Ending a Career in the Auto Industry: '30 and Out' (Springer Studies in Work and Industry)

  • 60 27 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

Ending a Career in the Auto Industry: '30 and Out' (Springer Studies in Work and Industry)

Ending a Career in the Auto Industry "30 and Out" PLENUM STUDIES IN WORK AND INDUSTRY Series Editors: Ivar Berg, Unive

734 18 12MB

Pages 282 Page size 198.48 x 314.88 pts Year 2010

Report DMCA / Copyright


Recommend Papers

File loading please wait...
Citation preview

Ending a Career in the Auto Industry "30 and Out"

PLENUM STUDIES IN WORK AND INDUSTRY Series Editors: Ivar Berg, University of Pennsylvania, Philadelphia, Pennsylvania and Arne L. Kalleberg, University of North Carolina, Chapel Hill, North Carolina WORK AND INDUSTRY Structures, Markets, and Processes Arne L. Kalleberg and Ivar Berg Current Volumes in the Series: THE BUREAUCRATIC LABOR MARKET The Case of the Federal Civil Service Thomas A. DiPrete THE EMPLOYMENT RELATIONSHIP Causes and Consequences of Modern Personnel Administration William P. Bridges and Wayne J. Villemez ENDING A CAREER IN THE AUTO INDUSTRY "30 and Out" Melissa A. Hardy, Lawrence Hazelrigg, and Jill Quadagno ENRICHING BUSINESS ETHICS Edited by Clarence C. Walton LABOR AND POLITICS IN THE U.S. POSTAL SERVICE Vern K. Baxter LIFE AND DEATH AT WORK Industrial Accidents as a Case of Socially Produced Error Tom Dwyer NEGRO BUSINESS AND BUSINESS EDUCATION Their Present and Prospective Development Joseph A. Pierce Introduction by John Sibley Butler THE OPERATION OF INTERNAL LABOR MARKETS Staffing Practices and Vacancy Chains Lawrence T. Pinfield SEGMENTED LABOR, FRACTURED POLITICS Labor Politics in American Life William Form WHEN STRIKES MAKE SENSE — AND WHY Lessons from Third Republic French Coal Miners Samuel Cohn A Chronological Listing of Volumes in this series appears at the back of this volume. A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.

Ending a Career in the Auto Industry "30 and Out"

Melissa A. Hardy Lawrence Hazelrigg and

Jill Quadagno Florida State University Tallahassee, Florida


Plenum Press • New York and London

Library of Congress Cataloging-in-Publication Data On file

ISBN 0-306-45336-3 © 1996 Plenum Press, New York A Division of Plenum Publishing Corporation 233 Spring Street, New York, N. Y. 10013 All rights reserved 10 9 8 7 6 5 4 3 2 1 No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher Printed in the United States of America


During the 1980s the news media were filled with reports of soaring unemployment. Especially hard hit were the heavy manufacturing industries and a region of the country that had come to be known as "the rustbelt/' Industries once touted as the main engine of domestic economic growth—automotive, steel—had become stagnant or decrepit, seemingly at risk of being swept aside by foreign competitors. Hardly a week passed without an announcement of another plant closing or descriptions of the fallout from plants already closed. Tens of thousands of workers were being laid off, and in contrast to previous cycles of layoff these workers often had no hope of recall. Their jobs no longer existed. "Downsizing" and "restructuring" had become the new watchwords. One strategy used by firms to manage their work force reductions focused on means of increasing the attractiveness of their pension plans—in particular, their ear/y-retirement plans. Each of the "Big Three" auto makers—General Motors, Ford, and Chrysler—adopted this strategy. The often realized threat of layoff and loss of recall rights provided an obvious "negative" incentive for workers to revise their expectations of continued employment in the auto industry. But because of the seniority principle in union contracts, this pressure fell primarily on younger workers. By augmenting the positive incentives—financial and otherwise—in their earlyretirement plans, the companies (and the union, the United Auto Workers) hoped to redistribute some of the job loss across age groups. Accordingly, under certain well-defined circumstances, eligible workers were offered the opportunity of early retirement with substantially enhanced pension provisions at unusually early seniority dates. Some of the workers jumped at the opportunity. Others preferred to stay on the job, and many of these workers did stay, but some of them, worried that worse might be coming, chose against preference and elected early retirement. The situation seemed well suited to a study of the retirement process under complex conditions and varied options. The more we pursued pre-



liminary investigations, the more evident it became that the train of events still unfolding in the auto industry provided a valuable opportunity to combine sociological and economic perspectives in a detailed study of factors involved in workers' decisions about the timing of retirement—decisions that for some workers entailed an abrupt, imanticipated departure from a "career job," for others a speedup of a long-planned event, and for most a heightened degree of uncertainty in the consequences of a decision not to retire despite the considerable inducements being offered. Rather than survey workers across all of the U.S. automakers, we chose to focus on the largest of the companies. General Motors, because it was still in the throes of restructuring. Despite having reorganized its production facilities and management structure during the early 1980s, GM continued to lose market share. Throughout the second half of the 1980s and well into the 1990s, GM was still closing plants and shedding droves of workers. If ever the decision to elect early retirement is made in a casual, leisurely way, this was surely not the occasion for it. Senior workers were feeling pressures from various quarters—not least of all, of course, from their junior coworkers whose jobs were on the line in both senses of that phrase. The transition to retirement has been the subject of a good deal of scrutiny since the first studies of adjustment to "old age" were conducted in the midst of the Great Depression. A number of the issues that motivated those earlier studies seemed equally salient in the present context. For example, how does the threat of unemployment influence the retirement decision of workers still in their 50s? Does the evident lack of employment opportunities for younger workers influence the decision? But the context of retirement decision making during the 1980s was also profoundly different as compared to those earlier times. For one, union-negotiated contracts in the manufacturing industries now included provisions regulating queuing behavior in a firm's internal labor market (e.g., layoff queues). For another, manufacturing workers now had a guaranteed retirement income in the form of a private pension coordinated with Social Security benefits. Seniority rights and federal legislation such as the Age Discrimination in Employment Act afforded senior workers substantial protections against an unwanted retirement, even in the midst of large and growing reductions of labor demand. Formal protections preserve options that might otherwise be withdrawn or radically amended. They do not, however, guarantee substantive rationality. Nor do they necessarily simplify the decision-making process, especially when the multiple options carry multiple, often conflicting pressures and implications. The availability of retirem^ent income has been one of the "pull" factors emphasized in numerous studies of retirement conducted during the past three decades. The pull of income security, which motivates voluntary withdrawal, is often defined in opposition to "push" factors such as job



dissatisfaction and health constraints. The conceptual problem with attributing retirement to either push or pull factors is that it assumes each sort of factor is constitutionally independent of the other sort of factor. It ignores interactions between positive incentives such as a guaranteed retirement income and negative motivations stemming from the fear of losing one's job and/or from anxieties about retiring "too soon" and not being able to find another job. Further, by concentrating on motivations of retirees, the decision-making processes of those who are subject to similar constraints and incentives yet do not retire are overlooked. The retirement options GM offered its workers provided a testing ground for theoretical models of the decision-making processes. An early-retirement program had been in operation since the 1960s. But during the 1980s—and in particular during the 1987-89 GM-UAW contract, which is the temporal frame of our study—General Motors announced the availability of "special early retirement windows" for specified categories of workers. This special plan waived some and altered other of the usual eligibility criteria for early retirement. At the conclusion of the 1987-89 contract period, we interviewed more than 1,700 male auto workers, all of whom had been eligible for early retirement under the terms of GM's longstanding early-retirement plan and some of whom had also been eligible to retire under terms of the special plan. While fewer than one in five of the sampled population elected early retirement under either plan, we oversampled each of the two categories of retirees, ending with nearly equal numbers of interviews of workers who retired under the regular early-retirement plan, workers who retired under the special option, and workers who continued their GM employment beyond the end of the contract period. Thus, we were able to examine comparatively the decisions of workers faced with different options in a largely shared context of immediate and impending shrinkages of labor demand. Chapter 1 introduces the framework of issues and describes the sampled population. Chapter 2 recounts some major features of historical context, including the development of retirement programs in private industry, characteristics of the employment system, and the deteriorating conditions of the domestic auto industry during the late 1970s and 1980s. In Chapter 3 we turn to the general and technical features of the retirement plans in effect during the 1987-89 contract period, compare them to features of previous plans, and evaluate the financial consequences for auto workers who elected retirement under one or the other of the two earlyretirement plans, in comparison to the financial consequences for workers who continued to defer retirement. Chapters 4, 5, and 6 explore various facets of the retirement decision. In Chapter 4 we estimate the relative strengths of various determinants and correlates of the decision (e.g., pension wealth, health capability, marital status, skill level, plant closure) and



compare outcomes between the two early-retirement plans. We also document some of the perverse consequences that arise from a strategy to accelerate the flow of early retirements as a means of dealing with excess labor supply. Chapter 5 is concerned with aspects of the workers' efforts to acquire information and advice pertaining to the retirement decision. In Chapter 6 we evaluate the extent to which issues of generational equity and altruistic intentions toward younger workers influenced the retirement decisions. Chapter 7 shifts to a consideration of postretirement characteristics and demonstrates how experiences of the retirement transition itself, as well as experiences of worklife before the transition, affected satisfaction with life in retirement. Finally, in Chapter 8 we summarize our results and discuss some of the ways in which private pensions are affected by public policy decisions. Although this study is specifically of auto workers, implications of the analyses have wider applications at a time when companies continue to cast off thousands of workers in a frenzy of mergers, acquisitions, downsizings, and restructurings. If the recent record is any indication of future trends, the decision-making processes here described for auto workers will be shared to some extent by large numbers of workers in other industries during the years ahead. As the employment relationship is increasingly individualized, the period at the end of an employment contract will ever more often say, "until further notice." On the other hand, generalizations are perforce limited by history: "Things change." Or, put less idiomatically, the process of deciding whether to elect early retirement at a given age is itself sensitive to historical context. Our auto workers are of a historical context which included an employment system that is now in decline. One difference already apparent concerns the fact that white-collar middle-management workers have recently lost much of their relative immunity to the effects of corporate downsizing and restructuring programs. In 1993, for example, middle managers accounted for 6 percent of the total work force and 22 percent of all layoffs (Marks 1994). The changing organizational technology of office work promises more to come. But middle managers are not protected by union-contract guarantees of seniority rights, income during layoff, and related benefits. Indeed, "laid off" in the white-collar world of middle management is typically just another euphemism for "fired." A parallel case study of the retirement behavior of middle-management workers in a large corporation undergoing major restructuring efforts would afford some interesting comparisons.



ACKNOWLEDGMENTS We are indebted to a number of individuals and organizations who have been especially helpful at various stages of this project. William Hoffman, Director of the Social Security Department, United Auto Workers, was an invaluable resource and collaborator in the initial stages of the project—from the construction of the sampling screens, to the formulation of interview questions, to the assembly of information about the retirement plans and the sample members' employment characteristics. A grant from the American Association of Retired Persons (AARP) Andrus Foundation supported initial stages of the research. Later stages drew on financial support provided by the Mildred and Claude Pepper Foundation. Able research assistance came from Deborah Andrews, Rachel Barich, and Kim Kruse. Portions of the text have benefited from the comm.ents of Richard Burkhauser, David Macpherson, and Joseph Quinn. Finally, the hundreds of auto workers who participated in the study: But for them and their generosity of time and understanding duringjiiterviews, this book would not be. Patches of their lives are reflected in their responses to our questions and thus in the analyses we have made of those responses. Because of the analysis techniques used, the heterogeneity of the workers' voices of experience has been transformed into the heterogeneity of population-based series of individual characteristics. Consequently, though we have written about the workers, little that was individually distinctive in their lives on the assembly line is recognizable in the text. We do, however, recognize their individual contributions, for which we are very grateful.

Contents Chapter 1 Internal Labor Markets, Plant Closings, and Retirement Introduction The Present Study Methods of Study

1 X 7 13

Chapter 2 Some Matters of Context


Introduction The Trend toward Early Retirement The Development of Private Pensions Collective Bargaining and Pensions Model Contracts and Pattern Bargaiiung Private Pensions as a Union Benefit Hard Times in the Auto Industry ACloser Look at GM, circa 1987 Early Retirement as a Siphon of Excess Labor

21 22 25 31 31 34 35 42 44

Chapter 3 The Financial Structure of Early-Retirement Pensions


Benefits and Incentives The GM Pension Plans The Early Years The 1987-89 Plan

49 54 54 56

xii Evaluating the Financial Incentives Comparing Monthly Benefits Calculating Pension Wealth Accrual Profiles Pension Characteristics of the Samples Predicting Pension Wealth Consequences of Delay

Contents 59 61 66 70 74 78 80

Chapter 4 Predicting Early Retirement Why Retire Early? A Basic Model of Early Retirement Pension Wealth and Related Variables The "Stickiness" Factor The Wage Effect and Occupational Status Other Incomes and Income Demand Health Some Background Factors Layoffs and Plant Closures Predicting Age at Retirement Regular versus Special Early Retirement Comparative Sensitivities to Pension Wealth SERP as a Contingent Avenue SERF and Age at Retirement Assessing the Basic Model

87 87 91 93 101 104 106 107 110 Ill 112 115 120 122 123 123

Chapter 5 Discussing Options in an End-Game


The Value of an Informed Decision Choosing Potential Informants What Advice Did the Workers Hear? Effects of Discussion on the Decision

135 138 147 151



Chapter 6 Solidarity and "Generational Equity"


Industrial Restructuring and Generational Conflict Young Bloods and Old-Timers Special Early Incentives Senior Workers' Attitudes Translating Attitudes into Behavior Why No Difference?

163 165 168 171 179 184

Chapter 7 Aspects of Postretirement Satisfaction


Satisfaction, Agency, and Autobiographical Memory Framing the Timing Problem Retiring Too Soon Status Satisfaction

191 196 199 209

Chapter 8 Inducing Early Retirement: Some Conclusions in Perspective . . . .


Revisiting Matters of Context Rationality as Plan and Behavior The Changing Parameters of Private Pensions

221 225 232







1 Internal Labor Markets, Plant Closings, and Retirement

INTRODUCTION A "firm"—that is, a signature (firtna), the name or title under which a company transacts business—is a legal entity organizing the complex processes by which conflicting interests and objectives of multiple, heterogeneous actors are maintained in a framework of contractual relations of exchange oriented by what the firm defines as a mutually agreed objective or output. The specific character of a firm is defined in part by an organizational technology that describes its output as a function of a structure of tasks or "jobs." This job structure is the core of a firm's internal division of labor and an internal labor market of renewable contracts. Workers flow into and out of the internal labor market via pathways established in some sort of employment system which regulates the pricing of labor time. The prevailing system of employment in the United States has traditionally been one in which firms have had great flexibility in making firmspecific adjustments in employment, that is, in altering labor supply to fit labor demand. When economic conditions are good, more workers are hired; when conditions worsen, unwanted workers are shed just as quickly. Thus, this system of employment is characterized by relatively high rates of turnover. Spells of unemployment tend to be short, but so too is the average length of job tenure. One consequence is that, with high rates of turnover, firms are more reluctant to invest very much in worker training. For a firm that requires a skilled labor force, high turnover rates are costly. An alternative system—commonly designated "closed," in contrast to the above-described "open" system—seeks to replace the flexibility of repeated cycles of "hiring and firing" with the rather different flexibility of repeated cycles of "layoff and recall."^ In effect, under the closed employment system a firm maintains an internal labor market into which workers 1


Chapter 1

are hired and from which workers leave, but also within which workers circulate between two statuses of "active employment"—^namely, "at work" and "laid off, but subject to recall." An obvious advantage to the firm is lower turnover, hence greater protection of the firm's training investments. To borrow Hirschman's (1970) felicitous vocabulary of analytic framework, the closed-system firm tends to substitute mechanisms of "loyalty" (i.e., worker attachment) for mechanisms of "exit" (i.e., firings, quits) as a means of regulating labor supply during the crests and troughs of the firm's "business cycle." One of those mechanisms of loyalty involves the notion of a "career job": a new employee is brought into the internal labor market with the twin understandings that circulations between "at work" and "laid off" will be more or less frequent but that a loyal employee can look forward to long-tenure employment in the firm (though this latter part of the agreement is implicit). Any worker on layoff (or "at work," for that matter) has the option of exit (quitting) for another, presumably better employment opportunity. However, the expectation of long-term employment with the current firm serves as a major part of the glue of employee loyalty—especially when, and insofar as, the given employment entails firm-specific training.2 Another part of the glue stems from a difference in reward structures. In the open-employment system worker productivity is treated as a given. The firm's aim is to match a wage to a given worker's evaluated marginal productivity; if both buyer and seller of the labor time agree to the offered match, the worker is hired or not fired; if not, the offer goes to the next candidate. By contrast, in the closed system productivity is treated as a variable that can be influenced by the wage relation; thus, the wagerate schedule of rewards emphasizes incentives designed to increase worker effort, ability, and, in the process, loyalty. Firms that use the "layoffrecall" scheme thus tend to pay somewhat higher wage rates, as an incentive to gain more effort and as an incentive to maintain loyalty in the face of periodic layoffs. The closed-system internal labor market poses some regulation problems of its own, and it is here that unionism—or a specific form known as "job-control unionism"—has been especially instrumental (and here also that the third of Hirschman's triadic framework, "voice," comes most into play). The internal labor market is typically characterized by a highly formalized, detailed contract which relates a ladder of sharply delineated jobs, task descriptions, and expectations, on the one hand, to workers' rights, obligations, and wage rates, on the other (the "job classification system"). The union controls the career ladder and career-income schedules through a set of seniority rules, which also govern the allocation of internal vacancies to candidates for promotion and the allocation of layoffs when the firm's management declares a need for layoffs. The rule of seniority essentially establishes an intertemporal "loyalty contract" in which agree-

Internal Labor Markets, Plant Closings, and Retirement


ments are made between past and future conditions. Thus, in the choice of layoff cycles over reduced earnings through reduced hours or reduced wage rates as a means of regulating fluctuating labor demand, senior workers have preferred to deflect the main cost of that regulative procedure to junior workers, while junior workers have been willing to accept the disproportionate cost in return for future compensations (or, if not, they exit the internal labor market). This job-control unionism retained some major elements of the tradition of "union voluntarism," as it was known in the days of Samuel Gompers (or "business unionism," in its later description), insofar as conventionally (i.e., narrowly) defined "economic interests of the workplace"—contract bargaining, short-term cooperative interests, industrial action at the plant, and so forth—continued to be emphasized in preference to more broadly defined interests of political-econom.ic reform.^ But job-control unionism more deeply engaged management's interests in molding and executing strategies for the regulation of labor supply, including the use of "no strike" provisions and the union's commitment to discipline rank-and-file initiatives. The U.S. auto industry, including both the manufacturing firms and the auto workers' union (the UAW),^ was prominent in developments of the closed employment system—as indeed it has been prominent in many other developments during this century. Ogburn's (1947, p. 4) testimonial 50 years ago may now seem quaint for his choice of comparatives, but not for the underlying point: "There is reason to remark that the inventors of the automobile have had more influence than Caesar, Napoleon, and Ghengis Khan." This impact has been due not just to the automotive vehicle as a physical device, with all of its manifold repercussions, difficult to exaggerate though they may be. In addition, the organizational technology of the vehicle's production—the assembly line and all that it has involved in the parceling of work tasks across an abstract grid of "time and motion," but equally importantly as an innovative corporate form of organizing the multiple relationships of employment—has had profound, if sometimes less visible, consequences for the textures of work experience during the second half of the 20th century. One auto maker in particular, not Ford but General Motors, had by the end of World War II established itself as the leading model of this new corporate form, a multidivisional organization that emphasized decentralized administration of semiautonomous units under a single overall coordination. GM became the setting of several wellknown studies of innovation in organizational technology (e.g., Drucker 1946; Chandler 1962). GM's plants also became the setting of several wellknown studies of blue-collar work life and labor-management relations— among them. Walker and Guest's (1952) study of GM's Framingham, Massachusetts, plant, then the most modern of all assembly line plants; Chinoy's (1955) study of GM's Oldsmobile plant in Lansing, Michigan (also the set-


Chapter 1

ting of Form 1973); and Rothschild's (1974) study of GM's Lordstown, Ohio, plant. And GM's political-economic clout became world renowned, not least because of an oft-quoted (and slightly misquoted) remark miade in Congressional testimony by Charles E. Wilson, long-time head of General Motors and, at the time of the testimony, Secretary-Designate to the Department of Defense in the first Eisenhower Administration: "What is good for General Motors is good for the country. "^ Shortly after the end of World War II, GM and the UAW completed a series of contract negotiations that rapidly developed the organizational form of the closed employment system. Details of those developments will be described in the next chapter; for the moment it suffices to say that most of the innovations were keyed to expectations of worker loyalty (from management's point of view) and to expectations of job security (from the union's point of view) in regulating the internal labor market. From today's vantage it might seem odd that a major firm in an industry in which most blue-collar employees were semiskilled at best would have had much of a stake in absorbing the costs of implicit long-tenure contracts, since the main advantage of such contracts from the firm's standpoint is in protecting investments in worker training. But two factors were important. First, the automotive industry had a long record of "boom and bust" cycles (given the price of the commodity relative to disposable income, consumption of the industry's product was extremely sensitive to interest rates, money supply, etc.), and those cycles could be expected to continue in the consumer-driven economy. So having a "warehouse" in which to store temporarily surplus labor during downturns was an attractive option—especially in view of the second factor, which has to do with worker training. Semiskilled though they were by today's standards, proficient assembly line workers ("production workers," as they are known in the industry language) were not readily available "off the street. "^ The required skills did involve some investment in training, in the conventional sense of occupational-skill training. But at least equally important were the socialbehavioral skills demanded by the organizational technology of the labor process: clock punctuality and spatial coordination across atomistic task performances, subordination to the physicalist logic of the machine system, and so on. A behavioral regimen taken for granted today was still alien to most of the rural migrants flocking to the industrial cities and filling factory rosters.'' This aspect of "training to the job" involved considerable investment from the firm's point of view. For an industry eyeing the prospects of rapid growth in consumer demand for a product with built-in "style obsolescence," having a labor force that could respond flexibly to the ups and downs of product demand represented a significant advantage. For the workers, on the other hand, that flexibility meant that, despite the comparatively high wage-rates which the industry paid, part-year work

Internal Labor Markets, Plant Closings, and Retirement


was rather frequent (more or less, depending on the state of the economy), and annual incomes were therefore often much lower than the high wagerates otherwise implied. At least since the mid-1930s unionized workers had been lobbying for a guaranteed annual income, but to no avail (see, e.g.. Fine 1969, pp. 60-61). During the quarter-century following World War II, manufacturing workers in general, and none more than auto workers, learned to appreciate regular increases in wage rates, in standard of living, and in the influence of their unions. Lacking enough of a cultural-historical perspective, they thought their relatively prosperous situation would last forever, or at least throughout any foreseeable future. The unusual conjuncture in geopolitical-economic relations which had put the United States in such an overwhelmingly dominant position in world markets of capital, and thus of wage labor, was just that—a highly unusual historical conjuncture. But for the large majority of people who were living the heady experience of world dominance from within the center, it increasingly seemed a natural state of affairs, an endlessly secure unrolling of "the American dream" (if only "external enemies" could be held at bay). That this internal prosperity was heavily subsidized—^by a geopolitical-economic dominance that ensured cheap energy supplies, external sources of cheap "raw materials," and abundant still undeveloped consumption markets for the steadily expanding production of goods and services—was an unwelcome message even when offered by "friendly critics," and more especially unwelcome when the support structure began to fray. Although the process of fraying was much too complex to be adequately understood in terms of any single event, and certainly began long before 1973, the "oil shocks" of the 1970s are by now a standard emblem of the profound structural changes that pulsated, sometimes violently, throughout the fabric of work life in a large segment of the U.S. population. And if the "oil shocks" became emblematic of the "causes" of those changes, the plight of "the manufacturing industries" (long regarded as the essence of "industry," the very meaning of the word), and of the plight of the automotive industry in particular, became the most visible emblem of the consequences. Where once most adults understood and accepted the central meaning of Charles Wilson's famed boast, they now were forced to make sense of iconic "brand name" industrial firms shedding tens of thousands of high-wage jobs in the United States and exporting labor demand to new plants in low-wage countries. By the 1980s the auto industry felt under siege on every quarter, now most immediately by the worst economic recession in decades. Industry troubles prompted the formation at MIT of an "international automobile program" of inquiry into the future of the industry (Altshuler, Anderson, Jones, Roos, and Womack 1984). Concerns about the survival of the auto industry in the United States were partly motivated by a recognition that


Chapter 1

motor-vehicle-related occupations directly accounted for 1 of every 11 members of the U.S. labor force: 1.2 percent in automotive manufacturing; 2.8 percent in sales and service; 1.0 percent in highway construction and maintenance; 4.1 percent in operating vehicles for commercial freight and passenger haulage (Altshuler et al. 1984, p. 7). And indirectly many other areas of the economy felt repercussions, since the auto industry accounted for more than half of the rubber and malleable iron, nearly half of the machine tools, and large portions of other industrial materials consumed each year in the United States. Alarms about the fate of this one industry were often indistinguishable from concerns about the fate of the U.S. economy in general. Auto makers' responses to their individual and collective predicaments were varied, though most of the responses have been popularly collated under the barely euphemistic phrase, "industry restructuring and downsizing." One response was to renew and expand experiments in "employee involvement" programs (also called "Quality of Work Life" programs), a newly furnished variant of a quasifamilialist corporativism designed at one level to maintain worker loyalty in the face of hard times but at another level to construe "industrial relations" in a way that would obscure the basic structural conflict between buyers and sellers of labor time by securing the conviction that all interests were fundamentally the same.^ Other responses, which formed the immediate and painful context of that renewed mission, involved the permanent closure of many plants, cutbacks and retoolings of other plants, assignments of labor demand to relatively low-wage countries, and excisions of large numbers of "excess workers" in the United States. Several of these responses will be taken up again in the next chapter. Here, for introductory purposes, we focus on one specific form of the auto makers' means of reducing the population size of their internal labor markets—namely, accelerating the flow of retirements. By the end of the 1950s, the auto industry boasted one of the most strongly attached labor forces of any industrial sector. The line between the traditional blue-collar orientation to occupational activity as "just a job" and the traditional white-collar emphasis on "career" had been blurred at least to the extent that large numbers of semiskilled and skilled workers had developed strong loyalties to an industry, and to specific firms within it, that seemed to promise careerlike ladders of advancement in a lifetime employment. Often the promise went unfulfilled, as Chinoy (1955) documented in his famous study. Automobile Workers and the American Dream. Nevertheless, the loyalty remained strong among auto workers who had gained some tenure in the employment system, in part because of comparatively high wage-rates and in part because an increasingly significant portion of their wages took the form of deferred compensation—a retirement wage and related retirement benefits. But what was once appreciated as a major asset of the industry—strong employee loyalty—had, by the 1980s,

Internal Labor Markets, Plant Closings, and Retirement


been transformed into something of a liability, at least in the view from company headquarters. The need to shed large numbers of strongly attached workers was complicated by the fact that this employee "stickiness" had been deeply institutionalized in the rule of seniority, one of the linchpins of the hard-won accommodation between management and unionized labor. Insofar as the reductions of labor supply were regulated strictly by seniority, the result would be an abrupt aging of the work force. Thus, from a standpoint largely shared (though for different reasons) by management and union leadership, senior workers who had reached eligibility age for early retirement were seen to be an obvious pool of candidates for removal from the internal labor market. Both management and union shared an interest in future auto workers, and younger workers had greater claim on that future than did older workers. The task, then, was to redistribute risk and opportunity—the risk of job loss, the opportunity for future job security—across age groups in a way that would implicitly contradict the "last in, first out" rule of seniority, without abrogating the basic principle of a seniority-regulated internal labor market. A "command" approach to the desired goal was out of the question, for reasons of legal contract (and behind that, management's interest in maintaining the union's cooperation). Achieving the desired goal would have to depend upon the right mix of incentives, incentives designed to decrease the exit costs for senior workers whose demand for the exit option was elastic. Since the 1960s, contract settlements between the UAW and the Big Three auto makers had included a lucrative package of benefits for workers who retired "early," that is, before the "normal" age of 65. Indeed, this early retirement plan, initiated in an agreement between the UAW and GM, was and continued to be one of the best in any industry Successful though it had been in "ordinary times," the package of incentives was not producing the desired volume of exits during the 1980s, when plants were being shuttered and tens of thousands of younger workers were heading the queue into permanent layoff. Additional incentives, targeted on senior workers in specific plants, would perhaps sufficiently augment the outflow of workers who could be expected to leave through the ordinary earlyretirement channel. THE PRESENT STUDY The immediate context of the present study began November 6,1986, when General Motors armounced that nine of its plants would be closed entirely, another two partially, by 1990, with most of the closures scheduled to be completed before the end of 1987.^ That dating of the context is in some respects a convenience, since these 11 plants were not the first of GM's closures; auto workers had seen or heard about such announcements


Chapter 1

before, and no doubt imagined that more would be coming, perhaps in their own neighborhood. But that is not to say that the workers had grown accustomed to the news, drawn up their separate bargains with fate, or suffered the burdens of job displacement any less for having the knowledge that others elsewhere had already gone through the experience. The workers directly affected by the closures announced on November 6, estimated at 29,000, did have several months in which to decide alternative courses of action. And senior workers usually had more latitude than their younger co-workers, since the typical practice in plant shutdowns was to cease operations in stages, during which time workers were "bumped" into layoff status according to seniority. But the several months were a period of uncertainties, anxieties, and animosities for most workers, young or old. As a result of previous negotiations with the union, GM agreed to offer a special package of incentives designed to reduce exit costs via early retirement for qualified workers as young as age 50 who were directly affected by plant closures. Officially termied the Special Early Retirement Plan (as distinguished from the long-standing Regular Early Retirement Plan), the package can be viewed as a test of management's ability to predict the relative proportions of "alert" and "inert" responses to alternative incentive structures (Hirschman 1970, pp. 24-25). At some point on a gradient of possible incentives, senior workers' elastic demand for the exit option would be sufficiently matched, from management's standpoint. The question was. Where that point? The chief "positive" incentive consisted in an offer of enhanced financial benefits to qualified senior workers who had not otherwise signaled interest in an early exit. Perhaps equally importantly, all of the positive incentives of the Special Plan were framed—and, one would expect, reinforced—^by the "negative" incentives of extended layoff, the uncertainties of recall or transfer to another, perhaps distant plant, social pressures (real and/or perceived) of expectation from junior co-workers who were first to bear the brunt, and appeals to worker solidarity in "sharing the burden." Did this Special Plan include the right mix of incentives? The brief answer is "No," inasmuch as management's hopes were not satisfied. For example, whereas GM's goal for 1988 had been to induce 7,500 early retirements by means of the Special Plan, only 3,600 workers actually did so. Similar ratios occurred the following years as well. But although management's goal was never met, large numbers of senior workers did respond "alertly" to the special package of incentives. And at the same time, many other senior workers elected an early end to their careers in the GM plants by taking the long-standing avenue of Regular Early Retirement, perhaps at least partly because of the pressures and uncertainties associated with the plant closures announced in November 1986. Our aim is to investigate within that context a variety of determinants of the early-retirement decision, using data generated from interviews with

Internal Labor Markets, Plant Closings, and Retirement


representative samples of auto workers who elected early retirement during the 1987-89 contract period and auto workers who were eligible to do so but declined. The basic question of our investigation is straightforward: What factors distinguished those workers who did retire from those who did not? An important dimension of that question consists in the structure of incentives in each of the early-retirement plans: Under what conditions did the incentives achieve the intended effect? It should be clear that the latter question is only a component of the former question. Hirschman's contrast between "alert" and "inert" responses to proffered incentives is useful, in that respect, for underlining the simple fact that an "incentive" is only as good as the intended recipient's evaluative perceptions of self-interest, options, and probable consequences of alternative courses of action within a trajectory of life events. Not that the general category of an alert response (i.e., "to retire") was exceptional for the population of senior auto workers; we can safely assume that virtually all of them hoped to end their GM careers by means of retirement. Nor that any of the workers were inert, in the sense of being oblivious to the various retirement schemes on offer; we can safely assume that the drama of an industry in serious trouble had caught the attention of all, giving the issue of alternative exit mechanisms more than usual salience. Rather, the degree to which any given worker was "moved" by the incentives of an early-retirement plan depended at least to some extent on a disposition to be moved—that is, on a readiness to retire now—and that disposition was no doubt a function of numerous factors, some of which had little or perhaps even nothing to do with the motivational field addressed by the incentives. Thus, our effort to assess the effectiveness of the incentives to retire, intended to alleviate some of the problem of redundancy in the firm's internal labor market, will turn on our ability first to grasp those individual dispositions and their conditions. We have at our disposal not only the data generated from interviews of the sampled population but also a quite substantial body of research literature—or rather, bodies of literature, for the present investigation stands at the intersection of three distinctive research topics and settings: the auto worker, plant closings, and retirement. We draw on all three bodies of previous research (and occasionally some others as well) in subsequent chapters. Here, because the collective volume of those literatures precludes anything like a systematic review, we offer some brief orientational descriptions. There have been many studies—some once widely heralded—of the behaviors, attitudes, and aspirations of workers in the automotive industry Prominent examples of the early research include Walker and Guest's (1952; Guest 1954) study of a newly opened plant in Framingham, Massachusetts, Chinoy's (1955) study of auto workers in a Michigan plant in 1946-47, and Kornhauser's (1956) study of auto workers' political attitudes


Chapter 1

and behaviors. Why auto workers? An important part of the answer was captured by Chinoy's (1955, p. 12) observation that they "work in a glamorous, [then] relatively new industry whose growth has dramatized the American tradition of opportunity, but whose present character makes it extremely difficult for them to realize the American dream." Rare today, no doubt, a depiction of the auto industry as "glamorous." But for half a century or more, no other image better symbolized the mass-production technologies of a consumption-driven economic order of style and prestige than did the image of "Detroit's latest models" rolling off the assembly line. Moreover, few other corporations have carried as much weight in the political economy of a nation as the major auto makers—especially General Motors, long regarded as the bellwether of Wall Street's impending future. The industry has been repeatedly approached as a quasi-laboratory for the investigation of various structural changes and trends (e.g., in ideologies of human management and workplace authority) for the simple reason that it has often been a pacesetter in those changes and trends. And increasingly throughout the quarter-century from 1948, auto workers were viewed as leading examples of the unprecedented growth of a "blue-collar middleclass prosperity." Once a leading figure in the drama of strikes and unionization, the auto worker of the 1950s and 1960s was more often featured as a personification of exchanges between "the good life" and "the alienated life" (e.g., Chinoy 1955; Blauner 1964; Form 1973). Likewise, studies of "plant closings, community abandonment, and the dismantling of basic industry"—to borrow the descriptive subtitle of Bluestone and Harrison's (1982) landmark treatment—have dealt with dramatic events and with the disparities that stand between expectation and realization. Ranging from Wilcock and Franke's (1953) survey to Aiken, Ferman, and Sheppard's (1968) study of workers displaced by the demise of Packard Motor Car in 1956, to Lipsky (1979), Aronson and McKersie (1980), Buss and Redburn (1983), Craypo and Davisson (1983), Newman (1985), Rothstein (1986), Perrucci, Perrucci, Targ, and Targ (1988), and Dudley (1994), this line of research has produced a second and partially parallel fund of insights based primarily on case studies of specific sites. Several of these sites have been auto plants and their surrounding communities. But even where not, the garnered insights generally apply to a broad range of industrial plants, since plant closures tend to share a number of important features in terms of legal issues (e.g., Rothstein 1986; Wendling 1984), economic, political, and social ramifications for the host community as well as for the displaced workers themselves (e.g., Dudley 1994; Flaim and Seghal 1985; Podgursky and Swain 1987), and psychological health effects among the workers and their families (e.g., Hamilton, Broman, Hoffman, and Renner 1990). Because of the concentrated volume of closures during the late 1970s and the 1980s, the imagery of abandoned industrial plants often

Internal Labor Markets, Plant Closings, and Retirement


served as a catalyst for a sort of "national ritual/' as Dudley (1994) put it, ritualized talk about deep structural changes in the nation's economy. Caught up in those changes and in the anxieties and conflicts being laid bare, men and women whose workplaces had been shuttered behind them voiced some biting critiques of the current state and prospects of "American industry," even if those critiques were sometimes heavily tinged with nostalgia for an era long gone (e.g.. Hamper 1991; Newman 1985). Studies of retirement, though of more recent vintage, have been commanding greater shares of the page space of social science research, as various aspects of population aging have moved to the fore in some of the discourse of public policy. But save for a very few relevant observations—for example, that age is one of the strongest predictors of the likelihood of reemployment among workers displaced by a plant closing (e.g., Aiken et al. 1968, p. 31; Wilcock and Franke 1953, p. 57)—studies of auto workers and plant closures have been mostly silent about retirement as a form of labor-force mobility and more particularly as a form of managing labor-force capacity within an industry. There is one large exception to that generalization, of course—the vanguard study of early retirement among auto workers by Barfield and Morgan (1970). The present research was designed in part as a replication and extension of Barfield and Morgan's investigation, though obviously in a quite different historical context. Their study, based partly on a sample of the general adult population of the United States and partly on a sample of auto workers during the mid-1960s, addressed a range of potential determinants of the early-retirement decision and subsequent attitudes toward the decision—for instance, age, marital status and household size, education, financial considerations, ability to keep up with the job, recommendations by co-workers and union or management officials, and so on. These are the main aspects of their study that we have sought to replicate and extend (Chapters 4,5, and 7). In addition, however, we examine (Chapter 6) another set of issues in connection with early retirement as an avenue of mobility out of the firm and industry—namely, issues concerning the distribution of unemployment risk or, conversely, employment opportunity across age groups. Barfield and Morgan's study was conducted during what were still halcyon days for the U.S. auto industry. The decision of a worker to elect early retirement, or not, had few if any negative implications for the continued employment of younger workers in a given plant or firm—or, so far as anyone then imagined, for later generations of younger workers as well. Times were different during the latter half of the 1980s, when the interviews that comprise the main data base for this study were conducted. On the one hand, the implications of early retirement for future generations of workers had become a prominent topic of public policy debates. Issues involved in the problem of how opportunity and risk, at collective as


Chapter 1

well as at individual levels, are and ought to be distributed into unknown (future), relative to known (present), conditions and capacities—issues sometimes gathered under the heading "generational equity"—were producing much heat along with some new insights and new research agendas. Various observers were noting, as did Brudney and Scott (1987, p. 20), som^ething "paradoxical" in the fact that while "institutionalized policies and pressures have encouraged older Americans to retire at ever-earlier ages,... now older Americans are being accused of knocking off work and enjoying a high standard of living at the [present and future] taxpayer's expense." On the other hand, the auto industry, highly prone to "boom and bust" cycles since 1919, had long used retirement (first, mandatory retirement; then early retirement) as a means of rejuvenating its blue-collar work force. When those halcyon days that were the historical context of Barfield and Morgan's study ended, as they had well before our study, the industry's effort to shed unwanted workers by expanding opportunities and incentives for early retirement were augmented by expectations—voiced by younger workers in the pages of Solidarity, the UAW's official newspaper, and elsewhere—that senior workers ought to retire early for the good of the current generation of younger workers. Thus, a question we address in addition to those shared with the Barfield-Morgan study asks whether workers retired early because they agreed with and were motivated by the sort of redistributive ethic featured in those expectations—that is, an ethic of redistributing the risk of indefinite layoff or permanent job loss among younger workers by converting some portion of it into early retirement among workers in their 50s. Did older workers' agreement or disagreement with that ethic, or with specific components of it, make any difference in the likelihood of their retiring early? As a working hypothesis, the question admits varying expectations about the effect of agreement. For example, did agreement increase the likelihood of retirement in contradiction of "self-interest" factors such as the size of the expected retirement wage? Did agreement modify the efficacy of one or another of those self-interest factors and thereby increase the likelihood of retirement? Or was agreement consequential only in the margin, discriminating probabilities of retirement only when self-interest considerations had been satisfied? We would not expect to see systematic evidence in support of the first of those three alternatives (nor, it would seem, did company management; otherwise they would not have expanded "self-interest" incentives). But the second alternative is plausible relative to the third. It is certainly easy to imagine that a great many junior workers had hoped, and had expressed their hopes in shop floor talk, that they could count on a climate of worker solidarity and "social acceptance" in which early retirement would be viewed as "the right thing to do" (see Frolich and Oppenheimer 1992).

Internal Labor Markets, Plant Closings, and Retirement


METHODS OF STUDY The event of central interest is defined as a completed decision to elect early retirement under the terms of one of the early-retirement plans that were part of the 1987-89 contract between General Motors and the United Auto Workers. The temporal frame of that event is the 3-year contract period, which began in October 1987. During that contract period General Motors announced the opening of a "special early-retirement window," effective from November 1987 through March 1988. The significance of that "window," briefly depicted in the preceding discussion, will be described in some detail in the following chapters; for the moment it suffices to repeat that certain workers in specific GM plants—plants scheduled for closure or sharp reduction—^were offered an opportunity to elect early retirement under a special plan which included some enhanced incentives, financial and otherwise. During that same interval, workers who satisfied minimum eligibility requirements, whether directly affected by plant closures or not, could also elect early retirement under provisions of the Regular Plan. Our interest is in both groups of retirees and in the relevant "at risk" population of auto workers who continued to defer retirement to some unspecified future date. The sampled population consisted of all male workers who were employed in GM's plants as production or skilled-trades workers as of October 1987, who were aged 50 or older as of October 1987, and who were eligible for early retirement under the terms of at least one of the early-retirement plans.io This population was stratified into workers who were still employed in GM's plants at the end of the contract period and workers who had elected early retirement under either of the early-retirement plans during the 3-year interval, 1987-89. Probability samples were constructed within each of the population strata, ii Individual sample members were initially contacted by mail, with an explanation of the study and a request for cooperation. Interviews were conducted by telephone between November 1989 and January 1990. The defining characteristics of the sampled population imply some obvious limits on the extent to which results of the analysis of retirement decisions can be generalized. First of all, the population is male, blue-collar, and of relatively long tenure in a major mass-production manufacturing industry. Most of the men were or had been employed in various types of production plants (assembly, trim, foundry, etc.) located in central-city or suburban places in the old "industrial heartland" (Michigan, Ohio, Indiana, Illinois, Missouri).^^ The age, gender, and ethnic composition of employees in industrial plants tends to vary across different plant locations (i.e., rural, suburban, central-city), across different regions of the nation, and across different types of plant (see, e.g., MacLennan 1985).


Chapter 1

Moreover, the fact that this is a study of retirement decisions over a short span of time during the late 1980s implies some restriction of the intersection between cohort biographies and historical periods. The average age of the men at date of interview was 57 (see Table A.l for other descriptive statistics). The large majority ranged between 53 and 61 years of age, which means that most of the men (1) were born either just before or during the Great Depression, (2) were too young to have served in World War II, though some of them probably served in the Korean War, and (3) began their paid-employment work lives (though not necessarily with GM) during the 1950s. This last trait means that most of the men entered the labor force when the union movement was at or near its apogee. The UAW in particular was then in the process of consolidating its organizational leadership at the national level, reining in the energies of rank-and-file militancy, and achieving a certain sort of accommodation with management. It was a time of "labor peace," to use one popular expression of the day, and of "consumer-oriented normalcy," to use another apt phrase. Flint, Michigan, which had been the site of so many overt contests during the preceding years and decades, was lauded by Look magazine as an "All-American City" in 1954 (Edsforth 1987, p. 218). Auto workers were being socialized in the politics of a corporate-style "union voice" as a means of resolving (or at least addressing) disputes between labor and management (see Edsforth 1987, pp. 200-219; Freeman 1989, pp. 169-196; Zetka 1995; Zieger 1986). That period of relative comity extended well into the 1960s, when the youngest members of our sampled population were beginning their employments with GM. If measured in terms of wage-rate gains and fringe benefits, the two decades were a promising time for blue-collar careers in the auto industry—assuming, of course, one could withstand the often punishing work assignments. That last observation points to another, very important caveat about the limits of generalization: the sampled population consists of survivors. That is to say, our analysis of retirement decisions pertains only to male auto workers who successfully fought against the physical and mental challenges of working in the plants long enough to have reached eligibility for early retirement. Many other men of the same birth cohorts (though how many others we do not know) worked alongside the members of our sampled population at one time or another—6, 8, perhaps even 10 or more years—and then left for employment elsewhere. Some of these latter men (again, the number is unknown), also in their mid-to-late 50s in 1989, could have been in employments where they, too, had opportunities of early retirement, perhaps under similar plans. In any case, we have no information specific to the "nonsurviving" erstwhile members of that larger population of men born October 1937 or after who were at one time or another, for one tenure length or another, employed in a GM plant. Our study necessarily pertains only to the survivors.

Internal Labor Markets, Plant Closings, and Retirement


Our analyses are guided in part by a theoretical perspective which can be described as a "rational-action" perspective. While that label, or some variant of it, has attained considerable prominence in recent decades by becoming attached to more or less distinctive, sometimes controversial groups of theory, we intend nothing special by it. In fact, the implicit (whea not explicit) point of departure of most social science theory has always been built around some version of a rational-action model of human behavior. Informed by the sort of egocentric bias that renders experience of "the world as it is from my point of view" (which in turn is situated within and dependent upon an "ethnocentric" or cultural-structural bias), this very general, highly abstract model assumes as central tendancy that human beings live their lives as consciously, sometimes self-reflexively deliberative beings engaged in subjectively meaningful, purposive behaviors (to use a vocabulary that has come to be stylized as "Weberian"). The deliberations tend to be episodic; yet an actor can usually supply a reasoned account of what he or she is doing at any given moment, how it fits within a biography, and why that rather than some other action is being undertaken. The deliberations tend to be consequentialist, in the sense that alternative actions tend to be sorted and weighed in terms of anticipated consequences; yet the sorting and weighing are obviously not the cerebrations of an omniscient, omnipotent, rational-calculating machine. Actions do often go wrong—perhaps more often than not, if measurement is precise enough and effects are traced far enough—although the degree to which they go wrong is not always very noticeable or important. Actions do produce consequences that the actor had not foreseen, and often could not have foreseen, much less intended. Yet actors frequently do decide among alternative actions or courses of action; these decisions usually matter, sometimes exiguously but other times in ways that far exceed the actor's perceptual horizon; and the decisions tend to be calculative, however imperfectly, in terms of expectable or anticipated consequences, on the basis of information and options at hand. Of course, to say that this general rational-action model is highly abstract is to acknowledge that it accommodates a great deal of variation on factors which are not describable, easily or at all, along a rational-irrational dimension. As numerous observers have insisted—including observers who have placed considerable weight on this very general rational-action model even as they have otherwise diverged just as considerably (e.g., Adam Smith, Karl Marx, Max Weber, Georg Simmel)—decisions and actions taken are often as much or more a function of forces of habit and sentiment than of the reasonings which are typically depicted as acts of rational calculation. Habits and sentiments as well as reasonings, and their consequences, are selectively institutionalized as reproducible and reproductive (not to say static) organization, which itself manifests a kind of rationality ("rationality as form") that need not be isomorphic with the ratio-


Chapter 1

nality of any given deliberative actor. Yet this reproductivity of organization—the tendency of an organization to persist—^not only creates but indeed is a particular selection and distribution of knowledges (or "informations"), options, incentives/disincentives, and so on, within which actors' deliberations, decisions, and actions exist as ranges of possibility arid probability, potentiality and actuality, anticipation and remembrance. To some considerable extent, in fact, that reproductivity is coercive in the very same instance that it is productive; it makes some of the decisions or choices "for" actors, as it were, often without their awareness. Thus, while our data analyses will draw upon some very specific, highly developed theoretical approaches within the general rational-action perspective—collectively referred to as utility theory or rational-choice theory—it will be at least as often the shortcomings as the main strengths of those approaches that provide orientation. Standard utility theory has been remarkably productive in generating a framework within which the dynamics of a variety of decisional problems can be rigorously studied. Indeed, as Plott (1987, p. 117) suggested, it has been one of the most productive of any "obviously false" theory yet devised. Given the complexity that each of us imagines herself or himself to be, the fact that a very simple algorithm, composed of a small handful of variables, performs as well as it does in a variety of contexts could be regarded as astonishing. ^^ x^g performance is of course always an approximation relative to the complexity of particular experience—a necessary cost of the parsimony. But in various model situations, where the options are well defined (as judged by the players, not just the theorist-observer), where the relevant information is reasonably accessible and unambiguous (again, as judged by the players), and where the players manifest socialization biographies that imply reasonably predictable understandings or prejudgments of the rules of the game, the approximate results of data analyses demonstrate that people do often tend to behave as, or as if they were, utility maximizers. However, just as approximations come in degrees, so do the various attributes of model situations. To list a few examples: (1) The range of what might count as a "utility" or "good" to be maximized allows much greater diversity than the typical model situations are capable of handling (e.g., not only nonquantifiables, in the sense of present limits of mensural technology, but also substantive rationalities that are by definition obscure in dynamics). Likewise, (2) actors' evaluations of information concerning options, constraints, risks, and other components of a decision dynamic are contextdependent, and often highly sensitive in that dependence, yet the typical model situations involve a rather narrow range of (often highly stylized) decisional contexts, such as gambling or lottery behaviors in an experimental setting. (3) Standard models assume that judgments about utility and judgments about event probabilities are independent; this important

Internal Labor Markets, Plant Closings, and Retirement


simplifying assumption allows reasonable approximations in some contexts, but in many others it is profoundly mistaken. And (4) standard models assume that actors experience uncertainties (of information, of constraints, of outcomes, etc.) in very well-formed ways, yet even in rather simple conditions uncertainties are often ambiguous or fuzzy, so much so that they do not conform to the sort of single first-order probability distributions that the concept of risk assessment implies.i^ In sum, when "the very ideal conditions" of the standard model situation "cease to hold"—to borrow Arrow's (1987, p. 201) locution—"the rationality assumptions become strained and possibly even self-contradictory. They certainly imply an ability at information processing and calculation that is far beyond the feasible and that cannot well be justified as the result of learning and adaptation/' Rationality is always contingent on a context of meanings; thus, "rational action," strictly speaking, is never a founding action. Obviously the rather broad rational-action perspective which we use as partial guide in the analyses of auto workers' decisions, being highly abstract, is empirically incomplete. The more closely one approximates the quotidian elements of the conditions and experiences of any given actor, the more that incompleteness is telling of the inability of a rational-action model to make sense of specific outcomes. We have already alluded to some of the constraints on options that formed part of the organizational setting within which the auto workers' deliberations and outcomes had the meanings they did have. We will consider some of those constraints at greater length and in greater detail in later chapters. But it should be evident that, for most of the analyses we undertake, those organizational circumstances must be treated as givens, not as variables, since our study deals with employees of one firm at one juncture in the firm's (and industry's) history. 1^ By the same token, in referring as we just did to the meanings of the workers' decisions, we are acknowledging a fabric of sentiments, habits, motivations, deliberations, and intentions that is empirically far richer than the threads of our data. We simply do not know all of the contents of the workers' meanings, nor can we assume that those meanings, if known sufficiently, could be ordered along a grand scale of commensurability. In moving toward a decision whether to retire "now" or "later," for example, the workers may well have been deliberating over conditions and prospects of event that varied markedly from person to person in substantive meaning—that is, deliberating not simply about different quantities on a grand scale of value ("the value of retiring now rather than later") but indeed about various and at least partly incommensurable values which they had to orchestrate in some fashion when deciding whether to retire "now, under these conditions and prospects," or "later, under (hoped for) better conditions and prospects."^^ As previously noted, we do not subscribe to the erstwhile standard assumption of decision the-


Chapter 1

ory that actors' "preferences" or value assignments are independent of their substantive beliefs. Nor do we find troubling an insight which has upset previously contented assumptions about tight linkages between valuation and choice behaviors—^namely, that the cognitive-evaluative processes involved in actors' assignments of value to alternative goods differ in importantly consequential ways from the cognitive-evaluative processes involved in making choices among the alternative goods. ^^ Although the data we analyze do not allow direct access to the processual dynamics of either sort of behavior, our interpretive inferences are based on the assumption that value assignment (or "preference") and choice are loosely coupled.

NOTES 1. Bear in mind that the polar terms, ''open" and "closed/' simply pertain to analytic models which have variable "descriptive adequacy" in relation to the employment practices of any particular firm. For more detailed recent treatments of employment systems, see Kochan, Katz, and McKersie (1986), Hounshell (1984), Katz (1985), Althauser (1989), Weiler (1990), Sayer and Walker (1992), Bridges and Villemez (1994), and Pinfield (1995). 2. Indeed, although the expectation has typically been founded on only an implicit understanding, recent court rulings have held that employers' signals of that implicit understanding, such as occur in employee handbooks or other official documents which serve to build employee loyalty, are legally enforceable implicit contracts (see Edwards 1993; Kochan, Katz, and McKersie 1986; Rosen 1994; Weiler 1990). 3. The policy of the American Federation of Labor stressed the importance of maintaining a careful distance toward political parties, the periodic contest for control of the state bureaucracy and its rewards, and any organized movement that was identified itself as (or was identified with) a class-based politics. Gompers' "voluntarism" notoriously included opposition to specific reforms proposed by and/or supported by the national government, including legislation concerning minimum wages, maximum hours of work, and old-age pensions (see Gompers 1914 and 1915; also, Hattam 1993; Horowitz 1978; Rogin 1962; Yellowitz 1965). 4. The UAW's official name is "The International Union, United Automobile, Aerospace & Agricultural Implement Workers of America"; we will refer to it by the more usual shorthand name, the United Auto Workers, or simply the UAW. Likewise, when referring to the top echelon of union leadership we will write "national," not "international." 5. According to the transcript of the Senate confirmation hearings, Wilson's remark, given in reply to a question about potential conflict of interests between his corporate background and the duties of his would-be governmental office, was that he could "not conceive of one because for years I thought what was good for our country was good for General Motors, and vice versa. The differ-

Internal Labor Markets, Plant Closings, and Retirement


ence did not exist. Our company is too big. It goes with the welfare of the country" (see Gordon 1995). Of all the famous words never quite uttered as they were later famously remembered, this instance involves a barely significant difference. But however one wishes to distribute the emphases across Wilson's recorded words, the basic statement being conveyed was anything but fantasy. GM then accounted for roughly half of the market in an industry that had already remade the social, economic, political, and physical geography of much of the world, with much more soon to come. Wilson, however, was not exactly the champion of labor that Gordon strives to depict (see Chapter 2, below). 6. While the standard image of "the auto worker" centers on the assembly line, in fact only about one of every five auto workers was actually "on the line" (a proportion that changed very little over the decades). However, the tasks of most "production workers" were strongly governed by the pace of the assembly line. 7. As Doeringer and Piore (1971, p. 13) noted, the various requirements of on-thejob training were one of the stimulants to the development of internal labor markets in large firms. In his study of the work force of a milling plant that had been outfitted with the latest in automated technology. Walker (1957, pp. 24-25, 40) found that "the battle between men and machines" continued for 3 years before an "equilibrium was established between the complex behavioral and technical forces involved." Even though many of the members of the new mill crew had worked together in older milling plants, the "change to the new mill nonetheless meant a change in both functional and social relations with fellow workers for most men" (see also Chinoy 1955, pp. 70-72,85-86). 8. The impulse behind these programs (to which we return in the next chapter) was hardly new. The auto industry in general, and GM in particular, had promoted a familialist image of "employee involvement" as early as the 1930s, in an effort to persuade workers of the irrelevance of independent unions (see, e.g.. Fine 1969, pp. 23-28,42-47,187-188). 9. One of the nine was Fisher Body No. 1, the massive plant which had been a key site of the famous sit-down strike in January and February of 1937, as a result of which GM recognized the UAW as official bargaining agent of GM's workers (see Edsforth 1987, pp. 170-189; Fine 1969). 10. The population was restricted to male workers because even with oversampling we were unable to gain a sufficient number of female production and skilledtrades workers aged 50 or older; they comprised a very small proportion of the total blue-collar labor force of the plants. 11. Additional information descriptive of the population strata and the sampling is available in the Appendix. 12. That five-state area accounted for three-fourths of the auto industry's labor force, and three of every five of the industry's facilities, during the mid-1980s. Michigan alone accounted for as many employees and facilities as the four other states combined. 13. The grounds for astonishment recede, however, when one considers that those algorithms, or the behavioral patterns which they partially formalize, are not simply ex post inventions of a sagacious theorist. They are integral to a broad historical-cultural formation in which formal, procedural, and substantive rationalities of instrumentalism are highly prized, not merely as a machinery of




16. 17.

Chapter 1 external relations and not only as an implement in a toolkit of applications to narrow technical problems of modeling or puzzle solving, but also as a set of characters by which people judge each other and themselves—indeed, a set of characters which people have increasingly become, in the historical production of competent actors. We are referring here to the technical concept of risk, which implies that one has a basis for selecting, from the range of all possible probability distributions, the single distribution that governs the given event. It is by virtue of this feature, for example, that one can calculate the risk of rolling a seven with a pair of balanced dice. Uncertainties greatly exceed risks, in volume, partly because of mensural limits, partly because of practical capacities of calculational skill, partly because of substantive rationalities, and partly because of nonrationalities. For a sampling of both earlier and more recent treatments, see Anand (1993), Edwards (1954, 1961), Ellsberg (1961), Koons (1992), Marschak (1975), Hausman (1992), and Savage (1954). Certainly we do not intend in that an endorsement of the available set of options as necessarily the optimal set from any particular point of view—management, union, employee, or any other. Rather, we intend simply a limitation of this study. Incommensurability defines a rather severe limit to the consequentialist principle of the rational-action model. For recent treatments of the limits, see, e.g., Raz (1986, pp. 322-326) and Stocker (1990, pp. 184-188,291-302). In the standard literatures of economics and psychology these differential processes are treated under the heading of "preference reversal" phenomena. For a sampling of treatments and evidences, see Grether and Flott (1979), Lichtenstein and Slovic (1971), Machina (1987), Tversky, Slovic, and Kahneman (1990); and for an assessment of the issues, Hausman (1992, pp. 227-244).

2 Some Matters of Context

INTRODUCTION The main events analyzed later in this book—workers' decisions about ending their careers in the auto industry via retirement—^were most immediately events of personal biography Workers who had contributed tens of thousands of hours of paid labor time during the course of 20 or 30 or 40 years of employment in the production facilities of General Motors grappled with the difficulties of scheduling a major transitional event of the life cycle: Should I retire now, or should I wait? For each individual worker it was primarily and largely the quotidian character of biographical experience that furnished the territory of that decision with intensities of meaning. Age, health status, family relationships, financial circumstances, the prospects of alternative or "second-best" employment in the local labor market, the attractions or anxieties (or both) of "quitting work"—these and similar other traits of personal experience levied their shares in the weight of the decision. But like all events of biography, these individual decisions were also nodes in a long and complex historical network of processes and events which had been instrumental in the specific biographical formations to a far greater extent than the biographies had been in forming even the most recent of the historical events and processes. Not that the auto workers were epiphenomenal to their own destinies. But nor were they in charge of the shadow world of their inheritances. With respect to the broad skein of conditions that had led to the actuality of particular options, contingencies, and expectations which composed the meanings of the decisional task, the workers had doubtless been "acted upon" much more than they had been determinative actors. The present moment of the decision came as a momentary end, a particular conjuncture in each biography, of developments in the political economy and public policy of a nation, in employer and union policies and practices concerning the regulation of labor in correla21


Chapter 2

tion with regularities of the life cycle, and in the vitality of one particular industry and, within that, of one particular company. Our intent in this chapter is to offer accounts of some of the more salient of those developments. The accounts are necessarily not constructed through the lenses of the individual biographies; the means of that are lacking. We offer instead synoptic accounts, brief and incomplete. They are derivative of, not substitutes for, standard histories of the topics addressed, and they are mostly tuned to the interests of this investigation of early retirement among auto workers during the 1980s.

THE TREND TOWARD EARLY RETIREMENT Best estimates indicate that labor-force participation rates among older men in the United States changed very little during the first four decades of this century (Ransom and Sutch 1986). Perhaps the secular trend toward lower rates began in the 1930s, but it was not until after World War II that persistent pronounced declines became the norm. Initially, the bulk of the reduction occurred among men aged 65 or older. But by the 1960s a second, parallel trend had begun, this one in the proportion of workers retiring "early" (i.e., prior to age 65), as observers were noting a steady growth in the relative numbers of "men who are well enough to work and who might get some kind of job if they were interested, but who prefer the leisure of retirement" (Epstein and Murray 1967, p. 105). Between 1970 and 1993 labor-force participation rates among men aged 62 through 64 fell by a third (from 69 to 45 percent). While much of that decline can be linked to the availability of Social Security benefits at age 62, notable decreases occurred also among men aged 60 or 61 (from 83 to 67 percent), and even among men aged 55 to 59 a downward drift in rates was unmistakable. Nor has the U.S. trend toward ever younger ages at which men retire been unique. Similar trends have occurred in Canada and in many European countries, although the pace of decline has varied from country to country (e.g., relatively small changes in Sweden, Norway, and Italy, but large declines in the United Kingdom, France, and the Netherlands; (see Layard, Nickell, and Jackman 1991, pp. 506-507; Quadagno and Quinn 1996).^ Most of the factors contributing to the downward drift in retirement ages have been articulated, directly or indirectly, in the "social reconstruction" elements of public policies concerned with problems of the regulation of labor (e.g., rising unemployment rates). Often these policies have deliberately promoted early retirement through a variety of legislative actions designed to reconstruct the incentive structures of public pension systems. The particular actions have not been uniform from country to country. But, generally speaking, eligibility ages have been lowered, requirements have

Some Matters of Context


been liberalized, and enhanced earnings-replacement schemes have been designed to encourage older workers to leave the labor force. Moreover, reforms of welfare programs not directly related to the public pension funds have enabled and sometimes encouraged early departures. For example, in the Netherlands, Germany, Sweden, and Great Britain disability insurance provides a source of income for unemployed and underemployed older workers, and as the eligibility criteria have been broadened rates of early retirement via a ''disability" status have increased (Guillemard 1991a; Guillemard 1991b; Jacobs, Kohli, and Rein 1991; Laczko and Phillipson 1991). Likewise, in some countries unemployment compensation programs have been altered in ways that encourage retirement; instead of providing only temporary income to ease workers through a period of layoffs, unemployment benefits have become a permanent pathway to retirement (see, e.g., Laczko and Walker 1985). Since the substantive as well as the procedural features of public pension schemes vary from country to country, and since the rate at which men have been retiring at younger and younger ages also has varied from country to country, it is perhaps tempting to conclude that the trend toward early retirement has been driven by liberalizations of the public pension systems. However, while provisions of public pensions have undoubtedly influenced early-retirement behaviors in each of the several countries for which systematic evidence has been assembled and studied, there is no simple pattern of correlation between variation in those provisions and variation in early-retirement rates. Indeed, Esping-Andersen and Sonnberger (1991, p. 230) concluded that cross-national differences in age-specific rates and trends of early retirement are greater than one would predict from differences either in the respective public pension systems or in the economic circumstances of the several countries. That the trend toward younger retirement ages has been widespread among countries of the North Atlantic region is indisputable. So, too, that it has been sensitive to more than economic fluctuations. Public policy formations, considered in historical-cultural context, have influenced retirement behaviors both directly and through complex interactions with changing contingencies of the market economy. But one should bear in mind that "public policy" in this regard has involved much more than the set of "entitlement programs" that provide direct income-replacement "transfer payments" such as public pensions and liberalized disability or unemployment compensation schemes. Instruments of public policy also operate in less direct ways—through tax codes, for instance, and through employment and employment-related regulations that encourage certain practices by employers and discourage others. The incentive structures pertaining to a worker's choice between continued labor supply and retirement are complexly molded at the intersection of these direct and


Chapter 2

indirect avenues of public policy instruments as they are manifested within specific market conditions, organizational cultures, and individual perceptions. A case in point can be seen in the United States, where the trend toward younger ages of retirement has been comparatively steep even though the public pension system has remained among the most conservative. The eligibility age for Social Security was lowered to age 62 for men in 1961 but only with a reduced-benefit provision (whereas in a number of other countries full-benefit eligibility occurs at age 60). And income from neither public disability nor unemployment compensation schemes is sufficient to make permanent withdrawal from the labor force an affordable option for most U.S. workers younger than 62. What, then, accounts for the large decline in participation rates among men who are still in their 50s? Has it occurred independently of instruments of public policy? Some observers have argued that it has. Esping-Andersen and Sonnberger (1991, pp. 244,247) concluded that the U.S. trend has been "largely a self-generated autonomous process happening independently of labormarket constraints,... not observably related to either push [i.e., rates of unemployment and labor-market contraction] or pull [retirement-incentive] effects," and that "the older worker has not borne the brunt of economic restructuration" practices. Since their conclusion was based on broad-grained data and a consideration of only the public pension system of retirement incentives, it is understandable that they found the trend-producing processes in the U.S. case relatively opaque. The accelerated flow of early retirements among men in the United States has occurred mainly in response to private-sector, employer-provided pension plans. As Esping-Andersen (1990, pp. 79-88,150-151) recognized elsewhere, unless one attends to the incentive structures of these pension plans—which form the primary and often the only source of income for workers who retire before age 62— the dynamics of early-retirement behavior in the United States cannot be adequately grasped. While employer-provided plans are properly termed "private-sector plans,'' that distinction can be overdrawn. Major features of the plans have been shaped by instruments of public policy, both deliberately and indeliberately. A few examples are noted in the next section, an overview of the development of private pension plans. Before turning to that overview, we should offer readers who are unfamiliar with the general structure of private pension plans as they exist today a brief description. Employer-provided plans are generally classifiable as either defined-benefit (DB) plans or defined-contribution (DC) plans. The distinction dates from the earliest private plans. While the two types of plan developed alongside each other, DB plans were increasingly preferred by firms during the early decades of this century (Latimer 1932, Vol. 1, pp. 43-48).

Some Matters of Context


In a DB plan the employer promises to pay each qualified worker a specified monetary benefit during retirement. Qualification is governed by 'Vesting" rules, the chief rule being that the worker must remain with the given firm for a specified duration before gaining pension rights. DB plans are not portable. In a DC plan, by contrast, the employer contributes a specified amount into a pension account, the amount calculated as a specified percentage of the worker's annual earnings. The worker may also make contributions to the account; under some DC plans this is compulsory. DC plans typically have shorter vesting periods, and they are portable. The size of the monetary benefit at retirement varies according to the value of the pension account at that date, which in turn is a function both of the accumulated contributions and of the success of investment decisions made by the individual worker or by the administrator of the plan (see, e.g.. Jeweler 1993; Schmitt 1993). Since DC plans are portable and place more of the decisional risk on the individual employee, they have been less instrumental in strategies of labor-force regulation. Historically, the longer vesting periods of DB plans, combined with the fact that they are not portable, have been effective in depressing labor-turnover rates. Until 1989, when federal legislation altered the rules,2 almost 90 percent of DB participants were subject to a 10-year vesting period (Clark and McDermed 1990, p. 26). In other words, orUy after 10 years of credited service with the given employer was a worker eligible to receive retirement benefits. Moreover, a vested worker who left the firm prior to a specified age was typically penalized by an actuarial reduction in the benefit entitlement.^ This, too, tended to depress labor-turnover rates. The pension market has been shifting from DB to DC plans in recent years, partly because newly enacted plans are more likely to be DC plans and partly because employment has been shifting away from industries that had initiated the DB type of plan (see e.g., Clark and McDermed 1990; Gustman and Steinmeier 1992). Also, it is worth noting, recent evidence indicates that participation in employer-provided plans has been declining among younger male workers, apparently because of the greater voluntarism of DC plans (see, e.g.. Even and Macpherson 1994).^ Since the present study is concerned with workers covered by DB plans, we will confine the following discussion to the development of DB plans, which still comprise the large majority of employer-provided plans.

THE DEVELOPMENT OF PRIVATE PENSIONS While not available to large numbers of employees (excepting corporate executives) until the middle decades of this century (Morrison and


Chapter 2

Roberts 1982, p. 97), employer-provided pension plans have been used as a component of strategies for regulating labor supply since the late 19th century. Apparently the first comiprehensive plan by a large private corporation in the United States was enacted by the American Express Company in 1875. The idea gradually spread to a few other large firms, especially the railroad companies (beginning with the Baltimore and Ohio in 1884), which set the principal pattern of benefit structures. According to this "railroad formula," an eligible worker's pension benefit was calculated at a base rate of 1 percent of the worker's wage, multiplied by the number of years of service, and was linked to a mandatory retirement age, usually age 65 or age 70. By 1900 a dozen plans were in operation. During the next two decades the number grew to 270, as many of the major industrial corporations—for instance. United States Steel, Westinghouse, Standard Oil, John Deere, Du Pont—established plans, mostly along the lines of the railroad formula. Du Font's plan, for example, paid a benefit calculated at a base rate of 1.5 percent of the worker's highest monthly salary, multiplied by years of service, with a vestment period of 15 years (Fischer 1977, pp. 165-167; Graebner 1980, pp. 120-125; Haber 1978; Latimer 1932, Vol. 1, pp. 20-26; Zunz 1990, p. 191). Because the record of evidence is sparse and partly anecdotal, systematic inferences about motives and intentions in the initial development and adoption of private plans are difficult to make. But judging from a survey conducted by the National Industrial Conference Board (1925), as well as various studies undertaken by state commissions and other investigators (see, e.g., Epstein 1922; Latimer 1932), it appears that the initial plans were responses mainly to organizational interests of the firm as internally calculated by company management. The union movement, still in its infancy, was occupied with more fundamental issues of survival and legitimacy, and more often than not viewed employer-provided plans suspiciously as an instrument of competition for worker loyalty. A few unions attempted to field their own pension plans (e.g., the International Typographical Union in 1908), but most of these were precarious ventures (Epstein 1922, pp. 193-210; National Industrial Conference Board 1925, pp. 66-67).^ There is in any case little to suggest that employees themselves had been clamoring for comprehensive pension plans. This is not to say that workers were generally unconcerned about postretirement income security. Surely they were concerned. But for most people these concerns were still framed within traditional forms of intergenerational transfer (i.e., reliance on one's adult children, relatives, etc.) rather than innovative forms of deferred compensation and public provision, which were widely viewed as "socialistic" or "communistic" invasions of the moral order. Insurance companies offered private annuity plans for individuals interested in "saving for old age," but very few people were attracted to them. Since the annuities typically gave

Some Matters of Context


extremely low rates of return (because of high "administrative costs" and other fees charged against the premiums), one might conclude that most potential customers were too astute to participate. However, state-administered voluntary annuity plans in Massachusetts and Wisconsin fared little better, despite significantly higher rates of return (Epstein 1922, pp. 274-279; Fischer 1977, pp. 168-169). Although life expectancy for men aged 60 in 1900 was 14.4 years among whites and 12.6 years among blacks (as compared to 18.4 and 16.2 years in 1988), most adult men apparently discounted their "distant futures" rather heavily, even when they had sufficient disposable income to invest in those future years, perhaps in the belief that they would not live to harvest long-term investments. Collectively as well as individually, of course, the dilemma stemmed from the limits of prescience: those persons who could least afford to make the investment at time^ were most likely the very people who at time^^^^ stood in greatest need of having made such investments (Latimer 1932, Vol. 1, p. 49; National Industrial Conference Board 1925, p. 66). Employer motivations in devising (or adopting) private pension plans, though undoubtedly mixed and varying from case to case, were generated mainly from organizational characteristics of the firm. The period from the 1880s to the 1920s witnessed a profound transformation in the organization of capital and thus in the organization of labor and employment relations, as a new corporate liberalism and progressivist ideology (and policies) remade the commercial, legal, and political landscapes (see, e.g., Horwitz 1992; Sklar 1988; Zunz 1990). A shift from proprietary to corporate, bureaucratic organizational forms entailed new understandings of an increasingly extensive domain of "employee management," some elements of which included attention to the costs of labor turnover in large firms that depended on increasingly (and internally) trained workers. Innovative interests in "personnel psychology" and "scientific management" soon came into play, as did other, related responses to the perceptual field of a bureaucratized organization of work. No doubt some employers offered pension plans because of genuine concerns for the well-being of their older workers, reflecting either remnants of the disappearing moral order of the proprietary form (in which employees were often family members, relatives, friends) or the moral order of a new corporate paternalism. But a major motive behind the early adoptions of private plans stemmed from employers' interests in reducing total wage bills by shedding the labor costs of older workers whose returns to company investments were diminishing—that is, by regulating the exit behavior of workers who had been loyal and valued employees but whose productivity had declined "because of age."^ Rather than follow what had been the typical practices—reassigning older workers to less demanding tasks (often at reduced wages) or simply firing them—employers learned that by offering a pension they could define an


Chapter 2

"expected retirement age" (indeed, often a mandated age)^ and thereby induce older workers to leave according to the employer's schedule, without damaging the loyalty of their most productive workers, those who had developed skills from long experience at the job but who had not yet entered their "declining years." The presence of a plan, it was thought, would alleviate anxieties about the future among the firm's most productive workers and thus bind them even more tightly to the firm, lower the turnover rate, and raise the quality of the firm's work force. Moreover, in virtually all cases the plan was explicitly defined as a revocable benefit (not a contractual obligation) subject to the private law of the firm; often the revocation trigger was linked to "inadequate service" including participation in strikes. The pensions were granted at the employer's discretion; workers had no guarantee that they would actually receive the pension benefits after retirement; older workers could (and sometimes did) have their pensions rescinded before they reached retirement age; and firms sometimes used the pension plans simply as a device for "dumping" unwanted workers (Fischer 1977, p. 167; Graebner 1980, pp. 121-131; National Industrial Conference Board 1925, pp. 14,24-40). The trend toward adoption of private pension plans accelerated during the period immediately following World War I, when a number of developments profoundly affected the relationship between buyers and sellers of labor time. None was more important than the rapid membership growth of a trade-union movement that sought to replace the "voluntarist" or "business unionism" tradition of previous decades. Confronted with the prospect of independent unions, many employers responded during the 1920s as they had before the war—^with unveiled hostility, using tactics such as strike-breaking, injunctions, and "yellow dog" contracts. The aim was simply to destroy "the unionist scourge/' mostly by frontal assaults. But other firms, especially the larger firms, recognizing the manifold costs of labor strife, sought instead to coopt workers through a new strategy of welfare capitalism, designed to make more workers more contented with company policy and less interested in striking or in joining independent unions (see, e.g., Bendix 1956, Chapter 5; Fine 1969, pp. 23-27). Employerprovided pensions—"industrial pensions," in the nomenclature of the day—comprised a part of that strategy More than 700 plans were in operation by 1930, mostly in large firms in the manufacturing industries. These pensions were still granted at the discretion of the firm and usually contained disclaimers of inviolable pension rights. Nevertheless, a growing number of employers realized that the fact of a pension-eligibility age, if observed with some regularity, lent valuable legitimacy to a firm's interest in taking advantage of the then existing labor surplus by replacing workers who were deemed "past their prime" with less expensive and (it was typically assumed) more trainable, younger workers. In effect, the presence

Some Matters of Context


of an "eligibility age" legitimated the use of a mandatory retirement age as a normal feature of employment relations in the bureaucratic organizational culture of the new world of work (see Levine 1988; Quadagno 1988). Most of the private plans in operation during this period were financially unstable. Already by 1925 warnings from the National Industrial Conference Board (1925, pp. 12, 101-122) and elsewhere emphasized that many of the existing pension funds had not been based on sound actuarial principles. As the industrial labor force grew and average wages increased, firms often found that their pension funds were seriously depleted. The situation worsened, of course, with the stock market crash of 1929 and the ensuing financial crisis. Some firms simply abandoned their pension plans altogether. The long years of economic depression further drained the trust funds of remaining plans, and the rise in bankruptcies eliminated pensions for thousands of other workers (Fischer 1977, p. 167; Latimer 1932). Partly because of the private-sector initiatives, the idea of systematic provisions for old-age income security had gained salience in the public consciousness. It was not a new idea. Thomas Paine had proposed a national pension system in 1796; the federal government had established a pension system for military veterans following the Civil War; and a handful of states had been considering (and in three instances—Massachusetts, Wisconsin, and Arizona—implementing) public pension systems. Efforts to legislate a comprehensive national system had begun soon after 1900, but the proposals always floundered against strong resistance from keepers of the public morality, who saw in all such proposals a corrosion of the moral fiber of "self-reliance" and "rugged individualism" (see, e.g., Epstein 1922, pp. 228-243). However, as doubts about the ability of the private sector to provide that old-age income security became more widespread in the midst of massive unemployment and poverty, support for a nationalized pension system gradually increased, even among some employers (e.g., Alfred du Pont). The result was the Social Security Act of 1935, which provided old-age insurance (OAI) for workers in certain industries at age 65. Not surprisingly, given the history of debates about public morality, negotiations of interest across competing (and changing) practical understandings of the difference between the private law of the firm and institutions of public law were present at the birth. To obtain sufficient support for the legislation from business leadership, for example. Congress allowed employers to deduct their^ OAI contributions as nontaxable business expenses. The Act also allowed employers to reduce their total pension costs by integrating OAI benefits with their existing pension plans. Whereas workers might have hoped that integration would increase the total package of retirement benefits, companies generally opted for a substitution instead (Dyer 1977a,b; Fischer 1977, pp. 168-184; Jacoby 1993).


Chapter 2

Table 2.1. Marginal Corporate Tax Rates and Pension Assets, 1920-80


Marginal corporate tax rate (%) paid by large firms

Noninsured pension assets: Book value (1980 dollars) per capita

1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980

10.0 13.0 12.0 13.7 24.0 40.0 42.0 52.0 52.0 48.0 49.2 48.0 46.0

2 8 35 60 82 124 209 444 773 1,203 1,478 1,435 1,517

Source: Ippolito (1986a, p. 25).

From the beginning, then. Social Security was linked to employer-provided plans in ways that often served the interests of the firm. However, that linkage allowed influence to flow in both directions. Employer-provided plans received favored tax treatment, as amendments to the Internal Revenue Code created tax advantages both for money paid into pension funds and for interest generated by those funds. But while those tax advantages encouraged eniployers to institute pension plans, they also gave the federal government a mechanism for regulating employers' management of the pension funds. Under two Revenue Acts, those of 1938 and 1942, the federal government began specifying the conditions under which contributions to pension trust funds would be tax deductible. The Revenue Act of 1938 forced companies to place pension funds in an irrevocable trust, thereby closing tax loopholes that had allowed employers to contribute to the funds during years when company earnings were high and then recapture earnings during lean years by revoking the trust. The 1942 Act allowed employers to deduct their contributions to pension plans only if the plan had been incorporated into a written contract guaranteeing that the plan would be permanent and for the exclusive benefit of the firm's employees and their beneficiaries. But as the marginal rate of corporate income taxation continued to increase (from 13.7 percent in 1935 to 24.0 percent in 1940, 40.0 percent in 1945, etc.; see Table 2.1), the preferential tax treatment associated with pension plans grew more and more attractive to employers despite the added regulations, especially since the ability to integrate the em-

Some Matters of Context


ployer-provided entitlements with Social Security benefits was retained (Quadagno 1988; Quadagno and Hardy 1996). COLLECTIVE BARGAINING AND PENSIONS The National Labor Relations Act of 1935 established a legal framework for the right of workers to organize on the basis of collective interests and to settle contract disputes with employers through regulated collective bargaining. Even though the legislation was not particularly innovative and had few, mostly dull teeth, it was stoutly resisted by many employers. General Motors executives, for instance, characterized it as ''thinly disguised class legislation" which would "promote the exploitation of the American worker for the benefit of a comparatively few professional labor leaders responsible only to themselves" (quoted in Fine 1969, p. 51). A major turning point in relations between corporate management and the labor movement occurred in 1937, when, following a sit-down strike at the Fisher Body Plants in Flint, Michigan, General Motors recognized the United Auto Workers as labor's official bargaining agent (see Fine 1969). Bear in mind that GM was a huge corporation, by the standards of the day, with assets of $1.5 billion and 69 plants located in 35 cities. The UAW victory was a monumental step in the legitimation of the union movement. Thereafter the proportion of unionized labor in the manufacturing industries grew rapidly, and by 1946 nearly 70 percent of all production workers in the major manufacturing industries were covered by a union contract (Piore and Sabell984,p.81). The story of unionism and its relations with corporate management in the United States is a complex one, well and variously told by many others, and our aim here is not to repeat or to add to any of those tellings. Rather, for purposes of our chief interest in this section—the union role in the development of pension plans—a highly adumbrated account of certain aspects of labor-management relations during the years following World War II will suffice. Model Contracts and Pattern Bargaining Three contextual features of labor-management relations during the immediate postwar period were of paramount importance (see, e.g., Cohen and Zysman 1987, pp. 138-139; Edwards and Podgursky 1986, pp. 19-22). One concerns the legal framework of workers' rights of association: while the National Labor Relations Act of 1935 (NLRA) recognized those rights and established a labor court system (the National Labor Relations Board), its provisions for enforcement were anemic at best, subsequently enroded


Chapter 2

in the Taft-Hartley amendments of 1947, and then virtually abolished from within the NLRB itself during the 1980s (Clark 1989, pp. 4-6, 151-170, 180-184; Horwitz 1992, pp. 235-237; Klare 1978). For all practical purposes, employer compliance with the NLRA and with Board rulings has been voluntary. Nevertheless, a number of NLRB rulings during the postwar years were instrumental in regulating a new "labor accord." A second contextual feature is the fact that the core industries in which the union movement registered its greatest success had large fixed investments in physical plants, concentrated in the Northeast, North Central, and then Pacific Coast regions. Steel mills, automotive assembly plants, and the like, were still relatively new and prosperous; capital costs made them virtually as difficult to move as coal mines and dockyards. The regional concentration of physically immobile facilities significantly aided the organizational efforts and bargaining power of unionized labor. By the 1980s, however, technological changes in the core manufacturing industries had rendered much of that physical plant obsolete, and firms anticipating the need to build new facilities could now use the uneven geography of industrial location and correlated union strength to their advantage by engaging in concessionary bargaining with local communities and with union locals. The third feature is the oligopolistic character of domestic markets in the major manufacturing industries (e.g., automotive, steel). The market in each of the industries was increasingly dominated by a small (and ever smaller) number of very large firms which competed with one another primarily on the basis of product cosmetics and marketing techniques rather than on the basis of price and product innovation. Within this context, a so-called Fordist system of regulation came to characterize labor-management relations (Harrison and Bluestone 1988; Kochan, Katz, and McKersie 1986; Sayer and Walker 1992). In contrast to the more tumultuous, confrontational approach which had been the rule prior to the national emergency conditions of the war years, the postwar "labor accord" featured an increasingly bureaucratized, orderly division of negotiative jurisdictions in the contract bargaining process. To the extent that the NLRA's recognition of the basic structural conflict of interests between labor ^nd management was retained, it was subordinated and tamed under a mostly implicit agreement that centered on a routinized wage-determination process and that assigned to unions the twin roles of regulating mobility through the job-classification system and enforcing "labor peace" in exchange for employer discretion in organizing the production process, in hiring, in controlling investment decisions, and in most other matters not directly involving shop floor management practices. The automobile industry was at the forefront of these developments, with firms

Some Matters of Context


in other mass production industries often adopting procedures irutiated in agreements between the United Auto Workers and the major auto makers. The process of wage determination consisted of five main elements. First, the model contract formula: initially established in the 1948 settlement between GM and the UAW, the formula included an "annual improvement factor" (AIF) which tied wage increases to advances in productivity and rekindled Henry Ford's fabled notion that good workers should be paid enough to enable them to consume the products they made, thereby contributing to the stability of the larger economy (Harrison and Bluestone 1985, p. 85).8 During the late 1960s and until the 1980s, the AIF amounted to 3 percent a year. The 1948 UAW-GM contract was also the first major industrial agreement to contain a cost-of-living adjustment (COLA) as part of a multiyear wage-determination formula. The COLA provision often meant that the AIF amounted to a real, rather than nominal, increment in wage rate (Katz and MacDuffie 1994, p. 200). Second, "pattern bargaining," which had two dimensions: when an existing contract expired, negotiations did not revert to a "zero-based determination" but instead assumed that previous settlements had established a pattern of expectations for future settlements; and a settlement reached with one major firm in a given industry set a standard to be followed by other companies in that industry (and often by companies in other, related industries). Third, federal labor legislation forced nonunion employers either to increase their wages to the level of relevant collectively bargained rates or run a high risk of having their labor forces unionized. Fourth, minimum wage legislation restricted the spread between top and bottom wage-rates, thus lessening wage-rate inequality. Fifth, wage-setting mechanisms in the public sector linked the movement of salaries paid government workers to the trend in compensation among unionized workers in the private sector (Piore and Sabell984,p.80). In addition, wage schedules were tightly linked to a job classification system in which component tasks of a technology of mass production were grouped into hierarchical "lines of progression" or "seniority districts," graded by skill levels, with wage-rates attached to characteristics of the job and therefore standardized across individuals holding the same job grade. In the GM plant which Chinoy (1955, p. 37) studied during the late 1940s and early 1950s there were already more than 500 separate categories in the job classification system. Training to the job for new workers was job-specific (i.e., no "cross-training" or multiple-skill apprenticeships) within a relatively short probationary period (usually 1 to 3 months; Chinoy 1955, pp. 38,42). Mobility up the job ladder was regulated by highly specific seniority rules which assumed a rigid equivalence between clock-time on the job and level of ability. Priority in filling higher-paying vacancies was de-


Chapter 2

termined by length of service. During periods of industry contraction workers with more seniority "bumped" those with less into the surplus-labor category ("laid off); when demand increased sufficiently, recalls from layoff occurred in order of seniority (Piore and Sabel 1984, pp. 113-114). Private Pensions as a Union Benefit Once the wage-determination process has been established (thus, for all practical purposes, removed from the table of collective-bargaining issues), unions switched their negotiative interests to other benefits. Pension provisions were an early candidate. Even before the routinized wage-determination process had been fully tested in later rounds of contract negotiations, pensions had become an issue. In 1948 the National Labor Relations Board ruled that pensions were a negotiable item in collective bargaining agreements, and immediately thereafter the large industrial unions demanded that employer-provided pensions become part of a total compensation package. In the 1948 contract negotiations between the UAW and GM, company officials, contending that they were financially unable to increase wages and benefits, resisted the union's demand for negotiations over pensions. Charles Wilson, then head of GM, complained about "this tendency to stretch collective bargaining to comprehend any subject that a union leader may desire to bargain on." If not stopped, Wilson argued, it would result in "the union leaders really running the economy of the country" (quoted in Atleson 1983, p. 148). But in 1949 Ford Motor Company agreed to provide a defined-benefit pension to workers who retired at age 65 with 30 years of service. Following the already established principle of pension integration. Ford's plan was designed as a supplement to each qualified worker's Social Security benefit, with Ford contributing enough funding to raise the total monthly benefit to $100. This plan became a pattern setter in the industry. The general concept spread to other industries as well, and by the late 1950s over one-half of all unionized employees were covered by integrated defined-benefit plans (Quadagno 1988). In the early 1960s a separate tier of pension benefits was added to the existing retirement plan. "Early retirement" benefits allowed workers to retire before reaching Social Security eligibility. Early retirement was advocated by trade unions, which sought both to protect older workers and to maintain employment levels for younger workers. When plants closed or relocated, older workers were given the choice of either moving to the vicinity of another plant or trying to find new employment—^neither an easy option for workers in their 50s or 60s. Further, technological changes in the industry disproportionately threatened older workers (Gordus 1980). However, the seniority system meant older workers retained greater job security than younger workers.

Some Matters of Context


In 1961, following a period of high unemployment. Congress amended the Social Security Act in order to provide actuarially reduced benefits to men at age 62. Three years later the first early-retirement provisions were added to the contract settlements between the UAW and the Big Three auto companies. These new provisions allowed auto workers to retire with reduced benefits at age 60 if they had at least 10 years of service and at age 55 if they had at least 30 years of service.^ The provisions also included "supplemental" benefits, paid until the worker was eligible for Social Security at age 65. However, qualification for this supplemental benefit was contingent on the worker's agreement to limit earnings from any postretirement employment, an agreement designed to encourage retirement not only from the auto industry but also from the labor force in general (Barfield and Morgan 1970, p. 166). It was after this new early-retirement option had been in effect for 6 months that Barfield and Morgan conducted their survey of auto workers aged 58 through 61. They found that financial incentives of the early-retirement plan were very effective in pulling older workers out of the labor force. Two-thirds of the respondents either had retired before age 65 or were planning to do so. Fully a third of the original sample had retired before age 62, the age of eligibility for reduced Social Security benefits. Thus, the early-retirement provisions benefited workers, companies, and the union. Older workers could retire early and still receive benefits. Firms could use pensions to weed out less productive workers, including those who had trouble keeping pace with the assembly line. The union could look forward to growth of membership among young workers, knowing that it had protected the rule of seniority and maintained the loyalty of its older (including retired) members.

HARD TIMES IN THE AUTO INDUSTRY Thomas Fitzgerald, a highly placed official of GM's Chevrolet Motor Division, opened his article in the Harvard Business Review by announcing that "[rjising costs and recessionary pressure have prompted the business community, as well as administrators of public agencies, to seek economies. One potential source of savings is in labor costs, but these have resisted reduction because of the downward rigidity of wage rates and the difficulties of increasing aggregate labor productivity." These sentences were written not in 1991 nor in 1981 but in 1971—well before the first of the famed "oil shocks" of the 1970s. The standard management discourse of efficiency, competitiveness, and profitability has perennially featured "excessive labor costs" and "inflexible labor markets" as chief impediments to better corporate performance. What Fitzgerald knew to be the case in 1971


Chapter 2

was still part of management's tool kit of facts and figures in 1981, but now the standard matter-of-fact knowledge was accompanied by emotive sounds of the claxon. Entire manufacturing industries—including the auto industry, which more than any other had stamped the conspicuousness of "rising middle-class prosperity" in the postwar world dominance of the United States—seemed in peril of being driven through the vanishing point. In the biographical perspective of a lifetime, blue-collar as well as white, the health of an industry automatically equated with the health of particular firms: the auto industry, for example, was General Motors and Ford and Chrysler. These were, after all, the actual sites of the lifetimes spent at work, blue-collar or white. Employees on both sides of the bargaining table were now speaking of labor costs and rigid labor markets with a somewhat revised understanding of the situation—though only "somewhat revised," since both sides of the table were ostensibly united (more or less) in a market of employment relations still intensely national (management speaking of "national competitiveness"; union leaders and rank-and-file members speaking of "jobs at home for American workers"), even though the capital markets which had been driving the technological innovations, the rebuilding of war-devastated infrastructures in Japan and Germany, and the increasing mobility of skills, public policies, and consuming bodies had dissolved much of the underside of our everyday maps of political geography. The 1980s were an especially difficult time for U.S. manufacturing industries in general. A recent summary of "the main developments of the 1980s in private-sector American industrial relations" tells of some of the consequences: "declining unionization, employers setting the agenda in bargaining, the spread of employee involvement programs and work place innovations of various types, increased decentralization in bargaining structures, declining real wages, and heightened concern for job security" (Voos 1994, pp. 1-2). Not only was the U.S. auto industry not immune to those processes; in many respects it was in the vanguard, just as it had been in the vanguard of developments during the decades of prosperity. In 1950 the United States accounted for 80 percent of world production of automobiles (Feldman and Betzold 1988). By 1989 the U.S. share had dropped to 14.6 percent. During the latter part of that period the declining relative shares combined with a reduction in the absolute volume of units produced: in 1960 the U.S. share of the world's passenger car production was 52 percent of 14.6 million units, or 7.6 million units; in 1988 the U.S. share was only 12.8 percent of 48.628 million units, or 6.2 million units (Stark 1989; Wards 1989). The effects on industry employment were profound. Until the 1980s many observers had believed that industries characterized by a high concentration of very large firms would experience relatively low rates of labor outflow during periods of contraction, because

Some Matters of Context


large firms tend to have stronger internal labor markets and relatively high per capita investments in firm-specific skills (Brown, Hamilton, and Medoff 1990, Medoff 1979). The recent record suggests otherwise (DiPrete 1993). Between the end of World War II and the late 1970s the auto industry grew increasingly oligopolistic—fewer but larger firms, each with a robust internal labor market. The number of production workers grew by about one-fifth during that period, reaching a peak in 1978. Between that peak and 1982, however, total employment in the industry (SIC 371) declined by 30 percent (from just over one million workers to just under 700,000). The number of production workers declined by 34.5 percent, from 782,000 to 512,000, with many of these remaining workers on indefinite layoff (e.g., the number fluctuated between 150,000 and 269,000 during the first 3 years of the 1980s). In late 1981 one-third of the UAW members at Ford and GM were on layoff; the gates to 22 Big Three plants had been closed, with more closures on the horizon. Industry employment rebounded in 883,500 in 1985, only to recede again during the next 6 years.^^ Correlatively, total UAW membership, which had stood at about 1.5 million in 1979, was down to 950,000 in 1990, a decline of 37 percent. Hourly wages suffered. In 1979, under the threat of bankruptcy, total collapse, and/or acquisition by a Japanese or European auto maker, Chrysler's workers agreed to a concessionary package that reduced their hourly wage rate $2.50 below the rate for Ford and GM workers and included the loss of the AIF, the automatic COLA, and paid personal holidays. Ford and GM followed suit in 1982. The nomial wage rate for a GM assembler in 1980 was $9.77; it was $13.15 in 1985 and $15.69 in 1990. But adjusted for inflation (CPI adjusted, 1982-84 base), those nominal rates translated into real rates of $11.86, $12.22, and $12.00, respectively (Harrison and Bluestone 1988, p. 87; Katz 1985; Katz and MacDuffie 1994; Morales 1994, Ch. 1; U.S. Bureau of Labor Statistics 1985). The crisis in the industry undermined major features of the postwar labor accord both directly and indirectly. Direct effects included management demands for concessionary bargaining with respect to wages and work rules, extended periods of layoff during which workers were vulnerable to loss of seniority and recall rights, and the closure of plants, especially in states that had a relatively strong history of unionism, at the same time that new production facilities were being built in states less sympathetic to unions (i.e., the so-called southern strategy).^^ Indirectly, structural changes in the national economy held up a refractive lens through which to review unionism's recent history. The fact that blue-collar workers gained dramatically larger shares of national income during the quarter-century following World War II had often been attributed primarily to ''union power." But in retrospect it was becoming evident that much of the gain had been due to demographic and structural changes in the economy—for


Chapter 2

instance, the growth of employees (i.e., a category of the bureaucratic corporate form) at the expense of shopkeepers, artisans, farmers, and the like. As large industrial firms grew increasingly attentive to issues of competitive advantage in globalizing markets, and increasingly attributed their failures of competition to inflexible labor markets and excessive labor costs, it became ever more apparent that "union power," even at the peak of union membership during the 1950s, had been dependent on at least the indifference, and often the cooperation, of the major industrial firms (see Clark 1989; Rothstein 1986; Sayer and Walker 1992). During the good economic times of earlier decades the primary issues in contract negotiations, in union representation elections, in the internal politics of union locals, and in the relationships of union locals to their national leadership had typically been formulated on a systematic national basis, with locally particular issues accorded secondary status. The systematic issues centered on improvements of wage-rate and benefit packages and the regulation of conditions in the internal labor market. Those issues did not disappear duiring the 1980s, but they were often supplanted by nonsystematic, locally specific issues of leadership, loyalty, and job protection, as the number of plant closures increased, unemployment levels rose, and local communities had to cope with the resulting tensions of job rationing (Clark 1989, pp. 102-104; Dudley 1994; Zetka 1995). The internal cohesion of unions in the manufacturing industries generally suffered from the presures of rationing, which often put union locals in a relationship of rivalry at the same time that union leadership at the national level was advocating "cooperative partnerships" between labor and management. As Clark (1989, pp. 70,185-187) noted, "many corporations depend upon their unions to help design and implement restructuring programs" for the rationing of jobs, even when it is apparent that rank-and-file workers are divided on the prudence of those ventures. So it was in the auto industry, where union influence over the work place diminished during the 1980s partly because of the expanded use of Quality of Work Life (QWL) programs and work teams, even though these cooperative ventures were designed and regulated by a committee of top UAW and management officials. While integrative schemes of labor-management relations such as QWL programs (also known as "worker participation" or "employee involvement" programs) have recently gained notoriety because of identification with an alleged "Japanese style" of industrial relations, they have a long heritage in the United States.i^ The NLRA (section 8a.2) set out a model for integrative schemes, although actual practices have often diverged from that model. Most common in the unionized sector, earlier versions of QWL programs were often implemented unilaterally by management as an alternative to collective bargaining—arguably in violation of

Some Matters of Context


the NLRA section (a possibility which the NLRB and courts have avoided addressing; see Kohler 1986, p. 500)—^but in recent decades union leaders (at the national level) have joined management officials and others in endorsing QWL programs as instruments that can achieve a ''democratization of the workplace" (see, e.g., Bluestone and Bluestone 1992; Ephlin 1988).i3 However, their most important achievement to date has been additional shrouding of the basic structural conflict between buyers and sellers of labor time. As one analyst put it, "employee involvement" programs in general are much like "a public defender's office that was established, funded, and staffed by the District Attorney" (Weiler 1990, p. 207). For rank-and-file workers in the auto industry, the immediately apparent effects of recent expansions of the QWL technology have included a massive revision of the job classification system. In some QWL programs, as many as 60 semiskilled production jobs have been collapsed into a single category, with wage schedules altered in the process. The revision was connected to the auto makers' relocation of production facilities to southern states which, partly because of "right to work" legislation, had lower wage rates. In one sense, this "southern strategy" failed; all plants owned by the U.S. auto makers were eventually unionized (GM's last holdout was unionized in 1984).i4 But the strategy succeeded in more subtle ways, as the southern plants frequently experimented with flexible forms of work organization, eroding principles of task specialization and the solidarity of pattern bargaining (Katz 1985; Katz and MacDuffie 1994; Piore and Sabel 1984, pp. 243-244). Another sort of southern strategy resulted in even stronger conflictive pressures on the union's role in job rationing. U.S. companies found it profitable to move some of their production components to countries such as Mexico and Brazil in order both to exploit the lower labor costs and to comply with demands by the governments of those countries that foreign companies expecting to sell products in the local markets had to have production facilities in those same markets (Piore and Sabel 1984, p. 198). The maquiladoras plants of Mexico, for instance, took in raw materials from the United States, processed them, and then exported the finished product back to the United States, with duty paid only on the value of the labor. This arrangement was originally made possible by a 1930 law which had been designed to let Detroit auto plants ship work to Canadian parts plants. Many years later, when the Mexican government sought to build new manufacturing plants for displaced farm workers, it used the law to encourage U.S. firms to set up border factories, throwing in duty-free imports of raw materials, low rent, and cheap labor. The strategy depended on keeping Mexican labor unorganized, and the Mexican government cooperated in antiunion campaigns. By 1989 GM operated 23 plants south of the U.S. border—the majority built just south of the Rio Grande—having


Chapter 2

closed 20 plants north of the border between 1982 and 1989 alone (Harrison and Bluestone 1988, p. 30). While this "outsourcing" to low-wage countries typically involved the production of low value-added goods, with most of the high value-added production retained in U.S. plants, the resulting externalization of labor demand contributed to a weakening of the internal labor market (Morales 1994, pp. 79-86). In response to the conditions of its weakening position, the UAW attempted the difficult feat of maintaining a cooperative orientation to the task of rationing jobs in a shrinking U.S. industry, while at the same time protecting wages and job security among remaining workers. The union's general strategy was to win agreements that would make plant closures and layoffs more costly to the companies. The strategy was only partially successful. During the early 1980s the union won agreement to a new Guaranteed Income Stream (GIS) program, according to which a worker who was placed on permanent layoff after at least 15 years of service would receive 50 percent of the previous year's earnings, plus an additional 1 percent for each year of seniority beyond 15 (up to a maximum of 75 percent of the previous year's earnings), until recall or retirement. Recall rights, eroded by long-term layoffs, were extended for workers who had not broken their seniority queues. The union also won agreement to an experiment in guaranteed employment at selected plants (four at GM, three at Ford): 80 percent of the existing work force would be guaranteed employment during the term of the agreement.^^ As well, management promised not to close any plants for 24 months, and workers won a profit-sharing plan and the establishment of retraining programs. The 1984 contracts with the Big Three included additional provisions designed to protect workers who "might be displaced from their jobs because of technological advances, corporate reorganizations, outsourcing, or other negotiated productivity improvements" (Katz 1985, p. 176). Displaced workers would be assigned to a Job Opportunity Bank Security (JOBS) program and receive full pay while awaiting transfer to another plant or retraining for other jobs in the company. This program was to run for 6 years. During its first 3 years of operation approximately 12,500 GM workers were enrolled. The year 1984 also saw significant recovery in the auto industry, with the Big Three reporting collectively an all-time record profit of nearly $10 billion. Union and trade publications would later refer to this turnaround as a "workerless recovery," however, since industry employment continued to lag well behind the levels of the late 1970s. Rather than recall laid-off workers, management increased overtime among presently employed fulltime workers. In 1982, for instance, production workers in the industry averaged 2.5 hours of overtime per week; by 1985 the average had reached 5.4 hours, and it remained in the range of 4 to 5 hours throughout the late

Some Matters of Context


1980s. The advantage to management was both greater flexibility in regulating labor supply during future downturns (i.e., by reducing overtime hours instead of laying off workers) and lower benefit costs. In 1988 the union successfully negotiated provisions regulating the use of overtime. Workers laid off because of volume declines had to be recalled in proportional numbers when production picked up after a slump. When overtime was used for any reason, the company had to pay into a training fund $1.25 for each hour of overtime exceeding 5 percent of "straight time." It was soon evident, however, that the intended penalty was an insufficient brake on ''overtime abuse." On the issue of "outsourcing," the union won contractual guarantees designed to protect the jobs of current UAW members with at least 1 year of seniority: if companies brought competing vehicles or parts into the United States, no auto worker making a similar product could be laid off. The guarantee also stipulated that workers could not be laid off except as a result of "carefully defined" reductions in sales. The language of the agreements was crucially vague, however, making the stipulations more flexible in enforcement than suited many adversely affected workers. One GM employee, a member of Local 22 (southwest Detroit), spoke for countless others when he protested in the pages of Solidarity, UAW's official newspaper (August 1989, p. 20) that "GM is announcing hundreds of layoffs each week because of a contract loophole permitting layoffs because of 'slow sales.' When we voted to ratify this contract, we were told our jobs would be counted in the secure employment level (SEL) program. Now we find out differently." The union's response to the complaint did not offer much solace: "The contract protects eligible UAW members against layoffs due to outsourcing and virtually any other reason except for carefully defined volume reductions linked to market conditions—in other words, poor sales. However, the contract also stipulated that as sales rise again, management must call back workers in the SEL instead of just assigning overtime—a major gain for us" (Solidarity, August 1989, p. 20). Downward pressure on labor costs in U.S. plants continued throughout the 1980s, despite improvements in production figures and profits. Large numbers of auto workers were shed from the corporate payrolls, and wage rates changed hardly at all in real (constant dollar) terms. In spite of the difficulties, however, the UAW did enjoy some success in protecting the comparatively favorable wages of remaining workers. In 1989 the average hourly wage for all production workers in the motor vehicle industry (SIC 371) was $14.25—considerably better than the $10.49 for production workers across all manufacturing industries and better still than the $9.66 for production and related workers across all industries. Compared to the corresponding 1980 averages, those 1989 wages were uniformly 44 to 45 percent higher in nominal (current dollar) terms and about 2 percent lower in


Chapter 2

real (constant dollar) terms. So the relative advantage in hourly wages enjoyed by auto workers at the beginning of the decade was maintained. Comparisons of average weekly earnings tell a rather different story, of course, because of the accelerated rate of overtime hours in the auto industry. Whereas average weekly earnings for production workers across all manufacturing increased a little more than 3 percent in real terms, for auto workers the increase was more than 10 percent (U.S. Bureau of the Census 1991, Tables 669,674). A Closer Look at GM, Circa 1987 As Morales (1994, pp. 85-86) has pointed out, the U.S. auto makers were initially uncertain how best to respond to the changing circumstances of the domestic industry. Would size reductions be sufficient? Or would basic revisions of production technology be required; and, if so, exactly what revisions? Initially, the general tendency was to try some of each approach. General Motors, the largest of the U.S. auto makers, was slowest to respond effectively with either approach, and by 1987 (the begiiming of the contract period that forms the context of the retirement decisions analyzed later in this book) GM's production efficiency was distinctly poorer than that of its main domestic competitor. Ford. Following the end of the severe recession of the early 1980s, GM expanded its production capacity and, unlike Ford and Chrysler, recalled a large proportion of laid-off workers. A reorganization in 1984 led to some consolidation of production facilities, experiments in new technological applications, and a joint venture with Toyota (NUMMI, or New United Motor Manufacturing, Inc.), but GM's share of domestic sales continued to erode nonetheless. From 1986 to 1989 GM's market share slid from 41 percent to 35 percent, four of the lost percentage points occurring in 1987 alone. Industry analysts estimated that each percentage-point decline reduced GM's optimal work force by about 10,000 hourly workers (Woodruff, 1990, p. 40). During the 3-year period from 1984, GM spent 12 percent of the $1 billion it had allocated for its participation in the JOBS program (the "job bank"), in order to retrain just over 12,000 workers. Ford had 2,000 workers in the program. In 1987 GM was paying 50,000 workers on indefinite layoff supplemental unemployment benefits (SUB) of 95 percent of the regular wages. These and related provisions served from the union's point of view as significant disincentives for GM to close plants, "outsource" to foreign or nonunion suppliers, or otherwise cut employment levels.^^ Yet GM's plants were operating at an average of 75 percent of capacity; Ford's, at nearly full capacity. GM was still in the midst of decisions to close plants— two in August 1987, another 11 scheduled for shutdown by 1990, and still others on the horizon—^whereas Ford had most of its paring and reconsoli-

Some Matters of Context


dation behind it, having closed only one plant (a forge plant, with 909 employees) since 1983. GM had continued to keep a higher proportion of its components production ''in house": whereas an average of 50 percent of the components of Ford's vehicles were supplied by other companies, the corresponding figure for GM was only 30 percent. In the 1987 round of contract negotiations the UAW selected Ford as pattern setter, with the primary goal of enhancing job security for remaining and future workers. The union recognized GM's parlous state. As the director of the union's GM Division put it, "GM faced declining market share, eroding sales, the announcement of nearly a dozen plant closings, a loss of public and investor confidence, and a crisis in its in-house components operations. Moreover, we in the union made clear our position that ... a large measure of GM's problems resulted directly from the company's lack of uniform labor policy and an absence of clear direction and followthrough from its top management" (Ephlin 1988, p. 65). Judging from the observations of industry analysts (e.g., Zellner 1987; Zellner with Bernstein 1987), the UAW decided to allow GM substantial maneuvering room by demanding less from Ford than probably could have been won, in order to avoid a confrontation with GM. In any case, GM agreed to essentially the same contract provisions as Ford. It was nonetheless an expensive settlement for GM. The UAW had wanted GM to continue to employ the same number of workers at the end of the contract period as it employed in 1987 (375,000, both active and inactive). GM, on the other hand, already had 50,000 workers on layoff, with an additional 37,000 due for layoff by the end of the contract period because of announced plant closings; and its market share had dropped by 4 percentage points in 1987 alone, with more slippage likely to come. In the end, GM agreed to a secured employment level (SEL) of about 315,000 with layoffs allowed only because of sales slumps (i.e., the "carefully defined" layoffs referred to earlier). GM also agreed to a ban on plant closings, excepting those already announced (Ford had agreed to a ban in 1984). Union officials reportedly expected that additional closures, though not yet announced, were a likely part of GM's reconsolidation plans (Zellner with Bernstein 1987). But they would be costly to the company, since affected workers would remain on the payroll regardless of whether jobs elsewhere in the company were found for them or not. In exchange for improvements in job security, the union promised to cooperate in efforts to increase productivity. Joint committees of union and management officials were directed to examine "Japanese style" work teams as a strategy for increasing productivity per unit of labor. Job classifications and work rules would need to be reduced or modified in order to create a more flexible regime. A new "Quality Network," developed and staffed jointly by management and union representatives, would work to


Chapter 2

enhance quality at all levels, from design to final assembly. One union official promoted these aspects of the agreement by avowing that "for the first time in the history of General Motors, union and management are working together to address the issues of quality and customer satisfaction" (Ephlin 1988, p. 65). The 1987 contract also included improved health care, disability, pension, and profit-sharing benefits. Pension benefits were raised by $4.20 per month per year of service, an increase of 19 percent over the previous contract. In addition, the agreement included improvements in some early-retirement options for qualified workers, details of which will be presented in the next chapter. Here, we consider briefly the relationship of those plans to GM's difficulties in reducing the size and cost of its internal labor market. Early Retirement as a Siphon of Excess Labor The practice of using early-retirement pension plans as an instrument for regulating labor supply began during the 1973-74 recession, when many companies added special incentives to their established early-retirement programs. Developed as a variant of "severance pay" allowances or bonuses paid to workers who met specified criteria, these incentive programs were typically made available during limited periods of time—or special retirement "windows"—to defined groups of workers who otherwise either had not attained early-retirement eligibility or, if technically eligible, had not qualified for a benefit level that would have made early retirement feasible. A parallel program of special incentives, known as "mutuals," operated in essentially the same manner, except that it could be offered at any time by mutual agreement between management and specified individual workers. During the 1980s the UAW negotiated expanded opportunities for senior workers to take early retirement, especially when they were adversely affected by plant closures or job transfers. CM, moving belatedly in middecade to achieve permanent reductions, used both indefinite layoffs and special incentives for early retirement. The latter mechanism was preferred for all sorts of reasons: not only did it better serve the corporate image in local-community relations and maintain a cooperative stance by the union; it was also less expensive. The contract required payments to GM's workers whether they were working or idle. Retired workers were paid at a rate that was roughly one-half the rate of the wage bill. Moreover, whereas wages had to be fully funded from current revenues, pension benefits were paid from a trust fund, which could be (and was) substantially underfunded relative to liabilities. Although it was impossible to know how many workers would have retired when they did, had the special incentive

Some Matters of Context


programs not been on offer, experience suggested that the programs were having some effect. In 1984, for example, about 7,000 GM workers elected early retirement; in 1986, 13,000. Further growth in those numbers was clearly in the interest of both the union and GM management, and so the 1987 agreement expanded the incentives of the Special Early Retirement Plan. Obviously no one knew exactly how many workers would be enticed by the incentives, but according to public reports at the time (e.g., Zellner with Bernstein 1987) approximately 35,000 of GM's workers were expected to qualify under terms of the program. The hope was that all or most of those 35,000 would elect to leave, since each who did leave could be subtracted from GM's secured employment level (SEL) of 315,000 workers without replacement.

NOTES 1. Part of the variation has been a function of different eligibility ages for state pension schemes. For instance, in 1983 only a third of the men aged 60 through 64 were labor force active in France, where the eligibility age was 60. On the other hand, the eligibility age in the Netherlands was 65, yet participation rates among men aged 60 through 64 fell from 74 percent in 1971 to only 25 percent in 1989; and among men aged 55 to 60, from 87 to 67 percent (Fiji 1991). 2. Vesting periods were limited to a choice between 100 percent vesting after 5 years (not 10, as prior to 1989) or 20 percent vesting after 3 years followed by 100 percent vesting after 7 years (Clark and McDermed 1990). 3. A more detailed discussion of actuarial reduction and related features of a DB plan is presented in Chapter 3. 4. Real contributions to pension plans by private-sector employees fell 36 percent during the 1980s. This appears to have been due not only to the shift from DB to DC plans but also to the growing use of the 401 (k) version of the DC plan—defined by Section 401 (k) of the 1978 Revenue Act—which allows participation on a more voluntary basis. Young workers are less likely to participate (Even and Macpherson 1994). 5. By 1920 the ITU, the United Mine Workers, and a few other unions were passing resolutions calling for the establishment of state pension systems or even a comprehensive national system. 6. Graebner (1980, pp. 127-128) quotes from an advisor's letter to Coleman du Pont, 1914: "I never heard of any plan except one that would assist in regulating salaries and that is a pension plan. If you people have not any such plan in operation, I strongly urge you to get busy and adopt one. The right sort of pension plan comes pretty near being a panacea for most of the ills that exist between employer and employee." 7. According to the results of the survey by the National Industrial Conference Board (1925, p. 72), about two-thirds of the plans which connected pension eligibility to a mandatory retirement age stipulated age 70 for men; nearly all of


Chapter 2

the remainder stipulated age 65. Mandated ages for female employees were sometimes younger. 8. In fact. Ford's fabled notion was itself mostly fable; few of Ford's cars were sold to his own blue-collar workers (see Foster 1988). 9. Senior workers with fewer than 30 years could retire as early as age 55 if the combination of years of age and years of service totaled at least 85. We explain these provisions in more detail in Chapter 3. 10. By 1994 the average employment figure had climbed to just over 850,000 (or about 15 percent less than the 1978 total), but much of this recovery occurred in "the Japanese transplants/' i,e., in U.S. plants owned by Japanese auto makers. Between 1978 and 1993 hourly wage jobs in GM, Ford, and Chrysler assembly plants dropped from 740,000 to about 400,000 (Katz and MacDuffie, 1994, p. 191). 11. As Wendling (1984, p. 11) noted generally, since "the reasons cited for closure in surveys and in court cases" often involved claims about excessive labor costs and work rules, the issues tended to be "amenable to resolution through collective bargaining." But often they were not. Bear in mind that under "job control" unionism only issues mutually regarded as negotiable were typically put on the table during bargaining sessions. The impact of plant closures on union strength extended beyond the auto industry, of course, and began before the 1980s. Schmermer's (1983) survey of Fortune 500 companies during the 1970s found that, whereas 52 percent of all of the facilities were unionized, 66 percent of the facilities that closed were unionized. The discrepancy was partly due to the fact that most of the closures occurred in mass-production industries, where historically the union movement had been most successful. 12. For example, the "open shop" or "American plan" movement of the 1920s involved initiatives designed to demonstrate that workers' rights of independent free association with respect to work place conditions were unnecessary (Bendix 1956). During the 1930s GM experimented with a number of "employee involvement" programs and utilized a "familial" rhetoric in efforts to socialize workers to company interests (see Fine 1969). During the 1950s and 1960s Douglas MacGregor championed a management theory ("Theory Y") that emphasized "employee involvement," and W. E. Deming promoted integrative schemes with missionary zeal over a long period of time, though to much better reception in Japan than in the United States. Resistance to the use of such schemes has often come from within management itself. For example, one GM official avowed—in a "give them an inch and they'll take a mile" argument— that QWL programs threatened management prerogatives (Fitzgerald 1971). 13. During the 1980s the heads of GM (Roger Smith) and the AFL-CIO (Lane Kirkland) were in agreement that the NLRA and its Labor Relations Board had become anachronisms and impediments to the development of new forms of labor-management cooperation (Clark 1989, pp, 250-251). 14. According to Clark's (1989, pp. 11-20) study, union successes in NLRB representation elections actually increased in the southern states during the 1980s, although the total number of contested elections continued to decline there as in other regions. Clark's data demonstrate that the historic "southern pattern" in employment relations was becoming the national pattern.

Some Matters of Context


15. The 1985 agreement for GM's Saturn Division plant in Tennessee (a right-towork state) provided 80 percent of the workers with job guarantees. The remaining 20 percent were defined as "associate members" of the union and lacked job guarantees (Jacobs 1994). Thus began the two-tier system of employment. 16. Whereas employment by nonunionized independent auto parts suppliers in the United States grew by one-third between 1978 and 1989, it declined by one-half among independent suppliers who were unionized and by one-third in the auto makers' own parts facilities (Morales 1994, p. 84).

3 The Financial Structure of Early-Retirement Pensions

BENEFITS AND INCENTIVES In popular imagination the idea of a "pension benefit"—whether the benefit payments of a private-pension scheme or those of a public scheme such as Social Security—most often elicits images of "reward" and "old-age security." The benefit of a pension is thought to be, at the individual level, a more or less proportionate reward for having completed a "productive work life" and, at the population level, a means of alleviating the risk of old-age poverty. That thought is correct so far as it goes—though the tendency sometimes is to confuse the "reward" aspect with a free good and to assume that "old-age security" is only a matter of money income. But the benefit of a pension—and here we must restrict the point to employer-provided defined-benefit pension schemes—redounds not only to the individual recipient and the population of recipients but also to the employing firm. In the case of defined-benefit pension plans, employers typically design the benefit schedule as an instrument of relative incentive, used to achieve certain policy goals in the regulation of the firm's internal labor market. When designed and implemented accordingly, the instrument of incentives can achieve an increase in employee attachment ("loyalty") during certain spans of "career age" and then an increase in employee exits during later ages. After joining the labor force of a firm, a new employee typically enters what amounts to a kind of "probationary period"—a period of trial and error, "training to the job," and so forth. From the firm's point of view, employee loyalty is being built and tested: some employees quit or are fired; those who survive the selection process embody investments by the firm (training costs, loyalty costs, etc.), which are roughly proportionate to the skill level of the employee. The contractual relation (whether based on 49



mutually genuine consent or simply on compliance or acquiescence), though formally symmetrical, assumes subordination of each employee's interests to the interests of the firm (i.e., the institution of private or intrafirm law, limited by institutions of public law such as the Age Discrimination in Employment Act). At some date enough seniority (i.e., loyalty) has been achieved to justify (from the firm's point of view) a longer-term investment in the employee. In terms of a firm's defined-benefit pension plan, that date is signaled by vestment rights: the employee has satisfied minimum requirements of the pension plan and now begins to accrue assets in the plan, a form of deferred compensation. From this date forward, it is in the firm's interest (ceteris paribus) to reinforce employee loyalty. ^ Various means are available. One is to pay "efficiency wages" (i.e., a wage rate higher than the prevailing market rate). Another is to distribute wage-rate compensation unevenly across the full tenure of a career. Third, loyalty can be reinforced by slowly adding value to the employee's pension-benefit account (i.e., graduating the deferred compensation). Theoretically each of these approaches should enhance loyalty. To some extent they are compatible with, even complementary of, each other. For instance, the second approach is less capable than the third, in the absence of mandated retirement ages, of inducing an abrupt shift from loyalty to exit. In effect, by adopting the second approach the employer has implicitly struck a long-term "tenure contract" with the postprobationary employee, centered on a wage schedule that disproportionately compensates high seniorities at the expense of low seniorities. Consider the theoretical model depicted as Figure 3.1, where V(t) is the value of a worker's marginal product at various tenure lengths, Wj^(t) is the minimum wage schedule at which the employee will sell labor-time to the employer (i.e., the employee's "reservation wage"), and W(t) is a wage schedule according to which the present value of wages paid during an entire tenure equals the present value of the employee's marginal product during that tenure.^ During early years of tenure the value of W is less than the value of V, though by declining differences; during later years of tenure the value of W is progressively greater than the value of V. The assumption is that postprobationary employees will continue employment with the given firm until date T, the tenure date at which the undercompensation during early years has been equalized by overcompensation during later years, which is also the tenure date after which the employee's reservation wage (Wj^) exceeds the value of his or her marginal product. Thus, T defines the efficient date of retirement from the employer's point of view. But from the employee's point of view, continuation of employment after T would be preferable (ceteris paribus); consequently, in absence of a mandatory retirement age or similar constraint, the employee is inclined to postpone retirement to a date later than T.

The Financial Structiire of Early-Retirement Pensions



W„(t) V(t)

Figure 3.1. A theoretical model of a wage schedule.

The third approach provides that constraint even in the absence of a mandatory age. The implicit tenure contract includes provision for deferred compensation in the form of a pension benefit, with a graduated schedule of asset accrual in the employee's benefit account. The rate of accrual is very slow during early and middle tenure years. But during later tenure years, as the firm's interest shifts from inducing loyalty to inducing exit, the rate at which asset is added to the now senior employee's pension account rapidly accelerates. Then, at date T (or at some related date at which the firm prefers exit to continued loyalty), the rate of accrual collapses to zero or even turns negative. Continued employment much beyond that date will mean that even though the worker would receive wage earnings exceeding his or her current (or even tenure-averaged) marginal product, the worker's total compensation will decline because of the zero or negative rate of accrual of deferred compensation (i.e., pension assets). This general framework can be applied to any number of employerpreferred retirement ages. Figure 3.1 assumes that the implicit tenure contract is negotiated early in the employment relationship, and the balance between the undercompensation and the overcompensation portions of the wage schedule assumes an exit at date T. By setting date T as the date at which the accrual rate in a "normal" pension plan collapses to zero, the em-



ployer induces exits at a "normal" retirement age, as noted above. But an employer may also offer another pension plan, one in which the date at which the rate of accrual collapses to zero is earlier than T. Of course, no employee agrees at the beginning of the implicit tenure contract to retire early and thus exit before T. An employee who, later in tenure (and eligible for early retirement), does exit before T does so presumably because he or she has decided that (ceteris paribus) the deferred-compensation incentives of the early-retirement plan sufficiently offset that part of the overcompensation portion of the wage schedule that will be forgone by retirement. In effect, employer and employee have renegotiated terms of the implicit tenure contract. But features of that renegotiation can differ, depending on characteristics of the early-retirement plan. When the early-retirement plan has been part of the employment relationship since the beginning of an implicit tenure contract (i.e., a "regular" plan), the renegotiation option is routine and relatively free of coercive features, and the employee has ample opportunity to engage in anticipatory behavior (''getting ready to retire"). However, under certain circumstances (e.g., plant closings) an employer may conclude that the actual schedule of exit behaviors, even with the incentives of an early-retirement plan in operation, is less efficient or less rapid than desired, in which case the employer may prefer to offer a new option "on the spot," with the intent of accelerating the shift from loyalty to exit among current senior employees to an even earlier date prior to date T. Here the option of renegotiation—specific to a more or less narrowly defined set of "unusual" circumstances and offered for a limited time—approximates what is sometimes called a "spot market" form of exchange relations (as distinguished from a long-term or "tenure contract" form of exchange relations). By offering "special" added incentives (relative to the incentives of the "regular" early-retirement plan), the employer abruptly and greatly increases the value of assets in the employee's pension account. In effect, the rate at which recent years of service accumulated credits to the account is recalibrated, so that a given year of service late in the implicit contract is worth more in the spot market situation than would otherwise obtain. What might appear to be a "bonus" paid to the especially early retiree in return for the exit is compensation for the wage earnings relinquished because of the especially early exit. Just as with the other retirement plans, the asset value of the account sharply declines if the employee does not exit at or soon after the desired date. But in addition, given the unusual circumstances under which the special incentives are offered, this "spot market" option to renegotiate the implicit tenure contract will typically include some coercive elements (e.g., the threat of permanent job loss) and little opportunity for anticipatory behavior by the employee. The foregoing is a thumbnail sketch of the operation of pension plans such as those which have figured in the GM-UAW negotiated contracts.

The Financial Structure of Early-Retirement Pensions


The sketch is oversimplified in some respects (corrected later in this chapter), but it serves well enough as an initial depiction of the financial-incentive structures that were built into the retirement plans at GM many years ago. The principal task undertaken in this chapter is an analysis of those financial-incentive structures, with particular reference to the early-retirement pension plans that were negotiated in the 1987-89 contract settlement. Because the retirement plans are characterized by a number of small but important distinctions in their financial-benefit structures, and because aspects of the financial analysis are somewhat complex, we have tried to simplify the presentation by imposing and following repetitiously throughout this chapter a sometimes arbitrary terminological regimen, albeit at some cost of monotony. For readers who are not acquainted with the work of actuaries or financial analysts and who do not have some familiarity with the internal structure of typical employer-provided defined-benefit retirement plans, there will be distinctions enough without having to wonder whether a substitution in wording signals intended difference or assumed synonymy. Thus, for example, we will refer to the set of plans as the "GM retirement plans," even though essentially identical plans were part of the negotiated contracts with Ford and Chrysler as well. Likewise, we will use the phrase, "employer-provided plans," even though the plans resulted from union-management negotiations and (since all plans herein considered are defined-benefit plans) the benefit payments amount to deferred compensation. The 1987-89 contract included two distinctive early-retirement plans: the Regular Early Retirement Plan (the older of the two) and the Special Early Retirement Plan; we will often refer to these by their acronyms, RERP and SERF. In addition, of course, the negotiated contracts have long included provisions of a Normal Retirement Plan (NRP) and a Disability Retirement Plan. Since the focus of attention in subsequent chapters is on the early-retirement decision, our analysis in this as well as in subsequent chapters largely ignores the Normal and the Disability Retirement Plans. However, because the calculus of decision making with respect to an opportunity of early retirement necessarily unfolded for all workers against the background of "normal retirement" provisions, our initial descriptions of financial characteristics of the plans will include some attention to the Normal Retirement Plan. As well, those descriptions will attend briefly to the financial characteristics of the plans in historical perspective—that is, from 1950 to the mid-1960s, the time of Barfield and Morgan's study, and thence to the 1987-89 contract, the immediate context of our subsequent analyses. To avoid confusions of changing vocabularly, we will follow the anachronistic practice of consistently referring to the Normal, Regular Early, and Special Early plans in just that nomenclatuire, despite the fact



that the Normal Retirement Plan was for a time the only retirement plan and, then during a later period, the Regular Early Retirement Plan was the only employer-provided avenue for early retirement. THE GM PENSION PLANS The 1987-89 UAW-GM contract is a very complex, multivolume document. The portion that regulates the terms of retirement contributes its fair share to the complexity of the whole. Much of the complexity reflects the general historical trend toward greater formalization and legalization of substantive as well as procedural details of contractually regulated relationships. But a significant part of the complexity is due to increased differentiations within the financial provisions of employer-provided pension plans. As the various interests in early retirement were developed and formalized in union-management negotiations, for example, the formulas used to calculate pension-benefit payment schedules became ever more complicated. Perhaps, then, the best way to facilitate an accurate yet reduced description of the financial and related characteristics of the retirement plans is to begin with a brief developmental account. The Early Years In 1950, GM's pension plan for wage-rate employees—in effect, the Normal Retirement Plan—^had relatively simple financial provisions. Activatable at age 65, the monthly benefit was calculated as the product of a single "basic benefit rate" ($1.50 in 1950) and years of credited service, up to a maximum of 30 years (Marples, 1965). The provisions included periodic revisions of the basic-benefit rate, in accordance with general rates of inflation and wage-rate increases, for newly retiring workers; the revised basicbenefit rates were typically assigned to previously retired workers as well. Thus, in 1953 the benefit rate was raised to $1,75; in 1955, to $2.25; in 1962, to $2.80. In 1955 the 30-year ceiling on credited service was eliminated. By the time of Barfield and Morgan's study (i.e., the 1964-67 contract period), the pension-plan provisions had become considerably more complicated, both in the terms of eligibility and in the calculation of monetary benefits. On the one side, GM mandated retirement at age 68. The main features of the Normal Retirement Plan were unchanged, though the basicbenefit rate had been reaugmented. On the other side, qualified workers now had the option of early retirement (in effect, the Regular Early Retirement Plan). Any worker with at least 10 years of service could retire as young as age 60. Indeed, a worker as young as age 55 could elect early retirement if the sum of his or her age and service years totaled at least 85.

The Financial Structure of Early-Retirement Pensions


This last provision, commonly called "The Rule of 85," is an important provision in the 1964-67 contract and in subsequent contracts. A brief description of the point of this provision, and of its relationship to another provision, commonly called "30 and out," will be helpful. The main point of the Rule of 85 is that it regulates early-retirement age by a seniority principle. Consider, for example, two 58-year-old workers, one with 26 years of service and the other with 27 years of service, contemplating an early retirement: the Rule of 85 gives priority to the more senior of the two workers. Note that at minimum eligibility age of 55, the Rule of 85 implies 30 years of service. In other words, eligibility for early retirement at that age— still a very young age for retirement during the 1960s—is reserved to workers who have very high seniority. This specific version of the Rule of 85 is known as the "30 and out" provision, which is assigned special significance in that it carries an automatic maximum monthly benefit. To repeat, then, imder terms of the 1964-67 contract a worker as young as 55 could elect early retirement if the sum of the worker's age and service years equaled or exceeded 85. Unless this worker had at least 30 service years, however, retirement prior to age 62 carried a penalty: the basic-benefit rate was reduced more or less in proportion to the number of years short of age 62. It was during the 1964-67 contract period, in other words, that GM initiated a two-tiered scheme for the calculation of early-retirement pension benefits. The first tier—which reproduced the concept of the "basic benefit rate," though with some modifications—consisted of a lifetime benefit component. The second tier, a new development stimulated by sensitivities to the age 65 provision of Social Security—consisted of a supplemental allowance component. The lifetime benefit component was calculated as the product of a basic benefit rate (which was now graded by age at retirement as well as by wage-rate) and years of service. The maximum basic benefit rate was set at $4.25 (i.e., $4.25 per month per year of service) for those retiring at age 62 or older; it decreased gradually as age at retirement declined, reaching $2.46 for those retiring at age 55 (the minimum-age criterion of eligibility). The supplemental-allowance component was added to the lifetime-benefit component for workers who retired prior to age 65. Payable until age 65, the supplemental allowance was calculated as the product of a "supplemental benefit factor" (which also was graded by age at retirement) and years of service up to a maximum of 30 years. Thus, for a worker retiring at age 60 or older, the sum of the lifetimebenefit component and the supplemental-allowance component amoimted to $13.33 per month per year of service; for those retiring at age 55, the sum reduced to $6.67; for those retiring at ages 56 through 59, the sum lay more or less proportionately between those limits. Since this Regular Early Retirement Plan, like the Normal Retirement Plan, included an elective provision for a survivor's benefit, the monthly benefit would have been slightly



smaller than is implied by those total rates for any worker who chose to purchase the survivor's benefit.^ But, ignoring that factor, the summed rates imply that a worker retiring at age 62 with 30 years of service would have received a monthly benefit of $400 until age 65; after age 65 the supplemental-allowance component would have been deleted, reducing the monthly benefit to $127,50 (the difference being approximately supplied by Social Security). In effect and by design, then, the supplemental allowance offered a financial incentive to retire at an age younger than the "normal" retirement age of 65, and this incentive was graded by age and by seniority. In general outline, the foregoing description of the basic eligibility and benefit structures of the Normal Retirement Plan and the Regular Early Retirement Plan applies to the sequence of negotiated contracts following 1964-67. The benefit-rate schedules were periodically adjusted in order to keep pace with inflation and earnings growth; pension recipients gained periodic cost-of-living adjustments; and there were some other changes as well. But in general outline the two plans remained in the 1987-89 contract essentially as we have described them for the 1964-67 contract. The main differences in the retirement packages of the two contracts are, first, the development of a second early-retirement plan (the Special Early Retirement Plan)4 and, correlatively, some modifications in the way early-retirement benefits were actuarially adjusted. The 1987-89 Plan The retirement package for nondisabled workers in the 1987-89 contract contained three separate plans: the Normal Retirement Plan (NRP), the Regular Early Retirement Plan (RERP), and the Special Early Retirement Plan (SERP).5 NRP eligibility occurred at age 65, assuming the worker had a minimum of 10 years of credited service.^ The NRP monthly benefit consisted of two components, the lifetime-benefit component (as before) and a "special age-65 benefit." The lifetime-benefit component (common to all three plans) was calculated as the product of years of service and a basic-benefit rate which varied jointly by year of retirement within the contract period and by a "benefit class code" that was determined by the job classification (essentially the wage rate) held for the longest period of time during the 24 months immediately prior to retirement.^ The second component, the "special age-65 benefit," was designed to offset the Medicare Part B premium. Thus, for example, a worker who retired under NRP in 1989 at age 65 with 30 years of service received a basic monthly benefit ranging from $745.50 to $768.00 (depending on the benefit class code); adding the "special age-65 benefit," which for that retirement date was $27.00, yielded a total monthly

The Financial Structure of Early-Retirement Pensions


benefit ranging from $772.50 to $795.00. Note that these values (and the values in all later illustrations) assume that the worker did not elect the survivorship option; if this option had been elected, the basic monthly benefit w^ould have been reduced by 5 percent (i.e., ranged from $708.23 to $729.60, plus the "special age-65 benefit''). The 1987-89 provisions of RERP were an updated version of the earlyretirement plan previously described for the 1964-67 contract. Workers aged 55 through 65 with at least 10 years of service, and workers of any age with at least 30 years of service, were eligible under RERP. Those who had completed the 30th year automatically qualified for the maximum monthly benefit under the "30 and out" provision; those who were short of 30 years qualified for proportionately smaller monthly benefits. The RERP benefit consisted of two components, as before: the lifetime-benefit component and the supplemental-allowance component. The lifetime benefit was calculated as the product of a basic-benefit rate and years of service—^just as under NRP, except that under RERP the basic-benefit rate was subject to an age-graded adjustment. A worker retiring via RERP at ages 62, 63, or 64 received 100 percent of the basic-benefit rate. For those who retired between the ages of 55 and 62, the basic-benefit rate was reduced by a small percentage (ranging between 5.6 and 6.7 percent) for each year of difference between retirement age and age 62. For example, those retiring via RERP at age 61 received 93.3 percent of the basic-benefit rate; at age 60,86.7 percent; at age 59,80.8 percent; and so on, down to 57.9 percent of the basic-benefit rate at age 55. However, for any worker who retired with a minimum of 30 years of service, or who otherwise satisfied the Rule of 85 (i.e., the sum of age and years of service equaled or exceeded 85), the full basic-benefit rate was restored once the retiree reached age 62. The supplemental-allowance component of the RERP benefit was, as before, restricted to workers who retired at ages younger than 62. Those who retired before age 62 with at least 30 years of service qualified for a supplement designed to raise their total monthly benefit to a specified maximum level; if such a retirement occurred on or after October 1,1989, for example, the specified level was $1,500 per month (i.e., the maximum possible sum of the lifetime benefit and the supplemental allowance was $1,500 per month). For workers who retired before age 62 with fewer than 30 years of service, the supplemental benefit was determined as the product of a supplemental-benefit rate (graded by age at retirement) and years of service. Under all circumstances, the RERP supplemental allowance was subject to three restrictions: it was payable only until the retiree reached age 62; the sum of the lifetime benefit and the supplemental allowance could not exceed 70 percent of the worker's preretirement monthly base earnings; and postretirement employment earnings in excess of a specified amount ($8,500 in 1988; $9,200 in 1989) were taxed at a marginal rate of 50



percent (i.e., for every $2 in earnings above the ceiling, the supplement was reduced by $1). This third restriction terminated at age 62. Finally, we come to the Special Early Retirement Plan (SERF), which was designed to augment the attractions of early retirement under specified conditions and to enhance the flexible use of early retirement as a means of managing problems of "excess" labor supply in the internal labor market. Part of that design was manifested in some differences in the regulation of eligibility. Under SERP, workers as young as age 50 (rather than RERP's stipulated age 55) could apply if they had accumulated at least 10 years of service and if they were, or were soon to be, laid off because of a plant closure (and there were no other plants "in the same labor-market area," the latter as defined by the given state's Employment Security Commission) or if company management otherwise offered a "mutually satisfactory" opportimity to retire xmder the Special Plan.^ Whereas eligibility imder RERP was routinely defined and virtually automatic once a worker had satisfied the minimum age and seniority requirements, eligibility under SERP was more flexibly defined and was for all practical purposes regulated as "company option," in the sense that management defined the opportunities and reserved rights of approval. The monetary benefit available to workers who retired under SERP contained some of the features of the RERP benefit and some different features. The SERP benefit consisted of three components: the lifetime-benefit component, a temporary-benefit component, and a supplemental component (though this last component was empty in certain cases, as explained below). The lifetime-benefit component of SERP was the same as the lifetime-benefit component of RERP, with one important difference: SERP retirees received 100 percent of the basic-benefit rate (multiplied by years of service) regardless of age at retirement. The temporary benefit—similar in form to RERP's supplemental allowance (though different in rate structure)—was calculated as the product of years of service (up to a maximum of 30) and a "temporary benefit rate" which ranged from $19.20 for retirements initiated during the year beginning October 1,1987, to $21.40 for retirements initiated during the year beginning October 1, 1989. However, the temporary-benefit component had a ceiling (ranging from $576 to $642, depending on retirement date), and it was payable only until the retiree reached age 62. Also, receipt of the temporary benefit did not entail any restrictions on postretirement employment earnings. The third component of the SERP benefit, a supplemental allowance, applied only to SERP retirees who left with at least 30 years of service; they qualified for a supplement if and insofar as the sum of the lifetime-benefit component and the temporary-benefit component did not equal the standard "30 and out" maximum benefit provided under RERP (i.e., $1,500 for retirements initiated during the year beginning October 1, 1989). This supplement was subject to the

The Financial Structure of Early-Retirement Pensions


same marginal tax on postretirement employment earnings as applied under RERP. Also as in the case of RERP, the sum of all components of the monthly benefit under SERP could not exceed 70 percent of the worker's preretirement monthly earnings. EVALUATING THE FINANCIAL INCENTIVES In principle, any senior worker who has satisfied or is on the verge of satisfying the eligibility requirements of an early-retirement plan faces a sequence of decisions: Should I elect early retirement now, in this period (i.e., this year), or should I wait until the next period (or the nth next period)? In practice, the deliberative process is no doubt neither that uniform across all eligible workers nor, for many of them, quite so routinely periodic. At any given date, some portion of near-eligible workers will have already decided to leave, while others will have decided to "wait awhile," and still others will have decided to continue their employments until age 65 or later, or "as long as my health holds up," or "as long as I can," and so on. Likewise among those workers who are already well past the date of first eligibility. And of course all of these decisions—except for the completed decision to retire—can be and often are revisited again and again. But regardless which of those specific descriptions fits the deliberative process of any given worker (again, excepting the completed decision to take early retirement), each eligible worker does in effect engage in the sequence of periodic decisions, even if only by default. And at each period the worker's decision, whether virtually or actually made, carries certain financial implications. In Chapter 4 we will address issues concerning the impact of financial factors on the auto workers' actual early-retirement decisions. Here our concern is with a broader, descriptive account of the financial-incentive structures of the early-retirement plans. On the assumption that eligible workers, when actually deliberating the early-retirement question, included reasonably accurate information about the financial implications of leaving "now" rather than "later," we want first to know what those implications were. Or, to put the point in another way, given that the retirement plans were designed with temporally distributed sets of financial incentives which were expected to influence workers' decisions about the timing of retirement, we want to have some imderstanding of those distributed sets before attempting to assess the extent to which the workers' decisions were in fact influenced by them. In this context, "financial incentive" obviously refers to a relationship between the monetary value of the early-retirement pension benefit and the earnings gained from continued employment in GM's plants. But that



relationship can be specified in different ways. The "present value"^ of the monthly (or yearly) benefit is a necessary starting point, and no doubt variation in that value is highly salient information in the decisional calculus; but it is only one side of the relationship. At minimum, we assume, the auto workers evaluated the present value of their expected benefits relative to the earnings income that would be replaced. This evaluation can be conceived as a gradient of substitution, at some point on which a threshold effect will occur: below the threshold, retirement is highly unlikely; above the threshold, retirement becomes increasingly likely. Indeed, in their study of auto workers Barfield and Morgan reported evidence of a threshold level of retirement income which most people seem to consider necessary to ensure a reasonably adequate postretirement living standard. Currently this level is about $4,000 per year; it is likely, though, that $4,000 is not an absolute figure, but one which reflects a current consensus about the minimum income necessary to provide reasonably comfortable living after retirement. Thus, the "threshold" level may shift upward over time as living standards generally rise— and this upward movement should be all the faster if price level increases are not kept within reasonable bounds (Barfield and Morgan 1970, p. 3). A standard approach to the measurement of that substitution threshold is the "replacement rate"—or the ratio of the expected pension benefits to the earnings forgone by retirement. The "replacement rate" threshold need not be uniform across all situations, of course; in fact, it typically is not uniform. Other factors, such as spouse's income, can affect the threshold value. In any given period an eligible worker deliberating the binary decision, "retire this period or wait until a later period," will presumably consider the replacement rate. However, the incentive features of the retirement plans here under scrutiny involve a good deal more than the expected amount of a monthly pension benefit in ratio to the current amount of monthly earnings from GM. In the first place, each of those terms of the replacement ratio represents a single period in a stream of income receivable over some definite but still unknown number of future periods.^o When workers deliberate the "now or later" decision, they presumably consider in some manner an expected trade-off between the two streams, since these streams are mutually exclusive. Second, each stream can be treated as having properties of wealth or asset holdings as well as having properties of income. When workers deliberate the "now or later" decision, they presumably consider in some manner the volume of wealth vested and accruable in the pension plan, relative to the volume of wealth potentially accurable from continued employment earrungs (i.e., private savings), with the understanding that (ignoring the probability of "dying on the job") a sub-

The Financial Structure of Early-Retirement Pensions


stitution will begin at some future date. Third, the lifetime-payment schedule of a defined-benefit pension plan is necessarily informed by and regulated by one or another set of specific assimiptions or stipulations about life expectancy, rate of asset accrual, and actuarial fairness; and competing plans offered by a given employer (e.g., NRP, RERP, SERF) may be differently structured in those respects, depending on (among other factors) employer interests, or negotiated employer and union interests, in using the plans as instruments for managing labor supply in the internal labor market. When workers deliberate the "now or later" decision, they are presumably sensitive in some degree to the variable implications of those assumptions and stipulations. The preceding paragraph introduces several important factors in the financial-incentive structure of the retirement plans. They will be developed and utilized later in this chapter. For the moment, we simply emphasize the general point that whereas the replacement-rate approach to measuring financial incentives (and workers' responses thereto) assvones that workers are primarily concerned with pension benefits as immediate income replacement, an alternative approach, which assumes that workers are primarily concerned to maximize assets within a remaining-lifetime framework, better captures the incentive structures of the plans (see, e.g., Burkhauser 1979; Fields and Mitchell 1984a; Gustman and Steinmeier 1989; Ippolito 1986a; Kotlikoff and Wise 1989a, 1989b; Mitchell and Fields 1982; Quinn, Burkhauser, and Myers 1990). Whether this alternative approach also better represents the actual deliberation—that is, the decisional calculus—performed by workers when they evaluate the financial aspects of the ''now or later" decision is a separate, empirical question, which will be addressed in Chapter 4. But there is no doubt that it provides a much superior evaluation of the financial-incentive structures of the plan—and therefore (whether the decision maker utilizes the information or not) of the financial implications of a decision to retire "now" (i.e., in any given period) under any given retirement plan. In the remainder of this section we will work toward that evaluation developmentally, as it were, beginning with some simple comparative descriptions of the present value of monthly benefits. Comparing Monthly Benefits A preliminary and partial understanding of the relative incentives/disincentives of different early-retirement plans can be gained from comparisons of monthly benefits under two versions of the Regular Early Retirement Plan, the version initiated in the 1964-67 contract and the version contained in the 1987-89 contract. Our point of departure for these comparisons is Table 3.1, which contains sets of benefit rates applicable to




^ S S! ^ ?^ ^ ^ ^ ^ ^ r:


\£> t>.* 06 ON rH en CO C*! CT) CfJ '^'




illlllliiii sli?SI5SSSS

I s





1 I

I rn o b- S t-H ON I I< 06 00 ON O O





T-4 H r4 r«i r4 r4


The Financial Structure of Early-Retirement Pensions


retirements at various ages under the contracts effective in 1961,1964, and 1989. Before proceeding with the illustrative comparisons, we should note some features of the table and the main assumptions behind these specific sets of benefit rates. First, the rates pertain only to versions of RERJP (i.e., the rates are not uniformly correct for the calculation of SERF benefits in 1989). Second, in 1961 pension activation could not begin until age 60, and the supplemental allowance had not yet been introduced. In 1964 eligibility for the supplemental allowance ceased at retirement ages older than 64 (in 1989, at retirement ages older than 61), and the allowance was payable until age 65 (in 1989, until age 62). Third, the left-hand portion of the table reports rates for the lifetime-benefit component of a monthly benefit; the right-hand side reports the sxmi of the lifetime-benefit rate and the supplemental-allowance rate. Fourth, the lifetime-benefit rates are adjusted for early-retirement age (e.g., at age 55 in 1989 the rate was 57.9 percent of the maximum rate of $26.50, or $15.34). Fifth, the combined rates (right side of table) are generally applicable to early retirement at seniorities younger than 30 service years; at 30 or more years of service the monthly benefit was automatically set at the maximum of $400 in 1964 (if retirement age was at least 60 but less than 65) and at the maximimi of $1,500 in 1989 (if retirement age was less than 62). Sixth, the rates are reported both in nominal terms and in terms of 1989 dollars.!^ Finally, the calculations on which the following comparisons are based ignore all future-year cost-of-living adjustments to the benefits and assume that the retirees did not elect the survivor's benefit option. Let's begin the comparisons by considering three hypothetical cases, each a retirement at age 60 after 25 years of service, but each during a different year, 1961,1964, and 1989. Name them Pat^^^j, Jesse^^^, and Lynn^^g^. In nominal-dollar terms, the monthly benefits compare as foUow: Peit^g^y $1.88 X 25 = $47.00 per month until death Jesse^^^: $13.33 X 25 = $333.25 per month until age 65, then $3.68 X 25 = $92.00 per month until death Lynn^^g^: $43.28 X 25 = $1,082.00 per month until age 62, then (under the Rule of 85) $26.50 X 25 = $662.50 per month imtil death In terms of 1989 dollars, the monthly benefits compare as follow: Patj^gji $7.33 X 25 = $183.25 per month until death Jesse^^^: $49.47 X 25 = $1,236.75 per month until age 65, then $13.66 X 25 = $341.50 per month until death Lynn^^g^: $43.28 X 25 = $1,082.00 per month until age 62, then (under the Rule of 85) $26.50 X 25 = $662.50 per month until death



Relative to Patj^^j, it is clear, Jesse^^^ enjoyed a substantial advantage because of changes in the policy governing early-retirements. The advantage was twofold. First, the lifetime-benefit component had increased 86 percent (from $183.25 to $341.50) in constant-dollar terms. Second, the supplemental allowance, not available in 1961, paid Jesse^^^ a "bonus" which raised his or her total monthly benefit to 362 percent of the lifetime benefit each of the 5 years that counted as "early-retirement years" (i.e., from age 60 to age 65). Relative to Jesse^^^, Lynn^^g^ experienced a somewhat more mixed set of advantages. On the one hand, the lifetime-benefit rate had been increased 68 percent in constant-dollar terms (from $13.66 to $22.98); but since Lynn^^g^ automatically qualified for the maximum lifetime-benefit rate ($26.50) because of the Rule of 85, the improvement was an even greater 94 percent. On the other hand, the supplemental-allowance rate in effect in 1989, though still amounting to a "bonus" (which raised the total monthly benefit paid Lynnl989 to 163 percent of the lifetime benefit), was payable orUy to age 62 (not to age 65 as with Jesse^^^) and was proportionately less lucrative than it had been (having declined 53 percent in constant-dollar terms). Thus, during each of the 2 years that counted as earlyretirement years for Lyruij^g^ (ages 60 and 61) the total monthly benefit was actually 12.5 percent lower (in constant dollars) than it had been during each of the 5 years that counted as early-retirement years for Jesse-^^^ (ages 60 through 64). While one should bear in mind that the comparative figures we have just examined are specific to a particular combination of age and service years, two generalizable points can be drawn from the comparisons. First, the different benefit schedules in effect for Jesse^^^ and Lyim^^g^ contained some incentives in favor of early retirement in each case. Second, the incentives were somewhat weaker for Lynn^^g^ because the rules of eligibility had been changed in ways that encouraged early retirement at ages younger than 60 (the age at which we had each of the three workers retiring), and this liberalization of the eligibility rules was "paid for," as it were, by some changes in the structure of benefit rates. The trade-off between benefit rates and liberalized terms of eligibility can be illustrated in part by comparing, across contracts, the maximum allowable early-retirement benefits and the criteria under which the maximum benefits were payable. (Again, we are restricting the comparison to RERP.) In 1964 the maximum allowable sum of the lifetime-benefit component and the supplemental-allowance component was $400 per month, which translates to $1,486 in 1989 dollars. In 1989 the corresponding value was $1,500. In other words, there had been virtually no improvement in constant-dollar terms. However, in 1964 the maximum monthly benefit was available only to workers who had at least 30 service years and were at

The Financial Structure of Early-Retirement Pensions


least 60 years old (but younger than 65), whereas in 1989 the $1,500 maximum was payable to workers of any age, provided they had at least 30 service years. Further, while in 1964 the maximum benefit was payable each month for a total of 5 years (ages 60 through 64), in 1989 the maximum benefit, though payable only until age 62, could have been paid for a much longer period of time. Thus, while the constant-dollar value of the maximum benefit had remained virtually unchanged, the eligibility criteria that were in effect in 1989 made proportionately more workers eligible for early retirement and lengthened the allowable interval of early-retirement benefit payments. Workers with high seniority gained the most from changes in the terms of the Regular Early Retirement Plan. Let's recall our three hypothetical workers; assume that each retired not at age 60 after 25 years but at age 55 after 30 years (and note that either combination totals 85). P^^96i pi'obably would not have retired at that age even with 30 years of service, because in 1961 pension activation could not occur at ages younger than 60. Jessej9^4 would have received a monthly benefit of $200.10 ($6.67 X 30) until age 65 and $73.80 ($2.46 X 30) thereafter; the corresponding values in 1989 dollars are $742.80 and $273.90. Lyim^^g^ would have received a monthly benefit of $1,500 (the automatic maximum at 30 service years) imtil age 62 and $795 ($26.50 X 30) thereafter. Now compare these monthly benefits for our revised versions of JessCj^^^ and Lynn^gg^ (retiring at age 55 with 30 service years) with the corresponding constant-dollar monthly benefits for the previous versions of the two workers (retiring at age 60 with 25 service years). It is evident that the revision was advantageous to L5mn^9g9 but disadvantageous to Jesse^^^. For example, by retiring at age 55 after 30 years, rather than at age 60 after 25 years, Lynn^^g^ received $1,500 per month during the first 7 years of retirement (i.e., to age 62) rather than $1,082 per month during the first 2 years of retirement (i.e., to age 62). Conversely, Jesse^^^ received (in 1989 dollars) $742.80 per month during the first 10 years of retirement (i.e., to age 65) rather than $1,236.75 per month during the first 5 years of retirement (i.e., to age 65). The foregoing illustrations—although only a sampling of the many possible combinations of age and seniority criteria, and restricted to only one of the two early-retirement plans available in the 1987-89 contract period (the Special Plan introduces still more variations)—give a reasonably good initial picture of the financial incentives favoring early retirement. The picture is incomplete, however, primarily because it has been drawn



only in terms of "single period" assessments of the present value of monthly benefits. It is time to shift our basis of assessment to the concept of pension wealth. Calculating Pension Wealth Imagine a time-line for a worker who was born in 1934. Assume this worker was hired by GM in 1959 at age 25; assume a life span of 75 years. 1934

















The half-century between hiring in 1959 and death in 2009 can be divided into two periods, employment and retirement. During the years of employment, GM provides a total compensation package consisting of current wages, health and life insurance, and deferred compensation in the form of pension rights. Retirement means cessation of the current-wage compensation and (assuming the minimum-age requirement is met) activation of the defined-benefit payments. Because the lifetime-benefit portion of the latter is payable until death, the younger the age at which the worker retires, the longer the worker's remaining lifetime and the longer the period of benefit payments. Thus, the younger the age at retirement, the larger the total sum of benefit payments—unless the size of the annual benefit is actuarially reduced, that is, adjusted for the probable length of remaining lifetime (which, bear in mind, is a function of a knowable value, age at retirement, and a value estimable from age-specific life-expectancy tables, projected age at death). We have already seen one facet of the actuarial adjustments—namely, in the age-graded benefit rates described in conjunction with some of the preceding descriptions and comparisons of monthly benefits. Actuarial adjustments can be made in different ways (or not at all), in accordance with different goals. Adjustments designed to achieve what is conventionally termed "actuarial fairness" attempt to equalize the value of benefits paid to workers across different remaining-lifetime payment periods. In other words, an "actuarially fair" schedule of adjustments would result in a range of benefit payments so that neither workers who retire "early" nor workers who retire "late," relative to some "normal" retirement age, would be either rewarded or penalized for the decision to retire "early" or "late." Thus, if benefits are actuarially fair, the lower monthly benefits paid to early retirees over a longer period of time and the higher monthly benefits paid to later retirees over a shorter period of time will

The Financial Structure of Early-Retirement Pensions


sum to the same value. A firm that has wittingly agreed to an actuarially fair pension scheme has effectively announced that it is indifferent to the timing of the retirement transitions of its workers; the benefit structure of its pension scheme contains no inherent incentives or disincentives favoring or disfavoring retirement at any particular combination of age and service years (aside from minimum eligibility criteria). As we have seen in the preceding illustrations, the negotiated contracts have contained some provisions that tend in the direction of actuarial fairness. The lifetime-benefit component has been anchored by the stipulation of what amounts to a "preferred" age of early retirement: age 62 in both the 1987-89 and the 1964-67 contracts. Whereas an early retirement at ages 62,63, or 64 yielded a lifetime-benefit component calculated at 100 percent of the basic-benefit rate, an early retirement at ages increasingly younger than age 62 yielded lifetime-benefit components calculated at proportionately smaller percentages of the basic-benefit rate (e.g., at age 55 in the 1987-89 contract the basic-benefit rate was 57.9 percent of the maximum rate of $26.50, or $15.34)—though only if the retiree left with fewer than 30 service years. In short, for workers who had not achieved the 30th year of service the incentives in favor of early retirement were age-graded in the direction of actuarial fairness. In other respects, however, the structure of benefits contained some much stronger incentives in favor of early retirement which tended away from actuarial fairness. We noted just above, for example, the fact that for workers who had accumulated at least 30 service years the basic-benefit rates were not age-graded in the direction of actuarial fairness. Consequently, the younger the age of retirement at the end of the 30th (or later) year of service, the larger the sum of remaining-lifetime payments. Other strong incentives were provided in the Special Early Retirement Plan, which until now we have been neglecting. A partial appreciation of the relative incentives of the Regular and the Special Early Retirement Plans can be gained from comparisons of total annual benefits payable to workers leaving at the end of the 20th service year and to workers leaving at the end of the 30th service year, under each of two early-retirement plans, at each of two retirement ages, age 55 and age 60, in 1989. These comparisons are shown in Table 3.2 (as with all previous illustrations, the calculations assume that the survivor's benefit option was not taken).i2 First, consider the annual benefit that was payable until age 62 (the first row of each of the two panels in Table 3.2) and then the annual benefit that was payable from age 62 forward (the second row of each panel). In general, it is apparent that whereas annual benefits paid after 30 service years were indifferent by retirement plan and by retirement age (because at 30 years the maximum monthly benefit of $1,500 was automatic), the bene-

Chapter 3


Table 3.2. Comparing Annual Benefits and Pension Wealth 20 years of service

30 years of service





Retirement age = 55 Annual benefit At ages < 62 At ages 62+ Pension wealth

$5,938 $3,682 $7430

$11,496 $6,360 $134,379

$18,000 $9,540 $205,017

$18,000 $9,540 $205,107

Retirement age = 60 Annual benefit At ages