e.g., for women: widows, spinsters, divorcees, and women otherwise separated from their husbands.8 For many purposes, however, the proper focus of attention is not on individuals. For example, the 2,136,000 couples, each of whose members is 65 or over and who appear respectively in the * W e shall subsequently use the simpler terms "unrelated males" and "unrelated females" to denote those with spouse absent.

Some Characteristics of the Aged

| 19

"male, spouse present," "female, spouse present," categories of table 3.4 represent only about two million cases of interest rather than twice that number. It is quite probable that some of the male members of these couples report incomes of $3,000 or more, whereas their wives report no income; it would obviously be fallacious to consider the male members adequately financed and their wives destitute. For all practical purposes, a couple is a unit of analysis as well as a unit in fact; income and resources as well as problems are shared. In this analysis, therefore, the notion of the aged person is replaced by the notion of the aged economic unit, and three kinds of economic units are defined: (1) the unrelated female, (2) the unrelated male, and (3) the couple. An aged couple is defined somewhat arbitrarily as a couple of which the male member is 65 years old or older. This definition excludes 153,000 females, 65 and over, whose husbands are less than 65 years old. It also includes by inference the 1,627,000 females who are not yet 65 but who are married to men 65 and over. Although arbitrary, this definition seems appropriate in terms of the structure of our society. It is more than mere convention which calls the male "head of the family." It is generally the male member of the couple who is the economic provider, and thus, so far as age affects the capacity of the couple to meet its economic problems, it is the age of the male member that is decisive.1 The concept of an aged economic unit is used in this study as more than a mere device for presenting results. When reporting the income, savings, or assets of an aged couple, we shall (unless otherwise noted) refer to the combined income, savings, or assets of the couple. When dealing with unrelated males or unrelated females, we report only the individuals' income, savings, or assets. Similarly, of course, when considering the needs of a couple we are concerned with the needs of two persons, not of one. (It should be noted that this aged economic unit is not necessarily a "family" or "household." An aged economic unit, while frequently also constituting a family, may be only a fraction of a family group.) A very large part of the problem of the aged is the problem of 4 There are, of course, exceptions to this, not only with respect to individuals, but with respect to economic problems. For example, so far as age affects the medical expenses of a couple, the older the wife, the higher these expenses are likely to be.

20

| Some Characteristics of the Aged

single aged women. Unrelated females constitute the largest single class of aged economic units—about 43 per cent of all such units —and outnumber unrelated males by more than two to one (see table 3.5)." The preliminary indication of economic position as given by the median income suggests that unrelated females not only constitute the largest group, but are faced by the most serious economic problems. T A B L E

3.5

M E D I A N I N C O M E OF A O E D E C O N O M I C U N I T S B Y S E X A N D M A R I T A L S T A T U S ,

T y p e of unit

Aged couples Single men (unrelated males) Single women (unrelated females) Total

Number (thousands)

Per cent

1952

Median income*

3.763 1,810

3»-4 I8.J

?I,387 662

4.23°

43- 1

273

9, 8 °3

100.0

» See Appendix A for discussion of procedure for computing median. Incomes refer to calendar year 1951

The differences indicated by table 3.5 are great enough to make it necessary to maintain the distinction among types of units for all analyses. Living Arrangement (Family Status) The standard Census Bureau classification of familial relationships is not well suited to the analysis of the position of older persons, whose precise roles in the family are frequently ambiguous. The usual arrangement is to classify older men and women according to whether or not they are the heads, or wives of the heads, of the families to which they belong. Such a classification can be misleading in an economic analysis, as, particularly with the group 65 and over, the designation "head of a family" is often an honorary title rather than an indication of economic position within the family.' The classification of living arrangements in table 3.6 makes it possible to determine whether it is likely that an aged person or couple is receiving nonmonetary support. The classification "not 5 The number of aged economic units is smaller than the number of aged persons because of the inclusion of two aged persons in most aged couples. 9 See p. 21 n.7, and p. 110.

Some Characteristics of the Aged

| 21

living with relatives" includes aged persons living alone, living with strangers, or (infrequently) living as a unit in a household containing their minor children. The next two categories are units in which the aged person (or couple) is designated the head of the family: in the first, adult children are present, in the second, they are not. The distinction is made because aged persons living with children are less likely to be the real economic heads of their families than aged persons so designated who are living with other relatives (brothers, sisters, and so on). The fourth and fifth categories cover persons who are not designated as family heads: the former, those living in the households of adult children, the latter, with other relatives. These five categories constitute a classification sufficiently flexible to permit combination into several summary categories, two of which are presented in the table. Among the significant features of this table are the figures showing the large absolute percentage of each group who are living alone. About five million of the aged economic units—more than half—were, in April, 1952, living alone or with nonrelatives (most of these were living alone). The percentage varied from 40 per cent for unrelated females to 70 per cent for couples: in view of the low median incomes of these units, these percentages seem surprisingly high (see chapter vi). Another large group lived with their children, a much smaller group with other relatives. The data on median incomes are also significant. Of the persons designated family heads, those not living with children (the first two categories in the table) have incomes that are, typically, significantly greater than those who are living with children. Of persons living with children, those designated family heads have significantly greater incomes than those reporting themselves as parent of the family head.1 7

This observation suggests that, at least in part, the census distinction between "family head" and "not family head" is significant. However, the significant difference in the incomes between the second and third groups in the table—respectively, "head, no children under 2 1 present" and "head, children 2 1 and over present"— suggests that in part the distinction is honorific rather than economic. Although the designation of "head" or "non-head" may be significant, it is also misleading. This is further substantiated by the fact that a significant percentage of those in the category "head, children present" were totally without monetary receipts. The percentage ranged from almost 10 per cent for couples to 3 3 per cent for unrelated females in this category. W e believe these considerations justify our use of a concept of family status that does not rely on the designation of "family head." Further evidence on this point is presented in chapter viii.

Si

^ r— »-• oo tj- ci ON TÍ- Í-N.OO OO oo ^û ci VO vr» OV«0 VO O CO ^ t*^ f^-'O Q co c ^J CO O co r -^vc Ti | f^ci i- o CO c tf^r-»vo o

.2 g

o s S.s

r -». Th ^o vr cih- coot-« . «o V-Ti^O C* H « oo n^i-viO

s




•o â Ü > 593 $ 561

$ S3»

$2,082 5

923

• Based upon ranking of five principal sources of receipts.

ference between the receipts distributions of men and women is a result of differences in the frequency and size of earnings. That earnings alone do not account for all of the difference is indicated by the comparison of unrelated males without earnings with similarly situated unrelated females. This characterization of the typically better economic position of the unrelated males should not obscure the fact that substantial

Sources of Receipts

|

105

groups among the males were no better off than their female counterparts. In addition to the 19 per cent without a major source of receipts more than 25 per cent received public assistance." Although, among those receiving assistance, the proportion relying on such aid as a sole major source was smaller than among unrelated females, it was nonetheless substantial (63 per cent).7 The T A B L E 8.9 UNRELATED M A L E S W I T H AND WITHOUT E A R N I N G S , 1 9 5 1

Without earnings

Category

With earnings Total Total

Percentage of all unrelated males. .

65.7

34-3

Median total income Mean total income

$448 $549

$1,440

$2,073

$I-Î499

$500 or more

9-7 $431 $Î3°

24.6

IOO.O

$2,029

$662

$2,680

$1,074

NOTE: This table treates incomes, not receipts. TABLE

8.10

COUPLES W I T H AND WITHOUT EARNINGS, 1 9 5 1

Category

Percentage of all couples Median total income Mean total income

Without earnings

With earnings Total Total

ÎI-Î499

Î500 or more 44.2

IOO.O

43-3

56.7

I2.5

$885

$2,162

$580

?2,695

$1,387

$1,142

$2,839

$737

$3,433

$2,101

NOTE: This table treats incomes, not receipts.

distribution of receipts from assistance as well as the total receipts associated with receipt of assistance are closely similar for unrelated males and unrelated females.8 In other words, males who received assistance were no better off than females who did. But the better position of the rest of the males is indicated by the fact that whereas 11 per cent of the aggregate receipts of unrelated males come from assistance, 19 per cent of the aggregate unrelated female receipts came from that source.9 " See 7 See 8 See 9 See

table table table table

8.8. 8.6. 8.4 andfigure8 . 1 . 8.3.

106

| Sources of Receipts

Couples.—As is clearly seen in table 8.8 and in a comparison of table 8.10 with tables 8.7 and 8.9, the superior economic status of couples dominates data concerning sources of receipts. Appreciably fewer couples had no major source of receipts, although the percentage that did is not negligible; correspondingly more had two or more major sources of receipts; and the average number of major sources was greater. As with unrelated males, earnings was the most frequent source of receipts as well as the source with largest typical size. But, in contrast with unrelated males, the frequency with which earnings were received was very large indeed relative to other sources (table 8.2). It was largely this predominant importance of earnings—a reflection of a greater degree of labor-force participation—that accounted for the greater receipts (relative to budget levels) of couples.10 The percentage of couples receiving assistance, although smaller than for either of the other groups, was large (17 per cent), and the recipients of assistance were in a position that was no better (and was very likely poorer)" than that of their "unrelated" counterparts. Summary.—A large part of the difference in the economic position of the three kinds of economic units has been described in terms of the numbers of sources of receipts and in terms of the size of the groups receiving a significant amount of earnings. However, the size of the group receiving assistance was appreciable for each kind of unit, and those receiving assistance had low receipts. Among unrelated females, those who received assistance did not stand out appreciably from other low income groups, but among males and couples they did. Since sources of receipts seem to play particular and consistent roles in the receipts of aged individuals 10 A couple, of course, represents two persons, not one. As a result, a direct comparison is complicated. Not only are the expenses of two persons relevant, but couples frequently get either two entitlements or a larger amount from some sources, chiefly pensions and assistance. Thus, the greater frequency of multiple sources and the higher receipts from each source are not necessarily indicative of superior economic position. But with respect to per capita earnings, couples are appreciably better off than unrelated males and unrelated females combined. In view of the relatively small earning power of aged females, one would not expect the distribution of earnings to show either a much greater frequency or much larger typical size for couples than for unrelated males. It shows both—owing of course to the differential rates of labor-force participation (see table 8.8). " Because couples consist of two persons. Compare the average receipt levels with the budgets previously developed.

Sources of Receipts

|

107

and couples it should be fruitful to examine separately each of the five major sources of receipts. This is done later in this chapter (see pp. 111-120), after a further examination of the significance of the differences among units. Are These Differences Significant? To what extent are the differences in the pattern of sources of receipts truly a function of differences in marital or sex status and to what extent a function of differences of such subsumed variables as socioeconomic status and living arrangement? A limited examination of the data on sources of receipts, alternatively crossclassified by the categories of socioeconomic status and living arrangement defined in chapter iii, provides an answer to this question. Attention is limited to selected principal major sources of receipts—earnings, pensions, assistance—and to "no major »12 source. Socioeconomic status.—Sources of receipts differ substantially with socioeconomic status (see table 8.11). The pattern conforms closely to a priori expectations. The professional and proprietary groups seldom were without a source of receipts, received assistance relatively infrequently, but had a high incidence of earnings; a large percentage of each of the farm groups had "no major source" and the frequency with which they received assistance was high; farmers and domestic service workers had the lowest incidence of pensions; pensions were more important for the factory classes; and so on. It is particularly evident that the eligibility provisions for social security (as they applied in 1951) played a large part in the resulting pattern. Although the classification of sources of receipts by socioeconomic status appears significant (it would have been extraordinary had it not!) it also appears that important net differences among type of unit remain. For the two categories in table 8.11 that contain all three types of units—"Professional and Proprie12 Asset income and dissaving are lumped with contributions, miscellaneous cash receipts, and other sources to form a residual category. This lumping is made appropriate by the relative infrequency with which these sources are the principal major source. Such small frequencies when cross-classified become so small as to be subject to large sampling variability, and thus contribute nothing to the analysis. It should be explicitly noted that for dimensions other than the one now considered, these excluded sources are individually significant, and will be discussed in the succeeding sections of this chapter.

I Sources of Receipts

io8 T A B L E

8.11

P R I N C I P A L M A J O R S O U R C E O F R E C E I P T S FOR U N I T S C L A S S I F I E D A C C O R D I N G

TO

SOCIOECONOMIC STATUS

(Percentages) Socioeconomic status Principal m a j o r source

Couples N o source. Earnings. Pensions. . Assistance Other.... Total

Total

No relevant work experience

Professional and proprietary

White collar

Skilled

Semiskilled

Unskilled

Farm

9-1

6.2

4-5

10.8

6.1

13-6

41.1

51.2

47-7

34

40.7

45.6

40.7

22.6

19.4

29.6

37

32-3

29.8

6.9

12.0

5-4

12.3

22.4

14.2

17.8

6.2

16.4

8 r

4.6

3

il.6

100.0

100.0

Skilled and semiskilled

Unrelated males N o source. . . Earnings.... Pensions. . . . Assistance... Other Total

18.5

12.6

15-9

2

3-5

21.9

22.8

31.6

22.5

2

3-i

!7-4

25.8

27.4

3

9-3 13.6

8.4

'3-9

33-8 13.2

H-3

6.0

16.2

100.0

100.0

100.0

r

20.0

100.0 White collar

Unrelated females N o source Earnings Pensions Assistance Other Total

32.1

40.8

7.8

I. I

6

'4

Other nonfarm

12.6

Service

18.2

!î-9

7-3

17-4

T

7-3

3-î il.8

2

11.0

33-5

2

29.0 11.8

14.6

Ï 3 Î

22.7

25-4

28.7

8.8

3-9

23-4

20.1

10.6

20.6

23.O

39-5

22.1

24-5

21.2

20.7

18.4

35-9 20.0

IOO.O

100.0

100.0

100.0

15.8

tary" and "Farm"—the previously identified pattern of differences among unrelated females, unrelated males, and couples was clearly present. As we move from the unrelated females to the couples, "no source" and assistance decrease in importance, earnings increase in importance.

Sources of Receipts

|

109

Living Arrangement.—Differences in the pattern of sources of receipts are likewise appreciable when the aged are classified by living arrangement (see table 8.12). Yet, important net differences by type of unit remain after such classification. The significance of the differences according to family status are thus not destructive of the generalizations of the last section, but they are of interest with respect to other problems of the aged. An important difference between the classification by living arrangement and that by socioeconomic status is that the association between living arrangement and sources almost surely reflects a line of causation from sources to living arrangement, whereas the association between socioeconomic status and sources was probably characterized by a causal chain running from socioeconomic status to sources of receipts. In any event, an association is evident.13 Those of the aged living in a family as parent or other relative of the head were very frequently without a major source of receipts and a low percentage had earnings, in sharp contrast to the group not living with relatives. This is not surprising but very likely accounts for much of the difference in living arrangements. An interesting comparison is that between those "not living with relatives," and those classified as "head, no adult children present" (living with relatives other than children); in each group the aged person (or spouse) was considered family head. Members of the first group were infrequently without a source of receipts but many received assistance. Couples in the second group had earnings almost as frequently, unrelated males and females had them much more frequently; all had assistance as the principal major source much less frequently. Both groups were apparently in a better economic position than those not living with relatives. It seems likely that these units were suprasupporting, that is, it seems likely that they provided some monetary support for the relatives who lived with them." 13 In the subsequent discussion w e shall use the words "frequently" and "infrequently" as relative to the totals in the first column of table 8.12. 14 Of course all of these groups are to some extent nonhomogeneous. The one important evidence of that nonhomogeneity in the generalization suggested above concerns (for couples only) the larger percentage having no source of receipts among the group that otherwise appears more favorably located. For unrelated females and unrelated males, the group "Head, no adult children present" seems in every way more favorably situated than the group not living with relatives.

no

| Sources of Receipts

The fact that a relatively high percentage of the group not living with relatives received assistance reflected, of course, the greater monetary need of these persons and the greatly lessened opportunities for receipt of nonmonetary support. It also very likely reflected implicitly the application of the needs test and the principle of family responsibility by public and private relief agencies. These data provide additional insight into the significance of the designation "family head" where aged persons were living with children. It was suggested in chapter iii that for the aged the designation family head includes both persons for whom the title was honorific (that is, persons who should perhaps be classified as parent of head) and persons for whom the designation was economic (and who therefore need not be distinguished from those in the category "head, no adult children present"). The extent of this heterogeneity can be judged by comparing the group "head, adult children present" with each of the other groups living with relatives (see table 8.12). For couples this group did not exhibit a pattern of sources sharply different from that of other family heads; unrelated males who were family heads and living with children had a pattern significantly different from the other groups—the frequency of "no source" and of "earnings" were both high relative to the totals. The pattern of sources for unrelated females was also significantly different, but was appreciably closer to the group who were not family heads than to the group of family heads not living with adult children. It thus seems likely that the Census designation of family by age of head without regard to the composition of the family introduces an important possibility of error if the family head is an unrelated male or female 65 and over, but only a small error if the family head is 65 and over, male and married. Conclusion.—There is a significant relationship between each of the tested variables and sources of receipts, but this relationship does not destroy the important net relationship between sources of receipts and marital-sex status.15 15 A third variable, place of residence, was also tested, but the chief significant difference there observed was that between farm residence and other types of residence, and this is so similar to the distinction (in terms of occupational background) between farm and the other classes of socioeconomic status that it has no independent interest.

Sources of Receipts

|

111 TABLE

8.12

P R I N C I P A L M A J O R S O U R C E o r R E C E I P T S FOR U N I T S C L A S S I F I E D ACCORDING TO F A M I L Y S T A T U S (Percentages) Living arrangement Category

Total

Not living with relatives

Head, no adult children present

Head, adult children present

Parent or other relative of head

Couples N o source

6.9

8

14 8

IO.3

39 4 21 I

4O.6

I7.9

!4-7

23 6 IO i IO 5

33-4

5 3 18 4

100.0

100.0

IOO 0

IOO 0

100.0

N o source

18 6

25 9

29.7

22 8

"•3 24.0

8 8

Earnings

44 I

Pensions

25 8

28.2

8 8

31 5 20 4

27.6

Assistance

19 3

21.0

8 8

14 8

20.6

Other

!3

5

15-5

29 5

7 4

9-7

IOO 0

100.0

IOO 0

100 0

100.0

32 I

16.2 10.8

33-7 8 0

Pensions

14 6

15-4

!3 9 20 9 l8 6

47.6

8

I

12.5

Assistance

23-4 22 I

31.6

l6 3

Other

26.0

3°-3

17 I 24 I

Total

IOO 0

100.0

100 0

IOO 0

44.0

Pensions

9 I 42 I 22 6

Assistance

12 0

12.3

Other

14.2

Total

Earnings

22.1

Ii

17.9 20.5

Unrelated males

Total

12.4

Unrelated females N o source Earnings

7

17

3-4

r9-3

17.2 100.0

The Role of Particular Sources Earnings.—If one were to advise a person as to the most probable road to a high level of receipts in old age, that advice would be: "Have earnings." For each type of unit, those with earnings as principal source had the highest average level of receipts. The differences in the frequency with which earnings were received by unrelated females, unrelated males, and couples accounted for much of the difference in economic position. The importance of earnings in explaining high receipts is indicated in other ways as

tíCo.* tjUMo O8 Ssl'-s-j c"« £•£ a]o * J -ti.H

vo d co es e) c© O os^

v i coso so t » \oSO oo

O SO

d oo oo tHO

vûoo H dOs COSO vo »a

oo C

o o S .S £ . 0 g O-S.sjs

4) ü 4J rt bO U (J 3

SÊji S » E 5 ;e O ^ u

sE a«i s* c

vr» «00 O 00

+

•o». ".5

OO N «

U u S O o ¡5

z o O W •J C uOJ —G u2 e* °e r-» o w M ro es

-Su-, E o eO.; o

O TJ CT\ Th

138

| Assets of the Aged

The most surprising feature was the high frequency of insurance in large rather than in moderate amounts for the group without income. The only other group for which this was true was the group with the highest levels of income, but there it is not surprising. These features suggest that the group without income was peculiarly nonhomogeneous with respect to previous incomes. Some, clearly, must have had high incomes when they acquired sizable insurance. They were, in 1952, probably living on accumulated savings or with their families. 8 T o summarize, it appears that so far as insurance alone was concerned, older women whose husbands were still living were only partly equipped with a significant cushion against the economic shock of a husband's death. A sizable group had only enough insurance to cover the direct and immediate expenses concerned with death; a sizable group did not even have that. Only a small group had insurance sufficient to provide even moderate support to a widow over a period of years. Liquid

Assets

Liquid assets potentially substitute for or supplement life insurance in meeting the financial burdens of widowhood. The existing amounts of liquid assets as previously tabulated (table 9.5) were obviously required for more purposes than merely providing a cushion against the potential shock of a husband's death. Indeed, these liquid assets were being dissipated at a rate of about a billion dollars per year to meet living expenses (table 8.3, p. 97). Supposing (incorrectly) that these assets were in such form that they would be wholly available at the time of death, the picture with respect to liquid assets would be quite similar to that for life insurance (see table 9.9). About 2,2 per cent had liquid assets of $5,000 or more, an amount that is presumed sufficient to provide several years of support for the widow as well as to meet costs of death and relocation. Another 16 per cent had liquid assets of less than $5,000, adequate for one to a few years' support. Thus, approximately 38 per cent had enough liquid assets to avoid immediate economic hardship occasioned b y the death of a spouse. A t 8 T h e data available on the use of savings are consistent with this hypothesis. Roughly one-sixth of the group without income have appreciable insurance. The top one-sixth of the dissavers with zero incomes dissaved in amounts of more than $1,000 per year. (See Appendix E , table 210.)

NONE

$1-2499 $2500-4999 INCOME C L A S S

$ 5 0 0 0 AND OVER

I NO LIFE INSURANCE S S S S j $ 1200-4999 • 199 $ 5 0 0 0 OR MORE W f f l h $ 2 0 0 - 1199 Fig. g.x. Face value of life insurance, by income, couples, 1951.

140

| Assets of the Aged

the other extreme, about 42 per cent had no liquid assets; in such cases the wife must be presumed likely to become immediately dependent if her husband dies and she has no other resources. The remaining 20 per cent had liquid assets of less than $1,000; amounts that might be characterized as possibly sufficient to adjust to economic plight but not to mitigate it. TABLE

9.9

COMPARISON OF L I F E I N S U R A N C E AND L I Q U I D A S S E T HOLDINGS OF C O U P L E S Life insurance*

Class interval

Zero

Liquid assets b Per cent distribution

Class interval

43-8

$1-199

3-5

Per cent distribution

Zero

42

$1-499

12 8

200-1,199...

28.0

500-999....

1,200-4,999.

H-7 10. i

1,000-4,999.

16

5,000 a n d u p

22

5,000 a n d u p

Total

Total

100. i

Face value of life insurance of male member of couple: from table 9.6. Adjusted value of liquid asset data. From Beneficiary Survey; see table 9.5 and Appendix D for adusted data and basis of adjustment. a

b

TABLE

9.10

C O M B I N E D HOLDING OF L I F E I N S U R A N C E A N D L I Q U I D A S S E T S OF C O U P L E S

(Percentages) Combined holdings computed assuming: Size of combined holding

Zero correlation (2)

Perfect positive correlation (3)

!9 II

42

0 32

II

6

1,000-4,999

36

26

22

5,000 a n d u p

32

33

24

Perfect negative correlation (I) O

$1-499 500-999

6

SOURCE: From table 9.9, as described in text.

Thus, considering liquid assets and insurance separately, only a minority of the female members of couples would be adequately provided for if the husband died. It is, of course, inappropriate to consider these two cushions independently, as they are potentially alternative means to the goal of security. Data to measure decisively the combined holdings of liquid assets and insurance are not available, but certain speculations are possible and it is also possible to define limits to the probable values of the combined

Assets of the Aged

|

141

holdings under three different assumptions (see table 9.10). It is assumed successively that there was (1) perfect negative correlation, (2) zero correlation, and (3) perfect positive correlation, between the holders of assets in varying amounts and the holders of insurance. That is, when assuming perfect positive correlation we assume those with the largest liquid assets had the largest amounts of insurance, and so on.9 It seems likely, for several reasons, that the second and third of these assumptions provide reasonable limits to the probable actual distribution: (1) The size of holdings of both life insurance and assets is positively correlated (though not perfectly) with the size of present income; both would doubtless be more closely positively correlated with past income. Either of these factors would lead toward positive correlation between holdings of insurance and of liquid assets. (2) Although, for individuals, substantial amounts of life insurance and liquid assets are alternative sources of security, it is unlikely that very many of those with large amounts of insurance would be without at least small or moderate liquid assets, particularly because liquid assets are a multipurpose source of protection and life insurance chiefly a single-purpose source. (3) Finally, the tabulation of total assets in table 9.1, which included liquid assets, life insurance, and all nonliquid assets, indicates that 18 per cent of the couples had total assets less than $500. This is an absolute minimum percentage for the combined holdings of liquid assets and life insurance in amounts of less than $500, and leads to the rejection of the data based on the assumption of perfect negative correlation. If zero correlation and perfect positive correlation are accepted as effective limits to the distribution of combined holdings of life insurance and liquid assets, only half (46 per cent to 59 per cent) of the couples had $1,000 or more of such resources. Combined liquid assets and life insurance of less than $1,000 seem likely to result in financial difficulty for widows for two reasons: (1) In most cases death of the husband will lead to a sharp reduction in income. Consider a couple, without earnings, both of whose members are 65 or over, qualified for benefits under OASI. They are receiving one and one-half times the primary in9 W e further assume that those within a class interval are evenly spaced and that the same assumption with respect to correlated holdings holds within class intervals as among them.

142

| Assets of the Aged

surance amount. With the death of the husband the widow is entitled to three-quarters of the primary insurance amount, or only one-half the amount received by the couple. Even for the aged it is unlikely that one can live half as cheaply as two. If, however, the husband had had earnings, as well as having been covered under OASI (as was generally true), these earnings would be replaced by a "survivors" benefit again amounting to three-quarters of the primary insurance benefit but representing a smaller fraction of previous income.10 Where the husband had earnings, but lacked coverage under Social Security or other pension plans, the reduction in family income would be even more drastic. (2) Since dissaving occurs among the aged, the amount of resources available to a couple at any given date is almost certainly greater than the amount that will be available when the husband dies, as, by the time of death, all or part of the liquid assets may have been expended to meet other contingencies or living expenses. It seems clear that most of the economic advantages couples now have will not be transferred to the widow if the male should die. If, as seems likely, the aged female has greater access to the labor force in the future, her vulnerability to the economic shock of a husband's death will be reduced: barring this, and changes in the structure of survivors provisions in pension plans, in holdings of life insurance or other assets, or in the conventional differential ages of husband and wife, the economic position of the aged female will continue to be made highly vulnerable by the death of her spouse. Sources of Vulnerability: Illness Poor health has been shown to play an important, indeed critical, role in the economic life of the aged by its effects on the ability to work. Secondary to this is the much narrower problem of medical expenses and their impact upon the resources of the aged. u A clear majority of those interviewed in the Follow-Up Survey (from 55 per cent to 70 per cent) received hospital or medical serv10 If the widow is under 65 no continuing benefit is received upon death of the spouse (unless minor children exist) until the widow becomes 65, at which time the full widow's benefit is received. In either case the lump-sum death benefit of three times the primary insurance amount (and thus under the 1950 law a maximum of $240) is payable to the widow. u W e here use the brief designation "medical expenses" to include all the expenses associated with ill health, including hospital expenses, drugs, nursing, as well as strictly medical expenses.

Assets of the Aged

|

143

ices during 1951 (see table 9.11). For more than 90 per cent of these persons, medical services required some cash outlay by either the aged person or his family. (Failure to receive medical service need not, of course, indicate good health.) TABLE

9.11

M E D I C A L 0 SERVICE AND M E D I C A L PAYMENTS' 3 OF AGED U N I T S , 1951 (Percentages) Couples c

Unrelated males

Unrelated females

30.8

44.8

36-9

69.2

íí-2 89.9

63.1

94.6 5-4

10. i

7-9

N o medical services R e c e i v e d medical services M a d e medical payments' 1 N o medical payments' 1

92. i

* " M e d i c a l " includes hospital, nursing, etc. b Payments by aged person or family. 0 Either member of couple. d Percentage of those receiving medical services. TABLE

9.12

A N N U A L M E D I C A L PAYMENTS OF A G E D U N I T S ,

1951

(Percentages) Size of payment

Couples

Unrelated males

5-5 30.6

IO. I

8.1

42.8

4°-3

io-149 150-299

340

26.9

15.2

6.7

3 1 -7 10.8

300 or more

H-7

J3-4

9.2

None ii-49

Total M e d i a n size

:

100.0 $91

99 9 $47

Unrelated females

100.1

?57

Typically, medical payments were small (the median annual payment was about $50 per person), but sizable minorities had moderately large medical expenses (see table 9.12). These minorities are not confined to those with the ability to pay, as judged by income; there is no sharp or consistent relationship between size of medical payments and size of incomes for those having medical services.12 12

A d e t a i l e d c r o s s - t a b u l a t i o n of t h e s i z e of m e d i c a l p a y m e n t s a n d size of i n c o m e

is p r e s e n t e d in A p p e n d i x E , t a b l e 2 0 6 . I t is b e c a u s e of t h e a b s e n c e of a n y s i g n i f i c a n t association b e t w e e n t h e s e v a r i a b l e s t h a t t h e t a b u l a t i o n s a r e n o t s u m m a r i z e d in t h e text.

144

| Assets of the Aged

For many the size of payments was large relative to income." These payments did not reflect the total social cost of medical and hospital services: many other forms of payments were involved, including medical and hospital insurance, free medical service, and private and group hospital schemes. Table 9.13 indicates the percentage of the aged receiving medical services from various sources. It should be noted, however, that even if these sources are treated as mutually exclusive, (which they are not) and assuming TABLE 9.13 NONCASH MEDICAL SERVICES" RECEIVED BY AGED UNITS,

1951

(Percentages receiving) b T y p e of service

Hospital or medical insurance"1 Free services® Other'

Couples 0

Unrelated males

Unrelated females

12.9

6.4

6.8

5-6

7-7

6.1

4-S

8.9

9.2

» Services not paid for b y aged persons or relatives, including medical, hospital, nursing, and drugs. Percentages are not exclusive. Person is included if any part of medical expenses are met in the indicated

b

ways.

0 Either member of couple. d Includes all prepayment or insurance plans (group or individual, nonprofit or commercial) whether or not t h e a g e d person pays the cost of the protection, i.e., includes policies paid for by former employer, union or relatives. e Services provided free of charge b y physician, hospital, sanatorium, clinic, etc. ' Medical expenses paid b y nonrelatives, b y a charitable organization, or b y a union or employer (except through a prepaid plan). Does not include amounts later reimbursed.

further that when such sources were used they involved no concurrent cash medical expenses (which is unlikely), less than onequarter of the aged received medical services for which they did not regard themselves as financially liable. In general, those who have liquid assets seem to be able to meet their medical payments from them. Medical expenses will, however, tend to recur and to increase and liquid assets will tend to decline over the years, but perhaps surprisingly, few of the aged have gone into debt to meet medical expenses. Less than 5 per cent of the sample reported any overdue medical bills or debts incurred in order to pay for medical services. Of the small number with such 18 Restricting attention to those persons or units with (1) no income but medical payments of $50 or more; (2) income under $500 and medical payments of over $300, we find that 10 per cent of the couples, 16 per cent of the males, and 27 per cent of the females with medical services were recorded. (See Appendix E, table 206.) For these people medical expenses were obviously significant relative to income. We do not have a detailed breakdown of medical expenses over $300, but it seems probable that in many cases medical expenses were very large and in such cases also constituted a significant burden.

Assets of the Aged

|

145

debts, less than one-quarter (about 1 per cent of the total aged population) had such overdue bills or accumulated debts of $200 or more (see table 9.14). Conclusion The generally poor economic position of the aged, evidenced by our examination of receipts, was not greatly mitigated by assets. Thirty-three per cent of the females, 38 per cent of the unrelated TABLE

9.14

M E D I C A L D E B T S OF A G E D U N I T S , A P R I L ,

1952

(Percentages) Size of debt

None $1-199 200 or more

Couples

95-6 3-3 i. i

Unrelated males

Unrelated females

97.2

95.8

2.0

3-5

0.8

0.7

males, and 58 per cent of the couples had some liquid assets. However, such information as is available (see appendix D) suggests that the holdings of liquid assets had a high positive correlation with size of income and hence the mitigating effects of asset holdings for those with low incomes can easily be exaggerated. The size of the available cushions against the economic shocks of old age seems frequently to be inadequate. The income advantage of couples is generally not transferable to the widow. Illness had its primary shock effect through impairment of earning capacity, but for a significant minority of the aged, medical expenses were large relative to annual income. Although the total amount of assets held by the aged seems large relative to annual medical payments, the large group without assets, the probable recurrence of medical expenses, and the use of assets for many purposes, including current living expenses, caution against too optimistic an evaluation of the adequacy of assets to meet medical expenses.

Chapter X

Conclusion

A great deal of attention has been given to the economic problems of the aged in recent years, all of it well intentioned, but much of it misdirected in terms of finding ultimate solutions to the basic problems. The primary purpose of this study has been to provide a factual platform for subsequent discussion. This chapter does not contain a comprehensive summary of the findings, nor does it present a set of recommendations for public policy. Instead it attempts to direct emphasis to the basic elements of the aged problem; elements which will repay research and which should condition social policy. As a first operating guide, it is essential to recognize that the "aged" are not a homogeneous group. The older population consists of several important subgroups, each with special characteristics and problems. Chief among them, in order of importance, are those who are victims of the following sources of economic distress: widowhood, illness, and obsolescence of their skills. Only through fundamental attention to the problems of these groups will much significant progress be made in changing the economic status of the aged. It is essential to suppress that basic stereotype of the aged that depicts a somewhat bitter old man who has been forced into involuntary retirement by age, thus depriving society of potential production and depriving the individual of income, of purpose, and of happiness itself. Such persons exist. But they are not typical, nor are they even close to the core of the aged problem. 146

Conclusion

|

147

It is obvious that much more attention should be given to the problems of aged women. Women substantially outnumbered men in the upper age brackets in 1952, and unrelated females were the largest single group in the aged population. Furthermore, the disparity between the numbers of men and women is expected to grow dramatically in the period when those presently living reach old age. But the mere size of the disparity presented in the demographic estimates underestimates the magnitude of the problem because of the tendency of women to marry men several years their senior. It will not be long before a majority of the economic units over 65 are unrelated females, the great majority of them widows. The particularly dismal economic position of unrelated females was shown by every examination of the data. The chief reason for the economic distress of aged women is that they lack employable skills, or pension rights or other benefits of past labor-force participation. The problem will doubtless diminish with time. The past few decades have witnessed an increasingly active degree of laborforce participation by women, who will, consequently, be better equipped to cope with the financial problems of age. As more women who have had significant labor-force participation reach 65, they will be better able to qualify for continuing employment, to qualify for substantial private or public pensions, or be more likely to have accumulated savings. But unless there is a rapid increase in labor-force participation the majority of aged women will continue to suffer from the disadvantages associated with lack of occupational background. One step in the solution of the problems of the aged is, therefore, the provision of employment opportunities for women of 40 or over. Here, as elsewhere, the problems of the aged can most constructively be attacked before they develop rather than after. Notwithstanding the substantial progress that can be made along these lines—and some important questions need to be asked: for example, "What proportion of her adult life need a woman spend in the labor force to be equipped for the period of aged widowhood she should expect?"—employment is not the solution for all of the problems of women, at least given a society with the structure and conventions of our own. It is also important that more attention be paid to the survivors' provisions of public and private pensions. In a real sense, these provisions are not the fringes of a pension

148

|

Conclusion

program, but the heart of it. It may, of course, be argued that solution of the problems of widowhood can be achieved through more adequate and more systematic provision by the husband during his working life in the form of private savings and life insurance. It is, in part, an ideological question whether such provision is to be preferred to provision through contributory (or other) pension plans, but the facts of life, at present at least, suggest that private provision is likely to be inadequate provision. The problems of widowhood are of course not entirely economic, but the psychological and sociological problems associated with the death of a husband are accentuated by critical economic difficulties. Although many aged widows now live in the households of their adult children, it is usually economic necessity rather than family ties that leads them to do so. The chief economic problems of men 65 and over are somewhat different: their main symptom is withdrawal from the labor force; their main causes are poor health and obsolescence of skills, frequently in combination. The causes rather than the symptom should be attacked. Although programs for reemployment of older men, operating either on the individuals themselves or through attempts to educate business and union leaders, have some scope and are worth undertaking, they are necessarily limited to a small proportion of the aged men not in the labor force. What is required is attention to prevention rather than to rehabilitation. This, in turn, implies a shift of focus to younger men. Medically, this shift has already occurred. The chief diseases of the aged have their origins at much earlier ages, and the nature of the aging process, including the development of degenerative diseases, is now being studied in the young and middle-aged. It is well to remember, however, that medical progress is not likely to change the character of the economic problems of the aged fundamentally. For although medical advances may be expected to lengthen working life, they may also be expected to increase life expectancy, so that the gap between retirement from the labor force and death will remain and may very probably increase. Obsolescence of skills is part of the price of social progress. At any given time the aged population will reflect the patterns of education, of training, and of developed skills acquired a generation or more before. So far as the economy has progressed, and has

Conclusion

| 149

therefore developed or demanded new skills, so far is it likely that the aged will be less well equipped than the young to cope with the new requirements. Although attention should be given to mitigating this obsolescence, it seems at least strongly probable that it can never be overcome.1 In considering how best to cope with obsolescence, it is well to remember that continuing education is likely to be more effective than rééducation, and that the earlier the age at which rééducation starts the more satisfactory it is likely to be. The current debate over automation has started a constructive discussion of this problem, but research is needed to suggest the most effective means of attacking it. Automation is, after all, only a dramatic phase of and phrase for the continuing process of change that has characterized our economy for more than a century. But this new attention has already led to the suggestion by a spokesman for a major union that what is required is scope for development of greater flexibility in job experience, even at the price of modifying time-honored principles of seniority.2 Even though discussion of automation is not directed primarily at the problem of the aged worker, it is directed at the problem of obsolescence, which is a fundamental aspect of the problem; the man of 45 rendered less useful by industrial progress today, will, in the absence of retraining, soon be left behind completely. Notwithstanding the progress that can be made, it is evident that large numbers of the aged will, in the future, as at present, be unable to provide satisfactorily for themselves. It is therefore apparent that the comprehensiveness and the generosity of programs of public or private support will be a major factor in determining the economic status of aged persons. Continuing expansion of the social security program, of private pension plans, as well as of programs of public and private assistance, seem central to the problem 1 Melvin W . Reder has suggested ("Age and Income," American Economic Review, Vol. 44 (May, 1954), pp. 661-670, especially pp. 669-670) that the problem may be transitory. He notes that the present aged averaged (median) about eight years of schooling, whereas today's youngsters average about 1 2 years. Since education and income appear positively associated, will not future generations, he asks, be better off? In a similar vein, he might have argued (but did not) that a chief cause of obsolescence was the shift away from agriculture in the period during which the present aged grew up. But, one might ask, will not every generation find itself behind the times in some critical characteristic and seriously unequipped because of it? 2 See the address by James Stern, consultant on automation to the U A W - C I O , at the Eighth Annual Conference on Aging, University of Michigan, Ann Arbor (1955).

150

| Conclusion

of income maintenance, which is, and will continue to be, a critical dimension of the problem of old age. The question of the form which efforts to provide adequate incomes should take is beyond the scope of this study, but it seems essential to indicate that some program is needed and that questions of form must not be permitted to defeat the objectives of every program. It seems appropriate to suggest that it is time for a careful reexamination of some of the provisions of existing public and private programs. Consider, for example, the so-called "retirement test" for eligibility for social security benefits. These provisions make ineligible for benefits those persons under 72 whose earnings exceed specified amounts in any month or year. The origin of these provisions (which have been periodically liberalized since the passage of the Social Security Act, most recently in the 1954 amendment) can be found in the economic conditions of the depression under which the act was originally passed. They were designed, in part, to encourage the withdrawal from the labor force of aged workers as a partial solution to the unemployment problems faced by all age groups. Out of the context of chronic unemployment, such provisions serve merely to limit the total income of persons receiving benefits. The incomes of beneficiaries are low enough that such a limitation might properly be reconsidered. Alternatively, if the size of benefits were made to increase with the length of covered employment and the age at which retirement occurred, it might both increase incentives to stay in the labor force and mitigate income inadequacies. This example is only suggestive: virtually all features of our present program of aid to the aged might profitably be reexamined. The question of cost must inevitably arise. A question frequently posed is whether an increasing number of aged persons combined with conventionally later entrance into the work force does not (or will not) constitute a heavy drain upon the product of the relatively smaller proportion of productive workers. Our findings suggest that whatever the cost, it is largely an inevitable one in terms of the economy as a whole. That it has not proven oppressive during past decades, when the proportion of aged increased sharply, suggests that it is unlikely to be oppressive in the future, when the aged population should be a more or less stable percentage of the total population—particularly when increases in productivity are

Conclusion

|

151

to be expected. It is essential to recognize that the real decision is whether income inadequacies of the aged are to be met by submarginal living, by informal support by others, or by well-organized public or private programs. In terms of drain on the economy, it makes no difference whether this support is provided by children, by the federal government, by private charities, by private pensions, or by state and local governments. Other issues, of course, affect the appropriate nature of programs of support, but mere computations that show, for example, social security payments as a percentage of payroll, only cloud the essential issue. The major conclusion is that the solutions of the problems of the aged are neither easy nor cheap. The aged do not constitute a great reservoir of unutilized labor whose members might benefit psychologically, socially, and economically by some magical lifting of artificial barriers to their utilization. If the contrary were true, the solution would be simple; it would be possible to benefit not only the aged but the economy as a whole by reducing the aged's "drain" on the productive population and increasing output as well. Rather, the extent of the labor reserve within the aged group at present is highly limited and may in the long run be increased only by finding solutions to the basic causes of the low rate of utilization of the aged in the labor force. Whatever progress can be made is, of course, worth seeking. But the problems of the aged will inevitably require continuation and expansion of programs designed to provide income (or its equivalent) for those of the aged who are not able to provide for themselves. In this sense the aged are a national problem. The dominant economic problem of the aged—the provision of adequate resources for old age—is in a fundamental sense simpler than their medical, sociological and psychological problems, in that it can be assumed by others. Wherever a solution is sought, it provides a challenge not only to research, but also to society. If once we succeed in solving this basic economic problem, new questions will arise. What will be the impact of the aged population upon the structural characteristics of our economy? What kind of goods will the aged demand? What sort of housing is appropriate? These questions and many others like them are compelling I as we move toward the time when an aged population will be able to exert an appreciable economic force in our markets, rather than

152

|

Conclusion

merely struggling to subsist. Originally we intended to consider these questions in this study. They now seem premature. It will be some time before these questions become pressing; much work must first be done on the urgent problems of today.

Appendices

Appendix A

Source and Reliability of Data

Source of the Data The basic data used in this study were collected by the Bureau of the Census in April, 1952, through the Current Population Survey (CPS) and from a special Follow-Up Survey, also in April, 1952, of persons 65 years old and over included in the income subsample, April, 1952, CPS. Data from both the CPS and the Follow-Up Survey were used in preparing the punch cards which provided the tabulations. The population from which the sample was selected was the population of the United States in April, 1952, but excluded members of the armed forces living in barracks, etc., on military reservations, and inmates of penal and mental institutions and homes for the aged, infirm, and needy. The sample consisted of approximately 15,000 households in 68 samples areas located in 42 states and the District of Columbia.1,2 Approximately 3,000 of these households 1 Thie main features of the sample design are described in Morris H. Hansen, William N. Hurwitz, and William G. Madow, Sample Survey Methods and Theory, Vol. I, Chapter 1 2 B [prepared by J. Steinberg] (New York: John Wiley & Sons, 1953), PP- 5 5 9 - 5 8 2 . In February, 1954, the Bureau of the Census introduced a redesign of the CPS, containing approximately the same total number of households but spread over 230 instead of 68 primary sampling areas. For a thorough discussion of the changes and their implications see: Hansen, Hurwitz, Nisselson, and Steinberg, "The Redesign of the Current Population Survey," Journal of the American Statistical Association, Vol. 50 (September, 1955), pp. 7 0 1 - 7 1 9 . 2 The sample used to collect income and related data in 1 9 5 2 was, in fact, a 60 per cent subsample of the total CPS. The total CPS sample consisted of approximately 25,000 households.

155

156

] Appendix A

contained one or more persons 65 and over. These households, containing a total of approximately 3,600 aged individuals, were included in the Follow-Up Survey. Population characteristics relate to April, 1952; data on income and on expenditures relate to the calendar year 1951. Data Collected The data of the CPS are now so familiar that no comprehensive summary of its coverage will be presented.8 In addition to information collected monthly concerning labor-force status, the annual (April) income report collects information on total income of individuals and families subdivided into three categories: "wage and salary," "self-employment," and "other." General classification information concerning sex, age, place of residence, and family status (among others) is also available. The additional coverage of the Follow-Up Survey may be seen by examination of the questionnaire reproduced at the end of this appendix.4 Source of the Inflated

Estimates

For many purposes it has been desirable to report the survey results in terms of the total population rather than in terms of the number of sample cases. For these purposes the sample totals have been inflated. The over-all inflation ratio (size of population : size of sample) is approximately 3,333. Individual inflations, however, can not be reliably made on the basis of this ratio alone, for reasons indicated in subsequent paragraphs. For most categories the inflated estimates may be crudely determined by using a ratio between 3,333 and 4,000. Two considerations mitigate against use of a crude inflation ratio. First, schedules on which there was partial or total nonresponse to certain questions need to be included. For the purpose of inflation, cases not reporting some characteristic (e.g., income) were distributed in the same proportion as those that did report 3

See Bureau of the Census, Current Population Reports, especially Series P-60 (income) and P-57 (labor force). 4 Certain data, for example, those of Section IV (Condition of Dwelling Unit), were collected for the benefit of other agencies (in this case, Federal Housing Authority) and have not been heavily utilized in our analysis.

Appendix A

| 157

the characteristic. Second, because of the use of a stratified sampling procedure, it was necessary to weight sample sizes to conform to the underlying nature of the stratification. For example, the sample included a disproportionate percentage of aged living in urban areas, and inflation procedures designed to correct this disproportion were employed. T A B L E A.L STANDARD ERROR OF ESTIMATED NUMBERS, APRIL, 1952, FOLLOW-UP SURVEY Size of estimate

Standard error

10,000 25,000 50,000 100,000 250,000 500,000 750,000 1,000,000 2,500,000 5,000,000 7,500,000 10,000,000

7,000 11,000 16,000 23,000 36,000 50,000 60,000 69,000 100,000 120,000 120,000 90,000

All inflated figures used in this study have been computed by the Bureau of the Census and have in some cases been adjusted to agree with independent estimates of the civilian noninstitutional population. Because of sampling variability the figures shown may differ from those that would have been obtained from a complete enumeration. Table A.i provides a metric of the errors due to sampling variability. The chances are about two out of three that the difference due to sampling variability between an estimate and the figure obtained from a complete census is less than the standard error of estimate. Computational Procedures Per cent distributions.—Distributions are computed on the basis of number of sample cases reporting the characteristics involved in the particular distribution. This procedure provides distribu-

158

| Appendix A

tions which: (a) in some cases differ from distributions based upon inflated estimates, (b) involve implicit assumptions regarding the treatment of those schedules on which there was partial or total nonresponse to questions. Differences between distributions based upon sample cases and distributions based upon inflated estimates are in general very small and well within the limits of expected sampling variability. The major exception concerns classification according to size of residential area, for which data sufficient to make alternative calculations is included in the appendix tables. The treatment of nonrespondents follows census procedure, but deserves slight additional comment. Nonresponse varied substantially by type of question, being very small with respect to characteristics such as occupation, family status, residence, nativity, and labor-force availability, and being greatest in questions concerning income, receipts, and their components. Complete income information was lacking for approximately 7 per cent of aged economic units, which compares with the 6 per cent nonresponse for all persons in the April, 1952, CPS. The maximum nonresponse percentage occurred for total receipts (11.5 per cent). The effect of these nonresponses upon the distributions is slight, as may be seen in tables in Appendix E. The procedure used implicitly assumes the (unknown) distribution of nonrespondents to be the same as that of the respondents; where partial information was available it was utilized. In effect, the procedure was to compute percentages on the basis of those reporting. Table A.2 should make the procedure clear: the population is to be divided according to a two-way classification (A,B), and those classified B are to be subdivided into categories B.i and B.2. Computation of medians.—The median value of a series is defined as a value such that no more than half of the distribution lies above it and no more than half lies below it. Loosely, it divides the distribution into two equal groups. For grouped distributions, application of this definition requires an assumption about spacing of the items in the median class interval; the standard assumption, here used, is that the items are equally spaced within the interval. Where the median value lies in an open-end class interval (e.g., $3,000 or more) no computation is made, but the median is described by the lower limit of its value (e.g., more than $3,000).

Appendix A

|

159

At one point the procedure utilized in this study differs from that used in Census studies, and leads, wherever median income is below $500, to a figure that is lower than the median obtained by Census procedure. We specifically utilize information concerning zero income, if reported. Census procedure combines all incomes under $500 into a single class interval before computation of medians, and thereby (given the spacing assumptions) distributes cases of zero income in the interval $o-$4gg. T A B L E

A.2

I L L U S T R A T I V E C O M P U T A T I O N OF P E R C E N T

DISTRIBUTION N u m b e r of sample cases

Total Category A Category B—total B.L B.2

Substatus not available

98 72

18

Status not available

Since, for the aged, absence of income is not uncommon and medians frequently do fall below $500 we have felt obliged to utilize the information concerning zero incomes. It may be noted that, under our procedure, but not under the alternative procedure, the median value may be zero. It must be reemphasized that comparability of our data with Census material is thus affected only where medians are below $500. Computation of arithmetic means.—Computation of means also requires spacing assumptions. We assumed that for all closed-end class intervals, the mean in that interval falls at the mid-point of the class interval. Thus for total income and total receipts and for earnings distributions all levels under $15,000 were so treated. For all other distributions of individual sources of receipts (e.g., pensions, assistance) closed-end intervals run to $2,500. This midpoint assumption, while conventional, probably typically overstates the true mean in any class interval owing to the J-shaped

i6o

| Appendix A

nature of most income and receipts distributions. A contrary bias is discussed below. For open-end class intervals ($15,000 and over; $2,500 and over; as appropriate) a more troublesome problem is posed. For, given no logical upper limit to the amount in the open-end class, there is no logical upper limit to the means. The most satisfactory method of surmounting this problem is to fit a Pareto curve to the frequency distribution and thereby to estimate means for the openend intervals. Chiefly because of the very small frequencies in our sample that fell in the open-end intervals, but also because of certain doubts as to the applicability of a Pareto distribution to the aged, this procedure was not applied, and we have instead made the following arbitrary assumptions: for class intervals "$2,500 and over" we have assumed the mean to occur at $3,000; for class intervals "$15,000 and over" we have assumed the mean at $17,500. These arbitrary figures for the open-end intervals are doubtless low. Their use was dictated by these considerations. First, since our interest in means is in the means of the entire distributions it seemed desirable to offset the upward bias of the assumptions used in closed-end intervals, discussed above. Thus, for example, the "true" mean for the interval $10,000-$ 14,999 is doubtless below the $12,500 used—perhaps as little as $11,000. The assumptions used for total incomes yield aggregates that are consistent with available independent estimates of these magnitudes. Second, the assumptions used are internally consistent: aggregates based upon the sum of particular sources of receipts are consistent with aggregates based upon total receipt distributions, notwithstanding the highly different class intervals employed. These reasons are pragmatic. We believe they yield aggregates that are, for our purposes, accurate indexes of magnitude. The Reliability of the Data The general context within which these data must be regarded is the accuracy of Census data. This problem has been discussed elsewhere.5 It is generally agreed that field-survey data measuring 6 See U. S. Bureau of the Census, Census of Papulation: 1930, Vol. II, Part I; S. F. Goldsmith, Appraisal of Basic Data for Size Distributions, Studies in Income and Wealth, Vol. 13, National Bureau of Economic Research (New York, 1951); H. P. Miller, " A n Appraisal of the 1950 Census Income Data," Journal of the American Statistical Association, Vol. 48 (March, 1953), pp. 28-43. See, in addition, the series of papers presented at the March, 1956, Conference on Research in Income and

Appendix A

| 161

income are subject to errors whose systematic effect is to understate the true aggregate of income. Do such errors operate with special force upon the aged population? There are several reasons for thinking they may. 1. The aged are predominantly a low-income group. Given a tendency for reporting units to understate their incomes, the effect will be most pronounced at both extremes of the income distribution and it is possible that a significant overstatement of the frequency below a specified low level of income will occur. 2. The aged receive proportionally a larger share of their total incomes from nonearned incomes, and these sources tend to be less fully reported than earnings.8 3. The special eligibility conditions concerning Old Age Assistance and Old Age and Survivors' Insurance benefits may provide incentives to understate income to an enumerator, notwithstanding the promise of confidential treatment. Despite these considerations, we believe the data of the FollowUp Survey to be of the same general order of accuracy as other Census-collected data on most of the measured characteristics. The use of Census procedures, enumerators, questions, and tabulation procedures seems the best guarantor of this assertion, but numerous available cross checks with other sets of data support it as well. Sampling variability.—Because of sampling variability, data shown in this monograph may differ from those that would have been obtained from a complete enumeration of all persons 65 and over. The standard error is a measure of the sampling variability. The chances are about two out of three that the difference due to sampling variability between an estimate and the figure that would have been obtained from a complete census is less than the standard error. The chances are about 19 out of 20 that the difference is less than twice the standard error and 99 out of 100 that it is less than two and one-half times the standard error. The allowance to be made for sampling variability in a percentage depends, in general, both on the value of the percentage and Wealth on the topic, "Appraisal of 1950 Census Income Data," to be published soon by the National Bureau of Economic Research. 6 See the papers presented at the March, 1956, Conference on Research in Income and Wealth: Selma F. Goldsmith, "Relation of Census Income Statistics to Other Income Data"; Leon Pritzger and Alfred Sands, "Post Enumeration Survey: Study of Income Distribution of the 1950 Census of Population." Data of the Follow-Up Survey also supports the assertion.

e* co v i h - o o os O O O O O O O O fi M

CO ^ ^O O O O 6 O ~

c*) co ' t VÍ m w M M M

vr» OO M n ^ V û OO CO O O O w w l-i M M W M

vo t>s, M 00 O d O O m M «

CO V» vr» « Cí «

N O vû c í v o Os h co O H N í í r t c i n n n n

TJ- Os (S l o f v O es c< co n n f i t

os n o oo co t*"- o c o v o O w c* c* n n - t ^

h VÛ ^ M »-i

^ H \Û co ^t" t

-t

oscí^o vi in vi

vr» vn O s » vn O ^ O M cí C O ^ V J S O tN N 0 0 oo

O O O O O O O O O •i c* V i O ^ o O O

Appendix A

|

163

the size of the base on which it is determined. If two percentages are of the same magnitude, the one based on the larger number of cases in the sample is, in general, less variable than the one based on the smaller number of sample cases. Table A.3 presents the standard errors of estimated percentages for inflated bases of varying size. TABLE STANDARD E R R O R

or

SELECTED

Distribution

All economic units. Married couples. . . Married couples, in urbanized areas of 250,000999.999

A.4

MEDIANS, APRIL,

1952,

FOLLOW-UP

Estimated median total unit receipts

3

778 1,491 1,912

SURVEY

Range of median (2 cases out of 3)

?

750-800 1

>430_1>i7°

1,720-2,160

For reasons discussed elsewhere in this appendix, the linkage between sample cases and inflated bases is approximate. The large sampling variability for sample groups under 100 (approximately 400,000 as an inflated magnitude) has been noted in tables. Distributions for sample groups of less than 50 have not been computed. Standard errors of means and medians can not be summarized easily because they depend upon the standard deviations of the distributions as well as the size of the base. Table A.4, indicating the standard error of selected medians, is, therefore, illustrative only.

Budget Bureau No. 41-5209: Approval Expires June 30. 1982 CONFIDENTIAL -The Information furnished Is accorded confidential treatment. it cannot be used for purposes of taxation. Investigation or regulation.

Sample Area

Form I1R-1 (3-12-52)

Line No. of Sample Person

U.S. DEPARTMENT OF COMMERCE BUREAU OF THE CENSUS

FOLLOW-UP SURVEY OF PERSONS 65 AND OVER

CPS Contrbl No.

Line No. of Enumerators respondent initials

Number of visits

Section I - WORK EXPERIENCE A. FOR PERSONS NOT IN THE LABOR FORCE B. FOR ALL SAMPLE PERSONS (except where previously ("U" or "X" in item 24 on Field Control Card) instructed to skip to Section II.) 1. Have you ever worked, not counting volunteer work or work around the house? I

I NO - (Skip to Section II)

I

I Yes - (Ask question 2)

2. How old were you when you last worked?_

(years)

1. What kind of work did you do longest during your working life?

2. In what business or industry did you do this work?

(For women, if less than SO above, skip to Section II.) 3. What was your last job or business?

3. Class of worker

(P, G, 0, or NP)

Occupation 4. How many years did ou do this kind .work?

Industry Class of worker. (P, 0, 0. or NP) 4. Why did you stop working? .

5. What is the longest yoti ever worked for any one employer? _

(years)

b. Did you ever have to leave a job because you reached the compulsory retirement age? 5. (Do not ask this question if *UH in item 24 on Field Control Card.) Are you wel I enough to work now? •

NO

I

I Yes

• I

(Ask:)

Are you interested in going back to work either full tine or part tine? 1 •

Yes - Full time regularly

2 •

Yes - Part time regularly

No I Yes (Ask:)

then did this happen? 7. test year • that is, in 19S1 - did you work at all? •

I

NO

I Yes (Ask:)

3 QI] Yes - Occasionally 4 •

NO

Comments on working

Footnotes and ex; '.¡mations

Comm-DC 28069

(a) How many weeks did you work full time last year? (b) How many weeks did you work part time last year?

GENERAL NOTE FOR SECTIONS II TO VII — DO NOT PILL THESE SECTIONS FOR WOMEN WHOSE HUSBANDS ARE ALSO IN THE SAMPLE. BUT SKIP DIRECTLY TO SECTION VIII. Section II - SOURCE OF INCOME (Fill this section only if an amount has been entered for income other than earnings in item 24 on Field Control Card.) "The enumerator «ho «as here earlier this month asked a few questions about income and was told that you [and your (wife) (husband)] had some income other than earnings in 1951. Now we would like to ask a little more about this Income. In 1951. how much money did you [and your (wife) (husband)] receive from ?" Sample Person Spouse (Enter amount or n o n ) A. Social Security (old-age or survivor's insurance benefit)?. B. Other government pension (Federal, State, local, military, veterans, railroad retirement)? C. Pension from private employer or union? D. Interest, dividends, annuity, or rent? E. Old-ag' assistance or other public relief? F. Regular contributions of money from persons outside the household? Q. Other («peej/y) Section III - MISCELLANEOUS CASB RECEIPTS A. Did you [or your (wife) (husband) ] receive any occasional cash gifts last year? •

No



Yes

(amount)

B. Did you for your (wife) (husband) ] receive any other money such as an inheritance or a lump-sua insurance payment? •

NO

I I Yes («pecit,)

(amount)

(type)

Section IV • CONDITION OF DWELLING UNIT AND TYPE OF FACILITIES (Fill this section only for married couples living in dwelling-unit quarters.) Condition of Unit 1 •

Not dilapidated

2 d]

Dilapidated

Piped Water Supply

• • • •

Coma-DC 28069

Hot and cold piped running water inside this structure Only cold piped running water inside tblB structure Piped running water outside this structure No piped running water (hand pump, wellj etc.)

Type of Toilet

• a n

• •

Flush toilet Inside this structure Flush toilet outside this structure Privy, outhouse, or chemical toilet NO toilet for this unit

ToiletExclusive Use

• 2d



Installed Bathtub or Shower • For this unit's exclusive use

For this unit's exclusive use

1

Shared with another unit

a D

Shared with another unit

No toilet for this unit

s D

No bathtub or shoner for this unit

Forn 1IR-1 (3-12-52)

Section V - SOURCE OF SUPPORT "Now I would like to get some information about your living arrangements." A. FOR SAMPLE PERSON LIVING ALONE OR WITH NO OTHER RELATIVES EXCEPT HUSBAND, WIFE, OR MINOR CHILDREN

B. FOR SAMPLE PERSON LIVING WITH RELATIVES NOT COUNTING HUSBAND, WIFE, OR MINOR CHILDREN 1. Are these living quarters owned or rented by someone living here?

1. Do you Tor your (wife) (husband)J own or rent these living quarters? 1

1

1 Owns (Ask:)

1 OWNED (Ask:) Do you [or your (wife) (husband)] own these living quarters or do the people you live with own them?

Are you. still making payments on a mortgage or purchase contract 1

1



NO



Yeo

Are you still making payments on a mortgage or purchase contract?

1 Rents (Ask:) What is the monthly rent (in these cniart.prs? (Enter amount or "rent free")

1

1



NO



Yes

1 Relatives he lives with own them (Ask:) Do you pay any rent for your quarters?

2. last year, in 1951, did someone take aver any of your Tor your (wife's) (husband's) J bills and pay them directly? We mean, for exaivle, bills for food, hospital, doctors, insurance, clothing or taxes. •

1 Person or spouse owns them (Ask:)

1

1 No

1

1 Yes (Monthly amount)

["""1 RENTED (Ask:)

No

What is the monthly rent on these quarters?

1 Yes (Ask:) (a) Did these bills amount to as much as $200? •

No



(Enter amount or "rent free")

Yes

Do you pay any part of this rent?

(b) Who was helping you out? Children



All.

2 [ID Other relatives

1

1 Part

3 •

I

I None

1 •

Nonrelatives such as friends and neighbors

2. Last year, in 1951, did you contribute your share of the living expenses of the family such as paying for food, utilities, laimdry, or other household costs?

4 (71 Other (specif,)

1

Footnotes and explanations

| Yes

|

| No

3. Last tear, in 1951, did someone take over any of your [or your (wife's) (husband's)3 bills and pay them directly. We mean, for exavle, bills for food, hospital, doctors, insurance, clothing, or taxes? • 1

No 1 Yes (Ask:) (a) Did these bills amount to as much as $200? •

No



Yes

(b) Who was helping you out? 1 •

Children

2 C D Other relatives

Coam-DC 28069

- 3 -

3 0

Nonrelatives such as friends and neighbors

4 O

Other (speeil,) For» IIR-1 (3-12-52)

S e c t i o n VI - MEDICAL EXPENDITURES A. During 1991 did you [ o r your ( w i f e ) (husband) J have any i l l n e s s which required: Yes

No

• • • •

• • • •

Physician or surgeon?

Regular purchase of dnigs? B. I f ' Y e s , " in any part of A, ask:

1. Did you or your family pay any part of the expenses for these services? (Do not include amounts l a t e r reimbursed by insurance.) •

No

I

I Yes

(Ask:)

How much was paid?

l O

Under $50

3 HZ] $150 to $300

2 C D $50 to $150

4 •

$300 or more

2. Was any part of these expenses taken care of by hospital Insurance or medical care plan? •

No



Yes

3. Was any part of the medical services provided without charge by a physician, hospital, c l i n i c , I

I No

I

etc.?

i Yes

4. Was any part of the expenses taken care of in some other way? •

No



Yes (specify)

C. Do you [ o r your ( w i f e ) (husband)] at t h i s time have any unpaid b i l l s or debts arising from the hospital and medical expenses you had in 1951 or e a r l i e r ? I

I No

I

I Yes

(Ask:)

Are they as much as $200? •

No

I

I Yes S e c t i o n V I I - SAVINGS AND ASSETS

A. There i s a l o t of interest in what savings people S3 and over may have t o f a l l back on to n e t emergencies. Do you [and your ( w i f e ) (husband)] have any of the following types of savings? 1. Money in bank or cash savings?

Q ] No

Yes - Amount $

No

QH Yes - Amount $

3. Stocks or bonds?

Q ] No

C D Yes - Amount $

4. Home or other property in which you have $3,000 or more invested?

r

Sample Person 2. L i f e insurance?

L ) No

[

Spouse

(face value of p o l i c i e s )

j Yes

B. Last year did you [and your ( w i f e ) (husband)] use up any savings, cash bonds, borrow on l i f e Insurance, or s e l l or mortgage property to meet expenses? I

I No

Yes

(amount) Section V I I I

- NATIVITY

Were you born In the United States? I

I Yes

I

I Ho

(Atk:)

Were you born of American parents? I

I Yes

I

I No

(Aak:)

How many years have you been l i v i n g in the United states? .

Comn-DC 26069

-. 4 -

Por» IIR-1 (3-12-52)

Appendix B: Part I

Fifty Year Projection of the Population

Our forecast of the size of the aged population and of its relationship to the total population is based on a fifty-year projection of the population prepared for the purposes of this study. This forecast should not in any sense be regarded as a prediction of the future course of the population of this country. It merely presents some possibilities that seem consistent with past relationships and with plausible assumptions about the major factors influencing population growth. The Bureau of the Census has published similar population forecasts for the period up to 1975. 1 We have adopted these projections for the period they cover and, as will be seen below, our forecasts for 1975-2000 are simply a mathematical extension of the Census Bureau projections. Thus the assumptions used by the Census Bureau have been built into our projections. We shall discuss these assumptions in this appendix but first we describe the mechanics of the process by which the Census Bureau estimates were extended for 25 years. Calculating Procedure The Census Bureau estimates of the future population and our extension of them were derived by the "cohort-survival method." The underlying idea of this method is that the persons of any given age 1 U. S. Bureau of the Census, "Illustrative Projections of the Population of the United States, By Age and Sex: 1955 to 1975," Current Population Reports, Series P-25, No. 78 (August 21, 1953)-

168

Appendix B

| 169

at any given time are the survivors of the group that was one year younger one year before, with adjustments for immigration and other minor factors. Thus, given any benchmark—for instance the 1950 Census of Population—and given assumptions about survival and immigration rates for each age group, it is possible to project the population by age step by step for as long a period as the survival and immigration assumptions are felt to be appropriate. The number of children in the youngest age group is projected in a similar manner on the basis of the female population and assumptions about age-specific birthrates. In this discussion we shall use the word "cohort" to denote a group of persons of the same sex in the same five-year age bracket. It is evident from the discussion in the preceding paragraph that the size of each cohort in each year is a function of its size five years earlier. We should, in fact, expect this functional relationship to be almost linear,2 since the major components of the cohort at the later date are the survivors of the cohort at the earlier date (whose number is approximately proportional to the size of the cohort at the earlier date) and the immigrants (whose number is roughly constant, independent of the size of the original cohort). We use the phrase "almost linear" because the relationship between the size of a cohort at two different dates cannot be expected to follow an exact mathematical formula, the major disturbances being: the discrepancy between the actual number of deaths in a five-year period and the number predicted from the death rates, fluctuations in the number and age composition of immigrants, changes in the internal age-composition of a five-year cohort, statistical errors in enumerations and reporting of age made at the beginning and the end of the five-year period. But these sources of disturbance are all relatively minor; one would expect and does find a very high linear correlation between the size of a cohort at any date and its size five years later. But now consider a case in which we do not have actual cohort sizes for a number of years but, instead, projections of these sizes. Then the successive sizes of a cohort are clearly related functionally; indeed, they are computed mathematically from each other. 2 That is, if we consider a number of cohorts of different sizes advancing through a specific five-year age-step, then there will be a single linear relationship which, if applied to the size of each cohort at the earlier age, will give a good approximation to its size at the later age.

170

| Appendix B

Furthermore, random fluctuations and enumeration errors which disturb relationships in actual populations do not affect mathematical projections. Thus the grounds for expecting a highly regular relationship connecting the size of a cohort at two successive dates are strengthened when the cohort sizes used are mathematical projections. This is the situation in our problem, which is to extend the Census Bureau forecasts for 25 additional years. TABLE B.l SURVIVAL RELATIONSHIPS, WOMEN, 6 5 - 6 9

(Thousands) Estimated size of cohort

Five-year survivors

1950

2,590

2,203

'955

2,750

2,362

Year

I960 1965 1970

!975

3.124 3,478 3,747 4,202

SOURCE: U . S. Bureau of the Census, Current Population

2,680 2,984

3,215

Reports, Series P-25, N o . 78, p. 6.

Table B.i and figure B.i illustrate this. Column one of table B.i gives the number of women aged 65-69 at five-year intervals as projected by the Census Bureau.3 The second column gives the number of women aged 70-74 at a date five years later than the one given in the stub. The last column is the difference between the first two. It should be noted that all causes of change—migration, errors in reporting ages, and the rest—are incorporated in this "net decrease" column along with the effects of mortality. Figure B.i is a scatter diagram in which the size of the cohort is plotted as independent variable and net decrease as dependent variable. The points for the last three years lie precisely on a straight line, but the points for the earlier years do not. This is to be expected, since the immigration and mortality assumptions used by the Census Bureau for the early years are not the same as those for the later years. This figure shows that the rather intricate mortality and immigration assumptions used by the Census Bureau, when held constant, implicitly define a linear relationship between the size of a cohort at a given age and its size five years later. This 8

Op. cit., p. 6.

172

| Appendix B

implicit relationship, moreover, is easily determinable from the data of table B.i. By using this relationship and the analogous ones for males and for different age groups, the course of the population can be traced out by five-year steps for as long as desired or as long as the assumptions made in the Census Bureau projections are felt to be applicable. T A B L E B.2 F E R T I L I T Y W E I G H T S U S E D IN PROJECTING B I R T H S

(Live births per year per 1,000 women) Fertility weights Age of woman

Higher fertility (Assumption A)

15-19. 20-24.

25-29.

3°~3435-3940-44. Total.

Lower fertility (Assumption B)

1939-194° average

no 56 17

61.2 52-3 *35-4 81.1 38.8 II.4

53-3 132.8 121.1 82.0

660

480.2

451.2

213 176

!

45-4 16.6

SOURCE: U. S. Bureau of the Census, op. cit., p. 3.

We have carried out this procedure up to the year 2000, using two sets of assumptions. Although the strategy just sketched was followed consistently, the details of the procedure had to be adapted to the peculiar needs of different age groups. Projecting the number of children under five.—Children under five are not the survivors of any preceding group, and thus the appropriate independent variable for predicting their number at any date is the number of women of childbearing age at the midpoint of a five-year interval terminating with the date for which the estimate is to be made. Fertility rates vary so much during the child-bearing ages, however, that the number of women between, say, 15 and 44 is too crude a variable to use. Instead we have used a weighted sum, using weights proportional to the age-specific fertility rates assumed by the Census Bureau. These weights are given in table B.2 and scatter diagrams for male and female children on one set of assumptions (Assumption A) are shown in figure B.2. The resultant regression equations are:

WOMEN OF CHILDBEARING AGE, W E I G H T E D (THOUSANDS)

Fig. B.2. Relation of women of childbearing age to children under five, high fertility assumption. (Source: Current Population Reports Series P-25, No. 78; number of women computed.)

174

| Appendix B

ASSUMPTION A Boys under 5, in thousands = -440.4 + 1.65447 x w e i g h t e d w o m e n childbearing age.

of

Girls under 5, in thousands = -368.6 + 1.57873 x w e i g h t e d childbearing age.

of

women

ASSUMPTION D Boys under 5, in thousands = - 4 4 3 + 1.202 x w e i g h t e d w o m e n of childbearing age. Girls under 5, in thousands = - 3 8 2 + 1 . 1 5 x w e i g h t e d w o m e n of childbearing age.

It should be noted that these formulae incorporate the effects of several factors in addition to fertility. In particular they incorporate the probability of a child's surviving from birth, at any time within the five-year interval, up to the date to which the estimate refers. They include also the effects of Census underenumeration, a rather surprising factor about which a few words must be said. The Census Bureau, and consequently we who follow them mechanically, have forecast not the actual population at various future dates but the population that would be enumerated if a Census were taken at those dates. Now it happens that there is some tendency for people to omit very young children when listing their families with Census enumerators; on the average the Census tends to miss about 436,000 young boys and 366,000 young girls.4 Consequently, the Census Bureau subtracts these amounts from their estimates of actual births, survivors, and immigrants before publishing their forecasts, and this underenumeration adjustment is reflected in our regression. In fact, it will be noticed that the underenumeration adjustment is practically equal to the constant term in our derived regression formulae. Both terms of these formulae, in short, have sound interpretations. Projecting the age groups 5-9 to 25-29.—The remarkable fact is that cohorts tend to grow during the early ages rather than to decline. This results from two factors. The first is the underenumeration adjustment: the babies who are forgotten when they are under five are reported when they enter the 5-9 age bracket, thus producing an apparent increase in the size of the cohort. The second factor is net immigration. During the early years mortality is so low that immigration, though not large, more than counterbalances 4

Op. cit., p. 4.

Appendix B

| 175

deaths, particularly for persons in their twenties. Thus the estimating approach illustrated for older women would overemphasize a distinctly minor cause of population change if applied to age groups under 30. Up to about 30, in fact, the death rate is less than 1 per cent per quinquennium, an essentially negligible factor in long-range population projection, and we have neglected it. Instead we have estimated the size of each cohort at each date by adding a flat amount to its size five years earlier to allow for the TABLE

B.3

ADJUSTMENTS FOR INCREASING COHORTS IN E A R L Y A G E G R O U P S

(Thousands) Assumption A

Assumption D

Age group Men

Under J 5-9

10-14

Men

Women

433 4

371

434

4

- 3

- 8

23 n o

- 7 O

100

20

100

34

7

34

20-24

0 20

25^9

7

3°-34

- 1 5

IS~I9

Women

2

- 1 5

37° 4

22 n o

2

EXAMPLE: T h e projected number of women aged 30-34 in 1980 is 34,000 greater than the number aged 25-29 in 1975 on either assumption. SOURCE: Derived from the Census Bureau projection cited.

effects of immigration and statistical adjustments. These adjustment quantities were derived from the behavior of the Census Bureau projections, and are given in table B.3. Projecting the age groups 30-34 to 70-74.—In the middle and later age-groups, the five-year changes in the size of cohorts are dominated by the effects of mortality, and the procedure already illustrated for women in the 65-69 year age group applies. Table B.4 presents the regression equations that were used. The constant terms are invariably small and, for the most part, meaningless. They reflect, mostly, the approximate nature of the simple relationships used to summarize a more intricate process and, in the younger age groups, incorporate the effects of migration. The slope constants measure five-year survival probabilities (along with the effects of minor disturbances) and their gradual decrease from age group to age group is just as expected, as are their general magnitudes.

176

| Appendix B

Projecting the age group 75 and over—The standard procedure had to be modified for projecting the highest age group because survivors do not graduate out of it at the end of each five year period but remain to be added to the new recruits from the age group just below. The projections of the oldest age group are therefore based on the sizes of the two oldest age groups according to the formulae: T A B L E B.4 P R O J E C T I O N E Q U A T I O N S FOR M I D D L E A G E G R O U P S Age group

35-39-•• 40-44... 45-49-•• 50-54... 55-59-•60-64.•• 65-69...

Men

0.9 + 6.8 + — 0.2 + I.5 + 3.8 + — 1.8 + 10.0 +

0.99061* 0.9772* 0.95726* O.92986* 0.8944* 0.8405* 0.7714*

Women

—0.6 3.6 1.3 3.7 —0.4 1.0 -2.7

+ + + + + + +

0.99645* 0.98911* 0.98017* 0.96761* 0.94872* 0.90955* 0.85875*

EXPLANATION: In each case x denoted the size of the cohort, in thousands, when it has attained the age shown and the formula gives the size the cohort will have when it is five years older. Thus the number of women aged 50-54 at any date, in thousands, is 1.3 + 0.98017 times the number of women aged 45-49 five years previously. T h e formulas apply for both fertility assumptions. SOURCE: Computed from the Census Bureau projections cited.

FOR M E N

Number 75 and over = 0.49617 x Number 75 and over 5 yrs before+ 0.70810 x Number 70-74, 5 yrs before FOR W O M E N

Number 75 and over = 0,54646 x Number 75 and over 5 yrs before+ 0.78759 x Number 70-74, 5 yrs before all quantities being expressed in thousands of persons. The constants in these formulae were computed so that the results agree with the Census Bureau estimates for 1965 and 1975. The formulae were verified by comparing their results with the Census Bureau forecast of the oldest age group in 1970. In both cases the discrepancy from the Census estimate was of the order of 1,000 persons, one-thirtieth of one per cent. These four procedures, adapted to the needs of the four different age levels, enabled us to project the population according to the assumptions made by the Census Bureau until the year 2000. The results are given in tables B.5 to B.10. (See pp. 180-185.) The sole

Appendix B

| 177

advantage of these implicit regression formulae over straightforward application of the cohort survival method is simplicity. The cost in inaccuracy is judged to be minor because of the plausible levels of all the constants and because check-projections to years not employed in calculating the empirical constants led consistently to errors amounting to no more than a fraction of one per cent. The Assumptions Underlying the Projections Population projections involve a good deal of arithmetic, but their essence lies in the assumptions that they have to be made before the arithmetic is begun. These assumptions specify the way in which the population at each date is related to the population at the next date, via the effects of mortality, fecundity, and migration. Since our projections are mechanical extensions of the projections prepared by the Census Bureau, our assumptions are the same as theirs. We shall therefore discuss briefly the assumptions made by the Census Bureau.6 The Census Bureau prepared projections based on four sets of assumptions, designated by the letters A through D. We have extended the series A and D, corresponding to the extreme sets of assumptions. All of the series use the same assumption concerning mortality. The general assumption is that age-specific mortality rates will continue to decline until i960 at the rate observed during the period 1939-1941 to 1947 and after i960 will remain constant. For the extreme upper age groups, however (white persons above 80 and nonwhites above 65), it was assumed that death rates observed during the 1940's would continue to be applicable. Thus these projections presume a very moderate amount of medical progress during the remainder of the century, and essentially no progress in dealing with the diseases of older persons, an assumption at considerable variance with the striking progress made during the first half of the century. Death rates in general were cut practically in half during the period 1900-1950; even for the age group above 85 a decrease of better than 20 per cent was achieved. There does not appear to be any reason to expect this trend to be arrested or even decelerated. As a result it appears that the mortality as5 For further details consult U. S. Bureau of the Census, Current Population Reports, Series P-25, No. 78 (August 2 1 , 1953) and Series P-25, No. 4 3 (August 10,

1950).

178

| Appendix B

sumptions used in making these projections tend toward understatement. In particular, the projections probably understate the number of persons in the upper age groups. Medical progress, however, is such an intangible process that we largely refrain from attempting to quantify the amount of the bias imparted by this pessimistic assumption.8 All the Census Bureau projections use the same set of assumptions concerning net immigration. The assumption is that net immigration will amount to 300,000 per year until July, 1955, to 280,000 per year from then until July i960, and thereafter to 240,000 per year, and that the age-sex distribution of the immigrants will be the same as that observed during 1946-1948. It is unnecessary to labor the point that the future course of immigration is an imponderable and that the assumption used, though unreliable, is as sound as any that can be made. If immigrants have the same mortality and fertility rates as the rest of the population, and if about a quarter of a million immigrants enter each year for the rest of the century, then the immigrants entering after 1950, together with their descendents, will contribute about 19,000,000 people to the population of the year 2000.7 This is about 6 per cent of the total population anticipated at the end of the century and about 10 per cent of the expected half-century increase in population. Thus immigration and fluctuations in it are minor factors in the future course of population and the conjectural nature of the estimate of immigration does not detract substantially from the reliability of the over-all population projections. The difficult and important assumptions are those relating to fertility, and it is with respect to those that the various projections prepared by the Census Bureau differ. Fortunately for our purposes, the future course of fertility will not affect the number of older persons for the rest of the century. But it does affect the proportion that older persons bear to the total population and we must therefore take it into account. The assumptions used in preparing the two series of forecasts that we have extended are as follows: Series A assumes that the age-specific fertility rates will remain at the level established durThis point is discussed further in Appendix B, Part II. Calculated b y assuming that each segment of the population grows at the rate of 1.63 per cent per year. The reproduction rate of recent immigrants is surely somewhat above this because immigrants tend to be recruited from persons in the most fertile years of life. 9 7

Appendix B

|

179

ing 1950-1953. Series D assumes that the fertility rates will decline linearly from the 1950-1953 level until i960, at which date they will reach a level determined from historical data by rather complicated averaging formulae. The fertility rates assumed for i960 are those of table B.2 where also it can be seen that the terminal fertility rates assumed for Series D are close to the average fertility of 1939-1940. The fertility rates are assumed to remain constant after i960. Neither the Census Bureau nor we can make out an adequate case for using just these assumptions. Indeed, we subscribe to the statement made by the Census Bureau forecasters, "Opinions differ considerably as to the future course of fertility in the United States. In the face of a lack of agreement as to the general trend and the absence of definite indications as to what this trend will be, a wide range, reflecting this current uncertainty with regard to future fertility, has been deliberately chosen."8 Empirical data on crude and specific birthrates available in the middle of 1955 show no indication of the fertility declines assumed in Series D. Still, those declines may show up in the next few years. The major reason for expecting age-specific birthrates to decline is that these rates were exceptionally low during the period 19301945 so that women in the second half of their childbearing period entered the postwar era with smaller families than they would have liked. The high birthrates for women aged 25 and over during recent years may therefore be a compensation for the low birthrates of women of all ages previous to 1945 and may fall back when the deficiency in family size has been made up.9 Our own preference is to regard the Series A projections, based on the highest fertility rates used by the Census Bureau, as the ones most justified in the light of current evidence. Current Population Reports, Series P-25, No. 43, p. 5. The foremost proponent of this line of reasoning is P. K. Whelpton, who bases his conclusions on what is perhaps the most exhaustive analysis ever made of fertility trends in this country. See his Cohort Fertility, Native White Women in the United States (Princeton: Princeton University Press, 1954), especially pp. 194-197. On the basis of his study of trends, Whelpton has forecast the number of children born to white women in the five-year period, 1949-1953. His low, medium, and high estimates, in thousands of births, were 14,030, 15,045, and 16,060 respectively. (Op. cit., p. 269.) Data not available to Whelpton at the time now show that his high estimate was almost exactly right and his others correspondingly low. This is just another instance of the persistent tendency of population to outstrip the expectations of forecasters in recent decades, but whether this tendency can be counted on in the future is very difficult to judge. 8

8

V I

H

V i

0 \ w

N



VO

V i sC

V i

r ,

t

fl

Oi

^

n

H

V O C O M

Oi

+

V I

0 \

H

V i

M

>-H

o

0 \ < 7 \ 0 0

V I

t

t n

Ci

Tj-

O

vo

Os

Co

m

v i T j - r t

m

o

o s o i o o s o

vo

V i

r ^

co

^

n

h

rj-

c*

I-I

o

ON

ON o o

ON c i

w

O

t

' t

n

n

t ^

^

^

N

rf-

ro

ci

co

n

OO V i

CO \ o

O r »

0

t-

ot

»-


i o

v ,

rj-

n

^

co

ON

On

co

t}-

t

" t

n

d

co

^ co

W O

t^ Jh "

« W

£

c^

>->

vo

co

•-< s o

C*

I-H

O

vo

Oi

hh

O

O

VO

\ o

ON

vo

so

oo

oo

ON 0 0

NO

OO

so

OO

O i C 7 i O O \ o

v>

^

so

V i

V I

^

^ ^

-vi"

t ^

Cv

co 00

co oo

«

on

00

O

o

t o t o

v n v i v i v i v i ' v i - ^ - c o c *

«

4)

S

Ov

VO

OV

oo

O v M

OQ

f ^ T t - 0 0

ve

V i

V i

o

V i

ci

V I V O

V i

V i

V i V O r i

ÍV

M



çl

ci

O

-a

«

E



^

O

P I¿> T «!R> M

M

«

¿

• O {•)

Os ^

FO

O ^

ON ^

ON

r r r

.

! ! ^ o * îr

nu

VO O »O _ "^•VRJWISOVÔ

—R

o

H

"

o H

V

os o© e* n Os OS Os Os OS Os 00

^ v » v o t-»

h e¿ W ta ei a * o J

«

i i0n o\ H 0

d

« m o\ t »o o\ h t ^ Os 00 vr, ^ CO ci -í-

co^nviroosrtoo r>.NO*OOVO ^ ^ ^

O Os fO fO

h

N ^ es n o n I s - ^ Ca ^ r> Os O* 0\oo f - OO Os OO so «O 4 4 n « «O

d ^

r » ci a co O ^ oo n vo o » r». r-. oo ^ 0\ ONOO r - s . » OS 00 SO l o v> TJco CO

co ci c* n 0\ ^ oo oo c* ' í ' c s ^ so Os os oo r - r ^ o s o o v o vo wi »o t n ^

d co

00 CS CO OS OO OO ^ ^ - s o VO Os co ^ OO 00 f-» Os 00 so w->vr>vr>vr>^t"^-COC
l o s o i o v i m 4 n n ci m c*

cooo f-s. rf oo O t ^ s o •-« v i m \o o ^ \û oo 00 VO VI V) I^SO l o l o V» Tf -^-COCOCi >H IM

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : ;

^

:

!

:

:

!

:

:

'.

:

!

:

!

:

!

:

I

: rt

!

= i , ¿ Jo ¿ Jn i> Jn ¿ ¿ ¿ O H H t i d n o f f w

:

:

I

tr*»

¿ü >

:

:

:

£ C

"

!

! ^

^

i, ¿ i, Ì £ v u e vo

O H

: > •. -ao

g bO «

§ u-t vo c

O H

£ Ph

t H tì t O H ri M « t Cl O ^ en ^ a o>oo O ON

ci a *

o _1

cl ¿ Uì Ol w C5 ä e - oí>. Ifí M > S ° CO g D O M ¡3 H~ lis Ol X H h O Z O P O OÍ

M M VO

d ro a» h ci a» CI vo cncjsoo vo ON oo oo Ñ o\co ^ vo 4 ih ih SO

o o\

en ko n >o H cot-t m ri co oo o ci 00 C7\ ON OO OO t^ t^. ON OO *0 In VI ivi



oo co oo ^ w oo n Noo \o n »o^o os o^o oo oo r-* r*oo wi vo

4

O OO Oo o í^cí oo Ï^Tf-OO OOOO N f^ ffl » vo

Ù

co f*» 'O O co

if O

rt "g

rj- OO O ^«O M vû OO fs «o H vo •-< c ^ ro vn O ^ O »o

fN VÌ 0\ « f^^^O r^O Vi Vi ^ co n t

vo o --tf- r-* r*» ^n o coowo on N VD f-» ON OO V£> WiVÛ VÛ «O lo "i" Tf CO C* O CO c* O t^O vo f-« ON OO vO

O co Os r-» O «o oo i-t « vi vi in ^ c^ fi co

"«f VO Vi ^ 00 00 vû V) VI

n ON fs H vr>0 ^ oo SO Vo vr> «-O ^ CO CS CS c*

o 00 VO «O «o

C< NO O vo oo ^lolo^-^cocort M




a> > o "fl C rt ^

— 2 f £

-

>

• O ; 4¿> Ws m rtt>0 so —

es -3" 2 (2

¿r

g H u5

^ ci vû oo »H c^ r-» t os o >-> ON ©s 00 00 OnOOVÌ) "t OO SO c* N N t ^o oo o « oo oo oo o» oo ^

ta &

o •J

ao . c* o on oo t-«. o

cooo O O oo is n «*joo oo oo O O Os r-» oo on o© so -^-w-ìoo h n o os oMs O O O O t oo N t ^ h covo n ^oo o oo os co ^o -t «^oo h- n h o O o oo so on M o O f») NVO Vi\o n N N H N n cf oo oo so ^ »ooo í^.ro'-' O »-1 os oo so o©

t oo O

C* O n Vi Vi t OO «n fN TJ- \O VO o r^^O -«hW-lOO N f i H O M t-H o Ol h >o fs

5 < o« X e M -< c^ m o O f-so t so

M

OO

O

t t

t*>» t C* t CO ON OO -i- SD woo i- n ON Tj- ^ t e e h r^N o >-< m N «o ^ vi

¡Z

Ö u S S-

v>MCoei0©r>»tsoc»M00ONt--.co'-'SO h o oo r-*. \o t t f^ f^ H r->. so co so n n H n fo H o t ON so c-5»-i o >-« i-H HH o os oo í— so vi n n

t t

O Oh O H u s H h O Z O o

oí CU

> o T3 C rt

> O c • Oí I Vi

t H o1 M

o» flt M 1 ^Î O M C*

^ u Tn h oun Tvo h o

r^

un ¿ »U ¿ vo (!) Jí VU wo uovo vß O n n 1 -

„» P-,

Appendix B: Part II

Effect of Error in Forecast Mortality Rates

In chapter ii and in the first part of this appendix, we gave reasons for believing that the mortality assumptions used in making our population forecasts do not allow sufficiently for the beneficial effects of progress in medicine and public health during the remainder of the century. Since mortality takes its toll primarily from the older members of the population, the effect of understating progress (or overstating mortality rates) is to reduce the estimated proportion of the population in the older age groups. Now, the main conclusion drawn from our fifty-year population projections was that the proportion of the population aged sixty-five and over is not likely to rise significantly in the next fifty years. The fact that the mortality rates underlying those projections are biased casts doubt on that conclusion. The task of the present appendix is to evaluate the seriousness (but not the extent) of the bias. For two reasons, primarily, we cannot give a very precise answer to the question, "How large an error have we incurred by underestimating future declines in mortality rates?" First, the question is vague and cannot be made specific without specifying both the likely over-all amount of improvement in mortality rates and also the age distribution of this improvement. Second, mortality rates affect population forecasts in a very complicated way mathematically. The relationship is so complicated that it is not practicable to present formulas showing population as an explicit function of assumed mortality rates. Instead, the usual method for showing 186

Appendix B, Part II

|

187

the consequences of different mortality conditions is to calculate the entire projection for each variant mortality assumption and to present comparative forecasts. Obviously this is a laborious and expensive task, too laborious to be justifiable when we are as much in the dark as we are about the possible future pattern of mortality improvement. For us to go through such a procedure would be as wasteful as computing road mileages to the nearest inch. In these circumstances we have adopted a very crude method of estimation to gain an appreciation of the order of magnitude of the error resulting from ultraconservative expectations about the future course of mortality. Theoretical Considerations We first formulate the mathematical relationship between agespecific mortality rates and population levels. Throughout the theoretical discussion we shall ignore (a) immigration and (b) the effect of changes in mortality rates on the number of births. Factor a is a minor influence on the trend of the American population. The neglect of factor b will lead us to somewhat overestimate the proportion of persons to be found in the upper age groups. Let denote the force of mortality at age x, i.e., the proportion of the population aged x that may be expected to die per unit of time. Mathematically stated, this definition is lix =

1 \x hm -



Az-»0

_

x

^x

ld_7 _ i 7 554

Unrelated males NR O P,W,W/0 All

998 901 804 914

Unrelated females NR 0 P,W,W/0 All

948 866 770 851

UA-2

1,723

1,516 1,414 1,628

973

878 784 891

923

846

749

828

UA-3

1,657 1 ,457 1,358

I,56o

937

846

754 857

889 813

720 798

All urban

1

,73J 1,524

1,421 1,636

Rural nonfarm High

Low

High

Low

1,258 1,104 1,032 I,l88

1,005 884 825 950

745

559 492

7°3

458 528

572

448

336

515 458

380

285

34° 399

300

980 885 790 899

714 642

931 851 754

678 618

836

Rural farm

574

654

547

608

523

656 612

542

426

494 437

367 325 37°

484

255

3'9 274 243 277

2o6 TABLE

I

Appendix

C

C.4

" S U B S I S T E N C E " BUDGETS (1951

Dollars) Residence

Family statu8 TJA-l

UA-2

UA-3

All urban

Rural nonfarm

Rural farm

High

Low

High

Low

881

704 619

522

391

4 59

Couples NR 0 p , w , w / o All

1,088 1,015 1,169

1,206 1,061

990

1,160 1,020 951

1,212 1,067 995

1,140

1,092

681 615 550 624

656 592 28 600

686 620

646 592 524 580

622 569 504

652 596 528

1,145

773

722 832

578

665

3 44 321

428 492

370

3 H

235

298 257 228 259

223 192 170 194

Unrelated males NR O P , W , W / 0 All

699 631 563 640

553

629

500 449

402 458

400 360 321 366

475

379

433

346

266 238 279

200 178 210

0

664 606

p , w , w / o

539

NR

All

596

559

00

Unrelated females

383

426

306 339

Appendix D

Comparison of the Follow-Up Survey of the Aged with the Survey of National Old Age and Survivors' Insurance Beneficiaries

This appendix is concerned with a limited comparison of the results of the Follow-Up Survey with those of the Survey of National OldAge and Survivors' Insurance Beneficiaries (hereafter designated as the "Beneficiary Survey") which was conducted in 1 9 5 1 by the Social Security Administration.1 Owing in part to consultations during the planning stages, and in part to the coordinating efforts of the Bureau of the Budget, the surveys were designed to permit maximum comparability of the results. There are, however, fundamental differences in the populations sampled, in the identity of the enumerators, and in the size of the samples of the two surveys. These factors are relevant in appraising the data of the two surveys and in determining the validity of the use of certain data contained in the Beneficiary study in conjunction with the data collected with the Follow-Up Survey. The Nature of the Samples The Follow-Up Survey attempted to sample the entire noninstitutional population 65 years old and over. It used the sample of households selected and used by the Bureau of the Census for the income subsample of the Current Population Survey in April, 1952, and consisted in an enumeration (by Census enumerators) of all 1 We shall not distinguish between the Follow-Up Survey and the Current Population Survey since our data are from both. See Appendix A.

207

208

| Appendix D

persons aged 65 or over included in the CPS income subsample.2 Approximately 3,600 aged individuals were thus included. The Beneficiary Survey was a one per cent random sample of the old-age and aged-widow insurance beneficiaries on the rolls in December, 1950. About 18,000 beneficiaries were interviewed by the field staff of the BOASI during the period November, 1951, through February, 1952. The sample included beneficiaries with benefit suspensions as well as those without and the data are tabulated for each group separately as well as for both combined." From these brief descriptions certain important differences between the populations are immediately evident. The Beneficiary study covered people on the beneficiary rolls in 1950 only and thus excluded: (1) aged persons who had never worked, (2) aged persons who have worked only in noncovered occupations (including, most importantly for this age group, the farm population, but also a variety of other groups), (3) aged persons under 75 who had not ceased working by 1950 and who had sufficient earnings or selfemployment income so that they did not qualify for benefits under the law.4 These considerations suggest that the population of OAS I beneficiaries will be more centrally located than the total aged population with respect to economic status: it will not contain many aged groups with very small incomes (mere qualification for old-age and survivors' insurance guarantees this); at the same time it will not contain many of the groups with the largest incomes because such incomes frequently occur in the noncovered occupations (e.g., in the professions of law and medicine) and also because of the importance, in the highest income classes, of those with large amounts 2 The regular CPS sample contains approximately 25,000 households; three-fifths of the CPS sample is designated as the income subsample. 3 Benefits could have been suspended in the survey year because of receipt by a beneficiary under 75 years of age of covered wages of more than $50 in a calendar month, receipt of net earnings in covered self-employment of more than $600 in a calendar year, or as a penalty for a violation of certain provisions of the Social Security Act. After September, 1952, the amount beneficiaries could earn in covered employment was increased by 50 per cent, and in 1954 was further increased to twice the 1950 limits. Also in 1954 the age limitation for unlimited earnings was lowered from 75 to 7 2 years of age. 1 The exclusions listed do not apply to aged persons whose spouse (or deceased spouse) qualified them for benefits. No doubt there are other biases involved in the use of the 1950 eligibility with respect to: (1) the exclusion of people who reached age 65 after 1950 but before the survey date and (2) the selection involved in the fact of death between 1950 and the survey date. Relative to the major factors listed above, these sources of bias seem negligible.

Appendix D

| 209

of earnings (see chapter viii, above) who were, because of large earnings, not qualified for benefits. The underrepresentation of the higher income groups in the Beneficiary Survey is somewhat, but not entirely, mitigated by use of the combined group with and without benefit suspensions, since those with suspensions typically have substantially higher incomes than those without.5 TABLE

D.L

U N I T S OF A N A L Y S I S IN T W O S U R V E Y S Beneficiary groups (Beneficiary survey)

Economic units (Follow-Up)

Classification» Number of sample cases

Per cent

Per cent

A. Unrelated females 1. Women beneficiaries—not married 1. Aged widows

4,788 2.235 2.553

27. i

43-1

B. Unrelated males 3. Male beneficiares—not married....

4.769

27.0

18.5

45-9

38-4

C. Couples 4. Male beneficiaries—married, wife entitled J . Male beneficiaries—married, wife not entitled 6. Women beneficiaries—married.... Total

4.7^9 8,104 4>5 r 3 3.082 509 17,661

» Lettered classification from Follow-Up Survey, numbered classifications from Beneficiary Survey.

This appendix examines the extent to which these features of the Beneficiary Survey, which lead it to give biased estimates for the total noninstitutional aged population, can be measured and the extent to which they affect the use of the data from the Beneficiary Survey in conjunction with data from the Follow-Up Survey. Units of Analysis The basic units of the Follow-Up Survey are aged economic units— individuals or couples. The units of the Beneficiary Survey are "beneficiary groups" which can be combined into categories closely similar to those of the Follow-Up Survey, as indicated in table D.i. 5 Compare tables B200 and C200 of the Beneficiary Survey. All tables of this Survey referred to are in "Selected Findings of the National Survey of OASI Beneficiaries, 1 9 5 1 " (June, 1953) and "More Selected Findings . . . " (January, 1954). U. S. Department of Health, Education, and Welfare, Social Security Administration.

2io

| Appendix D

For comparative purposes the percentage distribution classified by type of economic unit is shown for each survey. It is evident, and not surprising in view of the work-histories for qualification as a beneficiary, that the marital-sex status of the two populations differs markedly. This fact requires separate treatment of the marital-sex groups, and suggests further that the impact of the biases may differ sharply for males and for females.

Total Money Income "Total money income" in each survey relates to the income of the individual or the couple as appropriate. The components of income for each are sufficiently similar to the Census definitions of income that they are comparable without risk of error. Different class intervals used in the tabulation of the data make direct tabular comparison of the income distributions inconvenient, and we, therefore, present in figures D.i, D.2, and D.3 graphic comparisons of the income distributions. The upper portion of each chart compares frequency polygons of the income distributions. Income is plotted along the x axis and relative frequency density per $500 of income is plotted along the y axis.8 It is evident that there are persistent and systematic differences between the incomes of the two populations as reflected by the two surveys. The chief differences are graphically evident, and are also suggested in the tabular summary of frequencies in selected intervals of income in table D.2. The over-all pattern of difference is the same in all marital-sex groups and may be summarized as follows: (1) the aged population as a whole (as reflected in the Follow-Up Survey) has a much greater frequency of very-low-income receivers than does the population of aged on the OAS I rolls; (2) at moderate income levels—from $500 to $3,000—the reverse is true; frequency densities are greater for the beneficiaries than for the population as a whole; (3) at "high" incomes there is a return to the original relationship but one that is less clearly visible because of the smallness of the frequency densities at these high levels. These data confirm the a priori expectations: the Beneficiary 6 The data from the Follow-Up Survey are found in Appendix E , table 2 0 1 ; the data from the Beneficiary Survey is a consolidation of the data in Beneficiary Table C200. As plotted, the distributions are truncated at $5,000 for Beneficiary Survey and $6,000 for Follow-Up Survey. The excluded frequencies in no case exceed 5 per cent.

INCOME Fig. D.i. Comparison of the size distribution of income in two surveys of the aged, couples, 1951.

INCOME

INCOME Fig. D.2. Comparison of the size distribution of income in two surveys of the aged unrelated males, 1 9 5 1 .

PER CENT

INCOME

INCOME Fig. D.3. Comparison of the size distribution of income in two surveys of the aged, unrelated females, 1 9 5 1 .

214

| Appendix D

distribution is less dispersed with smaller frequencies at the extremes and greater frequencies over the central range. Since the systematic differences listed above are cumulatively offsetting, it is possible to find a single point in each distribution with cumulative frequency through that point that is the same for both surveys. The systematic variation of the frequency polygons is reflected by the ogives in the lower half of figures D.i, D.2, and D.3 and suggests the following rough points in common for the TABLE

D.2

INCOME DISTRIBUTIONS IN T W O SURVEYS: SELECTED CLASS INTERVALS (Percentages) CoupleB

Income

Beneficiary Under $500»

Unrelated males

Follow-Up

Beneficiary

4.8

18.7

500"-1,499

43-4

34-8

243 56.2

1,500-2,999

36.9

24.5

14.8

3,000-4,999

" • 3

14.8

3-6

7.2

3-7 1.1

100.0

100.0

$1,511

?I,387

5 , 0 0 0 and over Total Median Income B b

Unrelated females

Follow-Up

Beneficiary

Follow-Up

40.4

42.6

62.1

40-5

46.7

29.8

-3

8.9

5-6

4-7

1.2

3-»

0.6

i-7 0.8

100.1

100.1

IOO.O

100.0

$848

$662

$693

?4°3b

1 1

Class limit is ¿600 for Beneficiary Survey data presented. Combining all income receivers below i s00 into a single class.

two surveys. For couples, about 60 per cent of the frequency is below $1,900. For unrelated males, about 80 per cent of the frequency is below $1,500. For unrelated females, about gs per cent of the frequency is below $2,300. For dollar values much below those listed, the Beneficiary data will understate the relevant cumulative frequency (the size of the error will be measured by the vertical distance between the ogives) having income below that level. Above those listed, the reverse bias will occur. At or near these values the error will be small. The substantial range over which the pairs of ogives are reasonably close together, taken along with the systematic nature of the discrepancy between them (and its conformance to the a priori expectations), suggest that some interchange of data is feasible. The vertical distance between the ogives is used in the text to adjust given percentages for liquid assets found in the Beneficiary Survey for the population as a whole.

Appendix D

| 215

Estimation of Liquid Asset Holdings Liquid asset holdings were reported in detail in the Beneficiary Survey but were not included in the Follow-Up Survey. In view of the differences in the nature of the populations sampled, it would be inappropriate to use the data from the Beneficiary Study without considering the direction and the size of errors likely to be introduced. Our procedure is to adjust the asset data in the TABLE D.3 RELATIONSHIP B E T W E E N INCOME AND LIQUID A S S E T HOLDINGS OF THE A G E D , 1 9 4 6

(Percentages) Income class L i q u i d asset a m o u n t U n d e r $1000

Zero

48

?1-499

28

5,000

and over

Total Median amount of liquid assets

$3000 and over

25

4

9

16

AH income classes

3° r 3 r 3

26

9 1

28

33 23 23 12

26

9

100

100

IOO

IOO

$36

$800

500-1,999 2,000-4,999

$iooo-$2999

!

$2,428

$37°

SOURCE: J a n e t Fisher, " T h e Economics of an A g i n g P o p u l a t i o n , " tables 14, 16 (data from 1947 S u r v e y of C o n s u m e r Finances).

Beneficiary Survey upon the assumption that the nature and direction of the biases found for incomes (see above) apply in like manner to liquid assets. In text table 9.5 (p. 133) the data, both adjusted and unadjusted, are presented. We believe the adjusted data are substantially closer than the unadjusted to the (unknown) correct distribution of liquid asset holdings for the population as a whole in 1951. The reasons for this will be presented shortly; it may be noted that except for unrelated females the differences are comparatively minor, and no important conclusion is altered if the unadjusted data are used. For unrelated females, where the differences in magnitude are very large, the case for using the adjusted data is far the clearest because of the greater impact of the biases on the females included in the Beneficiary study. The chief difference between the Follow-Up and Beneficiary surveys' results on income was the smaller dispersion of the income

216

| Appendix D

distribution of the Beneficiary group. The chief consequence of that difference was that the size of the groups with very small incomes would be underestimated if reliance was placed upon the Beneficiary Survey. Three considerations support the contention that a similar bias is present in the liquid asset data of the Beneficiary study if it is to be applied to the total aged population. 1. Studies of the relationship between income and liquid asset holdings for the group over 65 indicate a positive and reasonably close association.7 (Table D.3 presents part of the available data TABLE

D.4

COMPARISON OF N E T WORTH (BENEFICIARY SURVEY) W I T H N E T V A L U E OF TOTAL ASSETS (FOLLOW-UP S U R V E Y ) , 1 9 5 1

(Percentages) Unrelated males

Couples

Unrelated femaleB

Asset holdings* Beneficiary

None (or negative)

Follow-Up

I6. J

12.4

S-4 3-7

Beneficiary

Follow-Up

Beneficiary

41.6

33-5

29.8

2J.9

6.4

II.4

14.9

5-5

9-9

5-3

9.0

9.0 41.2

II. I

3-3

42.4

41.9

100.0

100.0

100.0

5

i,000-2,999

10.8

5-3 3-3 7-4

3 , 0 0 0 or more

63.6

71.6

IO.3 30.I

100.0

100.0

IOO.O

ÎI-499 Soo-999

Total

12.

Follow-Up

* Attention is directed to the differences in the concepts of assets used in the two surveys. These differences are discussed in the accompanying text.

and is intended to be suggestive of the findings rather than a definitive reproduction of them.) This relationship is not surprising. It is, however, helpful; it suggests that since the two samples differ significantly with respect to income (differences we have measured) they will differ similarly with respect to liquid asset holdings (differences we can not measure directly). W e feel confident, that, on this basis alone, we can predict the direction of adjustment and have some notion of the size of that adjustment. Our assumption that the differences in the percentages measured by the two surveys will be equal for incomes and for liquid assets can, of course, err in either direction. 2. The only possible direct comparison of asset data from the two surveys is that presented in table D.4. This comparison is less * See for example, Janet Fisher, "The Economics of an Aging Population," Unpublished Ph.D. Thesis, Columbia University, 1950. (Data based upon Survey of Consumer Finances.)

Appendix D

| 217

satisfactory than the one presented for income since it reflects, in addition to differences in populations sampled, differences in the definition of the variables as well. The Beneficiary Survey tabulation concerns net worth: the net value of total assets, less both secured and unsecured borrowings and overdue bills. The FollowUp Survey concerns the net value of total assets alone. Importantly, the Follow-Up Survey does not deduct liabilities except those applicable to particular included assets. In addition, also importantly, the Follow-Up Survey includes, but the Beneficiary Survey excludes, life insurance in the value of assets.8 Both of these differences operate in the same direction: the Follow-Up Survey concept of total assets adds more and subtracts less than the Beneficiary Survey concept of net worth. Hence identical groups surveyed would show smaller frequencies at the lower asset levels in the Follow-Up Survey. These considerations, when applied to the data of table D.4, suggest that for unrelated females the approximately equal total assets position reflected in these data conceals a real advantage in the asset position of the Beneficiary group. For couples, the apparent slightly larger percentage (22 per cent compared to 18 per cent) of the Beneficiary study having assets in amounts under $500 would surely be reversed had identical definitions of total assets been considered. For unrelated males the apparent differences are sufficiently large that a priori it is not obvious whether a comparison of assets for the two populations using the identical concepts would reverse the direction of difference shown. Our conclusion about total assets," then, is that for females and couples certainly, and for unrelated males possibly, the sample of beneficiaries does not overstate and very possibly understates the percentage of units with small amounts of assets, if it is taken to refer to the population as a whole rather than to the group of beneficiaries. In other words, the sample of beneficiaries has no larger percentage of cases with small assets than the Follow-Up Survey. The significance of this fact with respect to liquid assets is due to the differing composition of total assets. To this we turn. 3. Total assets will not be a reliable relative guide to liquid 8 The value of life insurance used in the Follow-Up Survey is the face value of life insurance. ° While neither survey measures "total assets," the measures just described are so designated to distinguish them from "liquid assets."

218

| Appendix D

assets if holdings of nonliquid assets differ in the two surveys. The most significant form of nonliquid asset holdings consists of owned homes and, as table D.5 indicates, the incidence of home ownership is greater for the Follow-Up Survey sample than for the Beneficiary sample. In addition to having a greater tendency to own their own homes, those in the Follow-Up Survey showed a slight additional tendency to own them free and clear.10 TABLE

D.5

HOME OWNERSHIP: BENEFICIARY AND FOLLOW-UP SURVEYS (Percentages) Couples

Unrelated males

Unrelated females

Category Beneficiary

Follow-Up

Beneficiary

Follow-Up

Beneficiary

Follow-Up

Home owned

64.7

36.1

75-3

6

3-9

35° 65.0

38.6

35-3

73-7 26.3

24.7

Home not owned

100.0

100.0

100.0

IOO.0

100.0

100.0

81.7

«5-3

84.2

85.9

82.6

90.0

Total

61.4

Per cent of those with home owned who own it free and clear..

SOURCE: Beneficiary consolidated table A-301; Follow-Up Survey.

Since home ownership virtually assures total assets of more than $500, the difference in the size of the groups without owned homes reflects a difference in the size of the groups likely to be found with very small total assets. This consideration operates in the same direction as the differences in the definition of total assets in a comparison of the two surveys. Low values of liquid assets may be inferred, like income, to be more frequent in the aged population as a whole than in the population of aged OASI beneficiaries. Taken together, these three factors indicate the direction of adjustment of the Beneficiary data that is appropriate. Whether the amount of adjustment is correct is of course conjectural. W e believe it to be reasonable, but, as suggested earlier, the adjustment is not essential to the analysis of chapter ix. The actual technique of adjustment was to apply to the cumula10 These facts are almost certainly explained b y the fact that the Beneficiary Survey, b y nature of its selection, includes a much smaller proportion of the rural population, whose members have an appreciably greater tendency to own their own homes.

Appendix D

| 219

tive frequency distribution of liquid assets from the Beneficiary Study the percentage adjustments estimated from the relationship of income distributions in the two surveys. This adjustment was, of course, made separately for each type of economic unit and has the sharpest effect on the unrelated females. This, as we have seen, is appropriate in view of the significant differences in the size of the group of females included in the two surveys, and the biased selection (for our pin-poses) of the females included in the Beneficiary Survey.

Appendix E

Basic Data of the Follow-Up Survey

The body of data relating to the characteristics and financial status of aged economic units, heavily utilized in the present study, have aroused substantial interest and have already been borrowed by others working on problems of the aged. Among other users have been the Bureau of Old Age and Survivors' Insurance, The Bureau of the Census, and the National Industrial Conference Board. The present appendix consists of forty tables which contain a large part of the data collected and tabulated in the Follow-Up Survey. These tables are designed to supplement the text tables, which are analytic in character and summary in nature and therefore less than fully useful for purposes other investigators may choose to pursue. Where material is adequately covered in text tables it is not repeated in the appendix tables, nor do these tables include the data on aged individuals (as distinguished from aged economic units). The data included in these tables, while voluminous, are by no means comprehensive. Any selection of data will disappoint some. The complete tabulations remain on file at the Institute of Industrial Relations, University of California, Berkeley, California. Users of the data are referred to Appendix A for a discussion of the source and reliability of the data, and to the appropriate sections of the text for definitions of the concepts employed in the tabulation. Explanatory notes, both general and specific, are given below. The tables are gathered in four series, as follows: 220

Appendix E

| 221

Series 100 tables (Tables 101-115) These tables present the percentage distribution of aged economic units with respect to a series of sociological and economic variables, by type of economic unit. Series 200 tables (Tables 201-216) These tables cross-classify the percentage distributions of the 100 Series by the level of total unit money income. Series 300 tables (Tables 301-305) These tables are similar in scope to the Series 200 tables but use total unit receipts (instead of income) as the cross-classifying variable. Selected classifications only are presented because of the broad similarity of income and receipts. Series 400 tables (Tables 401-404) These tables cross-classify selected distributions of the population of aged economic units by the nature of the living arrangements of the units. List of Tables SERIES 1 0 0

101: Place of residence: distribution of aged economic units by type of unit, urban by type of unit, urban (by size of place) and rural: United States, 1951 102: Living arrangement: distribution of aged economic units by type of unit and living arrangement (family status): United States, 1951 103: Nativity: distribution of aged economic units by type of unit, native born, not native born (by years of residence): United States, 1951 104: Longest occupation: distribution of aged economic units by type of economic unit, longest occupational group: United States, 1951 105: Housing cost: distribution of aged economic units by type of unit, home ownership (by payment status) and home rental (by monthly rent): United States, 1951 106: Medical cost: distribution of aged economic units by type of unit, receipt of medical service, and size of medical payments: United States, 1951 107: Net value of total assets: distribution of aged economic units by type of unit, and amount of asset holdings: United States, 1951 108: Earnings: distribution of aged economic units by type of unit, and amount of earnings: United States, 1951

222

| Appendix E

109: Asset income: distribution of aged economic units by type of unit, and amount of asset income: United States, 1951 110: Dissaving: distribution of aged economic units by type of unit, and amount of dissaving: United States, 1951 111: Pensions: distribution of aged economic units by type of unit, and amount of pension: United States, 1951 112: Assistance: distribution of aged economic units by type of unit, and amount of assistance: United States, 1951 113: Contributions: distribution of aged economic units by type of unit, and amount of contribution: United States, 1951 114: Family income: distribution of aged economic units by type of unit, and amount of family income: United States, 1951 115: Source of receipts: distribution of aged economic units by type of unit, and number and kind of source of receipts: United States, 1951 SERIES 2 0 0

201: Place of residence: distribution of aged economic units by type of unit, and total unit money income, urban (by size of place) and rural: United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 202: Living arrangement: distribution of aged economic units by type of unit, and total unit money income, living arrangement (family status): United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 203: Nativity: distribution of aged economic units by type of unit, and total unit money income, native born, not native born (by years of residence): United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 204: Longest occupation: distribution of aged economic units by type of unit, and total unit money income, longest occupational group: United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 205: Housing cost: distribution of aged economic units by type of unit, and total unit money income, home ownership (by payment status), home rental (by monthly rent): United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages

Appendix E

|

223

206: Medical payments: distribution of aged economic units by type of unit and total unit money income, receipt of medical service and size of payments: United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 207: Net value of total assets: distribution of aged economic units by type of unit, by total unit money income and amount of asset holdings: United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 208: Earnings: relation between earnings and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by size of earned income Part II: Distribution of aged economic units by size of earned income for given levels of total income 209: Asset income: relation between asset income and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by size of asset income Part II: Distribution of aged economic units by size of asset income for given levels of total income 210: Dissaving: relation between dissaving and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by amount of dissaving Part II: Distribution of aged economic units by amount of dissaving for given levels of dissaving 211: Pensions: relation between pensions and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by amount of pension Part II: Distribution of aged economic units by amount of pension for given levels of pension 212: Assistance: relation between assistance and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by amount of assistance Part II: Distribution of aged economic units by amount of assistance for given levels of assistance 213: Face value of life insurance: relation between face value of life insurance and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by amount of face value of life insurance Part II: Distribution of aged economic units by amount of face value of life insurance for given levels of face value of life insurance

224

| Appendix E

214: Family income: relation between family income and total unit income for aged economic units, by type of unit: United States, 1951 Part I: Median total unit income by amount of family income Part II: Distribution of aged economic units by amount of family income for given levels of family income Part III: Distribution of all aged economic units by amount of family income and total unit income 215: Source of receipts: distribution of aged economic units by type of unit, by source of receipts and by total unit income: United States, 1951 (selected sources) Part I: Median total unit income by source of receipts Part II: Horizontal percentages 216: Source of receipts: mean total unit income by type of unit, by selected sources of receipts: United States, 1951 SERIES 3 0 0

301: Place of residence: distribution of aged economic units by type of unit, and total unit receipts, urban (by size of place) and rural: United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 302: Living arrangement: distribution of aged economic units by type of unit, and total unit receipts, living arrangement (family status): United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 303: Housing cost: distribution of aged economic units by type of unit, and total unit receipts, home ownership (by payment status), home rental (by monthly rent): United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 304: Net value of total assets: distribution of aged economic units by type of unit, and total unit receipts, amount of asset holdings: United States, 1951 Part I: Vertical percentages Part II: Horizontal percentages 305: Source of receipts: mean total unit receipts by type of unit, by selected sources of receipts: United States, 1951 SERIES 4 0 0

401: Residence: distribution of aged economic units by type of unit, by place of residence (urban, rural nonfarm, rural farm) and living arrangement: United States, 1951

Appendix E

| 225

402: Living arrangement: distribution of aged economic units, by type of unit, by total unit income, place of residence and living arrangement: United States, 1951 403: Source of receipts: distribution of aged economic units by type of unit, living arrangement and selected sources of receipts: United States, 1951. 404: Family income: distribution of aged economic units by type of unit, living arrangement and amount of family income: United States, 1951

Notes GENERAL NOTES

(These apply to all appendix tables, and are not separately noted on individual tables.)

A. Percentage distributions do not always add to totals because of rounding. Errors up to 0.2 per cent (e.g. 99.8) are frequent, and discrepancies of 0.3 and 0.4 per cent, while infrequent, do occur. B. Because of partial nonreporting there are, in effect, several different samples (subsamples): (1) All those reporting on a tabulated characteristic; (2) those reporting on a tabulated characteristic and also reporting amount of income (if any); (3) those reporting on a tabulated characteristic, amount of income, and also amount of nonincome sources of receipts (if any). Tables of the 100 series generally present all three; other appendix tables (and also the text tables) present the distribution which is most nearly appropriate. See Appendix A for a discussion of the importance of nonreporting. C. All tables are designated as referring to 1951. The survey was made in April, 1952, and data on sample characteristics (family status, occupation, residence, nativity, etc.) refer to the date of the survey; income and receipts data (and their components) are for the calendar year 1951. D. Unless otherwise indicated monetary tabulations for couples (e.g., income) refer to the combined amounts for both members of the couple. E. Inflated figures may imply a different percentage distribution than is reflected in the sample or subsample totals. See Appendix A. F. The computations of all medians and means reflect particular assumptions made with respect to the frequency distributions. See Appendix A. In particular, any median under $500 reflects a procedure that differs from usual Census procedures. FOOTNOTES

Distribution of those reporting income. of those reporting receipts. ° Number of sample cases between 50 and 99, subject to large standard errors. a

b Distribution

226

| Appendix E

" Number of sample cases below 50, and therefore too small for reliable presentation of percentage distributions. " Items do not add to total because of nonreporting. ' Classification is that of male member of the couple. B Class intervals beyond this point omitted because of small sample size. " A source of receipts is "major" if received during year in an amount of $200 or more. ' Distributions presented only for selected categories.

PLACE O F RESIDENCE: DISTRIBUTION O F AGED ECONOMIC UNITS BY T Y P E O F UNIT. URBAN (BY SIZE O F PLACE) AND RURAL: UNITED STATES, 1951 Sample p e r c e n t a g e s Type of unit and place of r e s i d e n c e

Number of units (in thousands)

P e r cent of total sample

P e r cent of s a m p l e reporting income

P e r cent of s a m p l e reporting receipts

100 0

100 0

Total

9, 803

100.0

Total u r b a n Total u r b a n i z e d a r e a 1 , 0 0 0 , 0 0 0 o r m o r e persona 250,000 to 999,999 p e r s o n s Under 250,000 p e r s o n s U r b a n p l a c e s , not in u r b a n i z e d a r e a Total r u r a l Rural nonfarm Rural f a r m

6. 155

64. 1 41.6 21.4 12.8 7.4 22.5 35.9 21. 3 14.6

3, 648 2, 123 1, 525

63 40 21 12 7 22 37 21 15

1 8 1 5 2 3 0 9 1

62 40 20 12 7 22 37 21 15

6 4 8 5 1 2 3 9 4

Couples Total

3, 763

100.0

Total u r b a n Total u r b a n i z e d a r e a 1,000,000 o r m o r e p e r s o n s 250,000 to 999,999 p e r s o n s Under 250,000 p e r s o n s U r b a n p l a c e s , not in u r b a n i z e d a r e a Total r u r a l Rural nonfarm Rural f a r m

2, 174

59.0 35.8 17.4 11.7 6.6 23. 2 41.0 22.3 18.7

1, 589 838 751

100 0

100.0

3 8 2 0 6 4 7 3 4

57. 7 34.7 17. 1 10.9 6.6 23. 1 42.3 22. S 19.8

100 0

100 0

58 34 17 11 6 23 41 22 19

Unrelated males Total

1, 810

Total u r b a n Total u r b a n i z e d a r e a 1,000,000 or m o r e persons 250,000 to 999,999 p e r s o n s Under 250,000 p e r s o n s U r b a n p l a c e s , not in u r b a n i z e d a r e a Total r u r a l Rural nonfarm Rural f a r m

1,033

777 457 320

100 0 60 40 22 11

0 5 3 4

58 39 21 11

1 0 2 1

58 38 21 11

2 9 1 3

19 40 23 16

5 0 6 4

19 42 24 17

1 0 8 2

19 41 24 17

3 8 6 2

Unrelated f e m a l e s Total

4, 230

100.0

100.0

100.0

Total u r b a n Total u r b a n i z e d a r e a 1,000,000 or more persons 250,000 to 999,999 p e r s o n s Under 250,000 p e r s o n s U r b a n p l a c e s , not in urbanized a r e a Total r u r a l Rural nonfarm Rural farm

2, 948

70.4 47.2 24.6 14. 3 8.3 23.2 29.5 19. 3 10.2

69.5 46.9 24.5 14.4 8.0 22.6 30.4 20.2 10.2

69. 1 46.4 24.0 14.5 7.9 .22. 7 30.9 20. 2 10.7

1. 282 828 454

All footnotes p r e c e d e Table 101. ( 22? ]

T A B L E 102 LIVING A R R A N G E M E N T : DISTRIBUTION OF AGED ECONOMIC UNITS BY T Y P E OF UNIT AND LIVING ARRANGEMENT ( F A M I L Y STATUS): UNITED STATES, 1951 Sample percentages Type of unit and living arrangement

-

Number of units (in thousands)

Per cent of total sample

P e r cent of sample reporting income

P e r cent of sample reporting receipts

A l l units

Total

9, 803

100.0

100.0

100.0

Not living with relatives Total living with relatives Total family heads Family heads, children 21 or older living in household Family heads, no children 21 or older living in household Parent of family head Other relative of family head

5,247 4, 556 2, 199

53.5 46.5 22.4

52.9 47. 1 22.6

52.5 47.5 23. 1

1,723

17.6

18.0

18.4

476 1, 724 633

4.9 17.6 6.5

4.6 18. 1 6.5

4.6 18. 2 6.2

Couples Total

3, 763

100.0

100.0

Not living with relatives Total living with relatives Total family heads Family heads, children 21 or over living in household Family heads, no children 21 or older living in household Parent of family head Other relative of family head

2,600 1, 163 1,020

69.1 30.9 27.1

69.2 30.8 26.9

68.9 31. 1 27.2

850

22.6

22.7

23.0

170 123 20

4.5 3.3 0.5

4.2 3.3 0.6

4.2 3.4 0.6

1,810

100.0

100.0

100.0

920 890 323

50.8 49.2 17.9

50.5 49.5 18.3

50.2 49.8 18.9

200

11.0

11.5

11.9

123 357 210

6.8 19.7 11.6

6.8 19.5 11.7

7.0 20.3 10.7

100.0

Unrelated males Total Not living with relatives Total living with relatives Total family heads Family heads, children 21 or older living in household Family heads, no children 21 o r older living in household Parent of family head Other relative of family head

Unrelated females Total

4, 230

100.0

100.0

100.0

Not living with relatives Total living with relatives

1,727 2,503 856

40.8 59.2 20. 3

39. 3 60.7 20.6

38.5 61.5 21.2

673

15.9

16. 7

17.2

183 1, 244 403

4.3 29.4 9.5

3.9 30.5 9.5

4.0 30.9 9.4

Family heads, children 21 or older living in household Family heads, no children 21 o r older living in household Parent of family head Other relative of family head A l l footnotes precede Table 101.

[ 228 ]

TABLE 103 NATIVITY: DISTRIBUTION OF AGED ECONOMIC UNITS B Y T Y P E OP UNIT, NATIVE BORN, NOT NATIVE BORN (BY YEARS OF RESIDENCE): UNITED STATES, 1951 Sample percentages Number of units (in thou sands)

Type of unit

P e r cent of total mple

P e r cent of sample reporting income

P e r cent of sample reporting receipts

9, 803

100. 0

100 0

100 0

7,909

80. 7

80 7

80. 6

1,894 572 1.052 220 18 32

19. 3 5. 8 10. 8 2. 3 0. 1 0. 3

19 5 10 2 0 0

3 9 6 3 2 3

19 4 5. 9 10. 6 2. 4 0. 2 0. 3

All unit 8 Total Head born in United States or born abroad of American parents Head not born in United States nor born abroad of American parents Y e a r s residence in United States: 60 or more . . . 40 to 59 20 to 39 10 to 19 less than 10 . .

Couples * Total Head born in United States or born abroad of American parents Head not born in United States nor born abroad of American parents Y e a r s residence in United States: 60 or more . . . 40 to 59 :. 20 to 39 10 to 19 less than 10 . .

3, 763

100. 0

100 0

100 0

3, 013

80. 1

80 2

80 2

19. 4. 13. 2.

19 4 13 2

19 4 13 2

750 162 495 86 7

9 3 2 3

8 1 0 5

8 2

0 5

0. 2

0 2

0 1

1,810

100 0

100 0

100.0

1, 363

75 3

75 7

75.7

447 152 235 46 7 7

24. 7 8. 4 12. 9 2. 6 0. 4 0. 4

24 8 12 2 0 0

3 5 5 7 4 2

24.3 8.4 12.6 2.7 0.4 0.2

Unrelated males Total Head born in United States or born abroad of American parents Head not born in United States nor born abroad of American parents Y e a r s residence in United States: 60 or more . . . 40 to 59 20 to 39 10 to 19 l e s s than 10 . .

Unrelated females Total Head born in United States or born abroad of American parents Head not born in United States nor born abroad of American parents Y e a r s residence in United States: 60 or more . . . 40 to 59 20 to 39 10 to 19 less than 10 . .

4, 230

100 0

100 0

100. 0

3,533

83 5

83 2

83. 2

16 6 7 2 0 0

16 6 7 2 0 0

16. 6. 7. 2. 0. 0.

697 258 322 88 11 18

All footnotes precede Table 101 [ 229 )

5 1 6 1 2 4

8 4 6 0 3 5

8 4 5 1 3 5

TABLE 104 LONGEST OCCUPATION: DISTRIBUTION O F AGED ECONOMIC UNITS BY TYPE O F ECONOMIC UNIT, LONGEST OCCUPATIONAL GROUP: UNITED STATES, 1951 Sample p e r c e n t a g e s Number of units (in thousands)

Type of unit

P e r cent of total sample

P e r cent of s a m p l e reporting income

P e r cent of s a m p l e reporting receipts

All units Total

9, 803

100 0

100 0

100 0

Total without work e x p e r i e n c e Total with w o r k e x p e r i e n c e Total p r o f e s s i o n a l , technical, and kindred w o r k e r s Self-employed Salaried F a r m e r s and f a r m m a n a g e r s Total m a n a g e r s , o f f i c i a l s , and p r o p r i e t o r s , except f a r m Self-employed Salaried C l e r i c a l and kindred w o r k e r s Sales w o r k e r s C r a f t s m e n , f o r e m e n , and kindred w o r k e r s O p e r a t i v e s and kindred w o r k e r s P r i v a t e household w o r k e r s S e r v i c e w o r k e r s , except p r i v a t e household F a r m l a b o r e r s and f o r e m e n L a b o r e r s , except f a r m and mine

2,459 7, 344

24 6 75 4

24 4 75 6

24 2 75 8

559 149 410 1, 531

5 1 4. 15

5 1 3 16

5 1 3 16

538 311 227 369 363 1, 102 1,054 418 533 458 419

5 5 3. 2 2 3 3 8 3. 7 11. 4 10 9 4. 2 5. 5 4. 6 4. 3

5 0 3 0 2. 0 4 0 3 6 11 6 11 0 4 3 5. 6 4 9 4. 1

4. 9 2. 9 2 1 3 9 3. 5 11. 7 10. 9 4 4 5. 5 5 0 4. 2

7 5 2 8

2 3 9 3

1 2 9 5

Couples * Total

3, 763

100 0

100 0

100. 0

Total without work e x p e r i e n c e Total with work e x p e r i e n c e Total p r o f e s s i o n a l , technical, and kindred w o r k e r s Self-employed Salaried F a r m e r s and f a r m m a n a g e r s Total m a n a g e r s , o f f i c i a l s , and p r o p r i e t o r s , except f a r m Self-employed

7 3, 756

0. 2 99. 8

0. 2 99. 8

0. 2 99. 8

282 113 169 1, 058

7. 5 3. 0 4. 5 28. 0

6. 6 2 5 4. 1 29. 1

6. 1 2. 1 4. 0 29. 2

304 197 107 152 176 769 474 17 186 93 245

8. 1 5. 2 2. 9 4. 0 4. 7 20. 5 12. 6 0. 5 4. 9 2. 5 6. 5

7. 0 4 6 2 4 4. 3 4. 3 21 2 12. 9 0. 5 5. 0 2. 6 6 3

6. 9 4. 5 2 4 4. 4 4. 2 21. 2 13. 1 0. 5 5. 0 2. 7 6. 4

C l e r i c a l and kindred w o r k e r s Sales w o r k e r s C r a f t s m e n , f o r e m e n , and kindred w o r k e r s O p e r a t i v e s and kindred w o r k e r s P r i v a t e household w o r k e r s S e r v i c e w o r k e r s , except p r i v a t e household F a r m l a b o r e r s and f o r e m e n L a b o r e r s , except f a r m and mine

....

All footnotes p r e c e d e Table 101.

[ 230 ]

TABLE 104 (Continued) LONGEST OCCUPATION: DISTRIBUTION OF AGED ECONOMIC UNITS B Y T Y P E O F ECONOMIC UNIT, LONGEST OCCUPATIONAL GROUP: UNITED STATES, 1951 Sample percentages Number of units (in thousands)

Type of unit

Per cent of total sample

Per cent of sample reporting income

P e r cent of sample reporting receipts

Unrelated males Total

1,810

100 0

100. 0

100.0

Total without work experience Total with work experience Total professional, technical, and kindred workers Self-employed Salaried F a r m e r s and farm managers Total managers, officials, and proprietors. except farm Self-employed Salaried Clerical and kindred workers Sales workers Craftsmen, foremen, and kindred workers . . . . Operatives and kindred workers Private household workers . Service workers, except private household . . . . F a r m laborers and foremen , . . Laborers, except farm and mine

4 1,806

0 2 99 8

0. 2 99. 8

0.2 99.8

71 7 64 424

3 0 3 23

9 4 5 4

4 1 0 4 3. 7 24. 1

4. 1 0.4 3.7 24.6

167 89 78 74 84 319 264 11 95 134 163

9 1 4 8 4. 3 4. 1 4 6 17 6 14 9 0 6 5. 4 7 4 8. 9

8. 3 4. 9 3. 4 4. 3 4. 7 17. 4 14. 8 0. 6 5. 3 7. 7 8. 5

8.3 4.8 3. 5 3. 9 4.7 17.8 14.3 0.6 4.9 7.9 8.7

Unrelated females Total

4, 230

100. 0

100.0

100.0

Total without work experience Total with work experience Total professional, technical, and kindred workers Self-employed Salaried F a r m e r s and farm managers • • • Total managers, officials, and proprietors, except farm Self-employed Salaried Clerical and kindred workers Sales workers Craftsmen, foremen, and kindred workers Operatives and kindred workers Private household workers Service workers, except private household . . . . F a r m laborers and foremen Laborers, except farm and mine

2,448 1, 782

57. 7 42. 3

57.3 42.7

56.9 43. 1

206 29 177 49

4. 0 4 1

9 7 2 2

4.4 0.6 3. 8 1.2

4.5 0.6 3.9 1.2

67 25 42 143 103 14 316 390 252 231 11

1. 6 0. 6 1 0 3 4 2. 4 0 3 7. 6 9 3 6. 0 5 5 0 2

1.7 0.6 1.1 3.5 2.4 0.3 7.5 9.5 6.3 5.7 0.2

1.7 0.6 1.1 3.4 2.4 0.3 7.5 9.8 6.3 5.9 0.2

All footnotes precede Table 101.

[ 231 ]

T A B L E 165 HOUSING C O S T : D I S T R I B U T I O N O F A G E D E C O N O M I C U N I T S B Y T Y P E O F U N I T , H O M E O W N E R S H I P (BY P A Y M E N T S T A T U S ) A N D H O M E R E N T A L (BY M O N T H L Y R E N T ) : UNITED STATES.

1951 Sample percentages

T y p e of u n i t a n d h o u s i n g c o s t

Number of u n i t s (in t h o u s a n d s )

P e r cent of t o t a l sample

P e r cent of s a m p l e reporting income

P e r cent of s a m p l e reporting receipts

All u n i t s Total

9. 803

100.0

Owners—Total F r e e and c l e a r Still paying Tenters—Total $0 $1 t o $19 $20 t o $39 $40 t o $59 $60 o r m o r e

5,011 4,363 648 4, 792 2, 654 560 1,055 345 178

51.8 45.0 6.8 48.2 26.7 5.7 10.4 3.6 1.9

51 44 6 49 27 5 10 3 1

0 3 7 0 4 7 6 3 9

51.1 44.4 6.7 48.9 27. 5 5.8 10. 5 3.3 1.9

3, 763

100.0

100.0

100 0

2, 753 2, 360 393 1,010 363 125 279 143 100

73.7 62.9 10.8 26.3 9.4 3.4 7. 1 3.7 2.7

73.2 62.6 10.6 26.8 9.8 3.4 7.2 3.8 2.7

100 0

100.0

Couples

Owners—Total F r e e and clear Still paying Renters—Total $0 $1 t o $19 $20 t o $39 $40 to $59

73 62 10 26 9 3 7 3 2

2 5 7 8 7 4 4 5 7

Unrelated males Total

1,810

Owners—Total F r e e and c l e a r Still paying Renters—Total $0 $1 t o $19 $20 to $39 $40 to $59 $60 o r m o r e • •

651 561 90 1, 159 605 171 312 55 16

100 0 36 31 5 63 32 9 16 3 1

100.0

1 0 1 9 8 7 8 5 0

• 36.0 31.3 4. 7 64.0 33.4 9.8 17.4 2.4 1. 1

100 0 36 31 4 63 33 9 17 2 0

3 5 8 7 4 9 2 4 9

Unrelated females Total

4 , 230

100.0

100.0

100.0

Owners—Total

1,607 1,442 165 2,623 1, 686 264 464 147 62

38.6 34.7 3.9 61.4 39.8 6.0 10.5 3.6 1.5

37.3 33.4 3.9 62.7 40.9 6. 1 10.7 3.4 1.6

37.4 33.5 3.9 62.6 41.2 6. 1 10.5 3.4 1.5

Still paying Renters—Total $0 $ 1 t o $19 $20 t o $39 $40 t o $59 A l l f o o t n o t e s p r e c e d e T a b l e 101.

[ 232 ]

T A B L E 106 M E D I C A L C O S T : D I S T R I B U T I O N O F AGED ECONOMIC UNITS B Y T Y P E O F UNIT,

RECEIPT

O F MEDICAL S E R V I C E , AND S I Z E O F MEDICAL P A Y M E N T S : UNITED S T A T E S ,

1951

Sample p e r c e n t a g e s Type of unit and m e d i c a l c o s t

Number of units (in thousands)

P e r cent of total sample

P e r cent of s a m p l e reporting income

P e r cent of s a m p l e reporting receipts

All units Total

9, 803

100.0

No m e d i c a l s e r v i c e in 1951 Total units with m e d i c a l s e r v i c e in 1951 $0 paid $1 to $49 paid $50 to $149 paid $150 to $299 paid $300 o r m o r e paid

3, 500 6, 303 471 2.338 1,976 747 771

36.0 64.0 4.7 23.4 20.4 7. 7 7.8

100. 0 35. 64. 4. 23. 20. 7. 7.

8 2 6 6 4 7 9

100 0 35 64 4 23 20 7 7

8 2 6 8 2 8 8

Couples Total

3. 763

No m e d i c a l s e r v i c e in 1951 Total units with m e d i c a l s e r v i c e in 1951 $0 paid $1 to $49 paid $50 to $149 paid $150 to $299 paid $300 o r m o r e paid

1, 130 2,633 146 815 881 404 387

100 0 30 69 3 21 23 10 10

8 2 8 2 5 5 2

100 0 30 69 3 21 23 10 10

2 8 8 6 4 7 3

100 0 30 69 3 21 23 10 10

5 5 6 9 0 8 1

Unrelated m a l e s Total No m e d i c a l s e r v i c e in 1951 Total units with m e d i c a l s e r v i c e in 1951 $0 paid $1 to $49 paid $50 to $149 paid $150 to $299 paid $300 o r m o r e paid

1,810

100.0

316 994 99 430 263 59 143

44.8 55. 2 5.6 23.6 14.8 3.7 7.4

100. 0 45. 54. 5. 23. 14. 3. 7.

1 9 5 3 9 3 8

100 0 45 54 5 23 15 3 7

1 8 4 5 0 3 7

Unrelated females Total

4. 230

100.0

100.0

No m e d i c a l s e r v i c e in 1951 T o t a l units with m e d i c a l s e r v i c e in 1951 $0 paid $1 to $49 paid $50 to $149 paid $150 to $299 paid $300 o r m o r e paid

1, 554 2, 676 226 1.093 832 284 241

36.9 63. 1 5. 1 25.4 20.0 6.8 5.8

36.7 63.3 5.1 25. 5 20.1 6.9 5.6

All footnotes p r e c e d e T a b l e 101.

[ 233 ]

100 0 36 63 5 25 20 7 5

4 6 2 8 0 0 5

N E T V A L U E OF T O T A L ASSETS: DISTRIBUTION OF AGED ECONOMIC UNITS B Y T Y P E OF UNIT, AND AMOUNT OF ASSET HOLDINGS: UNITED STATES. 1951 Sample percentages Type of unit and net value of total assets

Number of units (in thousands)

P e r cent of total sample

P e r cent of sample reporting income

P e r cent of sample reporting receipts

A l l units Total

9, 803

100.0

100.0

Units with no assets Total units with assets $1 to $499 $500 to $999 $1,000 to $2,999 $3, 000 or more

2, 211 7,592 953 685 804 5, 150

22.0 78.0 9.5 6.9 8.1 53.5

22.8 77.2 9.7 7.1 8.1 52.3

100 0

100 0

100 0 23 76 9 7 8 51

1 9 8 2 2 8

Couples Total

3.763

Units with no assets Total units with assets $1 to $499 $500 to $999 $1,000 to $2,999 $3, 000 or m o r e

488 3. 275 208 129 280 2,658

12 87 5 3 7 71

4 6 3 3 4 6

13 87 5 3 7 70

0 0 6 5 6 4

100 0 13 86 5 3 7 69

3 7 7 6 5 9

Unrelated males Total

1, 810

100.0

Units with no assets Total units with assets $1 to $499 $500 to $999 $1,000 to $2,999 $3, 000 or more

622 1, 188 109 179 164 736

33.5 66.5 6.4 9.9 9.0 41.2

100 0 33 66 6 10 9 41

7 3 0 1 0 3

100 0 • 34 65 6 9 9 40

2 8 1 9 1 8

Unrelated females Total

4,230

Units with no assets Total units with assets $1 to $499 $500 to $999 $1,000 to $2, 999 $3, 000 or m o r e

1, 101 3, 129 636 377 360 1, 756

A l l footnotes precede Table 101.

[ 234 ]

100.0 25.9 74. 1 14.9 9.0 8.3 41.9

100 0 26 73 15 9 8 40

9 1 3 3 2 3

100 0 27 72 15 9 8 39

2 8 4 3 2 8

EARNINGS: DISTRIBUTION OF AGED ECONOMIC UNITS® BY T Y P E OF UNIT, AND AMOUNT OF EARNINGS: UNITED STATES, 1951

All units

Amount of earnings

Couples

Unrelated males

Unrelated females

Number of units with earnings (thousands) P e r cent having earnings

3, 284 33.5

2, 132 56.7

621 34.3

531 12.6

P e r cent of those with earnings $1 to $499 $500 to $999 $1,000 to $1,499 $1, 500 to $1, 999 $2,000 to $2,499 $2,500 to $2,999 $3,000 to $3,499 $3, 500 to $3, 999 $4, 000 to $4,999 $5,000 to $5,999 $6,000 to $6,999 $7,000 to $9, 999 $10,000 to $14,999 $15,000 o r more

100.0 28.2 13.0 10.8 9.2 8.5 6.3 5.9 5. 1 5. 1 3.2 1.1 1.3 0. 8 1.4

100.0 22.0 11.7 11.0 8.4 9.5 6.9 7.3 6.9 6.8 3.7 • 1.4 1.7 1.2 1.7

100.0 28.4 15.3 9.7 12.5 9. 7 6.8 5. 1 3.4 2.3 3.4 1. 1 1. 1

100.0 53. 1 15.6 11.6 8.2 3.4 3.4 1.4 2.0 0. 7

1.1

0.7

Median earnings of those with earnings

$1,407

$1,820

$1.323

$471

Mean earnings of those with earnings

$2, 150

$2,514

$1, 885

$995

A l l footnotes precede Table 101.

T A B L E 109 ASSET INCOME: DISTRIBUTION OF AGED ECONOMIC 3 UNITS BY T Y P E OF UNIT, AND AMOUNT OF ASSET INCOME: UNITED STATES, 1951

All units

Couples

Unrelated males

Unrelated females

Number of units with asset income (thousands) P e r cent having asset income

2, 178 21.6

958 25.4

307 16.5

913 20.4

P e r cent of those with asset income $1 to $199 $200 to $449 $500 to $999 $1, 000 to $1,499 $1, 500 to $2,499 $2, 500 or more

100.0 20.7 30.7 22.6 10.1 8.8 7. 1

100.0 23.0 26.2 23.0 10. 5 11.3 5.9

100.0 27. 5 36. 2 16. 2 8.8 5.0 6.2

100.0 15.7 33.9 24.3 10.0 7.4 8.7

Amount of asset income

Median asset income of those with asset income , . .

$485

$517

$385

$508

Mean asset income of those with asset income

$813

$822

$673

$851

A l l footnotes precede Table 101.

[ 235 ]

DISSAVING: DISTRIBUTION O F AGED ECONOMIC UNITS a BY TYPE O F UNIT. AND AMOUNT O F DISSAVING: UNITED STATES, 1951 Amount of dissaving

All units

Couples

Unrelated males

Unrelated females

Number of units with d i s s a v i n g s (thousands) P e r cent having dissavings

1. 502 15.3

602 16.0

257 14. 2

643 15. 2

P e r cent of t h o s e with d i s s a v i n g s $1 to $199 $200 to $499 $500 to $999 $1,000 to $1,499 $1, 500 to $2,499 $2. 500 o r m o r e

100.0 21. 1 31.7 25.8 12.5 4.4 4.4

100.0 20. 1 28.2 26.8 12.8 6.0 6.0

100.0 25.0 37.5 18.8 12.5 6.3

100.0 20.4 32.7 27.9 12. 2 2.0 4.8

Median d i s s a v i n g s of those with dissavings

$473

$531

$399

$471

Mean d i s s a v i n g s of those with d i s s a v i n g s

$704

»782

$578

$681

All footnotes p r e c e d e Table 101

TABLE 111 PENSIONS: DISTRIBUTION O F AGED ECONOMIC U N I T S 3 BY TYPE O F UNIT, AND AMOUNT O F PENSION: UNITED STATES, 1951 All units

Couples

Unrelated males

Unrelated females

Number of units with pensions (thousands) P e r cent having pensions

2,854 28.8

1, 340 35.3

608 33.7

906 20.9

P e r cent of those with pensions $1 to $199 $200 to $499 $500 to $999 $1,000 to $1,499 $1,500 to $2,499 $2, 500 o r m o r e

100.0 7.0 29.6 36.6 14.0 9.9 2.6

100.0 7.0 18. 7 36.0 19.8 14.2 4. 2

100.0 5.5 32.3 39.6 14.0 6.7 1.8

100.0 8.0 44. 3 36.3 5. 1 5.5 0.8

Amount of pensions

Median pension of those with pensions

$681

$842

$653

$483

Mean pension of those with pensions

$838

$1, 001

$780

$634

All footnotes p r e c e d e Table 101.

[ 236 ]

T A B L E 112 ASSISTANCE: DISTRIBUTION O F AGED ECONOMIC UNITS BY T Y P E O F UNIT, AND AMOUNT O F ASSISTANCE: U N I T E D S T A T E S ,

A m o u n t of a s s i s t a n c e

All u n i t s

1951

Couples

Unrelated males

Unrelated females

N u m b e r of u n i t s with a s s i s t a n c e ( t h o u s a n d s ) P e r cent h a v i n g a s s i s t a n c e

2,177 21.8

618 16.6

475 25.4

1,084 24.9

P e r c e n t of t h o s e with a s s i s t a n c e $1 to $199 $200 to $499 $500 to $999 $ 1 , 0 0 0 to $ 1 , 4 9 9 $ 1 , 5 0 0 to $ 2 , 4 9 9 $2, 500 o r m o r e

100.0 7.8 41.4 45.7 4.2 0.9

100.0 7.1 33.9 43.5 12. 5 3.0

100.0 8.9 46.0 44.4 0.8

100.0 7.8 43.8 47. 7 0. 7

M e d i a n a s s i s t a n c e of t h o s e with a s s i s t a n c e

$508

$602

$468

$488

M e a n a s s i s t a n c e of t h o s e with a s s i s t a n c e

$563

$667

$512

$528

All f o o t n o t e s p r e c e d e T a b l e 101.

T A B L E 113 CONTRIBUTIONS: DISTRIBUTION O F A G E D ECONOMIC U N I T S 3 BY T Y P E O F UNIT, AND AMOUNT O F CONTRIBUTION: U N I T E D S T A T E S . A m o u n t of c o n t r i b u t i o n N u m b e r of u n i t s with c o n t r i b u t i o n s ( t h o u s a n d s ) P e r cent having contributions P e r cent of t h o s e with c o n t r i b u t i o n s $1 t o $199 $200 to $499 $500 to $999 $1,000 to $1,499 $1, 500 to $ 2 , 4 9 9 $2, 500 o r m o r e M e d i a n c o n t r i b u t i o n of t h o s e with c o n t r i b u t i o n s . . . .

Couples

1951 Unrelated males

Unrelated females

168 1.5

31 0.8

28 1.0

109 2.3

100.0 25.6 48. 7 20. 5

100.0 12.5 50.0 25.0

100.0 60.0 20.0 20.0

100.0 23.1 53.8 19.2

5. 1

12.5

$349

$424

All f o o t n o t e s p r e c e d e T a b l e 101.

[ 237 ]

3.8 $167

$349

TABLE

114

F A M I L Y INCOME: DISTRIBUTION O F AGED ECONOMIC UNITS

a

B Y T Y P E O F UNIT,

AND AMOUNT O F F A M I L Y I N C O M E : U N I T E D S T A T E S ,

All units

Amount of f a m i l y i n c o m e

93 9

P e r cent having family i n c o m e

100 12 20 11 8 6 5 6 5 8 5 3 4 3

P e r c e n t o f t h o s e with f a m i l y i n c o m e $1 to $499 $500 to $ 9 9 9 $ 1 , 0 0 0 to $ 1 , 4 9 9 $1, 500 to $ 1 , 9 9 9 $2, 000 to $ 2 , 4 9 9 $ 2 . 5 0 0 to $ 2 , 9 9 9 $ 3 , 0 0 0 to $ 3 , 4 9 9 $3, 500 to $3, 999 $ 4 , 0 0 0 to $ 4 , 9 9 9 $ 5 , 0 0 0 to $ 5 , 9 9 9 $6, 0 0 0 to $6, 999 $ 7 , 0 0 0 to $ 9 , 9 9 9 $10, 000 o r m o r e M e d i a n f a m i l y i n c o m e o f t h o s e with f a m i l y i n c o m e •

0 8 4 1 2 4 2 7 0 0 2 8 1 2

1951

96 2 100 8 16 15 9 8 6 7 5 7 5 3 3 3

0 6 4 0 E 0 6 4 1 8 0 4 6 5

$2,026

$1,849

Unrelated females

Unrelated males

Couples

95. 1

91. 3

100.0 15.8 25. 2 9.8 7.2 6.4 5. 1 5.3 4. 1 6.0 5.3 3.8 3.2 2.8

100.0 15.5 22. 1 7.9 7.2 4.8 4.0 6.7 5. 3 9. 1 5.5 4.0 4. 9 2.9

$1,462

$1.812

All footnotes precede T a b l e 101.

TABLE

115

SOURCE O F R E C E I P T S : DISTRIBUTION O F AGED ECONOMIC UNITS

b

B Y T Y P E O F UNIT,

A N D N U M B E R AND K I N D O F S O U R C E S O F R E C E I P T S 1 1 : U N I T E D S T A T E S ,

S o u r c e of r e c e i p t s

All units

Couples

males

1951

females

100 0

100 0

100 0

100.0

No m a j o r s o u r c e

20 6

9 1

18 6

32. 1

One m a j o r s o u r c e Earnings Pensions Asset income Assistance Contributions Other income Miscellaneous cash receipts Dissaving

55 17 12 6 15 0

9 5 0 5 7 4

53.0 6.0 10. 1 8.9 20.8 0.9 0. 1 0.9 5.3

Two or m o r e m a j o r s o u r c e s Earnings Pensions Asset income Assistance Contributions Other income Miscellaneous cash receipts Dissaving

23 6 7 3 2 0

Total

5 1 2 2 4 6

57.0 28 9 12 6 4 3 8 8 0 2

0 5 3 5

0 2 2 0

9 8 7 2 9 2

0 8 2 3

All footnotes precede T a b l e 101.

[ 238 ]

33 13 10 4 3 0

9 1 1 3 2 3

0 5 2 5

57 17 16 4 16 0

0 2 2 6 23 5 9 1 2

5 3 8 7 6

14.8 1.8 4. 5 2.8 2.7 0.2

0 9 3 2

1.1 1.7

P L A C E O F RESIDENCE: DISTRIBUTION O F AGED ECONOMIC UNITS BY T Y P E O F UNIT, A N D T O T A L U N I T M O N E Y I N C O M E , U R B A N (BY S I Z E O F P L A C E ) AND R U R A L : UNITED STATES, P a r t I.

1951

Vertical percentages Urban

Total unit i n c o m e

250,000 to 999,999

Under 250,000

Not i n urbanized area

Rural nonfarm

Rural farm

In u r b a n i z e d a r e a

Total Total Total

1,000,000 and over All units

N u m b e r of s a m p l e cases Total

2,730 100.0

1, 722 100.0

1, 114 100.0

576 100.0

341 100.0

197 100.0

608 100.0

597 100.0

411 100.0

$0 $ 1 t o $499 $500 t o $999 $1,000 to $1,499 $1,500 to $ 1 , 9 9 9 $2,000 to $2,499 $2,500 to $2,999 $3,000 to $3,499 $3,500 to $3,999 . . . . $4,000 to $4,999 $5,000 t o $ 5 , 9 9 9 $6,000 to $ 6 , 9 9 9 $7,000 t o $ 9 , 9 9 9 $10,000 to $14,999 . . $15,000 and o v e r • . .

21.1 20.3 23.7 10. 1 6.4 4.6 2.9 3. 1 1.9 2.3 1.4 0.6 0.7 0.4 0.5

22.2 15.8 22.9 10. 7 6.6 5.2 3. 3 3.9 2.4 2.8 1.6 0.3 0.8 0.6 0. 7

24.4 14.0 22.8 10.2 6.8 4.9 2.9 3.9 2.4 3.3 1. 8 0.3 1.0 0.4 0.7

25.7 11.5 22.6 10. 1 6.2 6.1 2.8 4.3 3.0 3.6 2.3 0.2 0.7 0.5 0.5

26. 1 15.5 22.0 10.3 7.6 2.3 2.3 2.3 2.1 2.6 2. 1 0.6 2. 1 0.6 1.5

17.8 18.8 24.9 10. 7 7. 1 6. 1 4. 1 5.6 1.5 3.6

18. 1 19. 1 23. 2 11.7 6.3 5.6 3.9 3.9 2.5 2.0 1.3 0.5 0.5 0.8 0.7

20.6 25.5 26.0 8.4 6.7 4.2 2.0 1.5 1.0 1.0 0.8 1.3 0.7 0.3

17.0 31.9 23.6 9.7 5. 1 2.9 2.4 1.9 0.7 2.2 1.2 0.5 0.2 0.2 0.2

Median income M e d i a n i n c o m e of t h o s e with income

$682

$762

$754

$785

$690

$770

$777

$576

$523

$904

$1,008

$1,048

$1,155

$987

$949

$972

$774

$704

115 100 0

69 100 0

245 100 0

233 100.0

203 100.0

8 9 15 8 15 7 3 5 4 7 4 1 3 1 3

8 8 8 15 11 13 8 13 2 8

9 3 6 3 6 2 5 9 7 5 9 8 8 0 8

10.3 10. 7 24.0 16.3 12.4 7.3 4.3 3.4 2.6 2.1 2. 1 2.6 0. 9

6.9 24.1 24.6 14.3 8.9 4.9 4.4 3.0 0.5 4.4 1.5 1.0 0.5 0.5 0.5

Couples N u m b e r of s a m p l e cases Total

1,045 100.0

609 100.0

7.5 11.2 19.4 IS.4 10.5 8.7 5.3 5.6 3.7 5.5 2.7 1.2 1.1 1. 1 1. 1

6.6 7. 1 15.9 15.4 10.3 10.5 5. 9 7.4 5.3 7. 1 3.3 0.8 1.5 1.6 1.3

Median income

$1,387

M e d i a n i n c o m e of t h o s e with income

$1,512

$0 $1 t o $499 $500 t o $999 $1,000 t o $ 1 , 4 9 9 $1,500 t o $ 1 , 9 9 9 . . . . $2,000 t o $ 2 , 4 9 9 . . . . $2,500 t o $ 2 , 9 9 9 $3,000 to $3,499 $3,500 to $3,999 $4,000 t o $ 4 , 9 9 9 $5,000 to $ 5 , 9 9 9 . . . . $6,000 t o $ 6 , 9 9 9 $7,000 to $9,999 $10,000 to $14,999 • . $15,000 and o v e r • . .

364 100 0 7 6 12 14 11 10 5 7 4 8 3 0 1 1 1

180 100.0 6.7 4.4 11. 1 18.3 8.9 12. 2 5. 6 7.2 6. 1 10.0 4.4 0.6 1.7 1.7 1. 1

$1,742

$1,8 69

$2,023

$1,736

$1,844

$1,494

$1,151

$1,901

$2,039

$2,159

$1,875

$2,028

$1,631

$1,309

A l l f o o t n o t e s p r e c e d e T a b l e 101.

7 6 7 7 7 0 5 2 3 0 3 7 5 7 5

7 7 7 9 6 0 7 0 9 7

4 7 21 16 8 10 6 6 5 4 2 0 0 2 0

7 9 1 8 5 7 5 7 9 8 6 8 9 4 6

0.9

$ 885

$ 955 {Continued)

[ 239 ]

T A B L E 201 (Continued) P L A C E O F RESIDENCE: DISTRIBUTION O F AGED ECONOMIC UNITS BY T Y P E O F UNIT, A N D T O T A L U N I T M O N E Y I N C O M E , U R B A N (BY S I Z E O F P L A C E ) A N D R U R A L : UNITED STATES, P a r t I.

1951

Vertical percentages (Continued) Urban

Total unit i n c o m e

Total

Not i n urbanized area

Rural nonfarm

98 100 0

127 100.0

88c 100.0

13.4 31.5 36.2 6.3 3.9 4.7 1.6

20.5 33.0 28.4 6.8 3.4 2.3 1. 1 2.3

1 8

10 2 23.5 24 5 19 4 6 1 4 1 2 0 4 1 1 0

5 3

1 0 1 0 1 0

In u r b a n i z e d a r e a Total

Total

1,000,000 250,000 to and o v e r 999,999

Under 250,000

farm

Unrelated males N u m b e r of s a m p l e cases Total

200 100 0

109 100.0

0 0 0 0 0 5 0 5 5 0 0

21. 1 15.6 27.5 6.4 4.6 5. 5 3.7 6.4 3.7 1.8 3.7

57 100 0

c

34 100.0

d

513 100 0

298 100 0

$0 $1 t o $499 $500 t o $999 $1,000 t o $ 1 , 4 9 9 . • • $1,500 t o $ 1 , 9 9 9 . • • $2,000 t o $ 2 , 4 9 9 • • • $2,500 t o $ 2 , 9 9 9 , - $3,000 t o $ 3 , 4 9 9 - - $3,500 to $3,999 • - • $4,000 t o $ 4 , 9 9 9 • - $5.000 to $5,999 - - $6,000 t o $ 6 , 9 9 9 • • • $7,000 t o $ 9 , 9 9 9 • • • $ 1 0 , 0 0 0 to $ 1 4 , 9 9 9 $15,000 and o v e r • •

15 24 29 10 5 3 2 2 1 0 1 0 0

15 19 27 13 6 3 3 4 2 0 1 0 1

0 4

0 7

Median income M e d i a n i n c o m e of t h o s e with income

$662

$780

$759

$742

$717

$846

$834

$571

$449

$794

$921

$914

$934

$900

$885

$938

$663

$620

94 100 0

265 100 0

237 100 0

120 100.0

8 6 8 7 1 7 5 9 2 6 4 6 8

4 1 5 8 0 7 4 4 0 7 7 3 3

18 17 29 11 6 3 4 4 2 1 2

1 5

19 19 26 17 5

3 3 3 5 3

5 3

0.8 2.3 1.6

2 0

Unrelated females N u m b e r of s a m p l e cases Total

815 100.0

550 100.0

5 6 8 0 3 4 9 9 5 3 3

36.3 21. 1 26. 5 6.1 4.0 1.7 1.2 1.2 0.5 0.5 0.4

37.8 17.6 27.6 6.9 4.0 1.6 0.7 1.3 0. 7 0.5 0.5

0 3

0. 1

0.2

1,172 100 0

287 100.0

169 100.0

39 14 27 6 5 2 0 1 0 0 0

40.2 18.3 24.9 8.9 3.0

4 3 9 3 2 4 7 7 7 3 3

28 26 31 5 2 2 1

7 6 9 3 1 1 1

33 28 24 4 4 1 2 1

2 3 2 5 2 9 3 1

34 36 22 1 2 0

6 7 4 7 5 8

31.7 44.2 18. 3 4. 2

$0 $1 t o $499 $500 to $999 $1,000 to $1,499 . . • $1.500 to $1,999 . . . $2.000 to $2,499 . • . $ 2 , 5 0 0 t o $2,999 • - • $3,000 t o $ 3 , 4 9 9 . • $3,500 to $3,999 . • • $4,000 t o $ 4 , 9 9 9 . • • $5,000 to $ 5 , 9 9 9 . • • $6,000 t o $ 6 , 9 9 9 • • • $7,000 t o $ 9 , 9 9 9 • • • $10,000 to $14,999- • $15,000 and o v e r • • •

35 26 24 5 3 1 0 0 0 0 0

0 2

0.2

0.4

0 3

0.6

Median income M e d i a n i n c o m e of t h o s e with income

$273

$324

$346

$372

$266

$400

$297

$210

$208

$614

$703

$744

$788

$732

$642

$6( 6

$446

$387

0.6 1.2 0.6 0.6 1.2

1 1 1 1

0 8

0 3

A l l f o o t n o t e s p r e c e d e T a b l e 101. [ 240 1

0 4

0 4

T A B L E 201 (Continued) P L A C E OF RESIDENCE: DISTRIBUTION OF AGED ECONOMIC UNITS BY T Y P E OF UNIT, A N D T O T A L UNIT MONEY INCOME, URBAN ( B Y SIZE OF P L A C E ) A N D R U R A L : UNITED STATES. 1951 Part II.

Total unit income

Number of sample cases

Horizontal percentages Urban In urbanized area

Total Total Total

250,000 Under 1,000,000 to and over 999,999 250,000

Not in urbanized area

nonfarm

Rural farm

A l l units Total $0 $1 to $499 $500 to $999 $1,000 to $1,499 $1,500 to $1,999 $2,000 to $2,499 $2,500 to $2,999 $3,000 to $3,499. . . . $3,500 to $3,999 $4,000 to $4,9998. . .

2, 730 575 555 647 275 175 126 78 c 85 c 51 c 64 c

100 0

63. 1

40 8

21. 1

12 5

100 0 100 0 100 0 100 0 100 0 100 0 100 0 100.0 100 0 100 0

66.4 49.0 61. 1 67.3 65. 1 70.6 71. 8 80.0 82.4 76.6

47 28 39 41 43 43 41 51 52 57

25.7 11.9 20. 1 21. 1 20.6 27. 8 20. 5 29.4 33.3 32.8

15 9 11 12 14 6 10 9 13 14

5 5 6 7 9 3 3 4 7 1

6. 1 6. 7 7. 6 7. 6 8. 0 9. 5 10. 3 12. 9 5. 9 10. 9

6. 6

3 1 3 5 4 7 0 8 9 8

7. 2

22 3

21 9

15. 1

19 20 21 25 21 27 30 28 29 18

1 9 8 8 7 0 8 2 4 8

21 4 27 4 24 0 18 2 22 9 19 8 15.4 10 6 11 8 9 4

12.2 23. 6 15.0 14. 5 12.0 9.5 12. 8 9.4 5.9 14. 1

Couples Total $0 $1 to $499 $500 to $999 . . . . . . . $1,000 to $1,499 $1.500 to $1,999 $2,000 to $2,499 $2,500 to $2,999. . . • $3,000 to $3,499 $3,500 to $3,999 $4,000 to $4,9996. . .

1,045 78 c 117 203 161 110 91 i N V n .to

«O «0 t- O

.m

io t* m in ro eó o tó i t- « f C- lí) O OI

• © to o • -h o o

SJ931JOM p j o q -sanoq a ^ A i J j sad^JOM p d j p u p i pue 93A.IlBJ9dO

H H N O lt) M M N

r>iD 000'»$ 666'eí o» oos'et

V

O tA O O en co S S

%

66*'et °> ooo'et

3

0

g P J
co

p

O)

w

g g«

C a Z U

H S fH a

1

O o O «9-

.

.

• m o) •

.

.

.

. co io

co e

. h n

• •

S 2

66»'Zt o; OOO'Jt

H 0 H Q

2 n «A

i t* n (o o

N i

o

'

. m o e* io n c

66»'1$ OI 000'It 666t OOSt

S 3

• «

66»t o» t$

> lA Ifl 9) f

JO BBtrj

CC O H H to 00 CO CO m ^ «

«

«

16

> lA f ) IS IO N N

m

H
0 0 0 0 0 0 0

o o o o » in to N m H

« to -h e- cj

.

. Tf co S 2

. m in h iH o» m ifl . n H ® ^

• co t« « n • • N ^ t" I H H IA « t i V A il) H • • H H N IO I Q A N n

o «e a N

o o ® m

< o ce co ce
( f i W IO 4 ' O O f

H ( 0 h>

^

O O

M

o O

«9. o

t



O H N N n H t - H

¡ Tf o

co m •

O «o o e s o ifi t o n to œ io . o t - o i v m t - H i o i o •

< f i N 0 » N I C N « ( O ( O

> CO O) A

(O

O

I f l ^ H N • e o io i N co t e •

O) N O)
^ n

a o> O) có

( ( c "

******* *

o > « » O o o o o o «»

O Xo oO O o oo o o « o o O lA O lA O A O e î c>î HoiOHg)iainti]in

co m in in

o

o o o

IO N H O) H H O) Ifl eo w

O) O) A O) O} o> Oí O) O) A O O) 93 O) ^ O) ^ O) O) O) ¡ « e - o o o o o o o * ® O Q © © © 0 © > _ o o o o o o o >g©tft©iO©irt© O H i A H H c o N m n ^

o

Ci

^

03 H co

i n

S.

V (S « S

P



.M . H

to 41 u ) n ! r*

o ci



£ P

e g I1 >. -a Ï

^ N H (CO 00t»CO CO O)

•M Ifl O) N O) N to i

11) OO (S NOl O Ifì -H O 11) N

t» to CC Iß H (

O OOOO! O OOO O

J
S ** O) «9- O S s S S § § O -Htf}^

0

O o H

S ** O

5

CO o CO o CO lA

Median family income of those with income

2 S

iti^iNoioiOHmion lA^ameencencov co h HNNn n « «9-

jaAo pus 000'0t$

(00}cecoo*eo oo>> H H N N (I) «o V 49«9gj 5» CT) O O O O 0 O p * o ** ** ** o o ° o o o o oo o o _ o o o o © mo g ©m ©m © o H II) H H N N n 1

© m o ^ c o ^ o N - n c v i o n o n i o n c o n a u i c«î

h

H N N n

co

. o

a) c- *

in

.

CO co CO

ffl

«

a

o

co i o »



«O «



«ft

^



. -H •«

t*

o

to N

co

co

• C4 t

CO

^

H

H

CO CO 00 co

©

co co m •



in o

a>

m

co h

«

co

n

t-



eg e - -«« i o

«

c - in tj
CO 00 o tto

os o

•g 66VS$ ooo'et «2 0 a? 666'Z$ Ol o 00S'Z$ >

i l ffl N — ï> "O O _N ete ® • * es ß "D «c -fig -5 u 3 J3 a• S jï o S Sus o m ï i Z fa ÛH O w

s -o I -g u C g .s « ¿2.5 > ci tm o a 5 £O " su» O