Territories of Profit: Communications, Capitalist Development, and the Innovative Enterprises of G. F. Swift and Dell Computer 9781503624030

Territories of Profit compares Dell Computer, the dominant computer manufacturer of the late twentieth century, and G. F

128 53 4MB

English Pages 304 [296] Year 2022

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Territories of Profit: Communications, Capitalist Development, and the Innovative Enterprises of G. F. Swift and Dell Computer
 9781503624030

  • 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

Advance praise for Territories of Profit “Territories of Profit brings together rigorous scholarly research and an acute awareness of the new technological and social environment in which firms operate nowadays. It is essential reading for both students and practitioners in business and policy making.” —Manuel Castells, Wallis Annenberg Professor of Communication Technology and Society,Annenberg School for Communication, University of Southern California “Gary Fields’s amazingly original book shows us why we need an ‘actorcentered approach’ to regional development. Reinstating the corporation and its behavior as central to the innovative process, he deftly slays several sacred cows of the new economy: that wealth is now, unlike before, generated from information rather than production; that free markets necessarily facilitate innovation; that egalitarian networks among companies have displaced oligopolists and unbalanced market power. Probing two historical pace-setters, Swift Meatpacking and Dell Computer, he documents parallels between two periods of communications revolutions and how each firm used administrative control, market power, and reconfigured industrial districts to achieve super-profits. Anyone who thinks that the digital revolution will bring apocalyptic societal and geographical change or a kinder, gentler networked world should read this book cover to cover.” —Ann Markusen, Professor of Political Economy, Regional Planning and Geography, Humphrey Institute of Public Affairs, University of Minnesota “Many books have recently been written about the ‘Communication Revolution’ and the ‘Information Revolution.’ This fascinating and highly original comparative study of G. F. Swift and Dell makes compulsive reading and serves admirably to illustrate the outstanding features of both revolutions, of economic theory, and of innovation studies.” —Christopher Freeman, Emeritus Professor and Professorial Fellow, SPRU, University of Sussex

“This book makes an important contribution to our understanding of the evolution of the global economy by inserting business organization into the discussion. The comparison of Swift and Dell is compelling, linking transformations by US industrial corporations that have taken place a century apart through the concept of ‘administrative control.’ It provides an antidote to the popular view that the ‘managerial revolution’ is a thing of the past and that the ‘market coordination’ of the economy has reemerged as the organizing principle.” —William Lazonick, Research Professor, INSEAD, and Professor, University of Massachusetts, Lowell “This manuscript is timely. The stock market let-down has forced a reexamination of the ‘new economy,’ which scholars arguably found hard to come to terms with during its heyday. There is nothing like a nineteenthcentury viewpoint to help put the twentieth into perspective.” —Richard Langlois, Professor of Economics, University of Connecticut Gary Fields’s Territories of Profit tells a fascinating story about two firms that distill the essence of successful profit-making—G. F. Swift, the meat packing and distributing giant of the nineteenth century, and Dell Computer. Fields leads the reader through a penetrating analysis of how firms develop the capacity to respond in particularly effective ways to the opportunities presented to them by technological change. He demonstrates that ours is not the only era in which revolutionary changes in information and communications technologies gave firms new strategic options. He also shows how important it is not to separate information from its uses in the production and distribution of goods. Not coincidentally, Fields expands our geographic imagination, demonstrating how space for production and new markets enters into firms’ strategies. By doing so he brings us to a new understanding of the real processes shaping the global economy. —Susan Christopherson, Professor of City and Regional Planning, Cornell University

Territories of Profit

Innovation and Technology in the World Economy

editors Martin Kenney, University of California, Davis/Berkeley Round Table on the International Economy Bruce Kogut, Wharton School, University of Pennsylvania

Urs von Borg, The Triumph of Ethernet:Technological Communities and the Battle for the LAN Standard Gary Fields, Territories of Profit: Communications, Capitalist Development, and the Innovative Enterprises of G. F. Swift and Dell Computer

Territories of Profit Communications, Capitalist Development, and the Innovative Enterprises of G. F. Swift and Dell Computer

Gary Fields

stanford business books An imprint of Stanford University Press Stanford, California

2004

Stanford University Press Stanford, California www.sup.edu © 2004 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved. Library of Congress Cataloging-in-Publication Data Fields, Gary, 1954– Territories of profit : communications, capitalist development, and innovation at G. F. Swift and Dell Computer / Gary Fields. p. cm. — (Innovations and technology in the world economy) Includes bibliographical references. ISBN 0-8047-4721-0 (alk. paper) — ISBN 0-8047-4722-9 (pbk : alk. paper) 1. Technological innovations. 2. Computer industry—Technological innovations—Case studies. 3. Meat industry and trade—Technological innovations—Case studies. I. Title. II. Series. HC79.T4 F54 2003 338'.064—dc22 2003019107 Printed in the United States of America on acid-free, archival-quality paper Original Printing 2004 The last figure below indicates the year of this printing: 13 12 11 10 09 08 07 06 05 04 Designed and typeset at Stanford University Press in 10.5/12 Bembo.

For Amy

Contents

List of Tables and Illustrations Preface and Acknowledgments 1. Introduction: Communications and Innovation, Business Organization and Territory

xi xv

1

Part 1. Theory 2. From Communications and Innovation to Business Organization and Territory: A Synthesis

27

Part 2. Rails, Telegraphy, and the Business Enterprise of G. F. Swift & Company 3. The Railroad and Telegraph as Commerce System and Market Space

63

4. Continental Divide: The Business Enterprise of of G. F. Swift & Company

91

Part 3. The Internet and the Business Organization of Dell Computer 5. The Internet as Commerce System and Market Space

139

6. Global Reach: The Business Organization of Dell Computer

165

7. Conclusion: The Rhyme of History

220

Notes References Index

233 247 267

Tables and Illustrations

Tables 3.1 3.2

Expansion of Rail and Telegraph Systems, 1848–1902 Percentage of Western Exports Shipped via New Orleans, 1839–1860 3.3 Shipments of Grain and Flour from Chicago, 1840–1861 3.4 Rail Receipts at Chicago for Various Commodities, 1852–1856 3.5 Rail and Water Shipments of Flour and Grain Received at New York, 1860–1890 3.6 Live Cattle Shipments from Chicago, 1852–1861 3.7 Relative Wholesale Price Indexes in Different Cities, 1866–1868 and 1889–1891 3.8 Urbanization, Railroadization, and Industrialization of the United States, 1850–1900 3.9 Size Rank of Largest Cities, 1820–1850 and 1850–1890 3.10 Growth of Manufacturing Employment in U.S. Cities, 1860–1890 4.1 Size of Beef- and Pork-Packing Industries, 1870 4.2 Comparative Expansion of Beef and Pork Packing, 1870–1890 4.3 Chicago Meatpacking Firms Ranked by Size, 1878 and 1884 4.4 Number of Branch Houses of Swift and Major Firms, 1878–1899 4.5 Number of Cattle Shipped from Chicago, Live and as Dressed Beef, 1878–1887 4.6 Manufacturing Industries Ranked by Dollar Value of Output, 1870 and 1900

65 74 75 76 77 81 83 85 86 87 98 103 109 111 112 114

xii List of Tables and Illustrations 4.7

Capital Value and Sales of G. F. Swift & Company, 1885–1904 4.8 Rank of Individual Firms Constituting the Beef Trust, 1903 4.9 Largest Slaughtering States by Dollar Value of Animals Slaughtered, 1860 and 1900 4.10 Output of Meat Products and Manufactured Ice, 1869–1904 5.1 Number of Host Computers Connected to the Internet, 1982–1990 5.2 Estimated Number of Individual Internet Users Worldwide, 1993–2001 5.3 Internet Domain Name Registrations, 1992–2001 5.4 Growth in Percentage of .com Web Sites, 1993–1996 5.5 Estimated Levels of Internet Commerce, 1996–2001 6.1 The Early PC Market in the United States, 1976–1983 6.2 Computer Hardware Production in East Asia, 1980–1990 6.3 Early Sales and Profit of Dell Computer, 1985–1990 6.4 PC Firms Ranked by PC Revenues, 1990 6.5 Computer Firms Ranked by World Market Share, 1994 6.6 Daily Internet Sales of Dell Computer, 1996–2001 6.7 Days’ Supply of Inventory at Dell and Compaq, 1994–2001 6.8 Computer Firms Ranked by World Market Share, 1997 6.9 Year-by-Year Percentage Increases in PC Shipments, 1996–2000 6.10 Computer Firms Ranked by U.S. Market Share, 2001

115 116 121 130 149 150 151 152 159 171 172 181 183 185 188 190 191 213 214

Maps 3.1 3.2 4.1 4.2 4.3 4.4 4.5 4.6 4.7 6.1 6.2 6.3

Travel Time (from New York) as Market Space, 1830 Travel Time (from New York) as Market Space, 1857 The Dressed Beef Enterprise of G. F. Swift, 1878 The Dressed Beef Enterprise of G. F. Swift, 1880 Expansion of G. F. Swift Production Facilities Top Five Slaughtering States, 1860 Top Five Slaughtering States, 1900 The Dressed Beef Enterprise of G. F. Swift, 1903 Centralization and Decentralization in the Dressed Beef Enterprise of G. F. Swift, 1903 The Dell Business Enterprise, 1990 The Expansion of Dell Assembly Sites, 1991–1999 The Dell Business Enterprise, 2001

78 79 106 110 118 122 123 124 128 182 204 212

List of Tables and Illustrations xiii

Figures 1.1

From Communications Revolution to Territorial Transformation 2.1 Schematic Outline of Communications Revolution 4.1 Schematic Outline of Beef Channels 4.2 Slaughtering and Meatpacking Output for Eight Cities, 1900 5.1 Number of Venture-Backed Initial Public Offerings and Amount Raised, 1978–2000 6.1. Schematic Outline of PC Channels 6.2 Dell’s Growth Before and During the Internet Period 6.3 Growth of Dell Sales and Employment, 1994–2001 6.4 Inventory of Dell and Inventory/Shipments Ratio of PC Industry, 1994–2000 6.5 Declines in Computer Prices, 1987–1999 6.6 Productivity Index for Dell, Compaq, and Gateway, 1994–2001

12 58 108 119 158 180 186 187 216 217 218

Preface and Acknowledgments

this study began as an effort to uncover historical analogies to Internet commerce. In the course of its development, it established two parallel worlds, one occurring in the late nineteenth century, marked by the railroad and telegraph revolution, the other occurring during the final years of the twentieth century, marked by the ascendancy of the Internet. Two business firms, G. F. Swift and Dell Computer, eventually emerged as the protagonists of these dual universes and became the focus of the research for this study. One of the primary aims in setting up this comparison and examining the impacts of a previous communication revolution was to temper some of the exaggerated claims, rampant when research for this book was undertaken, about the uniqueness of the Internet and the so-called information age. There is no denying this uniqueness, but in the absence of a serious look back in time, such claims are historically static with little insight into what is actually new. Perhaps more importantly, however, this study set out to challenge a far more dubious set of principles being formulated in academic, policy, and business circles on the convergence of Internet communication, perfect markets, nonhierarchical organizations, and innovation. By supposedly equalizing access to information, Internet technology, we were told, was not only converging with more perfectly efficient markets. The Internet was promoting a more innovative economy, less hierarchical business organizations contracting as equals through the price system, and a recession-resistant pattern of capitalist development different from the past. Challenges to these conceits form the threads of this study. The lessons to be learned from Territories of Profit are threefold: first, while the Internet is unique, its impacts on innovation and profit making of business users are intelligible only as part of a much broader historical revolution in communications; second, innovation driving capitalist development is not something that occurs through established markets and the price system but is instead the outcome of firms creating organizations and using mechanisms of power strategically in pursuit of profit; and third, in pursuing this mission of

xvi Preface and Acknowledgments

crafting more innovative routes to accumulation, firms use power and organization to reconfigure territory such that the continental economic empire building of the vertically integrated firm and the global market development of the network firm share similar histories. Numerous people played critical roles in this project as it evolved along this trajectory, and they were instrumental in helping me along the route. First and foremost, I extend my gratitude to the people at the Berkeley Roundtable on the International Economy (BRIE). Without the unwavering support of BRIE and the intellectually vibrant environment that it cultivates, I would never have been able to conceive, much less undertake, this project. I hope that my work has in some way provided a source of enrichment to BRIE and the community associated with it. I also want to thank colleagues in the Department of Communication at the University of California, San Diego. They contributed numerous insights as the book was being completed. Several individuals made unique contributions to this study. Martin Kenney played a pivotal role. He first suggested to me that I might profit by examining the personal computer industry and perhaps even Dell Computer. He was also a tireless critic of my work at every step of the way, as well as a formidable source of information on a broad array of topics. I especially enjoyed our intellectual sparring sessions and was enormously gratified by his dedication and commitment to my work. In an early draft of the project proposal, AnnaLee Saxenian observed that what was interesting about the issue of business organization was the way firms organized economic activity across geographical space. Consequently, the territorial component of this project owes much to her input. Her frank and candid assessments of the work as it progressed always provoked me to refine my thinking and sharpen my arguments. Manuel Castells was an ongoing inspiration, providing me with an opportunity to launch my idea for comparing Swift and Dell in a presentation to his class on the Internet three years ago. His input on this presentation, at the time little more than a concept, convinced me that the idea was feasible. Brad DeLong was a constant source of encouragement for my comparison. He challenged me to draw out what I had learned from juxtaposing the two periods, the two infrastructures, and the two firms. Other individuals, though not directly involved in this study, contributed in other ways to its success. Michael Watts provided me with exposure to a broad and critical literature on development. Jan de Vries introduced me to a level of historical studies and an historical literature previously unknown to me. Mary Yeager was kind enough to respond to my early queries about the meatpacking industry and raised obvious issues about comparing beef and computers. Susan Christopherson made invaluable suggestions on an early presentation of the concepts in this book at a conference of the Asso-

Preface and Acknowledgments

xvii

ciated Schools of Planning. Finally, I am grateful to Michael Teitz for his ongoing support over many years of mentorship and friendship. Mike was responsible for challenging me to extend my theoretical inclinations into an empirical direction and had many insights on the project itself. Numerous individuals helped me with critical feedback as ideas in this book began to take shape. Todd Goldman, Asha Weinstein, Isabelle Fauconnier, Matt Zook, Larry Barone, Larissa Muller, and Claudia Leal read my early material and were at times my harshest critics. They were also among my most encouraging backers. Several people responded directly to the manuscript with comments and criticisms that, because of their efforts, are now part of the book. William Lazonick encouraged me to elevate business organization to a more prominent position in the analysis. Richard Langlois’s comments on an early draft and his paper “The Vanishing Hand” were critical in helping me frame my own ideas. Richard John read draft after draft of material on the nineteenth century and tirelessly kept commenting on ways to improve the book. Paul Duguid reviewed and critiqued what I considered some of the more controversial elements in the analysis, adding a level of erudition in his comments that I found exciting as well as insightful. Francois Bar made invaluable suggestions on the Internet as a communications and commerce infrastructure. Jan Birger Vagn Lauridsen was especially helpful in critiquing my formulations of transaction cost economics. Finally, my good friend Peter V. Hall, who witnessed this book as it emerged from its primitive roots, provided an excellent critique of Chapter 2 as well as other ideas in the study. Specialized technical support was also an important part of this study. Jeremy Shaw provided crucial assistance with cartography and showed great sense in linking my arguments to actual map design. Tony Hicks did a wonderful job in the layout of the book especially with regard to the tables and figures. Claudia DaMetz lent critical ideas to the design of the cover created so masterfully by Rob Ehle. On a more personal but somber note, I am truly saddened that my good friend Marvin Rosen could not witness the completion of this study. Marvin was the first to teach me the value of history and historical materialism. I believe he would have been both critical and pleased with this project, but ultimately gratified that he played a decisive role in the pathway leading to it. One of the most difficult hurdles I confronted in this study was interacting with representatives of Dell Computer. Laura Tyson was kind enough to write me a letter of introduction explaining the value of my study. Nigel Johnson, formerly of the Eclipse Group, was indispensable in orienting me to the world of Dell, and he played the critical role in helping me obtain my first interviews inside Dell. My lifelong friend Charles Piller of the Los An-

xviii Preface and Acknowledgments

geles Times helped me contact key Dell personnel as well as computer industry analysts. Several individuals at Dell took an active interest in my book, including Stephen Cook, Daryl Robertson, Simon Wong, and Lance St. Clair, all of whom were enormously helpful in enabling me to understand aspects of the Dell business system. I am also exceedingly grateful to Patricia Katayama at Stanford University Press. Pat took an interest in the project at an early date and made certain that the process was efficient. Her interest at every step was making certain that the final product was the best it could be. It is difficult to sustain momentum on a book without ongoing feedback. My wife and partner, Amy Rakley, tirelessly endured my theoretical incantations and made critical suggestions on issues such as the perishability of beef and personal computers. In addition, she heroically reviewed the manuscript when I could no longer do it. At key points, she prevented my writing from straying into the often arcane and indecipherable world of the academy and made me see connections in my own work that I did not realize were there. Her observations and input were indispensable. Finally, my most heartfelt thanks go to Stephen S. Cohen. His unflagging support and supreme confidence in my work, even in moments when I was not at all clear how I could complete this study, were a constant source of inspiration to me. A couple of years ago Steve suggested that I use my interest in history to write a book on the Internet age. I followed this advice and am truly grateful for the support he has given me. I hope that this offering will fulfill his wishes. San Diego, September 2003

Territories of Profit

chapte r

1

Introduction Communications and Innovation, Business Organization and Territory

economies and technologies are historical creations. The way human beings organize their economic activity and the technologies available to them for producing, buying, and selling emerge from historically constructed settings and evolve in conjunction with the historical process. This study tells a comparative story of economic development and technological change that takes place during two historical periods. One period encompasses the early years of the mass-production economy in the late nineteenth century. The other focuses on the formative years of the Internet economy at the end of the twentieth century. Two business firms, G. F. Swift and Dell Computer—one from each period—are the central protagonists in this comparison. These two firms reveal parallel stories of innovation and economic transformation with a set of common themes. The narrative for these two stories begins by tracing how business users of transport and communications systems learn to reorient their competitive strategies and operational routines as the technology of these systems changes. It goes on to describe how, as businesses alter their strategies and routines and assume new capabilities, they transform the organizations through which they compete and seek profit. What this study compares are the business organizations that get created as firms use new transport and communications technology to innovate their competitive activity. At the center of this comparison is a story of how transformations in business organization are part of a general process of innovation within the firm and how the emergence of innovative enterprises reshapes the geographical territory of profit making.1

2 Introduction

The Comparison This comparison of innovative business organizations is organized around a central research problem: How does technological change in systems of communications and transport enable business users of such systems to recast their profit-making activities and innovate the organizations through which they produce, buy, and sell, and how do such enterprises reshape the development of territory for economic activity? This problem, in turn, seeks to explain the relationship between four phenomena: (1) transport and communications technology; (2) innovation within the firm; (3) the organizational structure of the firm; and (4) the geographical territory where the firm operates. While in formal terms these elements are related as “independent” and “dependent” phenomena, they provide a far more compelling set of analytical threads for comparison when ordered as a sequence of questions: (1) How does technological change in systems of transport and communications transform opportunities for profit making by firms? (2) How do changes in profit-making opportunities stemming from new transport and communications systems enable business users to innovate their strategies and routines for producing, buying, and selling? (3) How do these innovations result in the transformation of business organizations through which firms compete and make profit? and (4) How do these organizations assume territorial characteristics in the way they coordinate business activity within and across geographical space? To address the central research problem and the analytical relationships upon which this problem is built, this study examines two highly innovative business users of transport and communications systems as case studies. The first case is historical and focuses on G. F. Swift, a user of the rail and telegraph system for producing, buying, and selling during the late nineteenth century. The second case is contemporary and examines Dell Computer, a user of the Internet as an infrastructure for procurement, production, and distribution. Swift is the pioneering founder of the mass-produced fresh beef industry during the late nineteenth century. Dell is the contemporary developer of custom-built personal computers (PCs). These firms rank among the most innovative organizations of their respective time periods, ascending to this shared status by responding to different communications revolutions in a similar way. Both companies used technologies of communications revolutions to create enormously innovative production-and-distribution organizations that redefined competitive practices for business activity in their own time. In changing how products were built and distributed, the organizations created by these two companies redirected the routes by which goods trav-

Introduction 3

eled within and between firms to final customers. In the process, the innovative enterprises of Swift and Dell reconfigured territories for accumulation and profit. The focus on these two firms, however, is far from a worshipful paean to individual entrepreneurial heroism. This study seeks to tell a more fundamental story about innovation and capitalist development. Within this broad frame, Swift and Dell reveal how business organizations are inherently geographical in the way they occupy territory as they coordinate their competitive activities and how capitalist development, fueled by innovation and organizational change, is a territorial phenomenon linked to the activity of firms (Walker 1988; Markusen 1994). In effect, the experiences of these two companies raise one of the most compelling issues in the study of economic development: How is geographical space for profit making (re)configured from the innovative activity of business firms? The stories of Swift and Dell uncover answers to this puzzle. In addition to the comparative focus on innovation and capitalist development, this study raises issues of contemporary economic relevance by engaging with the economies of both present and past. First, at a time when economic policy makers throughout the world elevate models of the market as the source of innovation and economic growth, the experiences of Swift and Dell reveal innovative activity to occur largely outside the parameters of established markets and the price system in both contemporary and historical settings. As the business models of Swift and Dell make clear, innovative business organizations use administrative planning and corporate power—not the market—as the mechanism for generating profit and increasing returns. Consequently, the cases of Swift and Dell suggest that policies seeking to enhance innovativeness through the creation of more open markets may very well be wrongly targeted and even misguided. Second, this study establishes a more precise picture of global economic development by removing the historical isolation in which processes of contemporary globalization are often shrouded. While the phenomenon of globalization has its historically unique elements, the current processes of global market making represented by Dell reveal striking symmetry with the continental economic empire building of the late nineteenth century represented by Swift. In seeking innovative outlets for accumulation and profit, both firms reconfigure territory in a similar pattern of building organizations that spread and reconcentrate economic activity in new ways. Consequently, the comparison of Swift and Dell has special significance in providing a vantage for understanding the broader meaning of contemporary globalization and outlining historical pathways along which the contemporary global economy is being shaped by firms.

4 Introduction Admittedly, the comparison in this study draws upon a compelling model. Twelve years ago, Paul David, in a provocative and beguiling essay (1991), suggested that a comparison between modern-day computers and nineteenth-century electric motors provided a methodological pathway for explaining the so-called productivity paradox of the 1980s and early 1990s. In his piece, David accounts for the anomaly in the productivity statistics beginning in the 1980s, when computers entered the workplace, by referring to a similar lag in productivity growth following the introduction of electric dynamos in factories of the late nineteenth century. By way of the earlier example, he reasons that a period of adjustment was necessary following the introduction of new technology before productivity gains were possible. In the process, he succeeds in linking historically unique technological artifacts and the experiences they embody, across time. David’s theory of the productivity paradox derives from a specific type of comparative history described as parallel comparison (Skocpol and Somers 1980). As one of three comparative historical logics, parallel comparative history asserts a fundamental similarity among selected cases based on their conformity to underlying theoretical propositions. In their role as research subjects, Swift and Dell function as intensive (as opposed to extensive) case studies in a parallel comparison of business firms, with the business enterprise serving as the unit of analysis. Whereas extensive case study research seeks to discover common characteristics in a population of subjects based upon a large number of observations, intensive research reveals how a causal process uncovered from an engagement with theory manifests in a particular case and how the particular case, in reflecting this causal process, is representative of a broad-based and generalizable trend (Sayer 1992, 242–43). In this book, Swift and Dell are comparatively similar intensive cases representing a general causal process—the route from communications revolutions to innovation, organizational change, and territorial transformation—that connects different historical periods. In employing this logic of similarity between cases, David’s comparison emerges as a potent example of the use of history to explain recent economic outcomes. By casting technology and the economy of the late twentieth century into what he described as a “not-too-distant mirror,” David finds answers to questions about present technology that would not have been discernible from current observations alone. In the process, he provides insights about the perils of subordinating historical memory to the technological present. David notes the human tendency to lose sight of the past when confronted by the achievements of current innovation, an affliction he vividly describes as “technological presbyopia” (David 1991, 317). To sufferers of this malady, the technological future appears closer at hand than the historical path leading to it, and the afflicted, in their neglect of the past, ex-

Introduction 5

aggerate the sense in which the present is “unprecedented” and unique. David’s analysis suggests that such fixation on the future, and neglect of the historical route to present-day innovation, leads to a truncated and ultimately superficial engagement with technology itself. In the present study, the comparison of Swift and Dell—the rail and telegraph system and the Internet, and fresh beef and custom-built PCs—overcomes such afflictions. It uses history to provide insight into the economy around us.

Findings at a Glance When applied to the cases of Swift and Dell, this historically oriented comparative approach to innovation and economic development reveals several critical findings. First, the two cases reveal communications revolutions to be the catalyst for innovations in distribution and marketing. Such breakthroughs in the socalled sphere of circulation enabled Swift and Dell to elevate the role of logistics as the source of value creation and competitive advantage. These logistics-oriented process innovations led to the creation of organizations designed to coordinate the movement of materials, semifinished goods, and finished goods from suppliers to producers and customers in new ways. While both firms did succeed in using new infrastructure systems to create new commodities—dressed beef shipped long-distance and PCs configured through Internet communication—the logistics-oriented process innovations of Swift and Dell proved as formidable as new products in transforming the economies of their respective eras. With this emphasis on distribution and logistics, Swift and Dell reveal striking symmetry in attributes and aims. Dell uses the Internet to link customer orders with procurement, production, and delivery of PCs in creating an extremely innovative “directpull,” “just-in-time” production-and-distribution network. Swift, however, created a similar organization from rail and telegraph technology, using telegraphy to link orders from retail butchers with procurement of cattle supplies, disassembly, and final marketing. Admittedly the pioneering system of Swift operated in a more extended time horizon than the continuous orderand-fulfillment flow of Dell. Nevertheless, by linking customer orders to sources of supply and eliminating intermediaries between producer and customer, Swift’s network anticipates the basic idea of Dell’s “closed-loop” organization by a century. Second, this study on the logistics-driven organizations created by Swift and Dell challenges the belief that the mass-production age created wealth from goods whereas the Internet age creates wealth from information. Such partial truths ignore the ways in which Swift and Dell developed organizations that relied on both the processing of enormous amounts of real-time

6 Introduction information and the manipulation of high-volume flows of goods. Swift built its network not only from its rail-transported fresh beef. It created an organization on the basis of telegraphic information coordinating the movement of this product from its raw material form as cattle at stockyards to its finished form as edible beef at retail butcher shops. Real-time information exchange between the primary nodes in Swift’s network—stockyards, slaughtering facilities, and branch distribution houses—shaped a process of procurement, production, and sale that was modulated daily in order to balance conditions of supply and demand in the context of a highly perishable product (U.S. Bureau of Corporations 1905, 21). Information secured through telegraph technology was as critical to Swift in capturing value from its production and distribution as transformation and manipulation of the physical product. The production and distribution of the product, in effect, and Swift’s capacity to capture value from this activity of producing and selling were completely dependent on the capacity of the firm to manage information directing these flows. Dell, in turn, not only captures value from its logistics-oriented activity by the Internet-based information it maintains to link the primary nodes in its organization. It also makes profit and distinguishes itself from competitors by the way it links Web-generated information to the physical movement of supplies, semifinished and finished commodities from supplier factories, to supply logistics centers where components are staged for assembly, to its own assembly sites. While information is essential in coordinating these logistics, executing the movement of physical materials through these organizational nodes is as critical to Dell’s success as the Internet-generated information system that underlies how the hardware in this network is assembled and delivered. Third, the organizations of Swift and Dell reveal a similar geographical tendency to spread and concentrate in a pattern characteristic of industrial districts. From the rail and telegraph system, Swift built a production-anddistribution organization extending throughout the entire United States that obliterated the formerly localized markets of beef production and consumption. As it widened routes of distribution across the entire continent, the company decentralized slaughtering away from its original hub in Chicago and reconcentrated it in other locations in the Midwest. The result was a new system of meatpacking ensembles located in the center of the country, serving a vastly extended set of marketing routes. This new structure of production and distribution succeeded in articulating a national market space. Dell is building an organization on the basis of Internet technology with a similar geographical configuration. It has spread and decentralized computer assembly activity from its original hub in Austin, Texas, to locations around the world. At the same time, it has created concentrations of assembly activity in selected localities to build and sell products for the

Introduction 7

regions where assembly sites are located. In this way, Dell is playing a critical role in the development of high-technology industrial districts. In these place-based concentrations of manufacturing, both firms rely on critical relationships of territorial proximity between key nodes in their networks— stockyards, disassembly facilities, and branch distribution houses in the case of Swift and supplier factories, supply logistics centers, and assembly facilities in the case of Dell.2 The two companies actively shape these relationships of proximity in order to manage and control the movement of materials between these nodes and to execute their direct-pull high-velocity systems of production and distribution. In the process, both firms have emerged as agents in reshaping territory for profit making. Nevertheless, the pattern of elongation and concentration created by the two firms reveals a clear and obvious geographical difference. In the case of Swift, the elongation of distribution routes and the decentralization of cattle disassembly created a nationally based set of regionally concentrated production complexes. In the case of Dell, the spread and decentralization of PC assembly has created a globally based set of regional production ensembles.Yet, at the same time, the pattern of spread and concentration, while configured differently across time, appears to be the fundamental territorial manifestation of innovation, expansion, and the capitalist process. Finally, this comparison of Swift and Dell clears new ground in understanding firms as business organizations and networks that deploy mechanisms of power, coordination, and control as the basis for innovation. First, this study reexamines and clarifies the idea that economic activity and organizations in the current period are distinguished by the phenomenon of networking. Seemingly incontrovertible, the notion that firms and economic activity have only recently evolved into networks is an historically static formulation and consequently problematic. What, for example, is the appropriate designation of the myriad intermediate forms of subcontracting and networklike trade relationships in existence before the current period, observed by theorists from Ronald Coase (1937) to G. B. Richardson (1972)? Indeed, if there is one basic insight revealed in the history of economic life, it is that the activities of producing, buying, and selling have always occurred in networks (Braudel 1977, 1979). Far more critical in assessing whether the current period is one of networking is the task of identifying how the network linkages used by firms and individual economic actors to organize production and trade have evolved in different historical periods and uncovering the attributes of these connections—localized, regional or global, interfirm or intrafirm—inherent in economic activity. Business organizations, by their very nature, are networks with distinct characteristics that change over time. The integrated intrafirm organization of Swift is no less a net-

8 Introduction work than the dis-integrated interfirm organization currently being created by Dell. Furthermore, networks are not egalitarian substitutes for the hierarchical power relations of vertically integrated organizations (Barley and Kunda 2001). They reorganize relations of power. As networks evolve, they recast adjacent operations in the creation of products and services and redistribute where, among these activities, firms are most able to capture profit. Different steps performed within networks thus become contested with economic rents accruing to those firms capable of performing the most profitable activities and defending such rent-generating positions. Networking, in effect, retains the basic characteristics of linkage and connection common to economic life. At the same time, contemporary networking reproduces relationships of competition and cooperation between firms along with contests for power at the core of capitalist development. The proliferation of interfirm relationships within regions and across continents that seems to be driving the idea of the current period as one of networking is not entirely new. It is more accurately networking in a new form. Second, this comparison decisively reveals that while Dell and Swift reflect undeniable differences in structure, the integrated organization of Swift and the interfirm organization of Dell share the same need for mechanisms of power and control to mitigate risk in coordinating adjacent steps of procurement, production, and distribution and in innovating the logistics of these operations. Such mechanisms of power and control are fundamental to the innovative enterprise. In the case of Swift, these mechanisms, achieved through rail and telegraph technology, are exercised within the boundaries of the firm through the ownership of assets, that is, through vertical integration. In the case of Dell, these mechanisms, achieved through Internet communication, are exercised over other firms that lie outside the organizational boundaries of Dell but within the network of the PC maker. Dell is compelled to use such control in achieving what it calls virtual integration with its suppliers and logistic partners in order to coordinate the high-speed information and material flows needed in managing the logistics of its justin-time procurement, production, and distribution system. Dell interacts with other firms in its network not through markets and the price system but through an enforced structure of relationships—using the Internet as a tool to help facilitate this control. Consequently, this study takes issue with prevailing views of interfirm production networks as the organizational embodiment of ascendant market forces and presents an alternative view of interfirm networks as innovative organizations equally dependent on mechanisms of power, control, and administration used by vertically integrated firms such as Swift. Far from affirming the current period to be a revolution in production based on market

Introduction 9

coordination within interfirm networks, the comparison of Swift and Dell reveals that corporate power and administrative coordination lie at the foundation of the innovative enterprise and are deployed in different organizational settings. This underlying characteristic of the innovative firm, in turn, and the ongoing role of innovation in modern capitalist development, establish bridges between the late nineteenth and the late twentieth centuries.

Theoretical Framework Three distinct but often overlapping sets of literature provide the theoretical context for the comparison in this study. innovation as history The first set of literature, traceable to the influence of Karl Marx, employs a fundamentally historical approach in examining the phenomena of technological transformation and innovation within the firm.3 Developed systematically in the work of Joseph Schumpeter (1942, 1947) and elaborated more recently by theorists influenced by his notion of evolutionary economic change, this literature seeks to uncover the sources of innovation and its impacts on economic development within and across historical periods. In this framework, innovation is conceived more broadly than the accumulation of discrete inventions and new technologies. Innovation is the deployment and transformation of inventions into commercially viable products and profit-making activities and the diffusion of these new products and processes throughout the economy. It involves what Schumpeter described as the “creative response” of entrepreneurial firms and the adaptive response of other firms that, in trying to compete with innovators, imitate the original innovation. In seeking to explain how this process of invention, innovation, and diffusion occurs, theorists in this tradition focus on the process of learning within the firm. This process of learning to compete differently and choosing how to implement a new vision of profit making is what leads to the creation of new strategies, products, routines, and business organizations. These activities of learning about new profit opportunities and selecting alternatives for capturing profit in new ways transform patterns of competition and enable firms to create different trajectories of growth and development. This body of theory is used to position Swift and Dell as innovative firms. the firm as business organization and network The second set of literature examines how firms, as business organizations, coordinate their activity in both internal and external networks. The-

10 Introduction orists within this literature focus on two primary attributes of business organization. One group of theorists examines business enterprise as organizational linkages within and between firms. Such linkages result from the choices of firms on how to undertake and divide up the various activities in producing and selling a good or service and stem from a search by firms for “competitive advantage.”4 In this way, organizational linkages within and between firms reflect operational decisions on competing. Such linkages establish boundaries between firms and lie at the core of theories on business organization in terms of the degree to which firms internalize various economic activities and are integrated or the degree to which firms transact with other firms across markets for these activities and are dis-integrated. Business organization, and the networks through which organization becomes operational, whether intrafirm reflecting integration, or interfirm reflecting dis-integration, are thus the outcome of how firms choose to compete in economic space. A second group within this tradition extends the idea of business organization into the realm of territory. Using insights from the first group, these theorists reveal business organization to be inseparable from geographical organization and seek to identify how business organizations reflect geographical linkages in the coordination of economic activity, thereby embedding the firm in territory (Walker 1988, 385). This process occurs in the way firms organize the locations of their assets and coordinate the flows of activity, material, and information within and between these assets and between these assets and those of other firms in their production-and-distribution networks. The pivotal concern of these theorists is how the territory through which business organizations operate is constructed and reconfigured. While the starting point of this literature is the structure of the firm as a competitive unit in economic space, theorists in this tradition also emphasize how forms of business organization occupy a second analytical realm—geographical space. This body of theory is used to position Swift and Dell as creators of business organizations with attributes that occupy an economic realm of competition and a spatial realm of territory. communications as revolution The final set of literature focuses on the role of new communications systems in promoting discontinuities in the historical process. This phenomenon of communications as a revolutionary force, however, is not some inexorable driver of history and economic development. Innovative users of communications systems employ agency to discover attributes of communications revolutions. Through a process of learning, firms as historical actors put these attributes to use in recasting routines, organizations, and territory. Pioneered as an historical concept by Robert Albion, the communications revolu-

Introduction 11

tion represented a new era of “speed” in moving information, goods, and people across space (Albion 1932; John 1994). Although Albion acknowledged this phenomenon to be the prelude to the “Machine Age” and “Big Business,” his work elevated the communication revolution’s independent emergence. Nevertheless, his insight into the role of new transport and communications systems as prologue to transformations in economy and society established an enduring legacy. Theorists from Alfred Chandler (1977) to Manuel Castells (1996) have acknowledged the centrality of communications and transport in the process of historical change. Moreover, Albion’s depiction of the communications revolution as an ongoing phenomenon finds support in the more recent representation of communications as a “carrier wave,” a fundamentally unified historical pathway of communications breakthroughs beginning with the telegraph, demarcating different periods of innovation (Hall and Preston 1988). This notion of the communications revolution as “catalyst” and “carrier wave” is the foundation for the connections in this study between the competitive economic environment, the process of innovation and organizational change in the firm, and territorial transformation. It is what ultimately links the experiences of Swift and Dell.

The Argument From a synthesis of these literatures, this study illustrates how firms, in confronting revolutions in transport and communications technology, reshape the development of territory for profit making. Four key concepts provide the threads for this connection: (1) the communications revolution; (2) innovation within the firm referring to the creation of new products, the reorganization of routines to make new products, and the exploitation of new markets for selling new products and carrying out new routines; (3) the business organization of the firm consisting of the organizational and geographical linkages created by the firm to coordinate innovative activity; and (4) territory, referring to the geographical pattern of the informational and material flows coordinated by the business organization to carry out innovative activity. In this argument, the communications revolution is connected to territorial transformation through the process of innovation within the firm and the changes in business organizations made by firms as they undertake innovative activity (represented schematically in Figure 1.1). Revolutions in communications provide a pathway to innovation by enabling firms to recast the routes by which products circulate from their origins as raw materials, through the process of fabrication, to their final destination. As firms change their organizational structure to implement processes for redirecting these routes traveled by products from production to consumption, they reshape the territory in which they operate. It is in this role as catalysts for in-

12 Introduction

i

Territorial Transformation

~

! l l_ _ _ _ _ _ _ _ __

Communications Revolution

t Business Organization Innovation

figure 1.1 From Communications Revolution to Territorial Transformation

novation in logistics and business organization that communications revolutions have a defining impact in reconfiguring territories for profit making. theoretical contours From these outlines, the argument in this study begins from the concept of the firm framed roughly forty years ago by Edith Penrose (1959). For Penrose, the firm is a profit-seeking repository of human and material resources bound within an administrative framework. The mission of the firm is to grow by transforming its resources into profit-making activities. Such activities include creating new products and production processes and exploiting the market environment in new ways to accumulate wealth. Competition and the profit system compel firms to grow and accumulate profit in order to survive. Growth occurs when firms compete differently after acquiring new knowledge about profit making and adding what they have learned to their base of resources. When firms learn to compete differently, they acquire new capabilities derived from three types of knowledge—technical, functional, and managerial (Chandler 2001, 2–3). Technical capabilities are learned from applying broad-based scientific and engineering knowledge to products and processes while functional capabilities refer to knowledge specific to a particular product. Managerial capabilities refer to knowledge essential to coordinate the operations of the enterprise. Enhancement of capabilities is what enables firms to provide the market with new goods and services created from new processes and to exploit new territories in which to undertake this activity.

Introduction 13

Acquisition of new capabilities, however, compels firms to modify their organizational structure. Capabilities and organization evolve interactively in creating new routines. Consequently, the augmentation of capabilities is an organizational phenomenon. This enhancement of capabilities and the creation of new business organizations are the sources of growth in the firm. They represent an expansion in the firm’s resource base, not simply an increase in resources from a mobilization of more production factors. In this way, innovation, based on learning, and organizational change are integral to the growth process at the foundation of capitalist development. Firms are historically conditioned entities. They learn to make profit within environments in which the basic parameters for capital accumulation—technical, organizational, legal, political, and social—are essentially given. In these environments, established business practices, along with available technologies for profit making, and institutional frameworks condition the strategic, operational, organizational, and territorial choices of firms for accumulating. From these choices emerge products, routines, business organizations, and market structures. Although these environments condition competitive choices, they are at the same time constantly evolving and creating new opportunities for organizational learning. Consequently, firms’ strategic, operational, organizational, and territorial choices for accumulating profit are always open-ended and contingent. Learning in such environments is thus a function of both structure and agency. Similar to individual historical actors, firms “make their own history,” not exactly as they please but from “circumstances given and transmitted from the past” that condition both the range of choices for competing and the opportunities for competing differently.5 Among the most disruptive historical forces affecting the environment of profit making and competitive choice is the phenomenon of technological change in systems of communications and transport and the eventual buildout of such new systems. The railroad/telegraph system and the Internet represent different historical manifestations of this phenomenon. Two distinct though often overlapping groups of firms participate in the creation of communications revolutions. Spearheading this phenomenon are builders of transport and communications systems. This group encompasses an array of actors including inventor entrepreneurs, investors, government, and firms that undertake the build-out of infrastructure systems. Such infrastructure, when put into use, creates new systems of access across and within space, enabling people, merchandise, and information to circulate over distance and within distances of close proximity in new ways. These new systems of circulation and access alter conditions for accumulation and profit making. In the process, new infrastructure systems beckon to a second, more numerous group in the economy.

14 Introduction This second group consists of business users of transport and communications systems. In experimenting with the new technologies, these firms actually play an integral role in shaping the deployment of the new infrastructure itself. As these users learn to deploy new infrastructure in their business models and exploit the new systems of access and circulation created by such technologies, they innovate how they compete. This activity of innovation through learning is marked most decisively by a process of transformation in the strategy and structure of the business enterprise (Chandler 1962). Swift and Dell represent two such users. They exploit the profit-making opportunities of communications revolutions and create new standards for competing on the basis of changes in strategic outlook and innovations in business organization. Communications revolutions change the profit-making environment for transport and communications users by transforming a fundamental element in the economic system—the geographical structure of markets. Systems of transport and communications influence the territorial configuration of markets in two ways. First, technologies of transport and communications act to delimit market boundaries (Irwin and Kasarda 1994; Christaller 1933). They do so by influencing the costs to economic actors of producing and trading across distance. Market boundaries emerge at those locations where “costs of transfer”—the costs to economic actors of moving goods or securing information across territorial barriers—drive the prices of goods and services beyond their original value (Ohlin 1933, 100). The size of markets is thus dependent on the costs to and capacity of market actors to produce and exchange goods and services over distance and to communicate information needed to organize these activities (Du Boff 1980, 478). In this way, market boundaries correspond to territorial systems of access for producing and trading. Such systems of access are shaped and reshaped by the build-out and deployment of transport and communications systems and the ways that business users of this infrastructure actually deploy it in economic use. Second, in addition to delimiting boundaries between different market areas, technologies of transport and communications reshape systems of access within the boundaries of markets. It is users, however, in experimenting with the infrastructure, that act as the agents in this transformation. As routes and means of access within markets become reconstituted from breakthroughs in transport and communications, such new access systems alter the costs of economic activity and enable users to transform business routines in close proximity. In this way, technologies of communications and transport shape and reshape systems of market access across and within geographical space, thereby changing the environment of profit making for users of these technologies.6

Introduction 15

Embedded most fundamentally in the geography of markets and systems of market access are relationships between business firms and economic activities structured around time and space. In all periods of capitalist development, control over time and space is a centrally important strategic, operational, and organizational problem for the firm (Schoenberger 1997, 12; Harvey 1996, 240–41). Businesses are constantly engaged in reshaping their strategies, routines, and organizations in an effort to overcome the temporal and territorial barriers to accumulating profit. Communications revolutions create possibilities for firms to control time and space in economic activity differently so that what is inefficient or even impossible as a business model at one point in time may be viable as a profit-making venture in another historical moment. When firms grasp the opportunities for crafting new routines from breakthroughs in communications, and as firms create new forms of organization to implement such routines, their efforts lead to changes in the timing cycles and territorial arrangements of business operations. In this way, changes in business routines and changes in the economies of time and space are mutually reinforcing. A change in routine shifts the timing and territorial spacing of business operations. Changes in routines, in turn, present the firm with new challenges of control over these recalibrated and reconfigured time and space relationships. As the firm learns to use new systems of communication in creating innovative routines, and as the firm deploys such systems in assuming control over the time and space relationships recast by these routines, the innovative organization created in this process reconfigures the territory for economic activity and profit making. Communications revolutions, in effect, are control revolutions.7 They change the environment of profit opportunities by providing firms with different and more intensified types of control over space and time in economic activity. Such control in turn enables firms to create new routes to efficiency in the form of more innovative routines. As firms discover these routes to more efficient routines, they are compelled to assume new and different capabilities in order to capture the profit—the rents—from doing things differently. In recasting business practices and assuming new capabilities, the innovative firm reorganizes the structure of its enterprise and in the process reshapes the territory in which it assumes control over newly crafted routines. Not all firms are equally successful in learning about profit opportunities in economic environments modified by new transport and communications systems. As communications revolutions reshape markets and enhance systems of control over space and time in business activity, very few firms are able to grasp how to profit differently from the new economic environment. Such variation stems from the fact that the choices of strategies, routines, and business organizations made by firms do not derive from so-called per-

16 Introduction fect information of the most profit-optimizing pathway available in the market as assumed in rational choice models of human action. Variations and heterogeneity characterize strategic, operational, and organizational choices, even among firms confronting identical information and notions of profit opportunities (Dosi 1997, 1531; Metcalfe 1998, 35; Lamoreaux, Raff, and Temin 1999, 6–8). While most businesses adapt to the innovations of others, a few firms succeed in making choices that result in what Schumpeter described as creative responses in economic history. Swift and Dell are two such firms. empirical outlines Communications revolutions provided Swift and Dell with the technological tools to develop process innovations as the basis for profit making. While they used new transport and communications systems in conceiving new products, the innovative advance of both firms was the creation of organizations not only for producing but more importantly for distributing these products. Swift and Dell learned to deploy the technologies of communications revolutions in essentially becoming innovative logistics firms. In both cases, transformation into logistics-oriented enterprises was driven by a strategic vision of profit-making opportunities existing in a more direct relationship to the final customer. Such a shift in outlook derived from an understanding, however vague initially, of the potential for new communications systems to help forge this more direct route to the customer and bypass intermediaries entrenched in the two industries. The direct systems of marketing created by Swift and Dell served as their primary source of profit. Operationally, this direct route posed enormous challenges for the two firms. To solve the problems of making and marketing mass-produced fresh beef and Internet-customized mass-produced PCs and reaching the final customer in a more direct way, Swift and Dell had to reinvent logistics systems for the entire circuit of procurement, production, and distribution. This process of discovery occurred in both cases through experimentation and use—learning by “using.” In this learning process, resolution of one problem invariably uncovered other logistical difficulties that both firms had to overcome in moving toward a final and workable operational and organizational design.8 The outcome of this learning process was an enhancement of organizational capabilities and creation of enormously innovative logistics systems for mass-producing and mass-distributing products directly to buyers. These process-oriented innovations not only succeeded in establishing new standards of competition and new pathways of profit making in the meatpacking and PC industries. The innovations of both companies also diffused

Introduction 17

across different economic sectors, resulting in new patterns of growth and development throughout the economy. In both cases, communications revolutions—the rail and telegraph and the Internet—played pivotal roles in securing this innovative pathway of distribution to the customer. Organizationally, this direct path to the customer assumed two outcomes. As they created high-volume direct organizations, Swift and Dell faced critical choices with respect to systems of coordination and control over adjacent steps in procurement, production, and selling. These choices centered on the degree to which firms absorb sequential operations in procurement, production, and selling and the degree to which they contract with other businesses in allocating these tasks. In the case of Swift, the route to a more direct relationship with the customer occurred through vertical integration. In the case of Dell, this route occurred through an interfirm structure of organization, but one with a level of functional and operational integration sufficient for Dell itself to refer to its own organization as virtual integration. In spite of these structural differences, the enterprises of the two firms share a fundamental attribute of coordination and governance. In both organizations, the market is not the mechanism of coordination for organizing adjacent steps of procurement, production, and selling. Instead, both Swift and Dell used forms of corporate power and administrative planning to coordinate and control the activities within their networks, both inside and outside the formal boundaries of the firm. Consequently, although their formal structures differ, nonmarket mechanisms of coordination and control are fundamental to both firms. These nonmarket mechanisms proved critical to Swift and Dell in forging their direct path to the customer and securing an innovative route to profit making. Territorially, the organizations of the two firms that emerged from a more direct engagement with the customer reveal a similar tendency to extend and concentrate circuits of procurement, production, and marketing in new ways. This geographical pattern of spread and concentration contributes to the formation of industrial districts—the creation of new industrial spaces—and is fundamental to the process of innovation and capitalist development. As firms seek innovative modes of accumulation and acquire the organizational capabilities to implement these innovations, they reshape the territorial theaters in which the accumulation process takes place. In this context, firms, through planning and power, reconfigure landscapes of economic geography. They create differently ordered relationships of proximity between various economic assets under their control, both inside and outside their own boundaries, and recast flows of economic activity connecting these assets. Such reconfiguring is critical in implementing new and innovative economic routines and mastering the novel demands of time and space embedded in these innovations. This pattern of spread and concentration,

18 Introduction accomplished through the power of agency within the firm, links the experiences of Swift and Dell across time. Nevertheless, the two organizations and the territories they exploit reveal a paramount difference. The vertically integrated organization of Swift spreads and concentrates its innovative logistics activity over a fundamentally continental market space. The interfirm organization of Dell extends and concentrates its innovative logistics activity over a global market space. Whereas Swift played a central role in helping to articulate a unified national market, Dell is creating production complexes helping to articulate an emerging system of globalized markets. The organization of Swift consisted of three primary nodes: (1) stockyard facilities, where cattle supplies were stored and where Swift secured its cattle raw materials; (2) disassembly facilities, located immediately adjacent to the stockyard sources of supply; and (3) branch distribution houses, located throughout the country, where the firm shipped its product in order to supply retail butchers with fresh beef. In learning to coordinate the activity between these nodes and reach the customer directly, Swift created an early type of direct-pull just-in-time system of production and distribution. Orders taken at branch houses from retail butchers were telegraphed daily to Swift headquarters and provided the source of demand for procurement of cattle supplies. Stockyards served as warehouses for inventories of cattle supplies that Swift would purchase and “pull” into slaughtering plants for disassembly in accordance with orders received. In this way, Swift created an early, if rudimentary, form of product customization and differentiation on the basis of standardized cattle grades and cuts. Cattle pulled into slaughtering factories on the basis of order demand were butchered and sent to branch houses, where the various cuts and grades requested by retail butchers were distributed. One of the most critical core capabilities of this direct-pull system mastered by Swift was the balancing of supply and demand in close to real time, made possible as a result of the railroad and telegraph. Spatial proximity to raw material inventories was critical to the directpull system of production and distribution pioneered by Swift. Slaughtering facilities were located by the company adjacent to stockyards in a configuration linking sources of cattle supply and cattle disassembly. These slaughtering facilities, opened by Swift at other stockyard sites in the Midwest, represented new points of concentration in beef production and distribution. Branch houses, on the other hand, dispersed the product over a wide territory. In this way, territorial spread and concentration coexisted as part of Swift’s profit-making enterprise while Swift acted as the agent in configuring these relationships. As it built this national network of slaughterhouses and branch houses, Swift integrated into its own organization virtually all of the steps from pro-

Introduction 19

duction to final marketing of fresh beef and coordinated these steps through internal systems of administrative planning. In assuming ownership of these various functions, Swift eliminated—disintermediated—a large layer of the traditional wholesalers in the beef trade, enabling the firm to forge its more direct path to the buyers of beef. This organizational innovation of disintermediation is what provided the direct path to the customer. It was also the foundation of the company’s direct-pull system and its source of competitive advantage. The railroad and telegraph provided the infrastructure for coordinating this direct-pull production-and-distribution system built on disintermediation, territorial spread, and concentration. Dell’s production-and-distribution organization also consists of three primary nodes: (1) assembly plants, where Dell configures finished systems; (2) supplier factories, where components are manufactured; and (3) supply logistics centers (SLCs), located next to assembly plants, where components are stored as inventory. In learning to coordinate the activity between these three nodes and reach the customer directly, Dell, much like Swift, has successfully eliminated layers of traditional intermediaries and has created what is arguably the most innovative direct-pull just-in-time system existing in the world today. Customer orders processed through the Web initiate a process of pulling only those components from supply logistics centers into Dell assembly sites needed for configuration into finished products to fulfill orders received. Supply-and-demand balancing and the direct closed-loop relationship to the customer are the foundations of this just-in-time pull system and the source of the company’s competitive advantage over other PC firms. The Internet provides the infrastructure to forge this direct path to the customer and undertake the challenge of balancing supply and demand in real time. Like Swift, Dell is compelled to organize critical relationships of territorial proximity between its assembly operations and its sources of component supply in order to manage the logistics of its just-in-time direct-pull system. As a consequence, Dell has established supply logistics centers, where components are stored as inventory within twenty minutes’ driving time of assembly facilities at each of its six different computer assembly locations. Suppliers, through third parties, operate these warehouses and assume the inventory and carrying costs. In these warehouses, components originating from supplier factories located throughout the world are stored and “pulled” into Dell’s assembly plants as they are needed on a just-in-time basis. Internet messaging between Dell, supplier factories, and SLCs is the foundation of this direct-pull system. The virtually integrated organization of Dell, built upon relationships of cooperation and power, is the agent for implementing these relationships of proximity. Although formally separated from the suppliers in its network and

20 Introduction dependent on the external capabilities of other firms to help produce and deliver its products (Langlois 1990, 1992b), Dell is nevertheless compelled to enforce certain types of control over these other companies to ensure that its just-in-time system operates efficiently. As a consequence, Dell establishes these relationships of proximity organizationally, through mechanisms of administrative control, with its suppliers and network partners. Territorial relationships of proximity, created organizationally through virtual integration, are among the most critical value-creating assets in the Dell organization. Dell has extended and decentralized this direct-pull just-in-time production-and-distribution system across four continents. At the same time, the PC maker has chosen six regional locales in which to concentrate these operations. Global spread, regional concentration, and localized relationships of territorial proximity form a patterned geography of just-in-time production-and-distribution complexes fused together through Internet communication and virtual integration. Parallels in operational routines, organizational control, and territorial reconfiguration created by Swift and Dell are neither imaginary nor accidental. In very different historical contexts, Swift and Dell confronted similar underlying problems related to issues of control and coordination in the search for competitive advantage. The disruptive capacity of communications revolutions enabled both firms to discover profit-making opportunities in recasting the logistics of production and distribution and controlling these logistical routines in new and innovative ways. This study seizes upon patterns in this route from communications revolutions to innovation in routines and systems of organizational control to territorial reconfiguration in an effort to uncover underlying rhythms in capitalist development.

Why Swift and Dell? Despite the years that separate them and the different products that define them, both Swift and Dell reveal stories with striking symmetries that link communications revolutions and business innovation across time. First, Swift and Dell are comparable as two of the most innovative firms in their respective time periods. Both began as upstarts in industries with much larger, well-established companies. Both succeeded in outcompeting their older, larger rivals and transforming existing business practices in their respective industries. Both used breakthroughs in transport and communications to create process innovations linking procurement and manufacturing to distribution systems for reaching the final customer directly. These innovations redefined standards of competition in the meatpacking and personal computer industries as well as other industries in the economy of their respective periods. The organizations pioneered by Swift and Dell are, in ef-

Introduction 21

fect, paragons of innovation deriving from communications revolutions, one born with the rail and telegraph economy, the other the progeny of the Internet economy. Other similarities also link the innovations of these two companies. While dressed beef and personal computers may appear oddly matched, the two commodities actually share common attributes. Although they had already emerged as new products when Swift and Dell began to do business, both firms succeeded in transforming these products through new systems of distribution. As a result of the innovations in production-and-distribution networks, both products became accessible, affordable, and mass-consumed. Perhaps most important, however, dressed beef and custom-created personal computers share a fundamental attribute as “perishable” goods. They have a limited shelf life, after which they start to devalue and essentially spoil. This shared quality of perishability played a decisive role in motivating both Swift and Dell to transform the channels of distribution in the beef and PC industry and develop their innovative organizations for making and selling these products. The Swift case covers the initial years of the company from 1875 to roughly 1903. These dates encompass two significant developments in the American economy. First, this period witnessed the completion of a nationally integrated and standardized rail and telegraph infrastructure in the United States. Second, this period marked the appearance of the mass-production economy and the large-scale integrated industrial corporation connected to new production-and-distribution systems pioneered by firms such as Swift. Infrastructure, firm structure, and market structure evolved together during this period. The year 1903 as an end point is also not arbitrary. As the culmination of the first great merger wave in American history, this date, in most accounts of the period, brings the initial heroic period of the massproduction economy to a close. Whether by chance or fate, it also marks the date when Swift surpasses all of its competitors and becomes the largest meatpacking firm in the country and even the world. The Dell case covers the firm from its founding in 1984 and carries the story forward to late 2001. These dates also frame two critical developments. First, this period witnesses the creation of a mass market for the personal computer. Perhaps more significantly, the latter years of this period mark the development of the Internet as a communications and commerce system. Similar to the advent and expansion of the rail and telegraph, the emergence of the Internet as a commerce system has enabled firms to use the new infrastructure for producing and selling goods and services and coordinating business operations in entirely new ways. Dell has managed to reorganize its business model for producing and selling PCs in responding to the opportunities presented by the Internet. Much like Swift, Dell has used the new

22 Introduction technology to assume a position of first rank in its industry. The end point in this story of Dell is also not arbitrary. It was in 2001 that Dell ascended to the position of the largest PC firm in the world. Finally, the choice of Swift and Dell as case studies is the result of unique research opportunities presented by the two firms. Surprisingly, while there are a number of studies on meatpacking in the late nineteenth century, notably the work of Mary Yeager (1981), Margaret Walsh (1982), and Louise Carroll Wade (1987), and numerous references to the innovations of Swift in both general and specialized economic histories of the period (Chandler 1977; Cronon 1991), there is only one scholarly work on the firm of Swift itself, a dissertation written fifty years ago (Unfer 1951).9 The present study revisits these works along with primary material on Swift in an effort to uncover how the company created its pioneering dressed beef network from the rail and telegraph and the economic development impacts of this innovation. Dell, on the other hand, presents a different type of opportunity. With its modest beginnings and meteoric rise within the PC ranks, Dell has generated a type of modern business folklore. As a consequence, the company has garnered a large following in the business and trade press during the past five years. CEO Michael Dell added to his firm’s reputation with his own book about Dell and the business model he created (Dell and Fredman 1999). A plethora of how-to books on “business the Dell way,” mostly repeating insights from Dell’s own book, have followed (Saunders 2000). There is, in effect, an abundance of available information on the firm. Nevertheless, there are few scholarly studies of Dell (i.e., Curry and Kenney 1999; Kenney and Curry 2000a; Kraemer, Dedrick, and Yamashiro 1999; Albers 2000; Kraemer and Dedrick 2001). With the exception of the study by Toni Lynn Albers, who worked as an intern at Dell, these works reveal limited access to sources inside the company. As a consequence, much remains unknown about the specific mechanisms of Dell’s logistics-oriented organization. The present study overcomes this gap. Through interviews with Dell managers in supply chain operations, along with interviews with several key suppliers of Dell, this study provides an inside view of logistics at a firm shaping some of the most defining characteristics of economic globalization. Both familiar and new facts serve to position the two firms within an historically comparative framework. The comparison that follows reveals new insights about the two firms, the transport and communications systems they used, the innovations they created, and the past and present economies they helped to transform.

Introduction 23

Plan for the Book This study consists of three parts that follow this introductory chapter. Part 1, consisting of Chapter 2, establishes the theoretical framework for the study. This chapter critiques literature on innovation, business organization, and communications revolutions and from this critique sets out a taxonomy of the route from communications revolutions, to innovation within the firm, to organizational change, to territorial transformation. This taxonomy informs the comparison of Swift and Dell. Part 2, encompassing Chapters 3 and 4, focuses on the railroad and telegraph and the case of the G. F. Swift Company. Chapter 3 examines how rails and telegraphy created preconditions for Swift’s innovative enterprise in terms of the geography of markets and the pattern of urbanization during the mid to late nineteenth century. It outlines how the rail and telegraph system opened markets for more long-distance interregional trade and created a system of cities in which manufacturing and the consumption of manufactured goods became concentrated in large urban areas. Swift, in effect, not only relied on the rail and telegraph for the operation of its innovative organization but also built its organization from the interregional markets and entrepôts of consumption concentrated in cities that the rail and telegraph system helped to establish. Chapter 4 examines how Swift used the rail and telegraph infrastructure to create a mass-production and mass-distribution organization for fresh beef that revolutionized the meat industry. It analyzes how this innovation established new patterns of territorial development in the economy of the late nineteenth century that spread business activity nationally while at the same time concentrating development in new places, notably Chicago and other cities in the American Midwest. This chapter also examines how Swift was forced to confront the politics of interstate commerce in order to protect the far-flung market for beef it had engineered. Part 3, comprising Chapters 5 and 6, focuses on the Internet and the case of Dell Computer Corporation. Chapter 5 parallels Chapter 3, outlining how the Internet evolved from a communications system to an infrastructure for commerce and how the phenomenon of Internet commerce established the foundations for Dell’s innovative advance. Chapter 6, paralleling Chapter 4, examines how Dell is using the Internet to organize what is arguably the most innovative production-and-distribution organization of any current manufacturing firm. The research for this chapter focuses on the operational and organizational mechanisms Dell used to create its direct-pull procurement, production, and distribution organization and the role of Internet communication in providing the foundation for the innovative advances in this system. This chapter also describes the territorial outcomes of

24 Introduction Dell’s organization and how it has emerged as a paradigm of economic globalization. Chapter 7 subjects the two cases to comparative scrutiny, examining the innovations of Swift and Dell as uniquely tied to the two periods in question. At the same time, this chapter searches for common meaning in the innovations and enterprises the two firms created. This chapter adds to several current cross-disciplinary debates on the significance of the so-called information society, the nature of the firm and its relationship to innovation and economic development, the role of the business organization in the modern economy, and the role of logistics or, more broadly, the so-called sphere of circulation in production innovation and economic growth. In intervening in these debates and framing conclusions about the two cases of Swift and Dell, Chapter 7 also uncovers the meaning of the communications revolution from past and present. At the core of this revolution is a story of what occurs when the journey traveled by commodities from production to consumption assumes a different character and takes a different route. How fresh beef becomes mass-produced, travels across a continent in breaking the boundaries of localized markets, and in the end takes a more direct path in arriving at the butcher, and how the personal computer is custom-assembled in large volumes and travels across the globe in arriving on the desktop, and the economic development consequences of these routes, are the themes of the story that follows.

chapte r

2

From Communications and Innovation to Business Organization and Territory A Synthesis

Such phrases as the Romantic Movement, the Mercantile System, and the Second Hundred Years’ War have been of real value in helping students visualize and coordinate historical movements and influences. If there were a board of historians empowered to pass upon such labeling, one might propose to them another phrase—the “Communication Revolution.” Robert Albion, 1932 Whenever the economy or an industry or some firms do something that is outside the range of existing practice, we may speak of creative response. . . . Creative response changes social and economic situations for good [and] is an essential element in the historical process. Joseph Schumpeter, 1947 Firms are not islands but are linked together in patterns of cooperation and affiliation. . . . Co-operation may come close to direction when one of the parties is clearly predominant. G. B. Richardson, 1972

From Communications to Territory Technologies of transport and communications play a central role in economic life. Alongside financial systems and energy systems, transport and communications systems provide the basic infrastructure upon which economic activity takes place (Horwitz 1989, 7). Transport and communications technologies shape the parameters of efficiency by recalibrating the costs of basic elements in producing, buying, and selling. These elements include securing information from the market on opportunities for, and risks of, engaging in economic activity; shipping products across distance as well as

28 A Synthesis within distances of close proximity; and communicating with other agents in and across markets on opportunities for, and terms of, business activity. Historically, new technologies of transport and communications have promoted more efficient types of economic activity by creating new and less costly systems of access for firms both across and within markets. Such different forms of access have reconfigured the territorial limits of markets, blurring the boundaries of once-separated market areas while reshaping the spatial relationships between economic actors and activities within markets. In these environments where the geography of markets gets reshaped by breakthroughs in transport and communications and possibilities for economic activity are redefined, firms are presented with opportunities to change the way they seek profit. Such are the basic outlines of the stories at Swift and Dell. During the last half of the nineteenth century and the final years of the twentieth century, rail and telegraph technology and Internet technology created communications revolutions that assumed the role of what Schumpeter described as “leading sectors” in the economy (Schumpeter 1939). More than simply high-growth industries, these lead sectors ignited more widespread patterns of innovation among firms in both periods. This pattern included more than new products and new production processes. Businesses used breakthroughs in transport and communications to reorganize the structure of the firm itself. They developed pioneering forms of industrial governance that resulted in entrepreneurial types of business organization. These enterprises, in turn, used the power of administrative coordination rather than markets in creating distinct geographies of profit making. This chapter develops a taxonomy of this route from communications revolutions to innovation, organizational change, and territorial transformation across different historical periods. This taxonomy reveals how the innovations of Swift and Dell are not random acts of individual initiative but instead are catalytic elements in a broader process of economic development understood in the context of three often-distinct literatures on innovation, business organization, and communications revolutions. What follows is a critical examination of these three lineages in which communications and innovation, business organization, and territory are synthesized and recast to create a framework for understanding the role of the innovative firm in the development of the capitalist economy.

Innovation and Technological Change It is indeed an irony how the notion of innovation and the idea of markets have somehow become inextricably linked. So strong is this association that

A Synthesis 29

according to orthodox economic policy prescriptions, creation of the latter begets the former. While it seems incontrovertible that capitalist development occurs through a market process punctuated by instances of innovation, it is also true that these two concepts, innovation and markets, in many ways share at best an uneasy mutual affiliation. That the capitalist economy is driven fundamentally by the process of innovation rather than an equilibrated allocation process of market clearing is perhaps the greatest legacy left to economic theory by Joseph Schumpeter. In contrast to neoclassical economists, Schumpeter insisted that the capitalist process was not one of equilibrium in which markets adjusted according to the price system and laws of supply and demand. Capitalism instead was essentially a disequilibrium system in a state of continuous turbulence, driven by the innovative activities of firms and individuals in creating new products and processes, new business organizations and markets. Schumpeter crafted his celebrated metaphor of “creative destruction” to describe this process of innovation and the disruptive impacts of these activities underlying this phenomenon. This view of the market and the development process placed Schumpeter outside the economics mainstream. Economists, he contended, erroneously visualized the fundamental problem in economics to be one of “how capitalism administers existing structures,” whereas, according to Schumpeter (1942), “the relevant problem is how it creates and destroys them” (p. 84). This destructive and creative process of innovation was unevenly spread over time, tending to occur in periodized clusters or waves. Such unevenness gave an historical dimension to both innovation and capitalist development. In his work on business cycles (1939), Schumpeter argued that the process of innovation, with its cycle of creativity by entrepreneurs and diffusion to other firms, accounted for the uneven historical swings of recession and expansion in the capitalist economy. Schumpeter’s preoccupation with the historical dimension of innovation and the economy was an ongoing theme throughout his later work (Lazonick 1990, 1994). In summing up his approach to the study of the economy, Schumpeter wrote how “the subject of economics is essentially a unique process in historic time.” Nobody can hope to understand economic phenomena, he argues, without “an adequate command of historical facts and an adequate amount of historical sense or of what may be described as historical experience” (Schumpeter 1954, 12). It was this integration of history that distinguished Schumpeter’s approach to innovation and economic development. Schumpeter, however, was not the first to embrace history as the foundation for understanding the economy.

30 A Synthesis schumpeter and the legacy of marx While Schumpeter’s work on innovation is highly original, it derives much of its influence from Karl Marx. Four themes stand out in Schumpeter’s theory of innovation and economic growth that reveal this influence: (1) the decisive role of technology in capitalist development, (2) the disruptive and revolutionizing tendencies of technological change, (3) the crisisprone character of capitalism, and (4) the historical character of technology and the economy. Schumpeter himself critically acknowledged this legacy of “Marx the economist” in framing his own fundamentally historical theory of innovation (Schumpeter 1942, 21–44).1 Aspects of Marx’s work are therefore a logical starting point for profiling Schumpeter’s views on innovation and development. Marx, in Schumpeter’s view, was the first of the classical economists to recognize the role of technological dynamism in the development of capitalism and the first to understand the role of history in influencing both technological change and economic development. Much as Schumpeter drew upon the neoclassical work of L´eon Walras to reveal how equilibrium models of commodity flows did not represent the historical process of economic development, Marx drew upon the classical economists, primarily Adam Smith and David Ricardo, in critiquing the absence of history in the economic orthodoxy of his own day. “Economists explain to us the process of production under given conditions,” Marx wrote. “What they do not explain is how these conditions themselves are produced, that is, the historical movement that brings them into being” (Marx 1847, 199). According to Marx, capitalism leads to an immense expansion in productivity because the system of private property rights, together with market competition, creates historically unique institutions that generate powerful incentives for firms to innovate and accelerate the process of technological change (Marx 1848; Rosenberg 1982a, 8). These institutions of private property, along with competitive markets and the incentives they establish, make the capitalist class the first ruling class in history whose interests are linked not to maintaining the status quo but instead to overturning it by developing new technologies as a source of profit and accumulation. Anticipating the now celebrated passage of Schumpeter on creative destruction, and providing insights about the current period, Marx observed that “the bourgeoisie cannot exist without constantly revolutionizing the instruments of production. . . . Constant revolutionizing of production, uninterrupted disturbance of all social conditions, everlasting uncertainty and agitation distinguish the bourgeois epoch from all earlier ones. All fixed, fast-frozen relations . . . are swept away, all new-formed ones become antiquated before they can ossify. All that is solid melts into air” (Marx 1848, 111). This view of

A Synthesis 31

the capitalist process as one of incessant innovation and disruption stemming from new technology had an unmistakable influence on Schumpeter. Marx employed a fundamentally historical method in accounting for new technologies. He ascribed the catalyst for technological change to growing markets beginning in the sixteenth century. Such widened markets provided the environment in which firms could exploit new technologies as a source of profit and accumulation. In this way, Marx was decidedly not a technological determinist (Rosenberg 1982a, 36–38). Far from assigning technology an autonomous and independent role in transforming the economy, Marx attributed changes in technology to the enlarged horizon of possibilities for profit created by ever-growing markets as the economy evolved from early manufacture to modern industry (Marx 1848). Once established as historical outcomes, however, new technologies emerge in Marx as a central element in capitalist development. The conflicts between technology as a productive force and the social relations of production in terms of ownership and control over technology and the surpluses generated from it are, for Marx, what drive the economy. Technology plays a critical role in this rhythm of development, but it is not some ineluctable force. Capitalists make choices to innovate in order to compete more effectively and accumulate profit. Nevertheless, Marx acknowledged that as technology changed in conjunction with market expansion so too did business enterprise. He had a theory of firm concentration in which competition, innovation, the cheapening of commodities, and the scale economies of large enterprise developed in an evolutionary way. What Marx understood as the tendency of technology to develop alongside the enlargement of the capitalist firm, however, Schumpeter took one step further in equating the phenomenon of innovation with oligopoly. In Marx, capitalist development, technological change, and transformations in the size and organizational structure of business establishments were all part of the same historically driven process. A similar story would be told by Schumpeter—but with important differences. innovation and entrepreneurialism While in broad outline Schumpeter accepted these key elements from Marx in creating his theory of economic development, he added a critical idea about the process of innovation that separated him from Marx—the idea of entrepreneurialism.2 Schumpeter also distinguished different phases of the innovation process itself. According to Schumpeter, innovation consists of three distinct moments—initial invention of new products, processes, organizations, and markets; commercialization of these elements; and finally diffusion of these elements to other firms. In conceiving of innovation as a series of historically conditioned moments, Schumpeter was interested first in differentiating the behavior of firms at each of the three phases. Second, he

32 A Synthesis was particularly intent on tracing how the responses of entrepreneurial firms to the profit-making environment resulted in new business routines that challenged existing business practice, diffused to other firms, and transformed the entire economic system.3 For Schumpeter, innovation was both artifact and impact. Central to Schumpeter’s theory of economic development is the creative act of entrepreneurs in commercializing new technology and, in the process, launching innovation (Schumpeter 1947). It was only an act of entrepreneurship that enabled technical inventions to emerge from obscurity and assume the role of commercial artifacts (Freeman 1991, 304). Schumpeter, however, defined this process of innovation broadly. He conceived of innovation as “the carrying out of new combinations” corresponding to the new products, new methods of production and distribution, new forms of business enterprise, and new markets associated by Schumpeter with technological change (Schumpeter 1911, 1942). Entrepreneurialism acts as a disruptive force in the economy, challenging the competitive strategies and behavior of existing firms. The relatively short bursts of technological creativity by entrepreneurs, however, engender longer periods of assimilation and adaptation marked by imitation and complementary types of innovation. This process of diffusion has profound consequences for the economy as a whole. It completes the pathway of creative destruction along which are the new products, new operational routines, new forms of business enterprise, and new markets where firms seek profit. Schumpeter observed how new technological combinations marking the process of creative destruction were distributed unevenly throughout the history of capitalism. Such combinations tended to cluster in “swarms” that marked the beginning of the growth cycle (Schumpeter 1911, 223). As the economy moves outward along a new production frontier owing to the expansionary effects of new technology, the economic rents accruing to innovators that ignite the growth process are eventually competed away as other firms adapt to and imitate new innovation. This leveling of profit rates then paves the way for downturn and depression. Far from repeating, however, these business cycles redefine the context for the next round of innovation, expansion, and contraction. From this notion of business cycles and technological clusters, Schumpeter arrived at a long-term view of capitalist development punctuated by distinct industrial revolutions separated in time.4 He dates the first industrial revolution from the 1780s to 1842. The second occurs from 1842 to 1897, while the third begins in 1898 and corresponds to Schumpeter’s own time. Although time-specific, these revolutions share common features of transformation that act as drivers of the capitalist process. Schumpeter actually referenced a key aspect of the nineteenth-century communications revolu-

A Synthesis 33

tion in coining the term “railroadization” to describe the pattern of economic change associated with these features (Schumpeter 1939, 304, 325–51, 72–192; Andersen 1994, 26–62). In focusing on the railroads to illustrate his theory of economic change, Schumpeter built a model starting with an equilibrated system of competitive strategies, routines, business organizations, and markets that is “disturbed” by the innovation of railway-based transport networks.5 This innovation in the transport and communications sector of the economy provokes responses by business users of this infrastructure. Entrepreneurial firms among these users develop strategies, routines, and forms of enterprise that challenge the products, processes, organizations, and markets of other firms. These innovators create a new cost and pricing structure for economic activity and, more important, new activity itself. These changes in costs, prices, and types of economic activity are the basis of what Schumpeter described as new production functions in the economy. In order to compete, other firms adapt to the innovative activity of entrepreneurs and the production functions they establish. The result of these innovative and adaptive activities is a broader process of economic transformation. Schumpeter, however, was far from a technological determinist. He conceded that entrepreneurialism in the railway sector, which ignited such broad-based changes in the late-nineteenth-century economy, had a political edge. The leadership within particular groups of rail builders and the relationship of these groups to local, state, and national political figures played essential roles in promoting the viability of the railroad as a profit-making venture. These alliances between rail entrepreneurs and their political backers are what secured for rail builders the land and the rights of way necessary for rail building to occur in the first place. According to Schumpeter, such relationships not only were critical in promoting railroad development. Railroad entrepreneurialism tied to politics is what enabled the railroad to act as a catalyst for economic development. As railroads expanded, they triggered a range of innovations in other sectors of the economy because of opportunities to get involved in business activities supported either directly or indirectly by government. In this way, industries tied to rail building such as steel making emerged, while others, such as mail coaches, became extinct. Furthermore, according to Schumpeter, railroads supported by government homesteading promoted economic development in regions of road building ahead of population (Schumpeter 1939, 327–30). In effect, the visualization by individuals and firms of financial gain from railroad technology was an insufficient condition for launching the new infrastructure. For Schumpeter, the process of railroadization was an entrepreneurial as well as a political phenomenon. Initially, Schumpeter interpreted innovation to be an entrepreneurial

34 A Synthesis function of individuals (Schumpeter 1911; Freeman 1994). Later Schumpeter conceded that the entrepreneurial function had become increasingly socialized within the large capitalist enterprise. From the vantage of the mid-twentieth century, it was these firms that created the new products, processes, organizations, and markets of capitalist development. His “creative destruction” was a process occurring within these enterprises. The question still largely unanswered by Schumpeter, however, focused on what was actually occurring inside these enterprises to promote the innovation process. innovation as learning What Schumpeter conceded to be a still dimly perceived problem inside the firm emerged somewhat more illuminated several years later with the revelation that the innovation process is essentially a learning process (O’Sullivan 2000, 407). In many ways, the source for this now commonly accepted connection between innovation and learning derives from the work of Edith Penrose. In this link, Penrose uncovered the source of growth within the firm and throughout the economy. For Penrose, growth revealed an evolutionary process, at the core of which was the cumulative expansion of knowledge within the business enterprise (Penrose 1959, xii). As a pool of human and material resources bound within an administrative framework, the firm promotes growth by learning to transform its resources into new profit-making activities, that is, different products, more efficient processes, and even new ways of manipulating the market environment to serve its interests (Penrose 1959, xiii). Growth occurs when knowledge is added to this base of resources, and the firm subsequently provides the market with new goods and services in fundamentally new ways. In accounting for the socalled residual in the growth process, that is, the increment of expansion not attributable to increases in production factors, Penrose uncovered in the learning process one of growth’s critical missing links. As with Schumpeter’s growth concept, growth is generated within the enterprise. Innovation resulting from new knowledge is the catalytic agent for this expansionary transformation. These insights have spawned a more recent literature on innovation focusing on how firms learn, how firms act when they acquire new knowledge, and how the economy develops as a result of this process of learning and acting differently.6 One of the most compelling theoretical and empirical problems explored in this literature is how firms, in learning to generate profit differently in a given market environment, transform such knowledge into new capabilities. How, in effect, does the firm assume a role as an innovative enterprise? This route to innovative enterprise is distinguished by three primary elements (Dosi 1997, 1532). It begins from an economic environment with op-

A Synthesis 35

portunities for firms to learn about new alternatives for making profit and sufficient incentives for firms to exploit these opportunities. Second, this route to innovative enterprise is marked by the effort of firms to act on new knowledge by seeking to master capabilities lying outside their established business practices. Finally, this route approaches completion as firms create appropriate organizational arrangements to implement innovative advances. The environments where this learning process occurs are historical creations. Often referred to as technological “regimes” (Nelson and Winter 1982) or technological “paradigms” (Dosi 1982, 1984; Perez 1983), these historically conditioned environments share similarities with Schumpeter’s technologically based industrial revolutions.7 They demarcate periods of capitalist development based upon aligned sets of technologies, markets, business organizations, and institutions. As conceived by Carlota Perez (1983), periods of capitalist expansion reflect “complementarity” between dominant technologies and the organizational and institutional framework, while periods of contraction create crisis that promotes so-called creative destruction (pp. 358–61). This rhythm of the business cycle and innovation is the foundation for paradigm or regime change. These historically created environments also circumscribe what firms are capable of learning during a specific period. Based on the attributes of technologies, markets, and institutions in existence at a given moment in time, regimes or paradigms condition the types of economic problems firms confront, and therefore what firms in a given period can create. Nevertheless, historically conditioned regimes or paradigms leave open various forms of technological novelty and learning from one moment to the next (Dosi 1997, 1531). When one form of technological novelty gains momentum and becomes generalized among a broad population of firms, the nature of the regime or paradigm shifts. The result is transition from one historically periodized industrial revolution to another. What launches this learning process leading to the enhancement of capabilities and organizational transformation is the search and selection by firms for more profitable economic routines (Nelson and Winter 1982). Conceived broadly, routines refer to the “technology” of the firm, or what is often described as “getting things done.” Modifying the technology of the firm through learning—changing routines—involves a shift in the capabilities of the business organization and a shift in the organizational structure of the firm itself. This creation of capabilities and organization is the essence of the innovative process. When the process of search and selection is carried out by innumerable firms in imitating an initially successful innovation and the initial innovation becomes generalized, the overall economy evolves and there is transition from one historically periodized industrial revolution to another. Consequently, the route from learning within individual firms to

36 A Synthesis the development of innovative capabilities throughout a generalized population of firms occurs as part of an historical transformation. Organizational learning involves an investment by the firm in reorganizing its resources (O’Sullivan 2000, 407). Such investment, however, exposes the firm to risk and uncertainty.8 Firms that commit to learning and enhancing capabilities confront the risk of having to forgo a measure of both the use and exchange value of these resources as they are redeployed as part of learning process. At the same time, they confront uncertainty of two varieties: productive uncertainty and competitive uncertainty (O’Sullivan 2000, 407). Productive uncertainty exists because firms committed to learning have to figure out how to develop the capabilities in which they are investing before these resources can generate more profitable returns. Competitive uncertainty exists because even if a business develops a new product or better process it may not be superior to that of a competitor pursuing an alternative approach. Consequently, a critical element of the learning process involves the assessment by the firm of potential outcomes stemming from investments to acquire new capabilities. Firms learn to assess such outcomes in a variety of ways (Pavitt 1992, 220–21; Dosi 1997, 1532). They learn from direct experimentation with new products, processes, and entries into new markets in a process encompassing trial and error; they learn from competitors and other businesses, such as their own suppliers; they learn from other organizations and institutions, such as universities and government; and finally they learn from unsuccessful efforts at solving problems and even by failing at such attempts. Seldom do firms understand fully the exact trajectory of where the learning process will take them. As Schumpeter himself acknowledged, innovation is often the outcome of action taken without a complete understanding of future results (see O’Sullivan 2000, 407–9). In solving one problem to enhance capabilities, firms normally encounter additional problems unforeseen at the beginning of the learning process. For this reason, firms in the course of learning are often compelled to solve contingent problems that arise only after certain other difficulties have been overcome. What differentiates firms is the degree to which they are able to develop the capabilities to achieve a set of outcomes foreseen from an innovative advance. This process of organizational coherence, or “integration,” is central to the creation of innovative enterprises (Lazonick and Mass 1995, xv). Through their own agency to learn and make choices about technologies for getting things done, firms create their own levels of coherence that establish this differentiation. These choices, however, do not derive from some omniscient understanding of the most profit-optimizing pathway available in the market as assumed in rational choice models of human action. Firms seek solutions to problems and select alternatives for competing on the basis

A Synthesis 37

of imperfect knowledge about profit opportunities and an incomplete picture of the technological solutions available for pursuing these opportunities (Dosi 1997, 1531–32; Lamoreaux, Raff, and Temin, 1999, 6–8).9 This imperfect knowledge gives rise to variation in the choices firms are likely to make regarding strategy, routines, and organization. While business firms exist in the same world, they see the world differently, and they learn different things from the same world. As a result, they make choices that are not programmable but instead are highly contingent (Metcalfe 1998, 35). Some firms achieve much higher levels of learning and coherence than others. At the same time, the selection of competitive strategies, operational routines, and forms of business organization is not random. Because firms compete in historically conditioned environments, they make choices from a range of options that derive from such environments. Thus, while the parameters for the choices of firms are historically created, firms exercise agency in making their selections (Yates 1997).10 These choices drive the economic development process (Nelson 1998, 322). They not only provide the basis for innovation in the economy but also are the mechanism by which the innovation process diffuses, spreads, and transforms patterns of economic development. innovation as inducement While the literature on innovation as learning provides an analytical route from the microactivity in the firm to the increasing returns generated from new capabilities, it is less precise in specifying what in the market environment is providing the catalyst for acquisition of new knowledge. Here, as Giovanni Dosi (1997) insists, a valuable link may be made with the growth literature on inducements to innovation. From this perspective, innovation results from the responses of firms to specific transformations in the market environment. Changes in market demand, factor prices, even new technologies act as inducements on firms to compete in new ways. As a consequence, firms, in seeking the profit opportunities from different circumstances, learn new things and alter the supply of knowledge in the economy. The outcome of such collective organizational learning is innovation and growth.11 Inducements to growth and innovation, however, are not limited to changes in demand, prices, or technology. Inducements—as well as constraints—to innovation also exist outside the formal boundaries of the market within the realm of politics and institutional settings (Christopherson 1993; Zysman 1994; John 1998). Accordingly, innovation is more than a process of knowledge acquisition to alter routines for accumulating profit. Innovation involves the interaction of the firm with systems of economic rule making established through politics and structures of power related to

38 A Synthesis conflict and consent among groups and classes. Innovation, in effect, has two components (Mokyr 1990, 11): (1) an epistemological component involving the acquisition of knowledge and its transformation into new capabilities; and (2) a social and political component involving the influence on firms of institutions and the power struggles within society and within the firm itself. Such influences shape firm behavior and thus affect technological outcomes. This second dimension of innovation, according to its adherents, has been subordinated to the perspective on innovation as learning (Hughes 1983, x). For this group the learning environment is a product of politics, institutions, and struggle. Two closely related approaches to technological change, namely “contextualism” (Hughes 1983) and “social construction” (Bijker, Hughes, and Pinch 1989), seek to remedy this omission of politics and power relations.12 The model of innovative advance found in contextualism and social constructionism borrows certain features from Schumpeter, namely the ideas of invention and diffusion (called “transfer” in the social construction literature). To these two concepts, however, Thomas Hughes (1983) and Wiebe Bijker et al. (1989) add the notions of “reverse salients” and “momentum.” According to Hughes, reverse salients refer to critical technical problems where the line of innovative advance encounters bottlenecks in the form of knowledge gaps that if left unresolved preclude innovation (Hughes 1983, 14–17). For social constructionists, reverse salients also refer to constraints on innovation emerging within the social and political environment ranging from opponents of technological change including labor, to rule-making environments that create legal barriers to change. Critical to this group of theorists is the actor network, the medium through which individuals, groups, and classes interact and struggle with each other and through institutions to shape innovative outcomes according to their interests. Momentum refers to the phase in the innovation process when the problems of reverse salients are confronted and resolved, enabling innovation to strengthen. Central to this phase is the resolution of power struggles within actor networks, the outcome of which enables certain actors and groups with particular technological interests to prevail. Winners can choose to promote, thwart, or redirect the trajectory of innovation. Actors that prevail in these contests for power use the rule-making authority of politics and institutions to legitimize the chosen technological pathway corresponding to their interests. The critical aspect of this approach is that innovation is a contingent process shaped by choices, politics, struggles, and power.13 With similar concerns, but placing a greater emphasis on history, is a group of scholars who critique the idea of innovation as an outcome of the search for efficiency (Berk 1994; Roy 1997; John 1997; Sabel and Zeitlin 1985). Far more central than “the logic of efficiency” in understanding innovative outcomes is the role of politics, institutions, and relations of power. Equating

A Synthesis 39

innovation with efficiency, they argue, is akin to an ex post vision of the innovative process that suffers from what they insist is technological determinism. Moreover, they contend such arguments about innovation cast in the logic of efficiency are essentially restatements of neoclassical economic models. Innovation in these models is the result of efficient allocative outcomes. This overly determined, neoclassical vision of economic activity, they insist, has been overlaid upon the historical process and, much like neoclassical models, omits any substantial role for institutions, politics, and power struggles in the economy.14 Similar to the contextual and social constructionist approaches, these scholars seek to tell a story in which innovation, when it occurs, is induced and constrained by the social and political environment. Inducement is also the underlying theme in depictions of innovation as the outcome of the contentious relations and unequal power between workers and management. From this perspective, innovation is the result of class conflict and the ongoing efforts by management to gain greater levels of control over workers and work (Marglin 1974; Noble 1984). In pursuit of this aim, management searches for new technologies that empower managers with enhanced capabilities to reorganize work more efficiently with less resistance from workers. For their part, however, workers are endlessly seeking ways to resist these encroachments of management (Burawoy 1979). In this way, this interplay of managerial power and worker resistance induces continuous innovative advance. Finally, if conditions in the environment are what induce firms to learn and expand capabilities, then some of the most critical inducements to learning are the external economies and networklike interactive relationships of firms in so-called milieux of innovation or learning regions. Inspired by the insights of Alfred Marshall, this view of the innovation process derives from the observation that innovation tends to concentrate geographically in certain regional economies. In these place-based concentrations of economic activity emerge the interactive network relationships within and between firms that provide firms with the external scale economies—Marshall’s “mysteries in the air”—from which firms learn and innovate and from which regions become differentiated (Saxenian 1994).15 In this way, changes in the economic environment and conditions of concentration in the environment induce the process of learning within firms. The innovative enterprises that by definition are the agents of this growth process are also, it turns out, transformed by it and assume identities as new business organizations. The idea of innovation as inducement complements the idea of innovation as learning by emphasizing that the learning environment itself is circumscribed by the influence of politics, institutions, and power on the actors involved in economic activity. What firms learn and what they are able to implement is not the outcome of a mechanical functionalist process. Learning

40 A Synthesis by firms and the creation of new capabilities is a contingent process in which an innovative advance emerges under the influence of economic rules, institutional legacies, and the contested interests of different political, organizational, and class actors.

The Firm as Business Organization Business firms create forms of organization in the course of seeking profit. As firms learn about new ways to accumulate, they not only transform their routines for producing, buying, and selling. Firms reorganize to implement these new routines. In this way, organizational transformation is an integral part of the innovative process. This relationship between innovation and organizational change has its origins in Marx and Marshall. It also has a more recent lineage. In the late 1920s, economist Allyn Young (1928) observed that the marketplace consists essentially of “productive activities tied together by trade” (p. 533). He used this characterization as a starting point to determine how the relationship between the market, the division of labor, and innovative methods of production produces increasing returns and economic growth. His aim in revealing the outcome of this relationship was twofold. First, he wanted to demonstrate why forces counter to economic equilibrium “are more pervasive and more deeply rooted in the economic system” than is commonly realized (Young 1928, 533). Second,Young wanted to show how external economies, deriving from the extent of the market and the division of labor, provided the source for what he metaphorically termed “roundabout,” or innovative methods of production. In this effort, he not only drew upon the seminal insight of Adam Smith linking growth to the interplay of the market and the division of labor.Young also seized upon the observation made by Marshall of the business firm as a unit of organizational change and used this characterization to argue that innovation and economic progress derived from the capacity of firms to evolve in conjunction with changes in the market environment. Emerging from Young’s synthesis of Smith and Marshall is a marketplace of business firms evolving in organizational structure as they seek roundabout methods for producing and trading in an effort to generate increasing returns. While Young’s article provided a dynamic, even evolutionary view of economic development, his approach focused more on the aggregate economy than the business activity of individual firms (Lazonick 1991a, 294–95). It was Ronald Coase (1937), who, in a highly original article written roughly ten years later, asked a fundamental question about the nature of the firm that provided a critical theoretical insight on firms as forms of business organization. The issue that interested Coase was why and under what cir-

A Synthesis 41

cumstances a firm would choose either to produce on its own or purchase a given input in creating a product or service. “To make or to buy” was the essence of this choice. As a practical matter, in exercising this choice firms decide on the extent to which they internalize adjacent steps of producing, buying, and selling and the extent to which they contract with other firms in undertaking these activities. Such decisions situate firms along a continuum of discrete alternatives framed at each end by two basic types of business organization: intrafirm organization, in which the firm is highly integrated, and interfirm organization, marked by interactions among separate firms on the basis of price signals coordinated in markets. What Coase sought to uncover was the condition under which firms make the choice to internalize certain activities through planning and administrative coordination or contract with other firms through the price system, and the impacts of these choices on forms of governance and the boundaries of firms. integrated firms, intrafirm organization As a starting point in addressing this puzzle of organization, Coase imagined an economy “under no central control,” but unlike Young he focused his analysis on the individual firm in seeking to identify how the functions performed by firms are divided up among and between them. This issue led Coase to pose three basic questions: (1) When do firms produce for themselves internally, and when do firms purchase from other firms? (2) What types of economic organization derive from these decisions to make or buy? (3) What determines which activities a firm chooses to do for itself and which it procures from others? Coase, in effect, sought answers to why firms exist, what determines their internal organization, and what determines their boundaries. These questions, in turn, led Coase to address the puzzle of why, when there is a price mechanism for securing all goods and services in a specialized exchange economy, there should be any economic organization— that is, firms—at all (Coase 1937, 388). To solve this problem, Coase observed that the economy, although absent a central control, is only partially coordinated by the price mechanism. Firms employ a different organizing principle in which “conscious power” or planning is used to allocate resources. “If a workman moves from department Y to department X,” argued Coase, “he does not go because of a change in relative prices, but because he is ordered to do so” (Coase 1937, 387). Coase equated this power of planning to “entrepreneurs” and distinguished their activity of coordinating the operations of the firm internally from the activities of firms transacting through the price system. In the economy, he observed, “price movements direct production which is coordinated through a series of exchange transactions on the market.” In the firm, by contrast, “these market transactions are eliminated and in place of the complicated

42 A Synthesis market structure . . . is substituted the entrepreneur co-ordinator, who directs production” (Coase 1937, 388). For Coase, internalizing these transactions within the firm, and transacting in the market with other firms through the price system for the same goods and services are the two alternative methods for coordinating economic activity (Coase 1937, 387–89). For the answer to his central question of why there are firms, Coase proposed that there are costs to firms—transaction costs—of using the market and the price system to exchange goods and services. When the costs of coordinating transactions internally are less than the costs of using the market and the price system to transact for these items, the firm absorbs the activities represented by these transactions into its own organizational structure. As a consequence, the firm becomes more integrated and less reliant on the marketplace to secure the items needed to create a product or service. For Coase, a firm has a role to play in the economy if “transactions [can] be organized within the firm at less cost than if the same transactions were carried out through the market.” Firm boundaries are also established through this same mechanism of choice deriving from the costs of transactions. The limit to the size of the firm is reached “when the costs of organizing additional transactions within the firm exceed the costs of carrying out the same transactions through the market” (Coase 1991, 48). Managers of firms, he claimed, are preoccupied with the single overriding concern of transaction costs in calculating the trade-offs of using the market or absorbing business activities internally (Coase 1937, 404).16 This singular focus with transaction costs, however, compelled Coase to ignore other critically important aspects of business organization. Coase rejected a role for technology in the organization of the firm (Williamson and Winter 1991, 4). Coase also did not view politics, or contingencies in the historical process itself, as influential on the organization of the firm. His model is abstract and ultimately ahistorical (Lazonick 1991a, 195). Despite these omissions, in his 1937 article, Coase produced a seminal work with an enduring legacy. In posing basic questions about the structure of enterprise, he provided a theoretical starting point to account for the organizational variation in firms. He found in transaction costs a compelling, if one-sided, explanation for why firms were vertically integrated or why they interacted with other firms across markets. Coase’s pathbreaking approach to the nature of the firm, though surprisingly neglected for many years, has nevertheless provided the motivation for more recent theory on firms as forms of organization, most notably in the work of Oliver Williamson (1975). Starting from Coase’s dichotomy of the way firms either internalize transactions or transact through the market, Williamson used transaction costs to account for firms as representative of markets and hierarchies. Market-oriented firms were those that used the mar-

A Synthesis 43

ketplace to transact with other firms for inputs to make goods or provide services. These firms would also transact with other firms to distribute and sell their products and services. Hierarchies, by contrast, were those firms that assumed ownership over the input activities, the productive activities, and the marketing activities in creating and selling a product or service. Hierarchies, in effect, assume ownership and control over large portions of procurement, production, and distribution and are the equivalent of Coase’s directing “entrepreneur.” What Williamson did that differed significantly from Coase, however, was twofold. First, unlike Coase, Williamson proposed that transactions and transaction costs not only exist in market exchange but are also part of the operations internal to the firm. Consequently, transactions for Williamson exist both within firms and between firms contracting across markets. Costs of transactions result from uncertainty in exchange, which arises from three primary sources: (1) the self-interested opportunism of other parties to the transaction, that is, the potential of a party to a transaction to lie or cheat; (2) incomplete information or bounded rationality regarding the parameters of the transaction; and (3) asset specificity, that is, whether an asset necessary for a certain type of activity can be secured and controlled and whether such an asset can be redeployed for other use. The choices of firms to transact through the market or internalize transaction activities derive from their efforts to minimize the costs of these uncertainties. For Williamson, minimization of the uncertainties related to opportunism and bounded rationality suggests organization of transactions through markets. By contrast, minimization of the uncertainty related to being without assets necessary for a certain type of transaction or possession of assets without alternative uses compels firms to remedy such problems through integrated and hierarchical organization. Although Williamson accounts for three basic conditions that affect transactions costs, it is asset specificity upon which his predictive theory of organization hinges (Lazonick 2002, 11). Specifically, Williamson hypothesizes that “market contracting gives way to bilateral contracting which in turn is supplanted by unified contracting [hierarchical organization] as asset specificity progressively deepens” (Williamson 1985, 78; see also Lazonick 2002, 11). Second, unlike Coase, Williamson tested his approach to transactions costs and organization in actual historical situations. He was especially interested in the formation of corporate hierarchies and the process of vertical integration in the United States during the late nineteenth century. In this sense, Williamson’s interest in history aligned his work with Alfred Chandler’s approach. Nevertheless there is at least one important difference between the transactions-cost approach to firm structure elaborated by Williamson and the strategy-and-structure approach to the organization of the firm pioneered by Chandler. This difference focuses on the issue of the

44 A Synthesis relationship between innovation and organization. As Williamson concedes, the introduction of innovation complicates the assignment of transactions to markets and hierarchies (Williamson 1985, 143; see also Lazonick 2002, 13). It was Chandler who would more systematically make this connection between innovation and technology on the one hand and business organization on the other. For Chandler, the strategy of the firm, deriving from an “awareness of needs and opportunities created by a changing economic environment,” fundamentally influenced the structure of the firm, defined as the “design of organization through which the enterprise is administered” (Chandler 1962, 14).17 Beginning in the second half of the nineteenth century, the appearance of the vertically integrated enterprise reflected new strategies developed by management to produce and market goods in high volume. Through such strategic and organizational adaptations, management created what Chandler describes as “economies of speed” in an effort to achieve high-volume throughput. It was the ongoing efforts of firms to master economies of speed and high-volume production and marketing that account for the tendency of vertical integration to assume a dominant role in the organizational structure of late-nineteenth-century business enterprise (Chandler 1977). According to Chandler, there were good reasons for the link between the size of the firm, the volume of throughput, and the speed at which goods were produced and sold. Faster, high-volume throughput hinged on uninterrupted sources of supply and unimpeded sales of finished goods. Such requirements implied that functions once mediated by different firms using the market mechanism began to accumulate within the boundaries of a single firm using the “visible hand” of management (Chandler 1977). This form of administrative planning provided better forms of coordination between inputs and output in the environment of high volume and economies of speed. In addition, the need for management to secure more predictable sales outlets for the higher volumes of throughput, coupled with antiquated distributions systems, pushed numerous firms into marketing activities and forward integration. In this way, the strategy of the firm became linked to organizational structure through innovation that transformed production and distribution.18 In Chandler’s view, communications and transport technologies figured prominently in the emergence and development of this new business institution (see “Communications as Catalyst and Carrier Wave” below). The rail and telegraph system helped integrate formerly isolated localized markets into a geographically extended national market while at the same time concentrating market demand in cities. This market structure, in turn, created a new set of strategic opportunities for the firm. In the wake of more

A Synthesis 45

extended markets and mass markets in cities, producers had incentives to expand volumes in order to service this more extended and concentrated national market space. Rails and telegraphy furnished producers with the reliability and speed necessary to coordinate flows of supplies and finished goods long distances, as well as to and from urban factories in sufficiently high volumes to service this new market structure. As a consequence, firms became larger to take advantage of scale economies and assumed new capabilities in coordinating larger flows of inputs and outputs. Business firms also integrated backward into raw material suppliers and forward into marketing to sell their finished products. In Chandler’s model, changes in technology and markets created a new strategic orientation for producers based on high-volume throughput. Drawing upon Max Weber and Schumpeter, strategy became structure in the form of the vertically integrated, administratively coordinated intrafirm business organization. Chandler argued forcefully that the large corporation, coordinating its activity through internal administration rather than market transactions, was a more efficient form of business organization than the small-scale proprietary firm coordinating its activity through markets. Contrary to the claims of his critics, however, Chandler did not insist that the integrated corporation reflected an innately superior form of organization in economic life. His work seeks to explain the ascendancy of the vertically integrated firm in the American economy in a particular setting—the period of 1870 to 1920. For Chandler, this form of enterprise emerged around an historically conditioned set of efficiency objectives—economies of speed and economies of scale—that became realized through management control and vertical integration. It prevailed during the period for this historically specific economic reason.19 interfirm organization Ironically, when Chandler’s Visible Hand appeared in 1977, the large-scale, vertically integrated corporation appeared to be suffering the first serious challenge to its hegemony as a profit-generating institution since its initial creation in the late nineteenth century. Beginning in the early 1970s, large corporations in the industrial countries, especially in the United States, exhibited a precipitous decline in profitability that continued into the 1980s (Harrison 1994, 125–27). At the same time, a range of new firms, mostly from Japan but also from other countries in East Asia, along with smaller, more entrepreneurial firms, emerged as serious competitors to these previously formidable corporate organizations. Associated with a phenomenon described as the new competition, these firms were distinguished by their market-shaping, as opposed to market-reacting, behavior (Best 1990, 11). Interestingly, the economic challenge to corporate America represented by

46 A Synthesis Japan came from firms that were seemingly even more highly integrated than American companies. These Asian firms actually helped provoke this profit crisis by exposing the complacency and uncompetitive character of their once dominant American counterparts. The Japanese keiretsu and the Korean chaebol were business organizations integrated both vertically and across sectors, including finance (Gerlach 1989; Amsden 1989). Eschewing forms of market exchange, the keiretsu and the chaebol organized their business operations through tightly coordinated and highly administered relationships. In many ways, these organizations were the quintessential embodiment of the “visible hand.” Nevertheless, a very different story of this challenge—and one that has had a more enduring impact owing to the eventual slowdown and sustained malaise of the Japanese economy—has emerged with a focus on a far different organizational phenomenon. In this interpretation, the role of the large firm in economic development was being undermined by examples of place-based innovation deriving from clusters of medium-sized and even small firms, notably in Italy, Germany, and the United States. Such examples suggested the possibility of alternative models of economic growth and development to those driven by large-scale integrated enterprises. Together, this decline of large firms and the allure of alternative growth models based on clusters of smaller companies created what appeared to be a new environment for competing. By the mid-1980s, this change was apparently so pervasive that Michael Piore and Charles Sabel, in an influential book (1984), argued that the capitalist economy had arrived at what they termed a second industrial divide. Piore and Sabel compared this historical conjuncture to a similar moment during the previous century when mass production emerged from craft production. In the divide of the 1980s, however, the strategy that they advocated for relaunching growth was based upon a vision of transition to smaller— and more importantly—more flexible forms of business enterprise (Piore and Sabel 1984, 6). Not only did they find inspiration for this vision in certain industrial communities of craft production in the nineteenth century that, in their view, represented historical alternatives to mass production.20 Piore and Sabel were also able to reference the existence of smaller and medium-sized firms clustered in numerous place-based industrial districts as the living seeds of the new industrial order. What distinguished these communities, both past and present, were networks of firms based on relationships of cooperation and competition. Regardless of whether their vision was viable, Piore and Sabel had uncovered in these networks of firms an emerging trend in economic and organizational development. Following in the wake of Piore and Sabel’s book was an enormous amount of new theorizing about interfirm networks as a new form of business enterprise.21

A Synthesis 47

Much of this theorizing about regionally concentrated networks of firms derived from two overlapping convictions. First was the affirmation that the integrated firm was in a deep and perhaps irreversible malaise, its crisis the result of internal bureaucratic rigidities stemming from integration as an organizational form that precluded possibilities for innovation (Best 1990). These organizational characteristics, which at one time may have enabled the vertically integrated enterprise to compete effectively, now tended to act as blocks on innovative learning and the development of capabilities to enhance competitiveness. Implicit in this critique was the notion that as an organization the integrated firm possessed little capacity for adapting to a more competitive market environment. At the same time, part of this critique was the embrace of interfirm networking as a solution to the problems of the large-scale integrated firm. These network organizations, it was argued, promoted pathways of learning and adaptability that enabled them to innovate and compete. As institutions, integrated firms, from this perspective, were becoming competitively inferior if not obsolete, while enterprises organized from networks of firms were hallmarks of the future. Second, and perhaps more profound, was the conviction that existing theories of the firm provided little insight about interfirm networks as a specific organizational phenomenon. The dichotomy between contracting relationships undertaken through markets and contracting relationships organized administratively within firms, first developed by Coase and later refined by Williamson and Chandler, was from the perspective of these theorists insufficient as a framework for explaining the emergence and proliferation of interfirm contracting relationships in the aftermath of the profit crisis of the 1970s and 1980s. Signals for such a perspective, however, had already emerged independent of, and prior to, the competitive crisis of large firms. In a compelling article on industry and organization, G. B. Richardson (1972) argued that “by looking at industrial reality in terms of a sharp dichotomy between firm and market we obtain a distorted view of how the system works” (p. 884). His observations on forms of networking and contracting relationships between firms suggested first that business organization was highly open-ended and second that the choices made by firms on forms of organization represented a continuum, passing from pure market-type transactions through intermediate forms of cooperation to cooperation fully and formally developed within the same organization (p. 887). For this reason, Richardson was highly critical of the dichotomy between firm and market. Such a dichotomy, he claimed, “leaves out of account . . . the dense network of cooperation and affiliation by which firms are inter-related” (p. 883). Richardson’s insights into networks as unique forms of organization

48 A Synthesis resonated strongly in the more recent theorizing of network enterprises. “Neither Market nor Hierarchy” expressed this rejection of the lineage established by Coase (Powell 1990). In emphasizing the unique attributes of interfirm networks, Powell would write, “I do not share the belief that the bulk of economic exchange fits comfortably at either of the poles of the market-hierarchy continuum. . . . My aim is to identify a coherent set of factors that make it meaningful to talk about networks as a distinctive form of coordinating economic activity” (Powell 1990, 293, 300–301). From this framework, several contributions have attempted to explain the nature of interfirm networks as forms of business organization and issues of governance and coordination that supposedly make them uniquely innovative. Manuel Castells is perhaps the most emphatic in affirming the uniqueness of interfirm networks as an organizational form of business enterprise. Castells, however, has a far different point to make than simply distinguishing networks from either markets or hierarchies. He equates this organizational phenomenon—“linkages between economic agents”—with what he insists is a broader, historically unique networking phenomenon in the economy linked to “the information technology revolution” (Castells 2000, 5, 77). These linkages are essentially horizontal relationships in which the operating unit of the business is not really a firm but instead more a project enacted between nodes in networks (pp. 177, 214). For Castells, such ephemeral forms of organization correspond to the flexible nature of economic activity in the new millennium and the need of business enterprise for adaptability to compete in the restructured environment of capitalism dominated by the Internet. Project-oriented linkages can easily be transformed and reconstituted as business needs change and as conditions for profitability are redefined. Nevertheless, if, as Castells argues, these networks are unique to the information technology revolution, is it technology that is creating these forms of organization? And if networks are linked to the information technology revolution, what is one to make of interfirm networking before this revolution? Furthermore, apart from references to the power of information technology, it remains unclear what the mechanisms of coordination and governance are within this new organizational form that enable them not only to function but to function more innovatively than other forms of organization. In contrast to Castells is the literature describing the networking phenomenon as commodity chains. This phenomenon is defined as a network of production and labor processes the result of which is the creation and sale of a finished commodity. It differs from Castells in acknowledging the existence of commodity chains as forms of capitalist business organization dating from the early period of capitalism in the sixteenth century. In effect, the

A Synthesis 49

literature on commodity chains takes a more long-term view of interfirm networking and its role in the development of capitalism. At the same time, however, it acknowledges that commodity chains in each period possess unique attributes. This historically based perspective derives from a synthesis of two unlikely intellectual partners. On the one hand, theoretical inspiration for this literature derives from the “world systems” approach to capitalist development, in which different commodity chains spanning great distances across the globe fuel capitalist expansion beginning in the sixteenth century (Hopkins and Wallerstein 1994). On the other hand, this approach draws upon Michael Porter’s notion of the value chain defined as “an interdependent system or network of activities connected by linkages” that represent the various adjacent stages in the production and distribution of goods and services (Porter 1990, 41). In borrowing from both world systems theory and value chain theory, the commodity chain approach focuses on goods as a complete process of production, labor, and marketing. It seeks to reveal where the different parts of this process occur geographically and who controls the process (Gereffi and Korzeniewicz 1994, 2; Hopkins and Wallerstein 1994, 50). Commodity chains have three basic characteristics (Gereffi and Korzeniewicz 1994, 7). First, they have an input-output structure corresponding to a sequence of value-adding activities at different nodes. Second, they have a “territoriality” corresponding to the location of the various activities in the commodity chain and the way these activities occupy space. Third, commodity chains have a governance structure in the form of authority and power relationships within the network. These attributes, however, give commodity chains an historically specific character. During the late nineteenth and early twentieth centuries, certain types of commodity chains were internalized within the boundaries of vertically integrated corporations, where they coordinated mass production activities over national territories. What is distinct about the late twentieth century is the transformation of commodity chains into networks of independent firms organizing adjacent operations of procuring, producing, and selling around the globe (Gereffi and Korzeniewicz 1994, 7). Within these interfirm networks, profitability shifts from one node to another as an outcome of work organization and the distribution of power between the different nodes. Power, in effect, plays a key role in the governance of these organizations. Related to this approach but offering an analysis of more recent and specific forms of interfirm organization is the work of Tim Sturgeon (1997a, 1997b), who uncovers what he considers a new model of network organization: the turnkey production network. Functionally, this new entity is characterized by the separation of innovative capacity and production capacity marked by a distinctly new form of production outsourcing. Organization-

50 A Synthesis ally, this separation is represented by the emergence of a distinctly new entity: the contract manufacturer. Sturgeon observes that since the mid-1980s and particularly in the 1990s, large American name-brand electronics companies such as Apple, IBM, Hewlett-Packard, and indeed Dell Computer have been abandoning their internal manufacturing operations and turning to contract manufacturers such as SCI, Solectron, and Flextronics to build all or part of their products. These contract manufacturers fabricate and assemble the products of their clients through what is known as a turnkey contract (Sturgeon 1997a, 11). The contractor assumes responsibility for production, while design and marketing are retained inside the boundaries of the name-brand firm. The contractors themselves undertake this production activity through myriad subcontracting arrangements, dispersing and concentrating production in complex networks throughout the world. Costs, the diffusion of capabilities and skills to firms throughout the world, and the retreat of namebrand firms into “core competencies” are what drive the development of this new networking organization. Contract manufacturing and interfirm networking are thus part of the same organizational phenomenon. In this phenomenon, however, innovative capabilities in production migrate to the contractors while innovative capabilities in design remain within the name brand firm. By contrast, in the work of AnnaLee Saxenian (1994), networks of firms emerging from specific industrial systems are the sources of innovation and competitive advantage that differentiate firms within one region from firms in another and the regional economies where firms operate. Silicon Valley, according to Saxenian, is a network-based industrial system (p. 9). It is an innovative region because the industrial system upon which the region is built promotes horizontal and decentralized interfirm network relationships. These interactions, in turn, emerge from, and at the same time reinforce relationships built from mutual reciprocity and trust. Network-like ties between specialized firms enable multiple and spontaneous interactions to occur that create ongoing recombinations of knowledge and information sharing. Such network relationships are the basis for a process of collective technological learning (p. 9). While on the one hand reference is made to the increasing specialization and division of labor in Silicon Valley, there is, on the other hand, strong suggestion in Saxenian’s work that the interactions and relationships so central to innovativeness are not conducted at all through markets. Instead, cooperation between partners in the decentralized networks of Silicon Valley seems to more closely resemble nonmarket-coordinated interactions based on “embedded” relationships and the concept of trust within “business groups.”22 Consequently, Saxenian raises an interesting puzzle about the structure of organizations that are neither market nor hierarchy. While a highly specialized division of labor suggestive of markets

A Synthesis 51

drives the existence of interfirm networks, the constituent firms within these organizations appear to contract through relationships lying outside the market and the price system. In this way innovation, instead of being driven by markets, derives from relationships with a strongly nonmarket character. This issue of whether markets or administered relationships prevail within interfirm networks is central in understanding contemporary hightechnology value chains. In a highly original approach to this issue, Richard Langlois argues that increasing modularity in product design is creating the basis for the reemergence and proliferation of markets as the preferred mechanism of coordinating activity between firms. Langlois develops his case by acknowledging that the world of the late nineteenth century described by Alfred Chandler represented an industrial revolution, marked organizationally by vertical integration and governed by the visible hand of management. He insists that the current period is characterized by a revolution “at least as important as the one Chandler described . . . as profound as the one of the late nineteenth century.” In contrast to the enabling technologies of the rail and telegraph, this revolution has as its technological infrastructure the computer and the Internet. In place of mass-production processes, the current revolution is one based on product customization oriented around modular component parts. Perhaps most important, while the earlier revolution replaced Adam Smith’s invisible hand with the visible hand of managerial coordination, the modular revolution is marked increasingly by “coordination through arm’s-length trading on thick markets. . . . In this epoch, Smithian forces may be outpacing Chandlerian ones” (Langlois 2001, 2). Langlois calls this new form of governance the “Vanishing Hand.” His aim is to explain why the mechanism for coordinating mass production— Chandler’s visible hand—“is fading into ghostly translucence” and why markets are outpacing administered forms of coordination for economic activity. While the logic of this argument is compelling, there may be good reasons why, in an environment where competition has shifted away from the product and toward the efficiency of supply chain systems, the invisible hand of Smith may be too risky for coordinating complex processes of just-intime procurement and assembly among different firms (see Chapter 6). Instead, it is the visible hand of Chandler—conscious command and control— that appears more suitable as a coordination mechanism for high-risk just-in-time supply chain systems despite the fact that the different activities in these value chains are being undertaken within networks of separate companies.23 In this sense, the observation made thirty years ago by Richardson on the possibilities of firms exercising power beyond their own firm boundaries is prescient. “Planned co-ordination does not stop at the frontiers of the indi-

52 A Synthesis vidual firm,” he writes, “but can be effected through co-operation between firms. . . . Anti-trust legislation has checked vertical integration, but the same co-ordination is achieved through close co-operation between individual firms at each stage. . . . Co-operation may come close to direction when one of the parties is clearly predominant” (Richardson 1972, 895–96). For Richardson, interfirm networking may in some instances resemble the integrated firm. Forms of power and control are common to both types of organization. What is critical in assessing interfirm networks is not the organizational form per se by which coordination is achieved, but rather the way in which power and control are exercised through different forms of organization in achieving certain economic outcomes.

Communications as Catalyst and Carrier Wave There are few inducements to innovation and organizational transformation that equal the effects of a fundamental change in the means by which society communicates. As William McNeill (2000) writes, “major landmarks in human history” along with “the impulse to innovate” depended principally on improvements in communications that allowed messages to travel farther and more accurately across time and distance (p. 9). It was historian Robert Albion, however, who was the first to write extensively about the impacts of what he described as the “communications revolution” on economy and society (Albion 1932; John 1994). Albion originated this concept to describe the creation of an unprecedented “veritable age of speed” beginning in the late eighteenth century with which new transport and communications networks moved merchandise, information, and the public itself (Albion 1932; John 1994, 101).24 This preoccupation with speed led him to highlight the role of the communications revolution in the United States, where speed was critical in bridging the enormity of the nation’s continental-sized distances. Despite his insistence that the communications revolution emerged independently of industrialization—“it had performed wonders while our industries were still legitimate ‘infants’”—he acknowledged the impacts of new transport and communications systems on the growth of machine production and largescale business enterprise (Albion 1932: 718–19). Nevertheless, Albion did little to develop these links systematically (John 1994). It was left to a student of Albion, Alfred Chandler, to provide what has become a defining set of connections between new technologies of transport and communications, production innovation, and transformations in business organization. Similarly to Albion, Chandler emphasized the role of speed in the compression of geographical space as the defining impact of the communications revolution. Chandler, however, placed more importance on the tech-

A Synthesis 53

nological breakthroughs of the steam railroad and the electric telegraph in defining the communications revolution. Such an approach led Chandler to insist upon the mid-nineteenth century as the turning point in the communications revolution. More important, Chandler would make explicit the causal links between the revolution of rails and telegraphy, the emergence of the integrated corporation of the late nineteenth century, and the system of high-volume production and distribution built upon economies of speed. In building his model of organizational transformation on the rail and telegraph revolution, Chandler, in fact, was largely responsible for introducing transport and communications infrastructure into the vocabulary of historical analysis as a catalyst of economic and organizational change (John 1994, 102). What Chandler represents in terms of the revolution of rails and telegraphy has its counterpart in the work of Castells (1996) on information technology. Like Chandler, Castells seeks to demarcate an historically unique economic and social phenomenon. In contrast to Chandler, Castells finds this break in the late twentieth century, which he describes alternately as the “information age” and the “network society.” His point of departure in accounting for this phenomenon is the “information technology revolution, which emerged in the 1970s and is represented in its most recent manifestation by the Internet (Castells 1996, 5).25 For Castells, this revolution consists of the converging set of technologies in microelectronics, computing, telecommunications, and biotechnology (p. 29). These technologies and the revolution they have engendered have shaped the restructuring of capitalism since the 1980s (p. 13). For Castells, the source of the “new economy” created from this restructuring process is unmistakable. “This new economy emerged in the last quarter of the twentieth century,” he writes, “because the information technology revolution provided the indispensable, material basis for its creation” (Castells 2000, 77). Castells is convinced that the information technology revolution represented most decisively by the Internet and the informational economy spawned from it reveal decisive breaks with industrial society preceding it. While conceding that past forms of economy relied on the processing of information, Castells distinguishes these previous information societies from the informational society that has emerged only in the last twenty-five to thirty years. In the informational economy, productivity and competitiveness derive from the capacity of economic agents to generate and process knowledge-based information, whereas in the industrial economy the source of productivity derived from manipulation of materials and access to sources of energy (Castells 1996, 17, 66). The information technology revolution is thus an historical discontinuity on the same level as the industrial revolution of the eighteenth and nineteenth century (Castells 1996, 30). In Castells’s work

54 A Synthesis (1997, 244–45) is a call to historians of technology to compare and contrast the recent period of transformation marked by the Internet with analogous transformations in the past. Castells is certain that history will judge the current period of the Internet revolution to be one of epoch-making discontinuity. A compelling counterclaim to this perspective insists upon a largely unified “dominant line of development,” beginning with the telegraph and culminating in the extended electronic and computer networking technologies of the late twentieth century (Hall and Preston 1988, 37). This “carrier wave” of communication technologies provides links between the midnineteenth and late twentieth centuries. Comparison of the innovative enterprises created by Swift and Dell may very well provide an ideal test case for assessing these contrasting claims.

Toward Synthesis The starting point for comparison of Swift and Dell is Schumpeter’s observation that capitalist development is punctuated by waves of technological innovation beginning in the late eighteenth century. These waves demarcate distinct periods of industrial revolution in the development of modern capitalism. Each period is distinguished by a system of dominant technologies, around which cluster specific types of economic routines, business organizations, geographies of economic activity, and institutions providing legal frameworks of economic rule making. Competition and the search for profit compel firms in these periodized environments to seek more innovative ways of accumulating. This process of search is a learning process. In this learning process firms confront limits posed by existing economic routines, business organizations, market geographies, politics, and labor relations that condition pathways of accumulation. The learning process is an effort to overcome such limits by conceiving how to act differently about producing, buying and selling, and accumulating profit. Although each of these periods is unique, features common to this innovation process create comparable economic environments across time. Among the most disruptive historical forces transforming the economic environment is the phenomenon of communications revolutions. Two groups of firms create this phenomenon and act as agents for the process of innovation, in which the disruptive impacts of new transport and communications technology emerge, spread, and transform the rest of the economy. Igniting this phenomenon is the actor network of new transport and communications infrastructure builders that mobilizes the resources for construction of new systems. Extending this phenomenon are business users of the new transport and communications systems. These users complete a more

A Synthesis 55

sweeping set of changes in the economy by deploying new infrastructure to transform existing business models for profit making. The interaction of these two groups shapes the build-out of the new infrastructure and the pathway of transformation throughout the rest of the economy resulting from the choices of users in deploying the new systems to compete (Cohen, DeLong, and Zysman 2000). The roles of builders and users in creating the communications revolution and spreading its impacts reveal a certain identifiable pattern. This pattern starts with a breakthrough invention in transport and communications technology that is exploited and commercialized by inventor-entrepreneurs within the ranks of communications revolution builders. Although patent rights frequently protect the new invention, the patent process reveals limitations in stemming the entry of companies anxious to capitalize on the commercial potential of the new technology. Accordingly, the ranks of firms interested in transforming the new technology into built systems and constructing new infrastructure for such systems—rail, telegraph, and Internet firms—though initially small in number soon explode after this first stage of commercial success. Competing companies, sometimes numbering in the hundreds, invariably get involved in this early period of infrastructure creation and development. In the initial stages of new infrastructure development, multiple variants of the new transport and communications technology emerge among the different firms, unleashing intense competition for control over what is defined as the most technically superior system design. As these myriad firms compete to build out the new systems, the competitive process gradually gives way to a process of consolidation in which a small number of builder firms survive. During this process of contraction in the number of builder firms, the multiple variants of the new technology created at the outset of commercialization also diminish. The surviving firms compete ferociously to establish a dominant design or standard defining the path along which subsequent infrastructure development takes place. At stake in these standards wars is control over future profit making. On the one hand, individual firms seek to use their mastery over a particular technological design to set the terms for subsequent development of the infrastructure. On the other hand, firms that successfully develop a dominant standard are in a position to control terms of infrastructure access and use. The telegraph, the railroads, and the Internet all went through this process of standard setting—with clear winners and losers. The build-out of new transport and communications infrastructure has a transforming impact on system users in two ways. First, the build-out of the new transport and communications systems alters the strategic outlook of business users for competing and accumulating profit. Second, break-

56 A Synthesis throughs in transport and communications provide users with a set of tools for changing business routines and transforming organizational capabilities to accommodate such strategic shifts. Communications revolutions alter the strategic horizons of business users by reconfiguring the geography of economic activity. Such reconfiguration has two components. In the first place, the deployment of the new transport and communications systems provides users with new and different levels of access across and within markets for buying, selling, and producing. Second, these new levels of access create a different structure of costs in coordinating the movement of goods and securing information across distant markets and between areas of proximity within markets. As communications revolutions reshape the geography of economic activity, and as user firms confront opportunities for engaging in economic activity with a different structure of costs and a different map of access routes to profit, the strategic environment for firms assumes a different meaning. What changes for business users as access and costs get reconstituted are fundamental relationships in economic activity oriented around space and time. Shifts in the strategic orientation of firms deriving from transformations in the geography of economic activity ultimately involve a change in the way firms organize the territorial and temporal relationships in their operational routines. Throughout capitalist development, accelerating time in the circuit of capital from procurement of materials to final sale and exploiting new territories as sales outlets as well as places for procurement and production has driven the competitive activity of firms (Schoenberger 1997, 12; Harvey 1996, 240–41). If businesses are to exploit new geographies of economic activity and take advantage of reconfigured structures of access and costs, they must develop different, even intensified systems of control over territory and the cycle times of procurement, production, and selling taking place there. Technologies of communications provide firms with possibilities for such control. In this way communications revolutions are control revolutions. Consequently, in addition to providing the catalyst for transformations in strategic outlook, communications revolutions provide firms with technological pathways for developing the capabilities—operational and organizational—to exercise control over space and time in new and different ways. The economic environments of communications revolutions, where opportunities emerge for learning to control time and space in new ways, are without predictable outcomes. Certain firms learn about the profit-making opportunities associated with the communications revolution more decisively than do others and integrate the new infrastructure into their business models in accordance with their understanding of such opportunities. Nevertheless, the business models of such innovative firms rarely emerge fully

A Synthesis 57

formed. They evolve as incremental experiments in a process of learning by doing and learning by using. Gradually, through trial and error, firms create new systems of codified and tacit knowledge. This knowledge is the basis of new capabilities for producing, buying, and selling. It is also the foundation from which firms, in developing new capabilities, evolve into innovative enterprises. Business organizations are inherently territorial. They assume this territorial character in the way they choose to organize economic activities in geographical space. They organize their activities geographically in the way they locate their own physical assets, or nodes, and in the way they organize the flows of activity between these nodes and the nodes of other firms with whom they interact in producing, buying, and selling. While to some extent, the location of nodes and the configuration of flows reflect the capabilities of available transport and communications technology, nodes and flows also emerge as a function of the way firms choose to organize internally. Such choices involve the extent to which firms are integrated and absorb sequential steps in procurement, production, and marketing and the degree to which they are dis-integrated and contract with other firms in allocating these tasks. These choices of firm structure influence the locations of key assets and the routes by which flows of economic activity between these assets circulate. In this way, innovations in business organization deriving from the influence of the communications revolution and the process of organizational learning reshape territories of profit making. The outline of this route from communications to territory, which is depicted in Figure 2.1, can thus be summarized schematically as follows. From a given profit-making environment, new transport and communications technologies emerge, establishing the initial impulses of the communications revolution. These new technologies become commercialized by entrepreneurial infrastructure builders, who deploy and build out new transport and communications systems. When reaching a certain threshold, this build-out creates fundamental changes in the geography of economic activity marked most decisively by reconfigured systems of access across and within market boundaries and recalibrated systems of costs. Such changes in economic geography create opportunities for firms to accumulate profit differently. At the same time, the new infrastructure provides firms with tools to develop capabilities for implementing routines conceived from a new strategic vision of profit making. In this environment of initial infrastructure build-out, certain businesses successfully learn about and exploit the new infrastructure and the new structures of access and costs, space and time, created by these systems in developing business models to compete differently. Emerging from this learning process are initial experiments with innovative routines for profit mak-

Communications Revolution'

i

1-'-------Territorial Transformation

∆ Market Rules

Profit-making Environment'

Transport Communications Infrastructure Build-out

Communications Revolution ∆ Flows of Product/ Information

i

∆ Technology

Profit-making Environment

n

∆ Economic Geography

∆ Business Organization ∆ Business Strategies and Routines

figure 2.1 Schematic Outline of Communications Revolution

c

Experimentation Innovation (Users)

A Synthesis 59

ing. Gradually, as firms develop and implement new knowledge, they create new capabilities including work systems and forms of labor control, at times openly contentious, to undertake these innovative routines. As they enhance their capabilities, firms transform the organizational structures through which they carry out their operations and compete. Innovative business organizations, in turn, recast geographical landscapes for profit making by shifting the locations of productive assets and by rerouting flows of activities between these assets and the assets of other entities with which they interact in the course of procuring, producing, and selling. In certain instances, profit making in these reconfigured territories requires new systems of rule making to accommodate the innovations in routines and the business organizations developed for the new activities. Collectively, these changes—transport and communications technology, economic geography, business routines, and labor systems, organizational structure, territorial transformation, and market rules—produce a new profit-making environment (profit-making environment prime in Figure 2.1). This environment then paves the way for the next communications revolution—communications revolution prime—and the process continues.26

chapte r

3

The Railroad and Telegraph as Commerce System and Market Space

No doubt, the most significant object in the office was the ticker. . . . The offices of the ranches were thus connected by wire with San Francisco, and through that city with Minneapolis, Duluth, Chicago, New York. . . . During a flurry in the Chicago wheat pits, Harran and Magnus had sat up nearly half of one night watching the strip of white tape jerking unsteadily from the reel. At such moments they no longer felt their individuality. The ranch became merely the part of an enormous whole, a unit in the vast agglomeration of wheat land the world round. Frank Norris, The Octopus, 1901 As the organized type of society develops, the fusion of the various elements entails the fusion of the markets into one single market. . . . The result is that each industry produces for consumers who are dispersed over the length and breadth of the country, or even the whole world. Emile Durkheim, The Division of Labor in Society, 1893

Communications and Commerce When Swift began sending shipments of beef from Chicago to New England in the late 1870s, the communications revolution of rails and telegraphy—the precondition for the company’s long-distance system of production and distribution—was already in an intermediate stage of build-out. Two critical attributes of the infrastructure enabled Swift to reconceptualize the existing practice of beef slaughter and sale and create an enormously innovative network for producing and distributing beef in large volumes on a national scale. First, the rail and telegraph system had essentially become a commerce system. It had evolved into an infrastructure for moving freight and facilitating

64 The Railroad and Telegraph the exchange of information and messages needed to coordinate such movement of commodities. Such a development as a commerce system was not necessarily preordained. Nevertheless, soon after the initial deployment of the rail and telegraph systems, business uses—shipping goods and sending messages for buying and selling these commodities—assumed the dominant form of use on these new infrastructures. Second, Swift took advantage of the rail and telegraph not only as a freight-moving and transaction-facilitating system but also as an infrastructure with an increasingly national reach. This process of national integration, however, was not geographically uniform in its effects. At the outset of system expansion, linkages between the Northeast and the Midwest emerged far stronger than North-South connections. What resulted was a more vibrant set of trade relationships between the eastern seaboard and the agricultural heartland of the Midwest (Fishlow 1965, 262). Only after the 1870s did North and South become more integrated on the basis of rail and telegraph connections, though never to the extent of East-West connections. The evolution of the organization created by Swift reflected this spatial pattern. The firm relied upon the more established East-West links forged by the rail and telegraph system in launching its system of fresh beef shipments from Chicago to the Northeast in 1878–80. Gradually the firm expanded the geographical reach of these shipments so that the network extended into virtually every corner of the country. This chapter describes how the rail and telegraph revolution became a nationally oriented commercial revolution that provided the Swift Company with the foundations for its procurement, production, and distribution network. It examines how this communications revolution established an environment of opportunity for innovation. The rail and telegraph revolution created this environment by providing business users with a vision of how the geography of markets could be expanded and reconfigured. Rails and telegraphy enabled users to exploit an expanded and reconfigured structure of markets by redefining what was economically viable and efficient for firms across regions and within localities. This chapter also reveals how the politics and rule-making environment for these enlarged markets evolved as Swift and other large-scale firms sought to protect the national networks they had created. Finally, this chapter reveals how the rail and telegraph revolution rearranged trade linkages in the U.S. urban system and created mass consumer markets in cities that Swift used in building its national beef empire. infrastructure interconnection The commerce-oriented transport and communications system that provided Swift with the preconditions for its network began to take shape over the territory of the United States in the decades after 1850 (Fishlow 1965,

The Railroad and Telegraph 65 table 3.1 Expansion of Rail and Telegraph Systems, 1848–1902 1848

1852

1860

1870

1880

1890

1902

Rail track (thousands 5.9 of miles) Freight (millions of tons) 10.6 Telegraph wire (thou3.4 sands of miles) Messages sent (millions 0.5 per year)

12.9

30.6

52.9

115.6

208.2

252.5

17.6 23.3

46.5 56.0

147.6 133.6

338.9 291.2

691.4 1,200.7 848.8 1,307.0

1.4

5.0

11.5

31.7

58.4

89.7

sources: Alfred D. Chandler Jr., The Railroads:The Nation’s First Big Business (New York: Harcourt, Brace & World, 1965), Tables 1 and 2; Richard D. DuBoff, “The Telegraph and the Structure of Markets and Firms,” Research in Economic History 8 (1983): Table 1, p. 256; Edwin Frickey, Production in the United States, 1860–1914 (Cambridge: Harvard University Press, 1947), Table 13, p. 100.

262; Thompson 1947; Du Boff 1983; Table 3.1). Evolving from myriad locally based rail and telegraph operations, the national character of this infrastructure emerged most dramatically when the breach separating the Atlantic and Pacific coasts was finally enjoined by telegraph wires in 1861 and by rails in 1869. Although these two events captured the imagination of the country and inspired subsequent historical accounts of the rail and telegraph system, the story of how this infrastructure became a national system occurred in the nation’s interior. Alongside the bicoastal reach of rails and telegraph wires, expansion in the total mileage within the agricultural heartland of the country and the connections such interior growth represented were decisive in the creation of a nationally integrated rail and telegraph transport and communications system. Yet while expansion in rail and wire mileage provided the basis of a nationally oriented transport and communications system, it was the symbiotic development of rails and telegraphy that enabled this expansion to occur. Despite very different technological origins and initially separate development trajectories, the rail and telegraph infrastructure assumed its national orientation after 1850 as an essentially single transport and communications system. Telegraph wires expanded along railroad rights of way, and telegraphic communication became indispensable by the mid-1850s in enabling the railroads to operate safely. In 1849 the New York and Erie Railroad pioneered the use of the telegraph on its lines to control operations. Five years later, the telegraph was standard equipment for the scheduling of all rail operations. Railroads furnished transportation and materials for the construction of telegraph wires and operated telegraph offices from their depots. By 1870, U.S. railroads had funded and were managing two-thirds of Western

66 The Railroad and Telegraph Union’s twelve thousand telegraph offices (Chandler and Cortada 2000, 12). Although the capital and labor requirements to build and operate the two infrastructures were vastly different, rails and telegraphy became fused as the “twins” of nineteenth-century commerce (Field 1992).1 Prior to the Civil War, however, the rail and telegraph system suffered from numerous obstacles to interoperability. These problems were due in no small part to the large number of firms in each industry. In 1850, roughly five years after commercialization of the telegraph, there were already hundreds of telegraph companies (Jones 1852; Thompson 1947). During this same period, there were at least an equal number of rail firms, a number that eventually reached over two thousand (Chandler 1965). As a result of such unbridled competition, railroad track gauges, locomotives, and railcars varied throughout different regions of the country. Seven different track gauges were still in use in 1860 (Friedlander 1995). Similarly, telegraphy operated on competing Morse, Bain, and House technologies. Points of transshipment between competing lines, whether rail or telegraph wires, were sources of bottlenecks, where freight and messages had to be modified in order to accommodate the standards on the next leg of the route. These transshipment points added both time and costs to the movement of freight and messages over the rail and telegraph infrastructure. Without standards for interconnection, the rail and telegraph system was not able to fully exploit its technological capabilities for efficient long-distance transfer of goods and information.2 As the builders of these systems consolidated after 1850, they gradually addressed many of the most urgent problems of long-distance interconnectedness. Rail and telegraph builders began to establish uniform standards for track gauge and telegraph wires that eventually integrated rail and telegraph lines into a more unbroken network for freight and message traffic. They improved the technology of locomotives and freight cars that enabled the system to increase the size and speed of freight shipments. At the same time, builders agreed on construction standards for both track and wires. Steel replaced iron rails, and iron became the standard over less durable copper telegraph wires. Finally, railroads managed to overcome the problem of multiple local time standards in the United States, where in 1870 there were still roughly two hundred local times and eighty different railroad times in use. In 1883, railroads imposed a uniform time that enabled them to end confusion and operate more profitably (Kern 1983, 12). These ameliorations— technological, structural, and even social—resulted in more efficient overland movement of merchandise and facilitated the message transmissions for both rail companies carrying freight and business users of the infrastructure shipping and receiving merchandise.

The Railroad and Telegraph 67

interregional trade and mass markets The outcome of these improvements was a shift after the 1850s from water transport as the primary means of conveyance for interregional commerce to a rail- and telegraph-dominated system of overland long-distance trade. In 1850, despite gains by the railroads during the previous decade in hauling freight, boats and barges still dominated interregional transportation of bulk agricultural goods (Fogel 1964, 22). Of the estimated $1.46 billion in internal commerce in 1852, almost two-thirds traveled over water courses (Schmidt 1939, 818). Ten years later, however, the situation was reversed. By 1862, rails accounted for roughly two-thirds of the long-distance freight traveling from West to East. As rail use expanded, river routes declined. By the time of the approaching Civil War, the Mississippi had ceased to be an economically viable transport route used by grain merchants (U.S. Bureau of the Census 1864, clv). The decline in canals stemming from rail competition occurred more slowly, but the final outcome was little different from the fate of river routes. Although still widely in use during the 1870s and 1880s, canals were attracting very little business by 1890. Rails and telegraphy had subordinated rivers, canals, and lakes to a position of largely secondary importance in the nation’s internal trade. This rail and telegraph revolution was most fundamentally a marketing revolution in the movement of commodities and information. This marketing revolution made possible more rapid, larger, more geographically extended shipments of goods for delivery at definite times in specific places (Du Boff 1983, 255). Long-distance overland freight shipments by rail and long-distance real-time information exchange by telegraph thus became inextricably linked. As goods and information circulated over a rail and telegraph infrastructure increasingly national in scope, changes in the nation’s structure of markets and system of cities resulting from this circulation redefined the parameters for the profit-making activities of firms. The communications revolution of rails and telegraphy opened the boundaries between localized markets prevailing before the 1850s and created a wider, more spatially extended market system (Du Boff 1983; Yates 1986). This infrastructure enabled economic actors to expand their business operations to more distant locations by diminishing geographical barriers on freight and information movements between market areas and by reducing the time needed for commodities and information to circulate from one area to another. The resulting enlargement of market boundaries led to the establishment of a more geographically extended system of overland interregional trade. Alongside this change in the market geography of the country was an

68 The Railroad and Telegraph equally profound shift in the nation’s urban system, where market activity was concentrated. By the late nineteenth century, the United States had become a much more highly urban society. During the decade of 1860–70, the increase in the urban population exceeded the increase in the rural population for the first time in U.S. history (Pred 1966, 18). Existing cities expanded, and an enormous number of new urban settlements emerged from 1850 to 1890. Perhaps more important, cities after 1850, especially those in the top ranks of the U.S. urban system, assumed a more multifunctional role complementing their traditional commercial activities. Cities became centers of factory industrialization (Pred 1977, 85). With their burgeoning populations of businesses and consumers, cities were enormous concentrations of supply and demand. In these environments, raw materials entered, intermediate goods circulated, and finished products exited. It was in cities where these items were also consumed in unprecedented quantities by both firms and the urban public. In emerging as concentrations of supply and demand within production economies, cities evolved into entrepôts of mass markets. Ignited by rail transport and telegraphic communications, these changes in markets and cities served as platforms upon which Swift launched its fresh beef network. The firm linked procurement of cattle raw materials and production in the West with final marketing of the finished dressed beef output in the East through a distribution network built upon the foundations of wider markets and the system of rail-and-telegraph-based overland interregional trade. At the same time, cities were the points of concentration in Swift’s procurement, production, and marketing activity. On the buying side, Swift found mass markets for its cattle raw materials first in Chicago and then in other cities on the frontier of the cattle range in the Midwest. The firm established its production facilities and marshaled its factory labor force in these same urban centers. In selling its products, Swift used the expanding populations of cities, first in the East and later in the rest of the country, as centers of consumer demand and mass consumer markets. The firm in essence appropriated the impacts of the communications revolution on markets and cities. It crafted its own business model of using refrigerated rail transport and telegraphic communications from the market-shaping and city-building impacts of this revolution in creating an innovative enterprise linking the mass production and mass distribution of fresh beef. In forging the long-distance connections for its high-volume production and distribution system and focusing its activities in urban centers, Swift reinforced the very transformations in markets and urbanization that made its enterprise possible. In the process, Swift played a critical role in establishing the linkages between production and distribution activities that helped define a national market space.

The Railroad and Telegraph 69

Markets as Territory Markets, as territories where economic actors produce, buy, and sell, have boundaries that define them as places. The geographical size of markets is dependent on the costs to and capacity of market actors to exchange goods and services over distance and to communicate information needed to negotiate exchanges (Du Boff 1980, 478). Distance and the time needed by economic actors to bridge distance in transporting merchandise and communicating terms of an exchange act as limits on the size of markets. Market boundaries become fixed at those points where goods and information cannot circulate beyond barriers of geography in a timely manner. They also become fixed where the costs of moving goods or securing information beyond such geographical barriers drive the prices of merchandise to unmarketable levels, that is, when the cost of transporting merchandise or communicating the terms of exchange exceeds the value of the merchandise at its origin. These “costs of transfer,” that is, the costs of transferring goods and information over distance, limit the size of markets (Ohlin 1933, 100). In addition to this calculus of time, distance, and costs of transfer, two additional factors—technology and politics—shape the boundaries of markets. Historically, transport and communications technologies have conditioned the size boundaries of markets by controlling how merchandise and information circulates and by influencing the costs of such circulation (Du Boff 1980, 479). Politics also shapes the reach of market areas owing to the role of political authorities in establishing rules for market behavior (Polanyi 1944; Braudel 1979, 1977; Christopherson 1993; Zysman 1994). Political rule making influences the extent to which market actors can engage in commercial activity over politically divided territories. Perhaps more important, market rules influence the rewards and penalties economic actors can expect from such activity in terms of private gain or loss, thereby influencing the choices of whether market activity is worth pursuing in the first place. In effect, time and distance, technology and politics converge in creating a cost structure for market activity that establishes market boundaries. Time and distance, measured by the costs of using transport technology to convey goods and services from one location to another, and the costs to economic actors of using communications technology to secure information integral to an exchange act in concert with politics in shaping the boundaries of market areas. Such a view of markets borrows elements from a long-standing geographical tradition. In the 1930s, German geographer Walter Christaller (1933) synthesized ideas about transport, distance, and costs in developing a theory of what he called “central places,” where goods, firms, and populations concentrate. At the core of his theory was the “principle of markets.”

70 The Railroad and Telegraph For Christaller, the market was a territorial unit. It represented a range corresponding to an upper limit in terms of the distance beyond which essential commodities or “central goods” could no longer be exchanged in a cost-efficient manner. The central places that formed around markets, however, could be reshaped by two other secondary principles (Christaller 1933, 76–80). The first was the principle of transportation. Conveyance of goods through improved transport broadened the range in which such goods could circulate, therefore expanding the market area. The second principle was politics. Christaller conceded that markets and central places frequently emerged and grew in conjunction with administrative decisions of government. For Christaller, the three principles interact in defining the boundaries of markets where trade can occur (Christaller 1933, 76–80). Christaller’s theory has an undeniable elegance in its explanatory power. In focusing on distance, transport costs, and politics, Christaller succeeded in uncovering basic principles of market configuration. For Christaller, markets configured from these three elements exist as hexagonally shaped territories delineating central places of varying importance in a hierarchical system of equilibrium. This elegance and simplicity, however, also reveals certain weaknesses. The model suffers from an overreliance on mathematical abstraction. Paradoxically, not only is real geography absent from Christaller’s featureless landscape. Missing are meaningful references to the history of how these market areas emerge and how they change over time.

Markets Before the Railroad and Telegraph Prior to rail transport and telegraphic communication, markets for most products and economic activities in the United States were predominantly local in scope (Schmidt 1939, 820;Yates 1986, 151). These markets, especially in the less populated western areas of the country, were characterized by high levels of self-sufficiency and low levels of trade and exchange (Schmidt 1939, 818). In these areas of self-sufficiency there was an underdeveloped division of labor. Products originated, circulated, and were consumed within close proximity. Relatively few transactions occurred across market boundaries (Yates 1986, 151). Typical of this local market structure was the early meat industry itself (Walsh 1982). Virtually every town had its own abattoir where cattle and hogs were processed or where farm-slaughtered animals were further butchered (Clemen 1923). Although the slaughter and consumption of fresh meat possessed unique attributes that in the absence of efficient transport and refrigeration technology made it an overwhelmingly local activity, markets for other products, especially consumer perishables but also durable

The Railroad and Telegraph 71

manufactured goods, possessed similar localized patterns of production and consumption (Chandler 1977). Highly variegated conditions of supply and demand for goods and services existed between local markets. Perhaps even more profound were differences in the prices of commodities from one local market area to the next (Carey 1988, 216). Such variation in markets had profound impacts on intermarket trade. In the absence of well-developed transport and communications links, variations between markets constrained intermarket economic activity in two principal ways: (1) by limiting exchanges of information between economic actors separated by distance about prices, quantities, and types of goods available in different markets; and (2) by imposing formidable obstacles on the overland movement of goods between market areas. Trade under such circumstances conferred high levels of risk and uncertainty on merchants contemplating intermarket expansion. The sources of such risk and uncertainty derived from the capabilities of existing systems of communications and transport and the role of such systems in the activity of trade itself. Trade has to be preceded by the exchange of information between two types of economic actors. Economic actors on the demand side (purchasing retailer, wholesaler, or industrial enterprise) need information on supply sources and prices, while agents on the supply side (farmer, wholesaler, manufacturer) need to know outlets of final marketing and what buyers are willing to pay (Pred 1977, 38–39). Shared information, in effect, is a precondition to trade. If information cannot be exchanged over distance in a timely and cost-efficient manner, the consequences for markets are increased risk, reluctance by economic actors to engage in trade between distant points, and preservation of market boundaries. Similarly, if goods cannot move between buyers and sellers across geographical barriers owing to underdeveloped transport links, the consequences are the same—low levels of trade over distance and market boundaries that remain narrow. In the period before the introduction of the rail and telegraph system, intermarket trade imposed formidable costs on most merchants that exceeded the potential gains of such trade (Pred 1966, 163). In the absence of rail transport and telegraphic communication, these costs stemmed from shipping difficulties and information deficiencies related to the barriers of distance between markets. The costs of overcoming distance and the time needed to bridge distance in sending and receiving shipments and securing information to consummate transactions for such shipments placed obstacles on intermarket exchange. These cost constraints kept markets localized and hampered intermarket growth well into the 1840s (Du Boff 1983, 257). Despite the dominance of localized markets and the limitations on intermarket commerce, there was in place during the early nineteenth century a

72 The Railroad and Telegraph system of long-distance trade. This trade, however, did not go overland. Commodities could not move overland for any appreciable distance without raising their prices to unmarketable levels. In 1816, the freight costs for shipping corn by wagon 136 miles amounted to its selling price (Pred 1977, 66). Wheat, although less costly to ship by cart, still exceeded its selling price at a distance of 330 miles (Riley 1911, 94). Clearly, such costs for the overland movement of commodities represented a formidable constraint on intermarket trade. More significant than the costs associated with long-distance shipments was the time associated with shipping freight interregionally. In 1817, it took fifty-two days to ship a load of freight from Cincinnati to New York using available wagon and river routes (Slaughter 1995, 6–7). By contrast, in 1852 the same freight shipment from Cincinnati to New York took only six days on the Erie Railroad and its feeder lines (Slaughter 1995, 6–7). Owing to these circumstances, interregional trade during this period circulated overwhelmingly via coastal shipping between coastal port cities (Pred 1977, 66). In the first two decades of the nineteenth century, long-distance commerce between regions occurred along two principal routes: (1) between U.S. coastal ports and European (mostly British) ports; and (2) between northeastern and southern coastal ports oriented primarily on the cotton trade (Pred 1977, 66–70). In many respects, these two routes overlapped. Much of the cotton trade originating from southern coastal cities was shipped first to northern ports and then transshipped to British textile mills. This routing formed both a domestic coastal trade and a foreign export trade. Similarly, a portion of European manufactures entering U.S. eastern ports was sent to southern port cities. Inland from this very narrow corridor of long-distance coastal trade, however, localized markets prevailed. Interregional commerce was an export-oriented, coastal-dominated activity controlled largely by eastern merchants favorably situated between the southern market and Europe. During this period, the western interior of the United States as a marketplace was of marginal interest to eastern merchants or southern cotton shippers (Pred 1966, 18–19). This interior area known at that time as “The West” engaged in comparatively little interregional commerce. Isolated geographically from the other two regions of the country, less settled, and without direct access to the Atlantic and Gulf coasts where the nation’s interregional trade took place, the West had an even more localized and selfsufficient system of markets than did the other two regions (Schmidt 1939, 800). The region was overwhelmingly agricultural, although some manufacturing, oriented mostly for the farm, was part of the settlement pattern in the West from the very beginning (Page and Walker 1991, 282).

The Railroad and Telegraph 73

Trade in the West was overwhelmingly intraregional. Exchanges of agricultural commodities and locally produced manufacturers occurred primarily among the region’s river ports as the territory developed its own internal trade system (Pred 1977, 69). This intraregional system of trade reinforced the region’s localized markets structure in which producing, buying, and selling occurred in largely self-contained geographical areas. Farmers in the West were caught in a cycle of inertia. Lacking easy access to long-distance markets, western farmers had little incentive to produce surpluses in sufficient quantity for export outside the region (Schmidt 1939, 800–806). Without surpluses, western producers had little reason to seek long-distance interregional markets.

The Beginning of Interregional Trade After 1820, however, as population and economic activity continued to expand west of the Appalachian Mountains and as cities, notably Cincinnati and Saint Louis, became trading centers, this pattern of self-sufficiency was disrupted, and the West emerged as the source for a very different pattern of long-distance interregional trade. Sparked by the growth of Western agriculture, this new pattern of interregional trade diverted long-distance freight shipments away from the coastal routes and oriented it along an East-West axis. More important, as the West became the nation’s agricultural breadbasket, this East-West commerce shifted from water routes to the rails. Thus, in the broad sense, the route to the rail-and-telegraph-based system of longdistance interregional trade emanated from what was perhaps the most compelling feature of American historical geography during the early nineteenth century, the expansion of population settlements and economic activity in the region west of the Appalachian Mountains (North, Anderson, and Hill 1983, 111).3 While western agricultural development provided the stimulus for the rail-and-telegraph-based interregional trade system, long-distance freight shipments to and from the West began on water. Steamboats and canals provided western farmers with their first opportunity to sell in distant markets. These early forays by the West into long-distance trade occurred initially in the markets of the South (Schmidt 1939, 801, 806). By the 1830s, the South had become dependent on the West for grains and foodstuffs (Schmidt 1939, 803).4 In responding to this demand, however, western farmers also exploited opportunities to sell in the eastern market by shipping agricultural commodities on steamboats and barges downriver to New Orleans. There, western farm exports were either distributed to southern market centers for consumption on the plantations or transshipped to the East. In both cases, access

74 The Railroad and Telegraph table 3.2 Percentage of Western Exports Shipped via New Orleans, 1839–1860

Flour Meat Corn All foodstuffs

1839

1844

1849

1853

1857

1860

53 51 98 49

30 63 90 44

31 50 39 40

27 38 37 31

34 28 32 27

22 24 19 17

source: Albert Fishlow, American Railroads and the Transformation of the Ante-Bellum Economy (Cambridge: Harvard University Press, 1965), Table 39, p. 284.

to long-distance markets through the river trade enabled western agriculture to change from a largely self-sufficient activity to a surplus-producing industry that by the 1840s rivaled the great grain producers of Europe. This transformation of western agriculture into an export-oriented activity played the central role in shifting the mode of conveyance for longdistance freight shipments from water courses to the railroads. At the same time, the surplus-producing agriculture of the West, coupled with the population growth of the Northeast, diverted the primary direction of interregional trade along an East-West axis. The sequence of events that triggered this transformation in both the directional axis and the mode of conveyance for long-distance trade begins with the completion of the Illinois and Michigan Canal in 1848. Built in response to the expansion of grain production in the Illinois River valley, the Illinois and Michigan Canal enabled wheat and corn from Illinois to be routed to New York and the Atlantic through the Great Lakes, the Erie Canal, and the Hudson River. This Canal-Lake course also sent western grain through the Saint Lawrence to Montreal. The impact of this route was to divert Illinois and western grain intended for eastern markets away from the river route, bypassing Saint Louis and, with more far-reaching consequences, New Orleans (Table 3.2). Perhaps more significantly, it created a strategic point in this East-West trade at the southern tip of Lake Michigan, where a small but growing urban settlement and grain market existed. This settlement was the city of Chicago. During the first season of the canal’s operation in 1848, corn shipments from Chicago, situated at the key crossroads of the canal-lake route, increased eightfold from 67,000 bushels to 550,000 as farmers discovered the advantages of shipments away from the Mississippi and New Orleans (Table 3.3). Perhaps more significantly, the canal and lake route, in establishing a direct link between East and West, shifted the direction of long-distance in-

The Railroad and Telegraph 75 table 3.3 Shipments of Grain and Flour from Chicago, 1840–1861 (Thousands of bushels)

Flour and wheat Corn Total grain and flour

1840

1842

1845

1847

1848

1851

1855

1861

10

587

1,025

2,137

2,386

799

7,1150

23,886

— 10

— 587

— 1,025

67 2,243

550 3,002

3,221 4,647

7,518 16,634

24,379 50,512

source: U.S. Bureau of the Census, Agriculture of the United States in 1860, Eighth Census (Washington, D.C.: Government Printing Office, 1864), Table H, p. cxlix.

terregional trade along an East-West axis. This axis of trade, in turn, reinforced a convergence of interests evolving between the West and the East. With an expanding agricultural surplus and a resulting growth in incomes, the West became increasingly attractive to the East, both as a source of foodstuffs and as a market for manufactures. The East, in turn, represented a far more lucrative market outlet for western grain farmers than the South. An entirely new system of extended markets and long-distance trade was thus emerging on the basis of a new relationship between the regions of East and West in which the interior of the country occupied the primary interests of the nation’s business class (Pred 1966, 16–18; 1977, 66–70).

From Water to Rail- and Telegraph-Based Trade In bridging the two regions, the lake route actually served as a catalyst for development of the rail and telegraph trade system by creating a privileged position at the base of Lake Michigan for Chicago as a concentration point and primary market for the East-West grain trade. It was this position as a primary grain market, in turn, that made the city attractive as a railhead. The Galena and Chicago Union Railroad completed in 1849 marked the first step in this evolution. Illinois grain farmers now had access to the Chicago market via a rail route. During the 1850s, Chicago became a terminus for twenty-one different railroad lines. The four eastern trunk lines converged on Chicago as the rail system began to develop its long-distance East-West linkages. The remaining seventeen lines extended into the city’s tributary agricultural area as small agricultural towns competed fiercely for rail stations (Riley 1911, 88–89). The construction of these roads from Chicago into the West, however, did not occur as a massive building project ahead of demand. Nor did the

76 The Railroad and Telegraph table 3.4 Rail Receipts at Chicago for Various Commodities, 1852–1856

Flour (barrels) Wheat (bushels) Corn (bushels) All grain (bushels) Lumber (feet) Coal (tons) Hogs

1852

1854

1856

124,316 937,496 2,991,011 4,195,192 147,816,000 46,233 65,158

234,575 3,038,935 7,490,753 5,726,968 238,337,000 56,774 138,515

410,989 8,767,760 11,888,398 25,817,248 441,962,000 93,020 220,702

source: Elmer A. Riley, “The Development of Chicago and Vicinity as a Manufacturing Center Prior to 1880” (Ph.D. diss., Department of History, University of Chicago, 1911), 94.

railroads create settlement and economic development in open territory as some exogenous force. Instead, railroads followed grain cultivation. Rail infrastructure was routed to western areas that were already surplus grain producers (Fishlow 1965, 165–235).5 This geographical bias is reflected in the relationship between the rail routes and the grain-growing areas. By the end of 1853, more than 60 percent of new railroad construction in Illinois occurred in the eleven leading wheat- and corn-growing counties of the state, areas which represented only 25 percent of total statewide land area (Fishlow 1965, 173). From its location at the base of Lake Michigan, where it had become a primary grain market, Chicago established a hegemonic position as a rail center for a broad range of commodities (Table 3.4). In the process, it emerged as the crossroads for a new system of long-distance trade between West and East. It was the railroad and telegraph and the focus of this infrastructure system on Chicago that differentiated the new system of long-distance commerce from the earlier water-based system of interregional trade. A small trading post of 4,853 inhabitants in 1840, Chicago grew to 29,963 by 1850, when the first large shipments of grain began to arrive in the city. By 1860, as railroads linked it with the East and traversed its agricultural hinterland, Chicago exploded, its population reaching 110,000 people. It was the center of the nation’s grain trade, a commercial enterprise described at that time as a “revolution” and “one of the chief marvels of modern commercial history” (U.S. Bureau of the Census 1863, cxxxv, clvii; Fishlow 1965, 289). Central to this revolution is the story of how the rail and telegraph trade system succeeded in replacing water routes for moving grain from the western grain belt to the East (Table 3.5). This revolution, however, is also

The Railroad and Telegraph 77 table 3.5 Rail and Water Shipments of Flour and Grain Received at New York, 1860–1890 (Millions of bushels)

Water Rail

1860

1870

1880

1890

41.1 16.0

36.3 34.2

71.1 98.0

30.2 90.2

source: Louis Bernard Schmidt, “The Internal Grain Trade of the United States, 1860–1890,” Iowa Journal of History and Politics 20 (1922): 105.

the story of how Chicago emerged as the privileged location for this railand-telegraph-based activity. By 1862, as Chicago asserted its centrality in the rail system, freight carried from Chicago on the Pennsylvania, Erie, New York Central, and Baltimore and Ohio Railroads reached 6 million tons, or roughly two-thirds of the total internal trade of just over 9 million tons. “In one word,” notes the 1860 census, “railroads did what could not have been done without them” (U.S. Bureau of the Census 1864, clxvi). Chicago, in turn, was the epicenter of this historical transformation.

Advantages of the Rail and Telegraph System As a means of conveyance for freight, the rail and telegraph system had a number of advantages over water routes. The most obvious advantage was speed. The velocity of shipments by rail shrank distances much more dramatically than did any previous form of transport (Cronon 1991, 74). Not only did freight arrive faster. The railroads’ liberation from geographical and seasonal constraints gave it other advantages over competing river and lake routes in terms of reliability. The river route south was extremely risky for grain shipments. River navigation in the southern states during summer months, especially in drought seasons, was often impossible. At the same time, owing to weather, the risk of damage to grain and flour from overheating as well as moisture were of particular concern to grain shippers. Such risks imposed costs in the form of high damage insurance rates and warehousing costs where shipments were interrupted and rerouted due to nonnavigable waterways. Similarly, lake and river routes in the north during winter months were plagued by the uncertainty of freezing. Speed and reliability, in turn, had considerable impacts on costs. Although freight rates on rivers and canals were considerably lower than rail rates, the

78 The Railroad and Telegraph

6 weeks 5 weeks 4 weeks 1 day 3 weeks 2 days

3 weeks

3 days 4 days 5 days 6 days 1 week

© Jeremy Shaw

2 weeks

map 3.1 Travel Time (from New York) as Market Space, 1830 source: Adapted from William Cronon, Nature’s Metropolis (New York: Norton, 1991), 77.

greater levels of certainty and the faster speeds of rail shipments narrowed such differences (Maps 3.1 and 3.2). The extent to which shippers discounted differences on paper between freight rates on water and rates on the rails is perhaps best reflected by the ever larger volume of commodities that began to move over the rail and telegraph system. In addition to its operational advantages of certainty, speed, and low costs, the rail and telegraph marketing system was able to expand owing to the

The Railroad and Telegraph 79

1 day 4 days

2 days 6 days

1 week

5 days

4 days

3 days

© Jeremy Shaw

6 days

map 3.2 Travel Time (from New York) as Market Space, 1857 source: Adapted from William Cronon, Nature’s Metropolis (New York: Norton, 1991), 77.

benefits conferred by government on rail and telegraph companies. Both rail and telegraph builders profited enormously from privileged access to the system of entitlements—especially development rights and land grants— provided by government as incentives to expand their systems (Berk 1994). Politics thus worked in tandem with technology in giving decided advantages to rail and telegraph builders in opening markets and creating a railand-telegraph-dominated system of long-distance interregional commerce.

80 The Railroad and Telegraph

Rails,Telegraphy, and the Livestock Trade While the grain trade was decisive in creating the conditions for a rail-andtelegraph-dominated system of interregional trade, it was the shipment of livestock from the western range areas to eastern markets that revealed most clearly the capacity of the rail and telegraph infrastructure to extend market boundaries and concentrate trade in cities.6 Unlike grain and other bulk commodities, the livestock trade in live animal shipments did not have a competitive means of conveyance over watercourses. For the most part, livestock had never been transported any appreciable distance on boats or barges. Hogs were far too difficult to manage on such trips, while cattle were too large and unwieldy. In the absence of a viable means of live animal conveyance prior to the railroads, markets for the slaughter of live hogs and cattle were overwhelmingly local in scope. The large-scale, high-speed overland movement of livestock and the establishment of an elaborate system of long-distance trade in live animals were entirely creations of the rail and telegraph era (Fishlow 1965, 68). Before the completion of railroad trunk lines, however, there was some overland movement of livestock. Hogs raised on western farms in the Ohio and Illinois River valleys were driven overland relatively short distances to Cincinnati and numerous other, smaller packing centers in southern Illinois. There the animals were slaughtered, butchered, and packed during winter months. Seasonality in the trade was essential because the curing process needed the refrigeration of winter so that the meat would not spoil as it cured (Walsh 1982). In contrast to hogs, cattle were driven great distances overland from the western cattle ranges to eastern markets, especially during the 1840s. These celebrated cattle drives, however, often covering over a thousand miles, imposed significant costs on the cattle raiser and cattle shipper. Cattle lost weight on such drives—anywhere from 150 to 250 pounds—and had to be fed and watered along the route. Along with the labor costs of droving, the losses accruing to cattle shippers amounted to roughly twelve to twenty dollars per head, depending on the length of the drive and the quality of the stock (Fishlow 1965, 68–69). At an eastern price of eight cents per pound for a thousand-pound beeve, or roughly eighty dollars per head of cattle, such charges for long-distance cattle shipments amounted to as much as a quarter of the animal’s value at market. Beginning in the 1850s, the railroads began to offer competitive rates to ship live cattle from the western range areas to Chicago, and then from Chicago to eastern slaughterhouses. As a consequence, the business of live cattle shipping grew spectacularly during the 1850s (Table 3.6). The same forces that made Chicago the largest interior grain market in the country el-

The Railroad and Telegraph 81 table 3.6 Live Cattle Shipments from Chicago, 1852–1861

Cattle shipped

1852

1854

1856

1858

1860

1861

77

11,221

22,502

42,638

97,474

124,146

source: Louis Bernard Schmidt, “The Internal Grain Trade of the United States, 1860– 1890,” Iowa Journal of History and Politics 20(1922): 105.

evated the city to the greatest collecting point for livestock (Chicago Board of Trade 1864, 46). This development had actually encouraged the cattlegrazing business to migrate further west to cheaper grazing areas, where an increasing percentage of the nation’s cattle was being raised. As a consequence, live cattle shipments emerged during the 1850s as one of the most lucrative businesses of the railroads (Chandler 1988, 230). By 1860, livestock shipments accounted for roughly 33 percent of total eastbound tonnage and about 50 percent of eastbound freight charges (Fishlow 1965, 79). On the eve of the Civil War, as the railroads assumed more and more control of the long-distance cattle traffic, the cattle-driving business had actually become obsolete (U.S. Bureau of the Census 1864, cxxxi). With the railroad and telegraph, the market in live animals reached from the western plains to the Atlantic coast. Consequently, this rail- and telegraph-based system of long-distance trade not only redefined the livestock business but fundamentally reshaped the geographical structure of markets in which this business operated.

The Institutionalization of Interregional Trade The system of enlarged markets and East-West interregional trade became strengthened and institutionalized through the creation of formalized commodity exchanges. The establishment of boards of trade in Buffalo (1845), Chicago (1848), Toledo (1849), New York (1850), and Saint Louis, Philadelphia, and Milwaukee (1854) and the emergence of the Chicago Board of Trade as the country’s central commodity exchange marked the stages in this process of institution-building and market integration. Rails and telegraphy played a defining role in the way these institutions formalized a system of long-distance intermarket trade that was interurban as well as between urban and agricultural areas. The rail and telegraph system provided unique advantages to buyers and sellers of commodities in terms of shipping and communications that were essential in reducing the risks necessary for promoting intermarket activity. By accelerating the speed of shipments, providing new levels of certainty in

82 The Railroad and Telegraph transporting goods over geographical barriers, and conveying information to traders about supplies and delivery schedules, rails and telegraphy enabled both shippers and buyers to negotiate the transfer and arrival of commodities across different market locales. Pivotal to this more extended intermarket system of buying and selling were two critical innovations pioneered by the newly institutionalized commodities markets—the forward, “to arrive” futures contract and the system of standardized grades for commodities. Both were dependent on the rail and telegraph. Prior to these innovations, long-distance trade between bulk commodity shippers and buyers occurred on the basis of the consignment system. The high levels of risk to both parties in this system undermined intermarket commerce. In the consignment system, sellers of bulk commodities delivered by rail a small representative sample from their larger lots of bulk goods to eastern buyers, who would examine the sample and bid over the telegraph for the remainder. By the time the shipment reached the buyer, however, prices might have already changed. In combination with the system of standard commodity grades, the futures contract enabled buyers and sellers to complete a transaction by telegraph without need for a specially shipped consignment. The buyer would know what was being purchased because commodities of a particular grade would be fundamentally identical, while both parties could lock in a price through the futures contract on delivery of the grade specified. Furthermore, futures contracts provided buyers and sellers with other types of riskreducing flexibility. In contrast with the consignment system, futures contracts permitted commodities to be transported and delivered at chosen future dates when processors of agricultural goods—millers, wholesale butchers—were ready to refine them or when retail grocers were ready to sell (Du Boff 1983, 259). The impact of this system was to spread more uniform and less risky exchange practices over a wider territory while involving a more geographically dispersed pool of buyers and sellers in long-distance trade. As a result, commodities moved farther and faster while market areas, widened by the expansion of the rail and telegraph trade system and the institutionalization of commodities exchanges, became more integrated and uniform (Du Boff 1980, 479). Price uniformity across geographical areas generally reflects an absence of barriers to trade resulting from distance. When geographical barriers to trade diminish owing to such phenomena as new transport and communications systems, the result is greater levels of intermarket activity. As intermarket trade expands, prices tend to equalize across geography (Ohlin 1933). In the absence of specific data on intermarket trade, the movement toward price uniformity or dispersion across geography is actually a proxy for levels of intermarket activity in the economy. Thus commodity prices in different ge-

The Railroad and Telegraph 83 table 3.7 Relative Wholesale Price Indexes in Different Cities, 1866–1868 and 1889–1891 (New York = 1) Average

Baltimore Chicago New Orleans New York St. Louis San Francisco

Standard Deviation

1866–1868

1889–1891

1866–1868

1889–1891

1.099 1.055 1.068 1.000 1.102 1.133

0.976 0.969 0.919 1.000 0.944 0.980

0.212 0.219 0.166 0.000 0.250 0.318

0.083 0.073 0.117 0.000 0.087 0.170

sources: Jaime Jue, “Price Convergence and Interregional Arbitrage Opportunities in the Postbellum United States: Evidence from the U.S. Department of War” (Unpublished paper, University of California, Berkeley, Department of Economics, 1998), Table 1; U.S. Senate Committee on Finance, Wholesale Prices,Wages, and Transportation: Report by Mr. Aldrich [known as Aldrich Report], 52nd Cong., 2nd sess. Senate Report 1493, 4 vols. (Washington, D.C.: Government Printing Office, 1893), vol. 4, Table 18. note: These indexes comprise eight commodities for which comparative data are available (beans, candles, coffee, flour, rice, soap, sugar, tea).

ographical locales are a measure of the degree to which the U.S. economy was moving toward a more unified national market tied together by interregional trade. The effect of the rail and telegraph was clear. These infrastructure systems evened out markets in space (Carey 1988, 217).7 Commodity price data suggest that the prices of commodities in different regions, which had started to equalize during the antebellum period owing to intermarket economic activity from canals and steamboats, continued to converge at an even faster pace after the Civil War (Slaughter 1995; Jue 1998; Table 3.7). Variations in regional markets did not disappear entirely. There remained critical issues of interconnection on the rail system not fully resolved until the 1880s with the use of one single standard track gauge. In addition, the South, largely excluded from the system of East-West trade that had evolved in the decade prior to sectional conflict, was still not fully integrated into the interregional compact that continued to evolve between East and West after Reconstruction. Nevertheless, as the railroad and telegraph became more interconnected and interoperable in the post–Civil War period, and as these systems enhanced the overland transfer of goods and enabled information to become more widespread and evenly distributed, this infrastructure created an environment of greater certainty in the long-distance shipment of merchandise. The result was an enormous expansion in

84 The Railroad and Telegraph the movement of freight and information over rails and telegraph wires, and as a consequence, a greater equalization of prices across geographical space. As shipments of goods expanded, and as commodity prices equalized, a more nationally oriented interregional trade system emerged by the last quarter of the nineteenth century. It was this market structure and system of trade that served as a platform for the business organization of Swift.

Urban Impacts of Interregional Trade As the rail and telegraph system was reconfiguring market boundaries, it was also engineering the equally profound shifts in the nation’s urban system that served as a second platform for the network of Swift. These changes in the U.S. system of cities were marked most decisively by an increase in both the urban population and the overall number of urban places. As market boundaries enlarged, and as market activity assumed a more geographically extended character, trade and production actually became more highly concentrated in cities. The territorial spread of markets and the concentration of this spatially extended economic activity in cities were thus part of the same process of urbanization and economic growth. This pattern of urbanization and growth was the basis for one of the most far-reaching changes in the late-nineteenth-century economy, the advent of mass markets. The creation of urban mass markets, in turn, both business and consumer markets, reflected two important consequences of urbanization: (1) a shift in the internal size ranking among the nation’s cities, and (2) a change in the pattern of geographical linkages between cities. Mass markets also reflected an even more profound transformation occurring within cities themselves—the emergence of cities as manufacturing centers (Pred 1966, 1977). In this role, cities served as the focus of production economies, where production factors, supplies of materials, and demand for intermediate goods used in manufacturing became concentrated. As mass markets, cities became the centers of powerful transformative feedbacks in the economy. The growth of urban populations, the shifting rank of cities, the changing directional patterns of interurban trade, and factory industrialization emerged as mutually reinforcing catalysts for economic growth. At the same time, alongside this activity was the continuing expansion, standardization, and integration of the rail and telegraph system. Transport and communications, manufacturing and urbanization, and mass markets in cities thus evolved as interactive elements in a pattern of evolutionary economic change (Table 3.8).

The Railroad and Telegraph 85 table 3.8 Urbanization, Railroadization, and Industrialization of the United States, 1850–1900

Total population (millions) Urban population Cities with population over 10,000 Railroad mileage (thousands) Telegraph mileage (thousands) Index of manufacturing

1850

1860

1870

1880

1890

1900

23.1

31.5

39.9

50.3

76.1

92.4

11.3% 62

19.7% 93

24.8% 168

28.1% 223

35.1% 363

39.6% 440

9.0

30.6

52.9

93.3

166.7

206.6

12.0

56.0

133.6

291.2

848.8

1307.0*



16

25

42

71

100

sources: Allan R. Pred, The Spatial Dynamics of U.S. Urban-Industrial Growth, 1800–1914: Interpretive and Theoretical Essays (Cambridge: M.I.T. Press, 1966), 17; U.S. Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1970 (Washington, D.C.: U.S. Department of Commerce, 1975); Edwin Frickey, Production in the United States, 1860–1914 (Cambridge: Harvard University Press, 1947), 10–11. *Mileage for 1902.

urbanization and city rankings The urban population of the United States, which reached 5 percent of the total population only in 1830, began to expand dramatically after 1850, when rail and telegraph firms began to standardize their infrastructure and create a more uniform and nationally oriented transport and communications system.8 By the eve of the Civil War in 1860, roughly 20 percent of the population was living in cities. In 1890, the country’s 22.1 million urban inhabitants represented 35 percent of the total U.S. population. By the end of the century, roughly 40 percent of the nation was urbanized. Equally dramatic in this period was the increase in the number of urban places. In 1850 there were 236 cities, that is, places with a population of at least twenty-five hundred. By 1890, there were 1,348 such places. Even more profound was the increase in larger cities. Four cities in 1850 had populations in the range of fifty thousand to one hundred thousand. By 1890 there were thirty of these larger midsize cities. As early as 1820, the four largest cities of the Northeast—New York, Philadelphia, Baltimore, and Boston—had developed a system of large-city interdependence in terms of both imports and exports (Pred 1977, 67). These cities also dominated the trade with the South, where, unlike the Northeast, there was very little intraregional economic activity in the period before the rail and telegraph system. This southern trade with the cities

86 The Railroad and Telegraph table 3.9 Size Rank of Largest Cities, 1820–1850 and 1850–1890 1820 Rank 1850 Rank New York Philadelphia Baltimore Boston New Orleans

1 2 3 4 5

1 2 3 4 5

New York Chicago Philadelphia St. Louis Boston

1850 Rank

1890 Rank

1 19 2 7 4

1 2 3 4 5

source: U.S. Department of the Interior, Census Office, Report on the Population of the United States at the Eleventh Census: 1890 (Washington, D.C.: Government Printing Office, 1895), 370–72.

of the Northeast was controlled by Charleston and New Orleans. Such control established clearly defined patterns of interurban commodity movements between the southern and Northeast regions. Furthermore, New Orleans was also in a position to control commodity movements from the Ohio and Mississippi River valleys to both the Northeast and the South. These goods went by canal and river routes through the Ohio and Mississippi River systems to New Orleans, where they were then transshipped to other southern port cities or to the port cities of the Northeast. These interurban connections defined the geographical pattern of the interregional trade system before the rail and telegraph system came into use (Pred 1966, 1977). As the urban population increased after 1850, the ranking of first-tier cities within the nation’s urban system shifted in a dramatic way (Table 3.9). One city in particular, Chicago, grew unlike any other. Its emergence as the nation’s second-largest city by 1890 was vastly different from the pattern of rank size city growth from 1820 to 1850, which was marked by relative stability in the size rank of the nation’s very largest cities.9 The emergence of Chicago transformed the economic interdependencies and the geographical linkages between cities and the regions where they were located. In the process, Chicago helped reshape the pattern of both interurban and interregional commercial relationships that had been established in the mercantile era. As Chicago expanded by taking advantage of its position of primacy within the rail and telegraph system, and as it evolved into the nation’s premier commodities market, it became the focal point of the nation’s EastWest interregional trade system. Chicago successfully siphoned off flows of grain from its primary competitor, Saint Louis. It drove New Orleans and the river economy to a secondary role in the nation’s internal commerce system. It relegated the once dominant pork-packing city of Cincinnati to a

The Railroad and Telegraph 87 table 3.10 Growth of Manufacturing Employment in U.S. Cities, 1860–1890 Manufacturing Employment (Thousands)

New York Philadelphia Chicago St. Louis Boston Baltimore Pittsburgh Cleveland Detroit

Population in Manufacturing (%)

1860

1890

1860

1890

106.2 99.0 5.4 9.4 19.3 17.1 8.8 3.5 2.3

477.2 260.2 210.4 94.1 91.0 83.7 56.4 50.7 38.2

9.0 17.5 4.8 5.8 10.8 8.0 18.0 8.0 5.2

19.0 24.9 19.1 20.8 20.2 19.3 23.7 19.4 18.5

source: Allan R. Pred, The Spatial Dynamics of U.S. Urban-Industrial Growth, 1800–1914: Interpretive and Theoretical Essays (Cambridge: M.I.T. Press, 1966), 20.

secondary position, both as a livestock market and a meatpacking center. Its control of the nation’s internal trade and its emergence slightly later as an industrial center became sources of wonder even to the actors at the Chicago Board of Trade who made this happen.10 mass markets and manufacturing From 1860 to 1890, the top-ranking cities in the United States became characterized more by industrial and multifunctional activities and less by mercantile and trading functions (Table 3.10). This manufacturing activity also became increasingly concentrated in fewer, larger cities. In 1860 the ten largest cities accounted for 24 percent of all U.S. manufacturing value added. In 1890 the figure for the top ten cities was 38 percent (Pred 1977, 85). This change toward manufacturing was one of the most significant attributes of the urbanization process that enabled cities to emerge as mass markets. Owing to this transformation, the channels of interdependence among larger cities within the U.S. urban system increasingly involved commodity flows tied to both manufacturing activities and consumption. Such flows of incoming inputs and outgoing manufactures, both finished goods for final demand and intermediate goods for other factories, were not limited to durable products. These flows also included an exploding interurban trade in foods fabricated in factories to feed an increasingly urban manufacturing population (Pred 1977, 94). Such foodstuffs linked agriculture to industry

88 The Railroad and Telegraph and production to consumption in a reconfigured interregional and interurban trade and manufacturing system. Continued expansion and standardization of the nation’s rail and telegraph system, and the dramatic reductions in freight charges and communication costs that accompanied this build-out, enhanced the phenomenon of urban mass-production activity. From 1865 to 1890, costs for moving commodities on a per-ton-mile basis decreased by roughly 75 percent (Pred 1977, 94). These cost reductions, in turn, provided opportunities and incentives for firms to produce in high volumes. As cities evolved into concentrated sources of supply and demand for manufacturing activities within a system of enlarged markets, and as improvements in the transport and communications infrastructure enhanced the capacity of firms to produce in higher volumes, factory production became firmly anchored to cities. In this way, transport economies, information economies, mass-production economies, and urbanization economies were connected to a pattern of economic growth focused on urban mass markets and interregional trade. This sequence of transport economies, information economies, internal scale economies, and urbanization economies thus became mutually reinforcing, enabling the rail and telegraph revolution, mass production, interregional trade, and the urban system to evolve along the same trajectory.

The Politics of Market Space If this enlarged structure of markets and system of long-distance interregional and interurban trade begins with technology, it assumes its more complete form in the realm of politics. By the 1870s, when patterns of national market integration had become widespread, numerous state governments, responding to local business interests, succeeded in creating a web of regulatory barriers to internal commerce. These barriers were designed to protect local business firms in certain states from the competitive encroachments of firms from outside the area. As market areas enlarged and as the boundaries between markets blurred, business firms became more vulnerable to competitors residing outside the former market boundaries. Although efforts by local governments to protect local business interests by restricting commerce between states conflicted with the spirit of the commerce clause in the Constitution, states nevertheless took advantage of their own broadly defined authority known as states’ rights to circumscribe interregional trade within narrow limits. Such protectionist rule making on the part of state legislatures contributed to a specific type of market inefficiency. Protections extended to local business firms often enabled local merchants to assume what was effectively monopoly power in those markets where they operated. As late as 1875, the fed-

The Railroad and Telegraph 89

eral courts said nothing to disturb this impulse of states to intervene and protect local businesses, despite the efforts of the constitutional framers to create a single national market (McCurdy 1978, 635). States used the doctrine of states’ rights to implement protectionist legislation in two ways. First, certain states required nonresident salespersons to pay higher licensing fees than local merchants. For example, in the case of the Singer Sewing Machine Company, these fees aimed at restricting the company from setting up its network of retail outlets to compete with the firms of Wheeler and Wilson, and Wilcox and Gibbs. In a Supreme Court case in 1880, Webber v. Virginia, the Court, in acceding to the demands of Singer, struck down the rights of states to impose such fees. This decision revealed the extent to which the idea of a unified national market had become an integral part of the nation’s legal doctrine (McCurdy 1978, 642). The other mechanism implemented by states was the use of local inspection laws for food, most notably in the meatpacking industry itself. As the market power of Swift and other large interstate packers expanded during the 1880s, local butchers persuaded lawmakers in Minnesota, Indiana, Colorado, and Massachusetts to enact preslaughter inspections laws where beef and pork were sold. Although lawmakers claimed that their intent was to protect the public interest, these laws aimed at eliminating the ability of large packers to sell products transported across state lines. In this case, the remedy upholding the national market space was legislative. In 1891 the federal government authorized the Federal Meat Inspection service to conduct federal inspections of beef and pork produced for interstate sales. While these cases reflected the shift toward greater levels of long-distance trade, they also revealed the role of corporate political power in establishing policy in the marketplace. At the same time, the changing legal framework for interregional economic activity reflected new thinking among policyoriented academics about the market itself (Parrini and Sklar 1983). In this new outlook, not only was the country supposedly better served by longdistance commerce and a unified national market. Large oligopolistic firms emerged in this new view of markets as the most efficient organizers of economic activity within a unified national economic space.

The Platform for Innovation at Swift Innovation at the G. F. Swift Company would be neither wholly fortuitous nor the result of individual genius. It emerged on the basis of widespread trends in the economic geography of the nation ignited by the communications revolution of rails and telegraphy and its build-out. First, Swift would establish its organization for producing and distributing beef on the foundations of the rail and telegraph infrastructure and the

90 The Railroad and Telegraph system of interregional markets created by the build-out of this infrastructure. Such markets differed from the highly variegated and localized markets prevailing at midcentury. Consequently, by 1875, when G. F. Swift first conceived of ways to reorganize beef production and sale, he had at his disposal for this learning process a transport and communications infrastructure sufficiently developed and a market structure capable of supporting his idea for a new type of production and distribution organization. Infrastructure, market structure, and the interregional system of commerce created by the two thus provided Swift with a critical platform for innovation. Second, Swift built its innovative production and distribution network on the basis of profound changes in the nation’s urban system. These transformations, which were linked to the rail and telegraph revolution, elevated cities as manufacturing centers, as concentration points for interregional commerce, and as mass markets. In establishing its business, the G. F. Swift Company took advantage of these urban characteristics and exploited what was arguably the most defining event in the urban history of the nation during the latter nineteenth century, the emergence of Chicago as the nation’s second city. As the nation’s rail center and main commodities market, Chicago forged powerful economic linkages with other cities throughout the country. Equally important, Chicago also established critical linkages with a vast agricultural hinterland, including the cattle-raising areas of the western range. These interurban and urban-agricultural connections were fundamental for Swift. Furthermore, the firm profited from the development of Chicago and cities in the Midwest as mass markets for the purchase of cattle raw materials and as sites for factory operations. Perhaps even more important, Swift exploited cities as burgeoning centers of consumer demand for foodstuffs. The firm developed its beef network on the foundations of these cities, relying on the enormous expansion in the number of urban places as mass consumer markets for dressed beef.11 Eventually, the location of the branches in the firm’s dressed beef network resembled the map of U.S. cities with populations of twenty-five thousand inhabitants. Innovation at the firm was thus intimately connected to the nation’s urban history. It was upon these structures—technological, interregional, and urban— that Swift would emerge as an agent for change.

chapte r

4

Continental Divide The Business Enterprise of G. F. Swift & Company

I was determined to eradicate the waste of buying cattle which had passed through the hands of too many middle men and against which too many charges had accumulated. Gustavus Swift, 1890 We followed a few years after Mr. Swift. We had to learn the methods of dressed beef and experiment. The methods have been gradually growing. We have perfected them year by year, learning something every year as to how to do the business better. P. D. Armour, 1889 All along this east side of the yards ran the railroad tracks, into which the cars were run, loaded with cattle. “And what will become of these creatures?” cried Teta Elzbieta. “By tonight,” Jokubas answered, “they will all be killed and cut up; and over there on the other side of the packinghouses are more railroad tracks where the cars come to take them away.” Upton Sinclair, The Jungle, 1906

Gateway to Innovation In 1879, architects Daniel Burnham and John Root, pioneers of the Chicago architectural avant-garde of the late nineteenth century, completed a commission for the Union Stockyard Company to build a monumental entry to the company’s sprawling facility on Chicago’s southwest side. When it was built, the Gate of Burnham and Root celebrated the ascendancy of a particular industry—meatpacking—and a particular place—Chicago, which had evolved into “the great bovine city of the world” (Wade 1987, 56–57). The gateway, however, proved to be more than a celebration of previous ac-

92 G. F. Swift & Company complishment. It was both homage to industrial creativity in the recent past and prophecy of creativity to come.1 During the previous two years, Gustavus Swift had initiated a series of experiments in shipping slaughtered beef from Chicago to New England in his own refrigerated railcars. While not the first to try this technique—the Chicago Board of Trade had already conceded in 1876 how such shipments represented “important traffic” (Taylor 1917, 684–85)—Swift was the first firm aiming to build a business around such a system. By 1878, Swift, undaunted by the skepticism of other meatpacking firms that dressed beef shipments would fail as a business model, began forging an extremely innovative though still experimental system for slaughtering beef in one location and selling the dressed beef in markets located great distances from the point of slaughter. This innovation would have decisive consequences. It transformed the entire meatpacking industry within the next decade and recast the organizational and territorial pathways for the accumulation of profit in the U.S. economy of the late nineteenth century. communications and creative destruction Central to Swift’s innovation was a new type of business organization for mastering the control needed to produce and distribute the highly perishable dressed beef product. Such control, in turn, hinged on the development of the rail and telegraph system (Chandler 1992, 264; 1977, 392, 402). Although surprisingly neglected by Schumpeter, the dressed beef enterprise of Swift is among the most innovative business breakthroughs of the late nineteenth century. As an upstart among more established firms such as Armour, Swift assumed this entrepreneurial role and built this organization by developing a new product—beef slaughtered and butchered in one location and sold in another—and more fundamentally, a new process for making and selling this new product.2 In engineering these changes, the company occupies a position of centrality in what Schumpeter insisted was the essence of capitalist development—“creative destruction”—rooted in the phenomenon of innovation and technological change. What Swift succeeded in destroying was the market structure and system of beef production, consumption, and sale prevailing before 1875. In this system, butchers slaughtered cattle and sold beef to consumers living nearby, creating a highly localized structure of markets for cattle slaughter and beef consumption. Long-distance shipments of live cattle by rail from West to East, which coexisted with this pattern and provided supplies for eastern butchers, did little to undermine the fundamentally localized relationship of cattle slaughter and beef consumption. In these conditions of market localization, myriad intermediaries—wholesalers, jobbers, and merchants—were the dominant actors, creating numerous—and according to Swift—ineffi-

G. F. Swift & Company 93

cient steps in the process of beef production, slaughter, and final sale to retail butchers and consumers. By exploiting the rail and telegraph system to undersell existing butchers in local markets and make this new product affordable, Swift created the demand in eastern markets for his new product (Destler 1946, 42).3 As part of this effort to create a market, the company built a brand name, Swift “Premium,” and connected its brand to the newly discovered power of advertising in a concerted campaign to strengthen consumer preferences for Midwest-slaughtered beef among an initially skeptical public (Neyhart 1952, 154; Cockburn 1996). With an ever-increasing share of the consumer market for beef, Swift disrupted the relationships of supply and demand between buyers and sellers of beef that gave the beef trade and beef markets their local character. In the process, the company essentially destroyed the intricate system of intermediaries formerly prevailing in the beef trade. In place of these actors emerged a new type of business organization for coordinating the system of cattle slaughter and beef sale. the organization of swift The organization built by Swift for producing and marketing freshly slaughtered beef in high volumes over a national territory represented a revolution in logistics. This revolution elevated the role of process innovation as the route to profit making in the industry and transformed the business of meatpacking into a distribution business (Aduddell and Cain 1973, 97). This process-oriented innovation pioneered by Swift consisted of four fundamental elements. The core of the company’s innovative advance was the creation of a dispersed network of branch distribution houses for selling dressed beef to the firm’s final customers, retail butchers. This branch house network played a pivotal role in securing the demand necessary for the company’s dressed beef business to operate on the basis of high volumes. Through direct interaction with retail butchers, these branch houses provided the bridge between volume production, with its economies of scale, and mass consumer demand, which Swift had actually helped to create. It was the system of branch houses that enabled Swift to bypass the traditional intermediaries in the cattle and beef trade and create a more direct route to the customer. The direct system of moving beef from producer to retailer through the branch houses removed distribution bottlenecks created by the wholesale network that had constrained volumes and opened the channel for increased levels of throughput necessary for Swift’s system to operate profitably. Perhaps most important, this elimination of intermediaries enabled Swift to outcompete the two prevailing business models of the beef trade based on slaughter of cattle raised locally or alternatively on localized slaughter of cattle shipped

94 G. F. Swift & Company long-distance. Swift’s direct distribution system was, in effect, a nineteenthcentury form of disintermediation. It emerged as a pioneering example of institutional change within a wider reorganization in the nation’s wholesale networks as the mass production economy expanded and as vertically integrated manufacturers assumed a more dominant role in marketing their own products (Porter and Livesay 1971). Second, the high-volume system of cattle procurement, slaughter, and disassembly on the front end of the organization represented a pull system. Customer orders for cuts and grades of beef placed at branch houses by retail butchers and telegraphed daily to Swift headquarters were the basis for procurement orders of cattle supplies at the various stockyards to satisfy customer demand. Such a system “pulled” cattle supplies through the organization as orders were received instead of pushing finished product out of disassembly factories for sale to customers on the basis of demand forecasts. From its headquarters in Chicago, Swift & Company relied on telegraphic communication with its purchasing agents at the major livestock markets of Saint Louis, Kansas City, Saint Joseph, Saint Paul, and Fort Worth, where the company had slaughtering facilities, to convey orders to its buyers and receive messages on available cattle supplies. Swift, in effect, used the telegraph to balance cattle demand and supply in real time in this pull system while the railroad acted as the fulfillment mechanism for transporting supplies and finished product. Third, this organization incorporated a new structure of ownership and control over the adjacent steps in the production and distribution of fresh beef. A formerly disaggregated, locally oriented, and small-scale activity had given way to an industry controlled by vertically integrated, large-scale enterprises. As the first vertically integrated meatpacking firm, Swift was the forerunner of large-scale, nationally oriented business organizations. Along with five other packing firms, Swift was part of an oligopoly dominating the meat industry, a pattern that would characterize numerous other industries of the period (Chandler 1977). Finally, this organization created a new geography of economic activity incorporating tendencies of territorial spread and localized concentration. For the first time in history, shipments of fresh beef were transported great distances in refrigerated railcars. Such shipments helped to establish the foundations of a more uniform and integrated national marketplace, “annihilating space” in the process (Cronon 1991, 207–59). At the same time, as Swift was extending its network through the proliferation of branches throughout the continent, the company was concentrating economic activity in new locations. Chicago and then other cities in the Midwest where Swift expanded disassembly operations became place-based concentrations of slaughtering. Assuming the status of late-nineteenth-century industrial

G. F. Swift & Company 95

districts first observed by Alfred Marshall, these concentrations of slaughtering emerged as command centers of continentally organized industrial empires, gateways to a new territory of profit making.

Early Meatpacking In order to grasp the innovative and transformative impacts of Swift’s production and distribution network on the rest of the industry, it is imperative to understand the nature of the industry in which Swift initially operated. Until the last quarter of the nineteenth century, cattle slaughtering and fresh beef distribution were under the control of local butchers (Clemen 1923, 225). By 1890, however, local butchers in most places had become little more than retail conduits for a trade dominated by nationally organized firms. The business organization pioneered by Swift played the decisive role in this transformation by challenging the two most fundamental attributes of early meatpacking: (1) the decentralized markets of the industry, and (2) the dominance of pork over beef. In challenging these attributes, the organization of Swift recast the relationships between actors in the meat industry while creating a new type of firm, a more far-flung market geography, and a pattern of economic growth based upon high-volume, long-distance throughput. How the meat industry evolved from a locally oriented activity focused overwhelmingly on pork to a national activity in which beef rivaled and even surpassed pork in importance and how, as the beef industry expanded, the practice of shipping beef in carcass form great distances assumed ascendancy over live cattle shipments are the central themes in the Swift story of innovation and economic change.4 markets and meat Prior to the railroads and the refrigerated railcar, slaughtering and meatpacking was the domain of innumerable small farmers and firms that supplied local markets (Clemen 1923, 173;Yeager 1981; Walsh 1978, 1982). The geography and market structure of the industry was a mirror image of the localized markets that dominated the U.S. economy in the early nineteenth century (Pred 1966, 163–67). Virtually every town had its own slaughtering facility, usually located on the outskirts where urban and rural boundaries met, that provided for local consumption (Clemen 1928, 440). Large cities had multiple facilities. New York had more than two hundred abattoirs servicing various neighborhoods. Slaughtering and the consumption of beef were thus reflections of local markets that, in the absence of well-developed interregional transport and communications links, were largely self-contained.

96 G. F. Swift & Company Under these conditions, the overwhelming influence on slaughtering and meatpacking was seasonality. Grazing and feeding of cattle and hogs began in spring and lasted through the autumn. Slaughtering occurred at the end of the grazing and feeding season, while animals not yet mature were held and fed on corn during the winter. Because fresh meat was perishable, it had to be either consumed quickly after it was killed or preserved by salting, smoking, or curing. Meat preservation occurred in the winter, since the meat had to remain cool to prevent spoilage while the cure mixture penetrated the flesh, a process that lasted several weeks. Cured meat was eventually shipped on riverboats during the spring thaw (Yeager 1981, 5). This seasonal activity, undertaken by numerous small producers, supplied local markets under the guidance of the so-called invisible hand (Yeager 1981, 1). During the first half of the nineteenth century, consumption of pork far exceeded that of beef. The reason was that pork was more easily preserved and more readily eaten in preserved form than was beef. In addition, hogs were less expensive to raise than cattle (Yeager 1981, 4). Fresh beef was available, but conditions for the consumption of beef had severe restrictions. In the absence of rapid overland transportation, beef had to be butchered and sold close to where it was raised and slaughtered or it would spoil. This constraint limited the distance that slaughtered beef could be transported. Rarely did slaughtered beef, whether killed on farms or in abattoirs on the outskirts of towns, travel more than fifteen miles to market. Although river steamboats widened the meat trade after the 1820s, the growth occurred in the trade for cured rather than fresh meats. Because cured meats consisted overwhelmingly of pork, the expansion in meatpacking stemming from steamboats occurred in pork packing. Hauling of live cattle on steamboats for slaughter and sale in locales distant from where animals were raised was simply not technically or economically feasible (Yeager 1981, 4–5). The alternative to hauling live cattle was to ship freshly slaughtered beef by steamboat. Even this new form of transport, however, did not represent a sufficient technological advance to alter the local character of the beef industry. In the absence of refrigeration technologies, river steamboats did not provide enough speed to widen the market for freshly slaughtered beef. As a consequence, fresh beef remained a local product that could not be traded over long distances. Pork packing undertaken by numerous small and medium-sized firms and the transport of cured pork by river steamboats dominated the packing business in the early part of the nineteenth century.5 There was little in the way of an organized fresh beef industry in the United States. Cattle slaughtering for fresh beef was more an agricultural activity in which cattle merchants and local butchers acted as intermediaries for delivering beef from farms to consumers.

G. F. Swift & Company 97

the early beef trade Given the constraints imposed by distance and available transport, urban consumers in the population centers of New England and the Middle Atlantic states could avail themselves of fresh beef from two basic sources. One source was cattle raised on eastern farms. Networks of eastern cattle farmers, cattle merchants, and urban wholesale and retail butchers organized the beef trade that brought beef to eastern cities from nearby cattle farms. Sometimes the farmer would slaughter the cattle and sell the carcasses to a wholesaler, who delivered them to wholesale or retail butchers for further butchering and sale. At other times, the farmer would sell live animals to a wholesale agent, who would drive the live animals to a local abattoir for slaughtering or deliver them to a local wholesale butcher for slaughter and sale to retailers. Still another possibility was for the farmer himself to drive his stock to the local abattoir or to wholesale butchers for slaughter and sale. These networks of farmers, cattle merchants, and butchers became relatively well entrenched within localities. The other source for fresh beef was cattle made available in the East as a result of the much-celebrated long-distance cattle drives. The importance of cattle driving as a source of eastern beef consumption stemmed from the steady westward migration of cattle raising during the nineteenth century. Because it was extremely land-intensive, the raising of cattle gradually moved from the population centers of the East toward the frontier in search of open pasture. This migration increasingly separated the location of the country’s cattle from the human population and provided the fundamental stimulus for cattle driving. At their peak in the 1840s, cattle drives transported animals from Texas, the Great Plains, and the grain belt states to the eastern population centers on time-consuming and arduous journeys often exceeding a thousand miles (Clemen 1923, 174–79). Drives were seasonal, beginning anywhere from late February until June and terminating at eastern stockyards between April and August. Six principal actors played essential roles in cattle driving (Clemen 1923, 63–65). First was the cattle “grazer,” who raised the stock. Cattle “feeders” then purchased the stock in the fall and fattened the animals on corn during the winter months in corn belt states such as Illinois for marketing in the East during the following driving season. “Drovers” would drive the live animals to the East. At the end of the driving process, cattle were consigned by drovers to eastern cattle merchants. Slaughtering took place sometimes under the control of these wholesalers at eastern stockyards employing wholesale butchers, other times by large wholesale butchers in their own shops. Finally, wholesale butchers distributed dressed sides to retail butchers, who in turn would process the dressed carcasses for final sale to retail customers. The boundaries between these actors, however, often overlapped.

98 G. F. Swift & Company table 4.1 Size of Beef- and Pork-Packing Industries, 1870 Establishments Beef packing Pork packing Miscellaneous meat

36 206 17

Employees 435 5,551 499

Output ($ millions) 1.95 56.43 3.76

source: U.S. Bureau of the Census, The Statistics of the Wealth and Industry of the United States, Ninth Census, vol. 3. (Washington, D.C.: Government Printing Office, 1872), Table 8(c), p. 458.

Despite the high visibility of cattle drives and the creation of institutions along the trails for these drives such as livestock pens for animals and hotels and taverns for drovers, beef still remained a subordinate product to pork (Table 4.1). Pork packing tied to river conveyance was the overwhelmingly dominant trend in the packing business (Walsh 1978, 9). As late as 1870, cattle slaughtering and beef packing, as an industrial activity, was still extremely small compared to pork packing. Overall, the industry, with an output of $62 million in 1870, ranked as the eleventh-largest in the nation.

Rails and Cattle During the 1850s and after the Civil War, as railroads began to replace water as the primary means of conveyance for commodities, a profound transformation occurred in the beef trade. Railroads made possible the large-scale long-distance shipment of live cattle. This development, in turn, reinforced an already evolving geographical separation between cattle-grazing areas of the West and the population areas of the East. With railroads as transport and with the telegraph providing information on cattle prices and supply and demand conditions, the East became more viable as a market for long-distance cattle shipping from the western range. the stockyards In eastern cities, demand for fresh beef had outstripped the supply available from local cattle farms or through drovers.6 Rail transport enabled cattle from the West to be shipped “on the hoof ” to these burgeoning urban markets. Consequently, the 1850s and the period immediately following the Civil War witnessed the advent of a new business activity in the form of eastbound rail shipments of live cattle. This period also brought a powerful actor into the cattle and beef business—the railroads. In elevating Chicago to a position of primacy within the nation’s transport system beginning in the 1850s, the railroads created a new geography of

G. F. Swift & Company 99

concentration for the cattle trade. At this location emerged a huge cattle market and point of transshipment for the movement of cattle from the West to eastern slaughtering markets. Decisive in this process was the establishment in 1865 of the Chicago Union Stockyard Company. Capitalized at $1 million, the stockyard was spearheaded by the railroads and supported by Chicago meat-slaughtering firms with nine railroads subscribing $925,000, eight Chicago meatpackers contributing $50,000, and the public purchasing shares worth $25,000 (Wade 1987, 48–49). The aim of this project, located five miles southwest of downtown Chicago, was to create one large-scale centralized stockyard accessible to all the railroads as well as livestock dealers and meatpackers. Such a consolidated stockyard facility was entirely without precedent (Wade 1987, 50). Upon opening in 1865, the Chicago Union Stockyard was enormously successful both as a market for the local cattle trade and as a conduit for eastbound cattle traffic. By 1867, cattle were being driven from the ranges to so-called cowtowns of the West, such as Abilene, Kansas. These were essentially railheads where the animals were consigned to cattle merchants and then shipped by rail to Chicago. What emerged in the process was a differently configured territorial nexus between Chicago and the western grazing ranges, and Chicago and the East, with the Union Stockyards playing a vital role in the creation of this new geography. Central to this new system of cattle shipment was an entirely new institutional actor, the cattle commission house, which acted as a critical intermediary in the trade.7 Cattle commission firms were consignees operating in the Chicago stockyards. They helped make the market at the Union Stockyards by absolving cattle farmers and drovers of the need to travel with their stock to Chicago to secure their sale. Instead, cattle commission houses, working through local cattle merchants or their own agents in Abilene and other western shipping points, paid cash for delivery of stock prior to actual sale in the Chicago market and made a commission on the sale as a market maker. This system created new incentives for cattle farmers and drovers to make cattle available for sale to the Chicago market. In conjunction with the opening of the Union Stockyards, cattle commission houses provided critical institutional supports for Chicago’s cattle market and enabled it by 1870 to become “the great bovine city of the world” (Wade 1987, 57). one step forward . . . Despite advances in shipping cattle made possible by the railroads and the Union Stockyard facilities in Chicago, the new business of cattle shipping was not without problems. At the Chicago market, three issues confronted cattle shippers aiming to transport cattle by rail from sources of supply in the West to places of demand in the East. The first problem facing cattle buyers and shippers stemmed from the

100 G. F. Swift & Company system of intermediation in the cattle trade that had emerged following the creation of the Union Stockyards and the cattle commission houses. After 1865, localized cattle merchant networks that had formerly dominated the trade gave way to a more geographically dispersed and more numerous set of wholesale actors who entered the trade as adjuncts of the consignment system being perfected by cattle commission houses. As a consequence, myriad new jobbers, merchant wholesalers, consignees, and commission agents became part of the cattle-shipping business facilitating the movement of cattle from collection points such as Abilene to delivery points at eastern slaughterhouses. While this network made cattle distribution more efficient overall, the numerous exchange points in the network, between sources of cattle supply and shipper-buyers of cattle at the Chicago stockyards and between cattle shippers and retail butchers in the East, drove down margins for the business as whole. Shippers of live cattle purchased at Chicago for shipment East were therefore positioned within a selling chain with numerous steps that provided them with meager returns. Second, the cattle shipper confronted technical problems in successfully shipping live cattle on existing cattle trains. Cattle had to be fed and watered along the route. In addition, animals invariably lost weight during these trips, bringing in less money for cattle shippers at the point of sale. These difficulties, however, were far from the most onerous. A report by the Massachusetts Railroad Commissioners in 1870 describes in detail the ultimate risk facing cattle shippers in transporting cattle long distances by train: Cattle trains yield the road to most others and pass hours on the sidings; the animals are without food or water and often with insufficient ventilation in summer or shelter in winter; they are jolted off their legs and goaded till they struggle up, for they can not be permitted to lie down. They thus arrive at their destination trampled upon, torn by each other’s horns, bruised, bleeding, having in fact suffered all that animals can and still live. Under the most favorable circumstances they leave the train panting, fevered and unfit to kill; under the least favorable, a regular percentage of dead animals is hauled out of the car. (quoted in Clemen 1923, 195)

Finally, adding to the logistical challenges of shipping live cattle was perhaps an even more formidable obstacle—the fact that roughly 55 percent of the animal was inedible. This unusable part of the cow placed a freight cost burden on cattle shippers without the possibility of a return. As a result, the business of shipping live cattle was far more lucrative for the railroads. By the late 1850s, cattle shipments had emerged as the rail industry’s most profitable eastbound trade (Chandler 1988, 230). Upon completion of the Chicago stockyards, this trade expanded even more rapidly. Consequently, this was a business that the railroad companies aimed to protect. In order to maintain the cattle trade at profitable rates, the major live-

G. F. Swift & Company 101

stock-carrying railroads entered into pooling arrangements during the 1870s. The most notable of these agreements was the “Evener System,” in which the railroads and the largest cattle shippers worked in concert.8 Although the largest cattle shippers benefited from such cartel agreements, the fundamental problem of shipping live cattle—the freight charges levied on the inedible portion of the animal—remained unresolved even for the large cattle firms. Furthermore, despite cooperation between the largest cattle shippers and the railroads, these two groups did not meet in the marketplace as equals. The size of the railroads and their control over such a vital transportation service provided them with the power to maintain freight rates over cattle shippers.9 The need to overcome the cost problems posed by rail shipments of livestock compelled Swift to conceive of the production and sale of beef in a completely different way. In this reconceptualization, the idea of slaughtering and butchering meat in the West and shipping the dressed beef to the East over the rail system without spoilage was the basic problem that Swift would try to resolve. The key obstacles facing Swift in trying to overcome this problem were thus time and distance: the window of time before the meat spoiled and the distance between the disparate locations of slaughter in the West and sale in the East. The solution was both technological and organizational. Through trial and error, Swift came to understand that the solution to overcoming the barriers in shipping dressed beef was an integrated production and distribution strategy. On the technical side, the route to this integrated strategy was to link breakthroughs in refrigeration to the transport and communications infrastructure of rails and telegraphy. Organizationally, G. F. Swift confronted two issues. On the one hand, he had to confront the power of the railroads to set freight rates. Second, Swift somehow had to bypass the expanding channel of intermediaries in the rail-based cattle and beef trade. Dressed beef shipments on a large scale were incompatible with the organizational channels of distribution that had emerged in the cattle and beef trade (Porter and Livesay 1971, 168–73). Many of these distributors in the East were incapable of handling dressed beef, while others were openly hostile to the product, fearing the competition that dressed beef represented to their businesses with live cattle. Ultimately, the key to resolving these problems resided in the creation of a new type of business enterprise. Such an enterprise had to possess the capability of coordinating the various activities of procuring and slaughtering cattle, shipping dressed beef long-distance, and distributing the dressed beef in these geographically dispersed market areas. Only with such an organization could Swift integrate production and distribution and achieve the volumes necessary to make the

102 G. F. Swift & Company dressed beef project feasible. While the integrated organization pioneered by Swift represented the solution to this challenge, the initial steps leading to the creation of this organization actually began as experiments with the technology of refrigerated transportation. In these experiments, Swift was able to take advantage of earlier precedents in the technology of refrigeration.

Early Refrigeration By the close of the 1860s, the idea of adapting principles of refrigeration to transport systems for shipping perishable items was particularly compelling to early meat industry visionaries who influenced Swift. Initial efforts to ship refrigerated beef, however, were small-scale and still dependent on cold weather. In 1868 Henry Peyton Howard made an experimental shipment in his own steam freighter but soon discovered that weather conditions in the gulf area where he was operating were too unpredictable to sustain the business. Another pioneer, Thomas Rankin, shipped refrigerated beef from Denison, Texas, to New York City in refrigerated railcars but was forced to close his company after failing to achieve the volumes needed to make such a business profitable. The most viable precedent for Swift in terms of refrigeration strategies came from the meatpacking firm of George H. Hammond. In 1869 Hammond used a refrigerated railcar to ship eight tons of freshly slaughtered dressed beef from Detroit to Boston on the New York Central and Boston & Albany Railroads. Building upon this success, Hammond expanded his business after 1870, distributing his meat through existing networks of wholesalers and commission merchants. His railcar, however, had serious shortcomings, forcing him to confine his shipments to the winter months and thereby keeping his operation relatively small-scale. Because his business remained far too modest to compete with the live cattle shipment business, Hammond did not encounter resistance from the railroads. Nor did he meet with opposition from local eastern wholesale butchers since he did not undercut them on price. His western dressed beef cost the same or slightly more than locally slaughtered meat (Yeager 1981, 54). Nevertheless, Hammond’s business generated sufficient interest to lure other competitors into trying to perfect the system of grafting the technology of refrigeration onto transportation networks for shipping dressed beef from the West to the East. Firms run by Nelson Morris and T. C. Eastman experimented with refrigeration systems of their own. Others in the cattle and butchering business soon followed. By 1877, the ChicagoTribune wrote how the dressed beef business was now a significant activity (Taylor 1917, 549). Perhaps more important, the beef business in general was in the beginning stages of a large-scale

G. F. Swift & Company 103 table 4.2 Comparative Expansion of Beef and Pork Packing, 1870–1890

Establishments Employees Total output Pork, % of total Beef, % of total

1870

1880

1890

259 6,485 $62.1m 90.8% 3.2%

872 27,297 $303.6m 61.8% 27.8%

1,367 44,812 $564.7m 42.1% 34.4%

sources: U.S. Bureau of the Census, The Statistics of the Wealth and Industry of the United States, Ninth Census, vol. 3 (Washington, D.C.: Government Printing Office, 1872), Table 8(c); U.S. Bureau of the Census, The Statistics of Manufactures, Tenth Census, vol. 2 (Washington, D.C.: Government Printing Office, 1883), Table 7, pp. 464–65; U.S. Bureau of the Census, Report on Manufacturing Industries, Eleventh Census, part 1 (Washington, D.C.: Government Printing Office, 1895), Table 8, pp. 730–35. note: Pork and beef do not sum to 100 percent because of “miscellaneous” meatpacking products.

expansion (Table 4.2). It was in this environment that Swift would launch its own experiments in refrigerated dressed beef shipments over the railroads.

Entrepreneurial Opportunity G. F. Swift came to Chicago in 1875 to set up a cattle-buying office for his Boston-based wholesale meat business. At that time, refrigerated dressed beef shipments from the Midwest to the East, pioneered by Hammond and Morris, were expanding among other Chicago packers, although refrigeration techniques were still rudimentary. More important, there was no established business model capable of making the overland trade in dressed beef secure and reasonably profitable. The meat business was still dominated by live cattle shipments from the West through the entrenched channel of cattle merchants, brokers, and wholesalers. This experience from 1875 to 1877 of buying cattle in Chicago and shipping the animals East for his butchering business in Massachusetts brought Swift into direct contact with the economics and logistics of the live cattle-shipping business. It was the experiments in dressed beef, notably by his competitors Morris and Hammond, that especially drew his attention (Russell 1904, 23; Swift 1927). Not only was Swift looking for ways to avoid freight charges on the 55 percent inedible portion of the animals and eliminate the weight losses and bruising of the meat during shipment. Equally important to Swift was his determined effort to eradicate “the waste of buying cattle which had passed through the hands of too many middle men and against which too many

104 G. F. Swift & Company charges had accumulated” (quoted in Yeager 1981, 59). Because Swift understood the problem of supplying eastern markets with beef to be one of distribution—the damage to cattle from rail transport, the exorbitant freight rates on cattle, and the numerous intermediaries in the trade—he established as his objective the creation of a system for distributing his product differently. This objective, in turn, is what pushed Swift into experimenting with the idea of shipping dressed beef slaughtered in Chicago rather than shipping live animals. G. F. Swift seemed to possess the drive, skills, and temperament for this creative task. According to Charles Edward Russell, a radical critic of the Beef Trust, the “chief founder” of what was to become the dressed beef industry was big-boned, big-brained, resolute, indomitable, obsessed with that strange consuming passion for money-making that seems to be an exclusively American trait, a tireless worker, a devout Methodist, and of habits austerely and almost painfully correct. . . . He drifted westward to find some road to wealth. At the time Nelson Morris was experimenting with frozen meat in box cars, Swift was clinging desperately to the very skirts of the Chicago cattle market, a small speculator without capital or credit. . . . He was poor but was slowly forging ahead. The frozen-meat experiment came under his notice and he gave to it the searching scrutiny of an alert and powerful mind. . . . Mr. Swift studied this scheme and gradually unfolded in his mind a plan, having the prospect of enormous profits. (Russell 1904, 23)

In response to this challenge of seeking profit in a new way, Swift developed a business model linking four elements—refrigeration, transportation, communications, and organization—that succeeded in overcoming the fundamental obstacles to the long-distance trade in dressed beef. This business model, however, unlike that used earlier by Hammond, did not accommodate existing wholesalers in the industry. It obliterated them and created a cost structure based upon a branch system of distribution and high-volume throughput that exerted an irrepressible force on others aiming to compete in the industry.

Learning a New Business Swift launched his system in a series of experiments linking transportation and refrigeration. From the proceeds of his cattle-buying activity, he acquired sufficient capital for the purchase of his own slaughtering plant, which he used for initial trials of refrigerated dressed beef shipments (Yeager 1981, 58–63). For these early shipments, however, Swift used ordinary railcars, shipping the dressed product during winter months. He next rented refrigerated railcars from freight companies that were experimenting with

G. F. Swift & Company 105

the fresh fruit trade, but he was unable to lease enough of these cars to make the trade practical. By 1877, Swift decided to try to perfect his own refrigerator car as a first step in creating a dressed beef business. His partner at the time, however, James Hathaway, regarded the scheme as unsound, and the two dissolved the partnership. With thirty thousand dollars from the sale of his stake in the partnership, Swift formed a new enterprise with his brother, Edwin Swift. The aim of the Swift Brothers was to launch a business for procuring, slaughtering, shipping, and selling dressed beef. In 1878, Swift hired engineer Andrew Chase to help him perfect plans for an insulated and ventilated refrigerator car for the railroad. The design was highly innovative, with unique features that resolved some of the technical problems, such as moisture and ventilation, that had plagued cars built earlier for Hammond by William Davis. The central element of Chase’s design for Swift was a unique cold blast system that provided automatic circulation of dry air through the cooling compartment. Ice bunkers positioned in the upper corners of the car cooled the air, forcing warmer, lighter air to escape through ventilators. An advantage of these ice bunkers was that they could be filled from the exterior, making them easier and more efficient to operate. swift and the railroads When Swift approached the railroads of the Eastern Trunk Line Association to ask for assistance in building the cars, however, they rejected his request, fearing both the short- and long-term consequences such cars might pose to them as livestock shippers.10 Swift eventually found a builder in the Michigan Car Company, mortgaging his business as collateral for construction of ten cars. He then approached the Grand Trunk Railroad to enlist it as the carrier for his dressed beef. Unlike the railroads of the Eastern Trunk, the Grand Trunk Railroad had limited investments in the livestock trade and was therefore not threatened by Swift’s undertaking. Although its route east went through Ontario and Quebec, the Grand Trunk shipped into New England, where Swift intended to launch his initial dressed beef shipments. Furthermore, by giving the Grand Trunk his business, Swift offered the railroad a product capable of competing with the lucrative livestock-hauling trade of the other trunk-line rail carriers. There was another good reason, however, for Swift to give his business to the Grand Trunk. A competing railroad to Boston, the New York Central–Boston & Albany Railroad, carried the much smaller dressed beef shipments of his competitor, George Hammond. Swift’s refrigerated dressed beef shipments and the Grand Trunk Railroad were thus a natural combination. With a more technologically advanced refrigerator car and rail carrier in place, Swift began trial shipments of dressed beef from Chicago to New England in the winter of 1878. He established his first branch houses in Fall

map 4.1 The Dressed Beef Enterprise of G. F. Swift, 1878

Chicago

.......

Slaughtering facility Stockyard / cattle supplier Branch house

G

nk Railro a Tru d ran

d

© Jerem S h a w y

Fall River

Clinton

G. F. Swift & Company 107

River, Massachusetts, in a partnership with D. M. Anthony, who owned a wholesale butchering business there, and in Clinton, Massachusetts, where his brother Edwin managed the family slaughtering operation (Wade 1987, 106). These two locations formed the early foundations of the Swift beef network (Map 4.1). By 1878–79, Swift was sending shipments of refrigerated dressed beef on the Grand Trunk Railroad from a small slaughtering facility leased from Hammond to these two New England branches for distribution and sale. swift and the butchers Although Swift relied initially on the two branch houses in Clinton and Fall River for marketing dressed beef, his idea for expansion was to enlist existing wholesale butchers in New England as agents in a new system of distribution (Clemen 1923, 235). His strategy was to secure the cooperation of these actors as a means of gaining entry into local markets. His tactic was to buy stakes in these butchering firms. In this way, local butchers would be able to remain independent while becoming distribution agents for G. F. Swift, handling Swift’s dressed beef on a commission or consignment basis (Yeager 1981, 61). Swift used this strategy of alliances to build branches and secure larger sales volumes. Consequently, the early expansion of branch houses in the Swift network came from these consignment arrangements with existing wholesale butchers. More than mere conduits for sales, these relationships built by Swift represented a concerted effort to stem potential opposition to his expansion plans from New England butchers. Such opposition, however, was not long in coming. The hostility of eastern wholesale butchers to Swift stemmed from fear of being outcompeted by the new product and ultimately displaced. By slaughtering in Chicago and avoiding freight charges of shipping live cattle, Swift was able to sell dressed beef from seventy-five cents to one dollar cheaper per hundred pounds in New York or Boston than butchers who slaughtered rail-shipped animals (Yeager 1981, 62; Clemen 1923, 235). This differential provided Swift with a cost advantage over most eastern slaughterers of 3–5 percent for beef of equal quality. “You can ship about two cars and a half of live stock in one car as dressed product,” explained Albert McCurdy, a livestock commission merchant at the Chicago stockyards, “so that the cost of transportation is greater on the live stock than it is on the dressed product in proportion” (U.S. Senate 1890, 239). Although Swift’s dressed beef business was not profitable at the outset of establishing his branch system—he had high fixed costs initially for railcars and warehouses with cooling facilities—Swift was determined to maintain a pricing advantage over its competitors and endure possible losses in much the same way that modern firms use pricing to gain market share. Such price

108 G. F. Swift & Company Indirect Channel of the Older Beef Industry Cattle Grazers and Stockers

Cattle Commission Houses

Cattle Shippers

Wholesale Butchers

Retail Butchers

Swift's Direct Channel Cattle Grazers and Stockers

Swift

-··········>-

Retail Butchers

figure 4.1 Schematic Outline of Beef Channels

cutting, however, not only was aimed at securing customers but was also intended to convince butchers of the “wisdom” in joining Swift in the dressed beef trade. While some of these butchers joined Swift, on occasion the tactics used, in the face of resistance, were combative and ruthless.11 Yet even wholesale butchers with a more open approach to doing business with Swift were reluctant to handle the product because they were unwilling to make necessary investments in refrigerated warehouses for storing the dressed beef. Such butchering firms, while not actively opposed to the new product, rebuffed Swift’s overtures because they refused to pay what Swift regarded as a fair price for dressed beef, fearing that the product, in the absence of refrigeration, would spoil before they could sell it. Their reluctance and incapacity to handle dressed beef and the concerns that they harbored about being undersold by the new product and outcompeted turned most wholesale butchers into active opponents of Swift. It was thus the incapacity of and opposition from this eastern wholesale network that eventually compelled Swift to bypass existing outlets of meat distribution and begin setting up his own independent branch houses. This network of branch distribution, in turn, enabled Swift to outcompete and eventually subdue opposition from local eastern wholesale butchers by creating a more direct route to the final customer (Figure 4.1). As the competitive superiority of Swift’s system became apparent, other beef processors took notice. As a result, Swift’s branch house business model began to grow and diffuse to other firms.

Enterprise Expansion and Diffusion By 1880, Swift was operating twelve branches in New England. Two years later, Swift opened its first independent branch house in New York City, an event announced in an approving editorial in Harper’s Weekly. Editors at

G. F. Swift & Company 109 table 4.3 Chicago Meatpacking Firms Ranked by Size, 1878 and 1884 1878 Rank 1 2 3 4 5 6 7 8 9 10

Firm Chicago Packing Armour Fowler Bros. S. A. Ricker G. H. Hammond Tobey & Booth Davies, Atkinson & Co. Allerton Botsford & Co. Chapin & Cudahy

1884 Rank 1 2 3 4 5 6 7 8 9 10

Firm Armour G. F. Swift G. H. Hammond Anglo-American Libby, McNeil & Libby Fairbank Canning Chicago Packing B. F. Murphy Packing Hately Bros. Tobey & Booth

sources: Chicago Board of Trade, Twenty-first Annual Report of the Trade and Commerce for the Year Ending December 31, 1879 (Chicago: Knight & Leonard, 1879); Chicago Board of Trade, Twenty-sixth Annual Report of the Trade and Commerce for the Year Ending December 31, 1884 (Chicago: Knight & Leonard, 1885); Mary Yeager, Competition and Regulation:The Development of Oligopoly in the Meat Packing Industry (Greenwich, Conn.: JAI Press, 1981), 64. note: Rankings based on number of animals slaughtered and revenue generated from slaughtered animals.

Harper’s noted how other cities had benefited from the dressed beef trade and went on to describe the good fortunes of New York in gaining access to the lower prices of the dressed beef operators. The article concluded by marveling at how a side of beef leaving Swift’s plant in Chicago was hung on a railway hook, transferred to a refrigerated railcar, and pushed into the freezer of the branch house when it arrived in New York “without being removed from the hook on which it was hung when killed” (Harper’s Weekly 1882). During the 1880s, as Swift’s system of long-distance dressed beef shipments emerged as a competitively viable business model, two outcomes from that model began to change the structure of the industry. One outcome focused on Swift’s growth in relation to other meatpacking firms. As a result of its branch house system, Swift broke into the top ranks of the industry (Table 4.3). By 1884, after only six years in operation, Swift had become the second-largest meatpacking firm in the United States. Only Armour was larger. Second, and perhaps more significant, with the ascent of Swift among the top ranks of meatpacking firms, other packers began to imitate key elements of Swift’s production and distribution network. The most decisive element in this process of imitation was Swift’s branch house system. P. D. Ar-

map 4.2 The Dressed Beef Enterprise of G. F. Swift, 1880

Slaughtering facility Stockyard / cattle supplier Branch house

Chicago

G

nk Railro a Tru d ran

d

~··

© Jeremy S h a w

G. F. Swift & Company 111 table 4.4 Number of Branch Houses of Swift and Major Firms, 1878–1899

Swift Armour Morris S&S Cudahy All Firms

1878

1880

1884

1888

1895

1899

2 — — — — 2

12 — — — — 12

43 2 — — — 45

67* 10 9 2 1 89

138 125 61 31 28 383

189 152 87 42 47 517

source: G.F. Swift & Company, Branch House Dividends 1895, Swift & Company Records, Box 4, Chicago Historical Society; Mary Yeager, Competition and Regulation:The Development of Oligopoly in the Meat Packing Industry (Greenwich, Conn.: JAI Press, 1981), 63; U.S. Bureau of Corporations, Report of the Commissioner of Corporations on the Beef Industry (Washington, D.C.: U.S. Government Printing Office, 1905), 32. *Estimate (precise figure for Swift missing that year).

mour himself, in testimony before the Senate committee of 1890, conceded that his firm, along with others, had essentially learned the dressed beef business from Swift (U.S. Senate 1890, 432). In effect, Swift’s branch house innovation was undergoing a process of diffusion to competitors (Table 4.4).12 At the same time, this innovation was defining a new set of standards for entry and competition within the meatpacking industry. Nevertheless, despite this diffusion process, Swift retained certain advantages as the “first mover” of this process. Armour, Swift’s most powerful rival, did not open its first branch house until 1884. By that time, Swift was operating forty-three branches in a continuing effort to build a national distribution system (see Map 4.2) (Yeager 1981, 63). Perhaps the most decisive measure of change in the industry stemming from this process of innovation and diffusion is the relative shift that occurred in shipments of live cattle and dressed beef during the crucial transition years of 1878–87. During this period, dressed beef assumed a dominant role in the beef trade, completely overwhelming shipments of live cattle (see Table 4.5). With the exception of New York City, which maintained a thriving kosher industry, demand for beef in the East was being supplied not by live cattle but instead by beef slaughtered in Chicago. The impact of the new dressed beef product and Swift’s business model for distributing this product through branch houses was transforming how beef was being produced and sold. In an environment where the cattle and beef business was expanding and where beef consumption was on the rise, it was beef slaughtered in Chicago and shipped long-distance that was beginning to dominate the character of the trade.

112 G. F. Swift & Company table 4.5 Number of Cattle Shipped from Chicago, Live and as Dressed Beef, 1878–1887

1878 1879 1880 1881 1882 1883 1884 1885 1886 1887

Received

Shipped Live

Shipped Dressed

Shipped Dressed (%)

1,083,068 1,215,732 1,382,477 1,547,498 1,607,495 1,909,167 1,870,050 1,964,018 2,015,190 2,447,867

669,490 692,061 833,835 880,853 820,586 841,136 661,127 619,818 570,705 605,812

413,578 523,671 548,642 666,645 786,909 1,068,031 1,208,923 1,354,200 1,444,485 1,832,055

38 43 40 43 49 56 64 69 72 75

source: U.S Senate, U.S. Congressional Hearings, vol. 3, no. 829, Select Committee on the Transportation and Sale of Meat Products [known as the Vest Committee Report], 51st Cong., 1st. sess., 1890, p. 3.

Equally profound was the way these transformations reshaped the craft of butchering and the lives of individuals dependent on it. Gradually, during the course of the 1880s, as Swift’s business system spread and other packers established their own branch house networks, local slaughtering of cattle by wholesale butchers became an historical relic. In testimony before the Senate committee of 1890 on the meatpacking industry, Levi Samuels, a wholesale butcher from New York, provided a graphic description of this historical evolution (U.S. Senate 1890, 118–19). Q. [Senator Vest] What has been the effect of the dressed beef business upon the local butchers? A. [Samuels] In every section of the country, in nearly every small town, it has had the effect of completely stopping the slaughter of cattle by butchers. . . . Q. How do they [local butchers] operate now? A. Now the butchers, as a class, purchase Western dressed beef. Q. Suppose they do not choose to purchase the dressed beef, what is the result then? A. There is nothing else for them to purchase. Q. When you speak of Western people of course you do not mean all of us, but you refer to the Chicago people?

G. F. Swift & Company 113

A. I mean to say the dressed beef men. They sold their beef at a very much lower rate than the butcher could buy the cattle and produce the beef. . . . Q. In other words, they crushed out the business of the local butchers and then put their own dressed beef on the market? A. They crushed it out with their dressed beef. Q. Do you know of any specific instances where that has been done? A. Where it has not been done is the exception. In only a decade, the innovation of dressed beef had made local butchers such as Levi Samuels historical casualties. Many lost their livelihood as butchers. “Very few hold on [to the trade]” explained Mr. L. Leonard to the Senate committee (U.S. Senate 1890, 97). Although some former butchers were drawn in to the new slaughtering plants as part of a highly skilled “butcher aristocracy,” their presence was dwarfed by the number of semiskilled and unskilled workers in the cattle-killing gangs (Commons 1904, 3). From this process of stratification on the one hand and deskilling on the other emerged a new type of “butcher workman,” transformed into an industrial laborer (Brody 1964; Horowitz 1997). Where at one time a skilled butcher killed and dressed an entire animal, a newly created division of labor in the slaughtering plants turned a cow or steer into a detailed “map,” apportioning to each worker a tiny part of the total cutting operation (Commons 1904, 4; Brody 1964, 4). “Skill has become specialized to fit the anatomy,” observed economist John Commons in 1904 (p. 4). There were, for example, seventy-eight different occupations in the beef-killing gangs at the Chicago plants by 1904 (U.S. Bureau of Corporations 1905, 17–18). During this period in meatpacking, the number of wage earners in slaughtering plants increased from roughly eight thousand in 1870 to seventy thousand by the turn of the century (Horowitz 1997, 17). In effect, butchers assumed a new identity as part of a growing industrial working class. As Swift and the other large packers responsible for these changes became increasingly dominant, the meatpacking industry itself assumed a different role in the industrial structure of the U.S. economy. Little more than a decent-sized industry in 1870, meatpacking grew to a position of premier rank among other industries in the United States by 1900 (Table 4.6). Only the iron and steel industry was larger. Swift and the firms involved in this transformation operated national organizations. These enterprises redefined patterns of development in the economy of the late nineteenth century, creating a new geography of profit making. At the core of this geography was a very different fusion of spread and concentration.

114 G. F. Swift & Company table 4.6 Manufacturing Industries Ranked by Dollar Value of Output, 1870 and 1900 1870 Rank 1 2 3 4 5 6 7 8 9 10 11

Industry Flour milling Textiles Lumber Iron and steel Clothing Leather goods Machinery Sugar refining Tobacco products Furniture Meatpacking Pork Beef Miscellaneous

1900 Output ($ millions)

Rank

Industry

Output ($ millions)

445.0 380.9 252.3 199.5 116.5 157.2 138.5 119.6 71.8 69.1 62.1 56.4 1.9 3.8

1 2

Iron and steel Meatpacking Pork Beef Miscellaneous Machine shop products Textiles Clothing Lumber Flour milling Industrial machinery Boots and shoes Publishing Tobacco products

803.9 790.3 321.3 230.0 239.0 644.9

3 4 5 6 7 8 9 10 11

640.4 622.9 566.6 560.7 385.0 359.9 347.1 264.0

sources: U.S. Bureau of the Census, The Statistics of the Wealth and Industry of the United States, Ninth Census, vol. 3. (Washington, D.C.: Government Printing Office, 1872), Table 8c, p. 458; U.S. Bureau of the Census, Abstract of the Twelfth Census of the United States, 1900 (Washington, D.C.: Government Printing Office, 1904), Table 154, pp. 302–21; William Howard Shaw, Value of Commodity Output Since 1869 (New York: National Bureau of Economic Research, 1947), Table I-1, p. 52.

The Geography of Operational and Organizational Innovation With its network approaching seventy branch houses by the late 1880s, Swift was forced to seek additional slaughtering capacity in order to keep these marketing outlets supplied. After initially leasing slaughtering facilities and then purchasing its own plant at the Chicago stockyards in 1880, Swift opened its first expansion site for slaughtering at the Kansas City stockyards in 1888 (Unfer 1951, 57–59). Other disassembly sites soon followed. Swift opened plants in Omaha (1890), East Saint Louis (1892), Saint Joseph (1896), Saint Paul (1898), and Fort Worth (1902) as the number of its branch houses climbed above one hundred by the mid-1890s and close to two hundred by the turn of the century (U.S. Bureau of Corporations 1905, 32).13

G. F. Swift & Company 115 table 4.7 Capital Value and Sales of G. F. Swift & Company, 1885–1904 ($ millions)

Capital Value Sales

1885

1886

1890

1899

1902

1904

0.3 1.7

3.0 17.1

7.5 44.8

20.0 103.6

25.0 152.2

35.0 200.0

sources: U.S. Bureau of Corporations, Report of the Commissioner of Corporations on the Beef Industry (Washington, D.C.: U.S. Government Printing Office, 1905), 286; Federal Trade Commission, Report of the Federal Trade Commission on the Meat-Packing Industry, 6 parts (Washington, D.C.: Government Printing Office, 1919), pt. 4, pp. 25–27. note: Except for 1904, sales figures are estimates based on backward projections from the $200 million sales figure supplied by Swift for 1904 and relationship between capital value and sales.

A mutually reinforcing relationship between branch house proliferation and slaughtering capacity drove this growth. Nevertheless, in accounting for the catalyst in this expansion process, production and distribution were not equal. The mass-production economies in slaughtering, made famous in The Jungle by Upton Sinclair (1906), derived fundamentally from the innovation in distribution pioneered by Swift. This marketing revolution, which created a more direct route from producer to customer, made slaughtering a relatively small part of the beef business in comparison with distribution (Walker 1906, 493). “Essentially the packer is not engaged in production, but in distribution,” read a contemporary account of the meatpacking business. “The plant operations may therefore be looked on simply as processes preparatory to the distribution of meat and by-products in a more economical way than that afforded by the earlier method of sending the live animals to the consuming centers” (quoted in Aduddell and Cain 1973, 97). As Swift’s branch house and slaughtering network expanded during the 1880s and 1890s, the capital value of the company escalated dramatically. Incorporated in 1885 with a capital value of three hundred thousand dollars, Swift by 1904 had grown over one hundredfold, amassing a capital value of $35 million and generating revenues of $200 million (Table 4.7). At that time, it was one of the fifteen largest industrial corporations in the United States. Perhaps more significantly, in that year Swift emerged as the largest meatpacking concern in the United States, surpassing Armour in terms of revenues, slaughtering and distribution capacity, and market capitalization (Table 4.8). Indeed, by 1903, when the first great wave of merger mania in the United States had come to a close, G. F. Swift & Company was the largest meatpacking firm in the world.

116 G. F. Swift & Company table 4.8 Rank of Individual Firms Constituting the Beef Trust, 1903 Rank

Company

Cattle Slaughtered

Hogs Slaughtered

Market Capitalization ($ millions)

1 2 3 4 5 6

G. F. Swift Armour National Packing Morris/Fairbank Cudahy Packing Schwarzchild & Sulzberger (S & S)

1,578,215 1,255,366 848,884 761,179 469,228 559,200

4,079,756 3,451,892 3,101,425 1,247,393 1,347,675 623,598

35.0 27.5 15.0 6.0 7.5 4.4

source: U.S. Bureau of Corporations, Report of the Commissioner of Corporations on the Beef Industry (Washington, D.C.: U.S. Government Printing Office, 1905), xix–xx.

expansion and control This expansion in the Swift production and distribution organization, however, did not occur in arbitrary locations. The pattern of Swift’s expansion in slaughtering facilities paralleled the opening of new stockyard companies west of the Mississippi. In this process, Swift, as well as other large packers, was both follower and shaper of a new slaughtering geography. To some extent, these western stockyards where Swift migrated began to compete with Chicago as cattle markets, although Chicago maintained an overwhelming dominance as a cattle center well past the turn of the century. Nevertheless, Swift and the packers constituting the Dressed Beef Trust expanded into these growing cattle markets to gain control over access to sources of cattle supply for their growing national production and distribution networks. At their disposal in accomplishing this aim was their power as dominant firms. As large cattle purchasers, Swift and the Beef Trust came to control the practice of cattle buying and selling at the stockyards. One of the most critical tasks facing Swift and the other large packers was securing access to cattle supplies. By purchasing in volume, Swift and the other large packers essentially subverted the once independent role of the cattle commission business and transformed these merchant houses into unofficial purchasing agents for the Dressed Beef Trust. Given the opportunity to dispose of large consignments with Swift and the Beef Trust, these commission houses, which historically had facilitated the transfer of cattle from producers to buyers at the stockyards, now directed cattle purchases toward Swift

G. F. Swift & Company 117

and the largest packers. Cattle raisers suffered in this system from the low prices Swift was able to command. Perhaps more importantly, however, smaller independent cattle buyers and shippers were largely excluded from access to the best cattle supplies. In this regard, the remarks of Brewster Cameron, a cattle grazer and shipper, at the Senate committee of 1890 are particularly revealing. When asked by Senator Vest, the committee chair, how the cattle business was conducted in Chicago, Cameron replied that “the cattle are disposed of chiefly to what is known as the Dressed Beef Trust” (U.S. Senate 1890, 81). “The butchers of the country, the feeders, and all other buyers,” Cameron continued, “do not have a free opportunity to buy in competition with the Big Four.” Samuel Cady, a cattle commission merchant at the Chicago stockyards, echoed this theme when asked if there had been any change in the business of buying and selling cattle the last few years. “We do not have near as many buyers as we used to,” explained Cady (U.S. Senate 1890, 212). The reason? “I should express it something like the old expression that ‘the big fish are eating up the little ones.’” In addition to these informal mechanisms of control based on size, Swift and the other large firms benefited from more overt sources of power in the form of ownership interests in the stockyard companies themselves. Packers secured such ownership stakes in the stockyards in two related ways. The primary mechanism was for stockyard companies to offer stockholding as incentives to the packing firms to set up a plant at a particular stockyard site. Omaha, Saint Louis, Saint Paul, and Fort Worth extended such inducements to Swift. Similarly, Kansas City asked Swift to make an investment in the Kansas City Stockyard Company in exchange for subsidies such as free land for the Swift slaughtering facility. In both cases, ownership in the stockyard companies, through either subsidy or investment, provided Swift and the large firms with opportunities to gain greater control of the cattle markets at stockyard locations. While the ownership stakes themselves were relatively small (U.S. Bureau of Corporations 1905, 288), they nevertheless provided Swift with access to the governance authority at the stockyards that established systems of rule making for buying and selling livestock (Aduddell and Cain 1973, 106–7). In this way, Swift and the large packing firms transformed the stockyards from competitive cattle markets to highly controlled supply points for their national production and distribution systems. By expanding into these new stockyard locations, Swift influenced the development of a new geographical trend in slaughtering activity while changing the relationships by which cattle supplies were transferred from producers to the new slaughtering establishments. By the turn of the century, as a result of these expansion decisions by Swift and others, slaughtering had completed an ongoing migration to the

map 4.3 Expansion of G. F. Swift Production Facilities

Ft. Worth (1902)

Kansas City (1888)

St. Joseph (1896)

Omaha (1890) E. St. Louis (1892)

Chicago (1880)

Slaughtering facility (Year opened) Stockyard / cattle supplier

St. Paul (1898)

© Jeremy S h a w

G. F. Swift & Company 119 Chicago 248.8

73.2

Kansas City

67.7

Omaha St. Joseph

19.0

Indianapolis

18.3 12.3

St. Louis Buffalo

9.6

Cincinnati

9.5 0

50

100

150

200

250

Value of output ($ millions) figure 4.2 Slaughtering and Meatpacking Output for Eight Cities, 1900. source: U.S. Bureau of the Census, Manufactures, Twelfth Census, part 1, vol. 7 (Washington, D.C.: Government Printing Office, 1902),Table CXXXIV, p. ccvii.

nation’s Midwest in a pattern of regional concentration. In addition, slaughtering itself became highly localized. Chicago was the most dominant center, followed by Kansas City, Omaha, and Saint Joseph, cities where Swift expanded (see Figure 4.2). These concentrations of slaughtering represented the formation of what Alfred Marshall (1890) described as industrial districts (see Map 4.3). In these places, Swift and other firms profited not only from internal economies of scale based on high-volume slaughtering but also from what Marshall termed external economies deriving from place-based proximity to sources of supply, concentrations of labor and skill, and access to technical knowledge specific to the trade. Industrial districts of slaughtering and dispersed nodes of distribution thus coexisted in this new pattern of profit making. anatomy of the new geography This pattern of concentration for the slaughtering component of the meatpacking business replaced an older, far less concentrated geography for slaughtering. Although this older pattern included tendencies of both geographical concentration and decentralized spread, it was the latter that prevailed in the balance between the two. This older geographical pattern of weak concentration and strong decentralization was based upon two earlier attributes of the meatpacking business. One attribute, marked by the overwhelming primacy of pork in the meatpacking business, did, in fact, create places of concentration where pork was slaughtered (Walsh 1978, 1982).Yet, despite the concentration of slaugh-

120 G. F. Swift & Company tering in places such as Cincinnati—dubbed “Porkopolis” in 1850—pork slaughtering and packing also occurred in numerous river towns, especially in the Ohio River valley and central Illinois, as well as numerous towns in the East. In effect, while there were points of concentration in pork slaughtering, a highly decentralized structure of pork-packing towns, along with the slaughtering and curing of pork on numerous farms, acted as an offset to this tendency. The second characteristic that had strengthened the tendencies of geographical spread prior to the advent of dressed beef stemmed from the slaughtering of cattle in virtually all locales in the country. Because cattle slaughter and beef consumption had to occur in close proximity, the slaughtering of cattle was highly localized and geographically dispersed. This second attribute acted to decentralize slaughtering even more profoundly than pork packing and slaughtering, spreading it more uniformly throughout the country. Together these two elements created a geography of slaughtering marked more by decentralization than concentration. In effect, while there was concentration in slaughtering during the early period of meatpacking, the widespread slaughtering of animals throughout numerous towns and farms acted as a powerful countervailing force. Consequently, the pattern of concentration and spread for slaughtering that prevailed prior to the ascendancy of dressed beef was different from the geography of concentration and spread exhibited by the industry beginning in the 1880s, when dressed beef emerged as a dominant new product and business model. From 1860 to 1900, as slaughtering migrated into the nation’s midsection, the map of slaughtering activity changed dramatically (Table 4.9; Maps 4.4 and 4.5). This transformation is marked by the shift of slaughtering from its former base in Cincinnati and other river towns that had produced pork and from numerous local farms and local abattoirs spread throughout the country that had slaughtered cattle for beef consumption. This industrial structure had made New York, with the largest population, the largest slaughtering state. At the same time, the largest five slaughtering states accounted for 33 percent of total slaughtering activity. By 1900, slaughtering industrial districts in Chicago, Kansas City, Omaha, and Saint Joseph replaced this older, dispersed geography. One state, Illinois, and one city, Chicago, were overwhelmingly dominant, while the top five states in 1900 were doing close to three-quarters of all the slaughtering in the country. What had occurred during this period to change the geography of meatpacking activity so profoundly was the ascendancy of distribution as the source of territorial spread in the industry and the decisive element of the meatpacking business. Deriving from the rail and telegraph revolution and

G. F. Swift & Company 121 table 4.9 Largest Slaughtering States by Dollar Value of Animals Slaughtered, 1860 and 1900 1860 Rank

State

1 2 3 4 5

New York Illinois Ohio Pennsylvania Tennessee Top 5

1900

Value % of Total ($ millions) Slaughtering 15.8 15.0 14.7 13.4 12.4

Rank

7.4 7.0 6.9 6.3 5.8 33.4

1 2 3 4 5

State

Illinois Kansas Nebraska Indiana Missouri Top 5

Value % of Total ($ millions) Slaughtering 279.8 76.8 71.0 42.9 42.2

40.1 11.0 10.1 6.2 6.0 73.4

sources: U.S. Bureau of the Census, Agriculture of the United States in 1860, Eighth Census (Washington, D.C.: Government Printing Office, 1864), 186–87; U.S. Bureau of the Census, Manufactures, Twelfth Census, part 1, vol. 7 (Washington, D.C.: Government Printing Office, 1902), Table CXXXIII, p. ccvii.

the innovation of Swift, branch house distribution networks, and the logistics of sending dressed beef to branch houses located long distances from the places of slaughter are what elongated and spread the activity of meatpacking. As Swift expanded its network of branch houses during the 1880s and 1890s, and as it established additional slaughtering facilities outside Chicago to supply these branches, the company began to shape the foundations for a new geography of profit making in meatpacking (Map 4.6). Alongside the concentrated pattern of slaughter in industrial districts, the elongated activity of distribution through branch houses created the linkages for a new geography of concentration and spread. At the foundation of this concentrated and elongated geography, however, were innovations in business organization and operational routines deriving from a very different relationship—a more direct relationship—between Swift as a manufacturer and its customers. the direct system of operations The network of branch houses and slaughtering facilities built by Swift represented a direct system of producing and marketing (Rhoades 1929, 262). Fresh beef produced in high volume was sold directly by Swift to retail butchers without the involvement of wholesale intermediaries, who had formerly assumed an integral role in the process. This phenomenon of disintermediation gave the direct system of Swift its innovative character. In bridging the territorial separation of slaughtering facilities and consumption outlets, the system of direct marketing eliminated an historically entrenched

map 4.4 Top Five Slaughtering States, 1860.

Ohio 6.9%

Percentage of total slaughtering

Tennessee 5.8%

Illinois 7.0%

New York 7.4% Pennsylvania 6.3%

w

© Jere m y S h a

map 4.5 Top Five Slaughtering States, 1900.

Kansas 11.0%

Nebraska 10.1% Indiana 6.2%

total slaughtering

Percentage of

Missouri 6.0%

Illinois 40.1%

w

© Jere m y S h a

map 4.6 The Dressed Beef Enterprise of G. F. Swift, 1903

1 Branch house

2–4 Branch houses

5 Branch houses

Stockyard / cattle supplier

Slaughtering facility

© Jeremy S h a w

G. F. Swift & Company 125

group of wholesalers in the beef business (Porter and Livesay 1971, 168–73). Swift had essentially learned how to bypass these actors and capture that portion of the value that had formerly accrued to them in their role of facilitating the transfer of beef from cattle producers to consumers. In addition to disintermediation, Swift’s system of direct marketing had four other key operational attributes. First, Swift’s system was a pull system of procurement, production, and selling. In contrast to supply-driven systems, in which goods are produced and then pushed into various market channels for final sale, Swift’s system relied on actual demand in the form of orders from retail butchers to ignite the circuit of procurement, production, and sale. These orders for certain cuts, grades, and quantities from retail butchers served as the basis for pulling cattle supplies through procurement at stockyards, disassembly at factories, and sale as dressed beef at branch houses. In this sense, Swift’s pull system operated as an early type of just-in-time system. Second, Swift’s system operated on the basis of high-volume throughput. Dependent upon such volumes for balancing mass distribution and mass production, Swift’s system generated economies of scale by compressing time in the circuit of procurement, production, and final sale. Here the most formidable obstacle to time compression was in the slaughtering process. By the late 1880s, however, as a result of rationalizing and mechanizing the slaughtering process, ninety seconds is what it took to fell, slaughter, decapitate, and run to bleed a head of cattle (U.S. Department of Agriculture 1889, 365). With multiple killing lines running simultaneously, the most productive plants by the turn of the century were slaughtering as many as three thousand cattle per day (Commons 1904, 3). Third, because actual orders served to draw raw materials, semifinished sides, and finished cuts through the network, the pull system incorporated elements of an early if primitive form of customization. While it is important not to overstate this element, it is also critical to recognize how the different grades and cuts of beef ordered by retail butchers became part of a high-volume procurement and production system driven by the specific needs of retail customers. “In the marketing of beef there is the widest difference with regard to the character of the pieces sold,” noted the Bureau of Corporations (1905, 21). “A large proportion of the carcasses are sold as sides, but many retail dealers desire special parts . . . cut in a variety of ways.” Almost paradoxically, the highly standardized nature of the product—its capacity to be broken down into standard grades, cuts, and sizes—created parameters for both high volumes and customization. Finally, this direct-pull system relied upon an intricate process of supply and demand balancing in order to function. Demand and supply flows of product between the different branch houses, stockyard sites, and disassem-

126 G. F. Swift & Company bly factories constantly had to be modulated and adjusted as conditions changed. In fact, supply and demand balancing in close to real time provided the foundation for the direct-pull high-volume system oriented around customer orders. Much of what drove these operational process-oriented attributes was the time-sensitive nature of Swift’s product. Dressed beef was a perishable commodity. With a limited shelf life, dressed beef posed its most formidable set of production and distribution challenges with respect to time. Owing to the geographical separation of slaughter and sale, however, the issue of time was inseparable from the issue of distance. Compressing time and bridging distance were therefore the two most significant operational obstacles facing Swift. Confronted by this interrelated issue of time and distance, the direct-pull system of Swift was dependent upon the rapid transfer of information through telegraphic communications for balancing supply and demand and rapid transportation for moving product over distance through the network. Orders from customers had to be coordinated with the securing of cattle supplies, the slaughtering of cattle at disassembly sites, and shipments of dressed beef to branch houses for final sale. This process of supply and demand balancing in highly compressed time frames emerged as one of the most critical core competencies of the company. The telegraph provided the communications and information infrastructure for Swift in establishing this core capability. In describing the critical operational elements common to Swift and the largest packing firms, the Bureau of Corporations (1905) made specific reference to the challenge of supplying a customized and perishable product in high volume and balancing inventories of this product through telegraphic communication. “Differences in quality of animals and of their products are so great,” explained the Bureau, “that the closest supervision of the central office is necessary” in order to ensure the balancing of appropriate inventory in the network. “The central office is in constant telegraphic correspondence with the distributing houses, with a view to adjusting the supply of meat and the prices as nearly as possible to the demand” (p. 21). Operationally, the circuit of procurement, production, and sale in Swift’s direct-pull system began with the order-intake process and concluded with final sale at branch houses. These branch houses telegraphed orders for quantities on a daily basis to the company’s central office in Chicago. Depending upon the location of the branch, these orders were then conveyed by the central office to Swift’s seven slaughtering locations. Each branch house was normally assigned to a particular slaughtering plant upon which to draw product based on proximity to stockyard and slaughtering sites. Generally, the communication from branch houses was routed through the

G. F. Swift & Company 127

central office in Chicago (Chandler 1988, 53). At times, however, branches communicated their needs directly to the slaughtering plant to which they were assigned. In addition to refrigerated storage space for receiving dressed sides, each branch house included a sales office and a sales staff that procured orders from and sold the dressed beef to retail butchers, grocers, and other retail food outlets (Chandler 1977, 300). This sales staff contacted local butchers, conferred with them regarding their needs, and telegraphed orders to headquarters or assigned slaughtering plants either at the end or at the very beginning of the day. After receiving orders from its branch houses, Swift’s central office telegraphed instructions to its cattle buyers at stockyards as to the needs of the company in terms of breeds, grades, and quantities on any given day (U.S. Bureau of Corporations 1905, 21). Cattle generally arrived by train from transshipment points at night and would be held in stockyard pens. Procurement took place in the early morning with the aim of transferring cattle by rail spur from pens where they would be purchased to the killing floors of the plants no later than 11 a.m.14 From the slaughtering facilities, sides of dressed beef were shipped by rail to designated branch houses, where they were transferred to the chilled storage area. There, the dressed sides were matched with orders and further butchered by Swift employees in accordance with the orders received. Retailers would then call at the branch house and pick up their goods. As cattle moved from stockyards to the disassembly line in the various slaughtering facilities, the final branch house destination of the various sides and parts was already known (Chandler 1988, 237). Such a system was information-intensive and inconceivable without instantaneous communications provided by the telegraph. By the 1890s, Swift was spending two hundred thousand dollars per year on telegraphic communications to keep inventories in its network balanced in close to real time between cattle purchases at stockyards, production at slaughtering facilities, and orders from branch houses (U.S. Bureau of Corporations 1905, 207). Critical relationships of spatial proximity involving branch houses, stockyard sites, and slaughtering facilities were also integral to the direct-pull production and distribution network developed by Swift. First, branch houses required convenient railroad facilities for receiving finished goods. Consequently, branch houses were located where switching capability from railroad trunk lines could be expedited (Rhoades 1929, 270). Far more important, however, were the relationships between stockyard supply locations and slaughtering facilities. With its attributes of just-in-time procurement, driven by demand pull in real time, the dressed beef business avoided the buildup of cattle raw material inventories. Stockyards essentially assumed the responsibility—and costs—of holding inventories of supplies for the dressed

map 4.7 Centralization and Decentralization in the Dressed Beef Enterprise of G. F. Swift, 1903

5 Branch houses 2–4 Branch houses 1 Branch house

Information flow Stockyard / slaughtering facility

© Jeremy S h a w

G. F. Swift & Company 129

beef industry. A well-developed system, built over time, ensured that inventories of cattle supplies would be available in these locations to Swift and the other dressed beef firms for disassembly. In this way, inventories of cattle raw materials, which Swift essentially controlled without having to bear the carrying costs, and disassembly of these raw materials occurred in spatially adjacent locations. This direct-pull system, with its attributes of just-in-time procurement, product customization in high volume, and supply and demand balancing, redefined conditions for entry and competition in the business of meatpacking. Firms aiming to compete in the industry on a national level were compelled to imitate these elements. Implementation of these operational innovations, in turn, meant deploying the technologies of the railroad and the telegraph in fundamentally new ways. It also entailed the creation of a far different type of business organization. the vertically integrated organization The scope of the Swift operation in terms of branch houses and slaughtering facilities, coupled with the system of direct distribution to retail butchers, created unprecedented challenges of management and control. Confronting Swift was the complexity—and risk—of coordinating the long-distance flows of product and information between stockyards, slaughtering units, and branch houses and between branch houses and final customers. One of the most significant outcomes of this challenge involved a profound transformation in the organizational structure of the firm. The innovation of Swift in slaughtering and selling in disparate locations was as much an organizational phenomenon as it was an operational and processoriented set of changes. The solution pioneered by Swift to address this challenge of operational control and management of risk was vertical integration. Swift gradually absorbed the diverse and dispersed activities involved in the slaughter and sale of fresh beef within the boundaries of the company. The vertically integrated enterprise developed by Swift, though decentralized geographically, was highly centralized organizationally and operationally (Map 4.7). Major departments—procurement, disassembly, distribution, and accounting— functioned on the basis of instructions devised and communicated from headquarters in Chicago (Chandler 1988, 53). The reason for such central control stemmed from the logistical challenges of managing the various steps in producing and distributing a highly perishable product. In this process of organizational transformation, Swift not only assumed control over the operations most directly connected to cattle and beef processing but also incorporated a variety of ancillary but necessary steps into its own organization. This process of vertical integration represented an en-

130 G. F. Swift & Company table 4.10 Output of Meat Products and Manufactured Ice, 1869–1904 ($ millions)

Meat Ice

1869

1879

1889

1899

1904

Increase, 1869–1904

75.2 0.3

249.9 0.5

477.8 4.8

634.2 13.4

704.4 24.4

937% 974%

source: William Howard Shaw, Value of Commodity Output Since 1869 (New York: National Bureau of Economic Research, 1947), Tables II-1 and II-2. note: Meat products include fresh and cured beef and pork.

tirely new development in the history of American business enterprise (Chandler 1977). For the first time, a single company, in integrating production and distribution, began to systematically absorb a widening range of different activities within its organizational boundaries. Among the most notable of these ancillary operations was the Swift Refrigerator Transportation Company, which built refrigerated railcars for the firm’s dressed beef business (U.S. Bureau of Corporations 1905, 287). Instead of assuming the risk of having to purchase refrigerated freight services from the railroads, Swift instead produced and operated its own refrigerated railcar fleet (Raff and Temin 1991, 24–25). Similarly, rather than relying on the market to secure ice for cooling the refrigerator cars and branch houses, Swift became an ice manufacturer and distributor (Table 4.10). It purchased ice-harvesting rights on lakes all over Illinois and Wisconsin, sending iceharvesting gangs as far north as Green Bay during mild winters to ensure adequate supply (Unfer 1951, 36). It built immense ice storage houses near harvesting facilities and constructed icing stations along railroad routes to resupply railcars. Although ice was not a costly asset, the risk to Swift of being without it, as the largest single ice user in the country, was sufficient reason for the company to assume control over this seemingly minor detail in the production and distribution of dressed beef (Swift 1927, 191). Ice, in effect was an asset critically specific to Swift’s business. Much like other assets specific to the operation, it became a source of organizational integration within the boundaries of the firm (Raff and Temin 1991, 24–25). As the dressed beef business expanded and assumed a more complex character, the company became involved in business ventures that did not have a connection to the industry at the outset of Swift’s branch house innovation. An example of such a business was the Swift Fertilizer Works, established to make profits from the inedible by-products of cattle. This business in meat by-products compelled Swift to develop an industrial research

G. F. Swift & Company 131

business in the 1890s as the by-products portion of the industry, derived largely from cattle and classified as “miscellaneous meat products,” expanded to roughly the same size as the beef industry itself (Table 4.6). Animal-rendered margarine was one of the breakthrough products resulting from such activities in research and development.15 At the same time it was moving into new activities, Swift intensified its dominance over aspects of the distribution process that had initially escaped its control. Swift created “peddler car routes” outside its primary branch house distribution outlets in large cities; in the less urbanized areas of these peddler car routes, the company sold dressed beef butchered into retail cuts directly to the public, bypassing local retail butchers (U.S. Bureau of Corporations 1905, 30–32). Finally, the relationship of Swift and the stockyards, mediated through stockholdings in the stockyard companies and other subsidies (see above), reveals how Swift was able to extend control over the critical element of cattle procurement without assuming the costs of owning an asset outright. This “functional integration” with aspects of the stockyard operations, and the assumption of control over cattle buying that resulted from such integration, enabled Swift to minimize one of the greatest risks to its high-volume direct-pull system, that is, the ability to secure cattle raw materials. As a result of this relationship, sources of cattle supply at stockyards, while not formally integrated into Swift’s operation, became, in a practical sense, a part of the Swift enterprise.16 The company, in effect, sought control of the beef trade primarily through ownership of assets and creation of new businesses in a process of vertical integration. It also sought control over the beef business through mechanisms of market power and domination over terms of trade. Finally, Swift used investment and mechanisms of rule making and governance when such mechanisms were economically viable and politically desirable as in the case of the stockyards. This control, both formal and informal, extended the geographical reach of Swift into every region of the country. For Swift, there were several advantages to owning or controlling the adjacent steps in this high-volume production and distribution system (Aduddell and Cain 1973, 98–99). First was the fact that the creation and capture of surplus value in this network was dependent upon high volumes and internal scale economies. Such internal economies of scale, in turn, were generated from a division of labor difficult to find in any other industry of the period—a division of labor that was fundamental in creating the capacity to slaughter two to three thousand cattle per day in the most productive plants (Commons 1904, 3). Owning and controlling much of the channel, from sources of supply to slaughtering and distribution, possessed obvious advantages of certainty for organizing the steps needed to generate such levels of throughput. In this way, vertical integration and internal economies of scale

132 G. F. Swift & Company were mutually reinforcing. Second, the branch system of marketing eliminated the profit of the traditional wholesalers and consignees in the beef industry. Such value, historically captured by intermediaries, now flowed instead to Swift. Third, distribution and marketing was more efficient. As an integrated firm with its own centrally controlled procurement, production, and distribution network, Swift could match demand and supply of beef over a wider area and regulate product flows in accordance with shifting demand, supply, and price conditions. Derived from telegraph technology, this power of information reinforced its power of control through ownership of assets. Finally, as the turnover of product accelerated, owing to the highly organized integration of the branch, slaughtering, and procurement systems, and as demand and supply became more highly balanced, costs per unit of output were further reduced, thereby reinforcing the scale economies of high volumes. In the system pioneered by Swift, innovation was the foundation of a reinforcing set of relationships between economies of scale, the division of labor, and the extent of the market. As it turned out, however, the extent of the market was not a function of innovation and technology alone. The national markets created by the innovation of dressed beef also had a decidedly political dimension.

The Politics of the National Beef Market By the mid-1880s, vertical integration had become the standard for entry and competition in the meatpacking industry. In 1887, five of these firms— Swift, Armour, Hammond, Cudahy, and Morris—controlled 85 percent of the dressed beef trade, that is, the shipments of beef slaughtered at the major cattle markets of the Midwest and shipped long distance. Three years later, with the addition of Schwarzschild & Sulzberger, these firms controlled 90 percent of all dressed beef shipments. Although by 1903 these companies controlled only about one-half of the beef supply in the entire country (owing to the survival of farm-slaughtered beef and the role of local butchers in small towns and agrarian areas), their share in large eastern cities often exceeded 75 percent (U.S. Bureau of Corporations 1905, xxi). The system of centralized slaughter in Chicago and the national branch house distribution networks of the firms in this Beef Trust oligopoly established the operational and organizational standards for firms intent on competing for national markets. Nevertheless, even the smaller butchering firms that sold outside their immediate locality, but not on a national scale, and had managed to survive the competitive pressures of the dressed beef oligopoly began to adopt the organization and methods of the branch house system of production and marketing (Rhoades 1929, 265). Yet this transformation of

G. F. Swift & Company 133

the beef industry, in which the oligopoly of large-scale firms dominated, did not go unchallenged. First, resistance to the dressed beef trust emerged from the labor force in the large packing plants. In May 1886, thousands of workers in the Chicago slaughtering plants, despite being unorganized, answered the call issued by the Federation of Organized Trades and Labor Unions for a general strike to secure an eight-hour workday (Brody 1964, 13). Among packing firms, Gustavus Swift was recognized as a leader in breaking this strike of packinghouse workers (Swift 1927, 181). When the action of May failed and the packers, at the behest of Swift along with Armour, restored the ten-hour day in the packing plants, two more strikes involving nine thousand workers in Chicago occurred in October 1886. Again Swift was among the most aggressive in hiring strikebreakers to cross picket lines during these actions. A second round of strike activity occurred in the summer of 1894, after packing companies, in the aftermath of the 1893 depression, reduced workers’ wages. Unlike the earlier actions, however, this strike was accompanied by the use of troops from Fort Sheridan near Chicago, authorized by President Grover Cleveland, to keep trainloads of dressed beef moving from the packing plants to branch houses. Owing to this use of force, this action by packinghouse workers also suffered defeat. Other strikes followed throughout the remainder of the 1890s. Though sporadic and generally unsuccessful, these uprisings exposed the conflict existing in the stockyards between workers and owners of the packing firms that continued throughout this initial cycle of packinghouse unionization (Brody 1964, 13–33). While the trade union impulse represented a powerful source of resistance to Swift and the dressed beef firms, a second source of resistance emerged among smaller wholesale butchers. In 1886, wholesale butchers in several states formed a trade association, the National Butchers’ Protective Association. Its aim was to protect the livelihood of local butchers from the competitive encroachment of the large dressed beef firms. Taking advantage of still-existing sentiment against nonlocally slaughtered beef, this organization campaigned vigorously to defend the markets of local butchers by seeking passage of state laws requiring local inspection of cattle prior to slaughter. So successful were their efforts that they were able to persuade lawmakers in Minnesota, Indiana, Colorado, and Massachusetts to enact such state-level, preslaughter inspection laws. Under the provisions of these laws, animals would have to be inspected by state officials twenty-four hours prior to being butchered. Clearly protectionist in aim, the laws of these states were designed to block the interstate operations of Swift and other large packers and insulate local butchers from the competition of dressed beef slaughtered in Chicago and shipped long-distance. The large dressed beef firms chose to ignore the laws, instructing their

134 G. F. Swift & Company local agents to invite arrest and conviction (McCurdy 1978, 632). State authorities accordingly indicted the local sales agents of Swift and others. By 1890, the leading case, Minnesota v. Barber, was under review by the U.S. Supreme Court. In its judgment, the Court concurred with the argument of Armour, the defendant in this case speaking for the large dressed beef firms, that the preinspection law was established as a form of protection to the detriment of internal free trade. Furthermore, the Court held that such legislation represented a potential breach of the Commerce Clause of the Constitution. In the view of the Court, the Commerce Clause was intended by the constitutional framers to create a unified national market. Finally, the Court concluded that if government had a responsibility to protect consumers against the dangers of tainted meat, as counsel for Minnesota had contended, a remedy other than state-sponsored inspection would have to be created. The solution was federal inspection that preserved the idea of government regulation of the industry while at the same time maintaining the integrity of interstate trade and the idea of a national market (McCurdy 1978, 643–48). The outcome was the passage by Congress in 1891 of the Federal Meat Inspection Act. This legislation placed a new regulatory burden upon the industry. At the same time, however, this legislation legitimized the long-distance interstate traffic in dressed beef dominated by the large vertically integrated dressed beef firms. The new legal and regulatory environment enabled Swift and the large packers to secure the national markets they had created from innovation deriving from the rail and telegraph revolution. Yet, in legalizing the national markets of Swift and others and upholding the sanctity of interstate commerce, the Supreme Court in the Barber case and national lawmakers through the Federal Meat Inspection legislation gave new legitimacy to the importance of large-scale manufacturers (McCurdy 1978, 648–49). In the process, Swift and the dressed beef firms, as vertically integrated manufacturers, became the source of a new problem for the government. The issue was antitrust. Swift and the dressed beef firms faced ongoing confrontation from government in its role as a regulator of monopolies and trusts following passage of the Sherman Act in 1890. At least some of the political momentum for passage of the act itself came from the Senate Committee Hearings on the Meat Industry (U.S. Senate 1890). Its conclusions set the tone for much of the government’s antitrust activity involving Swift and the Beef Trust over the course of the next thirty years. According to the committee, “the principal cause of the depression in the prices paid the cattle raiser, and of the remarkable fact that the cost of beef to the consumer is not decreased in proportion, comes from the artificial and abnormal centralization of markets, and the absolute control by a few operations thereby made possible.”

G. F. Swift & Company 135

Other causes are also contributing factors, noted the committee, “but so long as the combine at Chicago can dominate the entire cattle interest of the country there will be the same result” (U.S. Senate 1890, 33). Throughout the 1890s and the first years of the new century, Swift and its erstwhile competitors entered into a variety of pooling arrangements and price-fixing schemes that became well publicized through journalistic exposés undertaken by the New York Herald and the Chicago Tribune (Gordon 1984). These accounts compelled a newly formed government agency, the Bureau of Corporations, to pursue an antitrust investigation of the beef industry in 1904. Relying upon data supplied by the Beef Trust itself, the report, issued the following year, concurred with the statements of the packers that no illegal monopolizing of the beef trade existed. This was only one in a long series of government investigations of the beef industry under the auspices of the Sherman Act. The most famous decision in the government’s lengthy battle against the beef industry—and one that emphasizes why these firms successfully evaded prosecution when the evidence against them appeared so overwhelming— involved the firm of G. F. Swift. The case in question was the 1905 decision in Swift & Co. v. United States (Gordon 1984). In this Supreme Court case, the Court affirmed that a party’s intent would subsequently be a material consideration in determining whether a combination constituted a restraint of interstate commerce (Gordon 1984, 245). In practice, this ruling had a profound impact on the enforcement of antitrust legislation. It placed a burden upon the complaining party to prove that firms, in this case Swift, intended to control terms of trade through illegal combination with other firms. Even under such rigid standards, however, Swift in this case received a judgment against it. Nevertheless, the injunction issued against Swift by Justice Oliver Wendell Holmes was so carefully qualified that the victory of the government was indeed “a pyrrhic one” (Yeager 1981, 184). The decision allowed Swift and the dressed beef industry sufficient room to conduct their operations as vertically integrated and oligopolistic firms. Furthermore, Swift and the other firms in the Dressed Beef Trust were permitted to combine for the purpose of limiting shipments and avoiding gluts in the various consuming centers of the nation (Yeager 1981, 184). Although forced to confront government regulators and the courts since the late 1880s, Swift had once again managed to prevail. For the first time, the Supreme Court in the 1905 Swift case recognized the implications of late nineteenth-century industrialization for adjudication of economic affairs (Gordon 1984, 279). Innovation pioneered by Swift had converged with politics and the courts in affirming the legitimacy of the large-scale business enterprise and the national territory where this new entity could make profit.

chapte r

5

The Internet as Commerce System and Market Space

Of all the applications emerging on the Internet infrastructure, Internet commerce will have the most profound impact. Peter Drucker, 1999

From Communications to Commerce When Dell began selling PCs over the Internet and reorganizing its procurement and assembly activity through Web-based interactions with its suppliers and logistics partners, the PC maker was able to craft these innovations on the basis of an historical transformation similar to the breakthrough that had enabled Swift to recast the system of production and marketing in the beef industry a century earlier. Like Swift, Dell has seized upon a revolution in communications as a platform for reorganizing its competitive activity. Just as Swift deployed the railroad and telegraph at the center of its extended and concentrated production and distribution organization, Dell has succeeded in transforming its organizational structure and territorial reach for profit making as a result of the revolution represented by the Internet. In both cases, new communications technology resulted in the development of commerce systems for producing, buying, and selling in fundamentally new ways.1 This chapter examines how this transformation of the Internet from communications system to commerce system occurred and how this process of commercialization created the foundations enabling firms such as Dell to compete differently. the contours of commercialization When Dell began experimenting with the Internet in 1996, the so-called network of networks had just passed through an initial period of takeoff as

140 The Internet an infrastructure for commerce supplementing its original role as a communications system (Reid 1997). This progression from an essentially noncommercial defense-oriented communications system to a commerce-oriented infrastructure is one of the most compelling features of Internet development (Kenney and Curry 2001, 45). Nevertheless, this pattern of commercialization is not unique to the Internet. The history of the Internet has certain features of creation and commercialization that are broadly similar to the pattern of earlier infrastructure systems.2 This pattern is represented in its initial creation phase by a technical breakthrough associated with the phenomenon of “invention.” Railroads, telegraphy, telephony, and the Internet all utilized breakthrough inventions for moving people, materials, and messages. Commercialization of breakthrough technologies, however, occurs primarily through the efforts of two types of firms. First, commercialization begins when such technologies become sources of profit for firms involved in construction and build-out of infrastructure systems. Second, commercialization advances when these technologies become sources of profit for firms that put the newly built infrastructure to uses in ways often unanticipated at the outset of creation and construction. With the exception of the telegraph, in which profit-making activity for telegraph firms and business users of telegraph systems followed only five years after initial demonstration of the new technology, a period of gestation separates the initial technological breakthrough from the commercial period. With the Internet, this period of gestation actually lasted almost twenty-five years, from initial demonstration in 1969 of the system’s core packet-switching technology, to a commercially oriented system marked by profit making in the build-out of the infrastructure and profit opportunities in Internet use by the early 1990s. Thus, while the Internet has its own unique history in evolving from a communications system to a commerce system, it shares three essential features of creation and commercialization with other infrastructure systems: (1) invention of the core infrastructure technology; (2) transformation of the technology into a built system providing profit to infrastructure builders; and (3) exploitation of the built system as a source of profit to a broader group of business users largely unconnected to the creation and construction of the infrastructure. These attributes, in turn, are marked by several key historical themes: infrastructure expansion and interoperability, creation of infrastructure standards, and the establishment of a legal and political rule-making environment for infrastructure deployment and use. It is in the role of infrastructure users that Swift and Dell emerge as key actors in this process.

The Internet 141

users as innovators Although not directly involved in creation and construction of infrastructure, business users are decisive in providing momentum for the buildout of infrastructure systems and in transforming invention into the broader process of innovation and economic change. As the economic viability of invention becomes clear, and as early investors along with technical experts promote initial infrastructure expansion, certain business users experiment in an effort to incorporate the new infrastructure into their business models and change the way they operate. It is this process of experimentation that results in commercial uses not anticipated by infrastructure builders at the outset. The rail and telegraph system was no more envisioned as the agent for the long-distance movement of commodities when it first emerged in the early part of the nineteenth century than was the Internet envisioned as a platform for buying and selling goods and services when it debuted as ARPANET in the 1960s. Such business use of transport and communications infrastructure originates at what has been termed the “consumption junction” of new infrastructure technology systems (Fischer 1992, 17). At this historical intersection, users of transport and communications systems come in contact with the initial variants of new transport and communications technology. As these actors begin to experiment and deploy the newly created infrastructure for profit-making ventures in ways unforeseen by the inventors and early builders of the new infrastructure systems, they become the functional source of innovation during communications revolutions. Alongside the initial creators of the actual technology, users provide the agency through which invention spreads and assumes the transforming attributes of innovation (Von Hippel 1988; Lundvall 1988). In this role as carriers of the innovation process, business users actually help to shape the development of new infrastructure, defining the goals and aims of newly deployed communications systems while influencing the trajectory of subsequent infrastructure deployment (Cohen, DeLong, and Zysman 2000). Business users in effect provide momentum for build-out of new communications systems as commercial enterprises. In concert with the inventors and developers of new infrastructure, business users of communications systems enable the impacts of communications revolutions to spread and act as catalysts for innovation in the economy. expandability and interoperability In ascending to its commercial status in the early 1990s, the Internet had to pass through two critical historical developments. First, the infrastructure had to expand. Second, the disparate pieces of the infrastructure had to in-

142 The Internet terconnect and become interoperable. Expansion and interoperability were thus interrelated.3 Similar to the development of the rail and telegraph, geographical expansion was integral in the evolution of the Internet from communications system to commerce system. Like the railroad and telegraph, the Internet after its initial creation developed as disparate communications networks which had to fit together and function as a system. Just as the disparate and often incompatible track gauges of the separate railroads and the different chemical and electrical technologies of various telegraph firms were made to interconnect in creating a unified transport and communications system, so too has the Internet assumed its role as a commercial infrastructure in stages marked by the efforts of Internet builders and users to enhance the interoperability of the new infrastructure. This process of expansion and interoperability, in turn, is dependent upon two additional elements: standards for infrastructure compatibility and rules for infrastructure deployment and use. Standards enable different elements of an infrastructure to function as a system while providing specifications around which new pieces can be added so the system can expand. Legal and political rules create the environment needed by businesses to participate in commercial ventures either in building or in using the new infrastructure. Standards and rules, in effect, provide momentum for new communications infrastructures to become more built out and interoperable.4 Several metrics are commonly used to provide indications of Internet expansion and interoperability. These indicators include the number of Internet users, host computers, Internet domain name registrations, and even the amount of venture capital expended on start-up firms. Modest growth in these measures began during the mid to late 1980s.5 The period of commercial takeoff in the early to mid-1990s, however, reveals several dramatic departures from previous trends, especially with respect to the number of users, the percentage of .com Web sites, and venture capital outlays. It was during this time, as expansion proceeded apace, that the Internet conformed to what has been termed Metcalfe’s law, named after the founder of 3Com Corporation, Robert Metcalfe. The basic idea of this principle is one of scale. According to Metcalfe, the value and functionality of a network will increase exponentially with each additional user. Consequently, at a certain threshold of nodes, a network’s value will exceed its costs so that the larger a network becomes in terms of the number of users, the more potential value it has to those users. Just as this principle drove expansion of earlier infrastructure systems such as the telegraph and telephone, so too is it also driving the expansion of the Internet (Moschella 1997, 106–7).

The Internet 143

the contingencies of internet commercialization Despite historical precedent and similarities between the commercial development of the Internet and infrastructure systems of the past, the passage of the Internet from communications system to commerce system was not preordained. Contingent events, occurring as the Internet evolved, were critical in pushing the Internet away from its military roots and creating the basis for an interoperable and commercially viable communications infrastructure. Four crucial developments proved decisive in this transformation. First was the decision by the defense establishment to split what was then the communications network of the Advanced Research Project Agency (ARPA) into two pieces. One piece was a military network aimed at continuing ARPA’s mission to create a communications system capable of functioning in a nuclear war. The second piece emerging from this split, however, was a nonmilitary communications system envisioned to serve a community of university and academic researchers. Placed under the administration of the National Science Foundation (NSF), this research-oriented communications network soon spawned numerous other civilian networks that together became the basis for the “network of networks” at the core of the Internet. Second, in the mid-1980s personal computers were transformed into devices that could interconnect and communicate via local area networks (LANs). Although this development removed some of the individual autonomy of personal computers, it created the foundations for a networking infrastructure based not on mainframes or even minicomputers, but on devices with potential for widespread use. After the mid-1980s, as PCs and PC-oriented workstations became interconnected mostly within firms, and as these local area networks proliferated, an expanding population of nonmilitary networks based on microcomputers created a new foundation for the Internet project. The networking of PCs and workstations was therefore a critical step in enabling the Internet to become widely accessible as a mass-based communications infrastructure. The third set of contingent events was the creation of two critical technological breakthroughs in the area of networking software—the World Wide Web and the Internet browser—that could be used in conjunction with personal computers. The most immediate impact of these two breakthroughs was to make the Internet far more accessible as a communications medium to both society and the business public. These tools, however, had an even greater impact in shaping a particular trajectory of growth and technological development. The creation of the Web and the Internet browser established a foundation—a technological paradigm—for an entirely new “ecology” of firms to emerge and develop myriad other Internet applica-

144 The Internet tions and products, from software, search engines, and portals to routers and broadband equipment that facilitated ease of Internet use (Hunt and Aldrich 1998; Kenney 2001). Fueled by venture capital, this swarming phenomenon of new technologies, created primarily by start-ups along with incumbents such as America Online, promoted a dramatic expansion in the number of new Internet users. In developing new tools for easy Internet access and use, these builder firms established an environment for a community of business users to emerge and experiment with the new Internet infrastructure as the communications element of a system for buying and selling goods and services. Dell Computer was among these early users. The final element contributing to the transition of the Internet to an infrastructure for commerce was political. By 1994, policy makers, influenced by a deregulatory free-trade environment, provided the first signals that rules for deployment and use of the emerging Internet infrastructure would be market-driven and would follow the lead of businesses. Although these rules are still evolving, both infrastructure providers and business users had a reasonably clear picture that the Internet was evolving along a market-driven, commerce-oriented pathway. Despite the fact that commerce appears as a phenomenon on the Internet only in 1995–96, the contours and contingencies producing this phenomenon began in the 1960s with the advent of the Internet itself.

Internet Origins and Architecture The Internet is a communications network of a unique type.6 It is most commonly depicted as a network of different communications networks. Nevertheless, it is perhaps more accurate to describe the Internet as a set of software overlays onto existing communications networks.7 On the basis of these software overlays, along with sets of common standards, the Internet is able to connect millions of different computers and communications devices. In this way it represents the convergence of computer and communications technologies. Although the backbone for Internet traffic incorporates existing telephone and cable television lines along with high-speed fiber optic lines connected by routers, the architecture of this network of networks represents an important departure from earlier communications systems. In earlier systems, communication occurs over dedicated connections from one point to another through a central switch. The Internet, by contrast, is based on decentralized and distributed nodes of computers all connected without a centralized point. This decentralized architecture derives from the Internet’s origins within

The Internet 145

the U.S. defense establishment. The Defense Department and its allied institutions, notably the RAND Corporation, were intent on responding to the development of Soviet military technologies in the aftermath of the Sputnik launch. What concerned the military above all was Soviet nuclear capability. Underlying the design of decentralized and distributed nodes was the idea that such an architecture, unlike centralized switching, would be able to withstand a nuclear strike.8 Military involvement affected not only the architecture of the Internet as a communications network but also the time lag that occurred between initial development of the packet-switching technology and eventual commercial use. Although the Internet was operating by 1969, it was only in the early 1990s that the network assumed its role as a system for mass communications and commerce. Prior to that time, it was a relatively obscure invention that, in its initial phase of expansion outside the defense establishment, was used by university researchers and not really commercialized in any significant way. Expansion and interoperability, however, would gradually occur in this military- and research-dominated environment. arpa and the internet As a first step in the Internet project, the Department of Defense created the Advanced Research Project Agency with a mandate to keep the United States ahead of the USSR in key defense and communications technologies. In 1962, ARPA formed the Information Processing Techniques Office (IPTO) to fund computer science research oriented toward promising military applications. By 1966 the director of IPTO, Bob Taylor, had begun to work on the problem of how to connect different mainframe computer systems so that these systems could share data and thus communicate with one another. ARPA allocated $1 million to Taylor to find a solution to this problem (Hafner and Lyon 1996, 12–13). A critical step in this effort emerged in 1967, when ARPA computer scientists developed the idea of inserting a subnetwork of smaller computers between the host mainframes, a concept later known as Interface Message Processors, or IMPs (Rosenzweig 1998, 1532). Once this idea of a network within a network was in place, ARPA eventually awarded a $1 million contract in 1967 to the computer consulting company of Bolt, Beranek and Newman (BBN) to develop the initial architecture for the ARPA network or ARPANET. Nevertheless, the contract for creation of the initial ARPA network had several other important antecedents that established some of the most important conceptual foundations for the nascent network (Rosenzweig 1998). The most important contributions came from Paul Baran, an engineer at the

146 The Internet Rand Corporation, and Donald Davies, a British physicist. In confronting the problem of how the U.S. government could maintain its communications capability following a nuclear exchange with the Soviet Union, Baran conceived of a communications network unlike telephony or even telegraphy that sent message traffic through dedicated physical lines. Critical to his concept was the idea of breaking communications messages into discrete “message blocks” that could be sent individually and reassembled at the reception point. He further proposed a system without centralized gateways, in which the switching nodes are decentralized, enabling messages to be rerouted in the event some nodes are incapacitated. Working independently of Baran but along a similar path, Davies focused his concerns on interactive computing and increasing the number of users who could access information from mainframe computers. In pursuing this aim, he developed a concept similar to Baran’s and gave “message blocks” the name “packets.” This idea of packet switching, in which data traffic is broken into discrete pieces through digital technology, sent through multiple paths instead of dedicated lines between two points, and finally reassembled at an end point, is the fundamental idea of Internet technology (Rosenzweig 1998, 1533). The first phase of the IPTO project aimed at creating a network of users at ARPA-funded university sites based on the concept of packet switching. The firm of BBN was responsible for developing the Interface Message Processors to enable the packet switching to operate between the four initial sites chosen for the project. These sites included the University of California at Los Angeles, the University of California at Santa Barbara, Stanford Research Institute, and the University of Utah. By 1970, these sites were expanded to include MIT, Harvard, and the University of Illinois. The following year, fifteen additional computer science centers funded by IPTO were connected to the ARPANET. This very modest pattern of growth in the number of host computers connected to ARPA continued throughout the decade so that by 1979 there were sixty-one such hosts (Tehan 2001; Kenney 2001, 16). By 1972, ARPA had demonstrated the feasibility of packet switching while creating a nationally extended though limited and little-used network (Rosenzweig 1998, 1536). An important dilemma, however, remained unresolved. As local area networks emerged alongside and independent of the ARPANET, computer scientists at ARPA realized that they confronted a problem similar to what they had to overcome in 1967 with incompatible computers. This time the problem was how to connect incompatible networks—not just incompatible computers—to one another. From this challenge of “inter-networking” came the project that would launch the Internet (Rosenzweig 1998, 1536).

The Internet 147

infrastructure standards Robert Kahn, a computer engineer who had moved from BBN to ARPA, spearheaded this “Internetting Project” to enable computers on different networks to communicate uniformly with one another and with the ARPA network. Kahn relied upon another computer scientist, Vinton Cerf, to help devise a standard language for packet-switching technology. Cerf and Kahn pointed out how previous language protocols for packet switching had addressed only the problem of communication on the same network. Their efforts introduced the idea of a protocol for communication on “different packet-switching networks” that became the basis for a new standard language called Transmission Control Protocol, or TCP (Zook 2001, appendix B). By 1977, ARPA began the first demonstration of the TCP idea in linking its three packet-switching networks. One year later, the new inter-networking protocol was divided into two parts and renamed TCP/IP. TCP consisted of the host-to-host linkages, whereas IP, standing for “Internet Protocol,” referred to the network-to-network links. A standard had developed for interconnection between computers operating on different networks. One of the most critical developments in the ascendancy of TCP/IP as the standard for Internet communication was support for the new protocol by the U.S. Defense Department. When ARPA came under more direct control of the Defense Department through the Defense Communications Agency (DCA) in the mid-1970s, the need for a standard communications language was paramount. As a consequence, the Department of Defense adopted the TCP/IP standard in 1980. The following year the Defense Department required all ARPANET hosts to implement the TCP/IP standard. This standard, however, did not go unchallenged. European telecommunications companies pushed an alternative standard to TCP/IP, the x.25. Nevertheless, support and funding by the Defense Department for TCP/IP enabled the American protocol to prevail over the European alternative in this battle of standards. In March 1981, the DCA provided an even more powerful incentive for adopting the TCP/IP standard when it initiated the requirement that all ARPANET hosts implement the protocol by January 1983 (Zook 2001, Appendix B). The creation and adoption of the TCP/IP standard and the requirement for its use by the DCA provided a powerful catalyst for growth of the ARPA network since, with the new language, different networks could connect to it.9 One source of growth was university computer science departments that sought access to the ARPA network. As the number of university-based users increased, however, the DCA decided in 1983 to separate military and

148 The Internet academic users by splitting the ARPANET into two networks. One network, MILNET, was for the military. The other was a residual ARPANET oriented toward civilian users in which the National Science Foundation began to play a major role.10 This civilian-oriented ARPANET attracted users from a second source of growth, the local area networks of business firms. As personal computers expanded in workplaces during the mid-1980s, and as networking of PCs became practical in 1985 with the introduction of the Intel 80386 microprocessor, the number of local area business networks expanded and emerged as a new source of demand for access to the ARPANET. Alongside LANs based on workstations, the growth of PCbased local area networks resulted in an increase in the number of networks connecting to the ARPANET. In 1982, the ARPANET consisted of fifteen networks. By 1986, the ARPANET included over four hundred networks (Zook 2001, Appendix B). This increase in the number of networks connected to the ARPANET was also reflected in a more dramatic increase in the number of network host computers. In 1984, host computers numbered slightly over one thousand but increased fivefold to a little over five thousand by the end of 1986 as new networks became connected to the ARPANET. The final years of the decade, however, witnessed the first growth spurt in the number of host connections as local area networks multiplied. By 1990, the Internet comprised roughly 313,000 host computers (Table 5.1). Nevertheless, while the number of Internet host computers was increasing and expanding the network, the number of actual users was still confined to a relatively limited community of academics and research scientists. Besides the connections of certain local area networks, the Internet had no real commercial characteristics. The transformation of the Internet into a commercial space, as well as a space of mass use, still awaited two crucial developments—the World Wide Web and the Mosaic, and later the Netscape Internet browser. These two elements would serve as the prelude for the Internet becoming a commerce system.

The Web, the Browser, and Web Commerce The breakthrough of the Web consisted of three critical Internet building blocks: hypertext markup language (HTML) to provide a format for Internet communication; hypertext transfer protocol (HTTP) to provide a guide for sending and receiving Internet communication; and the uniform resource locator (URL) to denote locations for finding information on the Internet. The creation in 1990 and release in 1992 of these elements was the work of computer scientists at the CERN European particle physics laboratory in Geneva under the direction of Tim Berners-Lee. Their efforts rep-

The Internet 149 table 5.1 Number of Host Computers Connected to the Internet, 1982–1990

Hosts

1982

1984

1986

1988

1990

235

1,024

5,089

56,000

313,000

source: Internet Software Consortium, http://www.isc.org/ds/host-count-history.html

resented a first step in making the Internet accessible for widespread use beyond technological elites (Weintraut 1997, xxic). The basic idea behind the Web and its relationship to the Internet is perhaps still best captured by Berners-Lee himself: The Web is an abstract, imaginary space of information. On the Net, you find computers. On the Web you find documents, sounds, video—information. On the Net, the connections are cables between computers. On the Web, connections are hypertext links. The Web exists because of programs which communicate between computers on the Net. The Web could not exist without the Net. The Web made the Net useful because people are really interested in information, not to mention knowledge and wisdom, and don’t really want to have to know about computers and cables. (quoted in Griffiths 2001)

Perhaps the most enduring contribution of the Web is that it provided the technical foundation for development of Web browsers, most notably Mosaic, and myriad other applications promoting both widespread use and eventually commerce on the Internet. The origins of the Mosaic browser derive from the NSF Supercomputing Center at the University of Illinois and two students working there, Marc Andreesen and Eric Bina, who began to explore ways of making the Web easier to navigate (Ceruzzi 1999, 303). By 1992, the two had written an early browser program they called Mosaic, which simplified use of the Web, and the two circulated this browser software freely over the Web itself. As a result, thousands of copies came into use and the number of individual Internet users climbed dramatically, from roughly one hundred thousand individual users worldwide in 1993 to approximately 3 million in 1994. In 1994, Jim Clark, the founder of Silicon Graphics Company, approached Andreesen, and the two created the start-up known as Netscape Communications that successfully commercialized subsequent versions of the Mosaic browser. In concert with the Web, the Netscape browser opened a new pathway of access to the Internet. Internet activity, which had expanded impressively after release of the Mosaic browser in 1993, exhibited an explosive increase in the number of individual users, expanding from 3 million in 1994 to 26 million the following year. These increases would con-

150 The Internet table 5.2 Estimated Number of Individual Internet Users Worldwide, 1993–2001 (millions)

Internet users

1993

1994

1995

1996

1997

1998

1999

2000

2001

.1

3

26

45

69

129

185

423

513

source: U.S. Department of Commerce, The Emerging Digital Economy (Washington, D.C.: Secretariat on Electronic Commerce, 1998), 7; U.S. Department of Commerce, Digital Economy 2000 (Washington, D.C.: Economics and Statistics Administration, Secretariat on Electronic Commerce, 2000), 5; www.zooknic.com note: The number of users in 1993 comes from U.S. Department of Commerce 2000, but Zook believes the number to be higher than 100,000.

tinue in the years following (Table 5.2). In addition to individual use, the Internet during these years was also becoming a magnet for institutional use, including businesses, educational establishments, and myriad different organizations. As a result, the number of Internet domain name registrations increased dramatically, paralleling the expansion pattern for individual users (Table 5.3). In the aftermath of the hugely successful introduction of the Netscape browser, which enabled the base of individual Internet users to expand, the company began work on a series of applications intended as a bridge from the browser to business. Netscape called this orientation “Integrated Applications” (I-Apps). The firm began to ship these commerce-enabling tools in 1995. The I-Apps, later renamed Commercial Applications, were among the first efforts at creating tools for Internet commerce (Reid 1997, 48). They were developed to enable businesses to sell goods or services through an Internet-based communications and transaction system. Two of the initial IApps, the I-Store and the Merchant System, were designed to create a catalog-type purchasing environment including an on-line infrastructure for clearing credit card transactions. The difference between these applications and catalogs, however, was that, instead of the mail or the telephone, the communications infrastructure for this type of buying and selling was the Web. This concept of commerce enabled by the Internet held enormous promise owing to the Internet’s strength as a “pulled” information channel (Reid 1997, 48). This potential was based upon the perception of parallels between pulling information supplied from a source on the Internet and buying goods sold by a vendor. If information could be pulled from the Web by Internet users, it seemed a short step to the idea first of substituting saleable goods for information, then replacing the concept of Internet

The Internet 151 table 5.3 Internet Domain Name Registrations, 1992–2001 (thousands)

Registrations

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

5

10

46

120

488

1,301

2,154

7,052 18,649 30,090

source: eMarketer, The e-Commerce B2B Report (New York: eMarketer, 2001); www.zooknic.com/domains/ counts.html note: Domains include .com, .net, .org, and .edu.

“user” with the concept of “buyer,” and finally changing the action of pulling information with a different type of action—a transaction. The challenge was how to shift this process of buying and selling in the physical world on to the Web infrastructure. The foundations for such a transition, however, were already beginning to emerge by 1994–95 in terms of users, both individual and firms, and applications. Internet traffic was continuing its dramatic rise with millions of new users, including businesses, emerging on the Internet, taking advantage of an explosion of new Web portals, search engines, and an array of Web-hosting services. Second, the proliferation of Web-based transaction applications being developed not only by Netscape but by numerous other software providers supplied tools for the Internet to become a business space for buying and selling (Kenney 2001, 23–24). Owing to these trends, the years 1993–95 demarcate the early commercial phase of Internet development and separate it from the precommercial, so-called B.C. (before commercialization) stage of the Internet (Weintraut 1997, xxxv). During this period, tools in the form of commerce-enabling Web applications stemming from the breakthroughs of Netscape created a more viable infrastructure for business activity and contributed to the overall build-out of the Internet infrastructure. These tools and the build-out of the Internet infrastructure enabled the burgeoning community of Internet users to orient to the Web in a commercial way. It is important to emphasize, however, that commercial use of the Internet and mass use did not evolve in opposition to one another. Both mass use and commercialization evolved more or less in tandem. In this environment, business firms, both start-ups and incumbents, began to create Web sites to take advantage of this business orientation of the Internet during this critical period of early commercialization. As a result, the number and percentage of .com Web sites, which was very small in 1993, started to rise dramatically in the next couple of years. By the beginning of 1996, Web sites associated with business firms represented roughly half of the sites on the Web. This percentage would in-

152 The Internet table 5.4 Growth in Percentage of .com Web Sites, 1993–1996

% of .com Web sites

June 1993

Dec. 1993

June 1994

Dec. 1994

June 1995

Jan. 1996

1.5%

4.6%

13.5%

18.3%

31.3%

50.0%

source: Mathew Gray, MIT, http://www.mit.edu/people/mkgray/growth/

crease even more dramatically by the time the decade came to a close (see Table 5.4). This period of rapid expansion in 1993–96 marked the beginning of the critical build-out stage for the Internet as an infrastructure for business (Weintraut 1997, xxxv). During this period of rapid build-out, the number of domain name registrations (a different measure than Web sites) in the .com, .net, and .org domains increased from roughly ten thousand in 1993, to two hundred thousand in 1995, to 1.5 million in 1997 (eMarketer 2001). During this same time period, the number of Internet host computers skyrocketed from 1.3 million in 1993 to 16.1 million by 1997.

Internet Commerce and Government Policy While the various Web-hosting and commerce-enabling applications for the Web began to appear in 1994–96, allowing the Internet to become viable as an infrastructure for commerce, and while early experiments in Internet buying and selling began to emerge by 1995, commerce on the Internet raised numerous legal and political issues for policy makers. If the phenomenon of Internet commerce was to become accepted and expand, it would require a rule-making framework to set standards for the conduct of buying and selling on this infrastructure. Rules on such issues as intellectual property, privacy, content distribution, and taxation of Internet transactions are still evolving as part of an ongoing process to create a uniform commercial code for Internet activity throughout the world.11 Two early sources, however, signaled businesses on the commercial direction of policy making for the Internet. One source was the Telecommunications Act of 1996. The other was a position paper released by the Clinton Administration in 1997 entitled Framework for Global Electronic Commerce.12 Although the Telecommunications Act of 1996 aimed at governing construction and deployment of telecommunications networks and providing new rules for the conduct of firms involved in infrastructure build-out, its impacts on the Internet were far-reaching. The Act in many ways affirmed that the world of data networking that had been emerging as early as the 1970s was different from telephone service and should be regulated

The Internet 153

differently (Bar et al. 1999, 2). The primary objectives of the 1996 act were twofold. First, the Telecommunications Act aimed to strengthen the already evolving convergence of communications and computer networking technologies that was providing the catalyst for Internet development and legitimize this convergence as public policy. Second, it sought to open competition for construction of the Internet infrastructure and to create open access to the network in contrast to the closed networks created by monopoly telephone companies. The pathway to these aims in terms of general policy was deregulation. The Act established a deregulatory environment for the telecommunications market, allowing firms from myriad different sectors, many of which were formerly precluded from involvement in telecommunications activities, to compete in expanding the build-out and interoperability of the Internet infrastructure. While the law focused on the infrastructure, its impacts were just as critical to the emerging community of business users. The legislation provided users with signals that the infrastructure would not operate haphazardly and arbitrarily but would expand in accordance with certain rules. If users were to conduct business over the Internet, they had to know that a framework was in place to promote the technologies enabling this infrastructure to extend its reach and become more interoperable. In a sense, the policy enabled users themselves to play a major role in driving the Internet revolution (Bar et al. 1999). Similarly, the Framework document gave businesses signals that Internet commerce not only would be encouraged but would actually receive special dispensations such as tax relief. The document outlined five major principles in which the government affirmed its interest in providing a predictable legal environment for Internet commerce. Nevertheless, the authors of the Framework were careful to specify that government involvement in establishing certain rules for these activities would not compromise the interests of business firms in a market environment relatively free of government oversight. Such involvement was to be one where government would not lead but would follow the practices and precedents—including the standards— established by businesses. In many ways, the document followed in spirit the deregulatory emphasis of the Telecommunications Act. Government was going to implement a more market-driven and less-regulated environment for the new activity. These principles were already the de facto policy of the administration before the document was released.

The Internet Retail Space As the foundations for an increasingly interoperable and commerce-enabled Internet infrastructure emerged in 1994–95, fueled by an explosion of ven-

154 The Internet ture capital investments in start-ups creating ever more applications and networking gear for the Internet, and as the signals for a favorable policy environment began to take shape more clearly, business users began a process of experimentation in buying and selling on the Internet. Most of these early firms engaging in selling products and services on the Internet were themselves venture-funded start-ups that began to sell a range of different consumer-oriented items to retail customers. In the process, these companies were pioneering new models of consumer-oriented Internet retailing. By far the most compelling of these early experiments in Internet retailing emerged in the form of a Seattle-based bookseller, Amazon.com. amazon and internet market space Amazon.com created the first large-scale business using the Internet as a communications infrastructure for retail transactions, initially selling books and later a wide array of different products to consumers. The firm established a Web site for Internet sales in July 1995 and became a billion-dollar seller of books over the Internet in a little over three years. During this period, Amazon succeeded in creating the most recognizable brand name in Internet commerce, one that rivaled the brand names of the world’s largest firms (Dodge 2001, 167). Its growth paralleled the expansion of the Internet itself. The sources for Amazon’s business model were twofold. The first was the enormous increase in the number of Internet users. The second was the availability of commerce-enabled Web tools developed by various software firms in the wake of Netscape’s successful Web products. Company founder Jeff Bezos provides a revealing admission of how he developed the idea for selling books over the Internet from these two sources: Two years ago [1994] nobody knew how many people were on the World Wide Web, but what was noticed was that Web usage was growing at 2,300 percent a year. I started thinking about what kinds of opportunities are there going to be in this new Web space. It was clear that the Web technology, even though it was rudimentary, would support primitive interactive retailing. I thought that the content areas were going to be very crowded, and it wasn’t clear what the revenue model was to make money off of content, so transactions-based business made sense to me (Southwick 1996).

From this rapidly expanding population of Internet users, coupled with the new commerce-oriented Web tools, Bezos successfully reconceived the notion of what is perhaps the most fundamental element in economic activity, the notion of a market. In grafting the Internet on to the notion of a market, the pioneering bookseller had created the idea of a selling space with boundaries of an indeterminate nature. The challenge confronting Amazon

The Internet 155

was how to use the Internet to sell in this market space and how to generate a profit from this new method of sales. The economic rationale for Bezos was that in reaching this market through the Internet and eliminating the costs of stores and employees the process of selling could be made more efficient and hugely profitable. By taking customers’ orders directly over the Internet and by eliminating the overhead costs associated with physical retailing, Amazon aimed to undercut prices of retail bookstores by marking up book prices from publishing houses at a fraction of the normal markup rate for retail stores. In addition, Amazon counted on creating a more convenient and efficient sales channel to its customers through Internet communications. It represented an experiment in seeking profit from communicating across a new type of space. internet efficiency, internet geography The business model devised by Amazon took advantage of the Internet in two fundamental ways. First, Amazon used the Internet to create certain types of market efficiencies. Second, the technology of the Internet enabled Amazon to reshape market geographies. It was the latter, however, that actually enabled the former. Changes in geography served as a new route to efficiency. The business model of Amazon provided one of the earliest and most powerful lessons in the way commerce on the Internet began to transform the concept of markets as territorial entities. This change in the geography of markets, in turn, had profound consequences in terms of redefining pathways to efficiency and profit making. The book business, and the way this business was transformed by Amazon, provides a compelling illustration of how this transformation in market geography changes the profit-making environment for firms. A bookstore represents a business with a market shaped most fundamentally by the concept of place. For the most part, customers of bookstores reside in the same geographical locale as the bookstore itself. While large bookstores might service long-distance customers and possess a more geographically dispersed customer base, in general the bookstore and its market share the same locality. Furthermore, in addition to bookstores, numerous types of businesses operate in this manner, their market and customer base conditioned by territorial boundaries. What the Internet as a commerce system did for Amazon, along with other Internet sellers, was to open the territorial boundaries that circumscribed and even protected the markets of established businesses. At the same time, the Internet subjected incumbent stores to competition from Internet firms located outside the market area. Early Internet auto retailing, flower selling, even the Internet grocery business provided examples of how the

156 The Internet territorial shape of markets for certain types of economic activity can change. In this sense, Internet retailing emerged little different than the innovation of mail-order catalogs pioneered by Montgomery Ward and Sears Roebuck in the late nineteenth century. These mail-order houses invaded the protected markets of retailers especially in rural areas and small towns, undercutting local merchants with low prices achieved through enormous volumes and economies of scale.13 Admittedly, the Internet did not dissolve distance as a factor in producing and trading, as many of its earliest disciples claimed it would. Internet merchants, Amazon among them, could not neglect the logistical challenge of warehousing inventory and delivering products to geographically dispersed customers. These Internet sellers turned to third-party logistics providers to manage warehouses and to the overnight-package-delivery industry, the “Airline of the Internet,” to provide the fulfillment for delivering goods ordered through the new communications infrastructure (Lappin 1996).14 Nevertheless, despite Internet technology’s inability to completely dissolve distance as a factor in business activity, this new infrastructure, much like earlier communications technologies, did provide the tools for firms to reshape the organization of territory for economic activity. By enabling firms to reach customers in geographically distant locales, the Internet redefined the meaning of economic territory and market space. Amazon used this space-transforming attribute of the Internet to accomplish a decisive efficiency objective. The firm took advantage of Internet communication to locate and reach an enlarged population of buyers for its product. In this way the firm relied on Internet technology to overcome one of the primary obstacles to efficient markets, the problem of incomplete information—the so-called problem of information asymmetry—in this case, incomplete information on the location of customers. Furthermore, the technology provided the firm with a channel of communication to reach this newly discovered customer base. In addition, Amazon used the Internet as a transaction-processing technology, enabling the firm to create greater levels of efficiency in the market by reducing its costs of transacting. Consequently, by using Internet communications to pioneer a more effective mechanism for matching buyers to Amazon as seller, and by using the Internet to process transactions from this newly created base of buyers in a much more automated and less expensive way, Amazon uncovered a new route to efficiencies in economic space. In this way, geography transformed though Internet communication provided a platform for efficiency. In addition to reconfiguring the market geography of buying and selling and redefining routes to efficiency, the most enduring impact of Internet sellers such as Amazon, despite its inability to turn a profit, was to legitimize the viability of business models for transacting and selling over the Internet.

The Internet 157

Although these early Internet-based selling models emerged primarily in the retail channel, where products moved from firms to consumers, forms of Internet retailing pioneered by firms such as Amazon also influenced the selling channels of nonretail businesses. Among these firms, Cisco Systems was arguably the most pioneering. It began selling its networking equipment to its business customers over the Web beginning in the second half of 1996. By the end of the year, Cisco had booked $100 million in Internet sales, which expanded to $1 billion in 1997 (U.S. Department of Commerce 1998, A3–11). Dell was also involved in these early efforts of manufacturers to develop Internet sales channels. By the late 1990s, numerous other firms whose customers were actually other businesses began to follow the example of Cisco—and Dell—in developing the capacity to sell over the Internet. This sales activity marked the beginning of a far more significant trend on the Internet, interfirm sales, where businesses buy and sell among themselves in what is commonly referred to as the business-to-business space. Within this space most of the value in the economy is created. It is also within this space that firms, especially manufacturers, confronted an enormous challenge as well as potential opportunity. This challenge focused on using the Internet to reorganize entire value chains for products, from procurement of supplies through manufacturing and assembly and finally distribution. Within a year after the initial explosion of Internet retailing firms in 1996–97, venture capitalists began financing companies with business models aimed at enhancing the efficiency of interfirm activity through Internet communications (Figure 5.1) (Kenney 2001, 36). As early as 1998, despite continued media attention on Internet retailing and firms such as Amazon, nearly 80 percent of Internet commerce was occurring between companies (Business Week, June 22, 1998, 130; Tedeschi 1999). In the process, the business-to-business space had emerged as the next wave of commerce on the Internet.

The Internet Business-to-Business Space While the business-to-business space on the Internet is distinct from Internet retailing, there are ways in which the precedents of Internet selling became models for reorganizing interfirm activity through the Web. The reason for this bridge between Internet retailing and the reorganization of interfirm value chains through Internet communications is that efficiency in the value chains for manufacturing and efficiency in retailing, despite obvious differences, share certain underlying similarities. Taken in its entirety, the value chain in manufacturing products involves an enormous amount of transacting in order to connect the creation of products with their sale and

158 The Internet Number of IPOs

Amount Raised



--a--

300

30

250

25

200

20

150

15

100

10

50

5

0

0 1980

1985

1990

1995

2000

figure 5.1 Number of Venture-Backed Initial Public Offerings and Amount Raised, 1978–2000 source: Mathew Zook, “The Geography of the Internet Industry: Venture Capital, Internet Firms, and Regional Development” (Ph.D. diss., Department of City and Regional Planning, University of California, Berkeley, 2001). Reprinted by permission.

consumption. Not only are there numerous transactions between different firms involved in this circuit from production to consumption. There are transactions within the boundaries of the same firm as part of the creation and distribution process. Transacting—trade—is therefore a fundamental element even in production. This shared characteristic between manufacturing and retailing—the process of transacting—is what enabled Internet retailing to provide certain precedents for the idea of enhancing the manufacturing process through Internet communication. If retailing could be made more efficient through Internet communication as firms such as Amazon were seemingly demonstrating, it seemed logical that opportunities existed to render interfirm trade, that is, entire business-to-business supply chains and production networks,

The Internet 159 table 5.5 Estimated Levels of Internet Commerce, 1996–2001 ($ billions)

Retail Business-to-Business

1996

1997

1998

1999

2000

2001

0.24 0.51

2.5 8.0

5.1 17.1

18.0 52.9

33.0 226.2

65.0 448.9

source: eMarketer, The e-Commerce B2B Report (New York: eMarketer, 2001); Forrester Research; NUA Internet Surveys.

more efficient in much the same way (Poirier and Bauer 2000). Such Internet-driven forms of enterprise between firms represented a new frontier in supply chain management and Internet commerce (Table 5.5) (Anderson and Lee 1998; Adhikari 1998). Indeed, certain manufacturers, notably Cisco Systems and General Electric, as well as Dell Computer, began to experiment in 1996–97 not only with Internet sales channels but also with ideas for reconfiguring the supply chain component of their business (U.S. Department of Commerce 1998). Such efforts by firms at reorganizing the procurement process through Internet communication, however, posed much more complex operational and organizational challenges than those faced by retailers using the Internet to recast retail sales channels. This complexity stems from the dominant role of interfirm activity in the economy itself. Roughly three-quarters of the transactional activity in the economy occurs along networks linking firms.15 Fulfillment of final demand to consumers represents but a small fraction of the transactions taking place between myriad firms that go into the creation of products and services. Reorganizing these linkages between companies through Internet communications effectively transforms the entire system of interfirm relationships in the economy. This dominance of interfirm trade in the economy and the potential efficiency gains from recasting this trade through the Internet are what attracted the efforts of venture capitalists, start-up firms, and existing companies to develop Internet-based business models for supply chain activity. In the process, the business-to-business space became the most coveted area of Internet commerce. In order to understand these efforts, however, it is critical to examine the foundations of interfirm value chains themselves. interfirm value chains The basic activities of firms in procuring, creating, and selling products and services take place in networks known as value chains (Porter, 1985).16 Each individual firm has its own value chain that defines the boundaries be-

160 The Internet tween it and other firms and reveals how the economic activity in the network—procurement, production, and selling—is organized. The manner in which firms organize and perform these various functions within their value chains is the source of the firm’s profitability and competitive advantage. What occurs along these networks is a process of value creation whereby goods and services move between different firms, or between different units of the same firm, and in this process of movement become transformed. Through the activities of initial concept, design, procurement, fabrication, marketing, distribution, and final sale, firms at these adjacent steps modify the attributes of goods and services, imbuing them with more value. Built in to each of these activities, however, is not only human labor. The various steps in the value-creation process also include an element basic to the labor process itself, the element of knowledge transfer, or information sharing. As products are transformed at adjacent steps in value chains, they are imbued with value measured by increments of time and information. Because the economic activity in value chains is essentially an information-and-knowledge-processing activity, communication is fundamental to the activity in value chains (Bar 2001, 36).17 The interfirm (and intrafirm) networks along which value chain activity takes place are, in effect, information and communications networks. Inputs of time and information added by firms to products or services to increase their value as they circulate within value chains are of two types. One type occurs on the production side of value chain activity. This input changes the properties of goods and services directly and involves the addition to the product or service of new materials or new knowledge and information. The other type of value-creating input occurs on the circulation side of the value chain. Often referred to as “the gains of trade,” this type of input adds value indirectly to a product or service by facilitating how products or services move through the various steps in the value chain creating value from what geographers term a change in location. Value on the circulation side is generated from information used to rationalize how products or services circulate from conception, to creation, to consumption. Despite these distinctions, inputs, whether on the production side of the value chain or on the circulation side, share a common and fundamental characteristic. Both types of inputs add value to goods and services by imbuing them with new knowledge and information through a process of communication. On the production side, communication transmits information that becomes embedded within products. On the circulation side, communications transmits information embedded within processes. As a result, communications-enhancing technologies provide a primary route for efficiency gains in the process of value creation. The Internet is currently providing such a pathway.

The Internet 161

internet value chains In reconfiguring distances between economic actors and reshaping market geographies, the Internet is recasting efficiency in value chains and enabling firms to reorganize business models as a result of three primary attributes (Malone,Yates, and Benjamin 1994, 67–68). These attributes of the Internet are (1) the digital communication effect, which refers to the larger volume of information that can be exchanged between businesses in less time at less cost; (2) the digital brokerage effect, referring to the number of alternative transactions that can be considered by firms in less time at less cost; and (3) the digital integration effect, referring to changes in the organization of businesses processes occurring at the interface of adjacent value-added stages of the value chain.18 All of these attributes enable firms to decrease the unit costs of coordination within value chains, including the costs of information processing and communication in tasks such as selecting and interacting with suppliers, settling contracts, scheduling and budgeting activities, and tracking flows of materials as they circulate from suppliers through production to final marketing. On the basis of these attributes, two different business models deriving from two types of firms emerged and began to recast interfirm value chains through Internet communication. The first model derived from the efforts of existing firms aiming to rationalize their procurement systems through Internet communication. In this business model, existing companies used the Internet to locate and negotiate with a far larger number of suppliers than was possible before Internet communication. The other business model that emerged in the Internet business-to-business space resulted from the growth of an entirely new business phenomenon, the so-called Internet market maker. Taking advantage of the Internet’s communications and brokerage effects, these start-ups created Web-based, on-line exchanges for aggregating and consolidating buyers and sellers in supply chains for entire industries or product groups (Business Week, March 13, 2000; Tedeschi 2000). In what was perhaps a paradox, these market makers involved in creating online exchanges emerged as new intermediaries aiming to reorganize interfirm buying and selling through the Internet and secure revenues from transactions. Among the companies in the first category, General Electric was arguably one of the earliest and most pioneering in developing a Web-based business model for procurement. As early as 1996, GE had already developed a pilot Internet-based procurement system for its lighting division using an extranet developed by GE Information Services. One year later, eight of GE’s operating divisions had implemented this on-line procurement system for some of their part and supply purchases (U.S. Department of Commerce 1998).

162 The Internet By the end of 1999, the company had managed to convert all twelve of its operating units to on-line purchasing. GE referred to its on-line procurement system as TPN, or “Trading Process Network.” The goal of GE’s network was to create efficiencies in procurement by relying on the capacity of Internet communication to aggregate large numbers of suppliers through the digital communication effect and create an auction environment with this extended group as a result of the digital brokerage effect. This process of market enlargement for supply sources, however, derived not only from the Internet’s communications effect in reaching large numbers of suppliers simultaneously but also from the Internet’s capacity to store, duplicate, and send information. These characteristics enabled GE to change the way it organized the work for sending out bid specifications, thereby taking advantage of the Internet’s digital integration effects. Bid specifications, including blueprints that formerly had to be copied and processed individually and then distributed to eligible bidders, were able to be stored digitally and sent as Web-based documents. Because the former process was complex and time-consuming, GE normally sent out bid packages to only two or three suppliers at one time. Through the Web, this process of sending bid specifications to suppliers was simplified to the extent that numerous suppliers were able to bid on GE projects. Suppliers registered to be on the network, enabling GE to reach this wider base of firms, in turn, providing GE with the opportunity to create on-line reverse auctions to drive down prices. The firm, in effect, was able to gain efficiencies in procurement by extending the market for suppliers and using this extended market to obtain the lowest possible component prices. The Trading Process Network helped GE to reduce the time required to identify suppliers, prepare bid requests, negotiate prices for parts, and award a procurement contract by 50 percent (U.S. Department of Commerce 1998, A3–27). By enlarging its market for suppliers and by enabling the firm to change certain operations in the procurement process, the Internet, through its communications, brokerage, and integration effects, has provided firms such as GE with cost savings in its supply chain. In many ways, the aims of the independent exchanges established in the Internet business-to-business space by start-up firms are similar to the trading network created by GE. These exchanges seek to create cost savings and efficiencies for firms by aggregating and consolidating buyers and sellers through the Internet’s communications and brokerage effects. The difference, however, between these exchanges and the Web-based trading network of GE is one of control. Whereas GE built an Internet business model in which it controls its own Internet procurement space, in the on-line exchanges, the business model and structure of control is different. With names such as e-steel.com, Chemdex.com, and PlasticsNet.com, these new firms

The Internet 163

created a business model in the early days of business-to-business Internet commerce based upon the idea of consolidating and aggregating buyers and sellers within industries while seeking to exercise some control over the transactions between these actors in order to generate revenue (Economist, June 26, 1999, 23). Thus, while the element of efficiency through aggregation and consolidation runs through both models, the on-line exchange is based on the idea of an intermediary exercising control over the activity and generating revenue from the transactions it orchestrates. In some ways, these intermediaries represented a type of gold rush in the business-to-business space of Internet commerce. Funded heavily by venture capital, these start-ups exhibited an explosive growth in number, especially after 1998. There were an estimated 250 of these exchange sites in early 1999. By the beginning of 2000, the number had expanded to roughly one thousand, while the amount of venture capital in these new companies increased from $1.8 billion in 1998 to $23.4 billion in 2000 (eMarketer 2001; Business Week, October 9, 2000). Many of these exchanges operated in the same industry. Metalsite and esteel competed against one another to organize transactions between steelrelated firms. Other sites such as VerticalNet and FreeMarkets organized buying and selling between firms in a range of related industries. What was similar in these start-up sites was their aim to attract established firms to their exchanges. Predictions about a massive migration of interfirm trade to these independent exchange sites, however, proved erroneous. Many of the existing firms, after initially experimenting with these exchanges, later abandoned them, preferring to establish their own sites rather than relinquish control to an intermediary (Kenney 2001, 37). Furthermore, these start-ups also suffered from the effects of the Internet downturn beginning in the spring of 2000. After their extremely buoyant start, in which over one thousand of these exchanges had emerged in the first quarter of 2000, by October of the same year roughly 90 percent were out of business with the survivability of the remaining independent Internet exchanges very much in doubt (Business Week, October 9, 2000).19 Many firms that tried using them at the outset of the Internet boom realized that relinquishing control to an intermediary was not in their interests (Business Week, December 4, 2000). Although Dell was among the existing manufacturing firms that experimented with independent Internet exchange sites such as FreeMarkets for procuring certain types of components, the PC maker largely abandoned these exchanges as a viable Internet supply chain strategy. Dell had a far different idea of how it could use the Internet in its business model. Despite the initial success of models for Internet commerce developed by firms such as General Electric, which essentially duplicated the

164 The Internet independent exchanges without relinquishing control, Dell rejected this brokerage-oriented approach to the Internet. Aggregations of myriad unknown firms through the Internet’s brokerage effects and transactions with such agents through markets, supposedly made much more efficient and free through this process of aggregation, had little to offer the PC maker as a way of making its supply chain system more efficient. For Dell, it was the need for strategic relationships with supply chain partners that played the decisive role in how the firm would deploy the Internet in its procurement, production, and distribution system. Dell would use the Internet as a mechanism to control these relationships and shape the creation of efficiency and competitive advantage in its procurement, production, and distribution organization.

chapte r

6

Global Reach The Business Organization of Dell Computer

The Internet offered a logical extension of our direct model. . . . This became the key to what I call a virtually integrated organization. Michael Dell Let’s face reality. If my largest customer had a new requirement, I’d listen to them. Michael Dell

The Internet as Process Innovation On the walls of a stairwell near the entry to the Dell Computer Corporation Tofper Manufacturing Center in Austin hang two large arrangements of picture frames, aligned row upon row, each frame encasing a patent awarded to the world’s largest PC firm. Embedded in this display is a story with an unmistakable aim. Represented in each frame is an idea. On the basis of merit—and a property right conferred by government—each of these ideas has ascended to a privileged status reserved for the phenomenon known as innovation. The “wall of patents,” as it is referred to by Dell, is a metaphor intended to convey Dell’s story as an innovative firm. Dell Computer has emerged as one of the most innovative firms of the current period on the basis of a process for producing, buying, and selling. This process-oriented route to profit, in turn, rests on three fundamental elements: (1) a different relationship with the customer that involves the practice of selling direct, absent intermediaries; (2) the notion of customization for the specific needs of end users and building customized products in high volume; (3) the use of the Internet by Dell to reorganize not only its system of direct sales but, more significant, the procurement and assembly opera-

166 Dell Computer tions underlying mass customization. By grafting its system of custom direct sales and procurement on to the Internet infrastructure, Dell has transformed these activities, creating what is arguably the most innovative procurement, production, and distribution organization ever built.1 Although it manufactures computers, Dell accumulates profit as a logistics firm. It is an organization of knowledge and routines extracting surplus not from production, but by managing the movement of product and information flows along a globally dispersed network of companies engaged in the various operations of producing and marketing finished PCs. As Dell’s vice president for manufacturing emphasizes, “We are not experts in the technology we buy. We are experts in the technology of supply chain integration. We have created this expertise with the Internet at its core” (vice president, interview, July 18, 2001).2 (See the appendix at the end of this chapter for a list of those interviewed both within and outside of Dell.) In this role as a logistics company, Dell’s most revered accomplishment consists of the degree to which it is able to balance supply and demand of product flows among the firms in this global commodity chain. “Supply and demand balancing,” insists this same Dell vice president, “is one of the most important capabilities at Dell” (vice president, interview, May 24, 2001). The key to this core capability in material balancing, however, lies in Dell’s capacity to process Internet information flows that it uses to manage the external capabilities of other firms. This mastery over material balance flows and Internet information flows has enabled Dell to create an organization differentiated from its competitors by the degree to which it has succeeded in accomplishing a singular aim: accelerating speed and compressing time in the movement of materials as they pass through the adjacent steps of customer order, procurement, production, and final product delivery (Curry and Kenney 1999). The most visible benchmark of this core capability in material balancing focuses on the levels of inventory Dell maintains. In 1994, when Dell launched its Internet strategy, it carried an average of thirty-two days of inventory in its procurement and production chain. By 1997, as Dell began to deploy the Internet more fully in its supply chain operations, the figure had shrunk to thirteen days. In mid-2002, Dell was carrying four days of inventory, while at the same time, Compaq, the firm Dell surpassed in becoming the world’s largest PC maker, held six weeks of inventory (Cook, interview, May 17, 2002; Business Week, June 17, 2002). In order to achieve this level of balance in demand and supply conditions, Dell has had to complement its Internet-based logistics activity with a different type of organizational relationship between itself and the other firms in its network. While this form of organization shares the dis-integrated interfirm structure of the networks organized by other PC makers, it

Dell Computer 167

differs in the degree to which its operations and its relatively small number of partner firms are functionally integrated by means of Internet communications. Dell refers to the structure of organization deriving from these collaborative relationships between formally separate firms as virtual integration. Although nominally separate, Dell and the firms making up its virtually integrated organization do not produce finished PCs on the basis of arm’slength interactions mediated through markets and the price system. On the contrary, Dell organizes highly structured relationships of collaboration between itself and its network partners on the basis of its power to manage and control these other firms through what Coase (1937) describes as “planning,” Williamson (1975) as “hierarchies,” and Chandler (1977) as the “visible hand.” Such mechanisms of control, typically associated with vertical integration, are in fact integral to Dell in securing the collaboration from its interfirm partners necessary to organize its high-velocity supply, production, and distribution chain and build custom-configured PCs in high volume on a just-in-time basis. The Internet plays a vital role in reinforcing Dell’s capacity to control these collaborative relationships.3 The Internet has provided Dell with the tools to create its own Web-based communications protocols as the foundation for collaboration with suppliers and logistics partners. The latter have little choice but to accept these protocols and integrate their operations around them if they want to remain within the PC maker’s organization. Dell, in effect, has successfully exploited the Internet infrastructure in creating a system of collaboration for suppliers and logistics providers that uses Internet technology as terms for entry into its network. In this way, Internet communication has enabled Dell to secure the benefits of organizational control associated with actual integration without its costs. Dell’s interfirm network, with its operational attributes of time compression in material balancing and its organizational attributes of virtual integration, is also creating a distinct geographical pattern of economic activity. This pattern is marked by the interplay of two spatial tendencies operating simultaneously, one of geographical spread, the other, concentration. Long distances separate the different regions where Dell has located its build-toorder operations. Long distances also separate Dell’s build-to-order operations from the locations of its key suppliers. On the other hand, Dell concentrates its build-to-order operations in selected localities where it organizes critical relationships of proximity between certain key nodes and activities in its network. Dell itself is the agent in creating these relationships, most notably in its requirements that suppliers maintain either factories or supply hubs within twenty minutes’ driving distance of Dell assembly sites. In this way, Dell is not only influencing the location patterns of firms in the

168 Dell Computer concentrations of high technology where it operates. Dell is also providing a compelling picture of how its Internet-based build-to-order innovation is concentrating and spreading the geography of economic activity across regions while defining processes of economic globalization as they actually occur on the ground. This Internet-driven, virtually integrated organization enabled Dell by 2001 to become the world’s largest and arguably most competitive PC maker. Perhaps more important, most other major PC firms have attempted to imitate aspects of Dell’s Internet-based business system. As Robert Cihra, computer analyst for ING Barings, prophetically affirmed just prior to the announced merger of Hewlett-Packard and Compaq, “the number one issue on the mind of virtually every major PC vendor is defense against Dell, and for Compaq, IBM, and HP, ‘Dell’ has truly become a four-letter word” (Cihra, interview, August 8, 2001). How Dell ascended to this position is the story told in this chapter. This chapter profiles the key operational, organizational, and geographical elements of Dell’s Internet-based procurement, production, and distribution organization.4 In organizing this profile, this chapter focuses on a recent supply chain integration project implemented at Dell known as DSi2. This project, touted by Dell as “the single biggest change in the Dell business system” is more accurately the culmination of an ongoing process of experimentation with the Internet that began in the mid-1990s and is still continuing (vice president, interview, June 5, 2001; interview, August 24, 2001).Yet at the same time the project represents the conquest of new terrain in the PC maker’s ongoing effort at recasting time and distance to become more profitable. The chapter is organized historically. Within this historical account of Dell’s ascendancy, however, are themes that tell a contemporary story of economic development. How the structure of the PC industry came into being; how Dell’s early business model challenged the competitive foundations of the industry; how Dell used the Internet to transform its original business model and create a uniquely innovative business organization; how, in assuming a position of competitive superiority and diffusing among other firms, this virtually integrated organization has changed the PC industry; and how the attributes of this innovative organization are redefining the territorial structures of the global economy are the contours of the story that follows.

Competition and the PC Before Dell When Dell Computer began operations in 1984, the PC industry, though still very young, had already developed an industrial structure with clearly

Dell Computer 169

defined terms of entry and competition and a dominant system of value creation and profit making. These attributes would provide Dell with opportunities for innovation on the basis of a very different vision of how to compete and make profit. In order to grasp both the initial phase of innovation at Dell, when the PC maker developed its custom direct business model, and the second phase of innovation, when the firm adapted this business model to the Internet, it is essential to identify the salient attributes of the industry as it evolved since the commercialization of the personal computer. ibm and the pc In the aftermath of initial commercialization of the personal computer in the mid to late 1970s, the most profound impact on the PC industry resulted from IBM’s decision in July 1980 to enter the personal computer market. In making this decision, the world’s largest mainframe computer maker committed to bringing a personal computer to market within one year. Despite its expertise in computing, such a goal was a formidable challenge. “The only way to get into the PC business,” explained William Lowe, director of the PC project for IBM, “was to go out and buy part of a computer company or buy both the CPU and software . . . because we can’t do this within the culture of IBM” (quoted in Langlois 1992b, 21). Consequently, the strategy pursued by IBM to develop the product within such a short time frame consisted of outsourcing component production and marketing the product through existing independent retail channels. This strategy enabled IBM to bypass its own traditional business model of using internally developed technology and its own internal sales force. In this sense, the IBM PC was a striking example of reliance on external capabilities to build a product and bring it to market (Langlois 1990, 1992b). While this production and selling system proved enormously successful for IBM, the impact on the rest of the industry derived from the firm’s competitive ascent and the resultant influence of its business model as a competitive standard for other PC firms. In this way, IBM was decisive in shaping the development of a production and selling system that would influence the way PCs were built and marketed for the next two decades. Dell would exploit the opportunities for new entrants created by the ascendant IBM business system and the inefficiencies in the marketing element in entering the industry and challenging the competitive standards on which the industry was based. Although in creating this system IBM drew upon internal sources for keyboards and circuit boards, most suppliers were firms outside of IBM. Tandon made disk drives; Zenith furnished power supplies; Epson made printers; SCI Systems, a contract manufacturer based in Alabama, stuffed the circuit boards; and China Picture Tube, part of the Taiwanese electronics

170 Dell Computer firm Tatung, made the monitors. The computer was assembled in Boca Raton, Florida, from these components. For distribution, IBM decided to market its PC primarily through retail computer outlets rather than its own sales agents. It even solicited input from the largest dealer, ComputerLand, on how best to accomplish this aim. In pursuing this marketing strategy through independent retailers rather than relying on its own sales organization, IBM successfully elevated the role of distribution as an element of competition in the overall value chain for the PC. Within the context of this strategy to rely on external capabilities, two procurement decisions by IBM stand out in terms of their impact on the development of the industry: the sourcing of the microprocessor and the operating system software.5 In shopping for a microprocessor rather than developing one of its own, IBM decided to use a sixteen-bit processor from Intel, the 8088, rather than the eight-bit processors used in existing PCs. Although this processor did use eight-bit external buses for which there existed supporting chips, the decision meant that the IBM PC could not use operating systems linked to 8-bit processors. Consistent with its strategy to use the market for necessary technology, IBM turned to the small Seattlebased software company active in the earlier phase of the PC industry, Microsoft, for the operating system. Microsoft, in turn, purchased a system from another software supplier, made some small modifications to it, and sold it to IBM. The product was known as PC-DOS. In a decision with profound implications, IBM allowed Microsoft to license the operating system to other PC makers as MS-DOS without having to share royalties. In purchasing the microprocessor from Intel and allowing Microsoft to license the operating system, IBM effectively ceded control of the two most critical elements of the PC architecture. The IBM PC had essentially evolved on the basis of an open and largely modular architecture and a global sourcing system for that architecture.6 Despite IBM’s having lost control over the architecture of its PC, and despite the lack of significant technological advance in the product, the IBM personal computer was hugely successful. Demand for the PC exceeded its own sales forecasts by 500 percent as the 13,533 units shipped during the final months of 1981 were far short of supplying the large backlog of orders (Langlois 1992b, 23). By 1983, IBM had captured almost 20 percent of the world PC market, with roughly 750,000 units shipped out of a worldwide total of 3.9 million (Langlois 1992b, 23; Steffens 1994, 167). Its expansion effectively pushed growth rates throughout the entire industry (Table 6.1). From a position as a relatively late entrant, IBM catapulted to a position just behind Commodore as the world’s second-largest PC maker.

Dell Computer 171 table 6.1 The Early PC Market in the United States, 1976–1983

Units sold (thousands) Value ($ millions)

1976

1977

1978

1979

1980

1981

1982

1983

17

41

121

181

246

380

792

1,764

36.0

74.5

223.8

302.5

495.0

936.9

2,002.3

4,718.0

source: John Steffens, Newgames: Strategic Competition in the PC Revolution (Oxford: Pergamon Press, 1994), 90– 126, 167.

large firms, suppliers, and clones Three critical developments followed from IBM’s success that affected the competitive structure and the development trajectory of the PC industry. First, IBM’s entry compelled other large companies from the office products and consumer electronics industries to enter the PC marketplace. These entrants included Xerox, Hewlett-Packard (HP), Texas Instruments, Zenith, DEC, and Wang Laboratories. This group also consisted of numerous foreign firms, among them NEC, Sanyo, Hitachi, Toshiba, and Fujitsu from Japan and Philips and Olivetti from Europe. Nevertheless, these firms exhibited widely varying levels of success. NEC, HP, Toshiba, Fujitsu, and Olivetti were able to gain between 4 and 9 percent of the U.S. market for differently priced PCs (Steffens 1994, 177). Other firms, most notably Xerox, were surprising failures in the PC market. Despite the qualified success of some of these large firms, however, by the beginning of 1984, IBM, Apple, and Tandy still dominated the PC market in the United States. Second, the growing demand for the product, coupled with the PC’s nonproprietary architecture, created an enormous market opportunity for suppliers of components, peripherals, and parts. An entire industry of specialized suppliers emerged to exploit these opportunities linked to the fortunes of the IBM PC. In addition, in launching its PC on the basis of highvolume production, IBM provided many of these component vendors with opportunities to scale their operations from the outset. As a consequence, these firms were able to develop production efficiencies and cost advantages very early after entry into the industry. Firms most notably in Taiwan but also in Singapore and South Korea entered this market and became skilled producers of these components.7 “IBM put its suppliers into the high-volume business,” and in the process “so bore their start-up and learning costs” (Ferguson and Morris 1994, 52). By the mid-1980s, Singapore and South Korea along with Taiwan had emerged as major world centers of production for the industry (Table 6.2).

172 Dell Computer table 6.2 Computer Hardware Production in East Asia, 1980–1990 ($ millions)

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990

Taiwan

Singapore

South Korea

80 110 170 430 1,040 1,260 1,739 2,890 4,001 5,046 5,886

— — 147 530 1,066 1,194 1,914 2,928 4,503 5,368 6,974

9 31 47 207 428 579 880 1,459 2,431 3,180 3,073

sources: Jason Dedrick and Kenneth L. Kraemer, Asia’s Computer Challenge: Threat or Opportunity for the United States and the World? (Oxford: Oxford University Press, 1998), 321;Young-Ja Bae, “Globalization, Institutions, and Industry Development: A Case Study of the Personal Computer Industry in Korea and Taiwan” (Ph.D. diss., University of North Carolina, Chapel Hill, 1998), 79, 148; Singapore Economic Development Board.

Finally, and perhaps most significant, demand for the PC, which IBM was not able to supply until eighteen months after it introduced the product, created an enormous market opportunity for other producers to clone IBM’s machine. The expanding supply base along with the PC’s open architecture provided these aspiring PC makers with a foundation for copying and building the product. The one potential hurdle for PC makers—duplication of the IBM input/output system software known as BIOS, which enabled the PC to transmit and receive data—proved far from insurmountable. Firms such as Compaq successfully reverse-engineered the BIOS specs in a way that insulated them from copyright and patent infringement regulations. Because of the market in components that had developed from IBM’s reliance on external capabilities, all that the clone makers needed to duplicate the PC was to buy the 8088 microprocessors from Intel, the MS-DOS from Microsoft, and the remaining hardware components from the expanding base of supplier firms. Furthermore, the interests of clone makers actually converged with those of Intel and Microsoft, whose fortunes were dependent more on an expansion in the number of PCs than on any one-way relationship with IBM. By 1983, Compaq was the largest among nearly one

Dell Computer 173

hundred clone makers that emerged in the wake of the IBM PC’s success. Two years later this upstart was fifth-largest firm in the PC industry and was ascending rapidly. Hundreds of other firms would soon follow, supported by a base of specialized suppliers numbering in the thousands (Dedrick and Kraemer 1998, 56). open standards and modularity IBM’s decisions to outsource most components of the PC and relinquish control over the microprocessor and operating system created a product architecture based on relatively open standards that had three enduring impacts on the nature of competition in the industry. First, open standards for PC components enabled the PC to become an increasingly modular product. Second, this attribute of modularity—the ability of PCs to be assembled from standardized components “much like Legos”—provided the foundations for PCs to be custom-produced in high volume (Langlois 2001, 26). Third, and perhaps most decisive, modularity and standardization diminished the role of technology as an element of competition in the PC industry and elevated the role of efficiency in procurement, production, and distribution of the product. Modularity, in effect, was the critical precondition of Dell Computer’s business model. The modular character of the PC diverged dramatically from the era of central computing that preceded it, during which mainframes and minicomputers produced by firms such as IBM or DEC had been proprietary “closed systems.” The architectures of these systems were largely incompatible with those produced by other vendors. As a result, firms producing mainframes and minicomputers tended to be highly integrated. Because these systems were closed and proprietary, firms building them had to produce components in-house and scale the production of these components sufficiently to achieve cost-efficient internal scale economies. Nevertheless, this integration did not include all computer components. Mainframe and minicomputer firms integrated vertically with respect to components of high value to reduce their risk and increase their control over production of computer systems. The openness of the IBM PC, by contrast, meant that firms defined standards for interoperability between different components more or less publicly and collectively. Such agreement on standards is basically “unsponsored,” arising de facto in a competitive environment, and is distinct from standards agreements that are negotiated and implemented by statute or political authority (David 1987; David and Greenstein 1990). As a result of this market-driven process of agreement, third-party vendors could design and build components on the basis of publicly defined standards that would operate together as a system. This openness enabled different hardware and

174 Dell Computer software firms to produce compatible and interchangeable components that could be integrated into the PC as a system created by the efforts of numerous separate firms. Assembled from components available in the market produced by a variety of firms, the IBM PC enabled a base of suppliers and imitators to proliferate. Although the microprocessor and operating system markets became virtual monopolies dominated by two firms, Intel and Microsoft, the remaining hardware segments of the industry evolved into extremely competitive businesses. Production of these components by the mid-1980s had already become a dis-integrated global activity, driven by the process of standardization and modularity established by IBM protocols. As a result, the industry became one of horizontal partnerships instead of vertical integration marked by the emergence of a large concentration of producers in East Asia. This industry structure created an entirely new set of management challenges for PC builders as coordination of the production system for PCs moved from inside to outside the boundaries of the firm (Langlois 1990, 1992b). At the same time, products built by different companies were virtually identical in terms of technology. The personal computer was, in effect, a product of globally organized interfirm production networks in which productive capacity, and to some extent even innovative capacity, had migrated to new areas of the world. The result was an industry of fundamentally similar products. From the outset, design and production of the PC’s standardized modular components evolved on the basis of the relationship between three basic elements making up personal computer systems: (1) semiconductor and microprocessor families, (2) operating and input/output systems, and (3) applications software (Steffens 1994, 121). Changes in any one of these elements created incentives for changes in the others. As these elements evolved, the result was a transformation in the hardware architecture in which these elements were housed. While all three elements were critical in promoting ongoing changes in the PC, the dominant driver in transforming the personal computer was the continuous improvement in semiconductor and microprocessor technology. Known as Moore’s Law, these advances doubled the number of transistors packed into microprocessors every twelve to eighteen months and forced the costs as a measure of performance for computing to plummet. From 1982 to 1998, the estimated cost per million instructions per second fell by a factor of five hundred (Rowen 2000, 194). Such changes in performance and cost resulted in average annual rates of decline in prices for microprocessors per transistor of 35 percent from 1985 to 1996, while for memory chips the decline was 20 percent over the same period (Curry and Kenney 1999, 11). This pattern of change enabled Intel, with a dominant position in

Dell Computer 175

the market for microprocessors, to set the key standards for the ongoing changes in PC hardware (Dedrick and Kraemer 1998, 73). Other components, however, most notably hard disk drives, developed along a similar trajectory. The average annual rate of decline in the price per megabyte of storage for hard drives from 1980 to 1989 was 30 percent. Such expansion in the capabilities of computer components compelled PC hardware firms to develop new generations of personal computers at regular and ever-shorter intervals (Steffens 1994, 151). As PC firms developed new products, however, the price for these products typically declined between 20 and 41 percent per annum over the life of each new PC model (Berndt and Griliches 1993; Curry and Kenney 1999, 12). As a product in a state of ongoing technological change, yet evolving along a pathway of modularity and standardization, the PC was susceptible to competitive pressures deriving not only from technology but also from the interplay of two fundamental variables, price and time (Curry and Kenney 1999). As soon as a new product iteration came to market, it was under constant downward price pressure, its value shrinking with the passage of time in anticipation of the next wave of new processing technology and application software. With technological change a constant and with price and time emerging more as the defining elements of the market environment for the PC, terms of competition for PC makers actually shifted away from technology to a very different aspect of the PC value chain. As product performance became increasingly standardized within the supply base and from one PC company to the next, and as product differentiation became difficult to sustain, distribution emerged as a primary competitive factor within the PC value chain (Steffens 1994, 259). This emphasis on distribution, in turn, would eventually elevate the role of logistics as a competitive variable in producing and selling PCs. Again, the legacy of IBM was a critical factor in this shift in the way it established competitive standards of the indirect channel. the indirect channel By 1983, the PC market in the United States reflected eight identifiable channels of distribution to end users (Steffens 1994, 160). These channels included (1) office products dealers; (2) value-added resellers and systems houses; (3) manufacturers’ own office retail stores; (4) wholesalers; (5) mail orders; (6) mass merchandisers; (7) direct sales primarily by the sales forces of larger PC makers, including IBM; and (8) computer specialty dealers. Of these, computer specialty dealers were overwhelmingly dominant, primarily because of the influence of IBM. By this time, PC makers had adopted IBM’s emphasis on the professional retailer as the preferred channel of distribution. In order to attract customers, especially business customers, com-

176 Dell Computer puter firms had to ensure that their sales channels had the same standards as those of the industry leader (Steffens 1994, 197). This indirect channel also represented a route to legitimacy for clone makers, notably Compaq. The decision by IBM to use professional retailers as its primary distribution channel and the influence of IBM on the rest of the industry changed the way many PC makers deployed resources. PC firms made investments in relationships with resellers and dealers and in marketing activities to support these relationships on the assumption that intermediaries provided manufacturers with better access to nationwide sales. PC makers in effect veered away from the earlier technology bias and toward a new emphasis on marketing and distribution activity (Steffens 1994, 197, 269). By 1984, this indirect channel of selling computers through intermediaries emerged as a defining element in establishing terms of competition in the industry. To compete, firms had to rely on these intermediaries. While there was value added to the PC by these entities as the product circulated from the manufacturer to the customer, there was also the inefficiency of an excess number of actors. So dominant was this system that even by 1987, with Dell and Gateway already selling direct, specialty computer dealers still accounted for 56 percent of total shipments, while collectively the various indirect channels accounted for 80–90 percent of all PC sales (Steffens 1994, 260). As a consequence, the idea of capturing the value created by intermediaries in the PC production and distribution chain presented a compelling potential opportunity for PC firms. entrepreneurial opportunity To capture the value created from intermediation in the PC value chain, the entrepreneurial firm had to reconceive and transform the relationships between four key elements at the core of the indirect system of distribution: (1) forecasting market demand; (2) building according to demand forecasts; (3) “pushing” finished inventory from the factory into the distribution and sales channel; and (4) waiting for customers to make purchases. While these attributes functioned together in creating a production and distribution system, demand forecasting served as the foundation for the other three and the source of the leakage in value from PC firms. Dependent on accurate forecasting of future demand, PC firms were vulnerable to the problem of unsold goods when forecasts went awry. At the same time, they were susceptible to missed opportunities when forecasts underestimated the market. Even with the best demand projections, however, the indirect push system of production was beset with an intractable problem for the PC manufacturer—the need to purchase and hold stocks of components and subassemblies in order to build finished systems and fill quotas for distribution outlets. Such challenges are not unique to the PC industry. They are

Dell Computer 177

symptomatic of all industrial activities that rely on demand forecasting and the sale of finished goods through intermediaries. These problems for the PC maker stemming from the forecasting system and the indirect channel, however, became exacerbated by the nature of technological advance in the personal computer itself. Driven by incessant advances in processor technology and software applications and the resultant shortened product life cycles, the PC is susceptible to ongoing downward price pressure as it circulates through adjacent stages of the value chain. Once assembled, the PC is constantly losing value because its components depreciate as time passes in anticipation of the next wave of technical improvements. These two characteristics—constant technical improvements and simultaneous downward pressure on PC prices—give the personal computer a perishable-like quality, similar to the products of industries such as fashion and even food (Kraemer, Dedrick, and Yamashiro 1999, 3; Kenney and Curry 2000a, 5).8 For this reason, the indirect channel of distribution was particularly well suited as a target for competitive challenge in the PC industry. This perishable-like quality of the personal computer and the depreciation in value of the product over time exposes PC firms to an especially vexing problem—the problem of inventory. If the product is held for any appreciable length of time in inventory, it is depreciating in value. Such a problem becomes increasingly acute for the PC maker at the stages in the value chain where the product moves from the final assembly point through the channel and eventually into the hands of the customer. Frequently, the period of time between final assembly and sale to the customer ranged from three to four months. During this time, the components in the PC, most notably the central processing unit, hard disk drives, and DRAMs would depreciate in value such that the selling price, by the time the system was actually purchased, would have to be lowered. This increment of time in the channel, in effect, represented a sizable loss for the PC maker. For this reason, competition in the PC industry as early as the mid-1980s began to evolve decisively along the path of speed and time (Curry and Kenney 1999).9 Such a shift created an opportunity for innovation focusing on the logistics of building the PC and distributing the product to the final customer. Any strategy for decreasing the period of time in inventory at each step in the value chain, especially between final assembly and sale, and accelerating the speed at which the build process and final marketing occurred held enormous potential as a business model in the industry. Already by 1984, the structure of the PC industry created opportunities for a business model based on speed and time. First, the industry was already organized on the basis of decentralized and globalized subcontractors. Virtually all computer makers had access to and were subcontracting from this

178 Dell Computer supply base located in East Asia, Japan, and the United States. Second, the PC components produced by this supply base had evolved into highly modular and increasingly standardized items. With uniform design, engineering, and technology, components were easily accessible to PC makers. Differences in the product from one firm to the next, whether branded or cloned, had narrowed so dramatically that, with the exception of Apple, design, technology, and engineering were of secondary importance as terms of competition in the industry. What differentiated these products were brand names. Nevertheless, even the IBM brand failed to stem the ascendancy of clones, most notably Compaq, but also the nonbranded clones. In effect, the PC had become a standardized commodity. It was built from components purchased from an accessible base of suppliers and distributed through a channel of actors that functioned on principles of demand forecasting and accepted the accumulation of inventory as inevitable. Admittedly, the process required enormous logistical coordination to build the product from this existing base of firms and market it successfully through the existing distribution channel.Yet, in evolving with an increasing emphasis on logistics and distribution, the PC industry was vulnerable to change from a business model challenging the principles upon which the prevailing system of logistics and distribution was organized. That challenge would emerge in 1984.

The Genesis of Dell A single core concept served as the inspiration for Dell Computer: “Sell computers directly to the end customer. Eliminate the resellers’ markup and pass those savings on to the customer” (Dell and Fredman 1999, 12).10 The target of Dell’s business model was thus not the technology of the PC but instead the industry’s indirect channel of distribution. A different relationship with the customer provided the underlying foundation for this business model. This relationship was the source of innovation at Dell and the catalyst for transformation in the company’s logistics-oriented production and distribution system. custom direct Michael Dell began his business in late 1983 by upgrading IBM personal computers from his dorm at the University of Texas and selling the customized PCs directly to businesses. In Dell’s own words, the superiority of these two notions—direct selling and customization—was obvious. What was less clear to Dell was the reason why existing computer firms were not producing and marketing their products in this way. Dell achieved several advantages from customized producing and direct

Dell Computer 179

selling. The most decisive advantage derived from the elimination of intermediaries in the PC value chain. Such disintermediation not only enabled Dell to capture that portion of the value taken by PC wholesalers in the process of distribution but also was critical in compressing time in the cycle of PC production and distribution itself and capturing greater levels of value by increasing the velocity of moving products from order through final sale. Second, the relationships with final customers built through direct selling provided Dell with a platform for one of the most important sources of sales in the industry, repeat sales. Because the technology of the PC was changing so rapidly, and because the product life of the PC was so short, customers buying a machine at any given moment were the best prospects for future sales. Furthermore, by eliminating product in the channel, Dell was in a position to offer its customers the latest technology without having to send products through a lengthy distribution process. Third, in producing PCs only after receiving orders, Dell avoided the inventory problems associated with faulty demand forecasting. Finally, in purchasing components and building finished systems only after receiving orders and payment from customers, Dell was able to carry a negative cash conversion cycle. It took possession of the customer’s money before paying its suppliers, thereby funding its own operating expenses. Excess inventory held by distributors of certain PC makers provided Dell with an early source of low-priced PCs from which the firm could turn a profit. Especially important was the “IBM gray market,” consisting of unsold PCs at certain dealers. “We would buy these stripped-down computers,” admitted Dell, “and sell them for a profit” (Dell and Fredman 1999, 14). In addition, Dell’s fledgling firm capitalized on another fortuitous opportunity to seed the business. An open bidding process in the State of Texas for PC hardware enabled Dell to compete for public contracts with more established firms. In winning several of these bids, Dell obtained a source of revenue to supplement its base of individual sales and began to grow rapidly. By early 1984, Dell was selling fifty thousand to eighty thousand dollars per month to customers in the Austin area. In May, Dell incorporated the firm. By the beginning of the following year, the company was building computers under its own brand name—PC’s Limited. Dell admitted to taking advantage of the three primary characteristics of the PC industry structure in entering the market relatively easily and building computers under its own brand. First, according to Dell, the burgeoning base of PC suppliers, both in the United States and in East Asia, provided the firm with access to the technology and engineering necessary to build the product (Dell and Fredman 1999, 23–24). Second, many of these components had become so technologically standardized that assembly into finished PCs had become a relatively low-skill activity requiring limited in-

180 Dell Computer Indirect Channel of the PC Industry

PC Maker

Suppliers

Distributors

Integrators, Resellers, Retailers

Final Customer

Dell’s Direct Channel

!~!

Suppliers 1·~

,, . ',. ''

,!:

:"

Assembly site Major supply site Supply flow

© Jeremy Shaw

Dell Computer 183 table 6.4 PC Firms Ranked by PC Revenues, 1990 Rank 1 2 3 4 5 6 7 8 9 10

Firm IBM Apple NEC Compaq Toshiba Olivetti Fujitsu Packard Bell Groupe Bull Commodore

Rank

Firm

11 12 13 14 15 16 17 18 19 20

Intel Tandy Acer Hitachi Siemens / Nixdorf AT&T Seiko Epson Hewlett-Packard Unisys Dell Computer

sources: John Steffens, Newgames: Strategic Competition in the PC Revolution (Oxford: Pergamon Press, 1994), 335; Jason Dedrick and Kenneth L. Kraemer, Asia’s Computer Challenge (Oxford: Oxford University Press, 1998), 57; Paul Pinella, “The PC Big 50 1990,” Datamation, December 1, 1991, 42.

difficult to keep in balance the two elements of its business model—low inventory and mass customization—which in other business systems are generally irreconcilable. Volume in Dell’s build-to-order direct system had thus reached a certain threshold where parts for orders had to be stored, revealing the vulnerability of the company’s business model in a high-volume setting. In addition to high volumes, these problems were traceable to the geography confronted by Dell in organizing supply chain operations for its system of mass customization. Long-distance linkages between suppliers, located primarily in East Asia but also Mexico and the United States, and assembly in Austin characterized the geography of this network (Map 6.1). Problems in coordinating the transport of these components over distance and accounting for the variation in component delivery lead times from suppliers conspired to push up inventory levels. Numerous separate warehouses in Austin, where these components were stored before assembly, exacerbated the complexities of the different lead time delivery schedules. “To our stunned belief,” writes Dell about the period around 1990, “we had quickly become known as the company with the inventory problem” (Dell and Fredman 1999, 37). tactical shift Dell attempted to remedy these inventory problems after 1990, as well as expand sales, by implementing two related and perhaps surprising modifica-

184 Dell Computer table 6.5 Computer Firms Ranked by World Market Share, 1994 Rank

Firm

Units (000s)

1 2 3 4 5 6 7 8 9 10

Compaq Apple IBM Packard Bell NEC Hewlett-Packard Acer Toshiba Fujitsu Dell

4,799 3,957 3,937 2,473 1,941 1,903 1,451 1,442 1,441 1,152

Market Share (%) 10.0 8.3 8.2 5.2 4.1 4.0 3.0 3.0 3.0 2.4

source: Gartner Group / Dataquest.

tions in its custom direct business system. First, Dell began to use a key actor in the indirect channel, value-added resellers (VARs), to reach certain business customers (Aragon, 1998). Second, Dell supplemented this approach to partnering with the VARs with an extensive program to market its computers through large retailers. The company negotiated deals with computer superstore CompUSA, big box retailer Price Club, office supply franchise Staples, and electronics chain Circuit City. These stores agreed to sell Dell PCs at the company’s direct prices. Indeed, as a result of these moves, Dell’s sales continued to expand. By 1994, the firm had generated annual revenues of just under $3 billion. Furthermore, with a world market share of 2.4 percent, Dell had ascended into the ranks of the world’s largest PC firms. Shipping over a million units, it now occupied the position of the world’s tenth-largest PC maker (Table 6.5). Nevertheless, this period of growth and ascendancy into the top ranks of PC firms also marked a negative milestone in the company’s brief history. Although Dell’s sales had expanded impressively, and although by 1994 it had become even more of a presence in the world PC industry, the company was not able to sell profitably through the intermediaries of the indirect channel. Its cost of sales rose dramatically, consuming profitability. Whereas in fiscal year 1991 Dell’s cost of sales were 66 percent of its total sales, by fiscal year 1994 the figure had risen to 85 percent (Dell Computer Corporation, 1994 Form 10-K, Item 6, pp. 13–14). Furthermore, its tactical shift into the indirect channel, though it expanded sales, exacerbated its difficulties with component procurement and excess inventories.

Dell Computer 185

As a result, Dell suffered its first quarterly loss during this period. For its fiscal year ending January 30, 1994, Dell reported a net loss of $35.8 million on its $2.9 billion in sales, the first and only annual loss, to date, reported by Dell (Dell Computer Corporation, 1994 Form 10-K, Item 6). Rising costs linked to larger-than-expected inventories in its procurement chain, a failed program with VARs, and too much product in the retail channel created the first significant crisis at the company (Wood 1993; Aragon 1998). Consequently, in early 1994, Dell made two critical strategic decisions. First, the company decided to abandon its experiment with resellers and concentrate on its origins as a direct seller. Second, and perhaps more significant, Dell began to experiment in adapting its custom-direct business model to an entirely new communications infrastructure—the Internet.

The Internet Dell’s Internet-driven business organization has taken shape in two overlapping and still ongoing phases. In the first phase, Dell used the Internet to transform its system of selling and order intake with its customers, enabling the PC maker to become one of the early pioneers of Internet commerce. The second phase consists of Dell’s deployment of the Internet to reorganize its procurement planning and the logistics of assembly and product delivery. Its aim in this second phase is to link its Internet-based system of customer order intake to an Internet-enabled process of procurement, production, and distribution with suppliers and logistics providers. What has occurred alongside this deployment of the Internet is an extraordinarily sharp acceleration in growth. From 1994 to 2001, revenues increased at Dell almost ten times, demarcating this period of sales growth from the early years of the 1990s while productivity per employee, measured by the differential rates of growth in sales and employment, also expanded dramatically (Figures 6.2 and 6.3). This period marks Dell’s ascent to the very top ranks of the PC industry. In addition to sales growth, this integration around the Internet has recast Dell in three profound ways. Operationally, Dell has used the Internet to create more rapid cycles in its system of pulling components from suppliers. In a build-to-order environment where PC systems are produced and delivered at rates that match the rate of sales, Dell has been forced to implement these compressed cycles precisely because of its explosive revenue growth after 1994 (Albers 2000, 3). In this sense, sales growth and time compression in procurement and assembly have been mutually reinforcing. Organizationally, in extending the Internet backward into its supply chain, Dell has created a new form of business enterprise with its interfirm partners to manage the more rapid cycles in its build-to-order system. This form of enterprise

186 Dell Computer Sales ($ billions) 35 30 25 20 15 10 5 0

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

figure 6.2 Dell’s Growth Before and During the Internet Period source: Dell Computer Corporation, Form 10K, 1992, 1996, and 2001.

relies on strategically organized relationships, not markets, as a basis of collaboration. Finally, Dell has created a distinct form of territorial organization in the places where it has established its build-to-order operation. In these places, physical proximity between supply sources and assembly sites, strategically imposed by Dell on its network partners, has resulted in a set of territorial relationships between operations and actors that are shaping the attributes of a new generation of industrial districts. These industrial districts, in turn, are helping to define the actual character of global production ensembles in the contemporary economy. While these attributes emerged most sharply following implementation of the DSi2 project, they began to take shape alongside Dell’s initial experiments in on-line selling, and deployment of the Web in supply chain management, and enterprise resource planning. on-line selling Dell had a decided advantage over other PC firms in selling its products over the Internet. Unlike the indirect channel of its competitors, Dell’s direct sales path was more easily adaptable to the direct links between manufacturers and customers characteristic of Internet selling. In this way, Internet selling represented a logical extension of the disintermediation at the core of Dell’s direct business model. Unlike its competitors, Dell did not confront the problem of alienating channel partners through direct Internet sales, and it was not impeded by the legacy of the indirect selling system. As a consequence, Dell was one of the first large manufacturers to set up

Dell Computer 187 Multiple of 1994 amount

12



10

D

8

Sales Employees

6 4 2 0

1994

1995

1996

1997

1998

1999

2000

2001

figure 6.3 Growth of Dell Sales and Employment, 1994–2001 (1994 = 1.0) source: Dell Computer Corporation, Form 10K, 1996 and 2001.

a Web site in 1994. The following year, customers were able to configure systems and obtain price quotes from Dell.com. By June 1996, Dell was the first firm in the PC industry to sell systems over the Internet. Within three months, Dell had become one of the largest Internet commerce firms. By the end of 1996, when Amazon.com was selling $15 million worth of books per quarter, Dell was already selling PCs over the Internet at a rate of $1 million per day equal to roughly $90 million per quarter, or six times the sales volume of Amazon (Table 6.6). Six years later, Dell was the largest online retailer of goods, accounting for 22 percent of all Internet retail sales (Tedeschi 2002). In expanding its Internet sales as well as increasing its total sales, Dell admittedly benefited from a mutually reinforcing relationship between the Internet and the PC. The expansion of the Internet as a communications revolution created a corresponding surge in demand for PCs as the device of choice for access to the Web. While the entire PC industry prospered from this relationship, Dell as an on-line seller profited from this innovation in two ways. It was able to take advantage of the increasing interest for Internet access not only as a demand stimulant for its own PCs but also as a means of scaling its system of order intake and innovating its sales system to meet this surge in demand. Nevertheless, while this transition to on-line selling succeeded in lowering Dell’s administrative costs per order from roughly $50 to $5, this innovation represented a relatively small portion of the overall operation. Beyond order intake on the front end of the value chain lies a more complex set of

188 Dell Computer table 6.6 Daily Internet Sales of Dell Computer (End-of-Year Totals), 1996–2001

Daily sales ($ millions) % of total Internet sales

1996

1997

1998

1999

2000

2001

1 7%

3 14%

11 32%

20 40%

40 50%

50 57%

source: www.Dell.com, Speech Archive of Michael Dell.

operations and relationships between Dell and its suppliers. These relationships and their integration with the customer would become the focus for the second phase of Dell’s Internet innovation.11 the beginning of the “second web revolution” By extending the Internet from the point of sale into the logistics of procurement and assembly, Dell’s “second Web revolution” represents a more far-reaching set of transformations than the Internet innovations focused on sales. In this second phase, the company has become an innovator in the realm of business-to-business Internet commerce. Dell refers to this type of Internet activity, linking itself to its customers, suppliers, and logistics partners, as e-commerce. As a logical extension of Internet-driven efficiencies in the process of order intake and sales, the second phase of Internet deployment at Dell represents the continuation of a single overriding concern—diminishing inefficiencies in the value chain associated with time. This preoccupation with time, in turn, compelled Dell to focus on ways of using Internet communication to enhance what is perhaps the company’s most critical capability— balancing demand and supply in a build-to-order environment. Dell’s effort to rationalize the process of demand and supply balancing had as its operational objective the compression of time at every stage of the procurement and build process in which product was held in inventory and acceleration in the number of inventory turns or cycles. Strategically, the company’s aim was to bring demand and supply into balance within a much tighter time frame by using Internet-generated information and communication as a substitute for inventory. This aim of compressing time in the process of demand and supply balancing is what drove Dell’s Web experiment upstream into supply chain management and enterprise resource planning. Dell dubbed this initial effort at integrating the Internet into the procurement and build processes the Genesis Project. Launched in 1994, Genesis aimed to create a standardized system of data management for all aspects of the Dell operation, using the Web as an information and communications infrastructure. Procurement planning, pro-

Dell Computer 189

duction schedules, order intake, product delivery, and accounting, along with the logistics linking these activities, were to share this common information platform. The project, however, did have a clear and measurable performance goal: to reduce the inventory problems that had emerged at Dell during the previous four years. Central to the project was a Web-based information system for enterprise resource planning designed by software firm SAP. The SAP/R3 platform was intended to enable enterprise resource planners to identify how the impacts of an order ripple throughout the entire operation. Functionally, its aim was to create simulations for balancing procurement levels, coordinating production schedules, and adjusting financial balance sheets. Two problems emerged with this project (King 1997). First, the SAP system had difficulty accommodating Dell’s regional organizational structure. Adopted in 1995, Dell’s regional organization hinged on the idea of dividing the world into regions and establishing expansion sites for final assembly that would process orders and sell to customers in the region where the assembly site was located. The SAP/R3 system was unable to manage these multiple assembly facilities. Second, the project did not appear sufficiently scalable to handle the growth rate at the company (Koncaba, interview, July 27, 2001). Genesis had been conceived in 1994, when Dell was a $3 billion firm. By the time Dell actually abandoned the still-unfinished effort in 1997, its sales volumes had quadrupled to roughly $12 billion. Furthermore, Dell had already concluded that the SAP/R3 architecture, with its large centralized database, was incompatible with one of Dell’s newer aims. Dell was committed to running a more flexible and less rigid system of enterprise resource planning on its own servers. This aim was incompatible with the structure of the SAP/R3 system. According to Jerry Gregoire, chief information officer at Dell in 1997, “having a full SAP suite and all of these tightly integrated applications going into the company at the same time didn’t make as much business sense as it did before” (quoted in King 1997). Nevertheless, one critical achievement emerged from this early effort at integrating the Internet into procurement and production. From 1994 until cancellation of the project in 1997, Dell did succeed in accomplishing its primary goal of decreasing inventory. By 1997, Dell had once again assumed leadership on this metric within the PC industry (Table 6.7). The company was also growing faster than its competitors in the PC industry. By 1997 Dell had become the third-largest PC firm in the world, surpassing such firms as Apple and Hewlett-Packard (Table 6.8). Despite cancellation of the SAP project, Dell would continue experimenting with the Internet for enterprise resource planning over the course of the next two years. During this period, Dell (along with Cisco Systems) became known as the firm with the most advanced Internet-based system

190 Dell Computer table 6.7 Days’ Supply of Inventory at Dell and Compaq, 1994–2001

Dell Compaq

1994

1995

1996

1997

1998

1999

2000

2001

32 67

21 59

16 44

13 32

8 29

6 31

5 36

4 42

sources: Dell Computer Corporation, 10-K Reports various years; Compaq Computer, 10-K Reports various years; Robert Cihra, The PC Industry, ING Barings Furman Selz, 1998; “HP and Compaq: It’s Showtime,” Business Week, June 17, 2002, p. 77.

for supply chain management. Nevertheless, as Michael Dell admits, these early efforts at Web-based supply chain management were still very much a process of trial and error in trying to create Internet links between Dell and its suppliers (Dell and Fredman 1999, 190). Furthermore, the company confronted the problem of trying to standardize competing, and not always compatible, information and communications platforms in terms of phones, faxes, and the Web. By the end of 1999, Dell had decided to embark on a far more ambitious project for integrating the Internet into its operation.

Dell’s i2 Business Organization Dell embarked on the DSi2 project with the idea of building its business “with the Internet at its core” (vice president, interview, May 24, 2001). In a strategic sense, the aim of the project was to enable Dell’s demand and supply planning, parts procurement, build-to-order production schedules, customer order intake, and product delivery processes to operate on a single but flexible Internet-based information platform. As a practical matter, this aim involved transitioning aspects of these operations still being conducted through phone, fax and even e-mail on to the Web. In addition, this project was designed to get all of Dell’s geographical regions to use this common Web-based information and communications system. Perhaps most important, the project reflected an effort to link itself more systematically with its supply chain and logistics partners. Although Dell developed some of its own proprietary software for the program, it turned to the supply chain firm of i2 to provide the overall architecture for the Web-based information system. The DSi2 initiative utilizes Web-based communication to address what Dell considers the most formidable challenge in its build-to-order business model—the challenge of supply and demand balancing in an environment of fluctuating customer orders over time. Although some form of supply and demand balancing confronts all manufacturers, most companies address

Dell Computer 191 table 6.8 Computer Firms Ranked by World Market Share, 1997 Rank 1 2 3 4 5 6

Firm Compaq IBM Dell Hewlett-Packard Packard Bell / NEC Gateway

Units (thousands)

Market Share (%)

10,596 6,958 4,464 4,297 4,116 3,261

13.1 8.6 5.5 5.3 5.1 3.7

source: Gartner Group / Dataquest.

this problem through inventory management. If orders spike upward at any one time, firms rely on inventory to balance the fluctuation. Dell, however, has rejected this approach as incompatible with its build-to-order pull system. It relies instead on what is called “burst capacity,” or what in more common parlance is termed “speed-up” to deal with ever-fluctuating demand.12 This strategy utilizes a small amount of unused capacity within the procurement and production process as a safeguard against large increases in demand at any one moment. When such increases occur, this unused capacity is deployed. In order to work in a build-to-order environment, however, burst capacity has to extend beyond Dell into the supply chain. Consequently, DSi2 is effectively a merger of the Internet and burst capacity extended throughout the Dell organization. The project is built around Internet communication to link the planning and logistics of procurement, assembly, and distribution for balancing supply and demand in close to real time. It also incorporates burst capacity, deployed throughout Dell and its supply chain and logistics partners, to deal operationally with unexpected increases in demand. Through this combination of informational and operational innovation, DSi2 has enabled Dell to drive down inventory while balancing the seemingly incompatible goals of customization and high volumes.13 The DSi2 project has two principal elements: global supply planning and demand fulfillment. The first, as its name implies, focuses on planning processes, and the second is concerned with the execution and logistics of moving parts and semifinished and finished goods through the Dell procurement, production, and distribution organization. This separation of planning from execution is one of the biggest changes implemented by the DSi2 project (Cook, interview, June 14, 2001).

192 Dell Computer operations: global supply planning Global supply planning involves two types of planning processes. On the one hand, it involves the creation and management of data and information used to forecast general demand parameters for components. Second, and perhaps more significant, global supply planning is a system of information exchange between Dell and its component suppliers to coordinate the process of ordering and procurement of components. The innovative advance in this process hinges on the way Dell has deployed Web-driven communication, configured through the i2 modules as well as its own software, to create an automated procurement system in which Dell is functionally integrated with its parts suppliers and logistics providers. This integration with suppliers through the Internet has resulted in a more centralized and standardized process for procurement planning. Global supply planning centralizes Dell’s inbound planning activity in the company’s worldwide procurement organization in Austin. At the same time, however, the global supply planning process creates “a single system of record” for procurement at all six of Dell’s regionally decentralized operations (Koncaba, interview, June 12, 2001; Kelly, interview, May 4, 2001). This centralized information system for what is essentially a regionally decentralized production system is one of Dell’s planning innovations facilitated by the Internet. Despite the emphasis on differences between its business model and those based on forecasting, and disclaimers about its own forecasting abilities, Dell does employ a type of strategic forecasting in its global supply planning process. These forecasts of unit demand and component requirements are the first step in creating a system of material balances, which aims to equalize supply and demand of parts within Dell’s build-to-order environment and is central in the overall system of supply and demand balancing. Thus, contrary to perception, “forecasting is critical” in Dell’s just-in-time producing and selling system (vice president, interview, April 17, 2001). Although Dell does not build products on the basis of demand forecasts, as do competitors that sell through indirect channels, without forecasts there would be no viable foundation from which to launch the build-to-order direct-pull system. In addition, strategic forecasting is crucial because some of Dell’s primary suppliers maintain lead times of eight to twelve weeks for production and delivery of components. Forecasting these cycles enables the company to position material in its supply chain so that it can be procured and pulled into the build process as needed. Nevertheless, Dell does not orient its business around this element. “We are not a great company at forecasting because we do not build our business model around forecasting,” insists Lance St. Clair, director of supply chain and materials management for Dell. “We are a pio-

Dell Computer 193

neer in using the Internet for e-commerce with our customers and suppliers to balance demand and supply in the movement of components through the supply chain and production system” (St. Clair, interview, July 20, 2001; January 10, 2002). Consequently, the DSi2 global supply planning system begins with twelve-month “Master Production Plans.” These forecasts tell operations departments how much business to expect in the next twelve months (inbound supply chain manager, interview, April 19, 2001). These forecasts, however, are continually updated as market conditions change. Forecasts for Master Production Plans are generated from two sources: (1) historical data and (2) the purchasing plans of relationship customers that buy at least $1 million worth of equipment annually from Dell. For the Master Production Plan, historical trends are disaggregated according to three main categories: (1) products (desktops, notebooks, servers, etc.); (2) market regions (the Americas, Asia Pacific, Europe, etc.); and (3) customers (large-scale relationship customers serviced by salespersons or smaller transactional customers). Because Dell grew so rapidly from 1994 to 2000, however, historical trends in these categories do not always convey accurate estimates of future demand. Furthermore, the marked slowdown in demand for PCs starting in 2001 raised additional questions on the use of historical trends as future indicators. This potential shortfall in historical data makes the information from relationship customers equally if not more valuable for intermediate and longterm forcasts. Dell is able to obtain this information on its relationship partners’ upcoming purchases from the customized Web sites created by Dell for these large-volume buyers and from information secured by salespersons working with these large accounts. This information, not accessible from historical trends, enables Dell’s largest accounts to participate in a direct way in forecasting future demand for materials. Master Production Plans, created from historical data and purchasing plans of large accounts, are transformed into materials requirements. Through the i2 module known as “Supply Chain Planner,” these material requirements are further refined into a broad-based material requirements plan (MRP), which is one of the principal outputs of the global supply planning process. The material requirements plan divides a projected level of output into a matrix of different components, automated for Dell by the i2 “Tradematrix” module. In certain ways, this matrix resembles a transactions table in what is known in regional science as input/output analysis. It provides Dell with the material requirements by component type for a certain level of final demand. Suppliers collaborate in this forecasting process through a secure extranet

194 Dell Computer called Valuechain.dell.com. Through this portal, component producers are able to verify Dell’s material requirements for the duration of Master Production Plans. Suppliers then commit to these requirements and take responsibility to produce parts as needed over the twelve months of the Master Production Plan. In this way, parts producers confirm supply and delivery of Dell’s component demand. Through Internet communication, suppliers in effect provide information necessary for completion of the Master Production Plan, which is constantly in flux as conditions in the build-to-order environment change. With the material requirements of the Master Production Plan as strategic parameters, the global supply planning process assumes its innovative character by creating procurement cycles matched to the fluctuations in customer orders in real time. Through an on-line tracking system for components, Dell is able to respond to changes in its material requirements and alert its suppliers of its new requirements. This tracking system, configured through a combination of i2 tools, generates what is called an “exception action report.” Created automatically over the Web, this report alerts procurement planners at Dell of an exception in one or more parts necessary to fulfill demand in that moment. Where an exception exists, a purchase becomes necessary. Through the Supply Chain Planner module of i2, the parts procurement planner/buyer is able to set up a requisition, get it approved, convert the requisition to a purchase order, and send the order to the supplier. The supplier receives an on-line notification of the purchase order via Valuechain.dell.com. The supplier, in turn, is able to respond to the notification and commit to the new order requirement, which is then received by the procurement planner/buyer at Dell. The advantage of this innovation is that it transfers the ordering process for parts to what Dell calls an “exception basis” away from spreadsheets and onto the Internet (St. Clair, interview, May 24, 2001). Just before DSi2 was instituted, certain aspects of this process were executed over the Internet, while others were undertaken by more conventional methods such as faxes and phone calls. By mid-2001, however, roughly one year after initial implementation of the program, almost 90 percent of Dell’s purchases from suppliers were occurring through Web-based interactions (vice president, interview, June 20, 2001). As a result of the direct relationship with its customers, coupled with the advent of Internet communication, Dell has created a new type of connection between the process of customer order intake and the process of components procurement from its suppliers. Forged on the basis of a new communications infrastructure, this link represents Dell’s ongoing search for ever-greater levels of balance between demand and supply in a build-to-order demand pull environment.

Dell Computer 195

operations: demand fulfillment Demand fulfillment in Dell’s Internet direct business model refers to how supplies are delivered to Dell’s production sites for assembly into finished products. The primary element of demand fulfillment consists of “pulling material to order” every two hours into Dell factories. Customer orders released to the factory floor for configuration into finished PC systems provide the input that initiates these material pulls. Dell schedules these orders to be built in two-hour cycles and pulls only those components required to fulfill orders for the given two-hour period. According to Lance St. Clair, demand fulfillment, with its emphasis on pulling material to order, is “the rocket science of the Dell supply chain system” (St. Clair, interview, May 24, 2001; January 10, 2002). A critical intermediate step, however, precedes this process of pulling materials into Dell’s assembly factories. This step consists of storing and staging components in sufficient quantities so that they can be pulled into the assembly process on a just-in-time basis. This staging process focuses on planning and executing the movement of components between two of the primary nodes in Dell’s network: (1) supplier factories located long distances from Dell assembly sites, and (2) supply logistics centers (SLCs) located in each of the six regional locales where Dell operates assembly plants. This staging activity is essentially a system for collapsing distance between the location of component supplies and the location of the assembly process. Such collapse of distance, in turn, plays a crucial role in enabling Dell to manage the compressed time cycles for pulling material into production on a just-in-time basis. In this way, physical proximity between supplies and assembly activity and time compression in the just-in-time pull system are mutually reinforcing. The reason this staging process occupies a position of such centrality in the Dell network stems from geography. Long distances separate the locations where components are produced and Dell’s assembly sites. In Dell’s just-in-time build-to-order environment, there is simply no way of eliminating the friction of geography when distance separates sources of supply and the activity of assembling supplies into finished products. In effect, Dell’s storing and staging of supplies is a response to the problem of risk. The greatest risk Dell faces in its business system is not only access to component supplies. The risk confronting Dell is access to supplies within specific parameters of time consistent with its build-to-order pull system. What Dell seeks through the staging and storing process is control over the risks it encounters in securing access to sources of supply in a just-intime build-to-order environment. The firm, in effect, reorganizes space as a substitute for inventory and a remedy against the risk of timely access to

196 Dell Computer component supplies. It creates territorial proximity between itself and its suppliers as a source of certainty and speed. Although numerous primary- and second-tier suppliers have established factories where Dell operates its assembly plants (see “Organization” and “Territorialization” below), usually at least 50 percent of the components at each assembly site—sometimes virtually all of them at sites such as Ireland or Brazil—are supplied from factories located great distances from Dell’s assembly centers. Dell requires suppliers that do not operate factories where Dell assembles finished products to operate through SLCs located adjacent to Dell’s assembly sites. In these SLCs, component vendors store parts as inventory. In general, Dell requires parts suppliers to maintain ten working days, or two weeks, of inventory at SLCs (Cook, interview, June 14, 2001). Consequently, there is inventory in the procurement, production, and distribution network, although compared to that of other PC makers, this level of inventory is relatively small. As explained more fully below, however, it is component suppliers that bear these inventory costs. The operational costs of the SLCs are assumed by suppliers and thirdparty logistics providers (3PLs) as part of an emerging trend in supply chain management known as vendor-managed inventory, or VMI. Dell is one of the pioneering firms in this area. Third-party logistics providers involved in VMI for Dell include BAX, Menlo Logistics, Ryder, IEC, and Eagle Global Logistics. Suppliers negotiate contracts for operation and management of supply logistics centers directly with 3PLs and pay what is called “pallet in/ out charges” to 3PLs for storage of component inventory (Cook, interview, June 12, 2001). While Dell closely monitors the ability of 3PL-managed supply logistics centers to provide the required services to Dell factories, the operation and management of SLCs is conducted independently of Dell (Kelly, interview, May 4, 2001). Nevertheless, this relationship between Dell and the SLCs is far from what would qualify as a market transaction between independent agents. In Austin, for example, the SLC built by Dell and leased to Eagle Global Logistics is directly adjacent to the Morton Topfer Manufacturing Center where Dell assembles finished goods, and is literally on the Dell property. Functionally, the SLC is completely integrated into the Dell assembly operation. Such an arrangement represents Dell’s strategy for simultaneously contracting out and passing certain costs on to other parties while retaining necessary control over an essential element in its operation. The staging process for the storage of inventory at SLCs begins when suppliers commit to the procurement order from Dell through the ValueChain portal. In making such commitments, suppliers take responsibility to position and store components in the supply logistics centers if they do not have a factory adjacent to the Dell assembly site.14 As part of their man-

Dell Computer 197

agement contracts with Dell, supply logistics centers are obligated to operate on the same Web-based i2 communications platform that Dell uses for global supply planning and demand fulfillment. In this way, SLCs are part of a three-way planning and fulfillment conversation taking place with Dell and suppliers on the same information infrastructure. Through ValueChain and the i2 tools, SLCs transmit information to Dell factories every hour on the ever-changing inventory of components stored in them and commit to material requests from Dell. At the same time, SLCs communicate with supplier factories on components needed to replenish inventories. Although the general principle of this replenishment system at SLCs is supply and demand balancing in real time, the system nevertheless functions on the basis of two weeks’ supply of inventory at all times. As orders from customers are received from the Dell.com Web site or by phone, and as they are queued and cleared by the credit department, they are downloaded to factories every twenty seconds and simultaneously transformed into a matrix of material requirements (vice president, interview, May 24, 2001). From this matrix of required components, the “Factory Planner” module of i2 creates a two-hour production schedule that organizes the “kitting” of parts for assembly as per order and a build schedule for orders in the two-hour window. At the same time, the schedule from the Factory Planner is converted to a different output within the i2 suite called “Rhythm Collaboration Planner,” which sends an automated feed to the SLCs on the components needed in the Dell assembly site for the upcoming two-hour build schedule. Third-party logistics providers managing the SLCs are given 1.5 hours to deliver the required parts to the dock at Dell’s factory. It is because of this 1.5-hour window that proximity between Dell and supply logistics centers is so critical. Dell then unloads the parts delivered to the dock on the assembly floor in thirty minutes. For the next two hours, assembly workers configure components into finished machines. These assembly workers generally put together a finished system from a kit of component parts in ninety seconds. The work on these highly automated assembly lines is extremely hectic and fastpaced. The entire cycle from order clearance to finished product is four hours. Only when material is needed within a two-hour period is it actually pulled into the Dell assembly site and configured into a finished system by Dell assembly workers. By pulling material to order over a Web-based communications infrastructure, Dell’s demand fulfillment system represents an innovation with an enormous operations cost benefit. First, because the dominant cost driver for Dell is the cost of goods procured externally from component producers, any advantage Dell can secure in terms of procurement costs will greatly affect its margins.15 Dell pays its suppliers for these components after they en-

198 Dell Computer ter Dell assembly plants from the SLCs. In the time between Dell’s procurement order and final delivery of parts following the staging of components in SLCs, suppliers assume the costs of falling component prices. Consequently, in an environment where the component costs are the primary drivers of total costs, and where the value of these components is falling over time at a rate of roughly 1 percent per week, Dell has an interest in securing these components at the last possible moment before they are assembled into finished machines. The SLC system enables Dell to accomplish this aim. When Dell assembles these components and ships the finished systems directly to the customer, generally within five to seven days from date of order, these components have a limited window of time in which to drop in value. The SLC system therefore provides a buffer to variability in demand and supply. It also shields Dell from the loss of value associated with PCs as perishable goods. These arrangements between Dell and its suppliers for producing PCs in such compressed time cycles, however, are not without certain contradictions and conflict. While Dell, according to one large supplier, operates what is probably the most efficient supply chain and inventory system in the industry (supplier 1, interview, May 24, 2002), Dell’s mission with its suppliers—ensuring continuity of supply—undercuts the primary aim in supply chain management, which is the reduction of inventory throughout the entire procurement and production process.16 “They [Dell] are much more focused on making sure they have continuity of supply than the other PC customers we serve,” insists another of Dell’s largest vendors (supplier 2, interview, July 2, 2002). This hedge against risk imposes certain limits on how far Dell is able to compress inventory levels in its build-to-order system. “It just isn’t possible to make this a totally efficient world,” concedes still another of Dell’s component providers. “The way that Dell balances this contradiction is by passing the inventory problem to suppliers with very clear expectations on the levels of inventory required in SLCs,” and when inventory passes to the control and ownership of Dell (supplier 3, interview, June 24, 2002). How Dell has been able to assume this power in its relationships with suppliers and create an extremely innovative and efficient procurement, production, and distribution organization is the story of how the PC maker has fashioned a new type of business organization. organization: the virtually integrated firm When it began implementation of its global supply planning and demand fulfillment system, Dell faced a difficult problem with its component vendors. For its DSi2 innovation to function, it had to ensure that its suppliers adopt the same Web-based planning and execution system being developed by Dell. This requirement meant that firms aiming to supply Dell had to op-

Dell Computer 199

erate on the same communications platform as Dell and develop the same interactive capabilities on the Web and the same standards of interoperability with the PC maker (Radjou, interview, August 6, 2001). Parts vendors and logistics providers therefore had to make investments in their own information systems compatible with Dell’s i2 system as a precondition to supplying and interacting with Dell. In speaking of these requirements on suppliers to make technology investments compatible with the systems at Dell, CEO Michael Dell is candid about the role and virtues of market power. “Dell in the U.S. is 50 percent larger than its nearest competitor and growing four times as fast,” he says. “Suppliers have a choice: Supply Dell, or lose market share. Let’s face reality. If my largest customer had a new requirement, I’d listen to them” (quoted in Perman, 2001). Such investment, however, was far more than a technological imperative. Convincing suppliers to operate on the same information and communications infrastructure as Dell was a first step in creating a new type of business organization. Dell refers to the form of enterprise it has created from these technical and organizational imperatives as the virtually integrated firm (Magretta 1998). Dell’s aim in establishing such an organization was to move away from market-oriented transactions for procurement of parts and logistics services. CEO Dell describes this phenomenon as one of replacing “the traditional ‘bid-buy’ cycle” (Dell and Fredman 1999, 180). What has emerged in its place is an organization built on nonmarket foundations. In its high-speed, Internet-driven procurement, production, and distribution network, Dell seeks to gain leverage over risk in replacing market-oriented bids for parts contracts with transactions undertaken through controlled relationships.17 Just as the mitigation of risk in component procurement is the aim of Dell’s ten-day replenishment system of supply at SLCs, so too is the mitigation of supply disruptions in procurement the motivation for Dell to pursue relational as opposed to market-based transactions with its parts vendors. Dell, in effect, has concluded that markets pose more risk of supply disruptions than relationship-based contracting. Virtual integration with a governance structure based upon nonmarket forms of administrative coordination is Dell’s response to such risk. In this sense, Dell’s reliance on administered relationships to link the adjacent operations in its network is similar to the dependence of vertically integrated firms in the late nineteenth century on administrative controls to organize their procurement and production systems. In much the same way that vertical integration coupled with administrative planning represented a response to the risk of managing complex procurement, production, and distribution systems without disruption, so too does the virtually integrated enterprise of Dell confront similar types of risk by deploying similar control mechanisms. The difference is that whereas manufacturing firms in the early

200 Dell Computer mass-production age tended to exert such control through mechanisms of administrative planning in concert with ownership of assets Dell exerts control through mechanisms of administrative planning in combination with assets owned by different firms. Although the asset structures of the two types of organizations are different—vertically integrated firms invariably owned the assets in their network, while Dell does not—the forms of control through administrative planning and the rejection of market-based interactions to accomplish operational objectives are fundamentally similar. Thus the idea that market forces are emerging as the mechanism of governance in interfirm production networks is far removed from the experience of the virtually integrated interfirm network coordinated by Dell. While the proliferation of the interfirm network as an organizational phenomenon is currently undeniable, the virtually integrated enterprise of Dell tells a far more revealing story about interfirm networking than the focus on its structural characteristics per se. The story of virtual integration pioneered by Dell is instead one of how power is exercised within networks of firms and how the exercise of such power mobilizes resources within the network for innovation and profit. Not only are relationships of administrative control compatible with interfirm networks. Such nonmarket relationships, in the case of Dell, are essential in enabling it to mobilize resources from both outside and within the enterprise in the search for efficiency. In using the Internet as the basis for these relationships, Dell has extended the idea of the closed loop with its customers, creating what in effect is an Internet-based closed loop with its suppliers. In this closed loop, transactions conducted over the Internet occur not through arm’s-length market-oriented interactions. Instead, they take place on the basis of coordinated relationships in which Dell plays the pivotal role. The relationship imperatives of virtual integration have had profound consequences on the supply base itself. First, virtual integration, requiring such close collaboration with component producers, has forced Dell to consolidate its supply base, a trend that actually started in 1994. As a consequence, Dell, which at one time worked with over two hundred first-tier suppliers, now has a core group of thirty to thirty-five primary suppliers. This small group provides close to 80 percent of the “material spend” for components used in Dell’s products (vice president, interview, April 17, 2001). An additional twenty suppliers bring the total to roughly 95 percent of Dell’s raw material costs. This relatively small supply base made it easier for Dell to coordinate vendors in its network around the same Web-based i2 systems for global supply planning and demand fulfillment platform. In this way, the Internet served a technical function of communication and control as part of an organizational mission structured around the idea of construct-

Dell Computer 201

ing controlled relationships as the basis for fulfilling the requirements of a high-speed just-in-time direct-pull system. The relationship between Dell and its suppliers and logistics partners and Internet technology play complementary roles in the establishment of this virtually integrated organization. Dell has made it clear to its component vendors that if they want to supply Dell they must submit to its technical requirements structured around the idea of virtual integration. Component vendors are candid in the way this system of relationship building reflects the relations of power in the Dell network. “Dell has a significant amount of power with its suppliers based on the current and future business levels they offer,” argues one of Dell’s large suppliers. “They know it and they use it” (supplier 2, interview, July 2, 2002). Nevertheless, this supplier is quick to add that the pressure from Dell on continuity of supply enables it to push its own suppliers for greater speed in the delivery of materials. Furthermore, Dell gives a clear accounting of what it expects from vendors. “One aspect that I like about Dell,” says this supplier, “there is clear delineation on control and ownership [of inventory].” As a result, this vendor is able to supply Dell with greater velocity than it does other PC customers. One important way Dell has used its leverage for relational rather than market interactions occurred in the aftermath of the Taiwan earthquake of 1997. As a result of supply disruptions with its Taiwanese vendors following this event, Dell initiated what it called the Supply Assurance Program. Its aim was to ensure that Dell would never be compromised by supply bottlenecks or shortages. Suppliers were obliged to endorse this program through agreements with Dell. “Dell dragged us over the coals to make certain that we knew about the risks in our own supply chain,” admits one of Dell’s suppliers, “but they did not impose upon us a set of best practices to correct it.” Instead, this supplier worked with the vendors in its own supply chain to become more efficient. The outcome, however, was somewhat contradictory, reflecting the sometimes divergent interests of Dell and its vendors. “As a result,” concedes this supplier, “we actually increased our available inventory so that Dell’s own supply chain system will not suffer even in an event as catastrophic as the Taiwan earthquake” (supplier 3, interview, July 24, 2002). Another example of how Dell uses its financial and market power relative to its suppliers is the way it pays these firms (information in this paragraph taken from Thurm, December 27, 2002). In the year 2000, Dell was paying its bills to vendors on average in fifty-eight days. By 2002, in the wake of the high-technology slowdown, Dell was forcing suppliers to wait an average of seventy-one days for payment. In order to secure these payment terms from suppliers, “Dell used its clout as the biggest maker of PCs to win the extended credit.” Not all of Dell’s suppliers, however, agreed to such terms.

202 Dell Computer SMTC Corporation, a contract manufacturer whose biggest customer was Dell, terminated its relationship with the PC maker because, as it explained in an SEC filing, its Dell contracts “were not generating sufficient returns.” Perhaps the most obvious way Dell has been able to prevail upon its supply base is its requirement that component producers stock SLCs with a tenday supply of inventory. According to a major supplier, “reduction in inventory levels at SLCs is the single biggest issue facing suppliers.” This firm insists that with the exception of Intel and Microsoft “Dell’s suppliers are bleeding.” For this reason, components producers have been trying to negotiate with Dell to reduce the ten-day requirement to five days. “Dell has resisted,” explains this supplier, “because they do not want the risk” (supplier 1, interview, April 26, 2002). In this regard, control through power is the route to risk avoidance, much like the vertically integrated firm. For Dell, nonmarket relationships with other firms crafted on the basis of power and control are the foundation of the virtually integrated network enterprise. “As a market leader in every aspect of the PC supply chain, Dell has the bargaining power to structure these relationships in order to sustain its build to order model,” insists Nigel Johnson of the Eclipse Group, a supply chain management consultant in Silicon Valley (Johnson, interview, April 24, 2001). This form of organization seeks other firms to operate functionally as units of Dell. In this way, virtual integration aims to capture the benefits of vertical integration—control over the adjacent operations required to build, market, and support a product—but without the asset requirements and expertise needed for such a comprehensive approach to competing. “We want to take advantage of the benefits deriving from vertical integration,” admits Dell’s vice president, “but the problem with vertical integration is that we have to become experts at many different product and process technologies. We want those advantages at Dell, but we want them virtually, through Internet communication” (vice president, interview, May 24, 2001). This form of organization and the structures of control that enable it to function in an innovative way have had profound impacts on the geography of the Dell enterprise. territorialization: geography and the dell organization Business organizations create economic territory in the way they coordinate and manage the assets and activities within the networks in which they procure, produce, and sell. Dell has fashioned an organization with a geography emerging from three basic sources: (1) the locations of the principal network nodes where goods are created and produced, and the territorial configuration of these physical assets; (2) the circulation of supplies, semifinished and finished goods, and information between principal nodes, and

Dell Computer 203

the territorial routes created by these flows; and (3) the power relations between Dell and other firms in its network that shape the locations of assets and how these assets operate and manage flows of product in the spaces where they are distributed. The innovations in global supply planning and demand fulfillment have enabled Dell to reconfigure the geography of its interfirm enterprise in three fundamental ways. First, Internet communication has provided Dell with a more centralized form of control, not only over global supply planning but also over the execution of fulfillment activities across its regionally decentralized organization. This more centralized form of control has resulted in a more standardized set of fulfillment systems at each of the locations where Dell concentrates its operations. The company has created what it calls a “copyexact approach” of the pull-material-to-order system in each of its regional locations (Cook, interview, June 14, 2001). This “copy-exact” approach standardizes the logistics of the Dell fulfillment system across space. Second, far from diminishing barriers of distance in procurement and production, the Internet-driven planning and execution systems have intensified the need for certain relationships of spatial proximity between key nodes in Dell’s network. Thus, while Internet communication has provided Dell with an infrastructure for managing the logistics of planning and execution in its globally spread procurement, production, and distribution system, the velocity requirements and the supply and demand balancing requirements of its Web-driven business model have placed an even greater premium on proximity between Dell and its suppliers and logistics partners. Such relationships enable the PC maker to manage the flows of materials in its just-in-time, pull material-to-order business system. Third, Dell has assumed an active role in shaping the geography of its organization by influencing the location decisions of its network partners. The PC maker has prevailed upon firms in its network to establish operations— either factories or, more commonly, the replenishment operations of SLCs—adjacent to the locations where it has chosen to organize its just-intime custom-build assembly activity. By controlling the locations of its network partners and by configuring relationships of proximity in these locales, Dell has actually shaped its own place-based external economies of scale. As it creates these external economies and reshapes local landscapes where the key nodes in its network operate, Dell is helping to fashion the industrial districts of contemporary globalization in which geographical spread and spatial concentration coexist (Storper 1997).

map 6.2 The Expansion of Dell Assembly Sites, 1991–99

Penang (1996)

Xiamen (1998)

Austin Nashville (2000)

Eldorado do Sul (1999)

Limerick (1991)

Assembly site (Year established)

© Jeremy Shaw

Dell Computer 205

The Geography of Assembly The locations of its assembly facilities are the primary drivers of the global and regional geography in Dell’s production network. The six assembly plants in the network are spread across the four continents of North America, South America, Europe, and Asia. Generally, each facility configures products for customers in a specific geographical area. These areas approximate Dell’s division of the world into four market regions: (1) the Americas (both North and South); (2) Europe, Africa, and the Middle East; (3) China; and (4) Asia-Pacific (see Map 6.2). Assembly sites in Austin and Nashville service customers in North America. Eldorado do Sul in Brazil is responsible for South American customers. The facility in Limerick, Ireland, builds PCs for Europe, Africa, and the Middle East. Xiamen, located directly across the strait from Taiwan, services China, while the Asia-Pacific assembly center in Penang, Malaysia, services the rest of Asia. Customer orders are automatically routed over the Web to the configuration center that services the region where the customer is located. In this way, Dell’s assembly sites are global in their geographical reach and regionally decentralized. The locations Dell has chosen for expanding assembly operations beyond its origins in Austin reveal a pattern with several common themes. The most important characteristic of the expansion sites in Limerick, Penang, Xiamen, and Eldorado do Sul is that they possessed existing concentrations of computer-related firms resulting in large part from government policy. Penang, Malaysia, is perhaps the best example of this phenomenon. Penang was the beneficiary of government policy in the late 1960s to attract foreign transnational corporations to Malaysia through a program of high technology targeting supported by the creation of “Free Trade Zones.” By the early 1970s, Penang emerged as a top electronics-producing region in Malaysia. Intel, AMD, Motorola, National Semiconductor, Siemens, and Hewlett-Packard established operations in Penang’s Bayan Lepas Free Trade Zone during this period that created employment for twelve thousand workers in the city-region in 1971–72 (Kahaner 1996). During the 1980s, a second wave of investment, this time including Japanese and Taiwanese firms, made Penang one of the most important centers of the electronics industry in Asia (Rasiah 2000). By 1992, Penang had over seventy-six thousand electronics workers with semiconductors and disk drives, the most important industry sectors (Kahaner 1996; Haggard, Li, and Ong 1998). When Dell began operations in Penang in 1995–96, the city-region had close to 120,000 employees in 148 electronics factories with roughly 32,000 in the hard disk drive sector (Kahaner 1996; Penang Development Corporation; Haggard, Li, and Ong 1998). Dell was now part of a computer-related high-technology cluster of companies, many of which

206 Dell Computer were major suppliers to Dell, such as Acer, Iomega, Komag, Seagate, Quantum, and Sony. This pattern, similar in Dell’s other chosen locations, reveals the importance of preexisting concentrations of electronics firms, including suppliers, in influencing Dell’s location decision. In this sense, the locations selected by Dell conform broadly to the preferences for proximity to other firms in the same industry. First observed systematically by Alfred Marshall, these agglomeration economies include access to specialized suppliers, a “constant market for skill,” and an environment enabling the spillover of technical knowledge (Marshall 1890, 267– 77). Such external economies of scale are now commonly attributed as drivers of high-technology concentrations. Yet, while Dell has taken advantage of existing agglomeration economies in these locations, it has also contributed in recasting these places as hightechnology concentrations by its own location decisions. In this sense, Dell’s location pattern exhibits what has been described as cumulative causation, or similarly, historical path dependence.18 In Limerick, for example, Dell began operations in 1992 with 184 employees. At the beginning of 2001, however, Dell employed five thousand workers at its Limerick complex, making it one of the largest high-technology employers in the Limerick region.19 Through such location choices, Dell has not only reinforced the high-technology character of these places but also succeeded in configuring a new set of activities upon the local landscape and, in the process, changing the nature of the places themselves. A second characteristic of Dell’s location pattern focuses on direct incentives. While Dell receives certain indirect benefits from the external economies where it has established its assembly locations, it has successfully secured a variety of tangible financial incentives from government in each of its assembly locations outside the United States. According to Dell and local development officials, such incentives have played a role in influencing its location decisions (Robertson, interview June 27, 2001; Wong, interview, May 24, 2001; Tobin, interview, February 6, 2001). In the case of Penang and Xiamen, Dell operates in government-created incentive zones, which provide the PC maker with exemptions from import and export taxes along with other material benefits.20 In the case of Limerick and Eldorado, Dell has similarly managed to secure incentives from local, regional, and even national authorities. Nevertheless, as is generally the case with such programs, it is inconclusive whether direct incentives actually influence location decisions or whether companies make such decisions on the basis of other factors and then simply profit from government largesse. Another characteristic of these locations is that while they all represent formidable concentrations of electronics firms, within this category they all

Dell Computer 207

share features as the world’s “second tier” of high-technology cities (Markusen, Lee, and DiGiovanna 1999). With the possible exception of Penang, these cities are one level below the concentrations of high technology found in places such as Silicon Valley, Boston, Singapore, or the area around Taipei. Dell is part of the growth trend that is extending electronics development to selected places outside of these first-tier electronics concentrations in creating newer high-technology industrial districts. Finally, Dell has shown a preference for locations near its most critical material supplier, Intel. The world’s largest semiconductor firm operates a production, testing, and supply center in Penang, roughly five minutes’ drive from Dell’s facility. Intel also operates similar facilities in Ireland and China, with relatively easy access to Dell’s configuration centers there. This pattern conforms to Dell’s preoccupation with avoiding supply disruptions and using proximity as a remedy for such risk. In effect, Dell has taken advantage of existing high-technology concentrations in making location decisions for its assembly operations. Its locations also reveal an effort to maintain physical proximity to Intel. In following these trends, Dell has emerged as an actor in reshaping the agglomeration economies where it has located into the world’s new generation of hightechnology industrial districts. Dell’s role in reshaping the places where it assembles PCs is more readily apparent in the way it participates in the geography of supply. The Geography of Supply In addition to assembly sites, Dell’s organization also assumes its territorial configuration of spread and concentration from a geography of supply. This geography derives from two sources: historically conditioned but constantly evolving location choices of supplier firms in siting their factories for component production and the decision-making power exercised by Dell to influence the locations of supplier facilities in order to optimize advantages of physical proximity in its build-to-order production system. The interplay of these two sources producing this geography is one of structure and agency in which Dell is the agent reshaping a structure to accommodate its operational objectives. Supplier factories creating this structure are spread throughout the world but have distinct geographical concentrations. East Asia, Mexico, and the United States are the largest concentrations where components are produced.21 Suppliers from these areas are themselves frequently part of dense subcontracting relationships in which the lead supplier firm and the firm (or firms) actually producing the components or subassemblies of the components are different entities based in different countries. This phenomenon of

208 Dell Computer cross-border production networks is common across a range of different PC supplies, from highly standardized components such as disk drives to semiconductors in which chipmakers in Silicon Valley typically subcontract fabrication and assembly to chip foundries in Taiwan, which in turn subcontract portions of this work to other specialty firms in different parts of Asia or increasingly China or even to their own subsidiaries in these areas. These networks of ever-shifting contracting relationships reflect the influence of two major trends. On the one hand, this phenomenon of subcontracting in which work is dispersed to other locations reveals the impact of what is called the “product cycle.” Originally developed by economist Simon Kuznets (1930) to explain business cycles, product cycle theory insists that the life of products passes through stages. This focus on stages was later given a more geographical orientation by Raymond Vernon (1960) and Ann Markusen (1985), who argued that stages in the life of products correspond to shifts in the location where such goods are produced. As products mature and become more standardized, production requirements for them become more routine. In this process of maturation and standardization, the production skills of firms from areas formerly unable to produce such goods eventually match the more easily mastered production requirements of goods in the more mature phase. What differentiates these newer entrants, however, is their low-wage and low-cost structure. As a consequence, firms originally making such products subcontract production to these newer firms outside the location of origin to take advantage of their low costs.22 Thus, from product cycle theory, the geographical spread of firms producing PC components in subcontracting relationships reflects an evolutionary stage of development in the PC. As the components of the PC assume standardized formats, production of these goods is able to spread to lower-wage producers. On the other hand, however, there is a second major variant of this story about contracting that supplements the emphasis in product cycle theory on low wages and geographical spread. This second approach shifts the focus from costs to capabilities. In this approach, cross-border networks of companies aim to exploit a diverse array of technological knowledge that is spreading outside the United States. At the same time, these networks are taking advantage of increasingly specialized skills and technical knowledge that are concentrating within and among particular firms located in specific geographical localities (Borrus, Ernst, and Haggard 2000, 2; Saxenian 1999). From this perspective, the geography of supply associated with subcontracting relationships is marked by the spread of capabilities to a new generation of firms that concentrate into new skill-based and knowledge-based regions. The process is one of creating new high-technology industrial districts.

Dell Computer 209

In effect, both product cycle theory and cross-border production network theory provide essential pieces of the story driving the pattern of spread and concentration in the geography of supply. One piece focuses on the search by suppliers for low costs; the other focuses on the search for skill. The geography of Dell’s supplier network is part of this dual phenomenon. This search by suppliers for costs and skills has given Dell’s supplier network its structural focus in East Asia, Mexico, and the United States. Within East Asia, firms in Taiwan, Singapore, Malaysia, Korea, Japan, and China supply most of Dell’s components. The companies producing components for Dell in these places are Asian firms such as the Taiwanese firms of Compal, Quanta, and Acer and the South Korean firm of Samsung. Asian firms producing for Dell also include such companies as Taiwan Semiconductor Manufacturing Company (TSMC), which is the primary foundry for Dell’s graphic chip supplier, NVIDIA. Dell’s suppliers in this region are also the East Asian subsidiaries of mostly American but also Japanese, European, and now increasingly Taiwanese-based transnational electronics producers. Because of the standardization and modularity of PC components, all of these companies are resorting to outsourcing PC component production to low-wage areas. In this process of migration to lower-cost areas, Malaysia and especially China are emerging as the locations of choice for fabrication of PC components. In addition to the East Asia region, the border area of Mexico is also an important center where production of components takes place. Invariably the same firms outsourcing production from subsidiaries in Asia are also sourcing production for PC components from their own border operations in Mexico. The Japanese firm Sony, for example, makes flat panel displays for Dell at its Malaysian subsidiary while also supplying Dell from a plant in Mexico. Samsung of Korea supplies Dell’s LCD monitors by outsourcing from the same locations. The United States is notable as the primary location for the two firms producing the two most critical components of the PC, the operating software supplied by Microsoft and the microprocessor and related semiconductor components supplied by Intel. Microsoft is in many ways a special case, not only because of the dominant control it exerts over the operating software but also because it is not really a material supplier. What it supplies Dell is intellectual rather than material property. Intel, in this sense, is far different, since the intellectual property it produces is embedded in a material product. Most design, development, and wafer production of Intel chips occurs at facilities in the United States. A significant amount of this same work, however, occurs outside the United States, in Malaysia, Israel, and Ireland. Low-wage assembly and testing is often subcontracted by Intel to firms such as ASE of Taiwan, which in turn sources some of its work from facili-

210 Dell Computer ties in Penang, Malaysia, and, increasingly, Shenzhen, China. In this way, chip production is a prime example of networklike subcontracting. Alongside this structural pattern in the geography of Dell’s supply base, however, is the role Dell has played in shaping the location behavior of its component vendors to accommodate its requirements for proximity to sources of supply for its build-to-order system. In much the same way that Dell obligated its vendors to operate on the same Web-driven i2 information platform, so too has the PC maker elevated “location” as a condition for entry into its virtual organization. This process has taken two forms. On the one hand, Dell has compelled suppliers to operate through the SLC system (see above), which is the approach it generally takes to address the need for proximity in its business model. Nevertheless, Dell has supplemented this approach by convincing certain suppliers to locate factory operations in the locations of its assembly plants. In pursuing this strategy, the company approached a number of its primary suppliers and requested that they service Dell in each of its regional assembly locations.23 In Penang, for example, Dell prevailed upon two of its contract manufacturers of motherboards, SCI and Jabil Circuit, to set up factory operations close to the Dell assembly facility (Wong, interview, October 2, 2001). While it may be that suppliers have good reason to follow Dell in these locations to benefit from the same economies of proximity, it is also true, as Michael Dell reveals, that suppliers and Dell negotiate about such location decisions. In this way, structure and agency—the objective forces pulling suppliers toward Dell and the actions taken by Dell to encourage such movement—are factors in accounting for suppliers’ decisions to locate near Dell. Consequently, whether forcing suppliers to operate through SLCs or compelling vendors to establish factory sites near Dell assembly centers, the motivation is the same: to use proximity as a response to the logistical challenges of material balancing in its just-in-time pull system of production and as a remedy for the risks of supply disruptions. Dell’s assembly sites and these supply sites, both factories and SLCs, are thus linked organizationally, technologically, and spatially. On the basis of these linkages, Dell has actually created more heavily concentrated bases of supply where it assembles PCs.24 Where Dell is surrounding itself with suppliers and supply hubs and organizing specific operational and spatial relationships between these entities and its assembly sites, it is playing a role in transforming the economic geography of those places. The PC maker is creating a set of fundamentally similar production complexes across space. “All of these factory complexes are set up much the same way,” says a Dell senior manager of global supply chain strategies. “We want consistency across the globe throughout the Dell system.” As it standardizes the logistics of the assembly system in each of

Dell Computer 211

these places, the copy-exact approach used by Dell for its pull-material-toorder model creates roughly equivalent spatial arrangements between supplier facilities and assembly complexes. According to Daryl Robertson, vice president and general manager of Dell Latin America, “We execute the same business model everywhere. It’s like McDonald’s. While there is some local customization [of production systems], we want to offer the same basic menu of products and services to our customers worldwide” (Robertson, interviews, June 27, 2001, and March 12, 2002). In employing this “McDonald’s-type” approach, Dell is creating a factory system that is reinforcing the idea of globalization as an essentially homogenizing force (Map 6.3). It is creating these uniform places as part of a pattern of spread and concentration. As the PC maker selects locations around the globe for its assembly operations, and as it organizes relationships of proximity between assembly and sources of supply in each of these selected locales, Dell has assumed the role of agent in crafting the territorial features of the contemporary regional economic world.25

The Diffusion of Dell’s Business Organization The efficiencies of Dell’s Internet-driven business organization have driven other firms in the PC industry into a pattern of imitation that Schumpeter characterized as the second element in the innovation process, the element of adaptive response (see Chapter 2). This adaptive response of firms to the business models of more innovative companies is what completes the transformation of entire industries and entire economies and is part of what Schumpeter described as “creative destruction.” Together the creative response of innovative firms and adaptive imitation by competitors drive the development of economies. There are few better contemporary examples of this innovation and diffusion process than the impact of Dell Computer on the personal computer industry. final ascent The motivation for this process of imitation is derived from Dell’s ascent to the top rank of the PC industry. In this position, Dell’s Internet-driven business system provided competitors with a model to emulate. From 1996 to 2000, Dell’s growth rate in units shipped was as much as four times the industry average and far exceeded Compaq, its principal competitor (Table 6.9). As the largest firm in the industry during this period, Compaq had good reason to be especially wary of Dell, although all PC firms were compelled to respond to Dell’s efficiency advantages (Kirkpatrick 1997a, 1997b; Aragon 1998). As a $2.9 billion company in 1994, Dell was a respectable competitor. As Dell began its seemingly irrepressible rise from 1994 to 2001,

KOREA JAPAN

MEXICO

Austin

UNITED S TAT E S

map 6.3 The Dell Business Enterprise, 2001

SINGAPORE

Penang

TA I WA N

MALAYSIA

Xiamen

CHINA

Eldorado do Sul

BRAZIL

Silicon Valley

Seattle

Nashville

UNITED S TAT E S

HUNGARY

CZECH REPUBLIC

Limerick

IRELAND

Assembly site Supply logistics center Localized supplier Global supplier

© Jeremy Shaw

Dell Computer 213 table 6.9 Year-by-Year Percentage Increases in PC Shipments, 1996–2000

Dell Compaq PC Industry

1996–97

1997–98

1998–99

1999–2000

62 42 16

65 21 15

51 17 22

27 8 15

source: IDC press release, January 2001; Gartner Dataquest press releases.

when it became the largest PC firm in the world (Table 6.10), however, every action of the company, from Internet selling to reorganization of the entire procurement, production, and distribution system, became a source of intense scrutiny by the rest of the industry.26 Other firms began to experiment with elements of Dell’s business model. This adaptive process of experimentation began to reshape the PC industry in Dell’s own image and likeness. in dell’s own image and likeness The adaptive response of firms to Dell had two primary components. The first was Internet selling. Almost two years after Dell began to sell its PCs over the Internet, Compaq, IBM, and HP also initiated Web sales of personal computers directly to customers. Unlike Dell, however, the initial target market for this effort was primarily consumer sales. A steady stream of press releases in 1998 along with print ads in major media, especially from Compaq and IBM, announced this Internet sales effort by Dell’s competitors.27 Nevertheless, Internet sales posed an almost intractable dilemma for Dell’s competitors that sold systems through the indirect channel (Kenney and Curry 2000a, 19). By marketing PC systems directly to customers over the Internet, Compaq and other purveyors of the indirect sales channel would be in direct competition with the very distributors and resellers upon whom they depended for the overwhelming bulk of their sales. This contradiction lies at the core of Compaq’s problem in trying to develop an Internet direct system of sales. When in 1998 Compaq made one of its frequent announcements of plans to sell direct to some of its corporate accounts, and direct through the Web to consumers, its channel partners, grasping the contradictory position of Compaq in direct selling, reacted with both hostility and skepticism. While Compaq’s resellers were unreceptive to the idea, at the same time, consensus among them was that Compaq would be forced “to sell 100 percent through the channel” (Hayes and Connolly 1998). Alternatively, Compaq and the other indirect firms could continue to avoid the In-

214 Dell Computer table 6.10 Computer Firms Ranked by U.S. Market Share, 2001 Rank

Firm

U.S. Share (%)

World Share (%)

1 2 3 4 5 6

Dell Compaq Hewlett-Packard Gateway IBM Apple

24.0 12.7 9.4 7.6 6.1 4.8

13.4 12.1 6.9 3.2 7.2 4.1

source: IDC press release, July 30, 2001.

ternet as a sales channel. Realistically, however, they could not simply allow Dell to dominate Internet sales and gain market share at their expense. As a result, in 1997–98 Compaq, HP, and IBM began cautious approaches to Internet selling. The second element initiated by these firms as part of an Internet orientation that was more far-reaching was to combine Internet sales with a build-to-order business model. Although inspired by Dell, this build-to-order business model was different from that pioneered by Dell. It was created specifically for the indirect channel. This new custom-build system was called channel assembly. Channel assembly represented a joint effort by indirect PC firms and their largest distributors, such as Ingram Micro and Tech Data, to respond to Dell by developing their own “pull” system for assembling PC systems.28 This response was essentially a compromise aimed at enabling indirect vendors to adopt an Internet selling system without alienating their channel partners. In the traditional indirect selling system, assemblers such as Compaq would make as many computers as their demand forecasts projected and ship them to distributors for final marketing. They would then hope that the forecasts were accurate and the distributors would be able to sell them. In this indirect system, distributors, although serving as the customers of the PC makers, had a potentially antagonistic relationship with the PC firms. Ownership rights to the end user were often zealously guarded by distributors and resellers. As a result, PC makers and their indirect distributors maintained more of a market relationship with each other. Channel assembly aimed to change this process first by creating a different type of relationship between the PC maker and the distributor, and sec-

Dell Computer 215

ond by dividing the assembly of the PC into two phases (Kenney and Curry 2000a, 22). In the first phase, the components of the PC that decreased in value more slowly than the other parts would be assembled by the PC firm. During the second phase, the components most susceptible to price decline—DRAMS, microprocessors, and hard disk drives—would be added to the box by the distributors when orders were actually received. Examples of the program at IBM and Compaq reveal how the influence of Dell has shaped the industry. IBM initiated its Authorized Assembler Program in 1997. Perhaps not surprisingly, it recruited an executive from Dell, Steve Martson, to manage the program. Compaq began the Configuration Partner Program of its Optimized Distribution Model during the same year. As part of this program, Compaq reduced the number of its primary distribution partners in the United States from thirty-nine to just four (Kenney and Curry 2000a, 23). The motivation for this reduction was to work better with fewer channel partners. Such a move was similar to Dell’s effort during the same time to reduce the number of its primary suppliers. These changes in the indirect system of assembly and the accompanying shifts in the relationships between PC makers and distributors resulted in systems of procurement, production, and distribution with geographical impacts that in many ways resembled the enterprise geography of Dell. First, channel assembly resulted in distributors establishing final configuration centers adjacent to the factory locations of PC makers in a program known as co-location. In this way, PCs assembled during the first phase of channel assembly by the personal computer maker could be more easily and more quickly transported to distributors for the second and final phase of the custom configuration. Second, the PC firms have essentially imitated Dell’s model of having components stored as inventory in warehouses located adjacent to configuration centers. “All of the major PC makers have followed Dell in using the SLC system of storing parts next to assembly facilities,” says one of Dell’s suppliers that also supplies other firms (supplier 1, interview, April 26, 2002). IBM has compelled its suppliers to establish replenishment services centers (RSCs) near the IBM’s fulfillment centers. These RSCs, much like Dell’s supply logistics centers, are either individually owned or leased by suppliers or managed by third-party logistics providers. Similarly, Compaq is using third-party logistics provider CSX in Houston to operate and manage a parts warehouse next to Compaq’s main campus. In effect, just as proximity emerged as a critical strategic element in Dell’s business system, for similar reasons it has also emerged as an essential element in the evolving business systems of Dell’s competitors. These changes did indeed make PC firms more efficient. Inventory levels throughout the industry assumed a downward trend similar to the trend

216 Dell Computer Dell Days of Inventory



PC Industry Inventory/Shipments Ratio

35

D

30

1.5

25 20

1.0

15 10

0.5

5 0

1994

1995

1996

1997

1998

1999

2000

0

figure 6.4 Inventory of Dell and Inventory/Shipments Ratio of the PC Industry, 1994–2000 source: Dell Computer Corporation, 10-K reports for various years; Compaq Computer, 10K reports for various years; Robert Cihra, “The PC Industry” (ING Barings Furman Selz, 1998); U.S. Department of Commerce, Office of Information Technologies, Computer Industry Indicators: 1992–2000 (Washington, D.C.: Government Printing Office, 2001).

at industry leader Dell (Figure 6.4). These levels of inventory reduction and enhancement of efficiency throughout the industry—improvements driven fundamentally by Dell—have also contributed to a much steeper series of price declines for computers during the Internet period (see Figure 6.5). Again, Dell as the most efficient and most dominant firm in the industry has played a critical role as the so-called King of Cutthroat Pricing in helping to drive this trend (Business Week, September 24, 2001, 92). From the second quarter of 1987 until the third quarter of 1994, price declines for computers revealed a trend of a 12.1 percent decline annually. By contrast, from the fourth quarter of 1994 through the fourth quarter of 1999 the trend decline for computers was more than twice as rapid, amounting to 26.2 percent (U.S. Department of Commerce 2000, 2). Nevertheless, as the rest of the industry tries to follow what Dell has accomplished, these firms are not always equally successful in duplicating Dell’s results. Many of them, in experimenting with new strategies such as channel assembly in the case of IBM and Compaq or even complementing a direct sales model with retail stores as Gateway has done, have encountered new and perhaps unanticipated difficulties (see, for example, Figure 6.6). First, these firms face ongoing challenges in grafting elements of Dell’s Internet direct model on to their own existing business models. In the parlance

Dell Computer 217 120~-----------------------------------.

100

Trend decline of 26.2 percent betvJeeo 9404 and 99Q4

50

25 Trend dedine of 12. 1 pen:ent betvJeeo 8701 and 9403

10

1987

1989

1991

1993

1995

1997

1999

figure 6.5 Declines in Computer Prices, 1987–1999 source: U.S. Department of Commerce, Digital Economy 2000 (Washington, D.C.: Economics and Statistics Administration, 2000), 2.

of the industry, Dell’s competitors face the problems associated with the “legacies” of their own business systems (Schrage 2002). Second, and perhaps more important, none of these firms appear to have the same capacity for execution of the Dell system, either planning or fulfillment, as does Dell itself. As a consequence, while Dell has forced its competitors into a process of adaptive response, it has retained decided advantages over them. Although they now more closely resemble Dell, they are not necessarily as efficient, productive, or competitive. In effect, Dell has forced competitors such as IBM and Compaq to overhaul the procurement and production logistics of a once-dominant producing and selling system. Through its competitive ascent, it has imposed its standards upon an entire industry. In the process, Dell has forced the industry to change. It has succeeded in defining the terms of competition in the personal computer industry, reinforcing the shift in those terms away from the product and toward the issue of costs and systems of logistics. Although these changes have resulted in the entire industry becoming more efficient, such transformations have enabled Dell, as the most efficient company, to emerge as the most competitive. Indeed, it is hardly an exaggeration to say that Dell has succeeded in remaking the PC industry in its own image and likeness. Others are trying to find ways to catch up.

218 Dell Computer Productivity Index 1.0

Dell 0.8

0.6 Compaq 0.4 Gateway

0.2

1994

1995

1996

1997

1998

1999

2000

2001

figure 6.6 Productivity Index for Dell, Compaq, and Gateway, 1994–2001 (measured by sales per employee) source: Hoovers Online Company Profiles.

Appendix: Interviews and Interviewees Access to interview subjects from Dell occurred in two ways. First, I assembled an informal network of individuals in the journalistic and business world who provided me with referrals to specific Dell individuals. Second, I employed an approach inspired by Dell itself. I went direct. From journalistic and business literature on Dell, I assembled a list of interview subjects relevant to the issue areas of my study. I then contacted these individuals from Dell via e-mail and explained what I was trying to do. Most of these individuals were extremely interested in trying to help me understand the Dell business system. Some agreed to let me refer to them by name. Others requested anonymity. Suppliers always requested confidentiality. From these contacts, interviews for this project assumed three forms: (1) face-to-face, (2) phone, and (3) e-mail. I also secured an invitation for a firsthand tour of the Morton Topfer manufacturing facility and the supply logistics center adjacent to this facility in Austin. There was no set questionnaire for the interviews. As I learned more about the Dell business model over the course of the interview process,

Dell Computer 219

which lasted roughly one year, my questions evolved and changed. In putting the Dell story together while conducting these interviews, I asked questions about particular elements as they emerged in my analysis. The goal was to get a sufficient amount of firsthand information not found in journalistic accounts of Dell and to create a scholarly story of how the PC maker operates what is arguably the most innovative business organization in the industry. dell interviewees Lance St. Clair, Director of Supply Chain and Materials Management Systems Stephen Cook, AFC Senior Process Engineering Manager Laury Johnson, Senior Manager, Logistics, Compliance and Procurement, Dell Brazil Gregory Kelly, Senior Manager, Dell Nashville Materials and Logistics Victor Koncaba, Senior Logistics Manager and Information Systems Architect Eric Michlowitz, Director of Supply Chain e-business Dan O’Donnell, Procurement Manager, Dell Europe Daryl Robertson, Vice President and General Manager, Dell Latin America Rosan Sison, General Manager, Dell Philippines Anna Belle Williams, Senior Manager, Worldwide Procurement Simon Wong, General Manager, Dell Asia Pacific Senior Logistics Manager for the Americas Inbound Supply Chain Manager Former Dell Executive Vice President of Manufacturing and Supply Chain Management other key interviewees Robert Cihra, ING Barings Nigel Johnson, Eclipse Group Robert Persuit, SJ Consulting Navi Radjou, Forrester Research Richard Tobin, Limerick Development Corporation Supplier 1 Supplier 2 Supplier 3

chapte r

7

Conclusion The Rhyme of History

it is often stated, more as conventional wisdom than verifiable truth, that history repeats itself. From this popular vantage point, outcomes occurring in one epoch, along with the actors behind such events, inevitably resurface at another point in time in a process fundamentally unchanging and immutable. Such a view of history, however, is at best uninformed and naïvely uncritical. More to the point, it is simply wrong. History never repeats itself, because every historical moment is unique. Nevertheless, the process of history does admit to a poetic quality that more accurately depicts its true character. History rhymes—not repeats—in revealing parallels between the events, actors, and outcomes from different periods. Implicit in this approach is the idea that the subject of history is both development and continuity. This rhyme of history has guided the comparison of Swift and Dell in this study. At the core of this comparison lies a fundamental issue in the study of economic development: how do economies grow and change, and what provides the catalytic agent in this process of transformation? The cases of Swift and Dell reveal most profoundly how firms creating innovative business organizations in response to communications revolutions reshape patterns of economic development and territories for profit making. The acquisition of new capabilities is what enables the innovative enterprise to implement the routines and create the forms of organization that transform the territorial configuration of economies. Among the most insightful observers of this process, Schumpeter argued that innovation leading to business cycles constituted the essence of the capitalist process. His approach to innovation and economic development, influenced most decisively by Marx, was fundamentally historical.1 Innovation occurred in waves that demarcated different historical periods. These periods were unique but possessed com-

The Rhyme of History 221

mon and comparable characteristics. In focusing on Swift and Dell, this study has uncovered parallel worlds of innovation. Although separated by a century, these parallel experiences provide insights into the underlying rhythms of economic growth and change over time while illuminating the character of both periods, past as well as present.

The Pattern of Innovation and Economic Change The pattern of innovation at both Swift and Dell derives from a similar historical origin—a revolution in communications. This shared phenomenon, one occurring during the late nineteenth century, the other occurring a century later, provided the historical foundations for the innovations created by the two firms. The railroad and telegraph and the overland system of interregional commerce created by this infrastructure established the preconditions for Swift to recast the system of beef production and distribution. Similarly, the Internet and the system of commerce evolving from this infrastructure enabled Dell to reorganize the production, distribution, and sales of personal computers. From this common platform of communications breakthroughs emerged similar sequences of innovation and economic development at the two companies. In both instances, communications breakthroughs altered the environment of profit-making opportunities, by reshaping the geography of economic activity and providing firms with new forms of control over the timing cycles and territorial arrangements of business operations. What emerged as opportunities for Swift and Dell were reconfigured systems of economic access across and within markets, a recalibrated structure of costs for engaging in economic activity across and within these reconfigured territories, and enhanced forms of control over the temporal and territorial organization of business activity. With access, costs, and control transformed by new infrastructure systems, both companies confronted economic environments replete with new choices for recasting business routines. At the same time, competition and the profit system itself provided the necessary incentives for Swift and Dell to exploit these reconfigured geographies and enhanced systems of control and learn routines for competing differently. This process of learning different routines led to the development at both firms of new capabilities, new forms of control over the products of work and the work process itself. Enhancement of capabilities, in turn, compelled both Swift and Dell to establish new sets of organizational arrangements, resulting in new forms of business organization—innovative enterprises—designed to implement innovative routines. Distinguished by these four elements—opportunities, incentives, capabilities, and organizational arrangements—this learning process is fundamental to what occurs inside the so-called black box

222 The Rhyme of History of the firm where innovation takes place (Rosenberg 1982a, 1994). The business organizations of Swift and Dell, however, were more than innovations in economic space. These enterprises reshaped territories for economic activity and profit making. They became geographically embedded in the way they coordinated flows of information and materials across space and in the way they organized certain relationships of proximity between key nodes in their networks in order to facilitate high-volume rapid turnover in circuits of procurement, production, and sale. These reconfigured territories represent geographies of innovation shaping the process of capitalist development. A similar route from the communications revolution, to innovation and the creation of innovative enterprises, to the reconfiguration of territory for profit making links the experiences of Swift and Dell across time. With both Swift and Dell, the learning process leading to the creation of innovative enterprise was of a specific type. Both firms engaged in a process of knowledge creation known as learning by using (Rosenberg 1982b). Similar to the acquisition of tacit skills and capabilities stemming from learning by doing, learning through use involves direct engagement in the productive process as a technology user. When firms learn by using, they engineer incremental improvements, through experimentation and trial and error, not in products, but in processes for making and marketing products (Rosenberg 1982b, 121–22). What Swift and Dell learned by using was how to deploy technologies of communications revolutions as the basis for innovations in the logistics of producing and selling. These two companies, in effect, learned, from use, how to make profit as logistics firms. The innovative advances of Swift and Dell evolved from the organizations they created and the capabilities they acquired in reinventing the logistics of producing and selling their products. The nature of the innovative advance and the legacy left by both companies upon the historical periods in which they operated derived precisely from this shared role as logistics companies. The Swift business system and the “Dell model” are essentially innovations in logistics and distribution. Although Swift and Dell emerged from this process of innovation as the largest and arguably most successful firms in their respective industries, it was the impact they exerted on the competitive behavior of other companies that enabled the two innovators to influence patterns of development in the economy. Once the competitive superiority of the business models created by Swift and Dell became known to their competitors, and once the two firms had ascended into the top ranks of the beef and PC industries based upon these business models, the organizations they built and the processes they pioneered became sources for the adaptive responses of firms aiming to compete with them. The integrated organization of Swift, built around dispersed branch house networks and concentrations of slaughtering facilities

The Rhyme of History 223

in the Midwest, was soon the basis of business models adopted by other firms in the industry. Similarly, the virtually integrated organization of Dell, with its Internet-based system of demand and supply balancing and inventory compression, has emerged as the competitive standard that other firms in the PC industry—with varying levels of success—have aimed to duplicate. In fact, the impacts of Swift and Dell as innovators were so compelling that these companies influenced firms in other industries beyond meatpacking and computing. A whole range of perishable goods industries imitated Swift (Chandler 1977). Dell’s influence extends beyond the PC even into industries as traditional as autos (Andrews 2000; McWilliams and White 1999). This process of diffusion, whereby innovations spread and become more generalized, is the complement to the creativity of entrepreneurial firms in promoting growth and change in the economy. It is what completes the process described by Schumpeter as creative destruction, in which firms imitate the creations of the successful innovator. As their business systems diffused and spread, the two companies succeeded in creating economic worlds in their own image. Finally, the experiences of Swift and Dell suggest that innovations in distribution and logistics, far from being the outcome of innovations in production, may very well be the source—the precondition—of production innovation and economic growth. What both firms succeeded in doing was to use distribution as a platform for creating accelerated systems of highvolume throughput. Conceived in this way, the innovations in distribution of Swift and Dell represent challenges to representations of industrial development in which manufacturers, aiming to speed production turnover, innovate production processes and then force adjustments in outmoded and antiquated systems of distribution to accommodate such systems of highvolume throughput. Indeed, the stories of Swift and Dell reveal the sphere of circulation to be far more central as a catalyst for growth—as well as a source of surplus—than is commonly conceived. Consequently, in terms of broad-based theory on economic development, this study uncovers a set of new facts to address an old and formidable problem in the historical literature. Originating in the work of Henri Pirenne (1925), this problem focuses on the extent to which trade, conceived broadly as the sphere of circulation, acts as the trigger of capitalist growth and change.2 This study has assigned to Swift and Dell a role as protagonists in this debate. By reinventing the relationship with the customer and creating organizations to implement such a vision for profit making, Swift and Dell elevated innovations in distribution as the pathway to time compression in production and sale and space reorganization as part of this acceleration of time. These two firms, in effect, provide an alternative way of looking at purveyors of “trade” in the growth process.

224 The Rhyme of History

The Contours of Parallel Worlds From this common platform of communications revolutions and innovation, Swift and Dell created business systems with parallel operational, organizational, and territorial characteristics. operations Operationally, the innovations of Swift and Dell consisted of direct-pull systems of production and distribution that operated on the basis of a similar principle. Customer orders, processed in real time from breakthroughs in communications technology, are the source for setting the system of procurement, production, and distribution in motion. In contrast to “push” systems, in which component supplies are stored as inventory and finished goods are pushed into the marketplace to be sold to customers on the basis of demand forecasting, the direct-pull systems of Swift and Dell relied on orders from customers already received to ignite the process of materials procurement and production. In both cases, communications revolutions enabled the two companies to establish the direct relationship with customers that provided the foundation for these pull systems. In the case of Swift, sales agents at branch distribution houses collected orders from retail butchers in the vicinity of the branch and then transmitted them on a daily basis to Swift’s headquarters, where they were broken down into purchasing requirements for cattle at stockyards in terms of quantities, cuts, and grades. The telegraph provided the essential communications links in this direct-pull system. In the case of Dell, orders from customers, both businesses and individuals, also pull the components and finished PCs through the cycle of procurement, production, and final marketing. Customer orders are routed to Dell’s headquarters and to various assembly sites, transformed into material requirements, and sent to supply logistics centers or local supplier factories. Components are then pulled from these facilities and delivered to Dell’s assembly site, where they are configured into finished machines. The Internet provides the essential communications links in Dell’s direct-pull system. While the two pull systems share a number of essential features, they differ in two important ways. First, in Dell’s system, customers are linked directly into the pull mechanism through Internet communication. For Swift, the telegraph was used internally within the firm. Although documentation is limited, the evidence suggests that retail butchers were not connected directly to Swift through telegraphic communication. Orders came to branch houses through face-toface sales calls with retail butchers located in the vicinity of the branch (Federal Trade Commission 1919, pt. 3, p. 127; Unfer 1951, 86). Consequently, the

The Rhyme of History 225

direct-pull system of Dell represented an advance over Swift in that Dell’s customers could be linked directly to procurement and assembly. Nevertheless, the direct-pull system in both cases operated on the basis of new communications technologies that enabled the two firms to use real-time information in the form of orders from customers as a substitute for the risk of forecasting market demand. In this sense, the business models of both firms were based on the principle of build to order. Second, the build-to-order pull systems of Swift and Dell differ with respect to the issue of customization. Dell’s system is, at its core, a system of mass customization. The firm relies on the Internet to build individually configured products in high volume as the basis for pulling materials through the process of assembly. Individual customer choice drives the build process. By contrast, Swift’s system was not organized around individually customized production. Nevertheless, Swift’s system does admit to an early type of customization made possible by telegraph technology. Consequently, while it would be an exaggeration to characterize Swift’s build-to-order system as a system of customized production, it would be equally wrong to characterize the Swift system as simply a mass-distribution system for an undifferentiated commodity. Myriad grades of beef from different varieties of cattle along with variations in cuts formed a range of product choices for retail butchers. Swift organized its procurement and disassembly activities on the basis of orders deriving from these choices. As orders collected from retail butchers at branch houses were telegraphed to Swift’s headquarters and then broken down into cattle-purchasing requirements, and as the various sides came off the disassembly line destined for certain branch house locations as per order, Swift to some extent was customizing its production in real time, knowing where and in what quantities these various cuts, sides, and grades were destined. In effect, Dell has perfected a system of just-in-time mass customization, a system with antecedents in the procurement, production, and distribution organization of Swift. In implementing their build-to-order pull systems, both firms essentially took advantage of a similar characteristic in the product—modularity and standardization—to create varying degrees of customization in their product offerings. It is precisely this quality of sameness in the standardized and modular elements that make up the PC and beef, that enabled Swift and Dell to create differentiated products for their customers. As a result of the modular and standardized nature of PC components, Dell has pioneered a system of mass customization by assembling these components “like Legos” (Langlois 2001, 26). Similarly, Swift took advantage of an extraordinary technical refinement in the division of labor for disassembly of cattle that resulted in semifinished dressed sides of a highly standardized and modular character that could then be “customized.” The company shipped these

226 The Rhyme of History standardized semifinished products to branch houses, where they were custom dressed into the precise cuts ordered by retail butchers. Swift and Dell also confronted a similar operational objective in organizing these pull systems: how to balance supply and demand flows between the different nodes in their networks in real time. Both companies used new communications technology to accomplish this aim. Swift relied on constant telegraphic messaging between branch houses, central headquarters in Chicago, stockyard purchasing offices, and disassembly sites to balance order demand from retail butchers processed at branch houses with purchases of cattle supplies and schedules for slaughtering, butchering, and shipping. Dell utilizes a similar system in linking nodes in its network through communications technology. Dell’s system of demand and supply balancing, however, relies on Internet messaging between company headquarters, assembly sites, supply logistics centers, and supplier factories. At the same time, Dell’s system incorporates an additional node in this chain of Internet communications that differentiates it from Swift. In Dell’s network, the customer is actually connected through new communications technology to the system of material balancing in the process of procurement, assembly, and distribution. It is the Internet that provides this connection. The fact that both Swift and Dell operated direct-pull systems of production made possible by technologies of communications revolutions helps dispel the commonly held belief of mass production as a system based solely on producing in high volumes. The case of Swift reveals the mass-production system indeed to be one of high volume. At the same time, however, Swift shows how mass-producers used enormous amounts of real-time information generated by the technology of telegraphy to modulate and control output in accordance with changing demand and to pull supplies as needed to meet shifting demand schedules. Communications and control were as important to Swift in its direct-pull system as they are to Dell in organizing its direct-pull just-in-time business model. organization The core of the innovative advances made by Swift and Dell were the organizations they created for producing and distributing goods. As they developed their direct-pull organizations, both companies essentially solved a similar problem in an effort to capture greater increments of value from production and distribution. Swift and Dell learned how to eliminate traditional wholesalers in the value chains of production and distribution. Their networks of production and distribution essentially disintermediated certain actors from the beef and personal computer value chains. In both cases, this process of disintermediation was forged on the foundations of communica-

The Rhyme of History 227

tions revolutions and the direct relationship with the customer that the technologies of these revolutions afforded. Swift built a highly integrated enterprise. In the process of integration, the company assumed ownership and control over most of the adjacent steps in the beef value chain. Vertical integration provided the firm with a response to the problems of risk in coordinating these geographically dispersed, time-sensitive operations. Swift was heavily exposed if any of the adjacent steps in the organization and the processes connecting them broke down. For this reason, the company assumed ownership over a vast array of different activities. Swift became a railcar builder and an ice harvester as part of its role as a cattle disassembler and dressed beef distributor. In organizing these activities, Swift utilized systems of administrative coordination—the visible hand—that replaced activities formerly coordinated through markets between different small businesses. As it experimented with these methods of administrative control, Swift built an organization without precedent virtually from scratch. Dell created an innovative enterprise with a different organizational structure. In contrast to the intrafirm network of Swift, the organization of Dell is an interfirm network that relies on the external capabilities of other firms. Without the technological expertise of firms outside the boundaries of Dell that supply the PC maker with virtually all of the components for PC production, Dell would not be in business. Consequently, Dell’s interfirm organization is built around a far more circumscribed set of core competencies, namely the assembly process, and the logistics of procurement, production, and distribution. Unlike Swift, Dell did not pioneer the interfirm structure of organization based on external capabilities. The disintegrated firm in the personal computer industry was already a well-established phenomenon when Dell entered the industry. What Dell did that was pioneering, however, was to use the Internet in building what it calls a virtually integrated enterprise. This form of organization provided Dell with certain benefits associated with vertical integration. First, the Internetdriven, virtually integrated organization of Dell enabled the PC maker to strengthen the direct relationship with its customers. Second, and more important, however, virtual integration has enabled Dell to extend this relationship with the customer in the other direction, thereby linking customers with suppliers and aligning procurement, production, and distribution more systematically around customer orders. Virtual integration, in effect, provides Dell with the organizational capacity to coordinate real-time supply and demand balancing in its high-volume build-to-order business system among the other firms in its network without having to assume ownership of the assets involved in these operations. Where Swift be-

228 The Rhyme of History came an asset-owning business enterprise, Dell has become an organization fundamentally asset averse. Nevertheless, in spite of these differences in organizational structure, both Swift and Dell employed a similar principle in organizing the movement of supplies and finished products through their networks. The two companies rely on the organizing principles of corporate power and administrative control rather than market coordination to ensure that materials and finished goods move within their networks from procurement through production to final marketing. Although Dell is a separate entity from the other firms that make up its network, it does not interact with these companies on the basis of markets and the price system in securing supplies and logistics services. On the contrary, Dell organizes the relationships of collaboration between itself and its network partners by essentially imposing upon them its own protocols—both technical and administrative—as a condition for entry into its network. Such use of force does not mean that only Dell profits from such relationships. Both Dell and its partners clearly benefit from this system of administrative control, but the idea that interfirm networks such as Dell’s reveal the flexibility and ascendancy of market coordination in the current economy is inaccurate. The need for Dell to exercise such control stems from the fact that the PC maker, in coordinating its high-speed build-to-order business model, confronts the same types of risk from disruptions at adjacent steps along the value chain as those faced by Swift. Just as Swift remedied such risks by taking control of virtually the entire value chain through ownership of assets, so too has Dell employed a mechanism for taking control of the value chain, but without having to assume ownership of the assets at these adjacent steps. Power and control are as much a part of the story at Dell as they were at Swift. Such forms of organizational control enable firms to act strategically and define the basic attributes of innovative enterprise. Indeed it is organizational control, not market control, that describes the process of resource allocation with respect to the enterprises most responsible historically for generating economic development (Lazonick 2003). In the logistics-oriented organizations built by Swift and Dell, Alfred Chandler’s visible hand has proven to be a more valuable asset than Adam Smith’s hidden hand. territory Business organizations are inherently territorial. They assume this territorial character in two ways. First, business organizations are collections of assets—network nodes—arranged territorially. Second, they are territorial in the way they route flows of materials and information between these assets and between these assets and the assets of other firms with which they interact in producing, buying, and selling.

The Rhyme of History 229

The organizations built by Swift and Dell reveal a similar geographical tendency to spread and concentrate in fundamentally new ways. Both organizations employ technology from communications revolutions to spread assets geographically and route flows of materials and information over long distances between these assets in establishing systems of long-distance control for accumulating profit. At the same time, both organizations reveal concentrations of assets and flows of material and information within specific places. In these places of concentration, relationships of physical proximity between key organizational assets, consciously engineered by Swift and Dell, play a critical role in their direct-pull systems. Swift used the railroad and the telegraph to build an organization extending over the territory of the United States that obliterated the localized character of beef slaughter and consumption while it eliminated the practice of shipping live cattle long distances. For the first time in history, cattle were being slaughtered in locations far removed from where they were being consumed as fresh beef. Alongside this pattern of spread, Swift and the firms it influenced consolidated slaughtering activity in Chicago and decentralized it in specific locales within the corn belt states of the Midwest. In the process, Swift and other large packing firms created industrial districts of slaughtering and meatpacking in Chicago and other Midwestern stockyard towns. The territorial pattern of this new and innovative way of producing and selling beef was one of a vast expansion outward, represented by distribution activity through branch houses and a powerful centripetal movement inward toward the center of the country for slaughtering. Swift assumed the role as agent in creating this territorial pattern. It was Swift that determined the location of branch houses and disassembly facilities. At the same time, Swift organized crucial relationships of geographical proximity between certain key nodes in its beef network. Swift established disassembly facilities at stockyards in which it invested in order to exercise control over cattle supplies. Branch houses, in turn, were located in virtually all urban centers such that the map of Swift’s branch house network and the map of urban America in 1900 were roughly the same. Both slaughtering facilities and branch houses, in turn, were systematically connected to rail and telegraph lines. This geography, with its pattern of spread and concentration, centralization and decentralization, was an integral element of Swift’s innovative business model. Perhaps most important, this geography of spread and concentration embedded in Swift’s organization provided the foundations for a national market in the United States. Dell is using the communications revolution of the Internet to build an organization with this same basic attribute of spread and concentration. In contrast to the nationally oriented focus of Swift’s beef network, however, Dell’s Internet-driven, virtually integrated organization is establishing new

230 The Rhyme of History standards for coordinating the logistics of procurement, production, and distribution on a global scale. The firm has spread and decentralized its assembly sites in an expansion pattern covering four continents. At the same time, the PC maker has chosen distinct regional concentrations of high technology for its build-to-order activity. In the process, Dell is actively shaping the agglomeration economies of contemporary high-technology industrial districts in much the same way that Swift shaped the industrial districts of slaughtering and meatpacking. In creating this extended and concentrated organization, the firm is playing an integral role in defining the territorial meaning of global markets and contemporary globalization. What Dell has established through its Internet-driven innovations in global supply planning and demand fulfillment is a set of fundamentally similar build-to-order production ensembles in different parts of the world, all functioning on the basis of the same territorial arrangement between key organizational nodes. In setting up these Internet-based production complexes, Dell has arranged the facilities in its network in relationships of spatial proximity in order to fulfill the highly compressed time schedules in its build-to-order system. Dell forces suppliers to have either a factory presence in each of Dell’s six global assembly locations or warehouse components in supply logistics centers near Dell’s assembly sites. Such proximity is essential so that the PC maker can “pull” parts from these factories or warehouses at two-hour intervals in accordance with its build schedules. Just as proximity was crucial to Swift in organizing the logistics of cattle supply procurement at stockyards and cattle disassembly, so too is proximity critical to Dell in coordinating the logistics of PC component procurement and PC assembly. Far from dispensing with the barriers of distance and defying the constraints of geography, the Internet in Dell’s organization has heightened the need for Dell to shape relationships of geographical proximity between certain nodes in its network. By enabling Dell to create procurement and assembly schedules in real time, the Internet has actually imposed new requirements upon firms for organizing space in the spaces of globalization. What Swift and Dell engineered through creation of innovative organizations were parallel systems of industrial districts. These systems of territorial spread and concentration, reflecting geographies of innovation where firms could compete in new ways, represent new territories of profit making.

Final Propositions Fernand Braudel, the celebrated historian of the Annales school, writes of three kinds of history: a “history of the world as it is being made”; a history of “conjunctures,” or sharp breaks; and a history of “structures” inquiring into long-term changes, which Braudel termed the longue durée (Braudel 1980, 74). In many ways, the comparison of Swift and Dell in this study

The Rhyme of History 231

combines these three historical time frames. In Dell, there is history still in the making, while in the comparison between the two, there is both the notion of a demarcation and, with a full century separating the two periods and the two protagonists, the possibility of viewing the stories of the two firms from a longer-term perspective. From the vantage of these three levels of history, the parallel worlds of innovation created by Swift and Dell provide the basis for an advance toward a set of propositions about the nature of the current period, and the broad meaning of the entire period covered in this study. These propositions are organized around a compelling debate in which this study is ultimately positioned: to what extent is the current period demarcated by a sharp break from what preceded it, and does the current period represent a so-called information age? Responses to this issue reveal enormously divergent and highly contested views.3 What follows is an engagement in this conversation. This study seeks closure in the paragraphs that follow by subjecting the economies and experiences of Swift and Dell to the discursive historical imagination in an effort to uncover a set of insights on the economic and technological meaning of contemporary society. The parallel worlds of innovation created by Swift and Dell are not historical accidents. Instead, the innovations in logistics created by these two firms derive most fundamentally from an historical trend that began in the nineteenth century with a sharp break from the past. Though history seldom admits to such clear-cut divides, in this case there is compelling support for such a break, represented by the advent of the railroad and the telegraph. Indeed, these were technologies truly without precedent (Drucker 1999; Carey 1988). Until that moment, there had not been a significant advance in the speed with which goods and information could travel overland or on the high seas since ancient times. Ships and humans on horses constituted the essential means and time frames for bridging distance in exchanging information and transporting people and merchandise. Rails and telegraphy transformed this historically long-standing paradigm of time and distance relationships. As a consequence, the profit environment for economic actors experienced a profound and irreversible shift. This shift elevated opportunities for profit making through control over recalibrated systems of time and space relationships in economic activity. The direct-pull logistics systems of Swift and Dell represent similar responses to this underlying trend. Beginning with the railroad and telegraph, firms in effect begin to confront an ongoing historical trajectory of transformation in relationships of time and distance. What emerges from firms in response to this ongoing set of shifts, reflected in the experiences of Swift and Dell, are innovations in profit-making activity built upon the recalibrations in time and speed and the reorganization of space and access that this trajectory presents. Furthermore, in addition to the experiences of communications and

232 The Rhyme of History transport system users, the technologies of railroads and telegraphy established foundations for human society to pose and resolve, in rapid historical succession, a series of related transport and communications challenges. Following almost immediately from the telegraph was voice telephony. After the telephone became widespread, it was not long before another major breakthrough occurred in the form of wireless radio broadcasting. Images came next through television, and then hybrids of symbols, voice, images, and wireless communication in the form of computer networking and eventually the Internet. On the transport side, the route from the railroad is equally compelling, leading to the automobile, air travel, and even the container ship. Undeniably each of these technologies is unique.Yet these technologies, beginning with the railroad and telegraph, have evolved along the same basic historical pathway. From an historical vantage lying somewhere between the longue durée and the not-too-distant mirror, these technologies appear to be clustered within a relatively short historical time frame. It is this historical cluster in its entirety, and the contrast with what preceded it, that makes compelling the idea of an information age with boundaries encompassing all of these technologies. To separate current information and communications technologies from earlier ones and to cloister the characteristics of the current period as somehow more distinct than previous communications breakthroughs disregards the ongoing historical relationship of communications systems and information to industrial society. Such a separation ignores how the Internet and its impacts have arisen as a new stage in an ongoing industrial and informational advance (Roszak 1994). Seen in this way, the entire period from the mid-nineteenth century to the present day is arguably a single communications revolution. This revolution began with the railroad and the telegraph. It is continuing to transform economy and society through the Internet. For business firms, the underlying theme of this communications revolution is one of control over recalibrated relationships of time and distance, providing firms with new capabilities for competing. Communications, capabilities, and control enable business firms to change how they conceive of profit making and how they act in pursuit of it. Within this historical space, the experiences of Swift and Dell are the manifestations of an ongoing communications revolution. In this revolution, business firms use new communications systems to accelerate cycle times in the turnover of capital and to master methods of control over the spaces where they spread and concentrate their operations. This revolution is not over. Swift and Dell represent different moments within it. It is likely to continue for many years to come.

Notes

Chapter 1 1. Territory in this study derives from the notion of the “region” developed by Harvey Perloff et al. (1960), who refer to the region as an area “tied by extensive interareal activity or flows” (p. 4). 2. On the role of proximity in modern manufacturing systems, see especially Gertler 1995. 3. Technology in this study refers to knowledge embedded in products, routines, and organizations for accomplishing purposeful and reproducible activity (Nelson and Winter 1982; Mokyr 1990, 275–76; Castells 1996, 29–30). Innovation involves an epistemological transformation—new knowledge—which leads to the creation of new products, new processes and organizations for making them, and new places where new products are produced, bought, and sold and where new organizations operate. 4. In this regard, see, among others, Coase 1937, Richardson 1972, Williamson 1975, Chandler 1977, Porter 1985, Powell 1990, Lazonick 1991a, Saxenian 1994, and Borrus, Ernst, and Haggard 2000. 5. The reference is from Marx 1851, but this interplay of structure and agency can also be found in Giddens 1984 with specific applications to business history by Joanne Yates (1997). 6. Technology deriving from transport and communications systems, is not the only influence on the territorial organization of markets. Markets also emerge on the basis of politics (Polanyi 1944; Christopherson 1993; Zysman 1994). Market boundaries expand and contract as a result of control over territory exercised by political authorities that set rules for economic activity and establish systems of entitlements, rewards, and costs on market actors in the areas under their rule. Such authorities condition the extent to which market actors engage in, benefit from, or abandon economic activity within the territory in question. In addition, politics in the form of entitlements and rules, as well as struggles between competing builder groups and actor networks, plays a critical role in influencing the actual development and deployment of communications systems. In this way, systems of transport and communications reconfigure the geographical organization of markets owing to the

234 Notes to Chapter 2 impact exerted by politics on the development of transport and communications infrastructure itself. 7. For a different set of perspectives on this notion of communications revolutions as control revolutions see Beniger 1986,Yates 1989, and Mulgan 1991. 8. On innovation as a learning process with goals seldom completely knowable in advance, see especially Kline and Rosenberg 1986, 297–98; Rosenberg 1994, 53–54; and O’Sullivan 2000, 407. 9. It is also surprising that one of the most celebrated works on the origins of mass production in the United States (Hounshell 1984) does not contain material on Swift or the meatpacking industry.

Chapter 2 1. Examples from the opening chapters of Capitalism, Socialism, and Democracy emphasize this point. Schumpeter observes that “Marx saw this process of industrial change more clearly and he realized its pivotal importance more fully than any other economist of his time.” As for the sources of his own historical approach to innovation, Schumpeter writes of Marx: “He was the first economist of top rank to see and to teach systematically how economic theory may be turned into historical analysis and how the historical narrative may be turned into histoire raisonnée” (Schumpeter 1942, 32, 44). Such passages contrast with the often static contemporary discussions of whether Marx was “right” in his analysis of capitalism’s attributes and tendencies. For Schumpeter, the picture of Marx was complex, resonating with both success and shortcomings. The literature on Marx’s impact on Schumpeter is vast, but see especially the work of William Lazonick (1991a, 1991b, 1994) and George Catephores (1994). 2. In developing his idea of entrepreneurialism, Schumpeter discarded two other key concepts in Marx. First, Schumpeter did not accept Marx’s emphasis on dialectics in history. Second, Schumpeter rejected Marx’s view that class conflict was the motive force in history and economic development. As a consequence, Schumpeter argued that capitalists did not achieve a preeminent position in the economy by exploiting the working class. On the contrary, he argued that the driver of capitalist society consisted of capitalists competing against and stomping upon themselves. Interestingly, Robert Brenner, in a recent analysis of the world economy written from a Marxist perspective, argues along the same lines, that the logic of competition— the horizontal relationships between capitalist firms—not class struggle, rules the rhythms of contemporary growth and recession (Brenner 1998). 3. Schumpeter did concede, however, that the entrepreneurial function and the process of economic change still required more detailed investigation in order to understand “the actual working of capitalism that we are but dimly perceiving as yet” (Schumpeter 1947, 156). 4. Schumpeter acknowledged that the idea of innovation cycles or “waves of innovation” had come from previous theorists, notably Nikolai Kondratief, Clement Juglar, and Joseph Kitchin (Schumpeter 1939; Hall and Preston 1988). Schumpeter’s theory has produced a separate debate on the timing and duration of long waves.

Notes to Chapter 2 235 Within this debate, however, Carlota Perez (1983, 359) argues that Schumpeter’s work does not actually provide a basis for long waves. She insists that Schumpeter’s theory is instead an account of the short-term cyclical movement of recession and recovery exhibited by the capitalist economy. 5. The idea of an economy in equilibrium may appear paradoxical in Schumpeter’s work since he aimed to distance his historical and evolutionary approach to the economy from neoclassical notions of equilibrium. Schumpeter explained, however, that his use of equilibrium was an analytical tool from which to launch his notion of technological disturbance. 6. Among the innumerable contributions to this literature see Rosenberg 1982a, 1994; Nelson and Winter 1982; Freeman 1982; Lamoreaux, Raff, and Temin 1999; Dosi 1997; Dosi, Teece, and Chytry 1998; O’Sullivan 2000. 7. Richard Nelson and Sidney Winter are explicit in making this connection. “The influence of Schumpeter is so pervasive in our work that it requires particular mention” (Nelson and Winter 1982, 39). 8. Risk and uncertainty are not identical. Risk is a known distribution of possible outcomes. Uncertainty reflects a situation in which the outcomes themselves are unknown (Knight 1921). 9. These notions of “imperfect” knowledge and “incomplete” understanding in no way imply that there exists in reality some state of perfect information to which firms aspire. Such a state exists only as one of the many assumptions of the economic world in neoclassical economics. 10. This interplay of structure and agency is the central idea in Anthony Giddens’s theory of structuration. He argues that historically conditioned environments shape—not determine—human action, which in turn reconstitutes those environments (Giddens 1984; Yates 1997, 161). The classic formulation of this idea comes from Marx, who observed that human beings “make their own history” but not as they please. “They make it from circumstances that are given and transmitted from the past” (Marx 1851). 11. Changes in the environment, however, do not mechanically produce innovation. Quoting historian of medieval technology Lynn White on the impact of new technologies on the process of innovation, David Hounshell (1995) points out that “a new device merely opens a door; it does not compel one to enter” (p. 210). 12. Although Thomas Hughes is typically categorized as a social constructionist, his approach reveals certain subtle differences, which he acknowledges in locating his views “somewhere between the poles of technological determinism and social constructivism” (quoted in Hounshell 1995, 215). 13. Within this context of reverse salients and momentum, standards and dominant designs play both a technical and social role in influencing the pathway of innovation. When standards or designs for certain products and processes become dominant and force other products and processes in the economy to adapt in order to function, such standards or designs can both determine and constrain innovation. In this sense, pathways for innovative advance become established owing to the difficulties of moving so many interdependent economic activities already functioning on the basis of the dominant standard or design to an alternative technological path.

236 Notes to Chapter 2 Certain standards or designs that become so thoroughly embedded in the economy—the QWERTY keyboard is the most well-known example, but the Microsoft operating system is equally compelling—can preempt innovation by inhibiting economic actors from exploring an alternative economic path. Standards and dominant designs are also sources of social and political struggles within actor networks— “standards wars”—because of the high stakes in control over dominant technologies. On dominant design, see Utterback and Suarez 1993 and Henderson and Clark 1990; on the process of standard setting, see David and Greenstein 1990 and David 1987. 14. Much of this critique, however, is directed at the work of one individual in particular, Alfred Chandler, discussed in more detail below. 15. AnnaLee Saxenian (1994, 6–7) emphasizes, however, that proximity alone is insufficient as an enabler of innovativeness and competitiveness. Instead, place-based concentrations of economic activity must have other attributes that together create an innovative industrial system. Nevertheless, for Saxenian, it is place, built from unique local histories, culture, and institutions, that differentiates industrial systems providing the source of innovative learning. 16. Despite this seemingly one-sided emphasis on the nature and boundaries of business firms, Coase was not without insights on the impacts of geography on organization. “Inventions which tend to bring factors of production nearer together by lessening spatial distribution tend to increase the size of the firm. Changes like the telephone and the telegraph which tend to reduce the costs of organizing spatially will tend to increase the size of the firm” (Coase 1937, 397). 17. Chandler points out, however, that strategies could be carried out through different forms of organization, although he insisted that the integrated corporation prevailed because it was the most efficient. 18. It was only in the aftermath of completing The Visible Hand that Chandler took an interest in Williamson’s work on transaction costs and its implications for his own emphasis on strategy and structure (see Chandler 1988). Chandler conceded the possibility that coordination of supplies, production, and marketing within the boundaries of the firm also resulted in lower transaction costs for the large corporate organization. In Chandler’s view, however, reductions in transaction costs were more an outcome stemming from more efficient coordination than a cause for organizational change. For Chandler, it was economies of high-volume throughput and economies of speed that created the basis for administrative control underlying the large integrated corporation, not costs of transactions. 19. Chandler’s critics, notably Gerald Berk (1994), William Roy (1997), Philip Scranton (1997), and Charles Sabel and Jonathan Zeitlin (1985), make two basic counterclaims to his argument. First, they are especially critical of Chandler’s efficiency argument, insisting that Chandler’s account suffers from technological determinism. Technology and efficiency, they argue, are insufficient explanations for the evolution of the large vertically integrated corporation. Second, because of this focus on technology and efficiency, Chandler (in the view of these scholars) neglects the political struggles at the center of industrialization and is oblivious to the fact that the choices made by firms about technology, strategy, and firm structure were

Notes to Chapter 3 237 politically, not economically, motivated. These shortcomings preclude Chandler from recognizing the diversity of outcomes during the late nineteenth century in terms of firm structure, regionalism, and technologies. There seems little reason, however, why Chandler’s argument emphasizing the primacy of technology and economics is incompatible with the view that politics is critical to the way firms make choices about competing. For an excellent overview of this debate, see Hounshell 1995. 20. This idea was developed more fully at roughly the same time in Sabel and Zeitlin 1985. 21. There is a vast literature on this topic of industrial districts and networks of firms. For an overview of some of the earlier theorizing, see Pyke, Becattini, and Sengenberger 1990 and Scott 1988. 22. In this regard, see the influence in Saxenian’s work of Marc Granovetter (1985) on “embeddedness” and his subsequent work on “business groups” (1998). Granovetter, in fact, makes the observation that nowhere do firms engage in relationships with other firms in a purely atomized fashion through the price system but in fact draw upon relationships that coalesce into identifiable social structures (Granovetter 1998). See also Adler 1998. 23. In a subsequent version of this thesis on the Vanishing Hand, Langlois builds a broader argument about the ascendancy of markets in which the modular revolution of the earlier paper assumes less of a role in driving the phenomenon, but his overall thesis regarding Smith and Chandler remains fundamentally the same (Langlois 2003). 24. Although Albion was the first to examine the phenomenon of the communications revolution systematically, earlier insights about the role of communications technologies in economic development can be found in the writings of Marx (1867, 1885) and Emile Durkheim (1893). Marx was perhaps the first to write about distance compression and market expansion stemming from advances in transport and communications technology. Durkheim (1893) observed how transport and communications technologies of the late nineteenth century tended to break down “segmented” or local markets in creating larger markets of an “organized” type (p. 305). It is also worth noting, with respect to Albion, that other scholars, most notably Elizabeth Eisenstein (1979), trace the lineage of the communications revolution much earlier to the phenomenon of printing during the fifteenth century. 25. Castells (1996) is careful to point out, however, that “technology does not determine society.” He insists instead that “technology is society” (p. 5). 26. Nevertheless, the process is not circular, implying historical repetition. Instead, the process is conceived as a spiral, representing parallel historical experiences within an overall context of development and change.

Chapter 3 1. Alexander Field (1992) concedes, however, that while rails and telegraphy emerged in tandem, “they were surely one of the most disproportionate pairings in the annals of economic history” (p. 401). He notes huge differences in capital and operating costs, total capital value of the built systems ($8 billion for railroads, $70–

238 Notes to Chapter 3 147 million for the telegraph industry), and total employment. The contribution of the telegraph lay in its capital-saving capability, enabling it to offset the capital-using bias of the railroad (p. 412). 2. In 1850, telegraphic messages from New York to New Orleans had to be rewritten four or five times at intermediate stations en route before being passed along to the next station. Interconnection problems between competing companies operating on different telegraph technologies created such bottlenecks (Jones 1852, 87). Similarly, railroads had to rely on imperfect interconnections. Hoists were used to lift freight cars off and on different wheel bases at points of transshipment between railroads operating on different track gauges. 3. In 1893, historian Frederick Jackson Turner proposed an enormously influential thesis on the origins of this expansion. He argued that the availability of free land in the West drove the boundaries of the American frontier in a continual westward direction. While Turner elevated the pioneering spirit of white settlers in this expansion, in truth, much of this free land owed its origins to government policy. A history of national legislation, culminating in the Homestead Act (1862), reflected government efforts to encourage settlement and development of western lands. By granting title at low costs to those willing to settle the area, along with ceding land to businesses, notably railroads and mining companies willing to exploit the area, government shaped western development. Critics, however, assailed Turner’s approach as an apologia for the Indian conquest (cf. Cronon 1991 on the debate). Despite the controversy of Turner’s thesis, there is little denying the expansion of population and economy in the Ohio and Mississippi River valleys and the Great Plains—and the impacts of this expansion on the creation of a rail- and telegraphbased system of interregional commerce. Ironically, Turner’s thesis coincided with the pronouncement by the superintendent of the 1890 census that the frontier had disappeared. 4. Admittedly the importance of the southern market to western agriculture and the extent to which the South was dependent on the West for foodstuffs remain the subject of debate. See especially Fishlow 1965, 276–88. Less debatable is how agriculture in the West evolved into a surplus-producing activity that acted as the catalyst for the system of East-West, rail-dominated long-distance trade. 5. Albert Fishlow (1965) points out that the sequence of railroads traversing undeveloped territory and inducing economic development “bears no resemblance” to the real world (p. 166). He attributes this idealized scheme of construction ahead of demand to Schumpeter, who, in Business Cycles, insisted that Midwestern rail construction “meant building ahead of demand in the boldest acceptance of that phrase” (Fishlow 1965, 165; Schumpeter 1939, 328). 6. For a more detailed analysis of the information presented in this subsection, see Chapter 4. 7. Price dispersion can be conceived as a measure of risk in intermarket trade, while risk is a reflection of barriers, among them, distance. The greater the differences in prices between regions, the greater the level of risk in intermarket economic activity. The higher the levels of risk, the lower the levels of intermarket trade.

Notes to Chapter 4 239 8. Urban places as defined in the censuses of the period are cities with at least twenty-five hundred inhabitants. Figures in the rest of the paragraph are taken from Pred 1977, 86, and U.S. Bureau of the Census 1975, 11. 9. During this early period, the major change in the city system occurred just outside the largest group, with the rapid growth of Cincinnati and Saint Louis. By 1850 they were the nation’s sixth- and seventh-largest cities, reflecting the primacy of an expanding river-based economy in the Ohio and Mississippi Valleys. By contrast, the rise in rank of cities such as Milwaukee, Cleveland, and Detroit from 1850 to 1890, though not as dramatic as that of Chicago, and the decline of New Orleans represented a declining river economy and a shift of fortunes to a rail-dominated internal trade system. 10. Even Mark Twain, in his semiautobiographical Life on the Mississippi (1883), remarked upon “that astonishing Chicago—a city where they are always rubbing the lamp, and fetching up the genie, and contriving and achieving new impossibilities. It is hopeless for the occasional visitor to try to keep up with Chicago—she outgrows his prophecies faster than he can make them. She is always a novelty; for she is never the Chicago you saw when you passed through the last time” (p. 398). 11. On this point, the metaphor of Fernand Braudel, which describes early modern European cities as “giant stomachs” that created opportunities for large grain and provisions merchants, is instructive. Braudel (1977) argues that expanding urban populations and accompanying consumer demand in cities were critical elements in the interactive development of urbanization and capitalist development (p. 28).

Chapter 4 1. For a shorter summary of themes in this chapter see Fields 2003. 2. On dressed beef as a new product see Kujovich 1970, 462. 3. In his study of forty-three entrepreneurial leaders of the nineteenth century classified into twenty-four categories, Chester Destler (1946) notes that five entrepreneurs of the period—Swift and Armour in meat, McCormick in reapers, Pillsbury for flour, and Duke with cigarettes—were notable for creating public demand for their essentially new products (pp. 29, 42). 4. Although statistically the value of pork products exceeded that of beef in 1900 (Table 4.6), the extraordinary increase from 1870 to 1900 in “miscellaneous meat products” is a statistical reflection of the ascendancy of beef. This miscellaneous category consists mostly of beef by-products so that when beef and miscellaneous meat are added together, as they should be, their value does in fact surpass the value of pork. 5. In 1843–44, over ninety-six western towns in Ohio, Kentucky, Indiana, Illinois, Iowa, and Missouri were involved in pork packing (Yeager 1981, 4). Because of the role played by riverboats in the pork-packing trade, the overwhelming majority of these pork-packing operations were located along rivers, primarily the Ohio, Mississippi, Illinois, and Wabash river systems (Yeager 1981, 5; Walsh 1978). 6. Rudolf Clemen (1923, 64–68) charts the westward migration of cattle raising during the 1840s, 1850s, and 1860s and the corresponding increase during these same

240 Notes to Chapter 4 decades in what he termed “minimum class” states, that is, states with cattle populations insufficient to satisfy their own cattle demand. 7. Information in this paragraph on the advent of the cattle commission houses and their centrality in the cattle trade is taken from Clemen 1923, 86–91. 8. The Evener System was devised by the three livestock-carrying railroads, the Erie, the Pennsylvania, and the New York Central, to divide the live cattle and hog traffic, guarantee each an agreed-upon portion of shipments, and stop the competitive rate cutting. Each railroad selected two major livestock shippers as “eveners.” If a railroad was unable to secure its quota of shipments through the normal distribution of livestock traffic, the eveners made up the quotas. As a reward for balancing the distribution, the eveners received a rebate of fifteen dollars per carload of livestock. The system existed from 1875 to 1879, when it broke down owing to the expansion of the dressed beef traffic (Yeager 1981, 30–40). 9. On this point, the Senate Committee Investigation on the Transportation and Sale of Meat Products concluded that “the enormous power wielded by the Trunk Line Association is almost incalculable. Every pound of freight and every head of cattle and hogs going either way across the continent must pay tribute to the roads comprising this vast combine” (U.S. Senate 1890, 17). 10. The Eastern Trunk Line Association was a cartel formed in 1877 of the major East-West railroads, which included the New York Central, Erie, Pennsylvania, Baltimore and Ohio. Its aim was to stabilize rates by pooling profits and traffic, thereby removing incentives for “ruinous” rate cutting. By 1879, the association included the major railroads operating east of the Mississippi. It also included the Grand Trunk, the major East-West Canadian Railroad that operated from Chicago to Boston via Detroit and Montreal. With no Canadian rivals, however, the Grand Trunk was able to circumvent the restrictions on rate cutting and traffic imposed by the association, forging its own competitive pathway. The passage of the Interstate Commerce Act in 1887 eventually prohibited such pooling arrangements (see especially Chandler 1965: 159–81). 11. Fitchburg, Massachusetts, provides a revealing example. Unable to convince the leading meat wholesaler of Lowe and Sons to distribute his beef, Swift set up one of his first branch warehouses in Fitchburg and used this branch to undercut Lowe and supply the town with dressed beef. When Lowe finally capitulated, Swift hired him to manage the Fitchburg branch (Swift 1927, 70–71). 12. The firm of Botsford is instructive on this point about the diffusion of innovation. Henry Botsford, it is reported, sent a circular to the directors of his company in the early 1880s imploring them to consider establishing a branch house distribution system much like that of Swift, warning them of the consequences if they failed to heed his words. The directors, however, decided against it. Botsford’s prediction came true. Although one of the larger firms in 1878–80, Botsford, as well as others, disappeared from the top ranks of Chicago meatpackers by 1884 (Table 4.3). On the story of Botsford, see Clemen 1923, 409. 13. By 1900, Swift owned some foreign branch distribution houses, mostly in England and Germany, but it was really after this period when the company began to expand its foreign sales in a significant way.

Notes to Chapter 5 241 14. Swift and the other large packers did not want their slaughtering gangs to be kept idle by lack of supplies. Workers generally reported at 7 A.M. but would be sent home without pay if cattle did not reach the killing floors by 9:00 or 10:00 A.M. and told to return two or three hours later (Commons 1904). By 1904, the Amalgamated Meat Cutters and Butcher Workmen, after threatening to strike, successfully negotiated pay for hours spent waiting for cattle to arrive from the yards (Brody 1964, 13– 33). 15. Beginning in the 1890s, Swift moved toward the establishment of a professionally staffed full-time research organization. Its first research lab dates from 1892 and was headed by W. D. Richardson, who later became the first editor of the Journal of Industrial Engineering. 16. Although Swift had integrated all of these steps in the fresh beef trade, the boundaries of the firm stopped short of backward integration into the raising of the livestock. Few incentives existed for owning the supply of cattle. As noted above, the concentrated power of Swift and other large packers enabled them to control terms of trade with the thousands of livestock producers (Walker 1906, 492; Federal Trade Commission 1919, 83). Such power ensured that the system of cattle sale established at the stockyards worked to the benefit of Swift. Commission houses had effectively come under complete domination of the large packing firms and took control of cattle shipped to stockyards by grazers and feeders and then made the stock available to Swift and other large packers to the exclusion of smaller packers and cattle buyers (U.S. Senate 1890, 131). Such a system of sale complemented the “subsidies” given by stockyard companies to packing firms for their investments in the stockyards and provided Swift with animals at acceptable prices (Arnould 1971, 22). Consequently, Swift and other large packers chose to secure cattle supplies across a “market” rather than absorbing such transactions into their own corporate hierarchies. This market, however, resembled a market in name only. Cattle markets at stockyards were characterized more by control mechanisms based on power rather than the socalled invisible hand.

Chapter 5 1. One major difference, however, separates these two commerce systems. By the late 1870s, when Swift was just beginning to create his beef network, rails and telegraphy had already supplanted water and wagon conveyance as the dominant commerce system. In contrast, the Internet as a commerce system and platform for Dell is still in a stage of relative infancy. Nevertheless, Internet commerce, although a small percentage of overall business activity, is already transforming how firms produce, buy, and sell. 2. In addition to Schumpeter’s notion of invention and innovation (see Chapter 2), this paragraph borrows from Thomas Hughes’s model (1983, 7–17) in revealing how infrastructure systems emerge from invention and grow through a process of technology transfer and “momentum” against the impediments of “reverse salients.” 3. François Bar, Michael Borrus, and Richard Steinberg (1995) point out that the notions of interconnection and interoperability, though commonly used inter-

242 Notes to Chapter 5 changeably, actually differ. Interconnection is a prerequisite for interoperability and is binary—you are either connected or not—while interoperability can increase gradually. The idea of interconnection derives from older communications technologies such as the telegraph and telephone, where the connection mechanism was a wire. Interoperability, however, presupposes a more complex level of compatibility needed for different systems to work across complex interfaces. The more complete the compatibility, the greater the level of interoperability. 4. Paul David (1987) makes a distinction between “standards agreements” that are negotiated and implemented by political authority and “unsponsored standards” that arise de facto among firms themselves in competitive environments. While many of these unsponsored standards emerge as optimal solutions to specific technological problems, it is sometimes the case that suboptimal standards, resulting from the specifications of a technology successfully deployed by a first mover, persist into the future despite the existence of a superior alternative developed subsequently. The classic case is the QWERTY typewriter keyboard, which emerged initially to slow down the speeds of skilled typists so as to avoid key jams (David 1985). Despite the existence of a faster keyboard, the QWERTY standard prevailed. Once such newly deployed technologies become established, the standards embedded in these technologies create what is called the “lock-in” effect. Lock-in occurs because as standards diffuse and become established a shift to an alternative, even if it is superior, is costly. Such feedbacks promote path dependence as greater numbers of users, faced with the same choice of high switching costs, accommodate their activity to the existing standard (David 1985; Shapiro and Varian 1998, 1999). 5. Internet statistics are notoriously inconsistent and must be approached with caution. Different sources report widely different numbers. They are useful only in time series as trend indicators rather than absolute measures. 6. Numerous accounts of the history of the Internet’s development exist, among them Hafner and Lyon 1996, Abbate 1999, and Norberg and O’Neill 1996, but among the better short accounts written from a nontechnical perspective is Griffiths 2001. 7. I am indebted to François Bar for this insight. 8. Owing to these origins, the Internet was less the product of inventor-entrepreneurs and more of an institutionally driven technical breakthrough. Although certain individuals played key roles in the development of Internet technology, unlike the telegraph and the telephone, the Internet is not as easily traceable to individual inventor entrepreneurs such as Samuel Morse or Alexander Graham Bell. 9. This victory of TCP/IP as the standard language of inter-networking is one of the reasons why the United States emerged as the dominant force on the Internet (Rosenzweig 1998, 1537). 10. As part of this decision to split the ARPANET, the National Science Foundation established an Office of Advanced Computing in order to create centers of supercomputing throughout the nation. Only five centers, however, were actually funded. The locations for these centers included Cornell, Princeton, Pittsburgh, University of California (San Diego), and the University of Illinois (ChampaignUrbana). The center at the University of Illinois would later play a decisive role in

Notes to Chapter 6 243 helping to transform the Internet into an infrastructure for mass communications and commerce, for it was at this campus that the initial versions of the Mosaic Internet browser were developed. 11. There are still significant differences between U.S. and European approaches to issues of Internet taxation and privacy. 12. Although the 1996 Telecommunications Act and the Framework appeared after Internet commerce had already debuted, the content and policy direction of both had already been signaled by 1994 in policy forums throughout the world by U.S. government officials (U.S. Department of Commerce 2000). 13. It is also interesting to note that organized opposition to Wards, Sears, and other national mail-order houses emerged during the early 1900s on the part of rural retailers and wholesale jobbers during debate over legislation to extend parcel post service in rural America. See Chandler 1977, 230–33. 14. “We’re on the fulfillment end of [Internet Commerce],” insists Federal Express CEO Fred Smith. “When the telegraph came along, there was a corresponding development of the rail system. The telegraph created the connections and the railroad allowed fulfillment. Well, today the Internet creates the connections, and we provide the fulfillment” (quoted in Lappin 1996, 284, 286). 15. The dominance of interfirm activity in the economy is reflected in the transactions tables of input-output analysis pioneered by Wassily Leontieff that reveal the extent to which intermediate trade between firms dwarfs final demand. 16. “Every firm is a collection of activities that are performed to design, produce, market, deliver, and support its product or service. All of these activities can be represented using a value chain, . . . the activities in a firm’s value chain are linked to each other and to the activities of its suppliers, channels, and buyers” (Porter 1985, 34, 36). 17. Admittedly, Bar does not link communications to value chains exactly in this way. Nevertheless, his point—that market processes and market activities are information-processing activities—is the source of inspiration for the connection between value chains and communications made here. 18. Although this model (presented in Malone, Yates, and Benjamin 1994) describes the impacts of new information technology in general, it retains an enormous amount of fluency in representing the effects on firms of the Internet itself. 19. Indeed by early 2003, almost all of these exchanges and auction sites for interfirm commerce had become extinct (Business Week, May 12, 2003, 60).

Chapter 6 1. There is a vast literature supporting this claim. From a journalistic perspective, see Rocks 2000; Perman 2001; McWilliams 1997; Dodge 2000; and Business Week, August 28, 2000, 90. From the supply chain literature, see especially Lee 2000, and from a scholarly perspective, see Curry and Kenney 1999; Kenney and Curry 2000a, 2000b; Kraemer, Dedrick, and Yamashiro 1999; and Kraemer and Dedrick 2001. 2. It should be emphasized that others in the industry are far more caustic in making this same point. “Dell doesn’t make computers,” says Scott McNealy, CEO

244 Notes to Chapter 6 of Sun Microsystems. “They’re not in the PC business any more than Safeway is in the food manufacturing business” (quoted in Business Week, September 24, 2001). In this sense, Dell fits a paradigm described as “The Computerless Computer Company,” taking advantage of the technical capabilities of other firms (Rappaport and Halevi 1991). 3. I am indebted to Navi Radjou of Forrester Research for emphasizing this point to me. 4. It should be noted that these elements are constantly in flux. One Dell engineer, walking me through the factory floor of the Topfer Manufacturing Center, insisted that Dell is involved in an “ongoing learning process” to improve its pull-material-to-order business system. Consequently, some of the process details given here may have changed slightly. The emphasis is the thematic meaning of these elements. 5. The following two paragraphs rely on Dedrick and Kraemer 1998, 51–53, and Langlois 1992b. 6. There is an extensive literature on IBM’s decision to create an open architecture and the role of Intel and Microsoft in the evolution of the PC. Among the more informative, see Chposky and Leonis 1988 and Ferguson and Morris 1994, as well as Dedrick and Kraemer 1998 and Steffens 1994. 7. Taiwanese firms tended to be smaller start-ups. By 1979 these companies in Taiwan were already fabricating components for central processing units (CPUs) and had successfully produced Apple II clones by 1981 (Bae 1998, 148). Firms in Singapore and Korea, by contrast, tended to be larger. The industry in Singapore grew as a result of investment by American electronics producers, while the Korean industry emerged within the large Korean chaebol, which had benefited from the transfer of technology from earlier investment by Japanese electronics producers (Callon 1995; Bae 1998). 8. The development by Intel of a two-gigahertz chip provides a powerful example of this phenomenon. As soon as it released this chip in late August 2001 at a price of $562, Intel reduced the price of its existing 1.8-gigahertz chips, selling for $562 in July, to $256 (Gaither 2001b). Such examples have compelled Stan Shih, CEO of Acer Computer, to liken PCs to fresh vegetables, while Michael Dell refers to PC components as “having the shelf life of lettuce.” 9. Martin Kenney and James Curry (1999) refer to the more recent manifestations of this phenomenon in the PC industry, but the importance of speed and time was already established in the industry by the mid-1980s. 10. Information in the following three paragraphs is taken from Dell and Fredman 1999, 10–15. 11. This notion of Dell’s “second Web revolution” is taken from Rocks 2000. A similar idea is also expressed in Boulton, Libert, and Samek 2000, 8. 12. Information on burst capacity and its role in the Dell business system is taken from Albers 2000. 13. “The Internet at Dell extends from our customers to our suppliers,” observes Lance St. Clair. “The use of the Internet and e-commerce represent ultimate customer choice and the ability to scale and deliver those choices to the customer without inventory” (St. Clair, interviews, June 20, 2001, and February 10, 2002).

Notes to Chapter 6 245 14. Suppliers that maintain factory operations near Dell assembly sites do not use SLCs but are obligated to supply factory sites on the same replenishment schedules as the SLCs. The most notable supplier to Dell that does not participate in the SLC system is Intel, which uses its own logistics centers to ship to its customers. 15. In fiscal year 2001, when Dell’s total sales amounted to roughly $32 billion, its material costs accounted for $26 billion, which emphasizes how small the gross margins are in the PC business. 16. At the same time, however, this supplier also suggests that Dell’s innovations in supply chain management are being copied by other PC makers. While these firms do not execute as well as Dell, this supplier insists that Dell’s advantage in inventory management is narrowing (see “Diffusion” below). 17. The early notion that the Internet and Internet commerce would eliminate the role of nonmarket relationships in transactions and create more purely marketoriented interactions among firms is contradicted by the experience of Dell. The Dell case reveals that in its build-to-order business model relationships, not market transactions, are more critical than ever. 18. On cumulative causation, see Myrdal 1957 and Hirschman 1958, while on path dependence as it relates to industrial location, see especially Arthur 1988. The similarity in these theories derives from the fact that both attribute location decisions in the present to the agglomerations inherited from the past. 19. Employment figures for Dell Limerick were furnished to the author by representatives of the Shannon Development Corporation and the Limerick Development Corporation. 20. The Bayan Lepas Free Trade Zone in Penang emerged from the Free Trade Zone Act of 1971, while Xiamen was one of four Special Economic Zones (SEZs), along with Shenzhen, Shuhai, and Shantou, created by China in 1979. 21. In late 1999 Dell started to procure low-end components from Eastern Europe, produced by electronics firms operating in the Czech Republic and Hungary (Williams, interview, August 30, 2001). 22. Markusen makes an original contribution to product cycle theory by shifting the argument to declining profit margins as the product becomes standardized to account for spatial dispersion. This insight on “profit cycles” is particularly appropriate in describing the PC industry. 23. Michael Dell describes how the PC maker went to its suppliers insisting they develop “the capability to service Dell around the world.” According to Dell, this insistence worked. “A vendor who started with us in Ireland knew we were building a manufacturing center in Malaysia, so it set up a plant next to our plant in Penang and then another in China. When we decided to expand operations in Round Rock, Texas, the same company added a plant there. Next stop: Brazil” (quoted in Dell and Fredman 1999, 178). 24. The extent to which Dell is able to draw on supplies produced locally, however, varies in each location. In Penang, local suppliers provide roughly 70 percent of Dell’s component needs, while in Limerick and Eldorado the percentage is much lower, ranging from 25 to 50 percent. 25. It is worth noting that in contrast to Swift, Dell did not confront a rule-mak-

246 Notes to Chapter 7 ing environment and local government jurisdictions hostile to its expansionist ambitions in securing its market territory. Nor did Dell have to contend with entrenched and hostile local interests and actors, as did Swift in the form of local wholesale butchers seeking government protection from the Dressed Beef Trust. Instead, Dell has found a global free-trade rule-making regime at its disposal from the beginning, and local governments around the world more than willing to extend favors to it in exchange for locating assembly operations in their jurisdiction. Admittedly Dell participates in a wide array of political activities oriented around issues of free trade, tariffs, and taxation to ensure that the free flow of goods necessary to operate its global production and distribution network continues. The point is, however, that Dell, unlike Swift, did not have to dramatically alter the politics of markets to accomplish its objectives. Dell was able to take advantage of a legal and social environment of global market politics already in place, whereas Swift was one of the pioneers in forcing through a set of political rule changes to accommodate a new type of firm operating in a new type of market territory. 26. “Now Everyone in PCs Wants to Be Like Mike: Michael Dell, That Is” (Kirkpatrick 1997b). 27. While the phenomenon of Internet commerce was pushing firms throughout the economy to experiment with Internet sales, PC firms would not have been oblivious to Dell’s example. 28. “In my experience,” says Tony Ibarguen, president of distributor Tech Data Corporation, “we’ve never had a rallying cry like the one we’ve had in the past year on supply chain costs that was motivated by Dell’s success” (quoted in Aragon 1998).

Chapter 7 1. On this point, see especially Lazonick 1991a, 1991b, and 1994. 2. On this literature, see Hilton 1976 and Aston and Philpin 1985. 3. For an extremely informative overview of the debates, see John 2001.

References

Abbate, Janet. 1999. Inventing the Internet. Cambridge: MIT Press. Adhikari, Richard. 1998. The ERP-to-ERP Connection. Information Week. October 19: 12SS–18SS. Adler, Paul S. 1998. Market, Hierarchy, and Trust: The Knowledge Economy and the Future of Capitalism. Paper presented at the University of California–Berkeley Forum on Knowledge and the Firm. Aduddell, Robert, and Louis Cain. 1973. Location and Collusion in the Meat Packing Industry. In Business Enterprise and Economic Change: Essays in Honor of Harold F. Williamson, edited by Louis P. Cain and Paul J. Uselding. Kent, Ohio: Kent State University Press. ———. 1981. Public Policy Toward the “Greatest Trust in the World.” Business History Review 55: 217–42. Albers, Toni Lynn. 2000. Avoiding the Perils of Cyclic Growth in Dell’s Direct Environment. Master’s Thesis, MIT Sloan School of Management and Department of Mechanical Engineering. Albion, R. G. 1932. The Communication Revolution. American Historical Review 37(4): 718–20. Amsden, Alice. 1989. Asia’s Next Giant: South Korea and Late Industrialization. Oxford: Oxford University Press. Andersen, Esben Sloth. 1994. Evolutionary Economics: Post-Schumpeterian Contributions. London: Pinter. Anderson, David, and Hau Lee. 1998. Synchronized Supply Chains: The New Frontier. http://www.ascet.com/ascet/wp/wpAnderson.html. Andrews, Fred. 2000. Dell, It Turns Out, Has a Better Idea Than Ford. New York Times, January 26, p. C12. Angel, David P., and James Engstrom. 1995. Manufacturing Systems and Technological Change: The U.S. Personal Computer Industry. Economic Geography 17(1): 79– 102. Aragon, Lawrence. 1998. Dell’s Channel Secrets Revealed. VAR Business, no. 1415 (July 20). Arnould, Richard J. 1971. Changing Patterns of Concentration in American Meat Packing, 1880–1963. Business History Review 45(1): 18–34.

248 References Arrow, Kenneth J. 1962. The Economic Implications of Learning by Doing. Review of Economic Studies 29: 155–73. Arthur, W. Brian. 1988. Urban Systems and Historical Path Dependence. In Cities and Their Vital Systems, edited by Jesse H. Ausubel and Robert Herman, 85–97. Washington, D. C.: National Academy Press. Aston, T. H., and C. H. E. Philpin, eds. 1985. The Brenner Debate:Agrarian Class Structure and Economic Development in Pre-Industrial Europe. Cambridge: Cambridge University Press. Bae, Young-Ja. 1998. Globalization, Institutions, and Industry Development: A Case Study of the Personal Computer Industry in Korea and Taiwan. Ph.D. diss., University of North Carolina, Chapel Hill. Bar, François. 2001. The Construction of Marketplace Architecture. In Tracking a Transformation: E-Commerce and the Terms of Competition in Industries. BRIE–IGCC E-conomy Project. Washington, D.C.: Brookings Institution Press. Bar, Francois, et al. 1999. Defending the Internet Revolution in the Broadband Era: When Doing Nothing Is Doing Harm. Working Paper 137. Berkeley, Calif.: Berkeley Roundtable on the International Economy. Bar, Francois, Michael Borrus, and Richard Steinberg. 1995. Islands in the BitStream: Charting the NII Interoperability Debate. Working Paper 79. Berkeley, Calif.: Berkeley Roundtable on the International Economy. Barley, Steve, and Gideon Kunda. 2001. Bringing Work Back In. Organization Science 12(1): 76–95. Bell, Daniel. 1979. Communications Technology: For Better or for Worse. Harvard Business Review 57(3): 20–42. Beniger, James R. 1986. The Control Revolution:Technological and Economic Origins of the Information Society. Cambridge: Harvard University Press. Berk, Gerald. 1994. Alternative Tracks: The Constitution of American Industrial Order, 1865–1917. Baltimore: Johns Hopkins University Press. Berndt, Ernst, and Zvi Griliches. 1993. Price Indexes for Microcomputers: An Exploratory Study. In Price Measurements and Their Uses, edited by Marilyn Manser and Allan Young, 62–93. Chicago: University of Chicago Press. Best, Michael H. 1990. The New Competition: Institutions of Industrial Restructuring. Cambridge, England: Polity Press. Biederman, David. 1999. Plug and Play? No Way. Traffic World, May 10, 19–20. Bijker, Wiebe E., Thomas P. Hughes, and Trevor Pinch, eds. 1989. The Social Construction of Technological Systems. Cambridge: MIT Press. Borrus, Michael, Dieter Ernst, and Stephan Haggard, eds. 2000. International Production Networks in Asia. London: Routledge. Boulton, Richard E. S., Barry D. Libert, and Steve M. Samek. 2000. Cracking the Value Code: How Successful Businesses Are Creating Wealth in the New Economy. New York: HarperCollins. Braudel, Fernand. 1977. Afterthoughts on Material Civilization and Capitalism. Translated by Patricia Ranum. Baltimore: Johns Hopkins University Press. ———. 1979. The Wheels of Commerce: Civilization and Capitalism, 15th–18th Century. Vol. 2. Translated by Sian Reynolds. New York: Harper & Row.

References 249 ———.1980. On History. Translated by Sarah Matthews. Chicago: University of Chicago Press. Brenner, Robert. 1998. The Economics of Global Turbulence. New Left Review, no. 229 (May-June). Briody, Dan. 1997. Dell Follows Industry Standards, but Its Direct-Sales Model Leads. InfoWorld, August 25, pp. 29–30. Brody, David. 1964. The Butcher Workmen: A Study of Unionization. Cambridge: Harvard University Press. Brown, John Seely, and Paul Duguid. 2000. The Social Life of Information. Boston: Harvard Business School Press. Burawoy, Michael. 1979. Manufacturing Consent: Changes in the Labor Process Under Monopoly Capitalism. Chicago: University of Chicago Press. Business 2.0. September 1999. The Web’s Trillion-Dollar Secret: Business to Business E-commerce, pp. 84–124. Business America. January 1998. Various Definitions of Electronic Commerce. Vol. 119(1): 44. Business Week. June 22, 1998. Doing Business in the Internet Age, pp. 121–72. ———. December 7, 1998. A New Cyber Order, pp. 27–31. ———. February 22, 1999. The Internet Economy, pp. 30–32. ———. March 22, 1999. e.biz, pp. E6–E49. ———. September 27, 1999. The e.biz 25, pp. E6–E49. ———. June 21, 1999. Information Technology Annual Report, pp. 117–52. ———. October 4, 1999. The Internet Age, pp. 69–206. ———. January 17, 2000a. Business to Business: The Hottest Bet on the Net? pp. 36–38. ———. January 17, 2000b. Customizing Compaq, p. 44. ———. March 13, 2000. E-Malls for Business, pp. 32–34. ———. March 27, 2000. Such Busy Bees in B2B, p. 50. ———. August 28, 2000. The 21st Century Corporation, pp. 75–214. ———. September, 11, 2000. Why B2B Is a Scary Place to Be, pp. 34–37. ———. October 9, 2000. The Big Kahuna of B2B Exchanges? Glen Meakem’s FreeMarkets Could Be One of the Few Survivors of the Bloody Shakeout Ahead, pp. 166E2–166E4. ———.December 4, 2000. Let’s Keep This Exchange to Ourselves: The B2B Boom Moves from the General to the Company-Specific, p. 48. ———. September 24, 2001. Dell, the Conqueror: Now the King of Cutthroat Pricing Is Looking Beyond PCs, pp. 92–102. ———. May 13, 2002. How E-Biz Rose, Fell, and Will Rise Anew, pp. 64–72. ———. June 17, 2002. HP and Compaq: It’s Showtime, pp. 76–77. ———. June 24, 2002. Who Will Master the Server Biz? pp. 92–93. ———. August 12, 2002. Whose Lunch Will Dell Eat Next? pp. 66–67. ———. May 12, 2003. The E-Biz Surprise, pp. 60–68. Callon, Scott. 1995. Different Paths: The Rise of Taiwan and Singapore in the Global Personal Computer Industry. Working Paper 105. Stanford, Calif.: Stanford University Asia/Pacific Research Center.

250 References Carbone, James. 1997. Build-to-Order Model Changes PC Industry. Purchasing 123(6): 68. Carey, James W. 1988. Technology as Ideology: The Case of the Telegraph. In Communications as Culture. Boston: Unwin Hyman. Cassidy, William B. 1997. Electronic Commerce on the Internet Is Real—Just Ask Dell Computer, Corporate Express. Traffic World, June 30, pp. 20–24. Castells, Manuel. 1996. The Rise of the Network Society. Vol. 1 of The Information Age: Economy, Society, and Culture. London: Blackwell. ———. 1997. The Power of Identity. Vol. 2 of The Information Age: Economy, Society, and Culture. London: Blackwell. ———. 2000. The Rise of the Network Society. 2d ed. London: Blackwell. Catephores, George. 1994. The Imperious Austrian: Schumpeter as Bourgeois Marxist. New Left Review no. 205: 3–30. Ceruzzi, Paul E. 1999. A History of Modern Computing. Cambridge: MIT Press. Chandler, Alfred Dupont, Jr. 1962. Strategy and Structure: Chapters in the History of the Industrial Enterprise. Cambridge: MIT Press. ———. 1965. The Railroads: The Nation’s First Big Business. New York: Harcourt, Brace & World. ———. 1977. The Visible Hand:The Managerial Revolution in American Business. Cambridge: Harvard University Press, Belknap Press. ———. 1980. The United States: Seedbed of Managerial Capitalism. In Managerial Hierarchies: Comparative Perspectives on the Rise of the Modern Industrial Corporation, edited by Alfred D. Chandler and Herman Daems. Cambridge: Harvard University Press. ———. 1988. The Essential Alfred Chandler: Essays Toward a Historical Theory of Big Business. Edited by Thomas K. McCraw. Boston: Harvard Business School Press. ———. 1992. Corporate Strategy, Structure, and Control Methods in the United States during the Twentieth Century. Industrial and Corporate Change 1(2): 263–84. ———. 2001. Inventing the Electronic Century:The Epic Story of the Consumer Electronics and Computer Industries. New York: The Free Press. Chandler, Alfred D., Jr., and James W. Cortada. 2000. A Nation Transformed by Information: How Information Has Shaped the United States from Colonial Times to the Present. Oxford: Oxford University Press. Chicago Board of Trade. 1864. Sixth Annual Statement of the Trade and Commerce of Chicago for the Year Ending March 31, 1864. Chicago: Tribune Company. ———. 1880. Twenty-first Annual Report of the Trade and Commerce for the Year Ending December 31, 1879. Chicago: Knight & Leonard. ———. 1881. Twenty-third Annual Report of the Trade and Commerce for the Year Ending December 31, 1880. Chicago: Knight & Leonard. ———. 1885. Twenty-sixth Annual Report of the Trade and Commerce for the Year Ending December 31, 1884. Chicago: Knight & Leonard. Chopra, Sunil, and Jan Van Mieghem. 2000. Which E-Business Is Right for Your Supply Chain? Supply Chain Management Review (July–August), 32–40. Chposky, James, and Ted Leonis. 1988. Blue Magic:The People, Power, and Politics Behind the IBM Personal Computer. New York: Facts on File.

References 251 Christaller, Walter. 1933. Central Places in Southern Germany. Translated by Carlisle W. Baskin. Englewood Cliffs, N.J.: Prentice-Hall, 1966. Christensen, Clayton M. 1997. The Innovator’s Dilemma:When New Technologies Cause Great Firms to Fail. Boston: Harvard Business School Press. Christopherson, Susan. 1993. Market Rules and Territorial Outcomes: The Case of the United States. International Journal of Urban and Regional Research 17(3): 274– 88. ———. 1999. Rules as Resources: How Market Governance Regimes Influence Firm Networks. In The New Industrial Geography: Regions, Regulation, and Institutions, edited by Trevor J. Barnes and Meric Gertler, 155–75. London: Routledge. Clemen, Rudolf Alexander. 1923. The American Livestock and Meat Industry. New York: Ronald Press. ———. 1928. The American Meat-Packing Industry. In Representative Industries in the United States, edited by H. T. Warhow. New York: Henry Holt. Coase, R. H. 1937. The Nature of the Firm. Economica 4(16): 386–405. ———. 1991. The Nature of the Firm: Meaning. In The Nature of the Firm: Origins, Evolution, and Development, edited by Oliver Williamson and Sidney Winter. New York: Oxford University Press. Cockburn, Alexander. 1996. A Short Meat-Oriented History of the World: From Eden to the Mattole. New Left Review, no. 215: 16–42. Cohen, Stephen S., J. Bradford DeLong, and John Zysman. 2000. Tools for Thought: What Is New and Important About the “Economy.” Berkeley, Calif.: Berkeley Roundtable on the International Economy. Cole, Arthur H. 1949. Entrepreneurship and Entrepreneurial History. In Change and the Entrepreneur, 85–107. Cambridge: Harvard University Press. Commons, John R. 1904. Labor Conditions in Meat Packing and the Recent Strike. Quarterly Journal of Economics 19(1):1–32. Cronon, William. 1991. Nature’s Metropolis: Chicago and the Great West. New York: Norton. Curry, James, and Martin Kenney. 1999. Beating the Clock: Corporate Responses to Rapid Change in the PC Industry. California Management Review 42(1): 8–36. David, Paul A. 1985. Clio and the Economics of QWERTY. American Economic Review, Proceedings 75: 332–37. ———. 1987. Some New Standards for the Economics of Standardization in the Information Age. In The Economic Theory of Technology Policy, edited by Partha Dasgupta and P. L. Stoneman. London: Cambridge University Press. ———. 1991. Computer and Dynamo: The Modern Productivity Paradox in a Not-Too-Distant Mirror. In Technology and Productivity:The Challenge for Economic Policy, 315–47. Paris: OECD. David, Paul A., and Julie Ann Bunn. 1987. The “Battle of the Systems” and the Evolutionary Dynamics of Network Technology Rivalries. Working Paper 15. Stanford, Calif.: Stanford University Center for Economic Policy Research. David, Paul A., and Shane Greenstein. 1990. The Economics of Compatibility Standards: An Introduction to Recent Research. Economic Innovation and New Technology 1(1): 3–41.

252 References Dedrick, Jason, and Kenneth L. Kraemer. 1998. Asia’s Computer Challenge:Threat or Opportunity for the United States and the World? Oxford: Oxford University Press. Dell, Michael, and Catherine Fredman. 1999. Direct from Dell: Strategies That Revolutionized an Industry. New York: HarperCollins. Destler, Chester McArthur. 1946. Entrepreneurial Leadership Among the “Robber Barons”: A Trial Balance. The Tasks of Economic History 6 (supplement): 28–49. De Vries, Jan. 1976. The Economy of Europe in an Age of Crisis, 1600–1750. Cambridge: Cambridge University Press. Dodge, John. 2000. Dell’s Internet-Based Plant Keeps Production Efficient. Wall Street Journal.com, September 26. http://cob.jmu.edu/williakc/mktg470/dailyupdate/netbasedplant.htm. Dodge, Martin. 2001. Finding the Source of Amazon.com: Examining the Store with “Earth’s Biggest Selection.” In Worlds of E-commerce: Economic, Geographical, and Social Dimensions, edited by Thomas R. Leinbach and Stanley D. Brunn. West Sussex: Wiley. Dosi, Giovanni. 1982. Technological Paradigms and Technological Trajectories: A Suggested Interpretation of the Determinants and Directions of Technical Change. Research Policy 11: 147–62. ———. 1984. Technical Change and Industrial Transformation. New York: St. Martin’s Press. ———. 1997. Opportunities, Incentives, and the Collective Patterns of Technological Change. Economic Journal 107: 1530–47. Dosi, Giovanni, Christopher Freeman, Richard Nelson, Gerald Silverberg, and Luc Soete, eds. 1988. Technical Change and Economic Theory. London: Pinter. Dosi, Giovanni, Renato Giannetti, and Pier Angelo Toninelli, eds. 1992. Technology and Enterprise in Historical Perspective. Oxford: Clarendon Press. Dosi, Giovanni, David J. Teece, and Sidney Winter. 1992. Toward a Theory of Corporate Coherence: Preliminary Remarks. In Technology and Enterprise in Historical Perspective, edited by Giovanni Dosi, Renato Giannetti, and Pier Angelo Toninelli. Oxford: Clarendon Press. Dosi, Giovanni, David J. Teece, and Josef Chytry, eds. 1998. Technology, Organization, and Competitiveness: Perspectives on Industrial and Corporate Change. Oxford: Oxford University Press. Drucker, Peter F. 1999. Beyond the Information Revolution. Atlantic Monthly, October, 47–57. Du Boff, Richard B. 1980. Business Demand and the Development of the Telegraph in the United States, 1844–1860. Business History Review 54(4): 459–79. ———. 1983. The Telegraph and the Structure of Markets and Firms. Research in Economic History 8: 253–77. ———. 1984. The Telegraph in Nineteenth-Century America: Technology and Monopoly. Comparative Studies in Society and History 26: 571–86. Du Boff, Richard B., and Edward S. Herman. 1980. Alfred Chandler’s New Business History: A Review. Politics and Society 10(1): 87–112. Durkheim, Emile. 1893. The Division of Labor in Society. Translated by W. D. Halls. New York: The Free Press, 1984.

References 253 Economist. June 26, 1999. The Net Imperative, pp. 1–40. ———. March 2, 1991. Dell Computer: It’s in the Mail. Eisenstein, Elizabeth L. 1979. The Printing Press as an Agent of Change: Communications and Cultural Transformations in Early-Modern Europe. Cambridge: Cambridge University Press. eMarketer. 2001. The e-Commerce B2B Report. New York. Federal Trade Commission. 1919. Report of the Federal Trade Commission on the MeatPacking Industry. 6 parts. Washington, D.C.: Government Printing Office. Ferguson, Charles H., and Charles R. Morris. 1994. Computer Wars:The Fall of IBM and the Future of Global Technology. New York: Times Books. Field, Alexander James. 1992. The Magnetic Telegraph, Price and Quantity Data, and the New Management of Capital. Journal of Economic History 52(2): 401–13. Fields, Gary. 2003. Communications, Innovation, and Territory: The Production Network of G. F. Swift and the Creation of a National Market in the U.S. Journal of Historical Geography 29(3). Fischer, Claude S. 1992. America Calling: A Social History of the Telephone to 1940. Berkeley: University of California Press. Fisher, Daniel. 2002. Pulled in a New Direction. Forbes, June 10, pp. 102–12. Fishlow, Albert. 1965. American Railroads and the Transformation of the Ante-Bellum Economy. Cambridge: Harvard University Press. Fligstein, Neil. 1990. The Transformation of Corporate Control. Cambridge: Harvard University Press. Fogel, Robert William. 1964. Railroads and American Economic Growth: Essays in Econometric History. Baltimore: Johns Hopkins Press. Freeman, Christopher. 1982. Economics of Innovation. London: Pinter. ———. 1990. Schumpeter’s Business Cycles Revisited. In Evolving Technology and Market Structure: Studies in Schumpeterian Economics, edited by Arnold Heertje and Mark Perlman, 17–38. Ann Arbor: University of Michigan Press. ———. 1991. The Nature of Innovation and the Evolution of the Productive System. In Technology and Productivity:The Challenge for Economic Policy, 303–14. Paris: OECD. ———. 1994. The Economics of Technical Change. Cambridge Journal of Economics 18: 463–514. Frickey, Edwin. 1947. Production in the United States, 1860–1914. Cambridge: Harvard University Press. Friedlander, Amy. 1995. Emerging Infrastructure:The Growth of Railroads. Reston, Va.: Corporation for National Research Initiatives. Gaither, Chris. 2001a. Despite Strong Sales, Dell Warns of Lower-Than-Expected Earnings. New York Times, January 23, pp. C1, C5. ———. 2001b. Intel Selling Two-Gigahertz Chips. New York Times, August 28, p. C4. Gereffi, Gary, and Miguel Korzeniewicz, eds. 1994. Commodity Chains and Global Capitalism. Westport, Conn.: Praeger. Gerlach, Michael. 1989. Keiretsu Organization in the Japanese Economy: Analysis and Trade Implications. In Politics and Productivity, edited by Chalmers Johnson, Laura D’Andrea Tyson, and John Zysman. New York: Ballinger.

254 References Gertler, Meric S. 1995. “Being There”: Proximity, Organization, and Culture in the Development and Adoption of Advanced Manufacturing Technologies. Economic Geography 71: 1–26. Giddens, Anthony. 1984. The Constitution of Society: Outline of the Theory of Structuration. Berkeley: University of California Press. Goldstein, Alan. 1997. Rivals Seek to Mimic Dell’s Success. Dallas Morning News, August 31. Gordon, David. 1984. Swift & Co. v. United States: The Beef Trust and the Stream of Commerce Doctrine. American Journal of Legal History 28 (July): 244–79. Granovetter, Marc. 1985. Economic Action and Social Structure: The Problem of Embeddedness. American Journal of Sociology 91: 481–510. ———. 1998. Coase Revisited: Business Groups in the Modern Economy. In Technology, Organization, and Competitiveness: Perspectives on Industrial and Corporate Change, edited by Giovanni Dosi, David J. Teece, and Josef Chytry. Oxford: Oxford University Press. Griffiths, Richard T. 2001. History of the Internet, Internet for Historians (and Just About Everyone Else). University of Leiden, Netherlands. http://www.let.leidenuniv.nl/history/ivh/frame_theorie.html. Hafner, Katie, and Matthew Lyon. 1996. Where Wizards Stay Up Late:The Origins of the Internet. New York: Simon & Schuster. Haggard, Stephen, Lim Pao Li, and Anna Ong. 1998. The Hard Disk Drive Industry in the Northern Region of Malaysia. Report 98–04. University of California, San Diego, Information Storage Industry Center. Hall, Peter, and Paschal Preston. 1988. The Carrier Wave: New Information Technology and the Geography of Innovation, 1846–2003. London: Unwin Hyman. Hamilton, Gary G., and Robert C. Feenstra. 1998. Varieties of Hierarchies and Markets: An Introduction. In Technology, Organization, and Competitiveness, edited by Giovanni Dosi, David J. Teece, and Josef Chytry. Oxford: Oxford University Press. Harper’s Weekly. October 21, 1882. Cheaper Beef, p. 663. Harvey, David. 1996. Justice, Nature, and the Geography of Difference. Oxford: Blackwell. Harrison, Bennett. 1994. Lean and Mean:The Changing Landscape of Corporate Power in the Age of Flexibility. New York: Basic Books. Hayes, Simon, and Byron Connolly. 1998. Compaq Takes on Direct Business in Response to Corporate Demand. Computer Reseller News. www.itnews.com.au/crn/ news/007-la.htm. Henderson, Rebecca M., and Kim B. Clark. 1990. Architectural Innovation: The Reconfiguration of Product Technologies and the Failure of Established Firms. Administrative Science Quarterly 35: 9–30. Hill, Howard Copeland. 1925. The Development of Chicago as a Center of the Meat Packing Industry. Mississippi Valley Historical Review 10(3): 253–73. Hilton, Rodney, ed. 1976. The Transition from Feudalism to Capitalism. London: Verso Press. Hirschman, Albert O. 1958. The Strategy of Economic Development. New Haven, Conn.:Yale University Press.

References 255 Hopkins, Terence K., and Immanuel Wallerstein. 1994. Commodity Chains in the Capitalist World-Economy Prior to 1800. In Commodity Chains and Global Capitalism, edited by Gary Gereffi and Miguel Korzeniewicz, 17–50. Westport, Conn.: Praeger. Horowitz, Roger. 1997. “Negro and White, Unite and Fight!”A Social History of Industrial Unionism in Meatpacking, 1930–90. Urbana: University of Illinois Press. Horwitz, Robert Britt. 1989. The Irony of Regulatory Reform:The Deregulation of American Telecommunications. Oxford: Oxford University Press. Hounshell, David A. 1984. From the American System to Mass Production, 1800–1932:The Development of Manufacturing Technology in the United States. Baltimore: Johns Hopkins Press. ———. 1995. Hughesian History of Technology and Chandlerian Business History: Parallels, Departures, and Critics. History and Technology 12: 205–24. Hughes, Thomas P. 1983. Networks of Power: Electrification in Western Society, 1880–1930. Baltimore: Johns Hopkins University Press. Hunt, Courtney Shelton, and Howard E. Aldrich. 1998. The Second Ecology: Creation and Evolution of Organizational Communities. Research in Organizational Behavior 20: 267–301. Innis, Harold Adams. 1950. Empire and Communications. Oxford: Clarendon Press. Irwin, Michael D., and John D. Kasarda. 1994. Trade, Transportation, and Spatial Distribution. In The Handbook of Economic Sociology, edited by Neil J. Smelser and Richard Swedberg. Princeton, N.J.: Princeton University Press. John, Richard R. 1994. American Historians and the Concept of the Communications Revolution. In Information Acumen:The Understanding and Use of Knowledge in American Business, edited by Lisa Bud-Frierman, 98–110. London: Routledge. ———. 1997. Elaborations, Revisions, Dissents: Alfred D. Chandler Jr.’s. The Visible Hand After Twenty Years. Business History Review 71: 151–200. ———. 1998. The Politics of Innovation. Daedalus 127(4): 187–213. ———. 2001. Rendezvous with Information? Computers and Communications Networks in the United States. Business History Review 75(1): 1–13. Jones, Alexander. 1852. Historical Sketch of the Electric Telegraph: Rise and Progress in the United States. New York: George P. Putnam. Jue, Jaime. 1998. Price Convergence and Interregional Arbitrage Opportunities in the Postbellum United States: Evidence from the U.S. Department of War. Unpublished paper. University of California, Berkeley, Department of Economics. Kahaner, D. K. 1996. Technology and Industry in Penang, Malaysia. Asian Technology Information Program (ATIP). www.cs.arizona.edu/japan/atip/public/ atip.reports.96/atip96.107r.html. Kanellos, Michael. 1997. PC Vendors Take the Direct Challenge. CNET News.com, June 20. http://news.cnet.com/news/0–1003–202–319914.html. Kenney, Martin. 2001. The Growth and Development of the Internet in the United States. Working Paper 145. Berkeley, Calif.: Berkeley Roundtable on the International Economy. Kenney, Martin, and James Curry. 2000a. We All Want to Be Like Mike: The Internet and the Personal Computer Value Chain. Paper presented to Workshop on E-

256 References commerce and Changing Terms of Competition: A View from the Sectors. Berkeley Roundtable on the International Economy (BRIE). April 28. ———. 2000b. The Organization of the Personal Computer Value Chain. A Proposal to the Alfred Sloan Foundation. ——. 2001. Beyond Transaction Costs: E-Commerce and the Power of the Internet Dataspace. Worlds of E-Commerce: Economic, Geographical, and Social Dimensions, edited by Thomas R. Leinbach and Stanley D. Brunn. West Sussex, England: Wiley. Kern, Stephen. 1983. The Culture of Time and Space, 1880–1918. Cambridge: Harvard University Press. King, Julia. 1997. Dell Zaps SAP. Computerworld, July. http://www.computerworld.com/cwi/story/0,1199,NAV47_STO5285,00.html. Kirkpatrick, David. 1997a. Why Compaq Envies Dell. Fortune, February 17, p. 26. ———. 1997b. Now Everyone in PCs Wants to Be Like Mike: Michael Dell, That Is. Fortune, September 8, pp. 91–92. Kline, Stephen, and Nathan Rosenberg. 1986. An Overview of Innovation. In The Positive Sum Strategy: Harnessing Technology for Economic Growth, edited by Ralph Landau and Nathan Rosenberg, 275–306. Washington, D.C.: National Academy Press. Knight, F. H. 1921. Risk, Uncertainty, and Profit. New York: Augustus Kelly. Kogut, Bruce. 1992. National Organizing Principles of Work and the Erstwhile Dominance of the American Multinational Corporation. Industrial and Corporate Change 1(2): 285–325. Kondratiev, Nikolai. 1935. The Long Waves in Economic Life. The Review of Economics Statistics 17: 105–15. Kraemer, Kenneth L., and Jason Dedrick. 2001. Dell Computer: Using E-Commerce to Support the Virtual Company. Irvine: University of California, Irvine, Center for Research on Information Technology and Organizations. Kraemer, Kenneth L., Jason Dedrick, and Sandra Yamashiro. 1999. Refining and Extending the Business Model with Information Technology: Dell Computer Corporation. Irvine: University of California, Irvine, Center for Research on Information Technology and Organizations. Kujovich, Mary Yeager. 1970. The Refrigerator Car and the Growth of the American Dressed Beef Industry. Business History Review 44(4): 460–82. Kuznets, Simon. 1930. Secular Movements in Production and Prices:Their Nature and Their Bearing upon Cyclical Fluctuations. New York: A. M. Kelley, 1967. Lamoreaux, Naomi R., Daniel M. G. Raff, and Peter Temin, eds. 1999. Learning by Doing in Markets, Firms, and Countries. Chicago: University of Chicago Press. Langlois, Richard N. 1990. Creating External Capabilities: Innovation and Vertical Integration in the Microcomputer Industry. Business and Economic History 19: 93–102. ———. 1992a. Transaction-Cost Economics in Real Time. Industrial and Corporate Change 1(1): 99–127. ———. 1992b. External Economies and Economic Progress: The Case of the Microcomputer Industry. Business History Review 66(1): 1–50. ———. 2001. The Vanishing Hand: The Modular Revolution in American Business. Unpublished Paper. University of Connecticut.

References 257 ———. 2003. The Vanishing Hand: The Changing Dynamics of Industrial Capitalism. Industrial and Corporate Change 12(2): 351–85. Langlois, Richard N., and Paul Robertson. 1992. Networks and Innovation in a Modular System: Lessons from the Microcomputer and Stereo Component Industries. Research Policy 21: 297–313. Lappin, Todd. 1996. The Airline of the Internet. Wired 4 (December). Lazonick, William. 1990. Organizational Integration in Three Industrial Revolutions. In Evolving Technology and Market Structure: Studies in Schumpeterian Economics, edited by Arnold Heertje and Mark Perlman, 77–97. Ann Arbor: University of Michigan Press. ———. 1991a. Business Organization and the Myth of the Market Economy. Cambridge: Cambridge University Press. ———. 1991b. What Happened to the Theory of Economic Development? In Favorites of Fortune:Technology, Growth, and Economic Development since the Industrial Revolution, edited by Patrice Higonnet, David S. Landes, and Henry Rosovsky. Cambridge: Harvard University Press. ———. 1994. The Integration of Theory and History: Methodology and Ideology in Schumpeter’s Economics. In Evolutionary and Neo-Schumpeterian Approaches to Economics, edited by Lars Magnusson. Boston: Kluwer Academic Publishers. ———. 2002. Innovative Enterprise and Historical Transformation. Enterprise and Society 3(1): 3–47. ———. 2003. The Theory of the Market Economy and the Social Foundations of Innovative Enterprise. Economic and Industrial Democracy 24(1): 9–44. Lazonick, William, and William Mass, eds. 1995. Organizational Capabilities and Competitive Advantage. Brookfield, Vt.: Elgar. Lee, Hau L. 2000. Creating Value Through Supply Chain Integration. Supply Chain Management Review (September/October). Lewis, Geoff. 1987. The Hottest Little Computer Maker in Texas: PC’s Ltd. Started in a Dorm. Now Its Clones Have Hit It Big. BusinessWeek, February 2, pp. 71–72. Lloyd, Henry Demarest. 1900. Wealth Against Commonwealth. New York: Harper and Brothers. Loasby, Brian J. 2001. Industrial Dynamics: Why Connections Matter. Paper prepared for DRUID Academy Winter Conference, Copenhagen. Lohr, Steve. 1997. Beyond Consumers, Companies Pursue Business-to-Business Net Commerce. New York Times, April 28, p. D1. ———. 1999. Computer Age Gains Respect of Economists. New York Times, April 14, pp. A1, C14. ———. 2002. On a Roll, Dell Enters Uncharted Territory. NewYork Times, August 25. Lohr, Steve, and John Markoff. 1998. Conquering the Internet. New York Times, November 24, pp. A1, C4. Lundvall, Bengt-Ake. 1988. Innovation as an Interactive Process: From User-Producer Interaction to the National System of Innovation. In Technical Change and Economic Theory, edited Giovanni Dosi, Christopher Freeman, Richard Nelson, Gerald Silverberg, and Luc Soete. London: Pinter.

258 References Maglitta, Joseph E. 1997. Special Dell-ivery. Electronic Business, December, 43–47. Magretta, Joan. 1998. The Power of Virtual Integration: An Interview with Dell Computer’s Michael Dell. Harvard Business Review 76(2): 72–84. Malone, Thomas W., JoAnne Yates, and Robert I. Benjamin. 1994. Electronic Markets and Electronic Hierarchies. In Information Technology and the Corporation of the 1990s, edited by Thomas J. Allen and Michael S. Scott Morton, 61–83. New York: Oxford University Press. Marglin, Stephen A. 1974. What Do Bosses Do? The Origins and Functions of Hierarchy in Capitalist Production. Review of Radical Political Economy 6: 33–60. Markusen, Ann Roell. 1985. Profit Cycles, Oligopoly, and Regional Development. Cambridge: MIT Press. ———. 1994. Studying Regions by Studying Firms. Professional Geographer 46(4): 477–90. ———. 1996. Sticky Places in Slippery Space: A Typology of Industrial Districts. Economic Geography 72: 293–313. Markusen, Ann R., Yong-Sook Lee, and Sean DiGiovanna, eds. 1999. Second-Tier Cities: Rapid Growth Beyond the Metropolis. Minneapolis: University of Minnesota Press. Marshall, Alfred. 1890. Principles of Economics. 2 vols. London: Macmillan, 1961. ———. 1919. Industry and Trade. London: Macmillan, 1921. Marx, Karl. 1847. The Poverty of Philosophy. In Karl Marx: Selected Writings, edited by David McLellan, 195–215. Oxford: Oxford University Press, 1977. ———. 1848. Manifesto of the Communist Party. In Selected Works in Three Volumes, by Karl Marx and Frederick Engels. Vol. 1: 98–137. Moscow: Progress Publishers, 1969. ———. 1851. The Eighteenth Brumaire of Louis Bonaparte. In Selected Works in Three Volumes, by Karl Marx and Frederick Engels. Vol. 1:398–487. Moscow: Progress Publishers, 1969. ———. 1857. Grundrisse. In Karl Marx: Selected Writings, edited by David McLellan, 343–87. Oxford: Oxford University Press, 1977. ———. 1859. Preface to A Contribution to the Critique of Political Economy. In Selected Works in Three Volumes, by Karl Marx and Frederick Engels. Vol. 1:502–6. Moscow: Progress Publishers, 1969. ———. 1867. Capital. Vol. 1. London: Lawrence & Wishart, 1974. ———. 1885. Capital. Vol. 2. London: Lawrence & Wishart, 1974. ———. 1894. Capital. Vol. 3. London: Lawrence & Wishart, 1974. Mayntz, Renate, and Thomas P. Hughes, eds. 1988. The Development of Large Technical Systems. Boulder, Colo.: Westview Press. McChesney, Robert W., Ellen Meiksins Wood, and John Bellamy Foster. 1998. Capitalism and the Information Age:The Political Economy of the Global Communication Revolution. New York: Monthly Review Press. McCurdy, Charles W. 1978. American Law and the Marketing Structure of the Large Corporation, 1875–1890. Journal of Economic History 38(3): 631–49. McNeill, William H. 2000. A Short History of Humanity. New York Review of Books 47(11): 9–11.

References 259 McWilliams, Gary. 1997. Michael Dell: Whirlwind on the Web. Business Week, April 7, pp. 132–36. McWilliams, Gary, and Joseph B. White. 1999. Dell to Detroit: Get into Gear Online. Wall Street Journal, December 1, pp. B1, B4. Meier, Richard L. 1962. A Communications Theory of Urban Growth. Cambridge: MIT Press. Mercador, Nicholas. 2001. Never Underestimate the Power of a Business Model. Business Model Insider, June 27. http://www.businessmodelx.com/cgi-bin/inclusions.cgi?id=63&type=1. Metcalfe, J. Stanley. 1998. Evolutionary Economics and Creative Destruction. London: Routledge. Minahan, Tim. 1997. Dell Computer Sees Suppliers as Key to JIT. Purchasing 123(3): 43–48. Mokyr, Joel. 1990. The Lever of Riches:Technological Creativity and Economic Progress. Oxford: Oxford University Press. Moschella, David C. 1997. Waves of Power: Dynamics of Global Technology Leadership, 1964–2010. New York: AMACOM. Mount, Ian, and Brian Caulfield. 2001. The Missing Link: What You Need to Know About Supply-Chain Technology. Business 2.0, May. Mulgan, G. J. 1991. Communication and Control: Networks and the New Economies of Communication. New York: Guilford Press. Myrdal, Gunnar. 1957. Economic Theory and Underdeveloped Regions. London: Duckworth. Nelson, Richard R. 1998. The Co-evolution of Technology, Industrial Structure, and Supporting Institutions. In Technology, Organization, and Competitiveness, edited by Giovanni Dosi, David J. Teece, and Josef Chytry. Oxford: Oxford University Press. Nelson, Richard R., and Sidney G. Winter. 1982. An Evolutionary Theory of Economic Change. Cambridge: Harvard University Press, Belknap Press. Neyhart, Louise A. 1952. Giant of the Yards. Boston: Houghton Mifflin. Noble, David F. 1984. Forces of Production: A Social History of Industrial Automation. New York: Knopf. Norberg, Arthur L., and Judy E. O’Neill. 1996. Transforming Computer Technology: Information Processing for the Pentagon, 1962–1986. Baltimore: John Hopkins University Press. Norris, Frank. 1901. The Octopus. New York: Bantam Books, 1971. North, Douglass C., Terry L. Anderson, and Peter J. Hill. 1983. Growth and Welfare in the American Past:A New Economic History. Englewood Cliffs, N.J.: Prentice-Hall. Ohlin, Bertil. 1933. Interregional and International Trade. Cambridge: Harvard University Press, 1967. O’Sullivan, Mary. 2000. The Innovative Enterprise and Corporate Governance. Cambridge Journal of Economics 24(4): 393–416. Page, Brian, and Richard Walker. 1991. From Settlement to Fordism: The Agro-Industrial Revolution in the American Midwest. Economic Geography 67(4): 281–315. Parrini, Carl P., and Martin J. Sklar. 1983. New Thinking About the Market, 1896–

260 References 1904: Some American Economists on Investment and the Theory of Surplus Capital. Journal of Economic History 43(3): 560–78. Pavitt, Keith. 1992. Some Foundations for a Theory of the Large Innovating Firm. In Technology and Enterprise in Historical Perspective, edited by Giovanni Dosi, Renato Giannetti, and Pier Angelo Toninelli. Oxford: Clarendon Press. Penrose, Edith. 1959. The Theory of the Growth of the Firm. 3d ed. Oxford: Oxford University Press, 1995. Perez, Carlota. 1983. Structural Change and Assimilation of New Technologies in the Economic and Social Systems. Futures, 357–75. Perloff, Harvey S., Edgar S. Dunn Jr., Eric E. Lampard, and Richard F. Muth. 1960. Regions, Resources, and Economic Growth. Lincoln: University of Nebraska Press. Perman, Stacy. 2001. Automate or Die: . . . Dell Computer Won’t Stop Remaking Its Business. Its Three-Part Strategy for Ultimate Victory? Web, Web, and Web. Business 2.0, July. Picot, Arnold, Christine Bortenlanger, and Heiner Rohrl. 1997. Organization of Electronic Markets: Contributions from the New Institutional Economics. The Information Society 13: 107–23. Pinch, Trevor J., and Wiebe E. Bijker. 1989. The Social Construction of Facts and Artifacts: Or How the Sociology of Science and the Sociology of Technology Might Benefit Each Other. In The Social Construction of Technological Systems, edited by Wiebe E. Bijker, Thomas P. Hughes, and Trevor Pinch, 17–50. Cambridge: MIT Press. Pinella, Paul. 1991. The PC Big 50 1990. Datamation, December 1, 42–43. Piore, Michael J., and Charles Sabel. 1984. The Second Industrial Divide. New York: Basic Books. Pirenne, Henri. 1925. Medieval Cities:Their Origins and the Revival of Trade. Princeton, N.J.: Princeton University Press, 1974. Poirier, Charles C., and Michael J. Bauer. 2000. E-Supply Chain: Using the Internet to Revolutionize Your Business. San Francisco: Berrett-Koehler. Polanyi, Karl. 1944. The Great Transformation. Boston: Beacon Press, 1957. Porter, Glenn, and Harold Livesay. 1971. Merchants and Manufacturers: Studies in the Changing Structure of Nineteenth-Century Marketing. Baltimore: Johns Hopkins University Press. Porter, Michael. 1985. Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press. ———. 1990. The Competitive Advantage of Nations. New York: Free Press. Powell, Walter W. 1990. Neither Market nor Hierarchy: Network Forms of Organization. Research in Organizational Behavior 12: 295–336. Pred, Allan R. 1966. The Spatial Dynamics of U.S. Urban-Industrial Growth, 1800–1914: Interpretive and Theoretical Essays. Cambridge: M.I.T. Press. ———. 1973. Urban Growth and the Circulation of Information:The United States System of Cities, 1790–1840. Cambridge: Harvard University Press. ———. 1977. City Systems in Advanced Economies. New York: Wiley. Putnam, George E. 1921. Joint Cost in the Packing Industry. Journal of Political Economy 29(4): 293–303.

References 261 Pyke, Frank, Giacomo Becattini, and Werner Sengenberger. 1990. Industrial Districts and Inter-Firm Cooperation in Italy. Geneva: International Institute for Labour Studies. Raff, Daniel M. G., and Peter Temin. 1991. Business History and Recent Economic Theory: Imperfect Information, Incentives, and the Internal Organization of Firms. In Inside the Business Enterprise: Historical Perspectives on the Use of Information, edited by Peter Temin. Chicago: University of Chicago Press. Raj, Winston. 2002. Brothers in Arms: Compaq’s Made-in-Asia Collaboration Project Tightens the Fragile Bonds Between Its Suppliers, Distributors, and Itself and Promises High Efficiency. CIO (March). Rappaport, Andrew S., and Shmuel Halevi. 1991. The Computerless Computer Company. Harvard Business Review 69: 69–81. Rasiah, Rajah. 2000. Regional Dynamics and Production Networks: The Development of Electronics Clusters in Malaysia. http://www2.ucsc.edu/cgirs/publications/cpapers/rasiah.pdf. Reid, Robert H. 1997. Architects of the Web: 1000 Days That Built the Future of Business. New York: Wiley. Rhoades, E. L. 1929. Merchandising Packinghouse Products. Chicago: University of Chicago Press. Richardson, G. B. 1972. The Organisation of Industry. Economic Journal 82(327): 883– 96. Riley, Elmer A. 1911. The Development of Chicago and Vicinity as a Manufacturing Center Prior to 1880. Ph.D. diss., Department of History, University of Chicago. Riordan, Teresa. 2000. Historians Take a Longer View of Net Battles. New York Times, April 10, pp. C1, C4. Roberti, Mark. 2001. B-to-B Startup Strikes Gold in Dell Deal. Industry Standard, June 12. http://www.thestandard.com/article/0,1902,21483,00.html. Rocks, David. 2000. Dell’s Second Web Revolution. Business Week, September 18. Rosenberg, Nathan. 1982a. Inside the Black Box:Technology and Economics. Cambridge: Cambridge University Press. ———. 1982b. Learning by Using. In Inside the Black Box:Technology and Economics, 120–40. Cambridge: Cambridge University Press. ———. 1994. Exploring the Black Box:Technology, Economics, and History. Cambridge: Cambridge University Press. Rosenzweig, Roy. 1998. Wizards, Bureaucrats, Warriors, and Hackers: Writing the History of the Internet. American Historical Review 103(5): 1530–52. Roszak, Theodore. 1994. The Cult of Information:A New-Luddite Treatise on High-Tech, Artificial Intelligence, and the True Art of Thinking. Berkeley: University of California Press. Roth, Daniel. 1999. Dell’s Big New Act: Michael Dell, Perhaps the Most Successful E-Merchant in the World, Wants to Teach You and Your Company How to Master the Internet. Business 2.0, July. http://www.ecompany.com/articles/mag/ print/0,1643,5577,FF.html. Rowen, Henry. 2000. Serendipity or Strategy: How Technology Markets Came to

262 References Favor Silicon Valley. In The Silicon Valley Edge, edited by Chong-Moon Lee et al. Stanford: Stanford University Press. Roy, William G. 1997. Socializing Capital:The Rise of the Large Industrial Corporation in America. Princeton, N.J.: Princeton University Press. Russell, Charles Edward. 1904. The Greatest Trust in the World. New York: RidgwayThayer. Ruttan, Vernon W. 1997. Induced Innovation, Evolutionary Theory, and Path Dependence: Sources of Technical Change. Economic Journal 107: 1520–29. Sabel, Charles F., and Jonathan Zeitlin. 1985. Historical Alternatives to Mass Production: Politics, Markets and Technology in Nineteenth-Century Industrialization. Past and Present, no. 108: 133–76. Saccomano, Ann. 1995. Dell Computers, SmithKline Beecham Outsource Logistics Management Needs. Traffic World, February 27, pp. 54–55. ———. 1998. A Study in Contrasts: First-Quarter Earnings of Compaq, Apple Show Inventory Can Mean Everything. Traffic World, April 27, pp. 44–45. ———. 1999. Dell Computer Builds Service. Traffic World, July 26, pp. 26–27. Saunders, Rebecca. 2000. Business the Dell Way:Ten Secrets of the World’s Best Computer Business. Oxford: Capstone Books. Saxenian, AnnaLee. 1990. Regional Networks and the Resurgence of Silicon Valley. California Management Review 33(1): 89–112. ———. 1991. The Origins and Dynamics of Production Networks in Silicon Valley. Research Policy 20(5): 423–37. ———. 1994. Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Cambridge: Harvard University Press. ———. 1999. Silicon Valley’s New Immigrant Entrepreneurs. San Francisco: Public Policy Institute of California. Sayer, Andrew. 1992. Method in Social Science. London: Routledge. Schmidt, Louis Bernard. 1920. The Internal Grain Trade of the United States, 1850– 1860. Iowa Journal of History and Politics 18:94–124. ———. 1921. The Internal Grain Trade of the United States, 1860–1890. Iowa Journal of History and Politics 19:196–245, 414–55. ———. 1922. The Internal Grain Trade of the United States, 1860–1890. Iowa Journal of History and Politics 20:70–131. ———. 1939. Internal Commerce and the Development of National Economy Before 1860. Journal of Political Economy 47(6):798–822. Schoenberger, Erica. 1991. The Corporate Interview as a Research Method in Economic Geography. Professional Geographer 43(2): 180–89. ———. 1997. The Cultural Crisis of the Firm. Oxford: Blackwell. Schrage, Michael. 2002. The Dell Curve. Wired, July. Schumpeter, Joseph A. 1911. The Theory of Economic Development. Cambridge: Harvard University Press, 1951. ———. 1939. Business Cycles:A Theoretical, Historical, and Statistical Analysis of the Capitalist Process. 2 vols. New York: McGraw-Hill. ———. 1942. Capitalism, Socialism, and Democracy. New York: Harper & Bros.

References 263 ———. 1947. The Creative Response in Economic History. Journal of Economic History 7(2): 149–59. ———. 1954. History of Economic Analysis. New York: Oxford University Press. Scott, A. J. 1988. New Industrial Spaces: Flexible Production Organization and Regional Development in North America and Western Europe. London: Pion Ltd. Scranton, Philip. 1997. Endless Novelty: Specialty Production and American Industrialization, 1865–1925. Princeton, N.J.: Princeton University Press. Shapiro, Carl, and Hal R. Varian. 1998. Information Rules: A Strategic Guide to the Internet Economy. Boston: Harvard Business School Press. ———. 1999. The Art of Standards Wars. California Management Review 41(2): 8–32. Shaw, William Howard. 1947. Value of Commodity Output Since 1869. New York: National Bureau of Economic Research. Sinclair, Upton. 1906. The Jungle. London: Penguin, 1986. Skocpol, Theda, and Margaret Somers. 1980. The Uses of Comparative History in Macrosocial Inquiry. Comparative Studies in Society and History 22(2): 174–97. Slaughter, Mathew J. 1995. The Antebellum Transportation Revolution and Factor Price Convergence. NBER Working Paper 5303. Cambridge: National Bureau of Economic Research. Smith, Adam. 1776. An Inquiry into the Nature and Causes of the Wealth of Nations. 2 vols. Oxford: Clarendon Press, 1976. Southwick, Karen. 1996. Above the Crowd: Interview with Jeff Bezos of Amazon.com. Upside 8(10): 29–34. Standage, Tom. 1998. The Victorian Internet:The Remarkable Story of the Telegraph and the Nineteenth Century’s On-Line Pioneers. New York: Walker. Steffens, John. 1994. Newgames: Strategic Competition in the PC Revolution. Oxford: Pergamon Press. Storper, Michael. 1997. The Regional World:Territorial Development in a Global Economy. New York: Guilford Press. Storper, Michael, and Richard Walker. 1989. The Capitalist Imperative. New York: Basil Blackwell. Sturgeon, Timothy J. 1997a. Turnkey Production Networks: A New American Model of Industrial Organization? Working Paper 92a. Berkeley, Calif.: Berkeley Roundtable on the International Economy. ———. 1997b. Does Manufacturing Still Matter? The Organizational Delinking of Production from Innovation. Working Paper 92b. Berkeley, Calif.: Berkeley Roundtable on the International Economy. ———. 2000. How Do We Define Value Chains and Production Networks? Paper Prepared for the Bellagio Value Chains Workshop, Bellagio, Italy. Swift, Louis F. 1927. The Yankee of the Yards:The Biography of Gustavus Franklin Swift. London: A. W. Shaw. Taylor, Charles H. 1917. History of the Board of Trade of the City of Chicago. 3 vols. Chicago: Robert O. Law. Tedeschi, Bob. 1999. The Net’s Real Business Happens .Com to .Com. New York Times, April 19, pp. C1, C6.

264 References ———. 2000. A Mad Dash to Create Electronic Marketplaces for Business-to-Business Transactions. New York Times, January 24, p. C10. ———. 2002. The 25 Top E’tailers and the Various Approaches They Use—From Catalogs to Hand-Holding. New York Times, July 22, p. C6. Teece, David, and Gary Pisano. 1998. The Dynamic Capabilities of Firms: An Introduction. In Technology, Organization, and Competitiveness: Perspectives on Industrial and Corporate Change, edited by Giovanni Dose, David Teece, and Josef Chytry, 185–211. Oxford: Oxford University Press. Tehan, Rita. 2001. Internet and E-Commerce Statistics:What They Mean and Where to Find Them on the Web. Congressional Research Service Issue Brief. http:// www.cnie.org/nle/st-36.html. Teresko, John. 2001. The Value of Velocity: Dell’s Rules: Disdain Inventory, Always Listen to the Customer, Sell Direct, and Eliminate Unnecessary Touches. Industry Week, October. Thompson, Robert Luther. 1947. Wiring a Continent:The History of the Telegraph Industry in the United States, 1832–1866. Princeton, N.J.: Princeton University Press. Thurm, Scott. 2002. “Five High-Tech Companies Are Amassing Big Piles of Cash.” Wall Street Journal, December 27, pp. A1, A4. Trommer, Diane. 1997a. As “Build to Order” Fires Up the PC Business, Can the Supply Chain Stand the Heat? Electronic Buyers’ News, December 15. ———. 1997b. Dell Unfazed by PC OEM’s Attempts to Follow Its Lead? Electronic Buyers’ News, no. 1088, December 15. ———. 1998. Dell Plans Links with Parts Vendors. Electronic Buyers’ News, June 8. Twain, Mark. 1883. Life on the Mississippi. Oxford: Oxford University Press, 1990. Unfer, Louis. 1951. Swift and Company: The Development of the Packing Industry. Ph.D. diss., University of Illinois. U.S. Bureau of the Census. 1864. Agriculture of the United States in 1860. Eighth Census. Washington, D.C.: Government Printing Office. ———. 1872. The Statistics of the Wealth and Industry of the United States. Ninth Census. Vol. 3. Washington, D.C.: Government Printing Office. ———. 1883. The Statistics of Manufactures. Tenth Census.Vol. 2. Washington, D.C.: Government Printing Office. ———. 1902. Manufactures. Twelfth Census. Part 1. Vol. 7. Washington, D.C.: Government Printing Office. U.S. Bureau of the Census. 1904. Abstract of the Twelfth Census of the United States, 1900. Washington, D.C.: Government Printing Office. ———. 1975. Historical Statistics of the United States: Colonial Times to 1970. Washington, D.C.: U.S. Department of Commerce. U.S. Bureau of Corporations. 1905. Report of the Commissioner of Corporations on the Beef Industry. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Agriculture. 1885. First Annual Report of the Bureau of Animal Industry for the Year 1884, “Live Stock and Meat Traffic of Chicago,” by Edward W. Perry, pp. 245–69. ———. 1889. Fourth and Fifth Annual Reports of the Bureau of Animal Industry for the

References 265 Years 1887 and 1888, “Meat Industries of the United States,” by H. C. Clark, pp. 359–78. U.S. Department of Commerce. 1998. The Emerging Digital Economy. Washington, D.C.: Secretariat on Electronic Commerce. ———. 2000. Digital Economy 2000. Washington, D.C.: Economics and Statistics Administration, Secretariat on Electronic Commerce. U.S. Department of the Interior, Census Office. 1895. Report on the Population of the United States at the Eleventh Census: 1890. Washington, D.C.: Government Printing Office. U.S. House of Representatives. 1863a. Report of the Commissioner of Agriculture for the Year 1862, “Beef and Beef-Cattle of the West,” by W. W. Corbett. 37th Cong., 3rd sess. ———. 1863b. Report of the Commissioner of Agriculture for the Year 1863, “History of Pork Packing in Chicago.” 38th Cong., 3rd sess. U.S. Senate. 1890. U.S. Congressional Hearings, vol. 3, no. 829. Select Committee on the Transportation and Sale of Meat Products [known as Vest Committee Report]. 51st Cong., 1st sess. U.S. Senate. 1893. Committee on Finance. Wholesale Prices,Wages, and Transportation: Report by Mr. Aldrich [known as Aldrich Report]. 52nd Cong., 2nd sess. Senate Report 1493, 4 vols. Washington, D.C.: Government Printing Office. Utterback, James, and Fernando Suarez. 1993. Innovation, Competition, and Industry Structure. Research Policy 22: 1–21. Varian, Hal R. 2002. If There Was a New Economy, Why Wasn’t There a New Economics? New York Times, January 17, p. C2. Veblen, Thorstein. 1904. The Theory of Business Enterprise. New York: Scribner’s, 1932. Vernon, Raymond. 1960. Metropolis 1985. New York: Anchor Books. Von Hippel, Eric. 1988. The Sources of Innovation. Oxford: Oxford University Press. Wade, Louise Carroll. 1987. Chicago’s Pride:The Stockyards, Packingtown, and Environs in the Nineteenth Century. Urbana: University of Illinois Press. Walker, Francis. 1906. The “Beef Trust” and the United States Government. Economic Journal 16(64): 491–514. Walker, R. 1988. The Geographical Organization of Production-Systems. Environment and Planning D: Society and Space 6: 377–408. Walsh, Margaret. 1978. The Spatial Evolution of the Midwestern Pork Industry, 1835–1875. Journal of Historical Geography 4(1): 1–22. ———. 1982. The Rise of the Midwestern Meat Packing Industry. Lexington: University of Kentucky Press. Weintraut, J. Neil. 1997. Introduction to Architects of the Web: One Thousand Days That Built the Future of Business, by Robert H. Reid. New York: Wiley. Williamson, Oliver E. 1975. Markets and Hierarchies. New York: Free Press. ———. 1980. Emergence of the Visible Hand: Implications for Industrial Organization. In Managerial Hierarchies, edited by A. D. Chandler Jr. Cambridge: Harvard University Press. ———. 1985. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. New York: Free Press.

266 References Williamson, Oliver E., and Sidney G. Winter, eds. 1991. The Nature of the Firm: Origins, Evolution, and Development. New York: Oxford University Press. Williamson, Oliver E., and William G. Ouchi. 1981. The Markets and Hierarchies and Visible Hand Perspective. In Perspectives on Organization Design and Behavior, edited by Andrew H. Van De Ven and William F. Joyce. New York: Wiley. Wood, Lamont. 1993. Dell Computer Tries to Get Back on Track. Electronic Business Buyer 19(10): 152–54. Wright, Carroll D. 1883. Report on the Factory System of the United States. U.S. Census, Census of Manufactures 1880. Tenth Census. Washington, D.C.: Government Printing Office. Pp. 527–610. Yates, JoAnne. 1986. The Telegraph’s Effect on Nineteenth-Century Markets and Firms. Business and Economic History 15: 149–63. ———. 1989. Control Through Communication:The Rise of System in American Management. Baltimore: Johns Hopkins University Press. ———. 1997. Using Giddens’ Structuration Theory to Inform Business History. Business and Economic History 26(1):145–83. Yates, JoAnne, and Robert I. Benjamin. 1991. The Past and Present as a Window on the Future. In The Corporation of the 1990s: Information Technology and Organizational Transformation, edited by Michael S. Scott Morton, 61–92. New York: Oxford University Press. Yeager, Mary. 1981. Competition and Regulation: The Development of Oligopoly in the Meat Packing Industry. Greenwich, Conn.: JAI Press. Young, Allyn. 1928. Increasing Returns and Economic Progress. Economic Journal 38: 527–42. Zook, Mathew. 2001. The Geography of the Internet Industry: Venture Capital, Internet Firms, and Regional Development. Ph.D. diss., Department of City and Regional Planning, University of California, Berkeley. Zysman, John. 1994. How Institutions Create Historically Rooted Trajectories of Growth. Industrial and Corporate Change 3(1): 1–41.

Index

Figures, maps and tables are identified by a parenthesis after the page number: 379(f), 45(m), 102(t). Notes are identified by page and note number: 244n6.

abattoirs, 95, 241n14 Acer Computer, 209 access: and circulation, 13, 14; and costs, 56 actor network, 38 adaptive capacity of integrated firm, 47 adaptive response, 9, 211, 213; of competitive firms, 222, 240n12, 245n15 administrative controls, 199, 228 administrative coordination: vs. contracting through price markets, 41, 199, 227; as foundation for innovation, 9; in lieu of markets, 28 administrative costs, 187 Advanced Research Project Agency (ARPA), 143, 145–46. See also ARPANET advantages of the rail and telegraph system, 77–79 agency: in choice selection, 37; in communications revolution discovery, 10; Dell decision making power on factory location, 207; learning as, 13, 36; from users, 141 agglomeration economies, 206, 207, 230 AISCs. See application specific integrated circuits Albers, Toni Lynn, 22 Albion, Robert, 10–11, 27, 52 Amalgamated Meat Cutters and Butcher Workmen Union, 241n14 Amazon.com, 154–55, 156, 187 AMD (Advanced Micro Devices), 205

America Online, 144 Andreesen, Marc, 149 Annales school, 230 Anthony, D. M., 107 antitrust litigation, 134–35 Apple Computer, 171, 178, 189 application-specific integrated circuits (ASICs), 180 Armour, P. D., 91, 109–10 Armour Company, 92, 109, 111, 115, 132, 133 ARPA (Advanced Research Project Agency), 143, 145–46 ARPANET (ARPA Network), 141, 145, 147, 148, 242–43n10 assembly plants, 19, 230 asset specificity, 43 Austin, Tex., 179, 183, 205 backward integration, 45 balance in supply and demand flows, 125, 203, 226; product flows, 160 Baltimore and Ohio Railroad, 77 Baran, Paul, 145, 146 barges, 73 Bayan Lepas Free Trade Zone, 205, 245n20 B.C. (before commercialization), 151 beef: vs. pork, 95, 98(t), 239n4. See also cattle; dressed beef beef-killing gangs, 113 Bell, Jay, 180 Berners-Lee, Tim, 148–49

268 Index Bezos, Jeff, 154, 155 Bijker, Wiebe, 38 Bina, Eric, 149 Boca Raton, Fla., 170 Bolt, Beranek and Newman (BBN), 145, 146 Boston & Albany Railroad, 102 Botsford, Henry, 240n12 boundaries: between different market areas, 14; of the firm, 42, 51 bounded rationality, 43 branch houses: consignments to, 108; cumulative numbers of, 111(t); development of, 240n11; as distribution network, 93; expansion of, 108; functions of, 126–27; imitation of, 109; locations of, 127, 129; network expansion, 112; organization and methods of, 132; for product distribution, 18; production and marketing system of, 132 brand names, 93, 178, 180 Braudel, Fernand, 230, 239n11 Brenner, Robert, 234n2 brokerage effect, digital, 161 builders of infrastructure, 54, 55, 140. See also users build-to-order business model, 225 build-to-order demand pull environment, 194 Bureau of Corporations, 125, 126, 135 Burnham, Daniel, 91 business: contraction and creative destruction, 35; enterprises as organizational linkages, 10; groups, 237n22; growth cycles and technological clusters, 32; users, 54, 55, 56, 144 business cycle expansion, 35 business organizations: changes of, 28; as collections of assets, 228; dichotomy vs. continuum, 47; efficient forms of, 45; of the firm, 11; interfirm networks as forms of, 48; literature on, 23; in modern economy, 24; research organization, 241n15 business-to-business commerce, 157, 158, 159, 162–63, 188 butcher aristocracy, 113 butchers. See local butchers; retail butchers; wholesale butchers butcher workman, 113 buyers, 151

Cady, Samuel, 117 Cameron, Brewster, 117 Canal-Lake route, 74 canals, 74–75, 86 capabilities, 12, 34, 221 capitalism: circulation, triggered by trade, 69, 223; commodity chains as business form in, 48–49; communications revolution effect on restructuring of, 53; creative destruction in development of, 92; incentives for capitalist profit systems, 221; innovation, as driver to capitalist economy, 29; innovation, leading to business cycles as essence of capitalist process, 220; markets, roles of, in capitalist economy, 29; new technologies as central element in capitalist development, 31; Marx, Karl on, 30–31; Schumpeter, Joseph on, 31–34, 234n3; technological dynamism, role of, in development of, 30; waves of technical innovation marking development of, 54 capital value and sales of G. F. Swift & Company, 1885–1904, 115(t) carrier wave, 11, 54 cartel agreements, 101, 240n10 case studies: the argument, 11–20; comparisons, 2–5; intensive and extensive, 4; existing, 22; findings at a glance, 5–9; interviews and interviewees, 218–19; time periods, 21 Castells, Manuel, 11, 48, 53–54, 237n25 catalyst, 11, 37 cattle: commission houses, 99, 116, 241n16; drives, 80, 81, 97, 98; drovers, feeders, grazers, 97; markets, 241n16; merchants, 97; and railroads, 98–102. See also dressed beef Cattle Commission houses, 16, 99, 241n16 central goods, 70 central place theory, 70 centralization: of G. F. Swift, 1903, 128(m); organizational, 129 central processor units (CPUs). See microprocessors Cerf, Vinton, 147 chaebol, 46, 244n7 Chandler, Alfred: on communications and transport, 11; critics of, 236n19; on influence of new technologies, 52–53; and interfirm relationships, 47; Langlois on, 51;

Index 269 strategy-and-structure approach to the organization of the firm, 43–45; on transactions cost reductions, 236n18; visible hand of, 167, 228 channel assembly, 214, 215, 216 channels: in beef industry, 108(f); in PC industry, 175, 180(f) Charleston, S.C., 86 Chase, Andrew, 105 Chicago, Ill.: on Canal-Lake route, 74; as center of meatpacking, 91; as concentration point for grain, 74–75, 75(t); emergence of, as nation’s second city, 90; G. F. Swift & Company headquarters, 94; growth of, 86; G. Swift early experiences in, 103; livestock concentration, 81; meatpacking firms ranked by size, 1878 and 1884, 109(t); as rail center, 75–76, 98; slaughtering concentrations, 119, 120; stockyards, 114; as viewed by Mark Twain, 239n10 Chicago Board of Trade, 81, 87, 92 Chicago Union Stockyard Company, 99 China, 205, 206, 209, 245n20 China Picture Tube, 169 chip sets, 180 Christaller, Walter, 69–70 Cihra, Robert, 168, 219 Cincinnati, Ohio, 72, 86, 120, 239n9 Circuit City, 184 circulation: affecting structures of market, 24, 67, 69; breakthroughs in sphere of, 5; between network nodes, principal, 203; new systems of, 13, 14; side of value chain, gains of trade, 160; sphere of (logistics), 24; as sphere of (trade) and trigger to capitalist growth, 223 Cisco Systems, 157, 159, 189 cities: concentration of manufacturing in, 87; concentration of trade in, 84; concentrations of slaughtering in, 119; high-technology, second tier, 207; interdependence of, 87–88; and mass markets, 68; as points of concentration, 68 Clark, Jim, 149 Cleveland, Grover, 133 clones, 172, 178 closed-loop organization, 5 closed-loop relationships, 181, 200 Coase, Ronald, 7, 40–42, 47, 167, 236n16

co-location, 215 Comdex, 180 Commerce Clause: of U.S. Constitution, 88, 134 commerce system, 63, 139 commercialization as stage of innovation, 31 commodities, 82 commodity chains, 48–49 commodity exchanges, 81 commodity prices, 83 Commodore, 170 communications, 160; as catalyst and carrier wave, 52–54; and commerce, 63–68, 139– 44; and creative distribution, 92–93; system, defense-oriented, 140; to territory, 27–28; and transportation, 2, 44. See also rail and telegraph systems communications effect, digital, 161 communications revolutions, 32–33; actors in, 54–55; as carrier wave, 11; as catalyst, 11, 37; as control revolutions, 15, 56; cycle of, 58–59; effect of, on business users, 56; and innovation, 221; intermediate stage of, 63; literature on, 23; from nineteenth century to present, 231; as ongoing phenomena, 11, 231; outline of, 57, 58(f); as a platform for reorganizing competitive activity, 139; and speed in geographical space, 52; to territorial transformation, 12(f) Compal, 209 Compaq: BIOS development of, 172; as clone builder, 178; Configuration Partner Program, 215; CSX, 215; Dell influence on production and distribution, 216; indirect channel, 176; Internet selling, 214; inventory levels of, 166; Optimized Distribution Model, 215; sales levels, 211; Web sales, 213 competition and the PC before Dell, 168–78 competitive advantage, logistics as source of, 5 component suppliers, 209–10, 244n7. See also suppliers CompUSA, 184 computer firms: ranked by U.S. market share, 214(t); ranked by world market share, 185(t). See also PC industry; specific firm Computerland, 170

270 Index computers: assembly operations, 205–7; components production, 172(t), 174; operating systems, 170; price declines, 217(f) concentration: of economic activity, 168; of manufacturing, 87; of slaughtering, 119, 120; of trade, 84 conjunctures (sharp breaks) in history, 230, 231 conscious power (planning), 41 consignment system, 82 consumption junction, 141 contextualism, 38 continental markets, 18 continuous order-and-fulfillment flow, 5 contours of commercialization, 139–41 contours of parallel worlds, 224–30 contract manufacturers, 50 controlled relationships and administrative controls, 199 coordinating transactions vs. market system, 42 copy-exact approach, 203, 211 core competencies, 50 corporate market, 180, 181 corporate power, 9, 89, 228 costs: of start-up and learning, 171; of transactions, 42, 43, 156; of transfer, 69; of transportation, 72, 77–78, 88, 101 creative act of entrepreneurs, 32 creative destruction: adaptive response as part of, 211; and business contraction, 35; with entrepreneurial enterprises, 34; as essence of capitalist development, 92; and process of disruption, 29 creative response, 9, 16, 32 cross-border production network theory, 209 Cudahy, 132 cumulative causation, 206 custom direct, 178–83 customer-direct sales, 155 customers, nature of, 193 customization, 125, 165, 225 cycles: business, 35; of creativity, 29; growth, 32; of just-in-time pull ssystem, 199; negative cash conversion, 179; of PC production and distribution, 179; of procurement, production, and selling, 56, 224; product, theory of, 208, 245n22; of unionization in meatpacking, 133

David, Paul, 4–5 Davies, Donald, 145, 146 Davis, William, 105 DEC (Digital Equipment Corporation), 173 decentralization in pork packing, 120 decentralized markets of beef industry, 95 decision-making power by Dell Computers, 207 declining profit margins, 245n22 Defense Communications Agency (DCA), 147 Dell, Michael, 22, 176, 183, 190, 199, 210 Dell.com, 187 Dell Computer Corporation: as agent for industrial change, 217; assembly sites expansion, 1991–99, 204(m); business enterprise, 182(m), 212(m); business organization of, 190–211; carrier wave theory, 54; case studies discussion, 2, 4–7, 22; communications revolution and, 11; component suppliers, 209; daily Internet sales (1996–2001), 188(t); decision-making power by, 207; early sales and profits, 1983–90, 181(t); factories near Intel factories, 207; as globalization paradigm, 24; gross margins of, 245n15; growth before and during Internet period, 186(f); growth of sales and employment, 187(f); innovations by, 3, 165; interfirm sales by, 157; Internet as structure for business by, 139, 140, 144; Internet exchange site experiments by, 163; Internet sales capacity, 157; inventory problems at, 189; learning within the organization, 16; as logistics company, 243–44n2; as a logistics firm, 166; marketing changes, 183–84; and market power, 202; market regions of, 205; organizational power and control by, 8–9, 20; organizational structure of, 227; origins of, 178–90; production and distribution system, 19, 21; profitability, 184; regional organizational structure, 189; as role of agent, 211; sales growth, 183–84; sales volume, 211, 213; spread and concentration of, 7; subcontracting by, 50; supply chain experiments by, 159; as user of technology, 14; virtual integration by, 17– 18, 167 demand and supply balance. See supply and demand balancing

Index 271 demand fulfillment, 195 dependent phenomena: in case studies, 2 depreciation, 177. See also perishable prices deregulation, 153 Destler, Chester, 239n3 diffusion (transfer), 9, 32, 223, 240n12, 245n15; concept in social constructivism, 38; of Dell’s business organization, 211– 17; in meat-packing practices, 229; process of, 111; as stage of innovation, 31 Digital Equipment Corporation (DEC), 173 direct closed-loop relationship with customers, 19 direct distribution system, 94 direct incentives, 206 direct marketing, 16; attributes of, 125 direct-pull just-in-time production-and-distribution network, 5, 18, 19 direct-pull production-and-distribution network, 19 direct-pull systems: Dell, 192, 201; origins and foundation, 19; relationship with customers, 224; Swift, 18, 19, 125–27, 129, 162; Swift and Dell compared, 224–26 direct selling, 165 direct system of operations, 121–29 disintegrated interfirm networks, 166 disintegrated interfirm organization, 8 disintermediation: advantages of, 179; of beef market structure, 19; by Dell Computer Corporation, 186, 226; as origins and foundation of direct-pull system, 19; by Swift’s direct sales, 94, 125 distribution: channels of, 175; as competitive factor, 175; indirect system of, 176; as source of territorial spread, 120–21 division of labor, 40, 50, 70, 113, 131, 225 domain name registrations, 152 Dosi, Giovanni, 37 dressed beef, 5, 21–22, 68, 227; expansion and control, 116–17; innovations in, 90– 93; investigations, 111–13; labor and politics, 132–35; operations and management, 120–21, 125–32; transportation and delivery, 101–9. See also cattle dressed beef enterprise of G. F. Swift (various years), 106(m), 110(m), 124(m) Dressed Beef Trust, 116, 116(t), 117, 132–33, 134, 135 Drucker, Peter, 139

DSi2 project, 186, 190–91; global supply planning system, 168, 193, 194, 198–99 Durkheim, Emile, 63, 237n24 Eagle Global Logistics, 196 early beef trade, 97–98 early markets, 70 early PC market in the United States, 1976– 1983, 171(t) early refrigeration, 102–3 East Asia, 178, 183 eastern beef distribution, 101 Eastern Trunk Line Association, 105, 240n9, 240n10 Eastman, T. C., 102 East Saint Louis, Mo., 114 East-West axis of trade, 74–75 e-commerce, 188 economic actors, profit environment for, 231 economic growth, patterns of, 88 economic incentives for business locations, 206 economic routines (technology), 35 economic space, 9, 10, 156 economic transformation process, 33 economies of speed, 44 economy in equilibrium, 235n5 efficiencies: of Dell Computer Corp., 211; in economic space, 156; Internet-driven, 186; market, by, 155; in procurement, 162; in production, 171 Eldorado do Sul, Brazil, 205, 206 embeddedness, 237n22 empirical outlines, 16–20 enterprise, expansion and diffusion, 108–14 entrepreneurialism, 31, 32, 33, 234n2 entrepreneurial opportunity, 103–4, 176–78 entrepreneurs, 41, 43 Epson, 169 equilibrium models of commodity flows, 30 Erie Canal, 74 Erie Railroad, 72, 77 established markets, innovative activity outside of, 3 e.steel, 163 Evener System, 101, 240n8 exception action report, 194 exception basis, 194 expandability and interoperability, 141–42 expansion: and control, 116–19; of Dell as-

272 Index sembly sites, 1991–99, 204(m); of G. F. Swift production facilities, 118(m); of the rail and telegraph systems, 1848–1902, 65(t) extensive case studies, 4 external capabilities, 20, 227 external scale economies, 40, 119, 206 extranet, 161 Federal Express, 243n14 Federal Meat Inspection Act (1891), 134 Federal Meat Inspection laws, 89 federal troops, use of, in strikes, 133 Federation of Organized Trades and Labor Unions, 133 final ascent, 211–13 final propositions, 230–32 firm: as business organization, 40–52; as business organization and network, 9–10 firm boundaries, 42, 51 first mover: advantages, 111; impact on standards, 242n4 Fitchburg, Mass., 240n11 flows (of activities), 57 Fort Worth, Tex., 94, 114, 117 forward integration, 45 Framework for Global Electronic Commerce, 152, 153 FreeMarkets, 163 Free Trade Zones, 205 Fujitsu, 171 functional integration, 131 futures contracts, 82 Galena and Chicago Union Railroad, 75 Gateway, 176 gateway to innovation, 91–95 General Electric (GE), 159, 162, 163 Genesis Project, 188 geographical barriers, 67, 69, 71, 82 geographical expansion of the Internet, 142 geographical linkages, 10, 11, 84, 86 geographical patterns, 11, 17, 86, 119, 120, 167 geographical separation: of cattle grazing and population, 98; of slaughter and sale, 126 geographical space, 2–3, 10, 14, 52, 57, 84 geographical spread, 120, 167, 203, 208. See also spread and concentration

geographical structure of markets, 14 geography: of assembly, 205–7; of economic activity, 94; friction of, 195; of innovation, 222; of interfirm enterprise, 203; of markets, 64; of networks partners, 203; new, anatomy of, 119–21; of operational organizational innovation, 114–32; of production and distribution, 6; of supply, 207– 11; velocity requirements and, 203 G. F. Swift & Company: as agent for change, 90; Chicago central office, 126–27, 129; core competencies of, 126; corporate size, 115–16; direct marketing of, 125; expansion of branch houses, 108; headquarters of, 94; network of, 126; owner of Boston wholesale meat business, 103; telegraph charges of, 127; time and distance obstacles, 101 globalization, 3, 24, 168, 230 global markets, 18, 230 grain: hub and shipments, 73–75, 75(t); Southern imports from the West, 73 gains of trade, 160 Giddens, Anthony, 235n10 Grand Trunk Railroad, 105, 240n10 Granovetter, Marc, 237n22 Gregoire, Jerry, 189 growth: models based on clusters, 46; of .com Web sites, 1993–96, 152(t); of manufacturing employment in U.S. cities, 1860–1890, 87(t) Hammond, 132 Hammond, George H., 102, 103, 104, 105, 107 Hathaway, James, 105 Hewlett-Packard (HP): early PC entrant, 171; Malaysian operations, 205; sales levels of, 189; Web-based sales, 213, 214 hierarchies: distinct from markets, 42, 167 high-technology cities, second tier, 207 high-technology industrial districts, 208 high-volume sales, 165 high-volume throughput, 44, 125 historical cluster, 231 historical path dependence, 206 history, 230, 231; as foundation for understanding the economy, 29; parallels in, 220; does not repeat, 220; of structure (longue dur´ee), 230, 232; of the world as it

Index 273 is being made, 230. See also geography; logic; theory Holmes, Oliver Wendell, 135 Homestead Act (1862), 238n3 horizontal partnerships vs. vertical integration, 174 Hounshell, David, 235n11, 235n12, 237n19 Howard, Henry Payton, 102 Hughes, Thomas, 38, 235n12, 241n2 hypertext markup language (HTML), 148 hypertext transfer protocol (HTTP), 148 i2 modules, 193, 197 IBM: Authorized Assembler Program, 215; beaten by Dell technology, 180; and clone development, 178; and direct channel sales, 176, 216; Internet selling, 214; open standards and modularity, 173–75; and PC development, 169–70; replenishment service centers, 215; Web sales, 213 ice: 130(t); role in dressed beef operations of Swift, 130; Swift as manufacturer of, 130 Illinois and Michigan Canal, 74 imperfect knowledge, 37 improvements in process for making and marketing products, 222 incentives for capitalist profit systems, 221 incompatible networks, 146 independent phenomena: in case studies, 2 indirect channel, 175–76, 184 indirect system of distribution, 176 inducement, 37, 39 industrial districts, 6, 186, 208, 230 industrial divide, 46 industrial revolutions, 32–33 inedible by-products, 130–31 inedible portion of beef, 100, 103 information age, 53 informational societies vs. information societies, 53–54 information management, 6 Information Processing Techniques Office (IPTO), 145 information society, 24 information technology revolution, 53 infrastructure: expansion and interoperability, 140; interconnection, 64–67; as preconditions for business innovation, 221; standards, 147–48 innovation(s), 240n12; advances in compo-

nents of, 37–38; based on processes, 165; concentration of economic activity, 168; as contingent process, 38; defined, 9, 233n3; and diffusion, 29; in distribution, 115; driven by interfirm relationships, 51; as driver to capitalist economy, 29; and entrepreneurialism, 31–34; within the firm, 2, 11; at G. F. Swift & Company, 89; as history, 9; as inducement, 37–40; inducement to, 37; leading to business cycles as essence of capitalist process, 220; as learning, 34–37, 39; ; through learning, 14; literature on, 23; in logistics and distribution, 222; and markets, linkage to, 28–29; moments (stages) of, 31; and organizational change, 40; as outcome of search for efficiency, 38; paragons of, 21; preempted by Microsoft standards, 235– 36n13; process-oriented, fundamentals of, 93; regional and geographic concentrations of, 39; and technological challenge, 28–40 innovative activity, 3 innovative capabilities, location of, 50 innovative enterprise, 9, 34–35 innovator, 188 input-output analysis, 243n15 inspection laws (beef), 89, 133, 134 institutionalization of interregional trade, 81–84 integrated firms, intrafirm organization, 41– 45 integrated intrafirm organization, 7 integration effect, digital, 161 integration (organizational coherence), 36 Intel: as supplier for clones, 172; as supplier to Dell, 202, 207, 215n14; Malaysian operations of, 205; IBM purchases from, 170; and open architecture, 244n6; perishable prices, 244n8; subcontracting by, 209; virtual monopoly of, 174 Intel 80386 microprocessors, 148, 170 Intel 8088 microprocessors, 172 intensive case studies, 4 interconnection, 241–42n3; of infrastructure, 64–65; and interoperability, distinction between, 241–42n3; standards for, 66, 83, 147; standards lacking, 238n2 Interface Message Processors (IMPs), 145, 146 interfirm activity, 243n15

274 Index interfirm networks: with external capabilities, 227; as form of business enterprise, 46; as forms of business organization, 48; for production, 8, 174 interfirm organization, 18, 41, 45–52, 200, geography of, 203 interfirm relationships, 237n22 interfirm trade 157, 158, 159 interfirm value chains, 157–64 interim partners to manage rapid cycles, 185 intermarket activity, risk reduction by promotion of, 81 intermediaries, 161, 162–63 internal scale economies, 131–32 Internet: architecture, 144–48; attributes of, 161; business-to-business space, 157–64; commerce on, 151, 152–53; commercialization, 143–44; communications, 167; Dell Computer’s use of, 184–86; described, 144; development of, 143; domain name registrations, 1992–2001, 151(t); efficiency, 155–57; history of, 140; as infrastructure for direct-pull just-in-time production-and-distribution network, 19; as infrastructure for mass-production and mass-distribution organization, 23; interoperability of, 141–42; market maker, 161; origins, 144–48; for production management, 165; as process innovation, 165–68; as pulled information channel, 150; retail space, 153–57; rule making on, 152; as sales channel, 155; technology, 146; as transactions process technology, 156; transformation of, from communications to commerce, 139; value chains, 161–64 Internet-based system for supply chain management, 190 Internet browsers, 143; Mosaic, 148, 149, 242–43n10; Netscape, 148, 149, 150; World Wide Web and commerce, 148–52 Internet Protocol (IP), 147 inter-networking, 146 interoperability, 140, 142 interregional trade: 68, 72–73; beginning of, 73–75; and mass markets, 67–68 interstate commerce, restraint of, 134–35 Interstate Commerce Act (1887), 240n10 interviews and interviewees, 218–19 intrafirm organization, 41 intraregional trade, 73, 85

invention, 31, 38 inventory, 177, 181; vs. burst capacity, 191; of Compaq and Dell, 190(t); from component vendors, 196; costs of beef, 127, 129; of Dell, 1994–2000, 216(f); levels of, 166; and mass customization, 183; problems of, at Dell Computers, 183, 189; reduction of, 215; at SLCs, 196, 197; of suppliers, 201 invisible hand, 96 IPTO project, 146 Ireland, 209 Israel, 209 Jabil Circuit, 210 Japan, 178, 209 Johnson, Nigel, 202 The Jungle (Sinclair), 115 just-in-time procurement, 125, 127, 129, 192–94, 225–26 just-in-time production and distribution, 20, 167 Kahn, Robert, 147 Kansas City, Mo., 94, 117, 119, 120 Kansas City Stockyard Company, 117 keiretsu, 46 knowledge, types of, 12 kosher industry in New York, 111 Kuznets, Simon, 208 labor: concentrations of, 119; control over, 59; factory (Swift), 68, 133 125; relations, 54; resistance of, 133; factory (Dell), 197; as reverse salient, 38; unrest, 133. See also division of labor labor unions, 133, 241n14 Langlois, Richard, 51, 237n23 LANs (local area networks), 143, 148 large firms, suppliers and clones, 171–73 large retailers, 183–84 leading sectors, 28 learning: as a function of structure and agency, 13; innovation as, 34–37; innovation through, 14; process of, 9, 10; by using, 16, 222 learning a new business, 104–8 learning process, 12, 34, 56–57 legacies, 216 legal and political rule-making, 140 Leonard, L., 113

Index 275 Leontieff, Wassily, 243n15 licensing fees, 89 Limerick, Ireland, 205, 206, 245n24 live cattle shipments from Chicago, 1852– 1861, 81(t) livestock, transporting overland, 80 livestock markets concentrations, 94 local area networks (LANs), 143, 148 local butchers: becoming retail conduits, 95, 96; pushed for preslaughter inspection laws, 89 local markets, 95 local market structure, 70–71 location, 160; as condition for entry, 210; Dell, 205–7; of network nodes, principal, 203; of subcontracting firms, 208; Swift, 116–17 lock-in effect, 242n4 logic: of competition, 234n2; of efficiency, 38; of similarity, 4 logistics, as source of value creation and competitive advantage, 5 logistics-oriented firms, 16 logistics-oriented organizations, 228 longue dur´ee (structure), 230, 232 long-wave theory, 234–35 Lowe, William, 169 Machine Age, 11 mainframe product line, 173 make or buy, 40 Malaysia, 205, 206, 210; component suppliers, 209; Intel chip facilities, 209; Penang, as center, 207, 245n20, 245n24 manufacturing concentration in cities, 87 manufacturing industries ranked by dollar value of output, 1870 and 1900, 114(t) margarine, 131 margins, reduction in, 100 market: actors, 233n6; boundaries, 14, 67, 69; competition, 30; development, 71; efficiencies, 155; expansion, 237n24; geography, 67–68, 155, 156, 161; integration and concentration, 44; power of Dell Computer Corporation, 199, 201–2; power of G. F. Swift & Company, 89, 131, 241n16; regions, 205; risk, 71; share, 98, 132; structure, 92; system vs. coordinating transactions, 42 marketplace, roles of, 40

markets, 42; for cattle, 241n16; expanded and refigured structure of, 64; geographical structure of, 14; and meat, 95–97; and politics, 233n6; before the railroad and telegraph, 70–73; roles of, in capitalist economy, 29; territorial limits of, 28; as territory, 69–70; World Wide Web as, 154 Markusen, Ann, 208 Marshall, Alfred, 39, 40, 95, 119, 206 Martson, Steve, 215 Marx, Karl: on human agency, 235n10; influence on Schumpeter, 9, 30, 31, 220; on innovation and organizational change, 40, 220; on market expansion and capitalist development, 237n24; on technology, 31; theory of economic development, 30–31 Massachusetts Railroad Commissioners, 100 mass customization, 183 mass markets, 68, 84, 87–88 mastering capabilities, 35 Master Production Plans, 193, 194 material balances, 192 material spend, 200 McCurdy, Albert, 107 McDonald’s type approach, 211 McNeill, William, 52 meatpacking industry: branch house innovations, 111; dressed beef shipments, 92; growth of, as an industry, 113; local inspections, 89; seasonality of, 96; Senate hearings on, 112; vertical integration in, 132 mechanisms of power and control for risk mitigation, 8 message blocks, 146 Metalsite, 163 Metcalfe, Robert, 142 Mexico, 183, 209 Michigan Car Company, 105 microprocessors, 146, 180; Intel, 148, 170, 172; perishable prices, 244n8; speeds of, 180 Microsoft: clones, mutual interest with, 172; as Dell supplier, 202; as intellectual property supplier, 209; PC operating system supplier, 170; standards preempting innovation, 235–36n13; virtual monopoly of, 174 middlemen, 103 MILNET, 148

276 Index Milwaukee, Wis., 81 Minnesota v. Barber, 134 modular architecture, 170 modularity and standardization, 173, 225 momentum: overcoming reverse salients, 38, 241n2; provided by standards and rules, 142, 235n13; for technological changes, by business users, 141; of technology, 35 Montgomery Ward, 156 Moore’s Law, 174 Morris, 132 Morris, Nelson, 102, 103, 104 Morton Topfer Manufacturing Center, 196, 218 Mosaic (browser), 148, 149, 242–43n10 Motorola, 205 MS-DOS, 170, 172 mysteries in the air, 39. See also Marshall, Alfred Nashville, Tenn., 205 National Butchers’ Protective Association, 133 national market, 229 National Science Foundation (NSF), 143, 148, 149, 242–43n10 National Semiconductor, 205 NEC, 170 negative cash conversion cycle, 179 Netscape (browser), 148, 149, 150 Netscape Communications Company, 149 networking, 6, 143, 148 network nodes, 203, 228 networks: of contracting relationships, 208; of firms as sources of innovation, 50; incompatible, 146; values of, 142 network society, 53 new economy, 53 New Orleans, La., 74–75, 86 new product development, 93, 239n3 New York, N.Y.: abattoirs within, 95; board of trade established, 81; dressed beef operations in, 109; dressed beef price advantage in, 107; early grain shipments to, 74; early refrigeration shipments to, 102; intercity dependence, 85; kosher industry in, 111; shipments from Cincinnati to, 72; slaughterhouses, reductions in, 120; Swift branch houses in, 108 New York and Erie Railroad, 65

New York Central–Boston & Albany Railroad, 105 New York Central Railroad, 77, 102 nodes (physical assets), 57 nonmarket mechanisms, 17 nonmarket relationships, 245n17 Norris, Frank, 63 NVIDIA, 209 offal (cattle by-products), 100, 103, 130 Office of Advanced Computing, 242–43n10 oligopolies, 94; in beef industry, 132, 135; development of, 89; equated with innovation, by Schumpeter, 31; Swift as part of, 31 Olivetti, 171 Omaha, Nebr., 117, 119, 120 one step forward, 99–102 on-line activities: infrastructure for clearing credit card transactions, 150, 163; material tracking, 163; procurement systems, 161, 162; reverse auctions, 162; selling, 186–87, 187(f) open architecture, 170 open standards and modularity, 173–75 operations: demand fulfillment, 195–98; global supply positions, 192–94, 224–26 opportunism, 43 opportunities and incentives, 35 order fulfillment, 243n14 order-intake process, 126–27, 185 organizational capabilities, 16, 34–35, organizational change and innovation, 40 organizational coherence (integration), 36 organizational control, 228 organizational learning, investment for, 36 organizational linkages, business enterprises as, 10 organizational relationships, 166 organizations, 226–28; arrangements to implement innovative enterprises, 221; creations to obtain value in production and distribution, 226; Dell Computer structure, 2, 139, 227; Swift, 93–95; 129–32; Dell as virtually integrated firm, 198–202 organizing principle of corporate power and administrative control, 228 output of meat products and manufactured ice, 1869–1904, 130(t)

Index 277 packets, 146 packet-switching networks, 147 pallet in/out charges, 196 parallels in history, 220 partner firms, smaller number of, 167 path dependence, 206 pattern of innovation and economic change, 221–23 PC-DOS, 170 PC industry: clones, 172, 178; computer firms ranked by U. S. market share, 214(t); computer firms ranked by world market share, 191(t); early entrants, 171; firms ranked by PC revenues, 183(t); firms ranked by PC revenues, 1990, 183(t); new entrants in, 169; open architecture, 172; PC value chain, 175; price declines, 173; sales, 1996–2000, 213(t); schematic of PC channels, 180(f) PCs and workstations networking, 143 PC’s Limited, 179, 180, 181 peddler-car routes, 131 Penang, Malaysia, 205, 206, 207, 245n20, 245n24 Pennsylvania Railroad, 77 Penrose, Edith, 12, 34 Perez, Carlota, 35, 234–35n4 perfect information, 16 perishability: of beef, 96, 126; of PCs, 177, 244n8; of beef and PCs, 21 perishable commodities, 21, 96, 126 perishable prices of computer components, 198 Perloff, Harvey, 233n1 Philadelphia, Pa., 81 Pinch, Trevor, 38 Piore, Michael, 46 Pirenne, Henri, 223 place and innovation in industrial systems, 236n15 planning, as hierarchies or visible hand, 167 planning to allocate resources, entrepreneurs in, 41 political rule making, 69 politics: in defining market boundaries, 69, 233n6; entitlements for rail and telegraph, 79; of interstate commerce, 23; of markets, 245–46n25; of market space, 88–89; of the national beef market, 132–36; and rule making, 64, 88, 140, 144

pooling arrangements, 101, 135, 240n10 population: expansion of, 68, 73; growth and distribution of, 85; growth of the Northeast, 74 pork, 98, 239n5 Porkopolis, 120. See also Cincinnati pork vs. beef, 95, 98(t), 239n4 Porter, Michael, 49 Powell, Walter W., 48 power of information and power of control, 132 power relations, 203 preservation of meat, 96 price: and time, 175; declines for computers, 216; uniformity and geographical barriers, 82 Price Club, 184 price-fixing schemes, 135 principle of markets (in central place theory), 70 private property rights, 30 process: and product, 92; improvements, 222 process-oriented innovation, 93 procurement and assembly operations, 165– 66 procurement, production, and distribution, 185 product customization and differentiation, 18 product cycle theory, 208–9, 245n22 production: coordinated administratively, 42; coordinated by market prices, 41; push system of, 176, 224; pull system of, 94, 125, 230 production and distribution network, 90 production side of value chains, 160 productivity paradox, 4 product life cycles, 177 profit cycles, 245n22 profit environment for economic actors, 231 project-oriented linkages, 48 protectionist rule making in politics, 88 proximity: of assembly in pull-systems, 230; of organizational assets, 229 pulled information channel, 150 pulling material to order, 195 pull-material-to-order system, 203, 211 pull systems, 94, 125, 230 push systems, 176, 224

278 Index Quanta, 209 QWERTY keyboard, 235–36n13, 242n4 rail and telegraph system: circulation of goods and information in, 67; as a commerce system, 63–70; competition with waterborne trade, 71–78; effect on commodity prices, 83; G. F. Swift & Company developments, 92–93, 98–101, 104–5, 107, 121, 126–27, 134; as infrastructure for direct-pull production-and-distribution network, 19; as infrastructure for massproduction and mass-distribution organization, 23; and the livestock trade, 80–90; in a pull system, 229; shipments of flour and grain, 1860–1890, 77(t); revolution of, 64; and standards, 55, 66; symbiotic development of, 65; ties to politics, 33 railroadization, 33 railroads: cartel agreements by, 240n10; cattle shipments, 98–102, 240n8; consolidation of, 66; cost reductions, 88; following grain cultivation, 76; grain to the East, shipments of, 76–77; interconnection and interoperability, 83; receipts at Chicago, 1852–1856, 76(t); reliability of, 77; transportation, costs of, 101 RAND Corporation, 145 range of markets, 70 Rankin, Thomas, 102 real-time information exchange, 6 refrigeration, 102, 104; experiments, 105, 107; of railcars, 94, 130; of warehouses, 108 regulatory barriers, 88 relationship customers, 193 reliability of railroads, 77 replenishment service centers (RSCs), 215 residual in the growth process, 34 retail butchers: in beef channels, 108(f); bypassed by Swift, 131; as customer of branch house, 93, 94, 224, 225; customization for, 226; in direct-pull system, 125; early trade of, 97; as a node in Swift’s production and distribution network, 5, 18, 100, 129; without middlemen, 121 reverse salients, 38, 241n2 rhymes in history, 220 Rhythm Collaboration Planner, 197 Ricardo, David, 30 Richardson, G. B., 7, 27, 47, 51, 52

Richardson, W. D., 241n15 risk management: mechanisms of power and control for, 8; and staging process, 195; theory of ownership of assets, 228; through cooperation between firms, 51; through vertical integration, 129–32; through virtual integration, 199–200 risks: and barriers to trade, 238n7; defined, 235n8; of inventory to suppliers, 198; of markets, 71; reduction of, by promoting intermarket activity, 81; in shipping, 77, 100 river trade, 74 river transport and pork, 239n5 Robertson, Daryl, 211, 219 Root, John, 91 roundabout (innovative) methods of production, 40 rules and rule making, 142; on Internet issues, 152; legal and political, 140 Russell, Charles Edward, 104 Sabel, Charles, 46 Saint Joseph, Mo., 94, 114, 119, 120 Saint Louis, Mo.: board of trade established, 81; livestock markets, 94; loss of grain trade to Chicago, 86; stockyard incentives to meatpackers, 117; as trading center, 73 Saint Paul, Minn., 94, 114, 117 Samsung, 209 Samuels, Levi, 112, 113 SAP, 189 Saxenian, AnnaLee, 50–51, 236n15 Schumpeter, Joseph: adaptive response, 211; on business cycles and innovation, 220; on capitalism, 234n3; diffusion as third step in business cycle, 223; innovation and creative response, 9, 16; innovation as creative destruction, 29–34, 92; leading sectors of the economy, 28; and the legacy of Marx, 30–31; on Marx, 234n1; as source for Chandler, 45; theory of economic development, 32; theory of innovation and economic growth, 30; on waves of innovation, 234–35n4 Schwarzschild & Sulzberger, 132 SCI Systems, 169, 210 search and selection for economic routines (technology), 35 Sears Roebuck, 156

Index 279 seasonality of meatpacking, 96 second Web revolution, 188–90 secure extranet, 193–94 sharp breaks of history, 230, 231 Sherman Act (1890), 134, 135 shipments of grain and flour from Chicago, 1840–1861, 75(t) shipping live cattle, 100 Siemens, 205 Sinclair, Upton, 91, 115 Singapore, 171, 209 Singer Sewing Machine Company, 89 size of beef and pork-packing industries, 1870, 98(t) size rank of largest cities, 1820–50 and 1850–90, 86(t) slaughterhouses, 95, 241n14 slaughtering: concentrations of, 119; for eight cities by dollar value of output, 119(f); meatpacking firms by size, 110(f); plants adjacent to stockyards, 18; plants as disassembly facilities, 18 SLCs. See supply logistics centers (SLCs) small-scale meat business decline, 94 Smith, Adam, 30, 40, 51 SMTC Corporation, 202 social constructionism, 38, 235n12 Sony, 209 the South: grain importer from the West, 73; lack of rail integration, 83; trade with the Northeast, 85–86 South Korea, 171, 209 space and time, 56 Special Economic Zones (SEZs), 245n20 speed: in production and distribution, 177; of railroads, 77; and time, 177 sphere of circulation, 5, 24 spoilage of meat, 96 spread and concentration: resulting from agency, 18; Dell’s interfirm network, 167; in direct-pull system, 19; of the economy by businesses, 3; geographical pattern of, 17; from a geography of supply, 207, 209; and globalization, 205, 211; globalization as, 20; with industrial district pattern, 6, 17; manifestation of innovation, 7; new geography based on, 113, 120–21; operations of Swift and Dell, 229, 232; as part of urbanization, 84; Swift & Company as agent of, 18

staging process and risk management, 195 standards, 142; agreements, 242n4; creation of, 140; for interoperability, 173; lack of, in telegraphs, 238n2; of Microsoft, preempting innovation, 235–36n13; setting of, 55; unsponsored, 173 Staples, 184 states’ rights, 88, 89 St. Clair, Lance, 192, 219 steamboats, 73, 74, 96 stockyards, 98–99, 241n16; in cattle procurement, 127; development of, 114; eastern, 97; expansion of, 116; as warehouses, 18 strategic forecasting, 192 strategic relationship with supply chain partners, 163 strategy-and-structure approach to organization of the firm, 43–44. See also Chandler, Alfred strikes, 133 structure: affecting suppliers’ decisions to locate, 210; and agency, 233n5; learning as function of agency and, 13; of history (longue dur´ee), 230, 232 Sturgeon, Tim, 49–50 subcontracting, 209–10 supplier factories: communications with SLCs, 197; as concentration of manufacturing, 7; Internet messaging to, 224; location decisions for, 207; locations of, 195; as a node in Dell’s organization, 19; order flow to, 224; proximity to Dell facilities, 207–11 suppliers: base of, 172; Intel as, 207, 215n14; inventory of, 201; inventory risk of, 202; in value chain, 194. See also component suppliers supply and demand balancing: Dell’s use of Internet information as inventory, 188; interfirm location premium, 203; of product flows, 166; Swift direct-pull system, 125; Swift vs. Dell, 226; telegraph communication in, 126; Web-based communication for, 190 Supply Assurance Program, 201 supply chain: geography, 183; management, 159, 190; partners, 163 supply logistics centers (SLCs), 19, 195, 196, 197, 198, 245n14 Swift, Edwin, 105, 107

280 Index Swift, Gustavus: business model of, 104; business organization created by, 92; description of, 104; early experiences, 103; experiments by, 92; innovations based on infrastructure changes, 90; as strikebreaker, 133; time and distance problems, 101 Swift & Company: as agent in spread and concentration, 18; as agent in territorial development patterns, 229; and the butchers, 107–8; distribution and retail sales, 109–10; Dressed Beef Trust and, 117; expansion and control by, 116–19; innovations by, 89–90, 92–93; investigations of, 111–13; labor and politics, 132–35; operations and management by, 121, 125–32; organization of, 93–94, 114–15; and the railroads, 105–7; transportation and delivery, 101–9; as user of technology infrastructure, 140. See also dressed beef; rail and telegraph systems; railroads; spread and concentration; telegraphy Swift & Co. v. United States, 135 Swift Brothers, 105 Swift Fertilizer Works, 130 Swift Premium (brand), 93 Swift Refrigerator Transportation Company, 130 Taiwan, 171, 209 Taiwan Semiconductor Manufacturing Company (TSMS), 209 Tandon, 169 Tandy, 171 Tatung, 170 Taylor, Bob, 145 TCP/IP, as standard, 242n9 technological clusters and business growth cycles, 32 technological dynamism, 30 technological paradigms or regimes, 35 technological presbyopia, 4 technology: defined, 233n3; and society, 237n25 Telecommunications Act of 1996, 152, 153, 243n12 telegraphy: as catalyst of economic and organizational changes, 53; change in time/ distance paradigm, 231–32; compared with Internet, 140, 241n1; cost of, 237–

38n1; and lack of standards, 66; role in procurement and order intake, 5, 126–27, 226; market integration, 81, 82; and opening of markets, 67; as precondition for Swift, 23, 45, 63, 64, 89; symbiosis with railroads, 65; systems, 66 territorialization, 202–4 territorial organization, 186 territorial relationships, 20 territory, 11, 228–30, 233n1 theory: of central places, 70; of cross-border production networks, 209; of economic development, by Marx, 31; of economic development, by Schumpeter, 32; of innovation and economic growth, 30; of product cycles, 208–9, 245n22; of structuration, 235n10 third-party logistics, 156 third-party logistics providers (3PLs), 196 time: and space, 15; and distance, 168 time zone standards, 66 Toshiba, 171 trade association of wholesale butchers, 133 trade: linkages, 64; as trigger of capitalist growth and change, 223 trade unions, 133 transacting, process of, 158 transactional customers, 193 transactions, 151, 199 transactions-cost approach, 43–44 transactions-cost reductions, 236n18 transfer. See diffusion Transmission Control Protocol/Internet Protocol (TCP/IP), 147 Transmission Control Protocol (TCP), 147 transport and communications, 70, 77–78, 104; breakthroughs in, 28; builders of, 54; business users of, 54; connection with innovation and business organization, 52; disruptive impact of, 54; infrastructure as catalyst of economic and organizational change, 53; role of, in economic life, 27; systems of, 11, 13, 14; urbanization effects of, 84. See also rail and telegraph systems travel time as market space, 78(m), 79(m) Turner, Frederick Jackson, 238n3 turnkey production network, 49 uncertainty, 235n8; in exchange, 43 uniform commercial code, 152

Index 281 uniform resource locator (URL), 148 unionization efforts, 133 Union Stockyard Company, 91 United States, 178, 183 urban impacts of interregional trade, 84–88 urbanization, 239n11; and city rankings, 85– 87, 86(t); consequences of, 84; economies of, 88; pattern of, 23; railroadization and industrialization of the United States, 1850–1900, 85(t) U.S. Department of Defense, 144, 145, 147 users: business, 54, 55, 56, 144; as buyers, 151; as innovators, 141; of the Internet, 140; providing agency, 141 U.S. Supreme Court, 89, 134, 135 value-added resellers (VARs), 184 Valuechain.dell.com, 194 value chains, 243n16; circulation side of, 160; defined, 49; described, 159–60; in manufacturing products, 157; production side of, 160; time-based inefficiencies of, 188 value creation, logistics as source of, 5 Vanishing Hand, 51, 237n23 Vernon, Raymond, 208 vertical integration, 17, 94; growth of Swift through, 129–32; vs. horizontal partnerships, 174; risk management through, 129 vertically integrated organization, 18, 129– 32, 199 VerticalNet, 163 Vest Committee, 112, 117, 240n9 virtual integration, 8, 17, 167 virtually integrated enterprise, 19, 198–202, 227 virtually integrated interfirm network, 200 virtual monopolies, 174 Visible Hand (Chandler), 45 visible hand (of management), 44, 46, 167

Wade, Louise Carroll, 22 wall of patents, 165 Walras, Leon, ´ 30 Walsh, Margaret, 22 water to rail- and telegraph-based trade, 75– 77 water transport, 67, 77, 96 waves of innovation, 234–35 Web-based communication, 190, 192 Web-based planning and execution systems, 198 Webber v.Virginia, 89 Web-generated information, 6 Web-sites, 151, 187(f) the West, 72, 73 Western exports shipped via New Orleans, 1839–1860, 74(t) Western foodstuffs, 238n4 Western frontier development, 238n3 Wheeler and Wilson, 89 wholesale butchers, 97; consignments to, 107; hostility of, 107; opposition from local eastern, 102; opposition to Swift, 108; resistance of overcome by Swift, 240n11; trade association of, 133 wholesalers, 226 Wilcox and Gibbs, 89 Williamson, Oliver, 42–44, 47, 167 World Wide Web (Web), 143, 148–52, 154 x.25 (European standard), 147 Xerox, 171 Xiamen, China, 205, 206, 245n20 Yeager, Mary, 22 Young, Allyn, 40 Zenith, 169