When Strikes Make Sense - And Why: Lessons from Third Republic French Coals Miners (Springer Studies in Work and Industry)

Citation preview

When Strikes Make Sense— And Why Lessons from Third Republic French Coal Miners

PLENUM STUDIES IN WORK AND INDUSTRY Series Editors: I v a r Berg, University of Pennsylvania, Philadelphia, Pennsylvania and A m e L. Kalleberg, University of North Carolina, Chapel Hill, North Carolina

WORK AND INDUSTRY Structures, Markets, and Processes Ame L. Kalleberg and Ivar Berg

THE BUREAUCRATIC LABOR MARKET The Case of the Federal Civil Service Thomas A. DiPrete ENRICHING BUSINESS ETHICS Edited by Clarence C. Walton ENSURING MINORITY SUCCESS IN CORPORATE MANAGEMENT Edited by Donna E. Thompson and Nancy EHTomaso INDUSTRIES, FIRMS, AND JOBS Sociological and Economic Approaches Edited by George Parkas and Paula England LIFE AND DEATH AT WORK Industrial Accidents as a Case of Socially Produced Error Tom Dwyer MATERNAL EMPLOYMENT AND CHILDREN'S DEVELOPMENT Longitudinal Research Edited by Adele Eskeles Gottfried and Allan W. Gottfried THE STATE AND THE LABOR MARKET Edited by Samuel Rosenberg WHEN STRIKES MAKE SENSE—AND WHY Lessons from Third Republic French Coal Miners Samuel Cohn WORKERS, MANAGERS, AND TECHNOLOGICAL CHANGE Emerging Patterns of Labor Relations Edited by Daniel B. Cornfield 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.

When Strikes Make Sense— And Why Lessons from Third Republic French Coal Miners Samuel Cohn Texas A&M University College Station, Texas

Plenum Press • New York and London

Library of Congress Cataloglng-in-PublicatIon Data

Cohn, S a m u e l . 1 9 5 4 When s t r i k e s make s e n s e - - a n d why ; l e s s o n s f r o m T h i r d R e p u b l i c F r e n c h c o a l m i n e r s ^ Samuel Cohn. p. cm. - - (Plenum s t u d i e s i n work and I n d u s t r y ) I n c l u d e s b i b l i o g r a p h i c a l r e f e r e n c e s and i n d e x . ISBN 0 - 3 0 6 - 1 4 4 4 5 - 3 1. S t r i k e s and l o c k o u t s — C o a l m i n i n g — F r a n c e — H i s t o r v - - 2 0 t h century. 2 . Wages—Coal m i n e r s — F r a n c e — H i s t o r y — 2 0 t h c e n t u r y . I. T i t l e . I I . Senles. HD5375.M615C64 1993 331.89'2--dc20 93-28082

CIP

ISBN 0-306-44445-3 © 1993 Plenum Press, New York A Division of Plenum Publishing Corporation 233 Spring Street, New York, N.Y. 10013 All rights reserved No part of this book may bereproduced,stored in a retrieval system, or transmitted in any form or by any means, electtonic, mechanical, photocopying, microiilniing, recording, or otherwise, without written permissionfiromthe Publisher Printed in the United States of America

Preface

Social scientists have not helped the working class make strategic decisions. Unionists need to know how to carry on industrial conflict so as to provide concrete economic benefits for their members. Should unions strike or not strike? Should losses be avoided at all costs, or can unions afford to take chances? Does economism gut the class power of workers or provide a pragmatic strategy for increasing workers' wage gains? We can say with great confidence that workers should join unions; there is now an exhaustive and compelling literature demonstrating that union membership provides a wide variety of economic benefits. We can say that corporatist class compromises lower income but increase job security and overall employment. Beyond that, however, we cannot say much. In particular, we can do little to advise particular unions in particular fixed institutional and political environments how they should handle the microtactics of individual confrontations. The United Farm Workers do not need a speech about the miracle of the Swedish industrial relations system. They need to know whether they should strike or not strike, and how their tactics should change if rival Teamsters come into the field. Unfortunately, medical research often has to start with rabbits before it proceeds to humans, and so it is with research in industrial conflict. The realistic prospects of doing a large sample analysis of contemporary American wage settlements that simultaneously estimates the effects of union tactics and economic factors are poor. One turns to earlier periods and other nations, though, and all these technical problems evaporate. This book contains an analysis of unions' strategic effectiveness in twenty separate labor markets over forty-five years each. Exogenous market factors can be almost perfectly understood and adjusted for. The differences in union tactics and strategy were major and profound.

Foreword

Popular collective action often confounds theories of collective action, especially when those theories adopt an aggressively individualistic perspective. According to a plausible and widely accepted theory, the temptation of any particular individual to ride free when others will pay the fare—not to assume the costs and risks of joining an action when one's own effort will make little difference to the outcome—undermines a wide range of possible actions. Yet people often commit their energies to worthy causes, and they sometimes risk their lives for shared objectives outsiders regard as imprudent or even insane. Violence in victory is selfcontradictory for warriors who have fought to seize enemy resources for the benefit of their own side; so, at least, says one sort of strategic theory. Still, looting and wanton destruction frequently break out after the battle is won. Another persuasive theory tells us that by disrupting production, hence the rewards of production, strikes attack the interests of both workers and employers; they constitute an expensive way of gauging the relative strengths of capital and labor when a cool appraisal of bargaining conditions prior to any strike could provide the same information at lower costs to all concerned. Yet strikes occur all the time. Workers and union leaders even think they gain from well-planned walkouts. Here Samuel Cohn enters the scene, great masses of evidence in hand, to argue that the persuasive theory has crippling defects. He uses exquisitely crafted tools to deconstruct the theory. Cohn is not content merely to tear down conventional wisdom and well-established theory. He insists on erecting in their place a quite different set of models—models premised on continuous interaction among workers, employers, and state officials; models representing not

vi

Foreword

merely the instantaneous but also the cumulative effects of strike activity; models incorporating the presence or absence of multiple unions; and models in which short-term rises in wages are by no means the only benefits that massed workers seek or should seek. Where the great majority of theorists have implicitly treated strikes as the doings of unitary actors—workers, unions, or some combination of them—he places bargaining among workers, unions, employers, and government squarely at the center of his analysis. He thereby demonstrates, among other things, how much difference the unity or segmentation of labor unions makes to employer strategies, and hence to the outcomes of strikes. Cohn began with a body of evidence that Edward Shorter and I accumulated years ago, as well as with a set of puzzles the Shorter-Tilly analysis of French strikes left unresolved. But he was not content merely to tidy up the cluttered space Shorter and I left behind. He not only refined and corrected our descriptions of miners' strikes, but he also expanded enormously the range of evidence concerning the mining industry and its workers. Furthermore, like most other analysts of industrial conflict. Shorter and I concentrated on the causes of strikes, giving only passing attention to their effects. Cohn has reversed that emphasis, constructing the most extensive and sophisticated analysis of strikes' outcomes that anyone has ever done. This move has brought him much closer to the strategic calculations of workers, unions, employers, and governments than Shorter and I managed to come. With meticulously analyzed evidence from thousands of miners' strikes, coupled with independent data on the same miners' wages and conditions of production, Cohn demonstrates the credibility of his new models under a wide range of conditions. He arrives, moreover, at some surprising results: the unimportance of the difference between winning strikes and losing them as compared to the sheer frequency of strikes; the identification of circumstances in which losing strikes produces greater gains than winning them; the considerable impact of strikes over other issues than pay on wage changes; and the greater worker gains from short than from long strikes. Cohn and his findings repeatedly challenge common sense by demonstrating that apparently obvious strategies—such as bread-and-butter business unionism—contain major contradictions. With exemplary clarity, Cohn lays out his ideas, sources, methods, findings, substantive conclusions, and strategic inferences. Informed readers are likely to wonder how widely conclusions drawn from the experience of French miners between 1890 and 1935 apply to other times and places: In our own American era of mobile capital and retreating

Foreword

vii

unionism, do the same strategic laws still prevail? Samuel Cohn's powerful analyses give us new reasons for wanting to know. CHARLES TILLY

New School for Social Research New York, New York

X

Preface

Questions of militant versus nonmilitant, short versus long strikes, economic versus political, unified versus divided—all these dimensions can be addressed. Admittedly, the conditions of French coal mining do not match those of tomato farming in the Rio Grande Valley or tire manufacturing in eastern Ohio. Precise results from parallel cases, however, are far more useful in social science prediction than is total ignorance. If a drug makes cancers disappear in rabbit livers, our best first guess is to try the medicine on similarly afflicted humans as well. This book would not have been possible without the generosity of many people. Charles Tilly was essential to the project: He provided machine-readable and photocopied accounts of all coal-mining materials in the Statistiques des Greves. He also read numerous drafts of early versions of the work, provided insight and inspiration, and provided sorely needed guidance on basic survival in France. Olivier Kourchid identified the primary archival sources for the northern coal miners and provided numerous oral insights on the history of the mines. Adrienne Eaton and Carol Conell collaborated with me on parallel analyses reported elsewhere that were essential to developing the final models of the present work. Kelley Winsborough, Ron Rodgers, Nan Yang, Stephen Petterson, and (again) Adrienne Eaton served as graduate assistants on the project and performed crucial functions in getting the initial data up and running. Kathryn Amdur provided an early diagnosis of some poorly conceived preliminary findings that, at the cost of adding two years to the duration of the project, saved me from some very serious errors. Her critiques spurred my trip to the Loire archives, in which were found the most informative qualitative materials in the analysis. Gerry Friedman single-handedly tutored me in the political and labor history of the p r e World War I period and shaped most of my thinking on early corporatist states. Pamela Oliver, Charles Halaby, Gerry Marwell, Sanford Jacoby, Richard Lachmann, Steven Rytina, John Markoff, and a number of anonymous referees (including one outstanding reviewer at Political Power and Social Theory) all provided early readings that produced major changes. The resulting book is not perfect, but, thanks to these people, it is much better. Daniel Schuman put me u p in Paris, making work at the Archives Nationales affordable. Last, Lynn Wallisch and Alexander Wallisch kept the long nights in Arras and Saint-Etienne from being lonely ones and provided merciful alternatives to the computer in Wisconsin and Texas. I owe great debts to them all. SAMUEL C O H N

Contents

Chapter 1 Union Choices Affect Wages French Coal Miners as a Critical Case A Brief Historical Introduction to French Coal Mining Overview of the Book Appendix: Maps of French Coal-Producing Regions

1 6 10 20 22

Chapter 2 Strikes and Wages: The Theoretical Relationship Strikes, Market Competition, and Managerial Logic: The SingleUnion Case Militancy, Steering, and the Prisoners' Dilemma: Strike Strategy in the Two-Union Case Predictions from the Model for French Coal Mining

27 28 34 40

Chapter 3 Strikes and Wages: The Statistical Relationship

43

Data and Methods Findings Conclusion

43 52 64 xi

xii

Contents

Chapter 4 Empty Victories, Hannless Defeats: The Relationship between Moderation, Strike Results, and Substantive Benefits

65

Operationalizing the Moderation of Demands Issue-Specific Success Rates Global Strike Outcomes The Effect of Strike Results on Wages Conclusion

67 70 79 83 85

Chapter 5 Is What You Ask For What You Get? Trade-Offs between Nonmonetary Benefits and Wage Attainment The Correlation between Wages and Nonmonetary Benefits Grievance Trade-Offs within Strikes The Effects of Pay and Nonpay Strikes on Wages Conclusion

87 93 99 102 103

Chapter 6 The State versus Managers: Contradictions of Early Corporatism in French Coal Mining 105 Managerial Indifference to Strikes Governmental Aversion to Strikes Factors Predisposing the State to Prefer Low Wages Factors Predisposing the State to Prefer High Wages Conclusion

107 113 120 126 135

Chapter 7 Militancy and Accomimodation: The Intimidation of Authorities by Unified Workers

137

Campaign for the Convention of Arras, 1889-1891 Gard, 1890 The 1892 Strike Law Carmaux, 1892 Montceau-les-Mines The National Strike Wave of 1902

138 140 141 141 143 146

Contents

Militancy Without State Support: Communist Strikes in the Loire

xiii

155

Chapter 8 Schism and Steering: State Control over a Divided Labor Movement

163

When Steering Equals Accommodation: The Response to Prewar Schisms Pure Steering in the Postschism Period The Continuing Institutionalization of Labor Relations Ethnic Divisions in the Labor Force Conclusion

164 175 186 191 195

Chapter 9 Limits of Militancy

197

Impact of Strikes Wages, Strike Costs, and Strike Duration Minor Role of Industrial Conflict

198 200 204

Chapter 10 Making Strikes Pay

213

Militancy Makes Sense Bona Fide Strategic Dilemmas

213 221

Cashing In on the Good Years

226

References

227

Index

239

Figures and Tables

Figures 1.1 Wages, Employment, and Prices in French Coal Mining . . . . 1.2 The Primary Coal-Producing Departments of France, 1890-1935 1.3 Regional Differences in Pay: Average Francs per Workday in French Coal Mines, 1890-1935 1.4 Regional Differences in Employment: Average Number of Workers in French Coal Mines, 1890-1935 1.5 Regional Differences in Productivity: Average Tons of Coal per 100 Workdays in French Coal Mines, 1890-1935 1.6 Union Affiliation after the 1992 Schism: French Coal Mines, 1890-1935

16 22 23 24 25 26

Tables 2.1 Prisoner's Dilemma Revisited: Wage Outcomes in a Two-Union Case by Strike Propensity 3.1 Parks Estimates of the Effects of Market and Conflict Variables on the Wages of French Coal Miners, 1890-1935 3.2 Parks Estimates of the Effects of Market and Conflict Variables on the Wages of French Coal Miners, 1890-1935, by Period 3.3 Parks Estimates of Period Interactions of Conflict Variables on the Wages of French Coal Miners, 1890-1935 3.4 Parks Estimates of the Effects of Period, Market, and Conflict Variables on the Wages of French Coal Miners, 1890-1935

37 53 55 57 58 XV

xvi

Figures and Tables

3.5 Residuals from Core Economic Model by Union Affiliation, 1923-1935 3.6 Mean Residuals from Core Economic Model by Strike Activity by Union Affiliation, 1923-1935 3.7 Mean Residuals from Core Economic Model by Strike Activity by Department and Strike Affiliation Simultaneously, 1923-1935 4.1 Theoretical Predictions of the Difficulty of Winning Different Types of Grievances 4.2 Failure Rates of Grievances by Type of Demand, 1890-1935 4.3 Logistic Regression of Strike Success on Selected Variables, 1890-1935 4.4 Parks Estimates of the Effects of Strike Results on Wages, 1890-1935 5.1 Mean Salary Rates per Day by Length of Workday: Individual French Mines, 1892 5.2 Regressions of Meant Salary Rates on Length of Workday and Mine Size, Individuals, and Mines, 1892 5.3 Parks Estimates of the Effects of Mining Fatality Rates and Control Variables on Wages, 1890-1935 5.4 Outcomes of Grievance Pairs, 1890-1914 5.5 Parks Estimates of the Effects of Monetary Strikes, Nonmonetary Strikes, and Control Variables on Wages, 1890-1935 6.1 Coal Supplies of Loire Metal Firms, January 1908 6.2 Estimates of the Percentage of Seats in the Chamber of Deputies Held by Leftist Parties 7.1 Contract Results of the 1902 National Coal Mining Strike by Region and Method of Settlement 8.1 Average Daily Salaries for Workers in the Four Largest Loire Mining Companies before and after the 1906 Strike 8.2 Quotes from Pas de Calais Public Administrators Concerning Steps to Bolster the Authority of the Reformist Union 9.1 Regression Estimates of the Effect of Strike Duration of Residual Wages 9.2 R-Squares for OLS Regression Models of Wages 9.3 Ratios of the Effects of Conflict Variables to the Effects of Market Variables

59 61 62 76 77 81 84 95 96 97 100 103 116 127 151 170 184 202 205 206

1 Union Choices Affect Wages

"Union choices affect wages." This sentence does not mean the same as "Unions raise wages." It is well known that union membership raises wages. There is now a very large econometric literature on this topic using a wide variety of analytic methodologies and controls, all of which show a positive union wage effect, although the size differs from one study to another (Lewis 1963; Parsley 1980; Freeman and Medoff 1984; Rosen 1969; Mishel 1986; Podgursky 1986). Not all unions, however, are alike. Some unions strike frequently; others hardly strike at all. Some unions win their strikes; others lose. Some unions are economistic and ask only for money; others have broad political and shopfloor agendas. Some unions are moderate; others openly espouse revolution. Some unions merge with other unions or work with rivals in a cooperative fashion; others break apart to maintain their own ideological purity. Surely these differences have an effect on the outcomes of labor negotiations. One might argue, following Hicks and generations of neoclassical economists, that all strikes are irrational (Hicks 1963; Reder and Neumann 1980; Machlup 1946). Perfectly rational bargainers ought to be able to foresee the results of strikes and therefore generate wage agreements that simulate the results of overt conflict while incurring none of the costs. But one could also argue, following many radical critics of the labor movement, that all strikes are rational (Piven and Cloward 1977; Lens 1959; Brecher 1977). Employers never cede anything unless subjected to overt pressure. Moderation gives managers a free hand to manipulate contracts to their own advantage; furthermore, passivity lowers workers' class consciousness, reducing their capacity to engage in major struggles at later dates. One might argue, from a narrow tactical perspective, that it does not matter whether strikers strike frequently or infrequently, but wheth-

2

Chapter 1

er they win or lose. This position has rarely received an explicit defense because it seems so tautological. It is implicit, though, in resource mobilization theory and histories that measure the progress of the labor movement by the results of landmark strikes (Shorter and Tilly 1974; Oberschall 1973; Foner 1964; Kriegel 1964). But one could also argue that strike results are irrelevant. In Hicksian theory, a strike settlement is the function of both the relative strengths of bargainers and the relative centrality of the issues being argued about. If strike results are driven more by one side's willingness to avoid conflict, then losses may actually increase the size of future settlements by demonstrating a willingness to engage in fanatic, suicidal struggle. Followers of Wisconsin school theorists, meanwhile, might argue that workers ought to confine their strikes to economistic issues (Commons 1913; Perlman 1928; Cochran 1977). Political and job control demands dissipate workers' power over many agendas. The more ambitious of these, such as radical social reform, are only attainable in revolutionary moments; in routine negotiations, they represent a waste of bargaining potential. One could also argue, though, that broad political and social agendas are a necessary part of intimidating employers into making concessions. If employers believe that unions with radical ideologies are liable to strike just on general principle, they will be less likely to believe that labor peace can be easily bought with desultory concessions. One could argue, again following American institutional economists and labor market segmentationists, that single-union representation is essential for obtaining meaningful wage gains (Sturmthal 1972; Lorwin 1954; Gordon 1982; Reich 1981). Labor's strength comes from unity, and unions that represent only a fraction of the labor force can be undercut easily by their passive rivals. But one could also argue, following British institutionalists and an assortment of writers from other traditions, that workers benefit from multiple representation (United Kingdom Royal Commission on Trade Unions and Employers Associations 1968; Clegg 1970; Ross 1948; Garmy 1934). Rival unions have to vie with each other for membership; this produces performance pressure which spurs both sides to bring home larger settlements. Furthermore, if a radical vanguard is essential for workers to make significant gains, such a vanguard could be suppressed in a bureaucratic and oligarchical giant union. These are all interesting theories. Surprisingly, however, social science research has produced little empirical evidence to adjudicate the conflicting views. There is a great deal of literature on the causes of

Union Choices Affect Wages

3

Strikes (Shorter and Tilly 1974; Cronin 1979; Edwards 1981; Cousineau and Lacroix 1986), but very little about the effects of strikes. We have a small and inconclusive literature on the effects of revolutions (Kelley and Klein 1981; Eckstein 1977), and there is a slightly larger literature on the correlates of strike victories (Knowles 1952; Ragin, Coverman, and Hayward 1982; Griffin, Wallace, and Rubin 1986; Edwards 1981; Snyder and Kelley 1976). Strike results, however, are not the same as the effects of striking. Result studies can differentiate those strikes in which the employer made concessions at the end of the stoppage from those in which no demands were granted, but this is not the same as knowing whether workers who struck received greater or smaller improvements in their standard of living than did those who abstained from striking altogether. The latter is the true effect of participating in a work stoppage. If one wants to find careful analytic studies contrasting the wages and benefits of workers who have adopted different strategies of industrial conflict, such studies are few and far between. One reason for our present state of ignorance is the disinterest that macrosociologists have had in workers' incomes. Income studies are generally considered to be the stuff of status attainment research. Although a handful of Marxists have explicitly considered wage determination per se, most of have avoided the topic as reeking of empty atheoretical quantification and an obsession with human capital theory, i This suspicion has been somewhat reasonable, given the tendency for many studies of income to focus on individual- or gross industry-level variables (Blau and Duncan 1967; Jencks et al. 1972; Sewell and Hauser 1975; Spaeth 1976; Bibb and Form 1977). The disinterest of macrosociologists in wages has been exacerbated by the traditional Marxist critique of economism. The pursuit of salary gains by organized labor has generally been viewed as a dangerous form of fetishism that distracts workers from making fundamental challenges to capitalist accumulation (Trotsky 1969; Mann 1973; Burawoy 1979). Wages increases are seen as trade-offs in return for hegemonic managerial control of the shop floor, the continued extraction of surplus value, and the passive acceptance of monopoly capital's domination of the state. Given the compelling quality of many of these arguments, it is easy to understand how Marxists have made a first priority of studying mass protest, class consciousness, and revolutionary politics and have downplayed the income issue. Writers in the Weberian social democratic tradition have not been 'Erik Wright (1979) and James O'Connor (1973) have been refreshing exceptions to this generality.

4

Chapter 1

much different. The primary concern of European labor corporatists and their American admirers has been the high economic growth, low unemployment, and extensive welfare-state benefits of those nations where social democratic class compromises have occurred. An explicit feature of such states has been a trade-off of wage restraint for full employment and government transfer payments (Cameron 1984; Friedland and Sanders 1986; Korpi 1978; Przeworski 1980). This has encouraged extensive scholarship on the origins of the welfare state, and limited interest in union wage gains. Money matters. The subject of workers' economic well-being is too important to relegate to secondary intellectual status. Undoubtedly, workers can improve their life chances if they create a welfare state or eliminate exploitation at the workplace; however, they can also improve their life chances if they increase their take-home pay. Improvements in salaries are far easier to obtain than fundamental transformations of political and economic systems. Workers themselves place a high value on increasing their wages. The fundamentally economistic nature of workers' aspirations has been noted by writers of both Marxist and non-Marxist perspectives; this is probably one of the few points about which both Lenin (1977) and Selig Perlman (1928) agree. Wages are the single greatest issue in industrial conflict. Studies of the United States, France, Britain, Belgium, Ireland, and the Netherlands have all shown that more strikes are fought over wage rates than over any other issue (Edwards 1981; Shorter and Tilly 1974; Cronin 1979; Walsh 1983). Wages dominated strike d e m a n d s even in syndicalist France and the France of the Popular Front, labor movements typically characterized as having class-based, political, and shopfloor concerns (Steams 1971; Galand 1937; LeFranc 1965). There is another, more subtle reason why macrosociologists have eschewed the question of tactical effectiveness. The rise of materialism in social scientific explanation has been accompanied by a reemphasis on the relation between rational choice and historical action (Hechter 1983; Coleman 1990). This emphasis on rationality has been a healthy tonic to functionalist accounts of social change that emphasized anomie, psychological maladjustment, and cultural disorganization in the face of rapid modernization (Smelser 1959; Komhauser 1959). Since resource mobilization theory, macrosociologists have successfully improved our understanding of world history by explaining social conflict in terms of reasoned responses to economic and political opportunities (Tilly 1978; Paige 1975; Skocpol 1979). These models are not compatible, however, with any data in which workers are consistently wrong or misguided in

Union Choices Affect Wages

5

the strategic decisions that they make, for then their level of participation in conflict would have an irrational component. Studies of tactical effectiveness are forced to confront worker irrationality explicitly. If two sets of workers face identical negotiating environments, and one set obtains a better outcome than the other, then the second set presumably made a poorer set of choices; their decision making was thus irrational. One can try to explain away the errors of the second workers by arguing that structural conditions were not quite identical, or by seeking some compensatory but unmeasured benefit. Contrary to the spirit of current writing on social movements, however, it may be that the workers simply messed up. How can strikers make bad tactical decisions? First, because protests are rare, strikers have limited opportunity to learn from experience. A manager of a firm obtains almost daily data on the effect of prices on sales volume and profits; he or she can experiment with raising and cutting prices and observe the effects. Workers may only experience one strike every 4 years. Given that circumstances change over time and among work sites, opportunities for true experiments in protest are nil. Second, the results of strikes are difficult to assess. Unionists need to know how the benefits they would have received had they struck differ from those they would have received had they not struck. They cannot directly observe both conditions: Either they strike or they do not; they cannot do both. Furthermore, they cannot be sure that any differences in wages or conditions before and after the strike are purely the result of the strike. Strikes frequently occur during contract expirations or changes in market conditions when wage levels would be revised anyway. The true effect of striking is not the change in wages that occurs after a strike, but the difference between the observed change in wages and what would have occurred purely from the employer's discretionary actions. Market forces might produce a wage increase or decrease regardless of striking; what the unionist needs to know is how striking changed the size of management's self-selected revision. This requires a multivariate analysis of both the economic and conflictual determinants of workers' wages, but such an analysis is simply beyond the capacity of most unionists. As such, labor spends most of its time in blind ignorance of the true effect of its strategies and tactics. Under such conditions, it is not surprising that suboptimal strategies can persist over time and that workers in identical settings can generate multiple interpretations of "ideal" union tactics. If one accepts the claim that workers' incomes are important in their own right, and that unionists can make bad decisions whether they are

6

Chapter 1

seeking to raise salaries or accomplish other goals, it then becomes important to determine how union decisions affect workers' incomes.^ Identifying the proper methodology and case material for an analysis of these decisions becomes the next task, and it is a demanding one.

FRENCH COAL MINERS AS A CRITICAL CASE The present study is an analysis of the wages of French coal miners between 1890 and 1935. This relatively obscure case was not chosen out of perversity or some private authorial fascination with Third Republic social history. Like a sea slug that receives attention from biologists for its unusually visible liver cells, French coal miners were selected because they represent a rare opportunity to provide a scientifically rigorous assessment of the effect of strike militancy on wages. Specifically, between 1890 and 1922, the coal miners maintained a unified labor movement. During this period, they followed a syndicalist strategy of holding large regional and national strikes with explicit political agendas. Their strategy strongly fits the traditional stereotype of politicized French unionism popularized by Ross and Hartman (1960) and Shorter and TUly (1974). In 1922, the miners' union split, as did nearly every other union in France. The breakaway union was communist; it adopted an explicit strategy of maintaining high levels of strike activity, demanding worker control over the shop floor and the revolutionary overthrow of the state, and using strikes as a tool for raising worker consciousness as well as obtaining immediate economistic ends. The mainstream union was reformist, espousing a social democratic strategy of using tripartite collective bargaining to achieve economistic gains peacefully. It eschewed striking, aggressively supported socialist parliamentary candidates, and traded cooperation on the shop floor for state interventions into labor negohations for the purpose of improving workers' conditions. There are no American-style business unions in this sample. There are, however, syndicalists, communists, and social democrats, three of the most important strategies of European unionism. More importantly for the Americanist, the unions varied enormously in their use of the strike tactic, notably in how often they struck, how often they won, and what they asked for. Questions of the relative efficacy of strikes and the significance of avoiding losses or concentrating on economistic demands are extremely central to understanding American strike tactics. A num^Scholars who reject the claim that economistic concerns are important could still use parallel inquiries to examine alternative utilities such as workplace control.

Union Choices Affect Wages

7

ber of more technical reasons for selecting French coal miners in this period are listed below. 1. There are high-quality wage data for French coal miners. To do an analysis of the economistic effects of class conflict, one needs good data on the main dependent variable, wages. Statistics on total wages paid are surprisingly difficult to obtain, particularly if one wants systemic series of within-industry variations. Data on wage rates per se will not suffice. Base wage rates are often weakly related to total take-home pay, because employers can manipulate bonuses, fines, surcharges, and job classifications to counteract concessions made on base rates per se (Chabert 1957). One needs the de facto (not the de jure) wage rate. One obtains this by taking the total wage bill and dividing it by the number of days worked. Although high-quality wage data are unavailable for most workers, they do exist in many periods and settings for miners. Most capitalist nations have maintained excellent statistics on business conditions in coal mining. Since the early nineteenth century, coal has been a critical asset both for economic growth and for military security. As a result, most capitalist countries have maintained detailed statistics on the economic conditions in coal mining. In the United States, these figures can be found in the U.S. Mineral Census and the U.S. Geological Survey. In France, these figures are located in the Statistiques de Vlndustrie Minerale, a collection of statistical material both on overall coal production and on every major determinant of coal production, such as total employment, productivity, the price of coal, and the number of days worked in a year. Among the statistics collected are the total salaries paid for all coal workers, broken down by year and locality, and the number of days worked per year disaggregated on a comparable basis. These data form the dependent variable for this study. 2. There are good data on the nonconflict determinants of wages. It is absolutely essential in a study of the effects of conflict on wages to purge the estimation of all nonconflict determinants of levels of remuneration. What could be worse than to claim that moderate unions were more likely to raise wages, only to discover subsequently that moderate unions were more likely to appear in firms that were productive, had high demand for labor, and sold premium products that commanded monopoly prices? At this point, it is important to introduce one of the key concepts of the book, the residual wage. I am not interested in raw wage rates, no matter how accurately measured. We are interested in wage rates net of market forces, which we define as residual wages. The true test of union effectiveness is the capacity to raise wages above the level that employ-

8

Chapter 1

ers would have paid voluntarily as a result of exogenous economic pressures. Employers rarely pay high wages out of pure altruism; they do, however, raise wages as a result of economic pressures. To determine a union's capacity to make an employer pay more than he or she would have because of market factors, one regresses wages on the most complete set of economic variables available that plausibly relate to remuneration. Residual wages are the residuals from this equation. Unless otherwise specified, every analysis that cross-tabulates wages with union tactics in this book will use residual rather than raw wages. To calculate residual wages, one needs good data on the economic determinants of levels of remuneration. This implies obtaining data on productivity, prices, unemployment, and the other variables suggested by neoclassical theory. This in turn virtually rules out doing a test of union effectiveness on national-level data. Accurately modeling the market determinants of wage rates for nation-states would be plagued by problems of changing occupational and industrial composition. There would be constant worrying about interaction effects resulting from real dangers of industry-specific market models or particularistic vulnerabilities to class conflicts, not to mention the difficulties associated with the noncomparability of different countries' economic and labor statistics. It is far more prudent to analyze a single industry with a stable occupational mix in a single nation. This way, one can analyze comparable workers with comparable classifications and a comparable currency, using data with consistent and well-understood recording conventions. Furthermore, the economic dynamics of a single industry are far easier to grasp intellectually than are the parallel dynamics for a large and diverse economy, with the former minimizing the risk of specification error in the control model. As was the case for wages, data on the economic determinants of wages at the industry level are hard to obtain, but mining again is relatively privileged in this regard. The same documentary sources that contain wage materials contain nearly every other relevant economic datum in mining as well. Material is available on productivity, volume of production, sales, prices, labor demographics, accident rates, and (for smaller samples) many other variables as well. The richness of econometric data on mining is simply not matched in any other industry. 3. There are better data on the history of industrial conflict in early coal mining than there is for contemporary nonmining industries. Obviously, in a study of the effects of strike tactics, we need data on just what those tactics were. This, fortunately, is not hard to find. National strike statistics usually provide a wealth of information on the details of overt industrial conflicts. These can usually, but not always, be supplemented by

Union Choices Affect Wages

9

union membership statistics, archival records, and secondary histories of both management and labor. For French coal mining, the primary source of data on strikes is the Statistiques des Greves, the statistical compilation of strikes used by Shorter and Tilly in their Strikes in France (1974). The present analysis has corrected these materials extensively with supplementary data from strike narratives included in the appendices to the Statistiques des Greves and secondary sources. Late nineteenth- and early twentieth-century strike data are far richer than is comparable material for the present era, containing material for individual strikes—on the identity of the employer, particular grievances, and the course of strike adjudication—that would nowadays be suppressed in the interest of protecting the confidentiality of the participants. Furthermore, France offers the unique advantage that police reporters routinely sat in on union meetings, taking detailed notes on the proceedings. This transcripts have survived, along with more customary union newspapers and managerial correspondence, allowing for a rich reconstruction of union history that would be harder to achieve in a non-French setting. 4. French coal mining is one of a small number of settings where significant within-industry variance in union strategy occurs. In order to assess the effectiveness of union tactics, there must be variance in such tactics. At the industrial level, significant differences in union strategies are very rare. In many countries, there is only one union per industry; in corporatist countries, such as Sweden or Germany, wages and conditions are largely (although not totally) determined by centralized national negotiations between one union and one employers' federation. In the United States, federal labor law has tended to favor unitary representation within bargaining units, which tends to promote single-union hegemony within companies and industries. Thus, regional or companylevel differences in union strategy tend to be reduced by the homogenizing effects of unified centralized leadership. As a result, it becomes essential to select a case where dual or multiple unionism is widespread. This is available for fleeting moments in the United States during some of the communist-noncommunist battles of the 1950s. These cases are sufficiently small and isolated, however, to defeat any attempt at large-sample multivariate analysis. There are more possibilities in western Europe, particularly in Britain, France, and Italy. 5. French coal miners are highly militant. The need for variance also dictates that the industry selected should be one with a high overall level of militancy. Many industries are not especially well organized; in these cases, the fundamental class-conflict-related differences in wages will be

10

Chapter 1

gross differences between the union and nonunion sectors. Although such effects are empirically important, the larger research question concerns the effect of variations in union activity, not the effect of having a union at all. The same logic suggests picking an industry that is relatively strike prone. A well-organized but quiet industry will have only a few strikes in its history to study. Though it would be possible to estimate the effect of striking versus not striking in such an industry, it would not be possible to analyze differences in strike strategy. If one wants to understand the consequences of striking over economistic as opposed to job control issues, or of having long rather than short strikes, a large and varied population of strikes will have to be examined. French coal mining fits these considerations perfectly. French miners, like all coal miners, were extremely militant; they have historically dominated national strike statistics with unusually frequent, large, and lengthy strikes. In France, between 1890 and 1935, the coal miners held an average of 16.8 strikes a year. This is substantially above the average for French workers in other occupational groups, particularly when one controls for the relatively small percentage of the French national labor force in mining. For every period for which Shorter and Tilly (1974) can provide data, miners rank either first or second in the population for man-days lost per 100,000 active population; their number of man-days lost are approximately triple those of the average French nonagricultural worker. Cronin (1979) and Kerr and Siegel (1954) have shown that this finding applies as well to most (although not all) industrial economies. Late Third Republic French coal mining thus represents a rare opportunity to study tactical effectiveness. The case contrasts unions with dramatically opposed conceptions of strategy; their choices— syndicalism, communism, and social democracy—represent important and much-imitated currents of Western unionism. Furthermore, all the technical data for an analysis are available, a condition that most emphatically does not apply to the overwhelming majority of settings one could consider. Necessity sometimes dictates the consideration of obscure examples; in this case, the obscure example gets at fundamental questions concerning the role of radicalism and moderation in producing benefits for the working class.

A BRIEF HISTORICAL INTRODUCTION TO FRENCH COAL MINING In order to understand the account that follows, it is useful to begin with some basic terminology and description.

Union Choices Affect Wages

11

The Major Unions Before 1922, the coal miners were represented by the Federation des Travailleurs du Sous-sol (hereafter FTSS). Although the title literally means "federation of underground workers," the union represented all workers employed in mines and quarries, whether above or below ground. In 1922, the union split, as did most of the major unions in France at the time. Communists left the mainstream organizations to start their own separate revolutionary unions. These were called unitaire unions—meaning "unitary" or "unified," an ironic choice given the circumstances of their birth. The original unions that the communists left became significantly more conservative as their leaders, who had long been of a more or less social democratic persuasion, found fewer challenges to moderate policies. The reformist unions were called confedere in reference to the General Confederation of Workers (CGT), the existing French national federation of unions. The communists formed the Confederation General des Travailleurs Unitairci (CGTU) as their national organization. Within mining, they formed the Federation des Travailleurs de Sous-sol Unitaire (FTSSU).^ Before the schism, the FTSS was very militant and strike prone. It had a successful history of holding large national general strikes to press simultaneously for wage increases and for such political concessions as pension reform or the 8-hour day. The miners' union was an important bulwark of the Socialist party; socialist parliamentarians from miners' districts were instrumental in pressing for both legislative gains for miners and benefits for the French working class as a whole (Gillet 1957; Trempe 1968; DoUeans 1953; Michel 1974). After the schism, the reformists adopted a social democratic strategy that prefigured the policies of the Scandinavian unions of the 1950s and 1960s. They forswore the strike tactic except in emergencies, and they continued to support socialist candidates, to lobby actively with state regulators of the mining industry, and to participate in tripartite collective bargaining. Their strategy was to use favorable government influence to force management to provide concessions; strikes were generally avoided in order to create goodwill with government authorities, and to giv^jilhe rare strikes that they did call particular emphasis and importance. Their policies were not unlike those advocated by Walter Korpi (1978): the transference of conflict to the political arena from the HSood general introductions to the history of the syndical schism in France can be found in Saposs (1931), Clark (1930), Dulot (1923), Garmy (1934), Kriegel (1964), Robert (1980), Labi (1964), LeFranc (1967), Martin-Saint-Leon (1923), and for the years after the schism, Amdur (1986, 1987).

12

Chapter 1

economic arena, the use of a political party as the primary representative of the working class, and the willingness to prevent overt industrial conflict from interfering with accumulation so long as workers received their fair share of the benefits produced by labor peace (Fine 1971).^ The communists, in contrast, were outspoken revolutionaries. The idea of a union seriously believing in an imminent workers' revolution in the advanced capitalist nations may seem misguided to the modern reader. This imposes a later twentieth-century ideology, however, on events of a completely different historical context. In 1922, the Russian Revolution had occurred a mere 5 years ago. Before then, no one had imagined that a workers' state could possibly exist in Russia. Because communism was ostensibly a solution to tensions within capitalism, it was presumably only a matter of time untU comparable revolutions occurred in the major capitalist economies. As a result, Marxist unions through Europe and the United States worked toward achieving such revolutions in the very near future. Unlike the reformists, the communists struck whenever possible. Strikes were viewed as occasions to raise class consciousness and to provoke confrontations with authorities, as well as to increase benefits for workers. Unitaires encouraged workers to engage in spontaneous strikes and attempted to develop local into regional strikes and regional strikes into national shutdowns. Job control and political grievances were explicitly pursued, as well as sympathy strikes both with other French miners and with workers in other countries. This is not to say that communists eschewed economic grievances. Unitaires supported high wages as a strategy for weakening the capital accumulation process. Unitaire wage demands were consistently higher than those of reformists, and they persistently decried the salaries specified in reformist collective bargains as being insufficient (Amdur 1986). The Objective Basis of Miners' Negotiating Power The miners benefited from having an extraordinary degree of class power. France was dependent on the continuous production of coal. It was able to produce only two thirds of its needs. The primary alternative suppliers were Great Britain and Germany, France's major commercial and military rivals (Oliver 1922; Lafitte-Laplace 1933; Schneider 1945; ^Undoubtedly, most of the confederes' positions were principled. Their moderation, however, is not unlike that ascribed to older and more successhil unions that have obtained comfortable, institutionalized relationships with authorities. See Lester (1958), Lens (1959), Walton and McKersie (1965), and in a political context, Michels (1962) for a discussion of union co-optation in other settings.

Union Choices Affect Wages

13

Peyerimhoff 1925). This undersupply made stockpiling of coal against industrial shortages relatively difficult, especially because coal deteriorates with outside storage. The miners' clout was further augmented by a relative labor scarcity. France has generally suffered from a shortage of manpower resulting from its persistent low fertility in the face of robust economic growth (Cross 1983; Gignoux 1942; Sauvy 1969). This was particularly the case in coal mining, where large concentrations of workers had to be attracted to remote locations far from existing centers of population (Reid 1985a; Trempe 1971; Beaubernard 1981). Unemployment was extremely low in most centers of French coal mining until the height of the Great Depression in the mid-1930s. Employment grew dramatically throughout most of the period, more than doubling between 1890 and 1930. Labor shortages were sufficiently critical that guest workers had to be imported in large numbers during the 1920s.^ The coal miners also enjoyed extensive state support.^ Throughout the entire period, the miners participated in tripartite collective bargaining involving their union, managers, and high-level state representatives. In many cases the state representatives could be expected to take the union's side; in nearly all cases, repression would be limited to the control of union violence as opposed to attacks on the union's right to exist or negotiate per se. In the event of strikes, the workers could count on the support of a significant parliamentary left, many of whom had been elected from mining districts on platforms of higher wages and benefits for coal miners (Trempe 1984; Perrot 1974; Friedman 1990). As if this were not enough to generate militancy, it must be remembered that the coal miners lived in remote towns with dense, occupationally homogeneous social networks. Like many artisanal French workers, they had a rich associational life centered around kinship, neighborhood cafes, union activities, and self-help organizations. Furthermore, company towns frequently generate radical workers. The issue is not, as Kerr and Siegel (1954) would have it, that such workers represent a mass isolated from the rest of society. The more banal explanation is that the employers of such communities tend to dominate every aspect of the economic, political, religious, and social life of the town. Tenants who are unhappy about rents or housing quality find themselves fighting the company, as do citizens who are unhappy about taxes or road repairs. Miners, customarily of an anticlerical persuasion, ^For more on ethnicity in the mines, see Chapter 8. ••The complex history of the miners' relationship with the government is covered in detail in Chapters 6-8.

14

Chapter 1

found themselves fighting the company because it funded and supported the local churches. As is generally the case in company towns where the local employer retains control over noneconomic functions, every social issue became a labor issue, and every labor issue became a struggle over individual autonomy and human liberty (Buder 1967; Chaloner 1973; Beaubernard 1981; Trempe 1971; Reid 1985a). Control over strategic assets, labor scarcity, political support, and the contentious nature of employer-dominated communities all contributed to creating a viable strike threat that miners' employers would have to take seriously at the bargaining table. The Significance of the Period The period analyzed in this book was selected for both practical and substantive reasons. Pragmatically, the years 1890-1935 are those for which the Statistiques des Greves reports detailed data on individual strikes. For coal miners, however, these dates are profoundly apt. Eighteen ninety represents the beginning of institutionalized labor relations in coal with the start of the campaign that led to the signing of the Convention of Arras, the first formal regional collective bargain ever signed in coal mining (Gillet 1957). The study ends in 1935, the year just before the explosion of the Popular Front. Nineteen thirty-six saw the unification of communist and reformist unions, the election of a communist-socialist government, and a major series of strike waves in French coal (LeFranc 1965; Hainsworth 1976). Before 1890, unions were barely recognized, received negligible government support, and wages were largely set by management fiat. After 1945, the mines were nationalized, and wage differences between establishments became a matter of proactively determined government policy. Between 1935 and 1945 were the years of the Popular Front and World War II, years of enormous idiosyncratic disruptions in coal mining. Thus, between 1890 and 1935, we get 45 years of industrial relations experience, all under one relatively consistent legal regime for adjudication of labor conflicts. The High Institutionalization of Labor Relations In late Third Republic French coal mining, we observe one of the first major corporatist systems for regulating the wages in a private sector industry. Sociologists customarily operate under the stereotyped perception that labor relations before World War II were uninstitutionalized, with little recognition for labor unions or government support for working-class causes (Snyder 1975). French coal mining belies

Union Choices Affect Wages

15

this superficial picture. The coal miners benefited from tripartite collective bargaining, the direct participation of government ministers on their behalf in wage negotiation, the presence of welfare-state legislation, the support of a labor party, and widespread union recognition in a format that strongly recalls the corporatist labor relations of Europe at midcentury. In this regard, French coal miners were absolutely atypical of other workers of their time, instead being highly representative of workers in modem capitalism. As noted earlier, these advanced labor relations were a function of the coal miners having strategic control over a resource that was central to the survival of any European government. Because the French, German, and British states were dependent on coal for both economic growth and military expansion, the governments took extreme pains to see that coal production would proceed without significant interruptions. Thus, throughout Europe, coal miners enjoyed a privileged place in labor jurisprudence that translated into early union recognition, special labor laws and welfare benefits, and unusually favorable mechanisms for the adjudication of industrial grievances (Marks 1989). Economic Trends Overall, the history of French coal mining before the Great Depression is a history of prosperity. This prosperity was periodized, however, with the years before World War I being characterized by growth and those after the war being characterized by parabolic expansion and decline. Figure 1.1 shows the trends in wages, employment, and real commodity prices for French coal mining. These figures include only the 20 departments upon which the statistical analyses are based (departments with discontinuous production histories and Alsace-Lorraine are excluded). For clarity of graphic presentation, each data point represents an average figure for a 4-year period. Using annual data does not change the substance of either the graphic or textual presentation, although the figure becomes substantially more cluttered. The figure shows slow, steady improvement in all of the economic indicators that are presented. The strong performance of coal mining before World War I was a response to the robust expansion of all the capitalist economies from 1890 to 1914. The 1870s and 1880s had been poor years for France. Pent-up demand and underinvestment produced much of the prosperity of the subsequent 25 years (Brogan 1967; Caron 1979; Mayeur and Reberioux 1984) French coal benefited as well from the rearming of France in preparation for World War I, as state investment

16

Chapter 1

Real W a g e s

Francs

2J5L5

Employment

smi&i^ 125.

Real Coal Prices

Francs

1890

1900

1910

1920

1930

Year All data points are four year averages.

Figure 1.1. Wages, employment, and prices in French coal mining

produced a modest Keynesian stimulus to heavy industry. After a miniature boom in the years from 1890 to 1892 and a period of stagnant coal prices but increasing employment from 1893 to 1897, there was a long sustained drive between 1898 and 1914 in which wages, prices, and employment grew at a dramatic rate. World War I was disastrous for the French coal mines. The war was fought on and in the mines themselves. Although coal is dispersed throughout France, the largest and most prosperous field is in the northern regions of Pas-de-Calais and Nord, the site of much of the heaviest fighting. The Germans seized most of the mines of Nord, while the battlelines ran right through some of the most important Pas-de-Calais coalfields. In many cases, whatever was not destioyed by artillery barrages was destroyed in 1918 when the Germans sabotaged as many French industrial facilities as they could before retreating. By the end of the war, nearly two thirds of the productive capacity of the northern basin had been destroyed (Gruner 1920; Georges 1921; Boulin 1927; Galand 1937; Hardy-Hemery 1984). This produced a major energy crisis in the years immediately following World War I. Coal had been generally scarce during the war, as combustibles were diverted from civilian use for military purposes. After the war, however, coal prices reached an all-time high, particularly after Britain temporarily stopped exporting coal in the interest of its own

Union Choices Affect Wages

17

reconstruction. Employment was low, but this did not reflect a lack of demand for coal as much as it did the temporary closure of much of the northern fields. The government viewed the restoration of the abused mines as essential to France's economic and military security; after a dramatic infusion of capital, the mines were rebuilt, job opportunities rose, and prices stabilized (Comite Central des Houill^res Frangaises 1921b; Oliver 1922; Peyerimhoff 1925). The prosperity that came from reconstruction was only temporary. Internationally, coal production had been steadily increasing in all of the advanced capitalist nations. New competitors, such as Poland and Czechoslovakia, were entering the world coal market at the same time that new technologies and Taylorite efficiencies were increasing the output of such traditional suppliers as Britain, Germany, and the United States. The result was substantial worldwide overproduction of coal. This problem, which would have been severe even had the conditions of the mid-1920s continued unabated, was exacerbated by two factors. The first was the technical evolution of modern societies away from coal to petroleum and electricity. Although some coal companies, notably Lens, were able to move laterally into electrical generation, the expansion of oil production and hydropower significantly reduced global demand for coal products. The second problem was the onset of the Great Depression, which produced an acute collapse in the industrial consumption of coal in both domestic and foreign markets. France entered the Great Depression relatively late. Employment and coal prices did not begin their terminal decline until 1930, and even in 1931, mining unemployment in Pas-de-Calais was less than 2%. Unemployment became catastrophic in 1932, however, and remained dismal until the era of the Popular Front (Comite Central des Houilleres Frangaises, 1925, 1930, 1935; Peyerimhoff 1931; France. Administration des Mines 1937). As Reid (1985a) has shown in his history of the mines of Decazeville, prosperity returned to a limited degree after the war, but in a larger sense the French coal mines never returned to their former well-being. Regional Differences There were overwhelming regional differences within French coal mining. There are five maps in the appendix to this chapter. Figure 1.2 identifies the 20 major coal-producing departments (provinces) in this study, along with the locale of some particularly important mines. Figure 1.3 displays the average pay of these 20 departments for the entire 18901935 period. Figures 1.4 and 1.5 provide parallel displays for total coal

18

Chapter 1

mining employment and productivity, respectively. Figure 1.6 identifies the union affiliations of each department in the postschism period. The primary contrast in Figures 1.3 through 1.5 is between Pas-deCalais and Nord as compared to all other provinces. Pas-de-Calais and Nord are adjacent departments bordering on the English Channel; they contain the single best coalfield in France. The sheer size and breadth of this northern coalfield dwarfs those of every other part of the country. These two departments were the two highest-paying departments in France, and their combined employment was greater than that of all other coal-producing departments. Pas-de-Calais was the most productive coal department in France, whereas Nord was in the top five. The northern departments, ironically, were late developers. Although the basin is contiguous to the Belgian coalfield, which was one of the first coal deposits to be developed in Europe, geologists had never located the primary veins on the French side. This was attributable to a geological anomaly that placed the French fields substantially to the north of where prevailing theories would have predicted them. As a result, throughout the eighteenth and nineteenth centuries, it was falsely believed that the Belgian field expired near the French border. Coal development therefore concentrated in the south of France, notably near the Loire (Gillet 1973; Gras 1922). After the discovery of the northern fields in the midnineteenth century, they were rapidly developed. From 1890 onward, nearly all of the employment growth in French coal mining was located in these northern fields. Older centers like Saint-Etienne in the Loire, Carmaux in Tarn, Montceau-les-Mines in Saone-et-Loire, or Decazeville in Aveyron saw very modest growth. With the exception of the coal shortage during and immediately following World War I, most of the remaining coal fields saw stagnation or absolute decline (Siegler 1909; Simard 1983; Leveque 1916). Pas-de-Calais and Nord were the center of reformist unionism in coal mining. Before the war, major moderate parliamentarians such as Basly were based in Pas-de-Calais; afterward, the northern field was the home of the confederes. Pas-de-Calais itself was almost entirely affiliated with the CGT. The Nord saw the active participation of a significant communist minority that provided a significant threat to reformist hegemony in the eastern portion of the northern field and led most of the most significant shopfloor disputes of the postwar period (Michel 1974; Trempe 1968; Hardy-Hemery 1984). Elsewhere in France, communists and reformists were somewhat randomly scattered among the smaller departments, with little correlation existing between union affiliation and prosperity. Map 1.5 shows

Union Choices Affect Wages

19

these affiliations. Of the three wealthiest remaining departments, the Loire was a communist stronghold coexisting with a weak and ineffective reformist minority. Tarn and Saone-et-Loire, by contrast, were reformist. Among the more marginal areas, Nievre, Cantal, Creuse, and Vendee were unabashedly communist, with little confeder^ influence. Aveyron was mixed. AUier was reformist, but with unitaire domination of one major center, Saint-Hilaire. The remaining areas were either reformist or ununionized. In many of these settings, communists had token presences.^ Previous Analyses of the Wages of French Coal Miners This study does not represent the first attempt to analyze statistically the wages of French coal miners. Henry Ludwell Moore in 1911 did a, by the standards of the time, pioneering study in empirical economics by demonstrating that French coal mining wages were related to productivity and prices. Frangois Simiand (1904) and his student M. Bourgeaud (1938) attempted to demonstrate that coal miners reduced their efforts after receiving pay raises, ultimately destroying the economic capacity of employers to pay union wage rates. As was generally the case in the early twentieth century, the econometrics of these analyses are tortured and crude. Side analyses I have performed using modern time-series cross-section techniques support the Moore argument that productivity raises wages, but not the Simiand claim that pay raises reduce future productivity. Because Simiand's position has been well debunked by American labor economists studying the union wage effect, a rigorous presentation on this issue has been not attempted here (Parsley 1980). Moore's position has been incorporated into the core economic models of the analyses. The other major discussion of French coal mining wages is Roland Trempe's voluminous treatment in Les Mineurs de Carmaux (1971). Although Trempe discussed the economics of Tarn coal mining in lavish detail, her wage discussion was psychological rather than econometric. She discusses with piercing sensitivity the meanings of wages for both workers and managers, and the multiple logics used by both sides to assess what levels of remuneration would be fair. For scholars interested in the formation of worker consciousness, no one in any discipline has more carefully formulated how economistic disputes become the vehicle for the creation of class identities. ^See Chapter 3 for a discussion of data sources and methods used in attributing departmental union affiliations.

20

Chapter 1

For the narrow econometrician, though, there is less here on the precise estimation of market and union effects on wages. Trempe is more than justified in her neglect here, because her case is so obvious. As will be seen from the recapitulation of Trempe's account of Carmaux in Chapter 5, after one decisive victory in 1892, Tarn miners' wages went sky-high and remained high for many years thereafter. In the Carmaux context, the union effect on wages is beyond dispute, and dickering over the exact magnitude of the raise is pedantic. In the context of all of France in the late Third Republic, however, arguing about the effects of different kinds of unionism is not at all moot. The size of the increase that can be attributed to Carmaux syndicalism is well worth comparing with the results of northern reformism or Loire communism. The present analysis borrows from Tremp^ by incorporating many concepts from her psychological discussions as variables in the core statistical models. Some of her items can be conveniently dropped, particularly when they can be reduced to an invariant managerial pressure to reduce wages or a generalized union pressure to raise them.

OVERVIEW OF THE BOOK The overall plan of the remainder of this book is as follows. Chapter 2 lays out the core theoretical argument that drives the rest of the book. A strong case is made for frequent striking as an optimal strategy for raising wages. This is because frequent striking gives a union a reputation for militancy, which in turn intimidates employers into making relatively generous settlements. An exception occurs when the labor movement is divided into radical and moderate factions; in these cases, militancy is often punished by a shifting of benefits toward more strikeaverse workers. Chapter 3 is the main statistical analysis that supports these claims. After a review of the empirical materials, the primary econometric models of French coal wages are presented. Chapter 4 examines a potential challenge to these findings. It may be argued by critics of Chapter 2 that the number of strikes is less important than the success of those strikes. In contrast, it is shown that winning strikes does not raise wages, and that in some cases losses can be more beneficial than victories. This is because strike results are as much a product of the radicalness of demands as of the relative strength of workers. By participating in kamikaze strikes on issues of high employer resistance, workers generate reputations for militancy. Chapter 5 shows that unions with broad political and workplace

Union Choices Affect Wages

21

agendas can achieve pay benefits at least as large as those who limit themselves to economistic demands. As will be shown, the effect of a union's choice of demands is limited, because workers rarely get to choose between alternative benefits. They either win everything or lose everything. Nevertheless, participating in political or workplace strikes facilitates creating a reputation for militancy. It is also shown that wages are not a product of compensating differentials for hazard or other shopfloor disamenities. The succeeding chapters attempt to ground the largely econometric analyses of the first half of the book into actual historical context. Sociologists will prefer the first half of the book; historians will prefer the second. Chapter 6 examines the general relationship between the unions, employers, and the state. It is argued that, perversely, French coal operators had little incentive to reduce strikes but great incentives to reduce wages. The state operated under precisely the opposite constraints. As a result, French coal employers only acted in the "normal" fashion to reduce strikes because of pressures placed on them from the government. Chapter 7 illustrates the materials in Chapters 2 and 3 by documenting accommodation and intimidation in the era in which the French labor movement was unified. Narrative accounts of the great prewar strikes are used to show how militancy actually produced wage benefits for workers. Similar processes are shown to have been operating in unitaire strongholds in the postschism era. Chapter 8 is the counterpart of Chapter 7, demonstrating the steering of benefits from radicals to moderates in settings where the labor force was divided. After a consideration of prewar schisms, where wage steering did not occur, a narrative account of the union schism is developed that shows how policies of rewarding passive workers developed over time. Chapter 9 contains a set of legitimate challenges to the conclusions of Chapters 2 and 3. A set of factors are identified that limit the benefits of policies of high militancy for workers. Readers uncomfortable with the conflictual tone of the rest of the book are advised to turn to this chapter for an analysis of the darker side of aggressive striking. Chapter 10 is a general conclusion and an extension of the French coal mining materials to a discussion of contemporary American and European labor movements.

22

Chapter 1

APPENDIX: MAPS OF FRENCH COAL-PRODUCING REGIONS

1 Pas de~L 1 Calais V L~ ^^^ ^

^ ^ ^, ^ - > . Nord ^

"l /(Moselle)

/

/Saone- ) / V Haute ^ / - \ i r V - »

c^/

Savoie\

Is6re

t.

SairnV"^

J L Loire- j^ / ^ Cantal i. Haute /

/ ^

V

j>

Allier T j

Creuse \ v\_,-o V \ r \ Loire}

/ rs^

^

rS"—v

J _/^ Montceau-1^ rt/tJ'es deteriorated, leading to major underreportage of industrial conflict (Robert 1980). 43

44

Chapter 3

of measurement error in the smaller coal-producing departments and in Alsace-Lorraine that would have seriously distorted multivariate analysis. 2 The analysis involves bituminous and anthracite coal exclusively, as opposed to lignite or bituminous schists. All coal workers are included in the sample, both skilled and unskilled, underground and surface. Dependent Variable Wages are measured as the total number of real francs (in hundreds) paid to labor in the department per workday. They include all monetary payments to workers, including bonuses, overtime, and other supplements. The fact that the data concern total payments to workers and not published wage rates is important. The calculation of workers' pay by employers was extremely complex. Employers set wages using piece rates based on such considerations as the weight or volume of coal produced, or the length of advancement of drilling or wood support. This base rate would be multiplied by an arbitrary factor (referred to as the "prime") negotiated in collective bargaining to represent changes in the cost of living and the demand for labor. The negotiated base rate would be then adjusted by the estimated difficulty of the mine site, or the age or occupational grade of the miner. A complex list of deductions could be made from this gross figure, either in the form of payments for supplies or penalties for malfeasance. Furthermore, there were fringe benefits, including but not limited to pension contributions, accident relief contributions, and free heating coal (Mangez 1925; Peyerimhoff 1931). As a result, although actual wage payments usually correlated with ^The data anomalies in the minor departments tend to take the form of rounding errors that can seriously skew estimates of small magnitudes. Arbitrarily changing rounding decisions can shift the residuals of small departments from extremely high to extremely low levels, producing wild swings in statistical results. There also are implausible discontinuities that would seem to be the result of clerical omissions. Alsace-Lorraine shows consistently high residuals. Despite the pleasure one may take in speculating on possible class-conflict-related causes of such residuals, the real basis of these wages was likely to have been more mundane. These factors include the high wages paid German miners before World War 1. Such wages would have been difficult to cut given the political pressure on the state to secure the loyalty of the former German miners to the new French regime. Other idiosyncrasies include a high percentage of foreign workers residing in Germany and a concomitant linking of wages to both the German labor market and the exchange rate between the franc and the mark. It is nearly impossible statistically to separate these local considerations from the more general variables of the all-France model. Therefore, with some regrets, both the part-time producers and Alsace-Lorraine were excluded.

Strikes and Wages; The Statistical Relationship

45

published wage rates, it was possible for an employer to adjust wages without adjusting wage rates by shifting workers among occupational titles, or by changing the schedule of supply costs and penalties. Using actual wage payments captures employers' non-rate-related adjustments of their wage bills and measures the actual labor costs experienced by management. The denominator of the wage variable is the total number of days worked. This measure is in contrast to the more usual alternative of measuring wage rates as total wages divided by the number of workers. Mining is subject to large random fluctuations in days worked per year, resulting from technical difficulties such as fires, floods, ventilation breakdowns, and shaft collapses. Such natural interruptions of the work cycle can be expected to have adverse effects on total wage attainments. In addition, the number of days worked per year will be lower during years with high levels of strike activity. It would be inappropriate in a statistical estimate of the effects of strikes on wages to include an artifactual negative correlation between these two variables. Measuring wage attainments by the total wage bill divided by the total number of days worked eliminates these problems.^ Economic Control Variables The analysis uses the following economic control variables: employment, productivity, coal prices, general price levels, national coal wages, the percentage of nonagricultural employment of the department, and year. Employment is included as a measure of the tightness of the labor market; it is measured in the conventional fashion as the number of coal miners employed in the department. Employment should positively correlate with wages.* Neoclassical economic theory predicts that wages are positively correlated with the productivity of workers (Hicks 1963). Productivity is defined here as the weight of coal produced divided by the number of workdays in the year. The workday denominator was chosen once again to make the measure independent of strike activity. Productivity also provides a partial control for variance in worker skill and corporate investment. Coal prices are included as a second measure of productivity. In ^Estimates of the amount of wages lost during strike days themselves can be found in Chapter 9. *It would have been desirable to have unemployment data as a measure of general labor market tightness, but these were unfortunately unavailable.

46

Chapter 3

neoclassical theory, productivity is measured in tern\s of the value, rather than the bulk, of workers' output. Workers who produce coal that can fetch a high market price should be more valuable than workers who produce cheap coal.^ Overall price levels are included to reflect the adverse effect on real wages of inflation in the price of consumer goods. Assuming fixed money wages, real wages will be negatively correlated with the price of general commodities. National coal wages are included as a measure of prevailing wage standards that employers would have had to meet in order to avoid losing their workers to out-migration. The wages of the local department are excluded from this national average. The percentage of employment that is nonagricultural is included as another measure of wage pressure from alternative employers. In general, urban wages tend to be higher than rural wages, and manufacturing wages tend to be higher than those offered by agricultural employers. The percentage of nonagricultural employment is an index of the prevailing wage outside of coal mining. Mine operators in heavily industrialized districts presumably would have had to pay a higher wage to prevent defections out of mining into such semiskilled occupations as factory work or construction. Time is included as a control to eliminate any unspecified effects of trend. It may be considered in part, however, as a market variable reflecting the general upward trend in wage rates. Since the Industrial Revolution, there has been a consistent tendency for real wages to rise as a result both of productivity increases and of the demand for labor outstripping population growth (Phelps-Brown and Browne 1968). Controlling for trend is necessary to eliminate statistically those wage increases that would have been the product of growing wage competition from French noncoal employers. Wages, employment, productivity, and coal prices are all taken from the Statistitjues de I'Industrie Minerale (1890-1935); general prices are from the Annuaire Statistique de la France (1966). The percentage of nonagricultural employment comes from the French census. Data for the years between decennial observations have been estimated through hnear interpolation. The analysis does not include measures of inflation or of exchange ^It would have been desirable as well to have data on profitability or capital investment to tease out the relationships among worker effort, technology, employer prosperity, and wages. Unfortunately, neither of these types of data were available for the period of the study.

Strikes and Wages: The Statistical Relationship

47

rates between francs and foreign currencies. Although institutional lags could conceivably cause wages to be relatively low in periods of inflation and high in periods of deflation, supplementary statistical analyses found no such effect. Exchange rates were excluded by design; including such measures might capture the acute crises in wage setting that were produced by the collapse of the franc in the mid-1920s (Gignoux 1942; Sauvy 1969). These effects, however, should be captured in the overall price indices, which do show the expected increases during the years of currency crisis. Union Membership The analysis also includes an index of union membership. The measure, unfortunately, is very crude. French union membership data are of substantially lower quality than those available for other advanced capitalist nations; they are arguably the worst in the developed world. The poor quality of French union data is the wholly understandable product of the legal and organizational basis of French unionism. In France, there are very few advantages to joining a union formally. Under French law, nonunion members receive all of the bargaining privileges and protections from employer sanctions that are available to union workers. At the same time, the terms of most union contracts are freely applied to both union and nonunion workers. At the same time, the unions themselves provide few organizational incentives for membership. Strike funds are generally low, and selective incentives such as union benefit programs are usually not available (Lorwin 1954; Saposs 1931; Friedman 1990). As a result, many workers who are wholly sympathetic to the union cause choose to free-ride and not pay their dues. This result is frequently exacerbated by union leaders who choose to agitate by pressing for spontaneous strike mobilizations rather than for union membership per se (Gallie 1983). As a result, it is not uncommon for the turnout rate for union strikes to substantially exceed union membership figures. In France, far more than in other nations, union membership statistics underestimate levels of union support. If one adds to this bias the substantial random errors that result from union bureaucracies' marginal record keeping on what leadership perceives as a peripheral matter, the result is very low-quality statistical data. Frost (1964) has analyzed the shortcomings of French union membership data for the early twentieth century in extensive detail; most of his conclusions are extremely pessimistic. Therefore, this study makes no attempt to make any fine-grained estimates of department-year changes in the density of union member-

48

Chapter 3

ship. The departments are coded in a crude dichotomous fashion as to whether or not they were union or nonunion. A department was viewed as being unionized if any of the following conditions applied: • At least 5% of the miners in the department were officially registered by local mining unions • A strike occurred in the department • Any of these conditions had existed in the department at any previous date during the period of this study This is a remarkably inclusive definition of union status. Reasonable objections to it can be made by any informed sociologist or historian. This all-encompassing operationalization was used as a conservative measure to create a bias against producing positive findings for strike militancy. The primary purpose of the statistical analysis is not to assess the impact of unionization on wages, but to assess the effects of strike strategy. Unionization is a control variable. As such, to avoid overestimating the effect of strikes on wages, and by doing so, artifactually supporting the theory in Chapter 2, every attempt was made to shift unspecified class-conflict effects to the union density variable rather than to the strike measures themselves. Following this logic, all strikes are assumed to take place under some form of union auspice and are coded as cases of a union being present. This gives worker militancy the properties of a Guttman scale— one can be unionized without striking, but one cannot strike without being unionized. There are multiple occurrences within the data of strikes in areas for which no record exists of union membership. My personal hunch is that the majority of these cases represent situations where workers had high levels of informal organization and associational solidarity, but eschewed affiliating with one of the Federations des Travailleurs du Soussol. Many of these strikes may have been associated with the presence of local leadership, the use of ad hoc strike committees, or the pooling of common financial resources, even if they went unrecorded by the Annuaire des Syndicats. By coding these settings as unionized, part of the benefits of these strikes is attributed to the presence of the shadow organization, and the return from the actual stoppage itself is minimized. The same logic is used to justify why once a department is considered to be unionized, it is never viewed as deunionizing. Informal organization often allowed workers to maintain some level of strike militancy even in the face of seeming demobilization. The best example of this is the period between 1894 and 1898. In 1893, the FTSS lost a major region-

Strikes and Wages: The Statistical Relationship

49

wide strike in the northern basin after attempts to expand the struggle into a national general strike had failed. After the defeat, by most accounts, popular support for the miner's union withered. In the Loire, the miner's federation became wholly inactive (Faure 1956; AN F 13791). In the north, the union did not vanish altogether; however, participation in organized agitation and formal meetings was extremely low (Pas-deCalais 93 M 221). Nevertheless, between 1894 and 1898, there were 14 strikes in the northern basin, where union activity was restricted, and 6 strikes in the Loire, where the union had dissolved. Most of these incidents were single-firm stoppages, where local militants could maintain continued pressure on management even without support from the departmental federation. Treating departments as always being unionized once they have achieved the initial criteria for being organized captures this mobilizing capacity of local leaders. It also avoids conflating the effects of militancy with those of plant-level organization. Union membership data for the prewar period come from the Annuaire des Syndicats for the years 1898, 1901-3, 1905, 1909, 1911, and 1912. Union membership data for the postwar period come from the annual reports of the Federation des Travailleurs du Sous-sol and the Federation des Travailleurs du Sous-sol Unitiare. In both cases, the index of raw membership was the number of annual subscription cards sold each year. Confedere figures are only available for 1931; unitaire figures are available for every year from 1923 to 1928 (Federation Nationale des Travailleurs du Sous-sol 1931; Federation Nationale Unitaire des Travailleurs du Sous-sol 1923-1928). Obviously there is a large amount of missing data, which is a further reason for not attempting detailed estimates of within-department changes in union membership. Union membership figures were divided by total employment to generate rough estimates of union density. This was then combined with strike data, and departments were then dichotomized as either union or nonunion using the rules discussed previously. Strike Frequency Two strike variables are included in the main conflict models: the average number of strikes per year, and the average number of pay strikes per year. All of these data are calculated from the tapes of Shorter and Tilly's Strikes in France (1974) and refer exclusively to coal mining disputes. Shorter and Tilly's data set is largely derived from the statistical compilations of strikes in the Statistiques des Greves. This source frequently provides detailed qualitative accounts of major strikes in appen-

50

Chapter 3

dices to the annual volumes. In many cases, the information from the statistical and narrative sections are inconsistent. I have assumed that the historical accounts, which are frequently very detailed and well researched, are generally more accurate than the statistical tables, which are subject to random clerical error. I have examined all of the supplementary narratives for coal strikes in this period and have corrected the statistical data to incorporate this new information. In many cases, the implied adjustments are substantial. For each department-year, the total number of strikes was calculated. Multiregional strikes were broken down into their component single-department elements. Strikes in which multiple firms went out were combined as one strike if they began on the same day or within 1 day of the inception of another strike. If the strikes began on noncontiguous days, they were treated as separate incidents. The Statistiques des Greves also provides a listing of all of the demands made during strikes. These demand data wiU be analyzed in some detail in later chapters; for this discussion, strikes were coded as being over pay issues or nonpay issues. Pay issues included salary offensives, salary defensives, and disputes over surcharges, late pay, and free coal. If a strike contained both pay and nonpay components, it was viewed as a pay strike. Mixed strikes were coded as pay strikes to provide a conservative bias against generating any effects for nonpay strikes. If nonpay strikes proved to be an important predictor of wages, one would need to show that this was not attributable to pay demands slipping into the nonpay sector. Pay strikes were analyzed separately to test the possibility that strikes over pay may have a more direct impact on wages than the total population of strikes (which contains nonpay incidents). Note that in the theoretical discussion of Chapter 2, the mechanism by which strikes generated favorable economic results for workers was by creating a reputation of militancy for the union that intimidated employers into making generous settlements. Presumably, strikes over workplace or political issues could have as much impact on a reputation for aggressiveness as pay strikes. Therefore, it was not predicted that pay strikes would be more highly correlated with strikes than total strikes. Preliminary analyses showed that the behavior of both strike variables was not linear. Not surprisingly, moving from zero strikes to one strike made substantially more difference in wages than moving from seven strikes to eight strikes. After various exploratory analyses, it was discovered that the best fit for strike variables could be obtained by recoding the strike variables into a five-category measure: zero strikes, one strike, two or three strikes, four or five strikes, and more than five

Strikes and Wages: The Statistical Relationship

51

strikes. This receding capture the intuitively appealing idea that the move to any kind of militancy from previous passivity would make a substantial difference, whereas once a large number of strikes have occurred, the marginal return of further striking is low.^ Statistical Method All of the equations reported in the subsequent analyses are Parks GLS estimates, rather than ordinary least squares (OLS) regressions. The Parks models are used because the data set takes the form of a pooled time-series cross section.'' There are 20 departments, measured over 40 years. In pooled time-series cross-sectional data, there are two violations of the assumption of the independence of errors, a critical assumption for OLS. The first is that there may be some unspecified variable correlated with region or department. If we failed to take into account some factor that led to low pay in the Vendee, for example, some 40 observations would be mismeasured, rather than just 1. The Parks estimation calculates the degree to which errors from the model are correlated with particular regions and consolidates these errors into one term. There are also likely to be errors associated with time. In traditional error-components or Wallace and Hussein models, time is treated in a fashion analogous to region, in which each date is viewed as an independent source of error. Such a specification is not especially realistic for wage data. An error-components model would take out error terms for 1890, 1891, and 1935, but would assume that there was no tendency for 1891 to be more similar to 1890 than to 1935. If there was an unspecified variable associated with "early" as opposed to "late" years, the assumption of the independence of the year errors would be violated. The Parks model makes the more realistic assumption that timebased errors take the form of autocorrelation. Wages are thus predicted from what they were in the previous period. This accurately captures the stickiness associated with wages; wage rates are often inflexible because of the presence of legal contracts and informal customs that make rapid changes in rates hard to negotiate. The importance of institutional factors of maintaining the long-term stability of wage rates is well known (Hicks 1963; Dunlop 1966; Marsden 1986). If, for some rea*The five-category recode also fits the data substantially better than do logarithmic or quadratic transformations. Tor a general discussion of the statistical issues associated with time-series cross-section data, see Kmenta (1971) or Parks (1967).

52

Chapter 3

son, wages in Gard were raised to an idiosyncratically high level in 1900, they are unlikely to return to base levels in 1901. The 1900 rate represents a precedent. Although Gard wages may return over time to a "normal" level, they will do so in small increments, with each step being somewhat in line with past practice. Autocorrelation nicely captures the problem of custom and contracts in wage rates. Parks models can be calculated either assuming a constant level of autocorrelation for the population or allowing the autocorrelation coefficient to change for each region. The present study uses departmentspecific rho's. These are generally quite high; the department-specific autocorrelations in the models presented here are usually above .75 and are frequently above .90. Parks models provide an adequate correction for this problem.8 The coefficients for these models can be interpreted in the same way as OLS regression results, with the one limitation that no statistic is provided for goodness of fit. Although various goodness-bf-fit statistics exist for GLS models, most suffer from such deficiencies as not fitting into a 0-1 metric, not being comparable between models, or not decreasing when variables are dropped from models (Greene 1990).

FINDINGS Table 3.1 shows the primary statistical findings for the entire 18901935 period. Model I shows the pure effect of the market variables on wages. All of the effects are highly significant and in the predicted direction. Note especially that coal prices are positively correlated with wages, and that overall prices are negatively correlated with wages. The data support the theoretical predictions of different effects of two superficially very similar variables. The first equation demonstrates the importance of market factors in determining wages; any attempt to assess the economistic impact of class conflict without taking these variables into account would have led to systematic biases and misspecifications. Model II shows the same equation with the addition of union density. Union density is significantly related to wages. The inclusion of union density does not change or reduce the coefficients for the market variables. This finding replicates the well-known relationship between ^In preliminary analyses, traditional error-components models were fitted to these data. The results were similar to those presented here, but the models reported time-based error components that were low. Given the high level of autocorrelation that actually exists in the data, this was attributed to the low capacity of WaOace-Hussein models to diagnose serially correlated errors.

Strikes and Wages: The Statistical Relationship

53

Table 3.1 Parks Estimates of the Effects of Market and Conflict Variables on the Wages of French Coal Miners, 1890-1935 I Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present Total strikes Pay strikes

1.20 14.7 4.15 6.81 6.65 10.1 -1.62 8.45 1.77 17.1 7.92 28.2 7.23 2.68

11 1.15 14.2 3.63 5.40 6.38 9.83 -1.51 8.44 1.76 15.6 8.18 15.6 6.23 2.48 5.35 2.73

III 1.16 14.3 3.59 5.32 6.37 9.82 -1.52 8.48 1.74 15.3 8.12 29.9 6.72 2.66 5.06 2.57 0.49® 1.45

IV 1.13 13.9 3.59 5.33 6.32 9.59 -1.50 8.30 1.75 15.6 8.17 30.0 6.31 2.47 5.38 2.73

0.10® 0.24

®Insignificant at .10; all other coefficients significant at .01. Note; Top figures are Parks coefficients; bottom figures are f-statistics. N = 800. 1914-1918 and 1931 excluded because of missing data. For ease of presentation, the coefficients for the following variables have been multiplied by the following constants: national coal wages, 10; productivity, general prices, union present, and strike variables, 100; coal prices, percentage nonagricultural, and year, 100; employment, 100,000. See text for sources.

the presence of unions and high wages; it also replicates the equally well-known finding that the zero-order relationship between unionism and wages is not a spurious result of market factors (Parsley 1980; Freeman and Medoff 1984). Model III includes all of the variables of Model II with the addition of the total number of strikes per department-year. The performance of this variable is wholly unimpressive; the magnitude of the strike coefficient is less than half the magnitude of the next weakest predictor in the model. Furthermore, the coefficient is not statistically significant at the .10 level. This weak performance is not the function of the inappropriate inclusion of a set of nonpay strikes in the measure. Model IV substitutes total pay strikes for total strikes. The pay strike

54

Chapter 3

coefficient is one fifth the size of the total coefficient. Although the total strike coefficient is significant at .2 (which implies an 80% chance of being unequal to zero), the pay strike coefficient is absolutely insignificant at any reasonable level. Furthermore, the inclusion of either strike variable leaves the magnitude of the control coefficients unchanged. If the analysis stopped here, the conclusion would be inescapable: Strikes are not related to wages. The pattern of findings in Table 3.1 is consistent with a Hicksian model of strikes. Indicators of the long-term bargaining strength of workers, such as market factors and union density, correlate with wages. Strikes per se, however, do not. This suggests that managers set their wages based on a pragmatic estimate of workers' overall negotiating strength; whether workers strike or do not strike is immaterial, because managers estimate the power of the union the same way regardless of whether a strike occurs. The problem with such reasoning is that the equations in Table 3.1 are misspecified, however, and the conclusions that come from this table are misleading. The equations fail to take into account the profound changes in strategy that are required when there is a schism in the labor movement. Because the French labor movement was unified before 1923 and divided afterward, there should be strong period effects that represent the different strategies required for different historical contexts. Isaacs and Griffin (1989) have argued cogently that long time series can severely misspecify causal relationships when they conflate social processes in different periods. The Isaacs and Griffin warning applies with great force to labor relations in French coal mining. Table 3.2 shows the effects of strikes on wages separately for early and late periods. The left equations divide the years before World War I (1890-1914) from the years after the war (1919-1935), and the right equations divide the years of a unified labor movement (1890-1922) from the years of the formal schism (1923-1935). Both pairs of equations tell very similar stories. In the early years, the effects of strikes on wages are strongly and significantly positive; in the later years, the effects of strikes on wages are strongly and consistently negative. There is a complete reversal of causal direction between the earlier and the later periods. These effects are substantively strong and are significant at the .01 level. Note that the effects of the other control variables remains the same, with the exception of employment during the postschism period. These findings are fully consistent with the promilitancy position. In periods of a unified labor movement, there were consistent, strong returns to striking. In periods where the labor movement was divided, strikes hurt the labor force, and management systematically rewarded quiet workers.

Strikes and Wages: The Statistical Relationship

55

Table 3.2 Parks Estimates of the Effects of Market and Conflict Variables on the Wages of French Coal Miners, 1890-1935, by Period Prewar- -postwar 1890-1913 Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present Total strikes N

1.32 24.0 3.63 10.3 53.6 21.7 -22.3 8.45 1.92 30.9 2.74 5.47 20.0 9.84 6.59 9.15 9.32 4.89 480

1919-1935 2.11 43.6 2.11 13.4 6.12 23.0 -1.49 13.2 1.45 21.1 8.35 49.0 -1.82 5.59 37.5 6.47 -3.30 2.73 320

Unity--schism 1890-1922

1923-1935

1.66 24.7 5.49 6.63 9.06 11.8 -1.32 6.11 1.46 21.5 7.88 38.6 6.29 4.53 2.57 2.68 13.8 5.32 560

1.70 12.0 -1.50® 0.76 3.78 7.42 -0.69 4.62 2.54 15.4 9.44 43.3 -2.58 6.87 5.51 3.40 -9.86 2.53 240

Note: Sources and conventions as in Table 3.1.

The theoretical discussions of Chapter 2 argued that the critical issue determining the appropriateness of striking was the unity of the labor force. The period effects reported here, however, could hypothetically be the result of many factors. The lower returns to strikes in the 1920s and 1930s are possibly attributable to the rise of communist unionism after the Russian Revolution. It is possible that "red" unions were highly ineffective strike tacticians, producing lower returns to strikes in the later period. It could also be that the governments of the 1920s and 1930s were consistently more hostile to labor than those of the prewar period. Gallie (1983) has argued quite correctly that the General Railway strike of 1920 produced an immediate and lasting repressive response from the French state from which the country's labor movement never recovered. Moreover, it could be that the change in strike effectiveness was a product of the war itself; possibly the German damage to the northern mines brought on a wave of conservative patriotism

56

Chapter 3

among miners and mine operators alike. The number of other plausible period effects that can be suggested is endless. One indicator that the schism itself and not some other time-based effect is responsible for the change in the effectiveness of striking is that the difference in the strike coefficients before and after the schism is more marked than the difference before and after World War I. The size of the period differential can be calculated by subtracting the late coefficient from the coefficient for the early period. The difference in coefficients for the prewar-postwar contrast is 12.6, whereas the difference in coefficients for the unity-schism contrast is 23.8. The schism differential is nearly twice the size of the prewar-postwar differential. If either government conservatism or the presence of communism or war was the issue, then the war-based differentials should have outperformed the schism. There is thus something about union divisions themselves that produces unusually strong differences in the viability of striking. Table 3.3 illustrates these arguments in another way. Table 3.3 presents models with all of the variables of either Model III or Model IV of Table 3.1, along with an interaction between strikes and period. The left pair of equations analyzes all strikes; the right pair of equations analyzes pay strikes only. Within each pair, the left equation periodizes strikes using the war, and the right equation periodizes strikes using the schism. The key finding is that for each dependent variable, only schism provides a statistically significant period interation. The best-performing equation in the tables is the second from the left, which uses total strikes and a split between unity and schism years to predict wages. Overall, total strikes had a strong effect, as predicted by the general model. In schism years, however, there was a statistically significant interaction that made the overall effect of striking negative. This equation implies that in unity years, the effect of striking is 9.5; in divided years, the effect is - 5 . 7 , for a difference for 14.2. If one analyzes total strikes but periodizes by the war, the effect of striking overall is weakly positive, but the effect of postwar years is statistically insignificant. No pay strike equation provides a positive overall effect of strikes over wages. The interaction effect for schism, however, is nearly twice as big as the interaction for war. The schism interaction almost approaches statistical significance. Overall, Table 3.3 shows that a schism periodization produces greater effects on the direction of the strike coefficient than a war-based periodization. It also shows that pay strikes do not predict wages terribly well, but total strikes do. This is entirely consistent with strikes communicating a position of militancy to management, with the content

Strikes and Wages: The Statistical Relationship

57

Table 3.3 Parks Estimates of Period Interactions of Conflict Variables on the Wages of French Coal Miners, 1890-1935 Total strikes

Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present Strikes Strikes x period (Early = 0; late = 1)

Pay strikes only

Prewarpostwar

Unityschism

Prewarpostwar

Unityschism

1.15 14.1 3.57 5.56 6.35 9.77 -1.49 8.17 1.75 15.3 8.21 29.0 6.23 2.45 4.66 2.36

1.17 14.4 3.53 5.20 6.40 9.86 -1.48 8.28 1.72 15.0 8.12 29.2 7.16 2.79 4.54 2.30 9.48 2.29 -14.2 2.24

1.12 13.8 3.59 5.51 6.33 9.54 -1.48 8.15 1.77 15.5 8.25 29.0 5.89 2.28 5.11 2.60 3.08® 0.52 -7.22® 0.86

1.14 14.2 3.54 5.19 6.36 9.63 -1.48 8.23 1.75 14.2 8.17 29.6 6.64 2.58 5.13 2.60 4.51® 0.97 -14.7E] 1.75

7. son 1.77 -6.84® 1.06

® = Insignificant at .10. El = Significance between .05 and .10. Note: Coefficients (except as noted) significant at .025. N = 800. Strikes and strike interactions in same metric; both are multiplied by 1,000 for clarity. Sources and conventions as in Table 3.1 except as noted above.

of demands being less important than a posture of radicalism, and with union divisions gutting the utility of aggressiveness by giving management a steering option. The theoretical discussion gave a very specific account of a steering mechanism by which employers shift benefits from strike-prone workers to passive workers. The equations thus far have only shown that during the years of union division, the returns to strikes were negative. This is a strong finding, but there are accounts one can give of the dynamics of schism that do not involve steering. It may be that the union division simply gutted workers' bargaining strength overall; strikes may have been negatively correlated with wages because miners

58

Chapter 3

simply no longer had the organizational strength to win benefits with strikes. If the schism lowered the returns to striking by crippling the organizational strength of workers as a whole, then one would expect that after the schism, residual wages would have declined. If the schism lowered the returns to striking by setting u p a steering mechanism, then residual wages after the schism should remain at the same levels as were found during unity. The schism argument holds that there is a transfer of benefits from the militant to the peaceful. The aggregate level of wages may easily remain the same; all that changes is their distribution. Table 3.4 regresses the control variables, strikes and period on wages. The schism is used to delineate the two periods. The term for schism years is miniscule and insignificant. This is a remarkable finding considering that there at least two artifactual considerations in the data that should have biased this estimation toward finding a negative period effect. First, the schism years include the Great Depression. It is entirely Table 3.4 Parks Estimates of the Effects of Period, Market, and Conflict Variables on the Wages of French Coal Miners, 1890-1935 Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present

Strikes Postschism

1.15 13.9 3.59 5.32 6.34 9.77 -1.53 8.32 1.73 29.4 8.11 29.4 6.71 2.65 5.17 2.36 4.76® 1.39 0.10® 0.31

® = Insignificant at .10; all other coefficients significant at .01. Note: N = 800. Period coefficient multiplied by 10 for clarity. Sources and conventions as in Table 3.1.

Strikes and Wages: The Statistical Relationship

59

Table 3.5 Residuals from Core Economic Model by Union Affiliation; 1923-1935 Affiliation

Mean residual

Nonunion Communist Reformist Mixed

-.253(36) -.030(84) .086 (96) .192 (24)

Contrast

Significance

Union vs. nonunion Reformist vs. communist Reformist vs. mixed Communist vs. mixed

.001 .026 .093 .002

Note: Residuals from Model I in Table 3.1. Significance levels calculated with simple f-tests.

reasonable that the trend effect would be reduced in a period with an economic crisis in the last four years. Second, time is included in these models as a control. As a result, to find a negative effect, wages did not actually have to decline over the period, but only to rise at a lower rate than they had during previous years. Despite both of these biases, no negative period effect was uncovered. This conservative test provides a strong confirmation of the argument that dual unionism per se did not gut the wage attainments of French miners as a whole. Dual unionism did, however, redistribute earnings among workers. Reformists earned more than communists; furthermore, areas in which communists and reformists were both active received more than any other. Table 3.5 shows the average residual wages for workers in nonunion, reformist, communist, and joint departments. Union affiliations were calculated from membership data given in the union annual reports. Unitaire departments were defined as any department in which at least 5% of the employed miners belonged to the unitaires and the majority of card-holding unionists were communists. Confedere departments were defined as those in which reformists made up at least 5% of the employed miners and the unitaires did not meet this condition. Mixed departments, thus, were those in which both the unitaires and the confederes had 5% of the workforce. In all of the mixed cases, confederes outnumbered the unitaires, making the communists a significant threat, but not the dominant union of the region. Any department

60

Chapter 3

in which neither organization had registered 5% of the workers was coded as nonunion. Note that Table 3.5 and the tables that follow report residual wages, not raw wages. Table 3.5 reports residuals from Equation I in Table 3.1, where wages are predicted from market variables only. I use an equation without union density because union status is one of the analytical variables in the table. The rest of the analyses in the chapter involve union workers only; these use residuals from Model II, which includes union density. These last analyses, because of their smaller N, use crosstabulations rather than Parks equations. Because only the years from 1923 to 1935 are used, there are only 240 cases, as opposed to 800 in the larger equations. Given the large number of variables already in the analysis and the detailed breakdowns required for demonstrating steering, equation specifications of these relationships run into problems of data attrition. The tables give a full accounting of the interrelationships of these variables (along with a graphic illustration of what conditions have small Ns and are intrinsically less reliable). The cost of using the tabular presentation is that these results may be misspecified if the strike variables have a causal impact on the control variables of Equation I. The equations in Tables 3.1 through 3.4 show that the coefficients for the control variables remain virtually unchanged when class-conflict predictors are added to the model. This suggests that errors resulting from joint causality among analytic and control variables are quite small, and that the estimates of the effects of class-conflict variables given in the tables will be fairly accurate. Table 3.5 shows that reformists earned substantially more than communists. This is precisely what would be expected under a steering model. Both groups earned more than nonunionists, suggesting that having an ineffective union was better than having no union at all. The group that earned the most was that of workers in mixed departments. In pure reformist departments, there was little chance that workers would defect to the communists, but in mixed departments, that threat was very real. Acceding to the demands of the primary (reformist) union was thus a strategy for building the credibility of the moderates and weakening the communist threat.^ Mixed departments therefore obtained residual 'Note that this would not apply in a department with a primary communist union facing a reformist threat. Even if employers gave the reformist minority concessions, overall wages should be low because of the steering of benefits away from the main unitaire union.

strikes and Wages: The Statistical Relationship

61

Table 3.6 Mean Residuals from Core Economic Model by Strike Activity by Union Affiliation; 1923-1935 Affiliation of department

No strikes One strike or more Significance

Reformist

Mixed

Communist

.119 (59) .034 (37) .187

.408(7) .104 (17) .009

-.010(46) .055 (38) .057

Affiliation of strike

No strikes One strike or more Significance

Reformist

Joint

Communist

.008 (206) -.010(34) .770

.006 (230) -.104 (10) .754

- . 0 0 2 (179) .027 (61) .538

Note: Sources and conventions as in Table 3.5.

wages that were more than twice as large as those obtamed by unthreatened reformists, and larger still than any other group. All of the differentials are statistically significant at the .10 level, and most are significant at .03 or lower. Table 3.6 shows the relative effectiveness of strikes, broken down by affiliation. Affiliation is measured in two ways. The top half of the table shows wages by strike activity within departments with different affiliations. The bottom half shows wages by strike activity by the actual affiliation of the union holding the strike. Thus, if communists struck in a mixed region, the strike would appear in the middle column of the top half and the right column of the bottom half of the table. Both halves show similar findings: Wages were lowered both by reformist strikes and by strikes in reformist areas. Because both types of workers received premiums for not engaging in strikes, militancy cost workers their peace bonus. Wages were also lowered both by strikes in mixed areas and by strikes jointly held by both unions. Managers punished reformists and neutrals who acted like communists. In contrast, wages were increased by communist strikes and by strikes in communist regions. In these areas, no steering was possible; faced with a militant union that could not be avoided, managers paid bonuses to co-opt the union. Many of these relationships are statistically insignificant. This, however, is because the regional tabulations include strikes by unions of

62

Chapter 3

Table 3.7 Mean Residuals from Core Economic Model by Strike Activity by Department and Strike Affiliation Simultaneously, 1923-1935

Reformist Departments Only No strikes One strike or more Significance Mixed Departments Only No strikes One strike or more Significance Communist Departments Only No strikes One strike or more Significance

Reformist strikes

Joint strikes

Communist strikes

.104 (82) -.021 (14) .079

.090 (93) -.024(3) .001

.105 (80) -.010 (16) .188

.181 (19) .238 (5) .723

.259 (20) -.138(4) .000

.356 (10) .076 (14) .011

-.043(75) .079 (9) .295

-.037(81) .160 (3) .342

-.075 (54) .052 (30) .098

Note: Sources and conventions as in Table 3.5.

different affiliations, whereas the affiliation tabulations include strikes in regions with different union compositions. Table 3.7 illustrates the arguments suggested by Table 3.6 and eliminates the heterogeneity problems by showing the mean residual wages attained by striking and passive workers, broken down simultaneously by both affiliation of region and affiliation of strike. Table 3.7 shows that in reformist departments, strikes of every affiliation were punished. Nonstrikers of every persuasion had mean residuals close to .10; strikers of every persuasion had mean residuals of -.02. The differences vary in their statistical significance, with the effect of joint strikes being highly statistically significant and the effect of communist strikes being insignificant. Overall, however, the detail supports the conclusion drawn from Table 3.6 that in reformist regions, any form of striking produced the withdrawal of a peace premium. In mixed departments, it was a matter of some indifference whether reformists struck on their own or desisted from striking. Joint strikes and communist strikes, though, were punished severely. The salaries in mixed departments in which reformists struck were slightly higher than those in which they desisted, although the differences are not significant. The residuals were universally low when the communists struck, however, and lower still when the reformists joined the communists in

Strikes and Wages: The Statistical Relationship

63

striking. The highest residuals were when there are neither joint strikes nor communist strikes. These differences are all statistically significant at the .01 level. Clearly, management was taking every pain to provide no conceivable reward to either communist strikes or cooperating with communists. Reformists could strike alone without penalty in the mixed areas, because management had to provide conspicuous rewards to reformists. They could not crush the reformist strikes as cavalierly as they could in purely reformist areas for fear of driving workers to the communist camp. Reformist strikes did not always provide benefits in mixed areas, because reformists often received their premium pay peacefully, but they were able to avoid the sanctions which were applied to militant communists. The situation was completely different in the communist areas. Strikes of every persuasion increased wages, but only communist strikes produced differences that were statistically significant. In no case was any type of worker punished for going on strike. The joint strikes and the reformist strikes did not produce significant effects, because confed^re actions were relevant to only a small statistical minority of the workers. (Note that communist strikes produced an insignificant penalty in reformist areas for similar reasons.) Because the communists were numerically predominant, their strikes raised wages. In the unitaire strongholds, there was no possibility of creating a lasting labor peace by making workers disaffiliate from the communists. Radical unionism was thus an unalterable fact of life. Steering was not viable within the department. Communists who did not strike could be perceived as tractable, reasonable, or suffering from organizational problems; these workers would not need to be bought off with concessions. Communists who struck frequently in the face of adverse conditions, however, could not be placated easily. Their wages reflected the greater payoffs required for strike avoidance. Note that a worker would have been better off as a reformist worker executing the reformist strategy than as a communist worker executing the communist strategy. Workers in reformist departments that experienced no reformist strikes had mean residual wages of .10, whereas those in communist departments that experienced communist strikes had mean residual wages of only .05. Furthermore, as Table 3.5 showed, reformists consistently outearned communists overall; the benefits of peace were better than the benefits of war. The union schism clearly created a situation in which the optimal strategy was to avoid strikes. When there was a single radical union, as occurred both before the

64

Chapter 3

schism and in unitaire departments after the schism, the best strategy was to strike as often as possible. Schism, however, destroyed the returns to this best local strategy. CONCLUSION The statistical analysis has shown the following: 1. 2. 3. 4.

Neoclassical variables predict French coal wages well. There was an orthodox union wage effect. Overall, there was no relationship between strikes and wages. This is because there were two completely opposite effects. In the preschism period, strikes raised wages; in the postschism period, they lowered wages. 5. Schism rather than the effects of government repression, communism, or war is likely to be the best explanation of the changing effect, because the slope of the strike-wage relationships shift more dramatically around 1923 than around 1919. 6. In the postschism period, mixed regions outearned reformists who outearned communists. 7. In communist regions, everybody benefited by striking. In reformist regions, everybody suffered by striking. In mixed regions, communist strikes produced significant wage losses.

All of these findings are consistent with the hypothesis of striking maximizing workers' income, so long as dual unionism does not activate a managerial strategy of steering. The discussion thus far has primarily inferred managerial intimidation and explicit strategies of steering from wage outcomes. These claims would be more plausible if the observed patterns of residuals correlated with explicitly formulated managerial actions and strategies. The next chapter provides further support for these arguments by providing archival evidence on the determination of worker wages and the formulation of employer strategies of labor control.

4 Empty Victories, Harmless Defeats The Relationship between Moderation, Strike Results, and Substantive Benefits

The discussion of strike strategy that has been presented thus far has been highly idiosyncratic. This is because there has been no consideration of the simplest strike strategy of them all. The "obvious" prediction that a naive observer might make would be as follows: Workers who win their strikes will make more than workers who lose their strikes; unions should engage in strikes they expect to win and avoid strikes they expect to lose.

If this argument were true, then strike volume per se would be relatively irrelevant, because what would matter would be the total number of victories relative to the total number of defeats. The argument that winners outearn losers appeals to common sense. It is invoked both implicitly and explicitly in the literature on strike results, in which the effects of strikes are operationalized as to whether the workers obtained satisfaction for their demands in the negotiated strike settlement (Friedman 1988; Edwards 1981; Ragin, Coverman, and Hayward 1982). The relation between strike victories and income has never been empirically examined. This is because a positive finding would be so obvious that the results would be unpublishable if the claim were to be confirmed. The standard wisdom, though, is probably wrong: Workers who win their strikes do not make more money than those who lose their strikes. There are good reasons why this should be so. The traditional argument would hold if strike victories were determined solely by the relative balance of strength between workers and managers. In this case, 65

66

Chapter 4

strong workers would win their strikes and on the basis of their strength, enforce more advantageous settlements on their employers. Variations in worker strength could produce a strong relationship between strike victories and substantive benefits. Strike victories, however, are not determined exclusively by the relative strength of the two sides. There is a second determinant of strike results—the relative moderation of strike demands. Hicks argued quite cogently that strike behavior is determined by the costs not only of striking, but of conceding the benefits that are in dispute. All other things being equal, management should be more likely to grant a strike demand that is relatively inexpensive than one whose costs are so great as to be prohibitive. By this logic, a set of workers could run up an impressive string of victories by striking only over trivial issues. Management would grant every demand; however, it is not clear that this would produce an increase in wages. In Chapter 2, it was argued that workers obtained economic benefits from class conflict by obtaining a reputation for militancy. Employers who are fearful of aggressive, resolute, uncompromising radicals are likely to make preemptive concessions to buy off future participation in class conflict. Confining one's demands to moderate requests that are palatable to the employer is likely to undercut one's reputation for militancy, and communicate a posture of conflict avoidance and conservatism. Employers could use such behavior to infer a high subjective cost of striking on the part of the union and lower the value of their preemptive payments accordingly. Even if this Hicksian portrait of managerial strategy is incorrect, there are compositional reasons for arguing that moderation is inversely correlated with workers' wages. Modest demands lower the returns that workers obtain from any particular victory. Walton and McKersie (1965) argue quite compellingly that within any negotiation, there is a maximum threshold of concessions that management is willing to make. Normally, this maximum is obscured intentionally for strategic reasons, in the hope that the union will overestimate the level of resistance it faces and make concessions that are cheaper than the threshold. A radical union is likely to make demands above the managerial threshold; this threshold will then be revealed as the union retreats from its position in hard-fought microsteps. A moderate union might make an initial offer on the unfavorable side of the threshold, or leapfrog the threshold in the process of eliminating its most "unwinnable" demands. Either of these actions will produce suboptimal settlements for workers. The argument here is not that there is a robust negative relationship between winning strikes and residual wages (although such a relation-

Empty Victories, Harmless Defeats

67

ship does appear in the present findings). The argument is that there are two factors affecting the correlation between strike victories and substantive benefits, strength and moderation. These forces work in opposite directions and produce effects that cancel each other out. Generating a history of victories may reflect a union's strong organizational base, or it may reflect a tendency to strike only over issues perceived as being winnable. A set of failed strikes may reflect declining labor scarcity or a deterioration in worker solidarity; however, it can also reflect a set of earnest struggles over ambitious issues by implacable and relentless shop-floor militants. Workers who continue to strike regardless of past losses prove they cannot be deterred by the results of one big battle. The employer of such workers should anticipate extensive conflict costs in the foreseeable future and make his or her bargaining decisions in this light. The data on French coal mining includes extensive material on workers' strike demands. Unlike many datasets that provide grievance material in somewhat useless, preceded classifications, the Statistiques des Greves restates the strike demands in literal form, allowing the investigator to recode them into theoretically meaningful categories. The listings also contain the outcome of each grievance independently, providing information on which demands were granted by management and which were refused. When the set of strike issues defies simple summary, the Statisticfues provide two- to three-page supplemental histories of each strike, allowing reconstruction of the complexities concerning each incident. The grievance data presented here make considerable use of these supplements. With these materials, it is possible to assess the relationship between moderation, worker strength, and strike victory far more rigorously than is possible for other workers. It is not possible to assess directly the effect of moderation on wages. However, one can estimate the effect of strike victories on residual wages. In these data, moderate demands are clearly associated with the likelihood of strike victory, whereas strike success is inversely related to wages. Both relationships are robust to the inclusion of an extensive battery of controls.

OPERATIONALIZING THE MODERATION OF DEMANDS There are two general techniques for measuring moderation, each of which allows for several variations. One strategy, the static approach, considers what types of grievances are nearly always viewed as central to managerial interests. Although employers in particular conjunctures

68

Chapter 4

might vary in how radical they perceive a given type of demand to be, there are nevertheless statements that can be made about employer resistance that are relatively constant across bargaining contexts. For instance, few managers will graciously accede to a request that they sacrifice their jobs. Another alternative is a dynamic approach that conceptualizes radicalism as the degree to which a demand differs substantially from those working conditions that have previously existed. Proposed changes that vary dramatically from present conditions in the workers' favor should be relatively likely to fail; those changes that represent either modest improvements or changes in favor of management should be more likely to succeed. The present study uses measures derived from both strategies. Static Measures There are at least plausible universalistic theories of managerial resistance to grievances: cost theory, strong managerial control theory, and weak managerial control theory. These can be summarized briefly as follows: 1. Cost theory. Employers, according to this model, should be more likely to resist demands that are relatively expensive to concede. Most neoclassical theories of strike behavior invoke this type of reasoning, with useful theoretical results (Hicks 1963; Farber 1978; Tracy 1986). In such arguments, all strike demands are reduced to the dollar cost to the employer of granting the concession, regardless of content. Employer resistance is then predicted by the difference between the cost of making a concession and the cost of continuing with the strike. 2. Strong managerial control. Some authors have argued that employers are especially reluctant to grant concessions that would reduce managerial autonomy in the workplace. According to this theory, managers are particularly likely to resist any incursions on their exclusive control over the labor process. Although managerial preoccupahon with control has been explained in psychological terms, it also reflects a reasonable fear that in the absence of such control, future technological decisions might be based on the vested interests of workers rather than on profitability. A concern with managerial control intensifies early resistance to collective bargaining in any form. In those settings where the union presence is unavoidable, managers then choose to grant relatively generous monetary concessions in return for maintaining complete autonomy in shop-floor matters. The greater difficulty of winning laborprocess-oriented demands has been argued both by human relations

Empty Victories, Harmless Defeats

69

theorists concerned with developing "constructive" industrial relations and by Marxist critics of bourgeois trade unionism. The claim that trade unions are economistic carries with it the implication that trade unions favor monetary concessions in part because they are encouraged to do this by the actions of their capitalist counterparts. If unions can win wage improvements but not job control, it is because employers are particularly unwilling to sacrifice control over the labor process (Herding 1972; Aronowitz 1973; Hill 1981). i 3. Weak managerial control. It is possible to construct a weaker version of this hypothesis. Managers may be willing to make some forms of joboriented concessions, but might be unwilling to grant concessions that threaten their personal prerogatives or careers. Most managers would be hostile to worker demands for the firing of supervisors or directors, even if they were open to union collaboration in technological issues. A strong version of this theory, therefore, would emphasize the relative unwinnability of shop-floor issues; the weak version would only argue that personal threats to managerial interests are resisted. Dynamic Measures An alternative conceptualization of the moderation of a demand is the relationship of that demand to the concessions that have already been granted by management. For any given level of worker strength, there is a relatively predictable level of benefits that can be obtained by labor; in a sense, this can be considered as a "stock" of potential concessions. Any individual strike is over whether management will add to the level of benefits that are already being offered. Thus, each particular strike demand represents a flow of further concessions. The odds of being granted a new concession depend on whether workers have already attained their full stock of potential concessions. Workers who are underrewarded net of their bargaining strength should be more likely to win their strikes, whereas those who are overrewarded net of their bargaining strength should be relatively likely to lose. The spontaneous raising of wages serves in many ways as a preliminary concession in future strike negotiations, because it represents an overt admission by the employer that the labor force has certain strategic advantages that 'In the present context, it is of particular interest to note that French employers are hypothesized as being extremely protective of their managerial prerogatives and authority, and thus as doubly unwilling to make concessions on workplace-oriented issues (Lorwin 1954; Sturmthal 1972). For an alternative treatment that characterizes managerial preoccupation with job control as being distinctly American, see Edwards (1981).

70

Chapter 4

warrant immediate economic attention. This argument holds with even greater force if these higher wages came as a resuh of a previous strike. The preemptive concession model thus predicts that strike success should be negatively correlated both with the past wages that workers have received and with any recent changes in their wage rates. Workers with high levels of past wages have received preemptive concessions in the past, and those who have experienced a recent wage improvement have just received a preemptive concession. Workers who have not received preemptive concessions are more likely to have employers who have underestimated their strength; thus, they are likely to win their strikes. Workers who have received preemptive concessions are likely to have employers who have made full adjustments for potential labor threats (and are thus unlikely to have the capacity to provide further benefits).

ISSUE-SPECIFIC SUCCESS RATES One strategy for testing the relationship between moderation and strike success is to see if moderation predicts the particular grievances that will be granted by management within an overall package of concessions. Many strikes involve more than one grievance. If workers make both an expensive and an inexpensive demand in a strike, a cost theory would predict that the former would be denied and the latter granted. Looking at individual grievances allows for the consideration of variations within strikes in the outcome demands; these variations are lost in analyses that focus exclusively on the outcome of entire strikes. The data allowed for the identification of 15 different types of strike issues. These were as follows: 1. Salary offensives (proactive demands for higher pay) 2. Salary defenses (reactive campaigns against managerial attempts to cut salaries) 3. Surcharges (disputes over bonuses, fines, and all other additions or deductions from workers' paychecks) 4. Heat (disputes over the amount of free coal workers could use for home heating) 5. Pension/sick benefits 6. Late pay/pay timing (disputes over nonpayment of salary or the length of time between normal paydays) 7. Income equalization (demands for job assignment or pay calcu-

Empty Victories, Harmless Defeats

8. 9. 10.

11. 12.

13. 14. 15.

71

lation procedures that would divide income more equally among workers) Work assignment/rules Safety Comfort/courtesy (requests for minor amenities at the workplace, such as clean drinking water, or courteous treatment from supervisors) Work hours (reduction of hours only) Punishment/layoffs (protests over punishments, suspensions, dismissals, and layoffs attributable to slack demand—in practice, disciplinary and economically oriented terminations are almost impossible to distinguish; both elements are present in many cases) Fire workers/ethnic exclusion Fire supervisors or management Union procedural (protests over union organizing rights and rights of representation)

These grievances were coded in terms of four criteria: cost, threat to managerial job control, threat to managerial job security, and latent bargaining strength. The last item is a control variable designed to capture compositional issue-specific correlations with access to concrete strategic resources. 1. Cost. The cost theory predicts that the following concessions should be relatively easy to win: comfort, safety, pay timing, heat, work assignments and rules, equality of income, requests to fire workers, requests to fire supervisors, and requests for union security. About salary, surcharges, and hours grievances no prediction can be made. The rationale for these predictions is that grievances requiring no immediate monetary expenditure should be relatively likely to gain employer acceptance. This would include comfort and courtesy requests, pay timing and format disputes, and demands concerning union privileges. In the French coal case, disputes over work assignments and work rules also tended to be over relatively low-cost issues. Comfort and courtesy grievances requests were usually for relatively inexpensive items, such as a supply of fresh water for lunch or the right to take a more convenient elevator to the work site. Granting requests for courteous treatment from supervisors was free; these usually required nothing more than a guarantee of future good intentions on the part of the employer. Disputes over the length of the pay interval or the day on which

72

Chapter 4

miners would be paid were easy to resolve as well. The final wage bill would remain the same under any system; all that was at issue was the size of particular constituent parts.^ The same argument applies to income equalization, because the dispute concerns the distribution of pay among workers rather than the total amount of wages paid in the aggregate. ^ Union security issues were relatively cost free in France; this is in substantial contrast to the United States, where these issues carried enormous implications for production expenses. In the United States, union recognition involves certification by the National Labor Relations Board. This provides workers with legal protection that they would have not otherwise enjoyed and involves a long-term permanent commitment to a union role in collective bargaining and wage setting. In France, there was no official certification or recognition process. The recognition of a union in one particular negotiation implied no legal or procedural commitment by management to involve the union in any subsequent negotiations. Furthermore, union workers did not enjoy any special legal privileges denied to nonunion workers. Under the law of 1892, both union and nonunion workers could request government conciliation and arbitration, and in neither case could the government force either bargaining or concessions on the part of management (Sturmthal 1972; Saposs 1931; Shorter and Tilly 1974). Thus, in the United States union recognition meant a permanent augmentation of worker's bargaining power, whereas in France it involved no sacrifice of managerial negotiating strength whatsoever. * Demands concerning the labor process tended to concern low-cost issues. The workplace issues in the data tend to be relatively minor; the following incidents are typical of the population of incidents as a whole. ^Occasionally a cost element would enter pay timing disputes if there was a serious absenteeism problem on the day after payday. Management often conceded more frequent paydays on the condition that the workers give a guarantee that attendance and effort would not diminish as a result. ^It could be argued that payment by group rather than individual performance lowered management's ability to punish nonperformers. This argument would hold in most forms of employment, particularly in office or factory work; however, it is less applicable to mining. Individual mining productivity depends on the ever-changing geological properties of the shaft more than on personal effort. A face that is easy to work in the morning can turn hopelessly intractable by the afternoon. Income equalization disputes are best seen as requests by workers for insurance against random environmental setbacks rather than as struggles over incentive systems. ••Union recognition became somewhat more expensive after the Matignon agreement in 1936, when French law incorporated a modest bargaining role for unions per se (Lorwin 1954).

Empty Victories, Harmless Defeats

73

At Buxieres-les-Mines (AUier) in 1907, miners struck to have the carts lubricated every 15 days. At Saint-Etienne (the Loire), workers asked to be rotated between gassy and nongassy sites. In some cases, the issues were nonzero-sum. In Isere in 1908, for example, workers wanted the wood piles for mine supports moved closer to the mining face, a request that was granted immediately. At Saint-Etienne, the workers wanted to be able to petition their supervisor rather than the company engineer for the right to be absent, a reform that undoubtedly would have saved hours of upper management's time. Some obvious exceptions have to be made to this generalization. In the late 1920s and early 1930s, the last 7 years of our 45-year period, job control strikes did reflect significant issues. During this time, Taylorism was introduced in force to the mines. Longwall operation was substituted for shortwall extraction; the practice of chronometry was extensive; and workers' pay became increasingly tied to the Bedaux system. Throughout the mines, there was a significant speedup that was associated with major reengineerings of the labor process. Although mining engineers had been intellectually aware of scientific management since the immediate postwar period (Laligant 1919; Langrogne 1919), it was only in the late 1920s and 1930s that Taylorism was implemented in any systematic or large-scale fashion (Viaud 1929; Barbier 1929; Audibert 1933; Hardy-Hemery 1970; Reid 1985a; Moutet 1986). This systematic immiseration of the labor process produced the customary response among workers—both passive and active resistance to the reorganization of work. The primary resistance to Taylorization occurred after the period of this study, during the explosive nationwide strikes of the 1936 Popular Front (Hainsworth 1976). During the early 1930s, though, there was a nontrivial number of smaller disputes, primarily led by communists, aimed at stalling the drive toward Taylorism and providing workers with protection both from the speedup and from firing (Hardy-Hemery 1984). Many of these disputes do not appear in the statistical records as work rule disputes. This is because the strikes began after workers were fired for violation of new work rules. The Statistiques des Greves lists these, therefore, as punishment or layoff protests. Although there were many disputes over punishments and job security, there were relatively few explicit demands by workers to institute alternative work procedures, or to roll back Taylorism per se. This does not mean, however, that workers passively accepted scientific management. The Popular Front uprisings and worker complaints during the earlier period show eloquently that workers profoundly despised what was happerung to their jobs. But because these conflicts were fought over managerial pun-

74

Chapter 4

ishments rather than elimination of the Bedaux system per se, the population of strikes that involved explicit demands over work rules continued to be dominated by relatively peripheral aspects of the labor process. Weaker exceptions were the battles over the control of company towns in such locations as Carmaux or Montceau-les-Mines at the end of the nineteenth century. Here, though, the battles were over not the physical organization of how work was done, but workers' right to support socialists, the 8-hour day, pension reform, and wages. All of these grievances appear in their own proper categories. Once again, although these issues involve worker control in a deep sense, the superficial category of disputes over work rules is left with a more trivial set of low-cost issues. Demands to fire either workers or superiors should be relatively inexpensive. In periods of low demand, when the mines are overstaffed, fulfilling such requests can be compatible with managerial plans to reduce overhead. When labor is scarce, the cost of replacing and retaining workers is more substantial. Nevertheless, most mining knowledge is general, rather than specific to individual firms. Turnover rates in French coal mines were extremely high, and miners routinely circulated between one mine and the next (Trempe 1971). This suggests that retraining costs were low, because the labor market at any given time consisted of a large number of pretrained workers who considered themselves readily employable at a large number of different workplaces. Because the number of workers involved in these demands was frequently low, the actual cost involved in firing workers and hiring substitutes is likely to have been minor. ^ Pay disputes of all kinds (salary offensives, defensives, surcharge disputes, and heat) and work hour disputes are difficult to measure in terms of cost. They all represented nontrivial expenditures, but the size ^Firing a supervisor is an extreme case of this phenomenon. Usually the workers wanted only one supervisor fired, most often either a lower-level supervisor responsible for an accident or the director of the company in the case of unpopular labor actions. Although one can imagine many reasons for resisting such requests, short-term cost is likely to have been a negligible factor. One exception to this principle would have been when miners demanded that an ethnic group that represented a large percentage of the labor force be fired. Discriminatory demands would be easy to grant when they involved victimizing individuals or ethnic groups that were only present in token numbers. Exclusionary demands are particularly likely to occur when minority members are introduced to a worksite, it being relatively easy to exclude the first Italian or Pole and thus set a formal precedent. Although the data on the ethnic disputes do not list the number of workers at risk of exclusion, it is quite likely, given the high probability of first-exposure grievances, that the number of workers involved was usually small.

Empty Victories, Harmless Defeats

75

of these expenditures varied with the size of the increases being sought. Heat was probably the least expensive of all of these, because the total amount of free coal that a worker received was a trivial percentage of total payments. Among the most expensive grievances were those involving layoffs and pensions. In both cases, management was being asked to subsidize unproductive labor. In the case of punishment and employment disputes, it was being asked to pay for workers that it had already defined as being dispensable; in the case of pensions and sick pay, it was paying workers who were producing nothing. Further, the cost of pensions and sick pay could be extremely large in an industry characterized by enormous fatality and accident rates. 2. Strong managerial control. This theory predicts that monetary demands will be easier to obtain than those concerning the labor process. The coding of issues by this criterion is obvious. 3. Weak managerial control. This theory merely claims that managers will oppose attempts to fire supervisory and managerial personnel. 4. Latent bargaining strength. Not all differences in the winnability of grievances are a function of variations in moderation. Some grievances should be associated with defeat not because of any particular objectionability to managers but because they tend to occur when workers' relative bargaining power is low. The best examples of these are defensive salary strikes and strikes over layoffs. Employers cut wages and dismiss workers when the demand for labor is low; under such conditions, the ability of workers to influence management by further withholding labor is questionable. Offensive campaigns are timed to the advantage of the union and tend to occur when management is seeking to attract rather than repel labor. Thus, one would expect offensive campaigns to be much more successful. There are two other factors that need to be considered. First, heat campaigns should be won relatively often. Heat is more likely to be requested in the winter, when workers' heating needs are greatest. This is also the period of peak seasonal demand for the coal mines, in which employment is likely to be highest and the pressure for production is likely to be the greatest. This correlation could be intensified by climatic factors; the most severe winters are likely to produce both additional heat strikes and greater coal sales. Strikes over delayed or tardy paydays are also likely to be easily won. In this case, the latent bargaining advantage is legal: Employers have a legal obligation to pay their workers for hours that have already been worked. Employers who pay late can, if necessary, be called into court and held liable to the traditional sanctions available to creditors. In

Chapter 4

76

Table 4.1 Theoretical Predictions of the Difficulty of Winning Different Types of Grievances

Grievance Salary offensive Salary defensive Surcharges Heat Pension/sick benefits Late pay/pay timing Income equalization Work assignments/rules Safety Comfort/courtesy Work hours Punishment/layoffs Fire worker/ethnic exclusion Rre supervisors Union procedural

Cost

Strong managerial control

E H E E E E E H E E E

E E E E E E H H H H H H H H H

Weak managerial control

Latent bargaining strength E H E E

H

H

Note; E = easy; H = hard; blank == no prediction.

the data, because of the small number of both types of incidents, latepay strikes have been merged with strikes over the length of the pay period, an issue not subject to favorable legal intervention. Nevertheless, the strength of the workers' position on late-pay strikes should give the combined category a favorable average. A summary of the predictions of the various theories can be found in Table 4.1. Table 4.2 shows the failure rates of the particular grievances within French coal strikes. All compromises are treated as successes.* The easy, intermediate, and hard categorizations are merely numerical descriptions of the observed failure rates. The rates clustered naturally •"A full discussion of this methodological decision would go beyond the limits of this chapter. The present treatment is analogous to that used by Snyder and Kelly (1976). In general, the factors that determine whether a nonlosing strike is a compromise or a victory are artifactual and bear little relation to the relative strength of the two parties. For instance, wage issues are highly likely to be compromised, because wages are measured continuously and thus allow for demand reductions; work assignment disputes, which are rarely compromised, are discrete entities for which intermediate options are less apparent.

Empty Victories, Harmless Defeats

77

into three roughly equal groups, with relatively small differences within groups and relatively large differences between groups. In general, the predictions of the cost theory, the weak managerial control theory, and the latent bargaining theory are all well supported, whereas the strong managerial control theory is disconfirmed. The strong managerial control theory predicted that managers would be more likely to resist demands involving interventions in the labor process than those that merely involved money. In fact, many interventions in the labor process were relatively easy to attain. Work assignment and work rule concessions failed less than a third of the time; income equalization demands, which involved shifting job assignments among workers, were also relatively easy to attain. Comfort and union security were the two easiest demands for coal miners to win; the high win rate for union security is inconsistent with the image of the manager militantly defending personal control of his firm. Further, most monetary demands were either moderately difficult or hard to win. On balance, monetary concessions were harder to obtain than were labor process Table 4.2 Failure Rates of Grievances by Type of Demand, 1890-1935 Failure rate Difficulty and grievance type

Percentage

N

Easy Comfort/courtesy Union procedural Late pay/pay timing Heat Income equalization Work assignments/rules

19.4 23.1 23.9 26.7 31.6 32.6

36 13 46 30 19 43

Intermediate Salary offensive Safety Surcharges Work hours Fire workers/ethnic

40.7 41.2 44.0 47.2 47.8

307 17 134 163 23

Hard Salary defensive Fire supervisors Punishment/layoffs Pension/sick benefits

57,4 62.4 64.4 74.2

61 56 197 66

78

Chapter 4

concessions. If trade unions sought economistic rather than labor process demands, they may have been facilitating the extraction of surplus value, but no one can accuse them of choosing easier rather than harder objectives. The other theoretical perspectives are well supported. Those grievances associated with latent bargaining strength tended to behave as predicted. Salary defenses were more likely to fail than were salary offensives: Defensive campaigns failed 57% of the time, whereas offensive campaigns failed only 41% of the time. Heat and pay timing were both relatively easy concessions to win; rescinding punishments and layoffs was relatively difficult. The net contribution of bargaining strength is less easy to assess, because cost theory made similar predictions on heat, pay timing, and punishment/layoffs. Only bargaining strength, however, can explain the inferior performance of salary defensives. The one prediction made by a weak managerial control theory— that managers will resist their own firing—is borne out. Requests to fire supervisors failed nearly two thirds of the time, a result that is only explainable by some form of managerial control theory.'' Cost theory succeeded with most of its predictions. The theory predicted, that comfort, union security, pay timing, heat, income equalization, and work assignments and rules would be easy grievances to win. They were, in fact, the six grievances in the sample that were least likely to fail, and they represent the entire cluster of grievances that were labeled as easy. Cost theory predicted that pensions and punishment/layoffs would be difficult to win, and those are the two most failure-prone demands in the data. The theory also argued that salary offensives, salary defensives, surcharges, and work hour disputes would be of indeterminate difficulty, because managerial resistance would depend on the actual size of the concessions being asked. Three out of these four grievances were of moderate difficulty; salary defensives were relatively difficult, although they were the easiest to win of the difficult grievances. The three remaining cost-related predictions were all disconfirmed, although only one of them was overwhelmingly awry. Safety, requests to fire workers and requests to fire supervisors were all predicted to be That demands to fire supervisors succeeded 37.6% of the time is actually quite remarkable. In many cases, the concession obtained was a compromise; the supervisor in question would be transferred to a set of workers who found him more acceptable. In other cases, the supervisor was already in trouble with management. Engineers whose incompetence produced mine collapses could expect problems with both the workers and the board of directors.

Empty Victories, Harmless Defeats

79

easy grievances.® Safety and worker firings were each of moderate difficulty, however, and demands to fire supervisors were the third most failure-prone grievances in the data. The difficulty of firing supervisors stems from managerial self-interest rather than the operation of cost considerations per se. Overall, out of 15 grievances, 11 were predicted correctly and only 1 was overwhelmingly incorrect, which can be considered a generally strong performance. Most of these findings can be explained only in terms of cost. The heat, late payment, and punishment effects could be the spurious product of latent bargaining strength, but the other successful predictions were unaffected by this theory and represent relatively pure effects. Grievances do seem to differ in their likelihood of being successful, making the moderation of demands an important factor to consider in studying the outcome of strikes.

GLOBAL STRIKE OUTCOMES The previous analysis only included static measures of moderation and crude measures of structural strength. A more complete analysis would also include dynamic measures of moderation and more explicit measures of power differentials. This can be done if one uses the results of whole strikes, rather than individual grievances, as the unit of analysis. All of the previous variables except those derived from the strong managerial control perspective are included in this analysis. This provides a test of whether the findings that were reported for individual grievances were the spurious results of other grievance- or strengthrelated factors. Each strike was coded for whether a grievance of a given type was present among the strike issues. Thus, the coding easypresent/hard-absent was applied to strikes with both single and multiple easy demands. Likewise, it was possible for a strike to be coded as easy-present/hard-present if demands of conflicting types were involved. Grievances involving latent bargaining strength, latent bargain^Those who would argue that work-rule and assignment demands are costly to employers because they involve fundamental issues concerning incentive systems can find some support for their position. Among the easy grievances, work-rule and assignment issues were relatively difficult to obtain, putting them in an intermediate position between truly easy demands and salary offensives. This location is consistent with the earlier discussion that argued that job control demands are heterogeneous, with some being relatively costly but others being cost free.

80

Chapter 4

ing weakness, and demands to fire the boss were treated in a comparable fashion. The dynamic variables are past wages and recent wage change. These are both expected to correlate negatively with strike success, because they measure the existence of preemptive concessions. Past wages are operationalized as the mean of average real per diem salary of coal miners 1, 2, and 3 years before the strike. Recent wage change is defined as the difference between the average real per diem salary 1 year before the strike and the same figure for the year of the strike. The measures were designed to be somewhat independent of each other by reducing their joint dependence on data from the period t — 1. Objective bargaining strength was measured using two indices. Favorable economic position was measured by change in total coal mining employment between the previous year and the present observation. Organizational strength was measured by the mandays lost during the strike. The number of mandays lost is mathematically a function of two components, the number of strikers who join the strike and the number of days the strike lasts. The number of strikers, in turn, measures worker strength in two ways. To the extent that strike size comes from simultaneous conflict in several firms, size measures the ability of unions and working-class organizations to coordinate offensives over a large area. To the extent that strike size comes from the participation of single large firms, size measures the primary sector status of the target, as well as the ability of employers to make generous concessions. Strike duration is affected by many factors, only some of which involve bargaining strength. Workers who are wholly unorganized are incapable of mounting long strikes. Workers who are inadequately organized may be able to mount long strikes, but lack any realistic possibility of winning them. Because strike duration is affected by other factors that have little to do with worker organization, however, a conservative approach was taken toward its use. In analyses not reported here, the number of mandays lost was divided into its two components and run separately; these equations nearly duplicate the ones presented. The analysis also includes a measure of mining fatalities, the percentage of workers who were killed in mining accidents in the department-year in which the strike occurred. Mining disasters had complex consequences for subsequent collective action in the mines; these accidents mobilized both the workers themselves and public support for the miners' cause. The mining fatality data come from the Statistiques de I'lndustrie Minerale. The unit of analysis is the entire strike. A strike was coded as a success if it obtained any concessions whatsoever from management.

Empty Victories, Harmless Defeats

81

Table 4.3 Logistic Regression of Strike Success on Selected Variables; 1890-1935 Bivariate

With controls

Variable

Logistic B

Standard error

Logistic B

Standard error

Fatality rate Employment change Days lost Past wage Wage change Low-cost grievance Grievance: Fire Boss Grievance with latent bargaining strength

.108 .134 .985 -.224 -.339 .657 -.430 .982

.043 .039 .305 .134 .254 .207 .304 .189

.100 .194 .642 -.295 -.682 1.290 -1.350 .575

.049 .045 .295 .156 .306 .294 .408 .212

Notes: N = 497. Null model: scaled deviance 677.6, 496 df; full model: scaled deviance 591.0, 488 df, p = .000. Significance statistics based on the decrease in scaled deviance obtained from contrasting the full model minus a model that excludes the variable in question. The coefficients and standard errors of some variables have been multiplied by a common constant for clarity. The variables and constants are fatality rate, 10; employment change, 1,000; days lost, 100,000.

Thus, if strikers made three demands and obtained only one, the strike was treated as a success. As before, both compromises and successes were viewed as successes for this purpose. The dependent variable is therefore a dichotomous dummy variable coded 1 if the strike obtained any concessions and 0 if it was a total failure. Table 4.3 shows a logistical regression of strike success on the explanatory variables. The evidence shows a strong negative relationship between strike success and both measures of wages. At a bivariate level, past wages and delta wages have a negative but weak relation to strike success. The Pearson correlation coefficients for past wages and delta wages are —.07 and —.06, respectively, with delta wages being significant at the .10 level and past wages being significant only at the .05 level. What is remarkable is that the performance of both variables improves markedly with the addition of statistical controls. The effect of past wages becomes more than 30% stronger when the rest of the variables of the core model are included, whereas the effect of delta wages more than doubles. The controlled coefficients of the two variables are significant at the .05 and .02 levels, respectively. These negative coefficients occur despite the presence of significant factors predisposing the data to produce positive results. In some cases, wages for f - 0 were measured after the strike had taken place. In these situations, one of the factors that would have raised wages would have been the very strike being considered. This would produce a positive

82

Chapter 4

artifactual correlation between delta wages and strike success for these cases, because strike success would be creating highly positive wage changes. This tautological relationship is so strong that it would seriously compromise the validity of any attempt to use a positive correlation between delta wages and strike success to support an alternative theory of strike success. There is, simultaneously, a latent positive correlation betw^een the wage variables and strike success that stems from the correlation between bargaining strength and wages. Because powerful, well-organized workers are likely both to have raised their pay in the past and to have been able to force employers to make recent monetary concessions, past wages and delta wages should be positively correlated with bargaining strength and, therefore, with strike success. Whatever force created a negative relationship between these variables and strike success was sufficiently powerful to overcome these two extremely nontrivial countervailing tendencies. The grievance content variables also produced significant results. Having a low-cost demand or a demand with latent bargaining strength increased the odds of a strike securing some concessions, whereas demanding that the boss be fired dramatically lowered the overall likelihood of winning. The three grievance variables are among the most statistically significant in the final model. Other analyses not reported here show that having a high-cost grievance or a demand with latent bargaining weakness did not significantly lower the chances for obtaining a concession. This is not a substantive negative finding but merely an artifact of how the dependent variable was measured. Because strikes were coded as being successful if they obtained any concession at all, the inclusion of any grievance that was easy to win correlated strongly with overall success. Demands that were hard to win were often irrelevant to strike success, because whatever concessions were to be granted would be obtained in other areas. In this light, the strong adverse effect of requesting the firing of the boss is especially interesting. This suggests that demanding to get rid of supervisors lowered the winnability of other grievances that were made during the strike. Such demands raised managerial hostility to such a level that even reasonable requests could expect rejection. Note that the coefficients for every wage and grievance variable except grievances with latent bargaining strength became stronger rather than weaker with the addition of controls. This suggests that none of these variables works primarily through one of the control variables as an intermediary, but that each one has robust independent effects. Thus, it is unlikely that the effect of wages on strike success comes from a tendency for workers with declining wages to have shorter, smaller

Empty Victories, Harmless Defeats

83

strikes or for workers with declining wages to come from districts with declining employment. Each of these hypotheses would have predicted that the controlled wage coefficients approach zero. The performance of the control variables is as predicted. All three strength variables have a positive relationship with strike success. In other analyses, other controls were considered, such as time period, turnout, and low-quality measures of government intervention. None had a direct effect on strike success; none changed any of the substantive findings that are reported here. The overall fit of the model is modest, though acceptable. No goodness-of-fit statistics are available for logistic regressions. Estimating the core model using OLS, however, generates an R^ of only .13. This is not a particularly strong performance. The model is supported by the fact that all of the relevant coefficients are significant and in the correct direction. Although moderation as measured here clearly predicts strike outcomes, it is also the case that a number of other variables would need to be considered in a fully specified estimation of French coal mining strike victories.

THE EFFECT OF STRIKE RESULTS ON WAGES All of the preceding analyses in this chapter are but a prelude to the central test. It was argued that strike victories do not raise wages, because strike results are as much the product of moderation as that of strength. Although the role of moderation has been substantiated, the critical issue is the correlation of wages and strike outcomes. Under ideal circumstances, it would be desirable to reestimate the equations of Table 3.2 with an additional variable for the average victory rate of strikes. Such a straightforward test, however, is not methodologically possible. The problem is that not all observations experienced strikes; it is possible to have department-years with winning strikes, departmentyears with losing strikes, and department-years with no strikes at all. Including the zero-strike years in a measure of strike results risks conflating inactivity with the victory rate. As an alternative strategy, one can regress wages on the number of winning strikes and the number of losing strikes. If victories increase wages and if defeats lower wages, then the coefficients for these two variables should be different, with the coefficient for winning strikes being positive and the coefficient for losing strikes either zero or negative. This empirical prediction is disconfirmed. Table 4.4 shows Parks equations predicting residual wages from the

84

Chapter 4 Table 4.4 Parks Estimates of the Effects of Strike Results on Wages, 1890-1935 1890-1935

Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present Total victories Total defeats N

1.15 14.3 3.60 5.51 6.33 9.73 -1.52 8.51 1.74 15.6 8.10 30.0 6.77 2.70 5.21 2.65 0.99® 0.25 8.65 2.23 800

1890-1922 1.64 24.4 5.51 6.52 8.92 11.6 -1.29 5.92 1.45 20.7 7.86 37.8 6.34 4.64 2.41 6.47 5.83 1.92 20.5 6.75 560

1923-1935 1.68 11.6 -0.95® 0.45 3.55 6.58 -0.65 4.17 2.50 14.1 9.45 42.7 -25.9 6.74 60.7 3.45 -10.0 1.97 -5.82® 1.27 240

Note: Sources and conventions as in Table 3.1. Total defeats in the 1890-1935 equation and total victories in the two period equations are significant at .05 but not at .01.

control variables of Chapter 3 and the number of winning and losing strikes. Results are presented for the whole period, and separately for the preschism and postschism periods. The victory and defeat variables have been recoded to match the coding of total strikes (i.e., 0,1, 2, 3 or 4, and 5 or more). In all three equations, the coefficient for winning strikes is less than that for losing strikes. In the 1890-1935 equation, victories had an insignificant positive effect on wages. In contrast, losses had a strong, significant positive effect on wages; the loss coefficient is nearly nine times as large as that of victories. In the preschism equation, both variables are significantly and positively related to wages. The beneficial effect of losses, however, is more than three times as large as that of victories. In the postschism equation, both variables have negative signs, but only wins significantly lowered wages; losses had no significant impact. The harm done by losses was approximately half of the harm done by victories, making losses less damaging for workers.

Empty Victories, Harmless Defeats

85

These equations are absolutely inconsistent with the claim that winning strikes is the key to obtaining higher wages. In every period, workers who lost their strikes outearned workers with large numbers of victories. The benefits of losing were reduced in the postschism years, when employers were able to resort to steering. In the one-union years, however, the gains from losing were sufficiently large to produce a positive global effect for the 1890-1935 period as a whole. In all periods, the greatest advantages came to workers who generated long strings of losses. This is consistent with the claim that being willing to strike despite a high probability of loss generates a reputation for militancy, which in turn translates into pre-emptive concessions. The highest paid workers seem to have been veritable kamikazes: workers who struck again and again even if their labor market position was weak, the demands were unattainable, and they had been trounced by management in previous encounters.

CONCLUSION The previous analysis suggested that there is no relationship between winning strikes and raising wages (or between losing strikes and lowering wages). If anything, it was more advantageous to be a loser rather than a winner. Losing strikes was associated with making radical demands, which included making expensive requests, challenging managerial job security, or striking after the receipt of recent concessions. Losing strikes was not especially associated with the presence of workplace rather than economistic demands. This analysis has a very important implication for labor history. Customarily, in accounts of social history, there is a tendency to evaluate the results of class conflict in terms of tactical victory. Analysts identify a major strike, determine if the workers obtained or did not obtain their concessions, and use this as an indicator of whether, in a deeper sense, workers won or lost. The present findings suggest that these assessments have to be tempered by a consideration of the centrality of the issues that were in dispute. Customarily, the major strikes that attract labor historians are major conflagrations concerning issues that are extremely important, as well as organizations and leaders that are particularly radical and outspoken. In such settings, tactical victories are likely to be substantive victories, because the concessions involved are hardly trivial. Tactical defeats, however, may not be substantive defeats. By workers' willingness to engage in a desperate struggle, they may communicate a posture of bona fide

86

Chapter 4

militancy that may induce future managers to make more generous concessions. Management may be unwilling to pay the costs that would be required to win a great strike a second time. The classic losses thus may have benefited workers if these induced management to defer future wage cuts or to provide judicious preventative wage increases. To assess the impact of strikes with complete confidence, the historian should not rely on immediate strike outcomes as his or her sole indicator. Long-term trends in residual wages have to be considered. Although a full program of multivariate analysis of wages may be impossible or inappropriate, at a minimum the historian needs to consider raw wages and to make a subjective assessment of the impact of other market forces. If losers receive preemptive concessions or stem their losses in declining markets, then their loss was only a short-term setback. If winners' subsequent wage offers are stingy, then it may be that the union's past victories were truly empty.

5 Is What You Ask for What You Get? Trade-Offs between Nonmonetary Benefits and Wage Attainment

The discussion in the previous chapters described a relatively simple world in which workers exclusively concerned themselves with monetary issues and attempted to maximize economistic utilities. Such models would appear to be challenged by the empirical materials presented in Chapter 4. Of the 1,211 grievances presented in Table 4.2, only 368 (or 30.3%) of these were salary offensives or defensives. Even if one treats disputes over late pay, surcharges, and free heat as monetary demands—as one rightfully should—this only raises the percentage of economistic demands to 47.7%. More than half of workers' demands were over nonmonetary issues.^ This raises a serious potential challenge to the findings of Chapter 3. It could be that differences in wage attainments are the product of differences not in the relative efficacy of class conflict, but in workers' preferences for cash over other benefits. Conceivably, a department with low wages could have a more effective strategy of industrial conflict than a department with high wages, producing a higher overall increase in worker utilities. It may have chosen, however, to take its winnings in the form of job control or shop-floor benefits rather than in salaries per se. The most explicit discussions of trade-offs between wages and non'This figure differs from the percentage of strikes over monetary issues, 56.5%. Many strikes fiad only one economic issue, a salary offensive, and many nonmonetary issues increasing the relative share of noneconomistic demands. 87

88

Chapter 5

economic benefits can be found in the literature in neoclassical economics on compensating differentials. The theory of compensating differentials holds that there is an inverse relationship between wages and the noneconomic amenities associated with particular jobs. The most important frequently mentioned alternative utility is job safety. Viscusi (1979) has argued compellingly that most workers place a higher value on physical survival than upon wealth. He then argues from this that workers who have to take jobs with high risks of occupational mortality must be paid a supplementary amount above marginal productivity to compensate them for the risk of physical harm. Although safety is the most commonly discussed alternative utility in the compensating differentials literature (Thaler and Rosen 1976; Olson 1981; Viscusi and O'Connor 1987; Hamermesh and Wolfe 1990), the approach has been just as profitably applied to inflexible hours, heavy lifting, noise, smoke, and demanding Work pace (Smith 1979; Duncan and Holmlund 1983). A parallel formulation can be called the theory of union choice. This theory would argue that workers' wage attainments are affected by the willingness of their unions to pursue monetary grievances. Economistic unions should make more money for their workers, whereas unions that emphasize job control or political agendas should make less money for their workers. These arguments are customarily put forward by leftist critics of the trade union movement. Orthodox Marxists, including Lenin (1977), Trotsky (1969), and Marx himself (1976) along with more modern authors such as Herding (1972) and Piven and Cloward (1977), have argued that bourgeois trade unions have sacrificed workers' legitimate interests in shopfloor issues, control over the labor process, and the development of a coherent socialist political program in the interest of the obtaining of immediate cash benefits. Put baldly, they claim that mainstream trade unions have been bought off. Marxist critics of trade unions do not have as their primary concern the prediction of workers' income. These authors are more concerned with explaining variations in workers' consciousness and the viability of generating an authentically socialist labor mobilization within capitalism. Their arguments, however, do make strong claims about income determination in reformist labor regimes. If a union preoccupation with wages can lower workers' attainment of shop-floor benefits, then a union preoccupation with shop-floor benefits can lower workers' attainment of wages. Presumably, if employers reward moderate unions with cash benefits, then they would withhold cash benefits from unions that pursue radical or nonmonetary grievances. Such actions could be punitive; however, they are more likely to be the result of employers responding to workers' precise requests. What workers ask for is what

Is What You Ask for What You Get?

89

they get: Those who strike for money get money; those who strike for seniority rights get seniority rights; those who strike for protective laws get protective laws; and those who strike for socialism get socialism (or alternatively, nothing). The evidence for both compensating-differential and union-choice theories is weak. The major writers on compensating differentials support their analyses with empirical results. Virtually all of the equations presented in this literature, though, are misspecified. The analyses use large-sample surveys of individuals to predict earrungs from such variables as age, race, and human capital stocks. The validity of these tests hangs on the accurate control of non-hazard-related determinants of wages. Data are never included, however, on such key determinants of wages as productivity, the monopoly status of firm, the capital-intensity of the workplace, or other indicators of the firm's capacity to pay. The omission of productivity in a neoclassical estimation of wages is particularly disturbing. Without any information on the demand-side determinants of wage rates, it is hard to interpret the meaning of the remaining coefficients. Because work settings such as mines and chemical refineries are both very capital-intensive and very dangerous, it is entirely possible that the coefficients associated with workplace danger are actually capturing unmeasured variations in primary sector status rather than danger per se.^ The literature provides almost no data to support the union-choice thesis. As was discussed in the introduction, there have been few systematic analyses of the comparative performance of unions varying on any dimension, let alone the dimension of grievance choice. This applies to both wages and nonmonetary issues such as job controls. Political attainments such as access to welfare-state benefits has been linked to socialist party strength (Stephens and Stephens 1982); however, the extent to which the economism of American unionism is the primary cause of weakness of socialist parties in America is a very live question for which both positive and negative arguments exist. This chapter provides tests of both the compensating-differentials and union-choice hypotheses using French coal mining data. Neither theory is supported; in fact, both fare rather badly. Their poor performance should not be surprising, for there are good theoretical reasons ^The equations in this book do not measure capital-intensity explicitly; however. They do include productivity, a strong correlate of capital-intensity. Mines that used compression drills and conveyor belts w^ere much more productive than those that used picks and horse-drawn wagons. The equations also include total employment, a direct measure of primary sector status.

90

Chapter 5

for expecting these two theories to fail. There are three major objections to theories of trade-offs between monetary and nonmonetary benefits, which are detailed below. 1. The logic of firm rationality suggests that nonmonetary disamenities should lower rather than raise wages if the disamenities lower aggregate productivity. The literature on compensating differentials tends to be written from the point of view of the worker. Danger, noise, and discomfort are viewed as negative utilities; workers have to be compensated for putting up with these disamenities. Wages, however, are not solely set by worker preference. The neoclassical theory of wages argues quite compellingly that employers set wages equal to marginal productivity. Consider the case of two jobs with equal productivity, with one job being significantly more hazardous than the other. It may be true that workers will prefer the safer job. In a purely competitive market, however, if the hazardous employer raises wages to compensate workers for their danger and does not receive supplementary productivity as a result, his or her goods will be uncompetitive relative to those of safer employers who pay lower wages. Thus, the cost of hazard pay is elimination from the labor market. Unfortunately for the hazardous employer, the cost of not paying compensating differentials is also elimination from the labor market if workers insist on hazard pay as a condition for taking on dangerous work. Neoclassical theory thus does not suggest the existence of compensating differentials; it suggests the withering away of dangerous work through market competition. The argument becomes more forceful when one considers the productivity consequences of dangerous work. Industrial accidents impose substantial costs on employers. At the most trivial level, they deprive firms of the continuous service of trained employees and represent an aggravated form of absenteeism and turnover. More significantly, accidents produce losses of capital as well as labor power. In some laborintensive settings, these losses may be restricted to the repair or replacement of broken machines. In many high-risk workplaces, however, potential capital losses can be catastrophic. Chemical plants, the merchant marine, nuclear reactors, air transport, and, significantly, coal mines, all represent settings where accidents can produce major disasters costing hundreds of thousands or millions of dollars. A coal fire not only hurts workers and damages the physical plant, it destroys the very natural asset that the mine was designed to develop in the first place. Even more serious than these are the costs of legal liability. Any party damaged by the accident, including the worker, has some form of legal redress against the employer. Even when companies settle lawsuits

Is What You Ask for What You Get?

91

for a fraction of actual damages, their overall payouts are large. The size of the Bhopal payment was undoubtedly greater than the marginal productivity of the handful of workers responsible for the accident. To be sure, liability insurance and worker's compensation reduce corporations' out-of-pocket expenses for their most dramatic obligations. Insurance, though, merely averages out risk. Employers of dangerous technologies pay high insurance premiums and contributions to worker's compensation as a routine cost of doing business. All of this suggests that all other things being equal, dangerous workplaces are less productive than safe workplaces. Even if the output per man-hour of safe and hazardous technologies are the same, the costs directly associated with accidents will lower the overall productivity of the latter. If wages are linked to marginal productivity, however, this means that wages must be lower in occupations and firms with high risks of mortality. This argument is not original. It appears in the works of the compensating-differentials theorists themselves (Viscusi 1979; Olson 1981). They use this logic to explain why older workers have lower rates of compensating differentials. Older workers are perceived as being more dangerous, so employers must hold some pay in reserve for the greater potential number of accidents. It is not explained why employers hold reserves for older workers and not for dangerous occupations as a whole. 2. Wages and benefits are determined more by variations in union strength than they are by variations in union choice; choice is only relevant to unions at a narrowly intermediate level of strength. The theory of union choice implies that unions have a choice about which benefits they receive. Such a claim carries with it a number of implicit assumptions. First, for choice to matter, the union has to be strong enough to win some of what it asks for. The content of demands is irrelevant for a trade union that is incapable of exerting any pressure on management. Second, the union has to be sufficiently weak that it cannot obtain everything it asks for. If the union were omnipotent, then it would never have to make choices. Thus, the theory of union choice implies that workers are strong enough to be allowed to win some, but not all, of their demands. Such a situation is not impossible; probably the number of settings in which unions have to make choices is numerous. Union benefits, however, will also be determined by the strength of workers relative to management. It is an empirical question what percentage of residual wages is explained by variations in worker strength and what percentage is explained by union choice among populations of workers of intermediate strength. It seems likely, though, that because

92

Chapter 5

union strength is relevant to all wage determination processes and union choice is only relevant to a few, the effects of strength will be more globally important. 3. Wages should be positively, rather than negatively, correlated with striking over noneconomic benefits, because noneconomic strikes generate a reputation for militancy. The previous chapters have argued that unions with a reputation for militancy are more likely to obtain high wages. Radical unions in turn are more likely to obtain a reputation for militancy. Thus far, the only indicator for militancy that has been used is number of strikes; however, another indicator is a willingness to use the union to pursue nonmonetary workplace and political agendas. The English unions of the 1950s and 1960s obtained a well-deserved reputation for contentiousness for the large number of job control strikes led by their shop stewards (Sturmthal 1972; United Kingdom Royal Commission on Trade Unions and Employers Associations 1968; Burawoy 1983). Technological changes and work reassignments were challenged routinely often by several different unions in succession. Similarly, the Lordstown strike of 1972 had a similar galvanizing effect by demonstrating that rank and file members of the UAW were willing to strike over personal fulfillment and boredom on the job. Political strikes can have the same impact. What has always made the communists seem radical is their explicit goal in earlier years of overthrowing the capitalist state. In France, their willingness to commit other industries to striking to have the railways nationalized, to coordinate French and German strikes to protest the invasion of the Ruhr, or to walk off the job to protest the war in Morocco convinced employers of the sincerity and commitment of the unitaires. Communist strike rates were high precisely because they organized such a large number of job control and political walkouts. If this claim is correct, then it poses a major challenge to the orthodox Wisconsin school legitimation of traditional American business unionism. Authors such as John Commons (1913) and Selig Perlman (1928) argued that business unions would be more effective than political or radical unions in meeting workers' economistic needs, which the Wisconsin theories viewed as primary. Business unions are supposedly better at raising salaries because they avoid wasting their resources on political or revolutionary demands that have no immediate impact on wages. The present argument, though, suggests just the opposite: By pursuing radical social programs, unions develop a reputation for militancy that enhances the capacity of workers to extract immediate concession. The choice is not money or revolution. The choice is both or neither. These three objections to the theory of trade-offs will be supported

Is What You Ask for What You Get?

93

below by empirical evidence concerning the wages of French coal miners. First, it will be shown that wages are positively rather than negatively correlated with the presence of nonmonetary goods, such as increased job safety or reductions in the workday. There is no evidence that the receipt of side benefits produces any reduction in earnings per se. Second, it will be shown that within strikes, trade-offs between monetary and nonmonetary grievances were rare. An analysis of grievance pairs shows that, in most cases, workers either lost all of what they asked for or gained all of what they asked for. Class conflict affected wages through the vehicle of increasing workers' objective strength or their reputation for militancy, rather than through discretionary union selection of benefits. Third, it will be shown that nonmonetary strikes had a greater relationship to wages than did monetary strikes. Wages were not raised through workers asking for money; they were raised by workers generating a reputation for militancy by striking for demands other than immediate wage increases.

THE CORRELATION OF WAGES AND NONMONETARY BENEFITS The theory of compensating differentials predicts that wages will be negatively correlated with the presence of nonmonetary goods. An ideal test of this model would regress wages on the sum total of all nonmonetary benefits received by French coal miners. This is obviously not practical. There are no data on the pace of work, the heat of the coal face, the noise of the machinery, or the volume of heavy lifting associated with various mines. There were also significant non-salary-based economic benefits received by miners that cannot be systematically measured, including company housing, subsidized medical care, and subsidized groceries. Although statistics are available on these for isolated localities and dates, the material is too sketchy to sustain large-sample analysis.^ ^Superficially, it would seem that the most prosperous basins had the greatest provision of corporate welfare. This is not surprising given that such programs could be very expensive. The largest company towns with the most lavish medical clinics seem to be in Pasde-Calais, Nord, and the Loire, although well-known company towns developed in such midsized fields as Carmaux, Montceau-les-Mines, and Decazeville. It might seem that the Loire would rely less on company housing because of the proximity of Saint-Etienne to the major field. In actual practice, firms in outlying areas, such as Cie. de Roche-la-Moliere at Hrminy, were well known for their highly developed company welfare programs, which included housing, clinics, and schools. Company paternalism probably exacerbated preexisting wage differentials rather than compensated for them, because marginal firms could not afford the additional investment in nonproductive facilities.

94

Chapter 5

It would be particularly interesting to regress wages on the provisions of political benefits. This would make it possible to examine direct trade-offs between economistic benefits and the extension of the welfare state. Such a desirable analysis, however, cannot be performed on French coal miners. Coal mining was so heavily regulated, and the coal miners' union was so active in political reform, that there simply is no variance in the rate of provision of political goods. Cross-sectional analyses are inappropriate, because welfare-state benefits were provided nationally across all coalfields. Longitudinal analyses are stymied by the fact that virtually every year, some form of coal mining reform was enacted by the state. The Annales des Mines, the official government yearbook of changes in mining laws and regulations, listed between 30 and 300 pages of revisions in coal mining labor law every year; nearly all of these changes were ameliorations in some form. Nearly all that can be said is that workers started with a relatively low level of legal protection, and this improved irregularly but steadily over time until they enjoyed a substantially higher level of coverage (Trempe 1984). Favorable regulation is thus correlated with time. As the equations of Chapter 3 showed, since wages also increased with time, workers seem to have made steady progress in both economistic and regulatory spheres. The analysis here is rough, the data could not support a more fine-grained consideration. Using this crude contrast, however, there is no evidence of a trade-off between political benefits and wages. A more sophisticated analysis can be made of the relationship between wages and the length of the working day. A reduction in the length of the workday was one of the most fundamental demands of the coal miners. This issue appeared in every major national strike wave before World War I and was the central grievance of two major strike waves in 1919. It even appeared as an issue in the strikes of the Popular Front of 1936 (De la Taille 1939). It is easy to see how length of the workday would be an issue in coal mining, given the extraordinary physical demands of coal extraction and the grim disamenities of the work environment. As if this were not enough, however, the length of the workday was also a fundamental economic issue. Workers were paid by the day, not the hour. As such, employers could demand overtime at will without providing supplementary payments. An employer who wished to lower the workers' effective hourly pay rate merely had to extend the workday an extra hour. The need to restrict companies' capacity to extract free overtime motivated the passage of several laws limiting the length of the workday. There were significant disputes over the passage and administration of these laws because, as will be seen in Chapter 7, the definition of

Is What You Ask for What You Get?

95

what labor constituted the workday was continually changing, with employers losing their rights to have transport time, changing time, and job preparation could as unpaid labor (Comite Central des Houilleres Fran^aises 1921a; Saposs 1931). By logic of compensating differentials, one would expect that workers with shorter hours should receive less pay per day. In a world with invariant hourly pay rates, workers who work shorter days should receive less money per day; this applies regardless of whether one views limitations in hours as an economistic or shop-floor concession. But in fact, the opposite seems to hold true: Workers with short hours received higher salaries per day. The evidence for this comes from a study of working conditions carried out by the Office du Travail in 1892 at the request of the Chamber of Deputies. A systematic survey was made of hundreds of workplaces in a large variety of French industries, including 52 coal mines. The sampling was neither random nor scientific. Some attempt, however, seems to have been made to ensure geographical representativeness, as well as the inclusion of both major and marginal employers. The study was primarily descriptive and includes little detail about the individual informants. Nevertheless, data were collected on both the size of the employer and the average salary rate. Note that the salary rate provided here is the formal salary rate stated in the collective contract or in informal practice, not the measure used in the analyses of this book (the sum of all funds paid out in salaries divided by the number of days worked). As such, the salary materials here are de jure rather than de facto; we do not know how they were affected by bonuses, fines, reclassifications, and other variations in the daily rate. Table 5.1 shows the average salary rate of the coal mines in the 1892 Office du Travail study, broken down by the length of the working day. These figures show that workers with shorter workdays tended to reTable 5.1 Mean Salary Rates per Day by Length of Workday: Individual Mines, 1892 Length of day

Mean salary rate

N

8 hours or less 8V2 hours 9 hours 9V2 hours 10 hours IOV2 hours or more

3.94 3.98 3.93 3.66 3.52 3.50

8 12 17 4 9 2

Source: France. Minist^re de Commerce, Office du Travail 1894.

96

Chapter 5

ceive higher rates of pay. There is a rough split between workers who worked 9 hours a day or less and those who worked longer days; the former received approximately 3.9 francs a day, whereas the latter received only 3.5 francs a day. The bivariate relationship presented here could be misleading. It is highly likely that the firms varied in their capacity to pay high wages, and that unmeasured variables such as productivity or primary sector status could account in part for the present findings. The data do not permit finely controlled analyses similar to those of Chapter 3. Employer size is available as a rough proxy for employer productivity and prosperity, however. Total employment has been traditionally accepted as an indicator of primary sector status (Wallace and Kalleberg 1981). The limitations of total employment as a complete indicator of these concepts are obvious. Table 5.2 shows ordinary least squares regression estimates of the effects of employer size and length of workday on daily wage rates. OLS regression was used rather than Parks because, unlike the datasets examined previously, this is a simple cross-sectional sample. At the bivariate level, employer size raises wages, and length of workday lowers them. Both relationships survive being included in a trivariate equation, although the length-of-workday variable is significant only at the . 10 level. The same inverse relationship that exists between wages and long workdays exists between wages and job mortalities. This is of particular note because occupational safety is the classic nonmonetary utility that is invoked in discussions of compensating differentials. Virtually every cite in the previous literature review on compensating differentials conTable 5.2 Regressions of Mean Salary Rates on Length of Workday and Mine Size: Individual Mines, 1892 Variable Length of workday

I

II

-.244 2.15

-.182H

Total employment R2

.08

III

.202 4.04 .25

1.78 .188 3.79 .29

El Significant at .10 Source: See Table 5.1. Note: N = 52. Top figures are regression coefficients. Bottom figures are (-statistics. For ease of presentation, employment coefficients and (-statistics have been multiplied by 1,000.

Is What You Ask for What You Get?

97

Table 5.3 Parks Estimates of the Effects of Mining Fatality Rates and Control Variables on Wages, 1890-1935

Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present

1 18901935

11 18901922

HI 19231935

IV 18901935

V 18901922

VI 19231935

1.14 14.1 3.64 5.38 6.36 9.81 -1.51 8.39 1.75 15.6 8.18 30.5 6.14 2.44 5.47 2.80

1.64 23.9 6.00 7.00 8.82 11.7 -1.26 5.87 1.47 22.5 8.09 39.3 5.33 3.73 3.50 2.73

1.64 13.0 -2.97 1.73 5.09 15.1 -1.09 9.53 2.76 20.9 9.01 50.4 -2.43 7.64 4.05 2.65

-1.48® 0.69

-0.50® 0.42

-20.2 4.88

1.15 14.2 3.60 5.30 6.36 9.80 -1.51 8.44 1.73 15.3 8.12 29.9 6.62 2.61 5.18 2.64 4.85 1.43 -1.44® 0.67

1.66 24.7 5.51 6.63 9.08 11.9 -1.33 6.16 1.45 21.6 7.88 38.6 6.31 4.54 2.58 2.69 13.8 5.27 -0.40® 0.37

1.69 14.8 -2.33 1.28 5.21 15.4 -1.07 9.80 2.76 19.5 9.10 54.6 -2.67 9.39 3.89 2.58 -1.29 5.14 -20.2 5.10

Total strikes Mining fatality rate

Note: Sources and format as Table 3.3. N = 800 for whole-period equations, 560 for preschism equations, and 240 for postschism equations.

tained an empirical analysis of the relationship between wages and job hazard rates. Mining fatality data are available in the Statistiques de I'Industrie Minerale. Although this source also contains material on accidents and injuries, the data reflect very strong period and artifactual errors associated with changing definitions. Although the definitions of both accident and serious injury are subject to legitimate dispute, there are fewer ambiguities in the assessment of mortality.^ Table 5.3 presents Parks estimates of the effects of the mining fatality rate (as measured at the department-year level) on wages, with the standard controls of the core ••There are minor random errors stemming from fatalities that occurred after long convalescences from injuries. Statisticians can be comforted by the fact that many mining deaths occur relatively quickly. Cave-ins, floods, explosions, and precipitous fires reduce the share of lingering deaths from falls or mechanical mishaps.

98

Chapter 5

equations of Chapter 3. Because strike activity showed significant period effects, equations are estimated for 1890-1935 as a whole and separately for the periods before and after the schism. Although there are significant period interactions in the effect of mortality on wages, in no case is there any appearance of a relationship that even remotely supports compensating differentials. For 1890-1935 as a whole and for the years before the schism, there is no relationship between mining fatalities and wages. This is true regardless of whether strike activity is included in the equation. For the postschism years, there is a significant negative relationship between fatalities and wages; workers with safer jobs had higher wages. The explanation for this finding must remain conjectural. It could be that government safety inspectors did less assiduous work in communist areas. There could also be a physical basis for the relationship. For example, there were dramatic breakthroughs in fire-fighting technology in the 1920s (Abadie 1927). Because these were more likely be adopted in the more technologically progressive mines, this could have produced a multiple correlation between firm prosperity, wages, and safety. It could also be that in the later years of this study, mine shafts were aging. Older mines are both less productive and more dangerous, as further extraction requires the development of smaller, gassier veins. This would produce a relationship between wages and safety that would increase over time. Possibly all of these processes occurred—or, equally possibly, none of them did. These ambiguities aside, none of the periods show a positive correlation between fatalities and wages. Such a finding, though, does not permanently discredit compensating-differentials theory. Advocates of compensating differentials might argue that such a test is too harsh. Mining fatality data include major catastrophes that can produce acute spikes in the death rates of particular years. It would be unrealistic to expect wages to soar in the year of the Courrieres disaster, for example, and return to normal levels the next year. Even so, reestimating the equations in Table 5.3 with average departmental fatality rates for the entire 1890-1935 period, rather than annual mortality data, does not change the present findings. Such a procedure eliminates any errors from one-shot catastrophic outliers. This negative finding suggests that regions with mines that were perpetually dangerous did not pay more than regions with mines that were generally safe. Viscusi (1979) has argued that dangerous work attracts individuals with a high tolerance for risk, who in turn are willing to accept lower levels of hazard premiums than would be required by the general population. It is entirely plausible that most coal miners were significantly braver than members of the average population. A coal mining culture

Is What You Ask for What You Get?

99

probably evolved with highly developed mechanisms for coping with danger. Although the hazards of mining may have produced workers with distinctive personalities, however, it did not produce compensating increases in pay. Employers preferred to set pay equal to marginal productivity and raise wages only when they were forced to by union pressure.

GRIEVANCE TRADE-OFFS WITHIN STRIKES The grievance data analyzed in Chapter 4 allows for the analysis of pairs of grievances as well as single grievances. The Statistiques des Greves lists all of the grievances for every strike. Thus, it is known not only which demands appeared and which were won, but also which demands appeared in conjunction with which others and which ones were sacrificed in return for other demands. Using such material, it is possible to reconstruct both the extent of grievance trade-offs that existed in French coal mining and any systematic patterns as to which of any two conflicting demands tended to be granted. This is a somewhat spectacular analytic opportunity. Unfortunately, it is only available for the years preceding World War I. After World War I, the French government adopted a simplified format for strike reportage that essentially only reported the primary grievance in a strike, rather than the full list of demands. Although in some cases secondary demands were noted, in more than 80% of the cases only a single grievance is listed, despite evidence from other sources that the list of demands was substantially more complex. For the prewar years, the statistical compilations are much better listing three or four grievances routinely; large strikes often received supplementary write-ups in which the full negotiating platforms were transcribed word for word. Table 5.4 lists the outcomes of all grievance pairs before World War I that involved one economistic demand and one noneconomistic demand. The primary focus here is on trade-offs between monetary and nonmonetary demands. Trade-offs between different monetary demands can be ignored, as these simply involve the form that the overall wage package will take. Likewise, trade-offs among nonmonetary demands are irrelevant to the present study, because these would not have affected wages. For the purposes of this analysis, all economistic demands were merged into one. This superdemand, called pay, was coded as having succeeded if any of its components were granted by management. The pay demand was coded as having failed if the employer granted none of

100

Chapter 5 Table 5.4 Outcomes of Grievance Pairs, 1890-1914

Grievance other than pay

No choice made

Pay wins. other loses

Pay loses. other wins

N

Work rules Work assignments Safety Comfort Punishment/layoff Ethnic exclusion Fire worker Fire boss Work hour Pension Equal income Union procedural

84.3% 83.4% 45.5% 75.0% 69.6% 100.0% 100.0% 60.8% 73.7% 73.4% 82.7% 76.6%

3.5% 11.1% 27.3% 5.4% 26.1% 0.0% 0.0% 34.8% 18.2% 5.3% 13.8% 0.0%

12.3% 18.0% 27.3% 19.6% 4.3% 0.0% 0.0% 4.4% 8.1% 21.3% 3.4% 23.3%

57 18 11 56 23 4 3 23 148 75 29 30

75.1%

14.9%

10.0%

477

Total

Note: Source and definitions as in Table 4.1. Pay grievances include salary offensives, salary defensives, surcharges, heat, and late pay/pay timing.

the components of the grievance. The definition of grievance success here closely parallels the definition of strike success that has been used throughout the book. The five grievances that were combined into pay include salary offensives, salary defensives, surcharges, heating, and late pay or pay timing. The heating measure was included to allow pay to contain disputes over payments in kind as well as in specie. The analysis then tabulated all pairs of grievances involving a pay and a nonpay grievance. If a pay grievance and four noneconomic grievances appeared in the same strike, then that strike produced four separate grievance pairs for analysis. Such a procedure maximizes the information that is available about the operation of every conceivable type of grievance pair. It has the liability that the cases reported are not technically independent of each other. This poses no problem for the narrowly descriptive analysis that follows; it does, however, preempt the use of techniques that would require significance tests. Table 5.4 shows the distribution of the results of each type of grievance pair. If both demands succeeded or failed, they were coded as representing no choice. In these cases, workers did not have to select which grievance they preferred, because both the monetary and the nonmonetary demand received identical treatment. The table also lists the percentage of cases in which the pay demand was granted and the

Is What You Ask for What You Get?

101

nonmonetary demand was denied, as well as the cases where the reverse occurred. The primary finding of Table 5.4 is that trade-offs were relatively rare. In more than three fourths of the cases, no choice among grievances was made: Workers either obtained both demands, or lost both demands. This finding is particularly marked for the classic shopfloor control grievances; 84% of the grievance pairs involving work rule disputes and 83% of the grievance pairs involving work assignment disputes resulted in no choice being made. In 100% of the cases involving ethnic disputes or demands to fire workers, the issues also were resolved without any grievance selection. For 10 out of the 12 grievances, no choice was made at least two thirds of the time. The two exceptions to the rule were safety disputes and disputes over firing the boss. These cases represented less than 8% of the total sample. The cases where choices were made are worthy of note. In general, workers selected the pay demands over the nonpay demands approximately 60% of the time. The relative evenness of this split belies any general theoretical argument that workers or unions unambiguously prefer monetary to nonmonetary concessions. Sometimes workers preferred money; sometimes they preferred better working conditions. Workers were particularly likely to select monetary benefits when the nonmonetary demand in question was being vigorously resisted by management. The grievances that show the greatest likelihood of pay winning relative to the other demand are punishment/layoffs and firing the boss. As was shown in the last chapter, these demands were particularly difficult to obtain overall. The grievance that was most likely to override monetary demands was union security. Here it seems quite likely that union officials were particularly willing to give u p cash benefits in the present in return for future strategic advantage. The remaining and less extreme findings probably should not be overinterpreted. The Ns in each category are relatively small; the number of examples of trade-offs are further reduced by the large percentage of nonchoice cases. As such, many of the intermediate findings are based on particularly small populations that are likely to be unreliable. Fewer than 25% of the known grievance pairs involving monetary and nonmonetary grievances entailed workers' making a choice between money and working conditions. This number is an overstatement of the amount of trade-offs that existed, for it by definition excludes all one-grievance strikes. It would be hard to characterize a labor movement as economistic or noneconomistic based on the properties of less than one fourth of the known choices.

102

Chapter 5

Believers in union choice might argue that the present test is unfair. The present analysis involves choices among demands that are formally articulated in strikes. Demand choice might occur in the grievance generation process rather than in the subsequent negotiating process. Economistic unions might exclusively make pay demands and censor all nonpay demands. Shop-floor-oriented unions might exclusively make nonmonetary demands and censor pay demands. The present data, however, do not suggest that censoring was especially prevalent. Table 5.4 lists 477 monetary-nonmonetary grievance pairs, a number substantially in excess of the 315 strikes that occurred between 1890 and World War I. The inclusion of both types of grievances in strikes was relatively common. It could be argued, though, that some editing of grievances did occur and that the downplaying of nonmonetary grievances reduced shop-floor gains while facilitating improvements in workers' wages. This claim requires more direct investigation.

THE EFFECTS OF PAY AND NONPAY STRIKES ON WAGES Table 5.5 shows the effects of pay and nonpay strikes on wages. The format used is similar to that of Table 4.4. Parks equations are run predicting wages from a full set of control variables, the number of pay strikes, and the number of nonpay strikes. If union-choice theorists are correct, then pay strikes should have a strong positive effect on wages, and nonpay strikes should have either a zero or negative effect. In contrast, the reputation-for-militancy theory would predict that nonpay strikes would be more highly correlated with wages than pay strikes, because the former represent a greater commitment to radicalism. For present purposes, a nonpay strike is defined as any strike in which absolutely no pay demands were made. Pay strikes comprise both strikes only involving monetary issues and those involving issues of both types. The present coding was selected to artifactually minimize the likelihood that nonpay strikes would have a positive coefficient, producing a conservative bias to the analysis. As is the case in the previous analyses, the number of both pay and nonpay strikes were collapsed into five categories (0, 1, 2, 3 or 4, and 5 or more). The findings in Table 5.5 support the reputation-for-militancy model over the union-choice model. For 1890-1935 as a whole and for each of the two major subperiods, nonpay strikes are more positively related to wages than are pay strikes. For 1890-1935 overall, pay strikes are insignificantly related to wages, whereas the effects of nonpay strikes are significant and positive. The effects of nonpay strikes are more than 18 times greater than the effects of pay strikes. The findings are similar for

Is What You Ask for What You Get?

103

Table 5.5 Parks Estimates of the Effects of Monetary Strikes, Nonmonetary Strikes, and Control Variables on Wages, 1890-1935

Productivity Employment Coal prices General prices Percentage nonagricultural National coal wages Year Union present Pay strikes Nonpay strikes

1890-1935

1890-1922

1923-1935

1.13 13.9 3.53 5.23 6.28 9.54 -1.50 9.54 1.75 15.0 8.13 30.1 6.42 2.53 4.99 2.51 0.44® 0.11 9.13 2.13

1.64 23.4 5.19 5.92 8.73 11.8 -1.22 5.69 1.47 19.6 8.08 19.6 5.42 4.08 2.40 2.41 0.57Q 1.75 1.70 3.97

1.69 11.6 -1.37 0.72 3.54 6.14 -0.71 4.38 2.45 14.3 9.34 41.4 -2.53 6.57 5.66 3.20 -1.71 3.35 0.16® 0.32

Note: Sources and format as Table 3.3. N ^ 800 for wfiole-period equations, 560 for preschism equations, and 240 for postschism equations.

the preschism years, except that nonpay strikes are only 4 times more effective. Pay strikes are almost statistically significant, although not nearly to the degree that nonpay strikes are. In the postschism years, pay strikes are negatively related to wages. This effect is strong and statistically significant. In contrast, nonpay strikes were harmless; the effect of nonpay strikes was not significantly different from zero. Some of this effect probably captures the greater willingness of communists to pursue political and shop-floor demands during this period. Communist strikes raised wages in communist areas and lowered or had no effect on them elsewhere.

CONCLUSION On balance, the data presented here provide little support for theories of compensating differentials or union choice. Wages were not inversely correlated with the provision of nonmonetary benefits. Union

104

Chapter 5

choices were found to be statistically rare. Workers did not make more money by striking for immediate economic demands. Strikes in which no monetary demands were made had more beneficial effects on wages than did strikes over pay. This discredits theories claiming that unions hurt their workers financially by pursuing broad-based social programs that have little obvious relevance to wages, salaries, or fringes. Arguments that American unions outperform European unions because they do not dissipate their energies on nonwage concerns need critical reexamination in this light. Of course, a true assessment of the impact of political unionism has to consider the relationship between capital, labor, and the state. It is necessary to ground the present econometric analyses in the overall political context of labor relations in French coal mining; it is to these larger state-related questions that the analysis now turns.

6 The State versus Managers Contradictions of Early Corporatism in French Coal Mining

The model described in Chapter 2 was an oversimplification of what actually occurred in French coal mining. The predictions from the model fit the data well; however. Chapter 2 suggested that wages were determined by the actions of autonomous private sector employers and unions. In fact, the state was an extremely active participant in the industrial relations in French coal mining. The primary labor contracts were negotiated in formal trilateral sessions with management, the union, and government representatives. Moreover, the state frequently imposed its will on the final settlement by exerting unilateral pressure on one of the two sides to make concessions. Because the state had substantial legal resources at its disposal, the party receiving pressure often had little recourse but to submit to the wishes of the public authorities. Thus, one cannot fully understand the process of wage determination in French coal mining without a coherent account of the logic of state intervention into the negotiating process, and how it would interact with the normal dynamics of two-party bargaining. The analytical task would be simple if the state and managers had the same interests. Then the union could be viewed as negotiating with two relatively similar authorities, and the theoretical predictions would be similar to those in the employer-authority case. The state, however, had parallel but different interests from employers. The French government was highly motivated to reduce strike volume, but at the same time, it vacillated in its commitment to reducing wages. In some cases, the state acted aggressively to curb labor costs; In others, it let wages float in the interest of industrial peace. It is not surprising that govern105

106

Chapter 6

ment officials who placed a high priority on social peace and a lower priority on wage restraint would force employers to end strikes by paying high wages. This produced many of the high returns to strikes that were predicted in Chapter 2 and empirically observed in Chapter 3. This counteracted a relatively idiosyncratic property of French coal mining management. Unlike most other employers, French coal operators had very little interest in reducing strikes. The unusual economic properties of the French coal market virtually ensured that companies could recoup the costs of industrial stoppages, no matter how long these lasted and under what market conditions they occurred. At the same time, the coal operators had a permanent, unambiguous interest in obtaining the lowest possible wages. Left to their own devices, the private companies thus would have been unlikely to raise salaries as a strategy for forestalling future industrial militancy. The inclusion of the state as a powerful negotiator in collective bargaining substituted the presence of a strike-sensitive, wage-neutral authority for a strike-neutral, wagesensitivity authority. This rekindled the interest in the control of strikes that one would expect in a more traditional private sector employer. Put differently, the model in Chapter 2 would apply to most private sector employers. It did not, however apply to French coal operators. It is only because of the participation of the state that the strategy of strike maximization became feasible. The active role of the government also facilitated the process of steering. Steering is most likely to occur when a single employer controls two or more workplaces with different union compositions. It is then possible for him or her to pay one set of workers more than the other based on their union affiliation. Steering becomes more problematic when the different workplaces are controlled by different employers. If the employer with the moderate workplace believes his or her union affiliation is secure, he or she may have little motivation to pay higher wages to provide an example for workers in other distant workplaces to adopt moderate unionism. For such an employer, paying high wages to reduce the militancy of the national labor force would be a public good that he or she would have to finance with private expenditures. The provision of a peace premium to the moderate factory would only occur under two conditions. First, if the employer of the moderate factory perceived a potential threat of radical mobilization in his plant, higher wages could be paid. Otherwise, there would need to be some external pressure put on the employer to pay the peace premium as a public good. An employer's federation negotiating a multiemployer contract could be the institutional agent to exert such pressure. Alter-

The State versus Managers

107

natively, the state could act as the organizer of such collective action, coordinating employer responses in the larger interests of the capitalist class. The French state was not narrowly instrumentalist. As such, the presence of corporatist labor-relations institutions presented conflicts between the logic of private capital and that of public authorities. Significant policies developed from every cause imaginable: capitalist agitation, working-class pressure, and the autonomous concerns of the state bureaucracy. Regardless of the origins of Third Republic state policies, however, the consequences of these policies often resembled the predictions of Marxist functionalists (O'Connor 1973; Offe and Ronge 1975). The short-term interests of individual coal companies were frequently overridden in favor of policies that favored accumulation at the national level and the facilitation of social reproduction through the minimization of industrial conflict. The state acted as an employer's proxy in attempting to reduce strike costs and, to a lesser extent, to lower the wage bill. Private French coal operators, with their own bizarre indifference to strike costs, would have been unlikely to replicate governmental strategies of labor control. The state forced them into a more conventional concern over militancy. This chapter is a prelude to the account of state and managerial attempts to control labor militancy that follows. What is presented here is a background guide to the motivations of both public and private authorities. The indifference to strikes of private capitalists is explored, as well as the origins for the strong governmental aversion to strikes. The institutional origins of governmental ambivalence concerning wage levels are also discussed. There were many factors predisposing state policymakers to push for lower wages; these were, however, balanced by an alternative set of high wage preferences. In subsequent chapters, the history of state intervention in French coal strikes is detailed as part of the larger narrative on the evolution of strike strategy.

MANAGERIAL INDIFFERENCE TO STRIKES Unlike the French government, the coal companies themselves were relatively indifferent to strikes. Prefectural documents contain multiple quotations manifesting deep concern about potential strike activity. This apprehensive sensibility is generally absent from managerial documents. Correspondence among members of the national or regional employers' federations may include informal estimates of the likelihood of striking, but relatively few warnings or premonitions about

108

Chapter 6

the adverse consequences of such activity. Even personal correspondence between company executives, such as may be found in the strike dossiers of the Cie. de Bethune or the general records of the Cie. de Montrambert et la Beraudiere, show little of the distress of the governmental records. In some cases, one even observes managerial braggadocio: In 1931, at a private meeting of the top leaders of the Comite Central des Houilleres (hereafter CCH), the coal employers association, with the minister of labor, Pigeot, head of the CCH in Loire, boasted about his capacity to force concessions on either confederes or unitaires regardless of whatever strike action they took (Loire 15 J 2449). Obviously, the documentary evidence presented here is hardly conclusive. In labor relations, intermediaries are far more likely than bargainers to discuss verbally the cost of strikes; active negotiators always feign indifference to strike costs as a strategy for obscuring their actual state of vulnerability. Confessions of weakness are more likely to be made verbally and confidentially than committed to paper for the historical record. Superficially, the claim that coal companies were indifferent to strikes would seem to be absurd. Since Peter Stearns's (1968) seminal essay on managerial responses to industrial conflict, we normally think of the French bourgeoisie as being committed to an all-out war to eliminate the strike and reestablish the authority of the patron in his workplace. French businessmen were unquestionably authoritarian and antiunion; they took all of the adaptive steps, such as buying strike insurance and using private thugs, that are carefully detailed by Stearns in his essay. However, there is a difference between the labor strategies of an employer who cannot tolerate participating in a strike, and those of an employer who will take as many strikes as is required to establish his authority over the labor force. The former either concedes rapidly, or, if he follows a Hicksian strategy, willingly participates in collective bargaining, mediation, and other information-sharing forums that allow for the peaceful simulation of industrial conflicts. Coal operators were of the second breed. They were not against the strike; they were against the union. Most social scientists and historians use some form of resource mobilization model in analyzing when employers will be willing to tolerate strikes. In a resource mobilization model, employers should tolerate strikes when they think they are going to win, and avoid them when they expect to lose (Tilly 1978; Oberschall 1973). Thus, we would expect employers to be willing to take strikes when they have powerful economic and political advantages at their disposal, such as unemployment, a

The State versus Managers

109

weak economy, stockpiled inventory, a sympathetic right-wing state, or a disorganized or divided labor force. In contrast, we would expect them to be unwilling to take a strike, and thus be wiling to make concessions when unemployment is low, the economy is strong, inventories are low, a social democratic state is in power, or the labor force is unified and homogeneous. What this simple formulation fails to take into account is that vulnerability to strikes depends not only on the absolute resource base of workers and managers, but on the capacity of employers to maintain revenue levels if sales become delayed. All a strike does is eliminate production and deliveries for a certain period of time. What matters is how much revenue the employer will obtain after the strike is over, and how this relates to the income flow he would have experienced with uninterrupted production. Consider the following scenario: A garment union can completely eliminate production for 2 months of a given year. Many clothing manufacturers face ironclad delivery deadlines from major customers; WalMart wants its spring clothing in January, not March. If the goods are not delivered in a timely fashion, the sale will be lost, and there is no alternative way that the employer can regain those revenues. Under these circumstances, if workers strike in early January, they have an effective threat, and management will be strike aversive. Now imagine that the company produces commercial aircraft. Until recently, this has been a relatively oligopolistic market where demand frequently outstripped supply. Airlines were used to waiting years for the delivery of jetliners, whereas aircraft manufactures usually sold every plane they made. A 2-month delay would inconvenience consumers, and the company would have lost the profits on 2 months of production and sales. The 2 months of lost production, however, would have made jetliners relatively scarce. Because the market price for new airplanes would thus be higher, some of the revenues lost during a strike would be recouped on later sales. In the French coal mining case, the capacity to recoup strike losses was even greater than that suggested in the hypothetical airline example. This was the conclusion that was drawn by the engineers of the French Mine Service after a comprehensive analysis of the finances of the coal mines in Moselle after the General Strike of 1923.^ If worker strength is what is required to produce a strike that can economically hurt management, then the February 1923 General Strike 'All of the following statistics in this section, unless otherwise noted, come from the French Mine Service's Report on the Mining Industry, 1923 (Moselle 10 S 1).

110

Chapter 6

should have crippled the companies of Moselle. France was still in the throes of the postwar energy crisis. Coal was scarce, and coal prices were high. Coal production and coal mining employment were soaring. The general strike in coal threatened to cut off the national energy supply just when consumption was rising (SIM 1922, 1923). The strike was internationally coordinated to coincide with a strike in the Saar and a slowdown in the Ruhr. Because of this, at the same time that domestic supplies would be scarce, German imports would be restricted by parallel labor actions. In some regions, the strike foundered on problems of low turnout. In Pas-de-Calais, Tarn, and Aveyron, fewer than 3% of the miners went on strike (AN F7 13791). In the Moselle, however, turnout was excellent, as the local independent union and the communists went out in unison.2 Moselle achieved 100% turnout by the second day. Fifteen days into the strike, it still led the nation with a 90% strike participation rate (Moselle 10 S 1). Although most of the nation struck for only 2 weeks. Moselle stayed out for more than 2 months, with good rates of turnout until the final days (AN F7 13795). Even so, the strike lost. The issue was not that the strike was undercut by coal shipments from nonstriking regions in France. The French Mine Service estimated that during the strike-year only 50 tons, less than 1% of Moselle's consumption, entered the province from other French coalfields. Nor was the issue that the strike had proven itself incapable of depriving Moselle of coal. The French Mine Service estimated that fully two thirds of the lost production of the strike was never made up in any way, representing an overall loss in regional consumption. The workers' critical weakness was an ample supply of substitute labor. Because of high unemployment in nearby Germany and the Saar, it was possible for the coal companies to fire fully 2,000 workers for participating in the strike, replace them all with new employees, and hire an additional 2,250 workers to cover general expansion. Of all workers hired in 1923, 64% were foreigners; nearly three quarters of these came from Germany or the Saar. Faced with a systematic campaign of firing that the workers were helpless to prevent, the unions had no choice but to go back to work on the unfavorable terms offered by management. The precise manner in which the workers lost had important impli^Although both unions went out together, the joint decision-making process was not altogether harmonious. The communists started striking a few days before the jointly agreed-upon deadline, while the independent union was still engaged in seemingly fruitful negotiations (AN F7 13793).

The State versus Managers

111

cations for the cost consequences of the strike. This was not a strike in which the workers had proven themselves incapable of economically damaging management and the local economy. The unions had effectively reduced coal sales and coal revenues for the period of the strike, while producing a collateral loss of energy consumption in Moselle as a whole. This was more of a lose-lose situation where labor inflicted costs on management, who in turn inflicted far greater costs on the workers themselves. What is remarkable is how quickly management recovered from the damages imposed on it by the strike. The French Mine Service argued that the immediate effect of the strike was to create a critical shortage of combustibles in the Lorraine. This meant that the mines would be able sell all of the coal they could possibly produce in the near future. Furthermore, prices and profit margins jumped significantly. In 1923, the price of ordinary coal increased 13% between January and April, when the strike ended. The price rose another 7% between April and December because of exogenous circumstances. A 13% increase in coal prices is very substantial, being roughly similar to what is observed in a coal basin moving from recession to mild expansion. The Mine Service attributed the entire 13% spring increase to the effect of the strike. Because of the increased sales at higher prices, 1923 was an extremely profitable year for the coal companies. The preceding analysis implies that the companies were able to make up for the production that was lost during the strike. This can be confirmed with data on productivity and absenteeism. One would expect in a year in which there was no production for 2 months that output per worker would go down. In most Moselle mines, however, the output per worker ratio for 1923 is virtually identical to that of 1922; at Sarreet-Moselle, one of the major coal companies in the department, output per worker actually rose by a small amount. The robustness of productivity can be explained to some extent by the technological improvements that occurred throughout French coal mining. The Lorraine mines were second only to Pas-de-Calais in their technological sophistication; in some aspects, such as fire control, they were the best in the nation. Productivity would have clearly risen as a result of increased use of pneumatic drills, jackhammers, conveyor belts, and hydraulic backfilling (Comite Central des Houilleres Frangaises 1935; Hardy-Hemery 1970, 1984; Reid 1985a; Moutet 1986). However, an additional component of the productivity increase came from overtime. If a coal firm wanted to increase output to take advantage of a fast-breaking sales opportunity, it simply asked its workers to work extra days. In 1923, the average Moselle worker put in 71 full

112

Chapter 6

days of overtime. This figure, over 2 months of overtime labor, becomes even more remarkable when one considers it was concentrated in the 10 months in which the mines were not on strike. This works out to 1.6 full shifts of supplementary work every week; because the normal workweek was 6 days, this implies working 7 days a week, with double shifts on some days. The strike only lasted 46 calendar days. If one excludes Sundays, the companies thus lost only 40 workdays during the strike. The 71 days of overtime that were worked covered for an additional 31 days above and beyond the downtime experienced during the strike itself. The massive increase in overtime undoubtedly made life miserable for the workers. Such intensification of labor, though, was not the only method available for increasing production. A more common solution to the problem of increasing output is to hire more workers, and the Moselle coal companies did this, too. During 1923, the number of workers employed in coal mining in Moselle went u p by 8.5%, a fairly substantial amount. There is an additional, somewhat morbid point to be added. Not all of the time that was lost in the strike represented productive labor. Coal mining generates a great amount of involuntary idleness. Some of this comes from technological mishaps such as floodings or fires, whereas other shifts are lost to absenteeism brought on by illness or injury. The strike actually had the perverse benefit of eliminating 2 months' worth of downtime attributable to shaft failures and injuries—the two months in which the mines were closed. The number of shafts lost to these causes was 29.8 per worker in 1922 and only 22.7 per worker in 1923. The Moselle coal companies were thus largely able to avoid suffering any financial losses from the General Strike. This is despite the fact that the turnout for the strike was large, the duration of the strike was more than 2 months, and coal was in critically short supply. It may be argued that the circumstance that allowed the companies to recoup their losses was an extraordinary market. Because of the strong demand for coal, they were able to make u p all of the lost volume at a premium price. During recessionary periods, the coal companies presumably would have been unable to sell unlimited amounts of coal, and the logic presented here would not apply. In a recession, however, workers have very limited capacity to hurt companies with strikes. Companies in such times are sitting on large stockpiles; there are few pressing deadlines for delivery. In recessions, strikes merely provide an opportunity for managers not to pay workers to produce goods that are already in excess supply (Rees 1954). Traditional business cycle theories of strikes have given us compelling reasons to believe that workers only have an effective strike threat against

The State versus Managers

113

management during upturns in the economy (Ashenfelter and Johnston 1969). Yet coal companies were also protected in upturns because of their ability to recoup both sales and revenues. The above argument should not be construed as a claim that strikes can never hurt employers. The invulnerability that has been suggested applies only to firms that will not lose money if delivery dates are retarded. Most firms are, in fact, subject to deadline pressure. In the garment example cited above, if the company cannot deliver its goods to the department store on time, the department store will obtain goods from other vendors in the present season and seek more reliable suppliers in the future. The cost of not making the deadline thus is a very significant loss of revenues. The threat that makes a strike effective is that other firms will obtain the sales that otherwise would have gone to the firm on strike. The threat of lost sales is a realistic one for most firms, but coal companies were somewhat protected from this. When an entire region went out on strike, as was the general case in the great departmental and national general strikes that characterized the period, alternative supplies could only be obtained from other regions. For the reasons discussed in the next section—nonsubstitutability of grades and inflexibilities in transport—it was not always possible to bring in large volumes of coal from distant regions. Thus, in a general strike, individual firms were somewhat protected from other firms stealing their markets. It is worth noting, somewhat parenthetically, that this would not have applied if unions had struck individual companies rather than entire minefields. If the workers of the Petite-Rosselle firm had gone out, and management was watching the Lorraine being provided for by its archrival firm, Sarre-et-Moselle, this would have been a strong incentive for it to settle quickly on favorable terms. Small one-firm strikes would not have been effective at pressuring the state for political concessions; they certainly would not have been effective at bringing about socialist revolution. For narrower economistic aims, however, small might have been better. Even so, for whatever reason, the miners struck across broad regions. This bought the companies freedom from strike costs; it also added considerably to the strike costs incurred by the state by raising both collateral unemployment and the expenses of policing.

GOVERNMENTAL AVERSION TO STRIKES Few sociologists or labor historians would seriously dispute the claim that governments in most liberal democracies have worked toward

114

Chapter 6

reducing strike activity. They vary in their success; Britain before Thatcher was arguably less successful in this regard than West Germany or Sweden. Few advanced capitalist governments, however, openly encourage stoppages or take no actions whatsoever toward bringing ongoing strikes to a conclusion. There is also little controversy as to why the state takes these actions. Strikes reduce national accumulation, although there is some debate concerning the magnitude of these effects. Strikes also are disruptions of social order that force public expenditures on various forms of social control. Regardless of the extent to which the public is actually inconvenienced, the state finds itself forced to provide for additional police protection and to absorb the costs associated with this repression. Most of these standard considerations affected French coal. In the late nineteenth and early twentieth centuries, coal remained the primary source of energy despite the growing importance of electricity and gasoline (Lafitte-Laplace 1933; Schneider 1945). A serious interruption of supplies could significantly disrupt industrial production. A number of conditions had to exist for a coal strike to produce significant collateral unemployment in other firms: It had to be large enough to cut off a significant proportion of the national energy supply; it had to be sufficiently sudden that customers could not plan for the strike in advance and stockpile enough coal to last them through the conflict; and importers and foreign suppliers had to be constrained in some way from providing alternative sources of energy. Put this way, it might seem that very few coal strikes would have a significant national impact. Most large national coal strikes occurred after extensive warnings. Walkouts would occur at contract expiration dates that were known months or years in advance; general strikes would occur weeks or even months after formal strike votes at national union conventions. Few of the major stoppages ever came as complete surprises. Admittedly, local strikes could break out without warning, but these were rarely large enough to affect a significant percentage of a region's coal supply. Furthermore, there were many willing foreign suppliers of coal: Britain, Germany, Belgium, and to a lesser degree Poland were all eager to increase their share of the French market (Comite Central des Houilleres Frangaises 1925; France. Administration des Mines 1937). Yet coal strikes empirically did produce collateral unemployment. Some problems stemmed from the nonsubstitutability among different grades of coal. Dozens of different types of coal were sold on the French market. Users of low-grade coal, such as residential consumers, could be relatively unfussy about the technical properties of what they bought.

The State versus Managers

115

Industrial consumers, and in particular transportation companies and metallurgists, had to be more selective. Different combustion processes required coals of very specific hardnesses and chemical compositions. Arranging for a substitute supplier for a specific coal was more difficult than finding a supplier for coal in general.3 Furthermore, coal tends to degenerate with handling, transport, and outdoor storage, and some applications require fresh, recently mined, locally produced product. Employers with these needs have less capacity to stockpile and to purchase imported goods (Schneider 1945). Lastly, transport difficulties often retarded shifts in the source of supply of coal. Most railway companies have little problem sending fully loaded foreign cars to any destination; in contrast, they incur significant costs sending empty cars back to the source of supply for refilling. Because coal mines send out a massive volume of bulk goods but purchase a significantly smaller volume of industrial products, they are always dependent on railway companies' willingness to deliver fresh supplies of empty rolling stock. Rolling stock shortages are particularly likely to occur when the return trip involves crossing international boundaries or the lines of competing freight companies. There is a constant temptation to load the empty cars with local freight and send them on a paying mission, rather than ship the empties back to another company's source of supply. Most rail companies resolve these difficulties with elaborate negotiated agreements on return arrangements. A major geographical shift in the coal market can suddenly create a huge redirection of rolling stock that is frequently incompatible with the return arrangements that exist between the new regions (Oliver 1922; PoUins 1971). Although these problems can usually be worked out over time, transport problems can impede the short-term substitution of foreign supplies in the event of an industrial stoppage. France was probably the most vulnerable to coal strikes in the years immediately following World War I. This was a unique situation resulting from an acute and unprecedented energy shortage. In 1920, France was only producing five eighths as much coal as it had produced in 1913; coal was so scarce it was rationed. The crisis had been created by the bad effects of World War I, in which the Germans had invaded the northern coal basin. During the war, most of the limited product of these mines went to Germany. At the end of the war, when they were being forcibly evicted by the allies, the Germans systematically destroyed the northern coalfields, dynamiting equipment, flooding shafts, and starting fires. ^This last factor should not be underestimated since railroads, the merchant marine, and steel companies represented three of the largest consumers of French coal.

116

Chapter 6

The remaining southern fields suffered from the manpower shortage brought on by mass mobilization and the rolling stock shortage created by the urgency of moving men and materiel to the front. The French increased their imports as best they could to make up for their productive shortfall. Britain and Germany, however, had their own constraints both on productive capacity and willingness to export. Coal was thus absolutely scarce (Oliver 1922; Lafitte-Laplace 1933; Comite Central des Houilleres Frangaises 1921b; Georges 1921). It is not surprising that in the middle of an energy crisis, the French state would be concerned about collateral unemployment and would take great steps to avoid a national coal strike (PC 23 J: Cahier des Greves, Cie. de Bethune; PC M 2382). Coal strikes, however, could shut down other employers under less dire circumstances as well. In the Loire, there was collateral unemployment in the general strikes of 1902 and 1906, although not in the Loire-only strike of 1900 (Loire 92 M 110; Loire 92 M 114; Loire 92 M 138; Loire 92 M 99). Some regional coal strikes could also produce significant damage to other firms. The best analysis of metallurgists' vulnerability to coal strikes was done by the Loire prefect in 1908. Nineteen hundred and eight was a good year for both coal and metal. The union used the cyclical upturn to press for an improvement in salaries and the length of the working day. No strike actually occurred in this offensive. Nevertheless, during the negotiations, the prefect surveyed 28 metal producers in the department to assess their vulnerability to a stoppage. Table 6.1 shows the result of this survey. Half of the metal firms had no more than a 2-week supply. Four firms were already out of coal and experiencing energy-related unemployment, suggesting the presence of Table 6.1 Coal Supplies of Loire Metal Firms, January 1908 Coal Supply

Number

Out of coal now 1 week or less 8 days-2 weeks 3 weeks-1 month More than 1 month Can substitute electricity

4 5 5 5 2 7

Total Source: Loire 93 M 23.

28

The State versus Managers

117

additional impediments to the obtaining of coal. The historical record does not elaborate on what these may have been, but rolling stock shortages or non-strike-related delays in production may have been responsible. Another five firms had 1 day to 1 week's worth of coal stockpiled. A coal strike would not have paralyzed the entire region; some firms could switch over to electricity, and one public firm, the Manufacture Nationale d'Armes de Guerre, was sitting on a 6-month stockpile. In contrast, more than 30% of the metal firms in the Loire had only a few days' worth of coal or less. These firms would have been highly vulnerable to a coal strike. As shall be seen in the next chapter, the government settled the dispute peacefully on very advantageous terms for the miners. The threat of collateral unemployment may have had something to do with the favorable quality of the government's intervention. The state was also motivated to avoid the social disruptions associated with coal strikes. The government usually responded to coal strikes with full military protection for the mines, with troops invariably being deployed at various locations around company facilities (Stearns 1968; Perrot 1974; PC M 2383). Despite the ominous overtones of confronting strikers with armed force, the goals of the military were relatively benign. Perrot (1974) has accurately characterized the policing of French strikes in the late nineteenth century as being oriented less toward repressing strikers for the act of striking than toward the containment of assaults against persons and property during otherwise legitimate protests. Trempe (1971) supports this in her history of the mines at Carmaux with substantial documentary evidence that troops were instructed to not interfere with the strike in any way unless a real threat to public safety emerged. Public safety, however, was generally defined as respect for the right to work. Strikers whose harassment of strikebreakers extended beyond verbal abuse could expect some form of police retribution. Coal strikers in France frequently did not play by the rules. The most famous example of labor violence in French history occurred during a coal strike, namely, the defenestration of Watrin in Decazeville in 1886. Watrin was a highly unpopular administrator with the selfprofessed goal of reducing workers' wages and increasing their work effort. During the strike, a mob of workers stormed his second-story office and threw him out the window. He died of injuries sustained from the fall (Reid 1985a). During the General Strike of 1902 in the Loire, many towns were turned into war zones. There were several dynamite blasts, and many more arrests for possession of explosives. A typical example was an incident in Ricamarie, where a foreman's house was dynamited during the night; fortunately, no one was at home at the

118

Chapter 6

time. Shots were fired through the windows of a nonunion carpenter. The dumping of coal trucks was widespread. The home of the director of the Houilleres Montrambert et de la Beraudiere was assaulted, although the attackers were rapidly dispersed by troops (Loire 92 M 110; SG 1902). Even the relatively staid northern region could turn violent. A general strike in Pas-de-Calais and Nord in 1893 produced no fewer than four dynamite attacks. In all, 217 workers were convicted for attacking either strikebreakers or the troops protecting them (SG 1893). This, however, was an extraordinarily disorderly strike for the region; it was one of the few cases where the reformist union voluntarily turned down state arbitration and experimented with breaking the rules. Most northern strikes were relatively nonviolent until the rise of communist provocations in the 1920s and 1930s. A major area of concern was providing maintenance to the mines during strikes. Most mines required regular pumping and drainage, without which the shafts would become flooded and unusable. Others required special ventilation or fire-fighting procedures. In most strikes, management and labor realized that they had a common interest in ensuring basic preservation of the shafts. The usual procedure would be for the strike committee to take responsibility for organizing and staffing caretaker crews (Trempe 1971; SG 1893; SG 1913); There were exceptions, however, such as in Albi in 1902 or in Cransac in 1913, where the strikers would attempt to cut off basic maintenance altogether (SG 1902, SG 1913). This would increase the urgency for management of running strikebreakers past the picket lines, increasing both the likelihood of violent confrontations and the potential economic damage of the strike. In the postwar period, communist strikes were generally more difficult to police than reformist strikes. Although violence was relatively restrained in the early communist strikes, by the late 1920s, the unitaires had become great advocates of chasse des renards: literally, fox hunting, but substantively, beating u p strikebreakers. In the Gard strike of 1929, a mob of 150 bludgeon-wielding workers pursued and harassed nonstrikers, who required protection from gendarmes. There were several successful attacks on strikebreakers, as well as a number of rockthrowing attacks on the military (AN F 13905). In the 1925 Villeboeuf strike in the Loire, the unitaires actually organized a campaign of chasse des renards in a public meeting, with the customary police observer sitting in the hall taking notes on the proceedings (Loire M Sup 502). In a 1935 strike at the Houilleres de Saint-Etienne, the communists not only attempted to cut off all maintenance work but organized patrols that went to the individual houses of all the firemen and pump operators, threatening those who crossed the picket lines (Saint-Etienne 6 F 18).

The State versus Managers

119

Compared to communist strikes, reformist strikes were relatively cost free. Both official and wildcat reformist strikers confined themselves to peaceful demonstrations and negotiations. This did not, however, make the government indifferent to the strike activity of reformists. Communists were perpetually trying to create a front unique of joint reformist-communist militancy. The reformists did join with the communists on 17 stoppages; when such actions were not possible, communists organized parallel, more disorderly demonstrations featuring speakers who condemned confedere leaders. In 1931, in Lens and Bruay, communists flooded reformist antiunemployment rallies, demanded the right to speak, and after being refused, started a parade and marched off to different locations for their own rallies (Pas-de-Calais M 2383). There was a third, more minor reason why the state opposed coal strikes. Such strikes could often turn into political embarassments for the government in power. There was a concern that middle-class voters would be frustrated by large numbers of industrial stoppages and would turn toward more extremist right-wing solutions. There was also a concern that the left would use strikes as a forum for its pohtical agendas and would cause defections from moderate reformers such as the socalled radical socialists. Some readers (I would include myself in these ranks) might question the extent to which Third Republic French politicians were vulnerable to the public opinion fallout from strikes. The prefects who administered wage negotiations, however, expressed concern about precisely these issues. In 1921, the government stalled a set of bargaining sessions in the Nord until after the completion of a strike in Lille textiles to minimize the political damage of having simultaneous strikes in one department (Pas-de-Calais M 2382). In 1925, electoral considerations were listed as a rationale for speeding up the conclusion of wage negotiations in the northern basin (Pas-de-Calais M 2375). Later in 1932, the prefect of Pas-de-Calais wrote a long letter arguing that a communist strike could have serious repercussions on upcoming legislative elections (Pas-de-Calais M 2383). Note that although all of these stories come from Pas-de-Calais, they involved three separate sets of ministers of the interior and three separate prefects. There seems to have been an institutional preoccupation that induced these authorities to limit strikes with whatever means were in their power.

The reader has seen how the firms and the state would have had divergent interests in minimizing industrial conflicts. Let us now turn our attention to the issue of wages. Management's preferences are so simple that they do not warrant a section-length discussion: Firms

120

Chapter 6

wanted lower wages. Lower wages meant lower costs of production, which at any given selling price meant higher profits. The wage preferences of the state, though, are more difficult to specify. There are many reasons (listed in detail below) why the state would have shared management's preferences for a cheap labor force. There were a number of opposite considerations, however, that gave the government incentives for raising rather than lowering wages. Public authorities' ambivalence about wage rates created the possibility of labor control policies where the state would prevent strikes at any cost, even if it meant significantly higher expenses for management.

FACTORS PREDISPOSING THE STATE TO PREFER LOW WAGES There were a number of factors that should have made the French state strongly prefer low wages. Had these been the only considerations taken into account by government policymakers, then the states' interests would have been virtually identical to those of the capitalists, and their negotiating stances would have been very similar. Coal as a Source of Tax Revenue Fred Block (1977) has argued that the interests of state bureaucrats and capitalists are generally similar because the state depends for its revenues on taxes, which are in turn dependent on economic growth. This argument is generally plausible for modern governments as a whole, and it is certainly applicable to the French state in regard to coal mining. The state technically owned all coal lands. Under the law of 1810, these lands were granted to private capitalists in perpetuity; in 1919, the law was amended to return ultimate title back to the state. In both cases, ownership could be stripped from the proprietor under a variety of conditions, one of which was nonpayment of financial obligations (Comite Central des Houilleres Frangaises 1921a). The French government imposed three direct charges on coal operators: a tax on the acreage of land covered by the concession, a tax on the profits of the coal company, and in various periods, a pithead tax on the coal as it was extracted. In 1931, approximately 10% of the final cost of production represented tax obligations. In part because it was a poor year for French coal mines, the share of the revenues of Pas-de-Calais coal companies received by the government in 1931 was nearly 50% greater than those received as dividends by shareholders (Lafitte-Laplace

The State versus Managers

121

1933). Note that these figures were prepared by the coal companies themselves and, as such, may have been overstatements. In other respects, however, these figures underestimate the dependence of the French state on coal revenues. Coal was also one of the most important bulk goods carried by the French railways and by the inland marine. Both of these enterprises were taxed heavily as well, providing a second fiscal interest of the government in maintaining coal production. Low wages would not have increased the state's receipts from the acreage tax, but the receipts from the profit and pithead taxes would have grown. Any factor that reduced the price of producing coal increased both corporate profitability and sales volume. State Ownership of the Saar Mines The state's proprietary interest in French coal mining increased dramatically at the end of World War I. As part of the Versailles settlement, the French government obtained the legal control over the mines of the Saar, which had been under German public control since the eighteenth century. The French were thus technically acquiring an asset of the Wilhelmine empire. The acquisition of the mines of the Saar moved the French state from being a coal rentier, collecting taxes from firms administered by private managers, to being an actual coal operator. The dominial mines were administered by French government officials working in close conjunction with the Ministry of Foreign Affairs, the Ministry of Public Works, and the Ministry of Labor. They were run as a profit-maximizing endeavor, with the proceeds representing a form of German reparations. Running the dominial mines gave the government the same labor relations problems experienced by the private sector. The 1923 General Strike, which set the entire tone for relationships between the unitaires and the government, began as a communist strike in the Saar (AN F7 13795). Most of the wage campaigns of the late 1920s and early 1930s also began in the Saar, with the pattern of the Saar settlement being frequently used as a basis for agreements (L 15 J 2449; L 15 J 2444; PC M 2382). Running the mines also gave the government the right to participate in the national employers' federation. Every important discussion by the Comite Central des Houilleres Frangaises included leading representatives from every coal basin; the Saar representative thus attended every wage, pricing, and marketing meeting after World War I. The impor-

122

Chapter 6

tance of this cannot be overestimated, because the CCH really did determine global managerial responses to union and regulatory pressures. On one hand, the inclusion of the Saar representative undercut any attempt by the owners to use the employers' confederation as a vehicle for opposing government intentions concerning wages and prices. On the other hand, such active collaboration would have coopted the Saar delegation as well, further inducing the government to follow private sector initiatives on union issues (L 15 J 2449; L 15 J 2444; PC M 2382). Energy and Economic Policy From 1890 to 1935, coal was the primary source of energy for France. It was not only the basic source of power for industrial production and residential heating, but the only source of power for mass transportation. Without coal, the railways would not run, and the merchant marine would be grounded. As such, maintaining a cheap supply of coal was important for national economic development. Part of the problem was simply assuring that there was enough coal to meet the needs of the economy. France has always consumed more coal than it could produce, with imports tending to represent about 30% of total consumption. Many government policies were oriented simply toward increasing production by opening mines or providing subsidies for capital investment. This was particularly an issue after World War I, when the Germans destroyed a significant proportion of the productive capacity of the northern basin. The government provided extensive, although arguably inadequate, funding to rebuild the mines (Comite Central des Houilleres Frangaises 1921b, 1925; Oliver 1922; HardyHemery 1984; Kuisel 1981). The state, however, was also concerned about the cost of production. Because the price of coal was an important determinant of the price of other industrial goods and of the overall cost of living, whenever the control of inflation became important to national economic planning, pressure was put on coal companies to lower their production costs. These preferences were made particularly explicit in the early 1920s during the financial crises that preceded Poincare's 1926 reevaluation of the franc. In 1921, 1922, 1924, and 1925, the government met with both management and labor ordering policies of wage and price restraint in the interest of reducing the overall rate of inflation (AN F 13791; M 318 M 20). Low coal prices would have also been important to any defense of the franc or of the national balance of trade. The more expensive coal

The State versus Managers

123

became, the more subject the French economy was to energy imports from the British and the Germans.* Foreign Policy Considerations Coal was a critical military supply as well as an economic resource. The mobility of soldiers depended on rail transport, which in turn depended on energy for locomotives. Furthermore, coal was required for the manufacture of the increasingly heavy armaments needed for warfare, including dreadnoughts, artillery, and eventually aircraft and tanks. The rivalry between France and Germany derived in part from territorial disputes over the coalfields of the Lorraine and the Ruhr. Disputes over German coal reparation payments were one of the most acrimonious issues in the negotiations over the execution of the Versailles treaty. Nonshipments of coal led to the unsuccessful French seizure of the Ruhr in 1923 (Brogan 1967; Bonnefous 1956; Boncour 1945). Coal was also a source of tension between Britain and France. Britain was the foremost supplier of foreign coal to France, supplying more coal to the city of Paris than did Nord and Pas-de-Calais combined. It was also the primary source of coal for all of western France (PC 15 J 2244). The quantity and price of the coal provided, however, was often an issue of dispute. During the coal shortages of the immediate postwar years, Britain was accused of withholding supplies from France and assessing foreign customers with surcharges far in excess of the cost of transport. Later, during the coal surpluses of the Great Depression, the British would be accused of dumping (Oliver 1922; Lafitte-Laplace 1933; Comite Central des Houilleres Frangaises 1935; France. Administration des Mines 1937). The tensions between France and her two neighbors would have been a strong incentives for the government to seek low domestic coal prices. Any factor that would have lowered the cost of production would

^Kuisel (1981) argues that before World War I, the French state was less likely to engage in economic planning. Certainly there would have been less pressure on coal companies to maintain low prices as part of either formal macroeconomic policy or any formal defense of the franc. Policies similar to those derived from economic planning, however, may have occurred through the activities of noncoal industrialists. The interests of these industrialists were taken into account during negotiations of coal mining wages. Employer federations, particularly in metalworking, were active lobbyists during the prewar years, and they would not have been neutral to high energy costs.

124

Chapter 6

have reduced the demand for imports and the concomitant dependence on these two hostile suppliers.^ Direct Representation of Capitalist Interests In the early days of Marxist thinking about the state, governments were viewed as being agents of capitalist interests because of both the placement of prominent businessmen in important public offices and the effective lobbying of state policymakers by important business organizations (Miliband 1968; Domhoff 1983). More recent developments in the theory of the state have rightly emphasized the state's dependence on accumulation for revenues, the role of working-class conflict, and the autonomy of many state sectors (Block 1977; Poulantzas 1973; Skocpol 1980). That said, however, there are many cases for which the old instrumentalist perspectives still provide insight. One of these is French coal mining. In the nineteenth century, the coal companies really did control political offices as their own personal fiefdoms. In 1891 in Pas-de-Calais and Nord, no fewer than two senate seats and five deputy seats were held by high-level administrators of coal companies (Gillet 1973). There was a comparable domination of parliamentary and municipal offices by company spokesmen in Carmaux, Montceau-les-Mines, and Decazeville (Trempe 1968; Beaubernard 1981; Reid 1985a). This, however, changed rapidly in the 1890s and early 1900s. The Socialist party, working closely in conjunction with the coal miners' union, mounted effective challenges to managerial candidates in local and national elections. By the middle 1890s, the companies were no longer even putting u p candidates against popular prounion parliamentarians in Pas-de-Calais. By 1905, municipal offices in many coal mining towns were held by Socialists. At the national level, coal arrondissements were being represented by such prominent socialists as Jaures, Basly, and Lamendin (Trempe 1968; Faur6 1956; Gillet 1973). The decline in managerial office holding was somewhat compensated for by a simultaneous increase in the development of formal associations of interest representation. During the 1890-1905 period, very ^There is an additional mechanism that would have linked foreign policy to hostility to labor, if not to low wages per se. The unitaires were fiercely antimilitarist opponents of French foreign policy. They vigorously protested French military intervention in the Ruhr and Morocco. They coordinated strikes across national boundaries; the 1923 General Strike was co-planned with a strike in the Saar and a slowdown in the Ruhr, and the 1926 strike was in sympathy with an English general strike. Conservatives generally accused the communists of being German collaborators. As a result, communists faced tremendous hostility from the Ministry of Foreign Affairs (AN F7 13795).

The State versus Managers

125

strong employers' associations developed in French coal mining. There were a number of false starts caused by nonparticipation by principal players and internal disagreements over the purpose and powers of these agencies; however, by the middle of the first decade of the twentieth century, employers had developed a strong national organization, the Comite Central des Houilleres Frangaises, and equally strong constituent organizations at the regional level (Duchemin 1940; Gillet 1973; LeFranc 1976; Friedman 1991). The importance of the CCH cannot be overestimated, particularly for the period after World War I. Nearly every government file dealing with the state and coal mining in the national or departmental archives after the war features substantial correspondence with the CCH or one of its regional affiliates. In the case of labor relations, such communications were a product of contemporary procedures of industrial conciliation. The prefect and the ministers of labor and public works formally consulted with both managerial and union representatives before the expiration of every wage agreement. In practice, although not technically de jure, the CCH became the official organ of management in national salary negotiations, because its membership was virtually identical to the composition of the company negotiating delegations. The CCH's role, however, was not narrowly limited to labor matters but extended to policy concerning tariffs, freight rates, marketing arrangements, and other aspects of energy policy.^ An instrumentalist could rightly argue that management would have been far better served to have its own representatives actually holding political office, rather than being dependent on lobbyists and consultation with policymakers. This is undoubtedly true. In many cases, however, the CCH was relatively effective at convincing policy influentials of the merits of its case. Harland Frechel (1990) has suggested that industrialists carry greater influence with the state during cyclical downturns, when their need for assistance seems to be externally validated. This observation seems to apply to French coal mining. The 1925 and 1931 wage negotiations occurred during recessions. Both times the government seems to have internalized management's position in toto (AN F 7 13793; PC M 2381; Loire 15 J 2449). Even during good years, CCH opinions were rarely disregarded. This influence, combined

^Listing every archival reference to the Comite Central des Houilleres Frangaises would be tedious. A detailed portrait of the broad scope of employers' association activities can be obtained from the Loire archives, which possess a fine collection of documents from the Comite des Houilleres de la Loire. These are located in Loire 15 J at various numbers in the ranges of the 2200s, 2400s, and 2500s.

126

Chapter 6

with the state's own numerous motivations for keeping coal prices low, provided multiple incentives for government policymakers to espouse low wages for coal miners.

FACTORS PREDISPOSING THE STATE TO PREFER HIGH WAGES In many respects, the French state acted as an employer's proxy. There were a number of factors, however, that weakened this agency relationship and made the state relatively tolerant of high wages. Working-Class Influence in the Chamber of Deputies Just as managers were able to defend their interests by making direct representations to government advocates, workers were able to defend their interests by making their own representations. Coal mining unions had some advantages in this regard, because after the turn of the century, they were largely successful in electing left-wing parliamentarians to office from coal mining districts. Of particular importance were Basly and Lamendin, the two highest-ranking officers in the coal mining union. Both were elected to Pas-de-Calais seats in 1893, which they then held without serious challenge until retiring in the 1920s. Known as the "miner-deputies," for 30 years they were the driving force behind every major piece of reform legislation for coal miners; during their tenure, the miners won numerous substantive improvements in workmen's compensation, pensions, and the length of the working day (Michel 1974; Trempe 1968; 1971). The coal miners could count on a lot more support than just the deputies from their particular districts. The entire parliamentary left was highly sympathetic to workers' interests; coal miners, who were viewed as the most oppressed and exploited of proletarians, were the subject of special attention and benevolence. The miners could also count on significant support from other parties. During the prewar years, a substantial proportion of the parliamentary center believed in solidarisme, the creation of overall societal improvement through the amelioration of conditions for the working class. This stance was common among petit bourgeois politicians, particularly those in the radical or socialist-radical blocs. Centrist reformers were frequently probusiness, antitax, and laissez-faire in their outlooks on general economic policy. But they did favor government conciliation of strikes. They also favored making concessions to workers as a strategy for eliminating grievances and producing industrial peace (Mayeur and

The State versus Managers

127

Reberioux 1984; Stone 1985). Even the far right could occasionally side with the unions on reform issues; Boulangerists, the Action Frangaise, and French fascists all made desultory attempts at garnering workingclass support (Irvine 1989; Mazgaj 1979; Sternhell 1986). Table 6.2 provides a very gross attempt to estimate the size of the workers' bloc in the Chamber of Deputies. The left column shows the percentage of seats held by leftists, and the right column shows the combined percentage of seats held by leftists and centrists. The quality of these estimates, though, is extremely low. These data are not suitable for multivariate analysis because of numerous systematic and random errors. The size of the parliamentary blocs was estimated by taking the results of every election between 1889 and 1936 and dividing the seats into left, right, and center, using information on the party that won the seat. Most industrial democracies have relatively well-defined parties with articulated distinctive platforms. As such, calculating the percentage of seats held by the British Labour party or the Swedish Social Democrats gives a meaningful estimate of the size of the parliamentary left. Table 6.2 Estimates of the Percentage of Seats in the Chamber of Deputies Held by Leftist Parties

Date of election

Socialist bloc only

Socialist and radical blocs

1889 1893 1898 1902 1906 1910 1914 1919 1924 1928 1932 1936

0 5.4 9.6 7.9 12.3 17.3 17.6 11.1 28.0 20.2 25.5 36.6

0 30.7 24.0 41.2 54.0 42.3 45.6 20.1 51.4 46.0 57.5 64.0

Source: Campbell 1958. Note: Socialist bloc includes Socialists, Communists, Revolutionary Socialists, Independent Socialists, Left-Wing Socialists, and Communist Socialists. Radical bloc includes Radicals, Socialist Radicals, Independent Radicals, and Socialist Republicans. So-called Right-wing Radicals and Independent Radicals are excluded. The Cartel des Gauches in 1924 was arbitrarily divided into half socialists and half radicals. This was a one-election socialist-radical joint slate.

128

Chapter 6

This is not a practical possibility for French data. French parties frequently were not associated with coherent ideological positions. Parties were reconfigured at virtually every election. Many parties lasted only one or two elections, despite featuring politicians of national prominence. Candidates chose their parties not on the basis of ideology but on the basis of electoral strategy. In many years candidates got elected by being on a winning slate.'' As such, great effort was expended mixing and matching candidates on the basis of personality—rather than platform—to derive a slate that would get most of them elected (Campbell 1958).^ There was no party discipline in chamber votes. Furthermore, there was little consistency in the actions of politicians over time. Millerand, Waldeck-Rousseau, and Clemenceau are all prominent national figures whose politics underwent major shifts to the right. Jaures arguably can be viewed as a petit bourgeous reformer who shifted from centrist to left. He had his beginnings as a radical, not a socialist (Bonnefous 1956; LeFranc 1963; Zevaes 1947). Table 6.2 shows on the left the number of seats held by socialists and communists. On the right, the number of seats held by the radicals is added to this figure. The radical party in France was a moderate petit bourgeois party, not extremist as the name would suggest. There are many minor parties that have been added somewhat arbitrarily to one group rather than the other. Furthermore, in 1924, the socialists and radicals formed a center leftist party known as the Cartel des Gauches. Those seats were divided 50-50 among the two parties for lack of a better strategy of decomposition. Although each individual number in the table is suspect, there is enough of a pattern in the data to suggest a few basic generalities. The size of the parliamentary left seems to have increased steadily during the period. Starting from a base of zero before 1890, the percentage of seats held by socialists and communists grew at a modest but constant rate. For most of the period, the miners could count on 10% to 20% of the deputies being from working-class parties, with this number being higher after World War I. If one adds the centrist parties to the bloc, the percentage of working-class supporters rises to between 30% and 50%.^ It is also worthwhile to note the dramatic drop in leftist representation in T h e r e was no universally applicable strategy for running elections in this period because the rules of balloting and allocating parhamentary seats changed extremely frequently. ^Consider the wonderful quote from Radical politician Paul Doumer, speaking during the election of 1902: "I am too well known to need to develop a programme" (Mayeur and Reberioux 1984, p. 221). ^Counting the radical bloc as labor supporters, however, may be more appropriate for the prewar rather than the postwar period.

The State versus Managers

129

1919 and the surge during the Popular Front election of 1936. These were extremely dramatic turning-point elections; the discontinuities shown here are legitimate and can be supported by parallel developments in French politics overall (Brogan 1967; Zevaes 1947; Gallie 1983). In general, there was a nontrivial base of union support among the deputies that could have been the basis for governmental willingness to tolerate high wages. Managerial Noncooperation with the Government A well-known feature of the French state at the turn of the century is that it frequently resolved labor conflicts by siding generously with the workers. Gerald Friedman (1990) has done a number of statistical analyses of French strikes showing that state intervention was the key to strike success during the syndicalist period. These findings are, of course, a more rigorous confirmation of earlier but cruder analyses presented in Shorter and Tilly's classic Strikes in France (1974). Such findings can be grounded in the politics of the period. Judith Stone (1985) and Sanford Elwitt (1986) have each written excellent books on bourgeois reform movements in middle Third Republic France that emphasize the growing concern by public policymakers about labor problems and the overall willingness of the French state to seek solutions through welfare-state legislation. The great centerpiece of the labor history of this era, Rolande Trempe's Mineurs de Carmaux (1971), deals explicitly and extensively with the favorable response of the republican state to the protests of miners against conservative managers (Trempe 1971). Stone and Elwitt provide sophisticated explanations of the reformism of the period in terms of the changing ideologies of a large number of political actors. The full exploration of the origins of these policies will not be attempted here. A gross but useful simplification of the accounts of Stone, Friedman, and Trempe is that the state sided with labor because management was rightly perceived as posing a threat both to the security of the republic and to industrial and social peace. A consistent theme of this period is managerial hostility to democracy and noncooperation with public industrial conciliation. Probably the most severe form of noncooperation was the continued managerial support through the nineteenth century for monarchist and Boulangerist opponents of republican democracy. The Third Republic lasted for 70 years and was only overthrown by the Nazi occupation of France. During the early years of the republic, however, the security of democracy was very tenuous. There were continual threats from

130

Chapter 6

Catholics, royalists, nationalists, and other conservatives favoring overthrow of the government and reestablishment of some form of autocracy. Of particular importance was the Boulanger affair of 1889. Boulanger was a conservative general who formed his own rightist party and won nearly 8% of the seats in the Chamber of Deputies on his first election. He subsequently was involved in revolutionary preparations that ended only at the last minute when he fled France rather than marching through Paris to seize government facilities. Some believe that only his tactical reluctance prevented the Third Republic from falling (Brogan 1967; Mayeur and Reberioux 1984; Irvine 1989). The Boulanger affair provoked a strong defensive counterreaction by supporters of the republic. Electoral laws were changed to neutralize some of Boulanger's stronger tactics (Campbell 1958). Furthermore, because Boulanger had been supported in some districts by working-class voters (although rarely in coal mining districts per se), the state began to build credibility among the lower classes by promoting general policies of social reform. Since many employers were archconservative royalists, pro-working-class policies provided a useful political buffer against the threat from this rightist elite (Friedman 1990). Employers were also obstacles to the implementation of state plans concerning labor conciliation. France, like many other capitalist states of the early twentieth century, sought to control industrial conflict by developing formal mechanisms of collective bargaining and public mediation (Sturmthal 1972; Saposs 1931; Clark 1930). Unions frequently supported these initiatives, despite nontrivial protests from anarchists and other revolutionaries (LeFranc 1967; Garmy 1934; Valois 1927; Michel 1974). Managers, though, resisted public conciliation vehemently. Their opposition was sustained, unified, and usually only broken by the threat of major governmental sanctions. As will be discussed in the next chapter, coal miners were far more cooperative with government peacekeeping initiatives than were their employers; as a result, public authorities frequently sided with the workers on the wage issues raised during these disputes. Welfare Functions of the State The 1890-1935 period saw a general expansion of the welfare state in France and throughout most of the developed world. Many early welfare-state programs took the form of governmentally mandated fringe benefits. By the 1930s, French law provided that coal employers fund disability relief, pensions, and family allowances (Villain 1918;

The State versus Managers

131

Mangez 1925; Lafitte-Laplace 1933); in 1935, legally required "social expenses" raised labor costs by a full 13.5% (France. Administration des Mines 1937). The French state had required since the royal edict of 1604 that mining companies make some provision for the medical needs of their workers. This concession was not derived from primitive class conflict; instead, it was an extension of traditional principles of French civil law. Mining is an extremely destructive activity that not only is dangerous to the workers, but ruins the use value of nearly everything in the immediate area. The works are unsightly; tunnels undercut the solidity of all preexisting structures above them; ground water gets diverted and polluted; and huge volumes of waste dirt are created that cover and foul large areas of previously arable land. The early mining acts specified that coal operators had to reimburse all parties for ancillary damages produced by the operation. The protection of the miners themselves was analogous to the protection of the general public provided for in these precapitalist regulations. The earliest disability and pension funds were voluntary responses to these requirements. They represented employer escrows of anticipated legal expenses and were often used to circumvent litigation and obtain quick settlements (Hatzfeld 1971). Throughout the nineteenth and early twentieth century, the requirements for disability and pension funds became more stringent and beneficial to the workers. Much of this was attributable to the political actions of the miners themselves. Pension reform was a major demand of many of the most important strikes of the period, including those of 1902, 1906, and 1912. This, combined with sustained legislative activity by the miner-deputies, produced substantial improvements (Trempe 1984). The miners' case was helped as well by large numbers of defaults by underfunded, mismanaged pension programs. Although the workers could rely on traditional civil sanctions for redress, the presence of conspicuous employer abuses strengthened the resolve of moderates to eliminate these disputes through proactive regulation (Hatzfeld 1971; Trempe 1984). There were a very large number of pension and disability reform bills, many of which affected administrative details as opposed to substantive benefits or reforms in the fundamental control of the programs; however, important ameliorations occurred in 1894, 1914, 1923, and 1930 (Saposs 1931; Lafitte-Laplace 1933). Family allocations, or bonuses for workers with multiple children, were required by law in 1930. In practice, because of informal government pressures, many coal companies had been paying these throughout the 1920s (Loire 15 J 2444, Hatzfeld 1971). There were many motivations for this policy, one of which was pronatalism. France suffered

132

Chapter 6

serious manpower losses during World War I; its capacity to defend itself against the Germans was limited by its smaller population and lower fertility (Gallie 1983). Coal operators would have been acutely sensitive to this concern, having seen the complete destruction of the northern coal basin during World War I (Georges 1921). There may have been economic motivations for family allowances as well. Workers with families were less prone to turnover than young unmarried celibates. The former would have been worth a premium to many employers during the 1920s, when labor was generally scarce and workers were leaving coal mining for construction and manufacturing (Lafitte-Laplace 1933). After 1930, however, family allocations became mandatory rather than voluntary. Disability funds and family allocations raised wages both directly and indirectly. Family allocations actually appear in the wage data. Pension and disability payments to workers not on the payroll are excluded from the figures; however, these funds were financed in part with "worker" contributions, namely deductions from salaries. Because real wages were negotiated on the basis of actual take-home pay, the workers' contributions were likely to have produced an increase in nominal salaries in order to maintain overall levels of substantive wages.^o Indirectly, these programs created a precedent within the government of public officials improving the welfare of miners by drawing on the funds of the companies. The prefects supervised the enforcement of the laws on disability funds, as well as the administrative arrangements concerning the local family allowance pools. Both matters generally involved some routine consultations with the local union. When these prefects came to supervise wage negotiations, working again with the unions, they would have had some experience in being representatives of workers' interests. A good illustration of the link between public welfare policy and wages can be found during the Great Depression in government attempts to reduce unemployment. Throughout the entire 1890-1935 period, prefects and other public authorities took vigorous efforts to protect workers from unemployment. In most years, unemployment was a negligible problem; the coal economy was booming for most of the prewar period and for a significant proportion of the 1920s. During the 1930s, however, unemployment became severe. In 1930 nearly 259,000 workers were employed by French coal companies, but by 1935 this number had '"There is generally a positive relationship between wages and fringe benefits, suggesting that workers do not subsidize nonwage benefits by accepting lower wages (Freeman and Medoff 1984).

The State versus Managers

133

fallen to 198,200. This number underestimates the hardship experienced by workers, because most miners retained their jobs but were only allowed to come to work 3 or 4 days a week. In October of 1933, at the relatively prosperous firm of Roche-la-Moliere-Firminy, workers had to take 8 days of unemployment out of a 28-day work month (Loire J 15 2294). Thus, miners who were still nominally employed had their disposable income slashed." The state's motivation for reducing the effects of unemployment was more political than financial. Technically, the state did pay for unemployment relief, but payments were only made to full-time workers who had totally lost their jobs and not to workers who were partially idled (Saposs 1931). Furthermore, most of the workers who were formally laid off were foreign workers, particularly Poles and Italians. The state had no financial obligations to foreign workers; even their transportation back home was paid for by the companies (Cross 1983). There was greater concern that mass unemployment would produce significant social disruptions. Throughout the 1930s, the communists led popular and well-attended unemployment marches. These became sufficiently disruptive that the prefect of the Loire banned all unemployment demonstrations in 1934 (Loire M Sup 545); the policing demands for unemployment marches soon came to limit the government's capacity to send troops to ordinary coal strikes (Pas-de-Calais 23 J). There was extensive concern that high unemployment would have adverse electoral consequences (Pas-de-Calais M 2383). There may have also been concerns that unemployment could fuel sympathy for French fascism, or encourage social pathologies such as crime and alcoholism. In earlier periods, when the effects of unemployment had been limited and regionalized, prefects had responded to local crises by creating informal labor exchanges. The prefect of Creuse, for example, found positions for displaced miners in Lavaveix by either finding openings in local industries or paying for their transfer to Nord and Pas-de-Calais (AN F 7 13795). When the coal industry began to suffer from overproduction in the late 1920s, the state responded by limiting imports, promoting sales, and developing regulated regional markets to divide equitably whatever demand was available. The effects of corporatist intervention were probably limited. There ''Because wages in this study are measured as total wages paid divided by total workdays, the dependent variable is not distorted by partial weeks. Partial workweeks, however, do vitiate the capacity of total employment to measure the overall demand for labor. This accounts for the otherwise inexplicable fact that employment is positively correlated with wages in the prewar period but uncorrelated with them in the postwar period.

134

Chapter 6

was little the state could do to reverse the global overproduction of coal or the rising competition from petroleum and electricity that had created the crisis. Furthermore, the state had joint responsibility for maintaining the health of not only the coal industry, but the merchant marine, the docks, the railways, and the state metal fabricators as well. The merchant marine and docks depended for their livelihood on robust English coal imports, and the railways and metal factories needed the cheapest possible coal in order to stay competitive. There were further conflicts of interest over transport rates, with the coal companies wanting cheap freight charges while the railway companies sought to maximize their revenue on coal hauls. Elaborate arrangements were worked out, such as shipping northern coal at subsidized rail rates to English Channel ports, where it could be shipped by water to western France to replace English imports. Even with subsidies, the French product was still more expensive. In most cases, though, the state faced zero-sum conflicts between industries that were not easily resolved (Loire 15 J 2462; Loire 15 J 2449). The state's primary mechanism for providing unemployment relief for coal miners was to force the companies to pay unemployment benefits. This was nominally a voluntary arrangement made by the coal companies at the encouragement of the government. In a 1933 speech to the Chamber of Deputies, however, Paganon, the minister of public works, had publicly threatened the coal companies with the imposition of new and severe coal taxes at the point of production unless they both created unemployment funds and financed them generously with their own money (Loire 15 J 2444). Privately, the prefects used informal pressure to reduce the amount of unemployment. All layoffs and reductions in the number of shifts worked per week had to be formally announced to the local prefect, who would usually require formal justifications and good-faith attempts at moderation (Loire M Sup 545). An interesting consequence of this is that nominal wages remained surprisingly high during the Great Depression. Before the crisis, these wages had customarily been adjusted annually, or even several times annually, as in 1923 or 1926. During the depression, wages were reduced once in 1931 and then again in February 1932. There were no further adjustments until the surge of wages during the 1936 Popular Front (Comite des Houilleres de la Loire 1934; Loire J 15 2279). It is hard to account for the absence of downward adjustments during late 1932, 1933, 1934, or 1935. During each of the intervening years, coal employment, production, and prices all fell.i^ ^^Productivity did increase as a result of Taylorism and personnel reductions (HardyHemery 1970; Galand 1937; Moutet 1986).

The State versus Managers

135

As subsequent chapters show, labor militancy, unemployment demonstrations, and the growing rapprochement between communist and socialists help to explain this. These developments, however, were linked to unemployment and the government's concern with the link between the immiseration and the radicalization of the working class. These perceptions were greatly fostered by the government's participation in welfare functions. The government's role in fighting unemployment and in administering employer-financed worker security programs had established an important precedent; prefects were accustomed to financially assessing employers to insure workers' well-being. These policies may have contributed to a public unwillingness to immiserate workers further with wage cuts, providing welfare-state programs with economistic consequences.

CONCLUSION This discussion has addressed many topics and has been relatively complex. The main point is that the state was highly motivated to reduce industrial conflict, whereas private coal companies lacked such motivation. Furthermore, the state was ambivalent about higher wages, whereas private companies overwhelmingly sought low wages. This made the state more likely than the coal companies to attempt to buy off industrial militancy by providing generous wage concessions. The state's aversion to strikes stemmed from concerns with collateral unemployment and the objectively high costs of policing relatively violent incidents. The firm's indifference to strikes came from the capacity to recoup revenues lost during strikes with subsequent sales at higher prices. As a result, workers could disrupt the regional and national economy, but they had less capacity to reduce the profitability of their own firms. The firms were motivated to reduce wages because of the high labor-intensity of coal mining. Although the state had many motivations to prefer low wages, including many congruencies of interest with the coal companies, these were counterweighted by the presence of socialists and prolabor reformers in government office, a concern with managerial intransigence in promoting strike activity, and a social welfare concern with avoiding the economic immiseration of coal communities. Though these concerns did not make the state a single-minded proponent of high wages, they weakened an instrumentalist concern with accumulation that otherwise might have prevailed. In this particular case, the active participation of the state in industrial relations greatly facilitated the correlation of strike militancy with generous wage outcomes. It is not clear whether this tendency would be

136

Chapter 6

replicated in other states. The participation of the Nazis in German labor relations, for example, probably did not lead to a similar positive relationship between strike activity and wages. Many liberal democratic states share the French Third Republic's concern with the minimization of industrial conflict, however, as well as that government's reluctance to use brute force in order to impose management's will on strikers. Because some of these governments also have working-class parties whose members hold public office and have welfare states that are financed in part by employers, some of the Third Republic's willingness to buy peace by imposing high wage settlements on firms might be replicable in a number of other settings. A more disturbing question is whether the mechanisms suggested in this chapter invalidate the logic of Chapter 2. If the strike-wage relationship is a function of statist intervention into industrial relations, it may not be necessary or even correct to argue that private sector employers acting on their own account can be intimidated by worker militancy into giving high wages. Technically, only an empirical examination of the correlation between strikes and residual wages in an industrialrelations system with minimal state intervention can provide a decisive answer to this question. Many North American industries would provide good tests of this hypothesis. I suspect that the two-party model of Chapter 2 is probably correct. The state was very important for producing strike-wage correlations in French coal because the private employers were relatively indifferent to strike costs. This indifference, though, is highly unusual. Most employers cannot recoup their losses from strikes during economic upturns; sales that are lost during strikes are obtained by competitors. Contractors, garment makers, metal manufacturers, and transport companies are all examples of firms that suffer permanent revenue losses when workers stop producing. These employers therefore have to take some actions to defend themselves against strikes. Although presumably they would prefer not to pay high levels of wages and benefits, a strongly organized work force with a stranglehold on production may give these employers very little choice. The state may be necessary to act as a brake on industrial conflict when employers show no interest in reducing stoppages on their own, but in most cases, both the state and employers have an interest in industrial peace. Double causation may make the relationship between strikes and wages overdetermined by parallel mechanisms.

7 Militancy and Accommodation The Intimidation of Authorities by Unified Workers

The first half of the book has demonstrated that there were two distinct optimal strike strategies in French coal mining, and that which of the two applied depended on whether the labor force was united or divided. The next two chapters are designed to demonstrate that the statistical results supporting this theory are not a methodological artifact, but instead have an observable empirical basis. In short, the regression equations are mirrored by the historical narrative. To some extent, the two strategies periodize into before and after 1923; thus. Chapter 7 primarily contains material from before World War I, while Chapter 8 mostly contains stories from the 1920s and 1930s. There were some unified labor movements after 1923, however, and some nascent schisms before 1923. These off-period accounts provide insight into the general characteristics of unity and division overall. The previous chapter suggested that the state played a major role in seeing that strike militancy was rewarded. The pre-World War I narratives that follow confirm this strongly. In most of the stories that are presented, workers received generous strike settlements because the state intervened overwhelmingly in their behalf. The willingness of the state to reward strikers with benefits will be here referred to as accommodation. Accommodation was sufficient but not necessary to create a positive link between strikes and militancy. The chapter ends with a discussion of strike strategy in unitaire districts after 1923. In these settings, there was a positive relationship between strikes and residual wages; nevertheless, the state was overwhelmingly hostile to the unitaire presence. The communists obtained their strike premiums with no 137

138

Chapter 7

state patronage whatsoever by creating a credible reputation of militancy. There is no attempt made here to provide a complete chronology of the hundreds of coal strikes that took place between 1890 and 1914, or even to provide a narrative of all of the major incidents. What is instead provided is a description of how accommodation operated, with the strikes being selected for their illustrative nature.

CAMPAIGN FOR THE CONVENTION OF ARRAS, 1889-1891 Eighteen ninety-one represents a significant milestone in the development of accommodation in French labor relations. ^ It is possible to argue that accommodation was born in 1884 with the legalization of unions, but 1891 is the date of the Convention of Arras, the first major formal collective bargain ever negotiated by French unions. The Convention of Arras was a regional agreement between the mining companies in the two major northern departments, Nord and Pas-de-Calais, and the miners' union, the Federation des Travailleurs de Sous-sol. Employers were only represented by an ad hoc negotiating committee, although subsequently the companies' bargaining arrangements would become more formalized. The most prominent company in Pas-de-Calais, the Compagnie de Lens, insisted on its own private parallel negotiations, and a few midsized companies (Bruay and Bethune) avoided having to pay higher salaries.^ With all its imperfections, however, the Convention of Arras was a stunning accomplishment for the union, representing the first time that workers had forced a wage and condition agreement on management that was legally binding and could not be single-handedly renounced by the company at a later date. The fact that it was a multiemployer agreement made this a decisive blow against individual, patronal control over firms. The creation of the Convention of Arras was a textbook case of a union working closely with a sympathetic liberal government to obtain favorable state intervention. The agitation began with a cluster of individual coal strikes in 1889. The offensive was not so much a general strike as an informal strike wave that began at Lens and then exploded

'Unless otherwise indicated, all empirical material in this discussion comes from Gillet (1957, 1973). 2A small number of miniscule companies such as Ligny-les-Aire actually never participated in the bargaining process until after World War I, although their wages and conditions frequently mirrored the results of the negotiations. The percentage of employment by these companies was so minor that the issue was never pursued seriously by the union.

Militancy and Accommodation

139

regionally as workers m parallel compariies imitated the success of the Lens workers. One factor precipitating the strike wave was rising coal prices and robust demand. Equally important, however, was that it occurred after a general election in which, in the north, a set of liberal republican candidates ousted a set of managerially supported rightists. Three of the defeated conservatives were major administrators of coal companies. At the same time, Basly, the prorepublican leader of the miner's union, lost his parliamentary seat in a working-class district in the Paris suburbs to an explicit Boulangerist. The republican government was very grateful to the miners for their political support, particularly in the light of the growing threat of a coup d'etat from Boulanger. The prefect of Pas-de-Calais was known to be frankly sympathetic to the miners' strike as a way of further solidifying the miners' electoral bloc and embarrassing the managerial opposition (Loze 1891). During the 1889 strike, the prefect met with the employers' federation (which up to this point had been narrowly a technical and marketing organization) and explicitly demanded generous wage concessions for the workers; management caved in, granting a 10% increase (Vuillemin 1890). History repeated itself in 1891. Basly was elected to the Chamber of Deputies from the very district where the director of the Compagnie de Dourges had lost the previous election. This time, management did not even try to oppose him with a conservative candidate. Subsequently, the union launched a salary offensive, starting with partial strikes at a few companies. Both the individual companies and the employers' federation refused to make any improvements. Managerial intransigence induced the miners to vote for a general strike in the basin. Both the minister of public works and the prefect of Pas-de-Calais insisted that management meet formally with the union in tripartite negotiations at the regional level. A committee of managerial representatives met with union negotiators at the prefecture at Arras; through a combination of bargaining and informal arbitration, they agreed to raise salaries another 10% and to sign the Convention of Arras (Loze 1891; Macqueron 1904; Mangez 1925). The role of prefectural and ministerial intervention in creating the convention is clear. It is also worth noting that in the negotiations, union leadership did everything it could to play the role of peacemaker, whereas management tended to be obstructive and uncooperative. Despite rising coal prices and soaring employment in the coal mines, management had made no overtures on its own to raise the published wage rates to reflect market conditions.^ The companies refused to meet with ^On the other hand, there was undoubtedly wage drift in the form of individual piecework bargains and reclassification.

140

Chapter 7

the union; in fact, the employers' federation wrote to the union inviting dissatisfied workers to quit and seek employment elsewhere. In contrast, Basly did everything in his power to take a public stance opposing strike action. He spoke vigorously and forcefully against striking at union meetings. It was a matter of open knowledge that union meetings were recorded by the police, with transcripts forwarded to the prefect.4 Over popular resistance, Basly imposed on the union the English system of formal, secret strike ballots before engaging in stoppages, specifically because it would slow down the implementation of general strikes. He also spearheaded the campaign against a strike vote in 1891, although he lost by a compelling majority. All this put Basly in the desirable position of being a conciliatory negotiator forced by a radical rank and file to press an aggressive case. Management, in contrast, had to bear the responsibility for their intransigence. This gave the government further reason to support the union in its demands while showing little sympathy to the employers.

CARD, 1890 An extremely interesting analytical contrast can be made with events in Gard, which also experienced a major strike in 1890.^ In this case, however, the prefect showed no support for the miners, made no attempt to facilitate negotiations with management, and repressed the strike with physical force. As in the north, the Gard coal owners were monarchist, and they represented a comparable political threat to the Third Republic. Gard differed from Nord and Pas-de-Calais in two important aspects, however, both of which produced similar effects. First, the mines in Gard were more economically marginal; one major employer was in the throes of liquidation and reconstitution. All of the companies suffered from geologically small, inaccessible deposits and geographical isolation ••During the Third Republic, French law required the opening of union meetings to police representatives. Government observers attended such meetings routinely and overtly. Police spies were unavoidable, there was no other forum in which militants could speak to workers in privacy, and actual levels of repression and harassment were low. Therefore, speakers at union meetings were relatively frank at recorded meetings. Basly, a moderate, spoke moderately. More radical militants used union meetings to espouse the violent overthrow of the state or to blast government incumbents. Overall, there seems to have been remarkably little self-censorship by labor speakers. 5A11 empirical material for this section comes from Gaillard (1976).

Militancy and Accommodation

141

from major markets, producing a stronger economic case for denying concessions. Second, and more important, was that the Gard movement was led by anarchists, whereas the northern movement was led by strike-averse reformists. The anarchists candidly advocated using dynamite in the mines; there had been two actual dynamite attacks on company property during the late 1880s. Furthermore, the anarchists rejected supporting republican candidates as a tool for improving working-class welfare, and they eschewed both arbitration and parliamentary reform. From a prefect's point of view, the Gard unions represented as severe a problem for social order as did management, and thus there was no reason to mediate on the union's behalf. This is in contrast to the Nord, where the union consistently demonstrated to the government its willingness to participate in conciliatory procedures.

THE 1892 STRIKE LAW The degree to which the republican government approved of the Convention of Arras can be judged by the strike law that was passed in 1892. This made the procedure used in settling the Pas-de-Calais strike of 1891 standard procedure for all strikes in all industries in France. Under the new legislation, any set of workers on strike could petition their prefect to set up a conciliation committee. This was a mandatory collective bargaining session, usually held at the prefecture. Bargaining could occur at the departmental rather than the firm level, and the prefect was expected to serve actively as a mediator (Shorter and Tilly 1974; Lorwin 1954; Saposs 1931). Management did not have to make any actual concessions, but avoiding doing so was often very difficult without governmental sympathy. Nearly every major coal strike in France after 1892 made extensive use of the provisions of this law.

CARMAUX, 1892 One of the most notable cases of state support for miners is the 1892 wage arbitration at Carmaux.^ Carmaux was a company town in a relatively isolated coalfield in Tarn in the south of France. It had a wellorganized labor force with a history of striking and ties to an equally radical set of glass workers. It also had an extremely conservative upper ^Unless otherwise indicated, all empirical material comes from Tremp^ (1971).

142

Chapter 7

class, which had dominated the political and cultural life of the town. Catholic institutions and ideology were vigorously supported as part of an explicit extramural strategy to promote docility and obedience among the resident miners. In 1892, the authority of the company suffered double blows: the passage of the 1892 strike law, and the election of a union activist as mayor of the town. Soon thereafter, the workers drew up a long list of demands. Management rejected most of these, including nearly all of the wage-related grievances. As a result, a strike commenced. The workers maintained impeccable public order. After an appeal by the workers to the prefect, trilateral negotiations were begun at the subprefecture at Albi. There management conceded on many points, but several major differences persisted between the parties. At this time, the prefect mandated arbitration; after two stages of arbitration, a settlement was imposed that specified generous wage increases for nearly every grade of worker (SG 1892). The defeat of management became even more pronounced after a second strike that year. In retaliation for the loss of the earlier wage strike, management fired a union leader. This activist, Calvignac, happened to be the newly elected socialist mayor of Carmaux who had just convincingly defeated a management candidate. The battle over wages thus spilled over into a class war over the fundamental control of the town. The workers reacted to the firing with an immediate walkout; the company then responded by announcing a refusal to take back many of the strikers. Unlike the first strike, which used legal procedures of conciliation and was quickly resolved in two weeks, the Calvignac strike turned into a war of attrition that lasted 3V2 months. The state intervened in this dispute. Most onlookers agreed that the company was punishing the workers for exercising basic democratic rights. After preliminary information-gathering procedures, the government presented the company with a compromise calling for rehiring Calvignac and all the workers. The only face-saving concession was the expected nonsupport of a worker demand to fire the company director. The company turned down this offer, however, as well as multiple offers to submit the dispute to arbitration. Finally, the Chamber of Deputies forced the company's hand. The deputies voted 449 to 38 for an order threatening the company with the loss of the right to its mineral concessions unless it accepted government arbitration. The company found itself forced to agree to formal arbitration. The arbitral sentence that was delivered exactly mirrored previous government offers. In particular, Calvignac was to be rehired, but then put on immediate vacation to allow for the full-time pursuit of his mayoral duties.

Militancy and Accommodation

143

It should be noted that the workers were then able to press the government for one further concession. The union had insisted that all workers be rehired, but the government had allowed for a compromise in which workers who were convicted of disruptions of the peace during the strike could be fired by the company. After the announcement of the arbitral sentence, the workers simply refused to go back to work. The strike leaders invited a set of prominent prolabor deputies, including Clemenceau and Millerand, to come to Carmaux and negotiate a solution to the impasse. Their visit produced a further concession: a blanket pardon for all workers indicted for strike-related offenses. This governmental action took away the last legal basis the company would have had for not rehiring workers. The state's intervention had now become completely one-sided; the workers were left with an overwhelming victory (SG 1892). Residual wages soared after these incidents. Tarn, which before the strike had had residual wages hovering around zero, became, net of market factors, one of the highest-paying coal mining departments in France. This boost lasted for many years; negative residuals did not appear in Tarn until 1926. No other coal company in France received the resounding series of defeats that were experienced by the Societe de Carmaux. State intervention explained much of what happened, but there would have been no state intervention without the workers actually going out on strike twice. Their willingness to push for further demands after state arbitration was another indicator of their high level of militancy and their readiness to engage in stoppages in pursuit of their goals.

MONTCEAU-LES-MINES Montceau-les-Mines provides two illustrations: one of what militancy could accomplish with the full support of the accommodationist state, and one of what it could accomplish under adverse political circumstances.^ There were two major strikes, one in 1899 and the other in 1901. The first strike was extremely similar to the 1892 strike in Carmaux. Like Carmaux, Montceau-les-Mines was a company town; the Societe des Mines de Blanzy dominated the religious, cultural, and political life of the town. Promanagement figures filled all of the major local offices, and as in Carmaux, these company candidates took conservative, antirepublican, proclerical positions. The working class did not TJnless otherwise indicated, the empirical material in this section is from Beaubernard (1981).

144

Chapter 7

share these views, and the mines had a history of anarchism and labor violence. Even so, the 1890s had been relatively quiet, and at the beginning of 1899 Montceau-les-Mines was unionized. Rising coal prices and economic expansion brought this peaceful period to an abrupt end. In June 1899, the Montceau miners spontaneously mobilized and demanded recognition of the union as their representative in wage negotiations, a reduction in the workday, and an increase in salaries. The government was relatively sympathetic to their position, especially in the light of many republicans' hostility to the politics of the Societe de Mines de Blanzy and its owners, the Chagot family. However, the level of government support for the miners rose dramatically because of a fortuitous political accident. In the early days of the strike, the government fell as a by-product of a conservative court decision in the Dreyfus case. The government that came into power was reformist and moderate, with a modest dependency on support from the left. The parliamentary socialists wished to use the occasion to provide a conspicuous demonstration of state support for the Montceau strikers. They began agitating for a withdrawal of the company's mineral concessions because of the latter's unwillingness to develop its mineral deposits. As in the case of Carmaux, any protraction of a strike on the part of management by not making immediate concessions to workers was interpreted by the left as a voluntary abandonment of production. The government as a whole largely adopted the socialists' position. The Inspector of Mines, representing the Ministry of Public Works, openly espoused the theory. He cited correctly the company's political and religious policies as key grievances underlying the restiveness of the working-class population. In a formal meeting with company representatives, he explicitly threatened to transfer the company's concessions unless management acted to bring the strike to an immediate conclusion. The Cie. de Blanzy had little choice but to capitulate. Nearly all of the workers' demands were granted, and the Montceau union obtained a dramatic victory. In the next 2 years, everything changed. In the elections of 1900, socialist candidates swept the management slate out of municipal office. The highly unpopular director of the company resigned, although he maintained a nominal position in the company and a salary. Though coal prices continued to rise, the demand for Saone and Loire coal declined. By late 1900, the company was facing the necessity of substantial layoffs. In late December, the company announced the dismissal of 450 workers. Workers reacted to this cut fairly slowly, but struck on January

Militancy and Accommodation

145

20. The strike began as a wildcat dispute over work allocation issues, then transformed itself quickly into an official union strike for a rehiring of the dismissed workers and a salary increase. Even the relatively small Christian union joined reluctantly in the struggle. This time, however, no governmental assistance was forthcoming. The republicans' concern with the use of corporate power to support conservative candidates was obviated by the successful election of socialists to local office and by the internal changes in the administration of the company. Likewise, the immediate economic distress of the Cie. de Blanzy made the Ministry of Public Works relatively sympathetic to the firm's positions on labor costs. As a result, the Chamber of Deputies formally voted to put no pressure on the company to settle the strike. With the absence of third-party support, the strike collapsed, and the workers returned to their jobs with no concessions. The previous story suggests that 1899 was a victory for workers and 1901 was a defeat, but an examination of the residual wages suggests a different evaluation. Although the workers obtained modest gains in 1899, they obtained substanhally more dramatic benefits in 1901! Residual wages rose between 1898 and 1899, almost entirely as the result of creating a new uruon. The gain in raw and residual wages that came from unionizing was fairly substantial; nevertheless, controlling for union status makes the wage gain disappear. There was thus no additional benefit that came from striking. In contrast, wages rose from 1900 to 1901, from 1901 to 1902, and every year thereafter until 1905. These increases are observed with either simple residuals or those that control for union status as well. The 1900-1905 residual wage gain is not artifactual. Employment declined continuously after 1900, and after 1901 this decline was paralleled in coal prices. Yet despite the deterioration of business conditions, raw wages actually increased every year from 1899 to 1904! These increases are hard to explain in economic terms. It is frequently argued that wages are sticky downward (Hicks 1963; Okun 1981); however, neither neoclassical nor implicit contract theory expects wages to rise during business downturns. Such a prediction is also inconsistent with the core equations in Chapter 3, which show a strong positive correlation among employment, prices, and wages.^ The robustness of wages in the face of declining employment conditions is likely to be the product of levels of labor militancy both in Saone"The rising residual wages are not a product of falling prices either. Wholesale prices were nearly constant in France between 1900 and 1905 (France. Institut National de la Statistique 1966). Furthermore, the increases are observed for money wages as well as residuals.

146

Chapter 7

et-Loire and in France as a whole. After management's defeat in the strike of 1899 and the rapid mobilization of workers to defend their interests in 1901, management had to be very careful about creating grievances. The 1901 strike had shown that workers were willing to strike even in business downturns. Furthermore, although the government had not come to labor's assistance in this particular case, the possibility of state intervention had been seriously discussed. In 1902, France experienced a general strike in coal, and the government intervened actively in the workers' behalf. Throughout this period, therefore, management was in constant danger of facing a successful strike movement. It is highly likely that management acted to forestall further conflicts. It reduced its levels of layoffs after 1901, and it also maintained both nominal and real wages. Saone-et-Loire did not join in the 1902 General Strike, although other nearby coalfields did. The combination of rising wages despite falling business conditions, and a labor force that abstains from participating in a major national mobilization suggests a policy of preemptive concessions. By maintaining good conditions for workers in a bad labor market, management avoided the kind of overt confrontations that had led to adverse interventions both in Saone-et-Loire and elsewhere. By creating a reputation for militancy, workers increased their wages through peaceful means.

THE NATIONAL STRIKE WAVE OF 1902

Between 1893 and 1898, there was relatively little strike activity in French coal mining because of an extended recession.' In contrast, the demand for coal was robust from the turn of the century to World War I. This produced a series of national strike waves, notably in 1900, 1902, 1906, 1912, and 1914. With the exception of 1900, these strikes were linked to a political mobilization to ensure the passage of some form of mining welfare legislation, usually involving either the 8-hour day or pension reform (Trempe 1984). In each of these cases, though, the political campaign was also used as a spearhead for a salary offensive and agitation for improvement in local working conditions. The 1902 national strike wave provides a good illustration of how the French accommodationist state responded to the largest and most threatening disruptions that the miners' union could generate. 'Except where noted, the empirical material for the following section is all drawn from SG 1902.

Militancy and Accommodation

147

There are at least five lessor\s to be learned from the 1902 strike wave: 1. The miners obtained substantial political benefits from their strike waves. 2. Although their economic offensive was tied to a political agitation, the substantive linkage between these two campaigns was tenuous. 3. The state proactively implemented an aggressive program of arbitration and mediation to resolve the dispute. 4. Employers who cooperated with the arbitration program obtained significantly more favorable settlements than employers who resisted the conciliatory process. 5. The miners obtained significant economic benefits from their strike wave by making wages sticky downward. Political Benefits The primary political concerns of the miners were obtaining an 8-hour day and better pensions. The 8-hour day continued to be an issue throughout the twentieth century up to World War II. In the earliest years, managers were reluctant to grant the concession in any form. Even when nominal 8-hour legislation was passed in 1905 as a result of the strikes, the implementation continued to be problematic (De la Taille 1939). With regard to pensions, in the nineteenth century, the employers wholly controlled the funds. Not only was there significant managerial underfunding, a perpetual issue in unregulated pension plans, but employers often denied retirement benefits in a completely capricious manner. Firing workers just before retirement and restricting eligibility to a small fraction of the work force were common practices (Trempe 1984). The agitation leading up to the general strike of 1902 was characterized by extensive campaigning for the 8-hour day, pension reform, and a legal minimum wage. These topics dominated union congresses, and the leadership of the Federation des TravaiUeurs Sous-sol extensively lobbied high government officials concerning these matters. In 1901, the minister of the interior, Waldeck-Rousseau, wrote to the union expressing a general willingness to commence legislation on the matters of the 8-hour day and pension reform, but not the minimum wage. The national committee of the union dropped this third issue and actively pursued the other two, with the union voting a general strike to push

148

Chapter 7

both issues. As usual, Basly and the rest of union leadership argued vehemently against a strike, but they were overruled by the rank and file. During this preparatory period, bills were filed on both the 8-hour day and pension reform; both ultimately became law. Arbitrary Linkage of Political and Economic Campaigns Although the salary offensives of the miners' union were ostensibly tied to political mobilizations, the actual linkages between these campaigns was tenuous. This is because, technically, there was no time between 1900 and 1914 in which some form of miner's-rights legislation was not before the legislature; one could pick a strike date virtually at will and claim that it was related to the passage of some mining bill currently before the Chamber of Deputies. There was a colossal volume of such legislation. Every year, the Annates des Mines, the official government publication on mining law and policies since the early nineteenth century, lists approximately 300 to 500 pages of changes in mining laws and administrative regulations. Even in the nineteenth century, at least 25% of these changes involved worker welfare, covering such issues as mining accidents, retirements, workers' housing, and organization of the workplace. The potential for political agitation was also increased by the relative slowness of the French legislative process. It routinely took years to get important laws passed. Bills had to undergo many votes, both in committee and in preliminary readings before the whole chamber. In the case of the 1902 strike, the pension reform that the miners sought had already been filed as a bill in 1901. The 8-hour-day reform the miners sought had been filed as a bill in early February of 1902, 8 months before the beginning of the strike. Both bills were finally passed in 1905, meaning that any date between late 1901 and 1905 would have been appropriate for a strike to advance either cause. Although the public demands of the strikers were for legislative improvements, the timing of the 1902 strike was dictated by the economic terms of previous collective bargains. In the northern basin and in the Loire, salary increases that had been negotiated in 1900 were coming to the end of their contractual term. In both basins, management had announced its intention of lowering these rates. These rates, called primes, were percentage bonuses added to the base salaries that existed when the first collective contracts were negotiated in a region. A prime of 10% meant that pay was 110% of base pay at some previous standard date. Although the miners demanded both of the above political concessions, as well as workplace concessions and a generous increase in salary.

Militancy and Accommodation

149

pragmatically the 1902 campaign was a salary defense designed to avoid a reduction of the prime (Mangez 1925; Comite des Houilleres de la Loire 1925). A similar use of political demands to distract from a salary offensive occurred in 1906, in the famous Courrieres strike. In this year, a major mining disaster killed more than 1,000 workers. The accident attracted national publicity, particularly when the miners struck immediately afterward. They received enormous popular sympathy, as well as extensive parliamentary concern over their safety and well-being. The postCourrieres strike won enormous wage concessions and furthered the cause of pension reform. Ironically, during this strike wave, the miners' union did not put forward a single demand concerning mining safety (SG 1906). Arbitration as the Preferred Government Solution Once the general strike had broken out and achieved a significant turnout, the Chamber of Deputies formally discussed the stoppage. Two days after the initial discussion, the chamber formally voted a resolution urging the government to "use all its influence to make the parties accept arbitration." The miners' union had already put itself on record as being wholly in favor of arbitration. What the resolution substantively implied is that the government was to pressure management to accept arbitration reducing patronal control over wages. Shortly after the vote, the president of the council met with union leadership and announced that he personally had begun negotiations with the companies and was in active communication with them. The Ministry of the Interior and the prefects implemented the arbitration policy with dispatch and vigor; in nearly every region that experienced a stoppage, a committee of conciliation was created within days. Furthermore, the justices of the peace and prefects mediating the negotiations made early and frequent appeals for both sides to submit their cases to arbitration. Where resistance to this proposal was encountered, the objections nearly always came from the companies rather than the union. Companies Who Resisted Arbitration Received Worse Contracts There were marked regional differences in the salary settlements of the 1902 strike. The regions with the least favorable new contracts tended to be those whose managements resisted the arbitration and conciliation process. Coal prices and the demand for coal had been falling since

150

Chapter 7

1900; this market decline was nationwide. One would therefore have expected a national downturn in wage payments that would have been reflected in a negotiated reduction of the prime. 1° Oddly enough, however, only in the most prosperous regions was management able to obtain a reduction in wage rates. In the more marginal areas, where presumably salary savings would have been particularly needed, the prime remained the same or rose. Greater management noncooperation with the conciliation process in marginal areas explains this apparent paradox. Table 7.1 shows all the salary settlements from the 1902 strike for which data are available. The most favorable settlement was that in the Nord and Pas-de-Calais; the least favorable that of Puy-de-D6me. In the northern basin, an agreement in 1900 that had set the prime at 40% was scheduled to expire in May of 1902. Near the date of the expiration, the company posted an announcement that after expiration, the prime would be summarily reduced to 25%. The union met with management to discuss the situation. At this negotiation, management introduced what would be a standard bargaining ploy of the postwar period, the Edeiman concession. An Edelman concession is when management makes a false and extreme negotiating demand and then immediately retreats to a more moderate position. The retreat gives the appearance of having made a concession, when in fact the second position represents the de facto original bargaining stance of the company. The term itself is a reference to a brilliant analysis of collective bargaining by Murray Edelman, the political scientist. In his essay "Conservative Political Consequences of Labor Conflict" (1969), Edelman argued that strikes and collective bargains in highly institutionalized settings can often be little more than dramaturgic ritual. The bargainers frequently have a clear notion of what the final settlement is going to look like very early in the bargaining process. To demonstrate to their constituencies that the present contract is the "best deal that can be obtained," however, both sides demand extravagant lists of concessions while engaging in strikes and condemning their opponents. Once a sufficient show of resistance has been made, they then "give u p " on their earlier platforms, making supposed "compromises" while in fact agreeing to what they wanted in the first place. This tactic, which was very common in French coal mining in the 1920s and 1930s, was rarely used before World War I. It did however make an appearance here. The day before the companies were scheduled to meet with the unions, they posted their intention of cutting the lowhere there was no prime in effect, one would have expected a comparable reduction in base rates.

Militancy and Accommodation

151

Table 7.1 C o n t r a c t Results of t h e 1902 N a t i o n a l Coal M i n i n g Strike b y Region a n d M e t h o d of S e t t l e m e n t Region

Change in prime

Pas-de-Calais/Nord Loire Loire-Haute

40 > 30 9 > 5> 4 > 3 0> 0

Puy-de-D6me

5 > 7

Allier

Increase (quantity unknown) 5>5

Tarn

Method Arbitration; early concessions Arbitration; management resistance Intervention by minister of interior; management refused arbitration Collective bargain; management refused arbitration Management refused to meet with prefect; no arbitration Arbitration; management resistance

Source: SG 1902. Note: A > B means a prime of A was ctianged to a prime of B. Several >'s in a row denote a multiphase change in the prime.

prime from 40% to 25%. The next day, as soon as the union walked into the room, they then announced that out of deference to the union, they would raise the prime back up to 30%. It is unlikely that management received any information in the 24 hours after the posting that made them sincerely reconsider their position. No strike action had been taken by the union. Although there certainly would have been grumbling from both union officials and the rank and file, such complaints have hardly have been surprising. There were also no official communications of disapproval from government officials. Whether there were unofficial pressures put on management by government authorities is unknown, but when management is assessing reaction to a controversial proposal, such measures are usually discussed privately with the relevant authorities rather than posted on the shop floor for all to see. The timing of the posting thus was not consistent with a change of heart by the companies; it was more consistent with management stating an extreme position so it could then publicly back down to show moderation and a conciliatory spirit. The northern miners waited several months to strike, in part because of the coordination required to execute the first nationally synchronized general coal strike in French history. The beginning of the strike was planned for August, when management was expected to make a parallel cut in Loire wages.ii I'The delay was also attributable to the natural strike averseness of the northern leadership, which took any excuse to wait.

152

Chapter 7

Once the strike actually broke out, conciliation committees were formed in Nord and Pas-de-Calais. There were minor administrative delays in getting these started because of benign difficulties experienced by both sides. Once they were formed, the prefects of both departments suggested arbitration. Although the companies did not immediately enter arbitration, they voiced full support of the principle, expressing concern only over which issues should be decided. They reduced the scope of arbitration by conceding on many of the nonsalarial points of dispute. ^2 Less than a week after the prefects' suggestion, the companies and unions were ready to go to the arbiters on wages, the single issue that remained. Following customary practice, the arbitration committee consisted of one delegate chosen by the union and one chosen by management. The result was a complete victory for management. The arbitrators accepted the managerial proposal of a prime of 30% with no amendments. The wage settlement was the same as that offered by the companies in the May negotiations and paid throughout the summer, representing a 7% cut in overall remuneration. In no other region were the companies able to obtain comparable reductions. The next best deal was negotiated by firms in the Loire. They were able to obtain a 5.5% reduction in overall wages, although this was phased in over a year and a half, with milder reductions in the interim. The Loire had originally been paying a 9% prime, which it wanted to reduce to 3%. The arbiters compromised by lowering the prime to 5% immediately, 4% a year later, and then 3% an added 6 months afterward. The 1900 prime convention expired in August 1902, approximately a month and a half before the strike began. The company directors did not engage in any Edelman concessions. They offered a flat 3% and claimed that the unions were fortunate that the prime was not eliminated altogether. Once the strike started, the prefect suggested conciliation committees and arbitration. The Loire companies followed the northern companies in accepting the principal of arbitration. Unlike the northern companies, who concluded their preparations in a week, the Loire negotiators delayed the process of arbitration a month, haggling over the choice of arbitrators. In the event of nonagreement by the company and union arbitrators, the final decision was to be made by a third arbitrator selected by both parties. The negotiators could not agree on a candidate i^The workers themselves conceded on some points for which it was known they had no government support. The most important of these was the demand for a minimum wage.

Militancy and Accommodation

153

for the third position; in this particular case, the union probably shared joint responsibility with management for the delays. The next problem emerged when management refused to accept candidates for third arbitrator offered by the union. Unfortunately, the first union candidate had been the prefect of the Loire. Such a rejection cannot have been well received either by the prefect himself or by his employer, the minister of the interior. Although the settlement that the Loire companies received was not as favorable as the deal in Pas-deCalais, it nevertheless represented a significant reduction in wages. Loire management had basically cooperated with government conciliation; the only minor differences were an absence of early concessions and a willingness to alienate important political figures. In most of the other regions where the 1902 strike took place, the companies were significantly more hostile to the mediation process. In these other regions, the workers struck in conjunction with the Loire workers. In the Loire-Haute, for example, turnout was nearly complete. The prefect arranged for tripartite collective bargaining, where agreement was reached on some minor issues. The prefect then proposed arbitration for resolution of further differences, but the local companies refused to participate. Upon this rejection, the minister of the interior called for a meeting with company representatives at the ministry in Paris. At this conference, he obtained commitments from the companies not to lower their bonuses throughout 1903. Technically, there was no prime to cut in the Loire-Haute; however, through the forced agreement to freeze salaries, the Loire-Haute was prevented from cutting wages at a time when other employers were obtaining rollbacks. Neighboring Puy-de-D6me underwent a similar experience. The Puy-de-D6me prefect worked in close conjunction with the Loire-Haute prefect. Trilateral negotiations took place; an impasse was reached; and once again the prefect suggested arbitration. In this case, the company preempted arbitration by granting a 2% prime increase after the proposal was made. The company maintained control over the setting of wages, but only by increasing wages during a period of falling prices. In AUier, managerial resistance to governmental intervention was even more sustained. The company of Saint-Hilaire refused to participate in collective bargaining, meet with the prefect, or even respond to the prefect's formal written communications. As a result, there was a stalemate that lasted for 3 months. Finally in January 1903, long after the strike had been settled elsewhere in France, the company was induced to agree to trilateral negotiations. In 1 day of bargaining, the company made concessions on nearly every point raised by the workers, and agreement was reached. As in Puy-de-D6me, management was

154

Chapter 7

forced to accept a wage increase despite absolutely deteriorating market conditions. Carmaux, in Tarn, seems to have been an exception. In 1902, management cooperated fully with the arbitration process (in marked contrast to their behavior in 1892); nevertheless, the preexisting prime of 5% was left wholly unchanged. The workers demanded that wages be frozen, and this concession was granted in full by the arbitrators. The paradox of a cooperative management not winning wage concessions is resolved when one considers that management had lowered the prime from 12% to 7% only 3 months before the strike. None of the other regions in Table 5.1 had experienced a comparable cut. The strike settlement prevented the Carmaux company from further lowering wages, but this ban had already been written into the preexisting contract (Trempe 1967, p. 464). In a very real sense, there were few substantive local issues in Carmaux in the 1902 strike; the workers were primarily displaying solidarity with a larger national mobilization. Because with only minor exceptions, all the workers were asking for was a preservation of a preexisting contract that reflected management's adjustments to market conditions, there was no reason for the company and the government not to go through the motions of giving the workers what they wanted. Strikes Benefited Workers by Making Wages Sticky Downward From the standpoint of economistic benefits, 1902 was a defensive strike. In some cases, notably in Puy-de-D6me and AUier, workers obtained positive gains. Generally, however, despite the presence of ambitious demands, all the workers could realistically hope for was to prevent management from lowering wages to reflect market conditions. They generally succeeded in this goal. Nearly everywhere, strikes resulted in collective bargains that stabilized wages into the next year. Some employers, like Carmaux, were able to obtain their cuts before wages were frozen, whereas others employers, such as those in the Loire-Haute, were caught napping. They lost their capacity to lower wages before they had made market adjustments. Conceivably, strikes were not necessary to obtain this end. It is possible to imagine the coal mining unions negotiating very long contracts, which would have produced wage rigidity without requiring constant striking. In French coal mining, however, most contracts had short durations; most of the preceding northern wage deals had lasted for less than 2 years (Mangez 1925). Settings with less institutionalization than Nord and Pas-de-Calais would have had no formal specifications of the

Militancy and Accommodation

155

duration of wage rates whatsoever. In the north, employers could plan to lower wages on the occasions of contract renegotiations, while in the less unionized south, employers operated on their own internal schedules. If the workers could move faster than management, they could thus obtain a favorable settlement. Lightning strikes in essence became a substitute for long contracts in settings where industrial relations were poorly institutionalized and wages were primarily set by managerial discretion.

MILITANCY WITHOUT STATE SUPPORT: COMMUNIST STRIKES IN THE LOIRE Chapter 3 showed that after the schism, strikes continued to correlate with residual wages in unitaire districts. In these regions, unitaires sufficiently dominated the local unions to preempt the practical possibility of converting the labor force to reformism by steering. One could steer benefits away from the area as a whole by not supporting the local labor movement. If local industrial conflicts produced problems of social control, however, management had little choice but to buy off strikers with concessions. Passive unitaires suffered from the full weight of managerial and governmental opposition; active unitaires could retain some measure of economic benefits. What is of particular interest is that the unitaires obtained their strike premiums largely without support from the state. In a few districts (notably Cantal), unitaires maintained a polite working relationship with both management and the prefecture (AN F7 13795). In most settings, though, the state was positively hostile to the communists. In the Loire, the communists spurned and abused local officials who offered the traditional resources of favorable state intervention. Afterward, the unitaires were left to fight their struggles on their own, with little in the way of political support. This discussion will focus on the Loire, one of the most important strongholds of unitaire syndicalism in French coal mining. Strikes were clearly rewarded in the postschism Loire: The correlation between strike activity and wages residualized for market factors and unionism is .49 (N = 12). The majority of unionists in Loire coal mining were communist, but there was a very significant minority reformist presence. Two towns, Firminy and Roche-Ia-Moliere, were staunchly reformist. They stood aside from the agitations that affected the rest of the department. The mines of these towns were operated by the Company of Roche-laMoliere-Firminy, the largest and most prosperous employer in the area.

156

Chapter 7

The mines were characterized by high wages and relatively generous employer-financed housing and benefit programs. They were, moreover, geographically isolated in small towns in the southwest of the department. Workers in Firminy and Roche-la-Moliere thus would have been cut off from most of the organized political and labor activity of the Loire. In contrast, most of the Loire coal mines were located underneath the city of Saint-Etienne and its immediate suburbs.^^i* Saint-Etienne had a long-standing tradition of radical unionism, and the unitaires were well integrated with other communists, including railroad and metal workers (Hanagan 1989; Amdur 1986). Before the war, Firminy had generally supported parliamentary reformism, whereas SaintEtienne had supported direct action syndicalists and anarchists. As a result, there had been an acrimonious history of Firminy not supporting Saint-Etienne strikes (Loire )2 M 138; Loire 92 M 139). These regional differences replicated themselves in 1923 when the Saint-Etienne miners supported the CGTU while Firminy and Roche-la-Moliere remained loyal to the CGT. After the schism, management continued its past practice of negotiating at a departmental level with the union representatives of the CGT. After the exit of the unitaires, however, the CGT only represented Firminy and Roche-la-Moliere. The communists were thus never present at the bargaining table. This was not attributable to communist unwillingness to participate. Throughout the period, the FTUSS submitted several formal written lists of demands to the Loire employers' federation and offered to meet in negotiating sessions. The CCH-Loire systematically ignored the unitaire communications, choosing to deal exclusively with the reformists. Note that in no case did the Loire prefects object to this ostracism. The state made no requests for management to discuss wage issues with the unitaires, and there are no records of any official displeasure with the CCH-Loire's conduct. The communists thus were cut out of the wage negotiating procedure with the full support and approval of the state (An F 7 13796; Loire 15 J 2449). Of course, the unitaires' willingness to participate in wage negotiai^It is no accident that the most prosperous employer in the Loire was located away from the central city. French coal companies were liable for damages to preexisting structures caused by loss of ground support. Property damage, and the concomitant restitutions, could be overwhelming in populated districts. The company of Villeboeuf was actually bankrupted in the 1930s by compensation owed to the owners of damaged Saint-Etienne buildings (Loire 15 J 2243). " I n 1927, only 26% of all Loire coal mining employment was in Firminy or Roche-laMoliere; 69% of Loire coal miners were in the Saint-Etienne area, and 5% were in a set of unitaire mines in the north of the department (Loire 15 J 2238).

Militancy and Accommodation

157

tions did not imply that the communists were willing to moderate their wage demands. During the February 1923 General Strike, the initial unitaire wage demand, which was never changed, was an immediate raise of 6 francs a day. Such an increase had never been given in the history of French coal mining; it was dismissed by most noncommunist observers as absurd. In 1923, however, both coal and coal mining labor were scarce, and market forces alone would have produced a substantial wage increase. The Loire companies offered the reformists 3 francs a day plus an improvement in family allowances. The one meeting held by unitaires and the reformists to discuss the proposal ended quickly in complete disagreement. The unitaires rejected the offer (purely symbolically, because the offer was not made to them and they were not involved in negotiations), but the confederes accepted it, and it became the official wage rate within the basin. The irony of the process is that these deliberations occurred during a unitaire strike in which the reformists were not participating. While these negotiations occurred, 75% of the Saint-Etienne miners were out on strike; almost none of the Firminy miners struck, i^ The reformists thus were negotiating the wage consequences of a strike being conducted by a completely separate union (AN F 13791). The events of February were repeated in November. The unitaires went out again; once again the reformists worked. Turnout rates for the fall and spring strikes were nearly identical, save for a moderate reduction in strike participation in the far north. Management gave the reformists a wage concession; the unitaires did not participate in the negotiations (AN F 13903). In 1925, the government attempted to incorporate the Loire unitaires into the traditional machinery of strike conciliation. The occasion was a strike in Villeboeuf, a district of Saint-Etienne, regarding two almost identical incidents on the same day in which a foreman had gotten into an argument with a Moroccan worker. In both cases, once the dispute started, the worker had punched the foreman, and the Moroccan had been fired. Subsequent investigation by the state mining engineer suggested that the workers had legitimate grievances. The foremen had been highly provocative, imposing extreme sanctions for relatively trivial infractions. Furthermore, the supervisors got the best of the resulting fistfights; in the most flagrant incident, the Moroccan had been severely beaten. The workers of Villeboeuf immediately went on strike, demanding i^The communist journal Humanite claimed a 20% strike rate for Firminy. Although Humaniii was reliable on most other issues, its strike participation figures often quintupled or dectupled estimates jointly confirmed from other sources (AN F 13791).

158

Chapter 7

the rehiring of the Moroccans and the firing of the supervisor involved in the aggravated beating. In a break with tradition, the company agreed to meet with the local unitaire leaders. Although the historical record says nothing about this point, it is entirely possible that the miner safety delegate for Villeboeuf was a communist. Miner safety delegates were generally drawn from the ranks of union activists; if this was the case, the company would have been legally obligated to meet with the delegate on matters concerning worker safety. Later the same day, the director met with the state mining engineer—again, standard practice in mine safety matters. After this meeting the company agreed to take back the workers. On the second day of the strike, the company agreed to transfer the foreman in question to a less desirable shift, where he would have been unable to supervise his former victim. Thus, by the second day, management had granted the substance of the workers' demands; furthermore, the state mining engineer seems to have been generally sympathetic to the workers' claims, and willing to intervene on their behalf. The CGTU did not accept any of these concessions. The prefect began to suspect that the unitaires were trying to prolong the strike as a basis for a regional political mobilization. On the first day of the strike, the unitaires wrote a letter to the prefect, laying out the legitimate bases of the workers' complaints. The letter also included a lengthy protest of the prefect's use of security troops to protect the mines during the strike. The unitaires had earlier in the day called for chasse des renards, the beating of any worker caught crossing the picket lines. In a formal meeting, they had actually worked out the arrangements by which this would be done. Unitaire activists were also working the Moroccan neighborhoods close to nearby mines in the hope of creating a regional ethnic mobilization. At the time, France was engaged in a colonial war in Morocco, a policy that was opposed vigorously by the French Communist party. In a letter to the minister of the interior, the prefect expressed fears that the unitaires were trying to create a linkage between these issues. His suspicions were confirmed when a few days later, national unitaire leaders arrived in Villeboeuf and began holding public meetings on the war in Morocco. The prefect ordered the state mining engineer to meet with the communists to try to obtain an amicable settlement. Negotiations were somewhat hindered by the union leaders arriving at the meeting with a correspondent from the communist journal Humanite, who proceeded to bargain actively with the engineer on his own account. The engineer offered a compromise by which the supervisor would be fired if found guilty of criminal acts by a tribunal. Management was later to spurn this

Militancy and Accommodation

159

compromise, but the unitaires rejected it on the spot, claiming they wanted the supervisor fired even if he was found innocent on all counts. The prefect then suggested arbitration. The arbitrator he put forward was Durafour, the minister of labor and a well-known socialist reformer. Durafour was the author of a recently passed 8-hour day act that was the most favorable work-hour legislation in French history. It is hard to imagine any government figure who could have given a more favorable hearing to a union. Despite a personal telegraphed appeal by Durafour himself and a reluctant acceptance by the company itself, the unitaires declined the arbitrator (Loire M Sup 548). i^ This relatively contentious strike ended with a fizzle. Despite the communists' best efforts, there was little sympathetic response from the Moroccan community, and the strike did not extend beyond Villeboeuf itself. The union then turned around and accepted the management offer to relocate the supervisor without further punishment. Whatever little sympathy the Loire prefect had for the unitaires was dissipated by this incident. In subsequent correspondence, the prefect noted that the Moroccan around whom the whole strike had begun did not show up for work the day after the strike. In the prefect's view the communists were simply political operatives, more interested in electoral gains than in any benefits that might accrue from sympathetic state intervention (AN F7 13795). The withdrawal of state support can most easily be seen in a strike that occurred the following year in Combeplaine, a very small mine in the north of the department. A worker had been fired after he refused to work alone at an isolated mine face. French safety regulations required that all underground workers have other workers in reasonable proximity; thus, the company's original request was technically illegal and the firing unsupportable. One hundred percent of the miners went out on strike. Arnaud, the leader of the Loire unitaires, wrote a long formal protest to the prefect citing the facts of the case and the relevant mining law. The state engineer made a formal inquest. He found Arnaud's presentation of the case to be fundamentally correct; early in the investigation, the mine operator himself admitted full guilt. Moreover, there was a substantive dispute between the employer and the mining engineer. The mine face was very shallow, being only 5 meters underground, and '*It is worth noting that net of any interest the unitaires had in prolonging the conflict, accepting Durafour as an arbitrator would have legitimized the role of parliamentary socialists in representing the interests of the working class, a position that would have been wholly unacceptable to the communists on ideological grounds.

160

Chapter 7

the site was well provided with safety equipment. The employer felt there would be no substantive problem with working alone, a position with which the state engineer disagreed violently. However, despite the fact that the employer was in violation of the law and showed no willingness to abide by the law, the engineer recommended that no action be taken against the firm. He advised the prefect that the company involved was a very marginal subcontractor who was close to insolvency and that "unless labor-management relations were improved, there was a risk that the pits would be closed." After the meeting with the engineer, the company fired both the original complainant and a leading union militant. Neither the prefect nor the engineer objected in any respect (Loire M Sup 548). The authorities apparently were willing to suspend the enforcement of workers' protective legislation if it meant teaching the unitaires a lesson. What followed was a sequence of continued communist militancy. Between 1926 and 1935, the unitaires struck nearly four times a year, a level of stoppages far in excess of that found anywhere else in France. Among these were three general strikes, in 1926, 1928, and 1931. The 1926 strike was in sympathy with the English coal miners who were out on general strike. No demands were put forward, and no direct concessions were obtained. There were, however, several independent wage negotiations during the year in which the effect of the strike might have been reflected.i'' The unitaires obtained a 45% turnout, a decrease from the results of 1923 (AN F 13793). At the end of 1928, the unitaires held a general strike on wages. In the early stages, they informally cooperated with the reformists. Just before the strike date, however, management offered a wage increase that the CGT accepted and the CGTU rejected. The unitaires went on strike alone and obtained a 57% turnout. Management formally reaffirmed to the prefect its intention not to deal in any way with the communists; the unitaires obtained no further concessions and came back to work defeated (AN F 137904). After the loss of 1928, the communists appeared to be in disarray. The Communist party attempted to remove Arnaud, the head of the union. The dispute substantively centered around concerns of Arnaud's political loyalty and his willingness to support leftist alternatives to the party. The party attempted however, to legitimate the purge by raising the issue of allocating responsibility for the 1928 defeat. There was a ^This was a year of hyperinflation caused by foreign exchange problems with the franc. Wages were adjusted virtually quarterly to take the declining value of the currency into account (Sauvy 1969; Loire J 15 2279).

Militancy and Accommodation

161

related conflict over whether shop-floor committees should be used as a vehicle for strike mobilization, a relatively transparent cover for circumventing local union leadership in the interest of producing greater party control of the Saint-Etienne coal miners. Unfortunately for the national party organization, the Loire rank and file was extremely loyal to Arnaud. Constitutional attempts to remove him failed repeatedly. Unitaire meetings during this period degenerated into bitter factional struggles, complete with jeering and fistfights (AN F 13791). Despite these distractions, the unitaire strike rate remained high; there were five communist strikes in 1930 alone. By 1931, the Great Depression, which came to France late, finally arrived at the French coal industry. There were national negotiations over the imposition of a wage cut. As in 1928, the unitaires and reformists made preliminary arrangements to combat the impending cuts jointly, and, as in 1928, the reformists came to an agreement with management at the last minute, just as they did in 1928. The unitaires were left to strike by themselves. The adverse market, the recent communist defeat, and the internal disarray of the unions made management relatively confident about its position. Pigeot, the head of the Loire employers' federation, personally assured the minister of labor that the reformists were resigned to accepting a pay cut and that the unitaires would not dare strike after having lost in an even more favorable market in 1928 (Loire 15 J 2449). In negotiations with the reformists he literally dared the unions to go on strike, arguing that a stoppage could not possibly help local miners (Loire 15 J 2239). Arnaud himself shared this general pessimism. Nineteen thirty-one represents one of the few cases where Arnaud was arguing for moderation and urging his workers not to strike (AN F 13791). The strike occurred anyway, however, because of the activity of the orthodox communists. The 1931 strike obtained no concessions, and in contrast to strikes supported by Arnaud, the turnout levels were extremely low (AN F 13905). Nineteen thirty-one represented the low point for the Loire unitaires. It was the first year since 1924 that there had only been one strike in the Loire, and residual wages reached the lowest level for the postwar period. Thereafter, both residual wages and strike frequency climbed in tandem. Arnaud separated from the CGTU, running an independent union for 2 years before joining the CGT in 1933. The CGT affiliation was nominal rather than substantive: In no case did the Firminy workers support any of Arnaud's stoppages. The policies of the Arnaud group were the same during the unitaire, independent, and reformist periods. Furthermore, during his reformist period, he struck at the same rate that he had as a unitaire. During this period the CGTU union maintained its

162

Chapter 7

formal presence, and the pure unitaires generally struck in support with the Arnaud group. The resurgence of residual wages was also affected by the increasing volume of unemployment demonstrations. As was discussed in the last chapter, both communists and reformists organized large demonstrations in the 1930s. The government responded to this threat to social order by pressuring companies to maintain their workers and to pay unemployment relief. Wages were kept high as a form of privately subsidized welfare; after the institution of company-paid unemployment benefits, discussions for lowering wages virtually disappeared from the private and public correspondence of the CCH-Loire (Loire 15 J 2244; Loire 15 J 2449). It would thus seem from the Loire experience that many sorts of collective action correlated with residual wages. For example, unemployment demonstrations produced favorable state interventions that indirectly affected wages. The most important form of intervention, however, was strikes. The high correlation between strikes and wages existed despite the virtual absence of state support for unitaire policies. In the Loire, the world of the first half of Chapter 2 prevailed. There was a unified set of public and private authorities opposing a militant labor force. This labor force generated a credible reputation for militancy and, as a result, was able to raise its wages through striking. The only twist on the Chapter 2 model is that the wage increases that unitaire workers received did not come through immediate concessions to a unitaire union; instead, management negotiated increasingly generous contracts with reformist representatives. Because unitaire workers were bound by the reformist contract, both reformists and nonreformists profited from the communist work stoppages. For both the pre-World War I FTSS and communists in unitaire strongholds, unified labor movements raised wages through striking. The situation changed significantly in settings where dual unionism was present. It is to these divided labor movements that we must now turn to see the use of steering and the gutting of the strike tactic.

8 Schism and Steering State Control over a Divided Labor Movement

Chapter 7 showed that militancy was rewarded in the prewar period, as well as among postschism unitaires. After the schism, however, workers generally made money by desisting from strikes, rather than engaging in them. As the statistical analysis of Chapter 3 showed, after 1922, residuals in reformist and mixed regions were inversely correlated with strikes. Furthermore, residual wages were twice as high in reformist areas where confederes did not strike as they were in communist areas where unitaires did strike. It was argued that these wage patterns were the product of a deliberate policy of steering on the part of French authorities: Both employers and the state had an interest in encouraging workers to join relatively moderate unions by punishing the communists and providing conspicuous benefits to the reformists. Thus far, skeptics might claim with some justice that little evidence has been presented to show that the state or managers were engaging in any explicit strategy of steering. A theoretical analysis has been presented showing that the state and employers would have been better off had they steered. Unfortunately, many such academic models have only limited relevance to what actually occurs. The only empirical support that has been presented thus far is in the correlative analyses of Chapter 3. Statistical findings, though can always be fairly attacked by readers who identify missing control variables that could confound the analysis and threaten the reported relationships. Therefore, the purpose of this chapter is to provide nonstatistical support for the steering hypothesis. The goal is to demonstrate that steering was explicitly adopted by the French government as part of a larger strategy of centralizing control over the industrial relations of French coal mining. This policy did not emerge full-blown at the first development of 163

164

Chapter 8

disunity among factions of the coal mining union. It was a learned response. Before the communists left the Federation des Travailleurs du Sous-sol, the state had dealt with divided unions before. However, these early splits were not nearly as serious as those that occurred between the confederes and the unitaires. The essential similarity between divided and undivided unions before World War I produced a comparable similarity of state response, with the French government accommodating militancy in both cases. The postwar years, however, presented the authorities with a labor movement that was turned against itself to the point of paralysis. Given the severity of the Bolshevik threat during the immediate postwar years and the extreme moderation that reformists adopted in response to international communism, the state was given a choice of profoundly different union movements that could be easily kept from uniting. This gave the government new options that produced an explicit policy of steering benefits to reformist workers while excluding communists from the formal industrial-relations process. Eventually, this was extended to a general policy of centralized control over the labor costs of the mines as a whole. As in the prewar years, managers who defied state policy continued to find the state supporting higher wages for their workers. Managers who cooperated, though, found a state willing to use its regulatory powers to enforce wage decreases as a deflationary measure. They also found a state that would accept Edelman concessions and formalistic participation in state conciliatory machinery as a substitute for bona fide bargaining. Capitalists did not manage to have their labor costs reduced; overall residual wages in the postschism period remained at the same level that had existed before the war. The process of granting those wages, however, became centralized and routinized into a formal legal machinery that rewarded the cooperative, punished the dissident, and linked labor costs to the needs of national accumulation.

WHEN STEERING EQUALS ACCOMMODATION: RESPONSES TO PREWAR SCHISMS The 1922 split between the confederes and the unitaires was not the first incidence of factionalism in the history of the FTSS. In the first decade of the twentieth century, there had been two splits, a relatively trivial separation between the main union and Christian unionists and a more substantive division within the FTSS between advocates of parliamentary reformism and advocates of direct action. Christian unionism

Schism and Steering

165

was not sufficiently important to justify a special state response; antistatism was a more serious matter. The early attempts of authorities to implement a strategy of steering, however, backfired into de facto accommodationism. Christian unionism, or, more pejoratively, yellow unionism, was developed by Catholic reformers in the late nineteenth century to provide believers with an egalitarian social program that could be an alternative to atheistic socialism. In contemporary France, the descendants of these old Catholic unions constitute the Confederation Frangaise des Travailleurs Chretiens (CFTC). These unions are militant, effective defenders of workers' interests. In the early twentieth century, however, these organizations were practically company unions. Christian unions taught that social conflict was fundamentally wrong. It was the moral responsibility of the employer to provide for the welfare of his employees. But in the absence of such good faith, most Christian unionists confined themselves to making verbal appeals instead of striking (Garmy 1934; Dolleans 1953). Although many French occupations had a significant constituency of jaunes ("yellows"), the impact of Christian unionism in coal mining was limited. The Annuaire des Syndicats shows that membership in these unions was very small. Christian union membership rarely exceeded 5% of the membership in the FTSS, much less 5% of the coal mining labor force as a whole. There were particular cases where Catholic unionism made deeper inroads. One of these was Montceau-les-Mines, where a yellow union was founded in 1899, a few months after the successful strike of the mainstream union. By 1901, the union had enrolled over a thousand members; that year's strike, the jaunes had probably organized at least 15% of the Montceau-les-Mines labor force.i These high membership figures, however, are extremely misleading. Most jaunes belonged to both mainstream and Catholic unions. Union membership in syndicalist-era France was not the exclusive affair that it is in the contemporary United States; unlike the American case, French workers did not formally have to elect a designated bargaining representative. Belonging to two unions was not unlike a modern environmentalist belonging to both the Sierra Club and the Audubon Society. The most important fact is that the Christian union provided very little independent input into the subsequent labor relations of Mont'As before, most of the material on Montceau-les-Mines comes from Beaubemard (1981). Beaubemard dismisses the Christian union's claim of 3,200 adherents.

166

Chapter 8

ceau-les-Mines. In 1901, the mainstream union unilaterally called a strike. The Christian leadership opposed the stoppage. The primary union, acting entirely on its own, was able to obtain a complete shutdown of the mines. The Catholics could not mobilize enough workers to keep even a small number of shafts open. The 1,000 or 3,000 members who supposedly were on the Christian rolls, in short, did not translate into 1,000 or 3,000 workers who were willing to cross a picket line to support management. A more serious rift was that between the parliamentary reformists and the direct-action syndicalists. The FTSS was significantly more reformist than were most of the unions of France. Because of this, it did not belong to the CGT, nor did it participate in most national or regional syndical congresses. In 1906, the CGT passed the Charter of Amiens, a fundamental document in French union history that rejected any collaboration with political parties or the state (LeFranc 1967; Saposs 1931). Such a policy was complete anathema to the FTSS, which was wholly dependent on its elected parliamentarians, legislative reform, and favorable state intervention into local strikes. The syndicalists of the wider French union movement rejected parliamentary reformism as a form of bourgeois co-optation. They argued that the restraint and compromise necessary for effective political action prevented more radical challenges to the social order that could produce an overthrow both of capitalism and state authoritarianism. The advocates of direct action thus prefigured the unitaires in viewing the reformist FTSS as having sold out to the employers (Loire 92 M 114). The opinions of militants in other industries was echoed in a minority tendency in French coal mining itself. In the north, Basly and his supporters survived a modest challenge from such forces. In most southern settings, such as Carmaux, the pressure from employers was sufficiently intense to motivate unionists to act in concert rather than as factions. In the Loire, however, the tensions between reformists and syndicalists generated a full-fledged schism in the union movement. This schism was based on the bona fide economic divisions that existed within the Loire-area coal field. The split, however, did not degenerate into the all-out mutual noncooperation that came to characterize the years after 1923. The two unions continued to plan activities jointly; furthermore, the moderates in some cases adopted the strategies and demands of the syndicalists. As a result, they benefited from the same accommodationist treatment that unified unions received elsewhere. The schism in the Loire ran from 1903 to 1911. Antiparliamentarianism was drawn to a large extent from frustration over managerial

Schism and Steering

167

perversions of state-generated conciliation mechanisms. In 1897, one of the parliamentary reformists, Cotte, had organized an inquiry into the shop-floor grievances of miners. Cotte had been working with socialist parliamentarians to provide an expose that could be formally presented to the Chamber of Deputies. The Loire companies systematically fired every worker who testified for Cotte; not surprisingly, his supply of informants dried up, killing the project. This suggested to many miners that the benefits of working through political channels might be limited in the face of unimpeded managerial power (Loire 93 M 22). In 1900, the Loire miners struck. The government imposed arbitration on the companies, and the workers obtained a generous settlement. Enthusiasm over this victory was tempered by subsequent attempts by management to renege on the spirit, if not the letter, of the agreement. The companies had agreed to both a salary increase and a reduction in the length of the workday. Many firms responded to the salary increase, however, by retarding promotions to keep workers classified in less expensive occupations; they also substituted boys for men. The workday limitation was weakened by forcing workers to come early to pick u p their equipment, or by preventing them from leaving in a timely way by barring them from elevators (Loire 93 M 22). Although the substance of the agreement was generally honored, the multiple petty acts of bad faith were inflammatory. Frustration with the FTSS and arbitration in general increased in the aftermath of the General Strike of 1902. In this case, a long, violent strike that was settled by arbitration produced a reduction of the prime and no immediate pension or workday benefits.^ Statistical analysis and historical hindsight suggest that the strike averted a more severe salary reduction; furthermore, the strike produced major pension and length of workday reforms in 1905. In 1902, though, the results of the strike looked thin. Saint-Etienne miners were particularly outraged by the union's administration of the strike fund, a significant proportion of which remained unspent at the end of the strike. The miners had seen union leadership tightly ration strike funds, push for arbitration, and then urge the acceptance of what was seen as a bad deal while the treasury still had ample resources. At the end of the strike, a crowd of militants confronted Cotte and beat him up, demanding either a return of the strike funds or a continuation of the struggle (Loire 92 M 114). In 1903, a rival union broke off from the FTSS. Its first few years were spent in basic organizing activities. The Loire miners' capacity for industrial conflict was vitiated by widespread demoralization over the ^See the previous chapter for a discussion of this negative result.

168

Chapter 8

1902 arbitration, as well as the continued weakness of the coal market. An attempt to obtain formal affiliation with the CGT stalled on the CGT's own concerns over accepting unions that did not represent entire occupations (Loire 93 M 22). By 1906, the passage of time, an upturn in the demand for coal, and recognition by the CGT revived the movement, making the new union a force to be reckoned with. The local insurgency, combined with a mobilization on the part of the national FTSS, produced a major strike in the Loire in 1906. During the strike, an important property of the schism became apparent. Support for the new union was centered in Saint-Etienne, whereas support for the old union was concentrated outside of the city, particularly in the concessions of Firminy and Roche-la-Moliere. (There was a small northern neck consisting of La Peronniere and Haute-Cappe that remained aloof from the conflict.) As was seen in the last chapter, for a variety of reasons, the firms of Firminy and Roche-la-Moliere were significantly more prosperous than were those of Saint-Etienne (Loire 15 J 903; Loire 92 M 138). It is not clear that Firminy's prosperity caused the reformism of its miners. Such an argument is tempting in the light of the fact that within French coal mining as a whole, moderate unionism was centered in wealthy Nord and Pas-de-Calais, whereas radicalism was concentrated in the more marginal mines of the south.3 Such an argument falls back on the traditional Leninist claim that the labor aristocracy is conservative (Lenin 1939; Perlman 1928). This logic is contradicted by multiple examples of highly radical wealthy workers, including communist American fur workers, the post-World War II leftist opposition to Swedish metal unions in prosperous plants, and for that matter, the militancy of the traditional French artisan (Foner 1964; Korpi 1978; Shorter and Tilly 1974). The relative leftism of Saint-Etienne is probably viewed better as a product of linkages between miners and other antistatist urban unionists, notably metal workers."* What matters for this discussion is not why the rich mines were less militant than the poor ones, but that the rich mines were less militant. The correlation of moderation and prosperity would have important implications for the redistributive consequences of striking. For much of 1906, the two unions continued to work together, de^One could advance this line of argument further by suggesting that the greater strength of communism in Nord as compared to Pas-de-Calais is attributable to the Nord mines' inferior market positions relative to those of their western neighbors. "•See Hanagan (1980) for a detailed account of interoccupational diffusion of militancy within the Loire.

Schism and Steering

169

spite their ideological differences. The salary offensive was directed jointly by a combined committee of leaders from both organizations. Early in the conflict, the prefect offered arbitration as a method for resolving outstanding differences between management and labor. Not surprisingly, the old union espoused this solution, but the new union opposed it. What followed was a combined election among members of both unions over whether there should be a strike or arbitration, as opposed to separate strike votes within the two unions. The election was hotly contested, and the results followed regional lines. Because Saint-Etienne was larger than the outlying areas, the prostrike faction won. At this time, the relationship between the two unions reached the point of greatest discord. In the light of the election results, the new union threw the members of the FTSS off the strike committee, filling the vacancies with antistatists. The miners of Firminy and Roche-laMoliere then signed a separate agreement with the companies and announced no further intention of striking. The form of this agreement was critical: The old union (formally the department federation of the FTSS) signed an agreement with the employers' association setting wages and terms for the entire Loire basin. Both the old union and management were, in essence, making a formal statement that the terms of the Firminy contract were to be applied across the Loire as a whole regardless of the outcome of the Saint-Etienne strike. The deal was pathbreaking. In return for the old union imposing its conditions on Saint-Etienne, management signed its first-ever regional collective bargain. All other wage settlements had been either directly negohated at the company level or forcibly imposed at the regional level by government arbitration. Such an agreement was predicated on the assumption that the Saint-Etienne workers would not be able to force their own terms on management, which turned out to be the case. Once the strike began, management made a formal proposal of arbitration to the new union. Aside from this, they neither bargained nor made concessions. The antistatists could hardly accept arbitration without acute embarrassment; furthermore, it was unlikely that government arbitrators would have supported the antiparliamentarians by providing them with a superior contract. As time passed with no concessions forthcoming, however, workers began to return to work, and there was a growing groundswell of opinion in favor of arbitration. In vain, the leaders of the new union petitioned the prefect to force management to negotiate seriously; they even approached Basly for his influence and support. They were rebuffed severely by both parties. The new union thus returned to work

Chapter 8

170

Table 8.1 Average Daily Salaries for Workers in the Four Largest Loire Mining Companies before and after the 1906 Strike

Company Underground Men Nonstriking firms Montrambert-Beraudiere Roche-la-Moli6re-Firminy Striking firms Mines de la Loire HouUleres de Saint-Etienne Women Nonstriking firms Montrambert-Beraudiere Roche-la-Moliere-Firminy Striking firms Mines de la Loire Houilleres de Saint-Etienne Boys Nonstriking firms Montrambert-Beraudiere Roche-la-MoIiere-Firminy Striking firms Mines de la Loire Houilleres de Saint-Etienne

Salaries before strike

Salaries after strike

Absolute change

Percentage change

5.35 5.00

5.57 5.23

.22 .23

4.03% 4.50%

4.80 4.80

5.03 5.03

.23 .23

4.68% 4.68%

2.18 1.83

2.28 1.93

.10 .10

4.48% 5.32%

1.66 1.55

1.76 1.65

.10 .10

5.85% 6.25%

2.85 1.87

3.05 1.97

.20 .10

6.78% 5.21%

1.66 1.87

1.76 1.97

.10 .10

5.85% 5.21%

Source; Loire 92 M 138.

with nothing in hand except the Firminy agreement (Loire 92 M 117; Loire 92 M 138). One might superficially conclude from this chain of events that because the Saint-Etienne workers lost overwhelmingly, they suffered economically from going on strike. Salary figures, however, suggest that just the opposite occurred. The Saint-Etienne workers obtained a greater increase in residual wages than did their more peaceful counterparts in Firminy. Table 8.1 shows the average level of raw wages for underground men, women, and boys earned by workers in the four largest Loire coal companies before and after the 1906 strike. Two of the companies did not strike and were party to the original Firminy negotiations, whereas two of the companies did strike. There is a common pattern to the distribution of wages within each

Schism and Steering

171

demographic category. First, nonstriking firms paid consistently more than striking firms; this was true both before and after the 1906 strike. This illustrates the arguments made earlier that non-Saint-Etienne-based companies enjoyed significant market advantages, which in turn produced higher market wages. Second, both striking and nonstriking firms received identical absolute increases in their salaries. Discounting the idiosyncratic exception of the Montrambert boys, absolute increases within each demographic category were within a percentage point of each other. This was the product of the universal application of the Firminy agreement. These two regularities produced a third: The striking firms received higher percentage increases in their salaries. Anytime a fixed increment is added to quantities of different magnitudes, there will be a higher percentage increase for the lower quantity. This higher percentage increase suggests a higher residual wage. The prestrike wages were likely to have been substantially affected by the underlying demand for labor, productivity, and coal prices experienced by the different firms. The greater percentage increase for firms with low base wages implies that workers in these firms were receiving a higher monetary return for these market factors. Put differently, workers in poor firms were getting a level of salary increases that was appropriate for rich firms.^ Of course, it could be argued that the global rate reflected the level of salary increase appropriate for the poor firms, and that the nonstriking firms were relatively exploited. Such a phenomenon was common in European industry-level bargaining in the 1960s, where the industriallevel contract was designed to be sustainable even by the most marginal companies, whereas more prosperous firms negotiated firm- and plantlevel supplements (Sturmthal 1972). The figures in Table 8.1 are observed wage payments that take into account whatever firm-level adjustments were made to the Firminy agreement. If salary payments were "appropriate" for strikers and "exploitative" for nonstrikers after such adjustments, this would still provide the relative advantage to strikers that is being argued here. It is more likely, however, that the Firminy agreement reflected market conditions for the stronger firms. At the time that the Firminy ^There were two exceptions to this pattern. For some reason, Montrambert boys were paid 50% more than any other juveniles and received a double wage increase. Removing them from the sample restores the larger pattern. The other exception is wages for aboveground men. These wages were almost identical across companies with the exception of the Houilleres de Saint-Etienne, which for some reason paid these workers well. There is no tendency for striking firms to get higher wages than nonstriking firms in this one group of occupations.

172

Chapter 8

agreement was negotiated, the official position of both the old and the new unions was that both groups would be going out on strike in accordance with the official strike vote. Management was trying to induce the rural unions to break their plans and not take a viable strike. Though it is conceivable that a set of strike-averse leaders might have accepted a sweetheart deal, the old union leaders had a long and credible record of militancy, including participation in the general strikes of 1900 and 1902. It is unlikely that management could have induced them to break syndical unity with the offer of a pittance. Some latent steering may have been occurring as well; a generous, peacefully obtained contract would have been an inducement for Saint-Etienne workers to opt for arbitration over striking. The widespread defection of the Saint-Etienne workers during the strike suggests that such a process may have actually been going on. The response of authorities to the 1906 strike was unintentionally accommodationist. Although every overt attempt was made to punish the antistatists and support the old union, the effect of the Firminy contract was to reward disproportionately those workers who were supporting the rebel union by going on strike. It can be argued that it was low pay rather than militancy per se that produced superior percentage increases in salary, and that the correlation between militancy and pay is spurious. Without the aggressive policies of the Saint-Etienne syndicalists, however, it is unlikely that management would have been driven to make sizable concessions. It had never voluntarily agreed to a regionwide flat-rate contract before; only the extraordinary circumstances surrounding the divisions in Loire-area unionism moved it to make this innovative response. The rejection of arbitration by the Saint-Etienne antistatists induced authorities to bolster the moderate union at all costs, even if it meant the unintentional rewarding of the more militant, poorer work sites. The implicit accommodationism of the 1906 strike became explicit in the government's handling of issues concerning the 8-hour day. In 1905, a new 8-hour law was passed as a result of 5 years of union agitation. To most modern readers, 8 hours implies 8 hours from the arrival at one's place of employment until the time one leaves one's place of employment. The 1905 law, however, only restricted to 8 hours the time a miner could be at the coal face itself. The time that a miner spent arriving at the pit, changing his clothes, getting his tools, waiting for the elevator, riding down the shaft, and walking to the coal face was not counted as part of the workday. In a typical deep shaft, this preparatory labor could easily take 2 hours, with a similar time spent at the end of the shift. (Delays at the end of the day could be even longer because of the time required for washing up.)

Schism and Steering

173

The passage of the 1905 law led to a number of controversies over the exact interpretation of the workday. At first, the government stood steadfastly by management and refused to make any concessions. A month after the 1906 salary offensive, the workers at Verpilleux in SaintEtienne began to put in only 8 hours in the mine itself. ^ They explicitly conceded management's position on getting tools and washing up, but included the elevator trip and travel underground as part of the 8-hour day. After posting two formal warnings, the Houilldres de Saint-Etienne locked the workers out and began to hire substitutes. The union sent a delegation to management that was turned away. The company argued that having been fired, the representatives were merely former employees, and as such it had no obligation to bargain with them. The company offered formal arbitration to the union leaders, but the antistatists who ran the Saint-Etienne union declined. The union's refusal to arbitrate was motivated on pragmatic as well as ideological grounds; given the unambiguity of the 1905 law, it is unlikely that the workers would have prevailed. The Verpilleux workers returned to work with nothing. Worse still, two important antistatist leaders, Escalier and Benoit, were fired. The prefect confined his intervention to sending in troops to maintain order (Loire 92 M 139). The state's policy of doing nothing was predicated on the inactivity and moderation of the reformist union. To everyone's surprise, however, in 1907 the moderates adopted the tactics of the antistatists and began to agitate for an immediate overriding of the 8-hour law. Throughout the fall, Bouchard, the leader of the old union, had been threatening to start a salary offensive because of improving conditions in the Loire coal market. In late December, he instead told workers to return their lamps half an hour early. His rationale was that the 8-hour law implicitly assumed an hour of rest as part of the workday; because most workers actually rested a half hour or less, by law they were entitled to leave early. Both the companies and the prefect denied the validity of Bouchard's reasoning; his initial logic was rarely referred to in either the subsequent negotiations or the settlement. This action by the moderates, however, produced an immediate and dramatic accommodationist reaction. The minister of labor—Rene Viviani, a prominent socialist—took a train to Saint-Etienne and personally attended a meeting with the Loire prefect, reformist union leaders. •*Actually, they only put in 9 hours in the mines. The 8-hour limit was phased in over time, with the legal workday being reduced to 9 hours in 1906, 8V2 hours in 1908, and 8 hours in 1910. The timing and speed of the phase-in was generally accepted in these early disputes.

174

Chapter 8

and the companies. The union demanded a reduction of 30 minutes in the workday, whereas both management and the prefect argued for maintaining the literal interpretation of the law. The minister suggested a fairly one-sided compromise by which the companies would grant the shorter day but would invest in a relatively expensive relocation of wood depots to make the workers more productive. When this proposition was refused by the companies, the minister essentially ordered the companies to cut the workday by 15 minutes and invest in the new depots. He argued that the consequences of a strike would be severe and that both workers and managers had to make concessions; however, only management gave anything up. Upon receiving the minister's directive, the companies devised two separate plans for cutting 15 minutes from the workday. Both were rejected by the union, as the reformists demanded and obtained a full 30-minute reduction. The only nominal concession that was made to management was the right to term the new workday an "acceleration" of the changes promised to the workers in 1910 (Loire 93 M 23). State policy during this schism had the effect of steering workers from the antistatists back to the moderate union. In terms of real benefits, the Saint-Etienne miners gained from both their own strike of 1906 and the reformist hours campaign of 1907-1908. The new workday applied to all mines in the Loire, without regard to union. In terms of visible concessions obtained, however, the new union appeared remarkably inferior to its rival. Neither the 1906 salary strike nor the SaintEtienne protests over the 8-hour day had produced any new concessions for the antistatists; in contrast, the reformists were able to obtain major new benefits without striking. The anarchist faction was wholly discredited. Some Saint-Etienne activists tried to save face by blocking the ratification of the Viviani deal at a key meeting. ^ Because of overwhelming rank and file support for the shorter workday, though, the Saint-Etienne vote was ignored, and the hours deal was approved. After this incident, the new union was for all intents and purposes irrelevant. The antistatists were involved in no further significant negotiations. The leaders of the two factions continued a war in the press on and off in the subsequent years; however, the two groups were moving toward conciliation. In 1911, there was a formal reunification of the rival organizaTewer than 3% of the authorized voters attended the meeting scheduled for the SaintEtienne ratification vote, allowing a small number of anarchists to swamp the election and obtain a narrow victory.

Schism and Steering

175

tions, leading to important combined agitations in 1912 and 1914 (Loire 93 M 23). The tactics used in the schism in the prewar Loire were more akin to whipsawing than to the classic pattern of steering observed after the war. The moderate union remained militant; by threatening a strike in 1906, it obtained an unprecedented regional agreement. By making sincere strike threats concerning the length of the 8-hour day, it was able to win instantaneously workday concessions that previously had been obtained only through multiyear parliamentary campaigns. In essence, the parliamentary reformists outperformed the advocates of direct action by taking direct action themselves. The government steered benefits in the direction of the moderates and, by doing so, effectively destroyed the antiparliamentary opposition. In the short term, the cost of this victory was the provision of significant costly benefits to the members of both unions. In this case, steering became accommodation. Accommodation occurred by accident in 1906, when strikers obtained higher residual wages than nonstrikers, then deliberately when the workers were given generous hours concessions not to strike. To the extent that the activity of the moderates was motivated by the need to compete for credibility with the antistatists, workers were receiving benefits indirectly as a result of the activities of the new union. In later years, steering would not produce accommodation; the existence of a moderate union that was absolutely and totally strike averse gave the authorities a party to reward whose participation in strikes was only minimal.

PURE STEERING IN THE POSTSCHISM PERIOD The 1922 schism was far more severe than any previous division among coal miners. Previous splits had involved minorities of insignificant strength or had been localized to single regions. This time, the division was national, with sizable forces on both sides. In previous splits, unionists of both tendencies had been willing to engage in strikes, but this time, the majority faction was committed to passivity. Previously, there also had been some joint coordination between the two organizations, but in the 1920s the reformists were committed to a policy of absolutely independent action. Between September 1924 and February 1926 alone, the confederes turned down no fewer than six separate requests by the unitaires to form a united front {Humanite, February 4, 1926).

176

Chapter 8

With the schism came an end to the great national coal strikes that had been a feature of French industrial relations since the turn of the century. Even during the 1919-1922 period, when within the CIT, the reformers (majoritaires) and the communists (minoritaires) were continually arguing among themselves, the FTSS was able to mobilize two general strikes in 1919 and 1921 that obtained major wage and condition improvements for coal miners throughout the country. After 1922, there were no well-supported general strikes until the mobilizations of the Popular Front. The pattern for the future became apparent during the Unitaire General Strike of February 1923, which produced a turnout of only 4% in Pas-de-Calais, less than 50% in the Loire, and 24% in Gard. The replication of these dismal rates in the November strike only confirmed labor's new weakness (AN F7 13903). Reformists, communists, and the authorities all viewed the unions as being impotent. Consider the following quote from Bartuel, the new secretary of the confederes, written the week that the schism was announced: "The situation created by all of these divisions [between communists and reformists] . . . as well as the presence of Polish miners makes any attempt at protests over the workday or working conditions impossible" (Rappel, 2/23/1922).^ Communists were less likely to confess weakness, because their position was that workers always had the potential to win great victories given ideologically proper leadership. During the Great Depression, Humanite ran stories of unions that had extracted concessions from managers during economic collapses (AN F7 13791). Minor unitaire leaders, however, sometimes gave analyses that deviated from the party line. Alain Tournier, the head of the unitaires at Decazeville, delivered this pessimistic assessment to a union meeting in 1925: "Partial strikes [i.e., strikes involving unitaires only] are doomed to failure. The companies profit from these by systematically firing unitaires. Before we make recourse to the strike tactic, it is necessary to have the unification of all workers" (AN F713793). Authorities agreed that the unions were weak and divided and employers congratulated themselves over the fiasco of 1923 (Pas-de-Calais 23 J). Police noted the tepidness of reformist wage claims in this period and predicted further moderation (AN F7 13792). The policy of steering was also the product of the bourgeois reaction to the Bolshevik Revolution of 1917. In the late twentieth century, it ^Bracketed texts in these quotes and elsewhere are my own supplements to fill in the sense of the quotation from the larger context of the selection.

Schism and Steering

177

is easy to view this revolution as an incident that would be largely confined to the Soviet Union. This was not at all obvious, however, to observers at the time. Leftists and rightists alike viewed the possibility of worker uprisings under advanced capitalism as very real (Maier 1975; Gallie 1983; Kriegel 1964). In France, these hopes and fears were made concrete by the Railway Nationalization Strike of 1920. This was a communist campaign to secure the nationalization of railways by the French state as a prelude to the full assumption of state control over industry. Industries were scheduled to strike in strategic waves: first the railwaymen, then workers in heavy industries, such as steel and mining; then workers in support industries such as municipal transportation and printing, and so forth. The communist minority of the CGT had designed the strike plan and, through careful organizational work, had gotten approval for its actions. The strike was run somewhat reluctantly, however, by reformist leaders of the CGT as a whole. The strike failed miserably. Turnout was low, even among the railwaymen, and many of the auxiliary unions provided only marginal support (Kriegel 1964; Clark 1930). Moreover, the presence of a national, centrally coordinated assault on the private administration of property galvanized the resistance of the French authorities, who initiated a sustained policy of controlling labor and repressing the left. As a result, the CGT was temporarily outlawed, and thereafter the influence of unions on public policy was dramatically reduced (Gallie 1983). The coal miners' union had supported the 1920 General Strike in only a limited way. Even modest participation, however, was too much. Central leadership had in rote obedience to the CGT given instructions to strike according to the national schedule, but support for mobilization was weak at both the executive and rank and file levels. The miners had recently won major wage concessions in their own mines-only general strike; there was little enthusiasm except among the communist minority for striking again on behalf of another industry. Nevertheless, approximately one third of the miners in the northern coalfields went out on strike. This was not enough to stop coal production or bring down the capitalist economy. It was, however, a good result in the light of the 5% to 10% turnouts that would be typical of unitaire general strike calls after the schism. The Loire, in contrast, obtained nearly 100% participation, as did a small number of other future unitaire centers (Pas-deCalais M 2382; AN F7 13791). After the strike, there were massive punitive layoffs. The Comite Central des Houilleres, with full government approval, intended to eliminate political strikes in the mines by using

178

Chapter 8

the complete range of coercive resources at its disposal. The FTSS petitioned Millerand for clemency for some of the strikers. The response from Millerand can only be described as nasty (Loire 15 J 2444). The post-1920 reaction, however, did not produce a systematically repressive regime whose sole response to worker organization would be the iron fist. Gallie (1983) is correct that the state was far friendlier to labor before 1920 than afterward. Nevertheless, welfare state legislation continued to be passed. The 8-hour-day law that was passed in 1919 was maintained in the 1920s, despite concerted employer offensives to reverse it. In 1922 and again in 1932, family allowance laws were passed that forced employers to pay premiums to workers with large families. Several major social insurance bills were passed throughout the 1920s, culminating in full sickness and old age coverage in 1930 (Saposs 1931; Loire 15 J 2444). In the coal industry in particular, miners' pensions were ameliorated several times during the 1920s. Furthermore, the FTSS, which was still united, obtained major wage gains in 1921 after a national strike. The national dependence on coal prevented the state from adopting wholly despotic policies. The syndical split among the coal miners could be exploited, however, to weaken the control of unions over the national energy supply, as well as the capacity of the communists to repeat the experiment of 1920. Authorities did what they could to isolate the unitaires. Among the most important policies was to systematically refuse to negotiate with the communists. Regardless of whether the workers in a given mine were represented by communists or reformists, the contracts that regulated pay and conditions were negotiated exclusively by the confederes. In areas where the reformists were dominant, such as the northern basin, the exclusion of communists merely created an American-style bargaining environment in which workers were represented by a majority agent. But in areas such as the Loire, where the communists were in the majority, such exclusion led to the de facto nonrecognition of the legitimate representatives of most of the workers, and to "show" contracts negotiated by small reformist minorities of the labor force. The communists were nearly always willing to bargain; in the north, the Loire, Moselle, Gard, and Aveyron, communists made formal wage proposals to the employers' federations. Such proposals were rarely acknowledged, let alone responded to (AN F7 13795; AN F7 13796; Loire 15 J 2449). In areas where there was no reformist presence whatsoever, management continued to negotiate reluctantly with the communists. This was the case in the Vendee, in Cantal, and at Saint-Hilaire in AUier (AN F7 13902; AN F7 13795). However, these examples were exceptional. All

Schism and Steering

179

of these were very small mines in remote rural areas; the settings were too isolated and invisible to involve any significant risk of creating a pattern or precedent. Furthermore, although the state disapproved of dealing with communists, it looked with even greater disfavor on the formal violations of French mining law that would have been necessary to refuse to meet with any workers' representatives. Even the presence of a nominal reformist union, however, was enough to disenfranchise the communists and set u p exclusive bargaining relationships with workers more to the taste of the authorities. The second strategy that was used to weaken the communists was to punish reformists who actively made a united front with the unitaires. This policy was used to great effect in the Unitaire General Strike of February 1923, in which most reformists scrupulously avoided providing any support to the communist mobilization. There were, however, two significant exceptions in Gard and Moselle. In Gard, the local confederes went out on strike with unitaires. In Moselle, the moderates were christian unionists rather than affiliates of the FTSS. Moselle like the rest of Alsace-Lorraine had a union structure based on German models, having been recently part of the Prussian state. The Moselle unions set policy independently from their counterparts in the rest of France. In 1923, both the christians and the communists were enthusiastic about striking. Plans for a regional strike were developed by a joint committee of union leaders. Similar events transpired in both regions. Salary negotiations took place between the local moderate union and the employers' federation. The communists, in a deliberate attempt to discredit the confedere bargaining teams, started their na? mal general strike while these negotiations were still in progress ana before any significant impasse had been reached. As soon as the strike beg^.a, the employers in both Gard and Moselle retracted offers that had been on the bargaining table and waited for the conclusion of the stoppage. The strike in Gard was over in a week; it was apparent by the second day that a major defeat was inevitable. In Moselle, workers stayed out for nearly 3 months after the rest of France had returne'.' to \v'ork; this extended strike produced no greater concessions, and, as was the case in Gard, the workers came back to work with nothing in hand (AN F7 13792; AN F7 13795). The consequences of participating in the strike were severe. The Moselle mines fired more than 1,000 strikers, replacing them with workers from outside the region (Moselle 310 M 20). The most important consequences, however, were economic. Elsewhere in France, the reformists had been negotiating for a pay increase. Management agreed to a generous improvement in wages, particularly in Nord and Pas-de-

180

Chapter 8

Calais. Gard and Moselle were excluded from these settlements and forced to work at their former salaries. Neither area was allowed to keep the offers that management had put forward in the aborted negotiations. A police informer noted that the salary increases given to peaceful reformists were an explicit attempt to "paralyze the movement of the unitaires" (AN F7 13791). The strategy had its desired effect. Unionists in both areas were demoralized. Disgruntled activists complained frequently about the good contracts that were given u p in 1923. Interunion relations were particularly acrimonious in Moselle because of the exceptionally long walkout there. Unionists dwelled for years afterward on the issue of who was responsible for the debacle of 1923. The matter was a common topic of formal union discussions; even in 1926, union business meetings could be broken u p by fistfights between activists rehashing the issues of that dispute (AN F7 13792; AN F7 13795). The bitter relationship between the christians and the communists forced the unitaires to act entirely on their own. The communists in 1925 attempted a 24-hour general strike by themselves, but they were unable to obtain any concessions. Management further humiliated the union by firing the longtime communist leader, Lucien Witz (AN F7 13795). In 1931, Moselle communists considered joining a national unitaire strike against pay cuts but decided not to participate. Their reasoning was that to be successful, they would need at least a 75% turnout, and they could not guarantee such a turnout without the active support of the Christians (AN F7 13791). The early successes of steering induced authorities to apply this policy elsewhere in France. In Aveyron in 1925, the communists were experiencing significant growth in membership from a small initial base. A set of wage negotiations were in progress, with the workers threatening a strike. The prefect of Aveyron urged management to grant some form of concession. He wrote: An improvement in the daily rate, even if minimal, would have the effect of stopping the strike movement. It is necessary to add that the departmental communist party is paying close attention to these negotiations with the fervent desire that they fail. I have in vain been calling the attention of the mining directors to this especial point. The prefect's efforts, however, do not seem to have been made in vain. Management raised its offer after the selection was written. After some delay, the workers accepted the deal and did not strike (AN F7 13793). In Gard, dismay over the disastrous outcome of the 1923 strike and internal infighting between the confederes and the unitaires debilitated

Schism and Steering

181

the strength of the union movement. Police reporters noted that union membership was low, that strike funds for both organizations were absolutely empty, that leaders were practically begging for dues, and that the rank and file was alienated by the ideological divisions within the union movement. The communists continued to be strong particularly in centers such as Besseges, however, and held important positions as miner safety delegates and administrators of disability funds. The Gard comparues, acting without government prompting, moved to consolidate the position of the confederes. As the prefect reported to the minister of labor: Up to today, the employers' federation has taken account of the split among the miners between the CGT and the CGTU in making its decisions. The weakness of the second faction and the unwillingness of the first to cooperate with them permitted the companies the liberty of dealing with workers' demands as they saw fit. However, since the initial split of the unitaires and reformists . . . the successive conquests of the Communist party in the occupational and political arenas, namely, their success in recent elections for disability fund trustees and industrial arbiters, in cantonal elections, and in the recent by-election of Ales, has broken the equilibrium between the two factions, with a growing superiority among the extremists . . . . [There was a preemptive pay raise that was less than what the unitaires demanded but was equal to reformist demands. Management's concessions to the confederes] can be explained by the fact that the CGT delegation appears to have lost its self-assurance and its authority (AN ¥7 13793) In 1929, the Gard confederes rejoined the unitaires and engaged in another joint general strike. They were able to obtain substantial participation, with turnout exceeding 65% in most mines. The strike was a total failure, however, and participants returned to work under their original conditions. As a result of substantial violence during the strike, local managers had begun to institute a program of punitive firings. This was aborted, though, upon the receipt of a coded telegram from the company directors in Paris. The local engineers were told that although they could make public statements threatening reprisals, they were not to implement any severe program of punishments. The directors explicitly specified that worker morale was to be maintained in order to ensure reformist victories in the upcoming miner delegate elections (AN F7 13905). The presence of steering can be most graphically illustrated with material from Pas-de-Calais in the 1930s. Pas-de-Calais was a reformist

182

Chapter 8

stronghold that for most of the 1920s escaped any significant unitaire influence. Whatever communist activity existed was limited to the community of Polish immigrant workers, and even among the Poles, communism was less influential than it was in the Nord or the southern fields. But in the 1930s, a rump group within the union, called the "CGT of 32," began to agitate for a more militant posture and greater cooperation with the communists. At the same time, a number of workers at the local level were electing communists to positions of responsibility supervising pension and disability funds, as well as making united fronts with the unitaires for local strikes over workplace-related issues. In one case that drew a great deal of attention, the reformist workers of Avion (Pit # 4 of Ostricourt) joined u p with the communists on a job control strike and then switched loyalty and joined the communists permanently (Pasde-Calais M 2383). These early initiatives were the harbingers of the social explosion that was to come in the great mining strikes of the Popular Front. During the period of the present study, these activities were nothing more than omens; however, they aroused a great deal of concern among public authorities. Contemporary discussions of these developments made very clear that state regulators valued the contribution to social peace made by the reformists and were willing to manipulate salaries in order to keep the confederes strong. Table 8.2 provides four lengthy quotes from the internal communications of the prefecture of Pas-de-Calais. The selections are presented in thematic rather than chronological order. Bracketed texts represent paraphrasing of intermediate quotes; for clarity, I have changed the paragraphing from that of the original and italicized selections of particular importance. Most of the selections are from correspondence from the subprefect of Bethune to the prefect, but the sentiments expressed are not confined to that one individual. They are echoed elsewhere in the files from which these selections are drawn, in police reports throughout the district, and in the prefect's own reports to his superiors. The selected quotes show several themes. The first is that any increase in communist strength or weakening of reformist strength was something to be viewed with great consternation. Quote I refers to the social threat that communism poses to large industrial populations. Quote II is a paean to the stability and predictability of industrial relations that comes from having a single union with predictable, moderate leaders. A weakening of their position is viewed as leading the way to demagoguery and mob violence. Quotes III and IV explicitly argue for the importance of maintaining the authority of reformist leaders; Quote IV holds that such unionists help to keep the working class calm

Schism and Steering

183

and reasonable. Quote III links the confederes to political stability as well. Second, the political authorities used an absolute deprivation model to explain both worker allegiance to the reformists and the willingness of miners to abstain from industrial disruption. These quotes assume a world in which prosperity produces a contented, docile labor force, whereas poverty produces disgruntled radicals. There are repeated references to the adverse effects of unemployment. The prefects also note popular discontent over the growing Taylorization of the coal face and the intensification of work discipline. The prefects strongly echo the accounts of the embitterment of labor relations at the shop floor given by Hardy-Hemery (1970, 1984), Moutet (1986), and Hainsworth (1976). Third, the prefects explicitly linked wages to worker happiness. Every quote makes some form of correlation between the docility of the miners and the provision of some sort of concession on salaries. Pay was not the only issue with which the government concerned itself; Quote II in particular is drawn from a larger essay whose predominant thrust is the importance of reducing unemployment. There can be little doubt from these selections, however, that the prefects felt that pay concessions were the key to social peace. Fourth, there are oblique references throughout the quotes to direct government interventions toward reducing pay cuts and maintaining workers' income; these passages, in particular, have been italicized for clarity. The government terminology for such interventions was "measures to improve the situation in the coal industry." These records contain little information about corporatist policies in the northern fields. As was seen in Chapter 6, though, in the Loire during this period, the government generally attempted to reduce unemployment and forestall wage cuts by offering various forms of governmental assistance to the coal industry. These included public purchase of coal, regulated marketing schemes, and negotiated decreases in freight charges. The state did not unilaterally impose high wages on employers, as it did in the 1890s and 1900s. Authorities could negotiate for delays in the implementation of adverse economic policies, however, by asking firms to wait for state market interventions to take an effect (PC M 2387). This is the sense of Quote II, where government "measures" to support the industry are traded off against a reduction in the "brutality" of company measures, and of Quote III, where government assistance is explicitly tied to a palliation of salary cuts. Quote IV notes that public opinion crediting the government with obtaining wage benefits for workers was an important contributor to social peace. The Pas-de-Calais example illustrates several important principles

184

Chapter 8

Table 8.2 Quotes from Pas-de-Calais Public Administrators Concerning Steps to Bolster the Authority of the Reformist Union 1. May 5, 1933 a. Sub-prefect of Bethune to prefect [Preliminary discussion of united front between CGT and CGTU in Avion, which after a promising start has suffered a mild setback.] Nevertheless, a movement in favor of unity must not be viewed with indifference. One must take account of the daily work of unitaire propagandists in the miners' milieux. The spirit of the miners is evolving slowly but surely to the unity desired by all of the dues payers. It must be further noted that unemployment, the reinforcement of work discipline, the acceptance of salary cuts, and the successive failure of strikes and revolutionary demonstrations are indisputable arguments in favor of unity. [Memo finishes with discussion of united fronts at several locations.] b. Prefect's comments on above [Prefect adds information on other united front activity missed by subprefect.] All such joint committees end up being dominated by the unitaires. [Prefect notes the substantial publicity given such committees by the communist newspaper Humanite.] Acting on their own, such committees have little influence on the labor movement. It is a different story altogether when such committees receive support from the left. Communist methods are too well known for there to be any doubt that joint committees can have serious consequences for social peace in a working class region with over 400,000 workers. II. June 30, 1931 Subprefect of Bethune to prefect [Discussion of discontent among the miners. Among causes cited are unemployment and a Taylorite increase in work discipline. There have been several communist victories in elections for disability fund trustees. Many of these elections were in reformist strongholds.] The unavoidable conclusion is that morale among the workers in the coalfield is in complete disarray. The miners have lost whatever confidence they have ever had in their leaders. A strike movement is taking place of a spontaneous unorganized basis. . . . We risk reexperiencing the incidents which occured in the era when the old syndicat was not yet constituted in which the speaker of the moment would collect workers on street corners and lead them against the neighboring grocery or the house of an unpopular engineer. It is essential that measures be taken as rapidly as possible to improve the situation in the coal industry . . . so that the worker on receiving his fortnightly

Schism and Steering

185

Table 8.2 (Continued) pay will perceive a tangible amelioration and will become calm. I am doing my best to obtain from the directors of the companies a guarantee that their measures will not be put too brutally into force. III. January, 4, 1932 Subprefect of Bethune to prefect [Discussion of 1932 salary negotiations.] The leaders of the old union have two motives: one, to act quickly to cut off any communist maneuvers, and two, to avoid a strike. They intend to explain to the workers that a strike will produce no result. Unfortunately, the precedent of last year casts doubt on the viability of this plan. I have tried to learn the intentions of the companies. The government alone must judge if it is in a situation to influence the coal market sufficiently to reduce the importance of the reduction of salaries that has been suggested. . . . I believe that if a cut is applied, no matter how feeble the rate, it will result in a partial strike, great discontent in the work force and a new reduction in the authority of the confedere union, and no doubt, a repercussion in the legislative elections. IV. December 27, 1933 Police reporter to subprefect of Bethune The local workers' unions have organized a campaign . . . to march on the prefectures and subprefectures. In our sector, confed^res and unitaires, who have reconciled for the purposes of this meeting, have enjoyed a great success. The indifference which has marked the workers' responses to the call of their unions has disappeared. The unions had been agitating throughout December; the situation had not seemed serious because the miners were not disposed to listen to the militants. [Police reporter explains the peace by listing beneficial economic conditions that have resulted in low rates of unemployment]. Furthermore, it was rumored that the companies had given in to government pressure and would grant the payment of family allocations and unemployment benefits. . . . Anger increased when the management of the Cie. de Anzin announced they would cease to honor the current wage agreement as of January 1. . . . Under these conditions, the reformist union can no longer dominate the situation. It is impossible for them to make the workers understand that they must be calm and reasonable. Source: Pas-de-Calais M 2387.

186

Chapter 8

of steering. First, it shows that public authorities sought to raise the salaries of reformist workers to keep them from cooperating with communists. The quotations here strongly echo the theoretical arguments of Chapter 2. Second, they reacted particularly strongly to credible communist threats. The atmosphere of crisis that permeates these quotes has no analogue in Pas-de-Calais writings of the 1920s, when the reformist control of the north was complete. Authorities steered particularly strongly in response to successful communist organizing drives in otherwise moderate areas. Thus, it is not surprising that mixed departments enjoyed residual wages that were substantially higher than those found in either reformist or unitaire districts. Third, state policymakers explicitly justified these wage increases as tools for ensuring social peace. They were concerned with the disruptive and political effects of communism; the economic effects of high wages play a less prominent role in these discussions. Authorities were willing to impose high costs on the private sector as a strategy for simplifying their own jobs as maintainers of social order. Steering was a strategy for redistributing wages, not for lowering them. To the degree that low pay was correlated with social discontent, reductions in wages were to be avoided.

THE CONTINUING INSTITUTIONALIZATION OF LABOR RELATIONS Accommodation and steering were fundamentally different policies. There was, however, an important unifying principle behind both of them: the continuing institutionalization of labor relations by the state. In both periods, French state policymakers attempted to control industrial militancy by directing workers' protests into formal legal channels. It is hard to say whether the use of public mechanisms of conciliation reduced the volume of industrial conflict. Most contemporaries probably would have argued that state intervention had a palliative effect; it seems unquestionable that most prefects believed their actions were reducing industrial conflict and promoting social peace. A modern observer might argue that accommodationism at least was a spur to greater militancy on the part of workers seeking favorable governmental attention and that even steering could be manipulated for parallel purposes. Regardless of whether institutionalization truly lowered the number or severity of strikes, however, it certainly increased the state's control over the outcome of strikes. By making prefects and ministers party

Schism and Steering

187

to all of the major labor disputes in the nation, the government obtained a lever with which to control wages, working conditions, and indirectly prices and output. In an era where centralized control over the economy was weak, the state's active role in labor relations could be viewed as a substitute for Keynesianism in the regulation of accumulation. Why the state favored institutionalization could be the subject of extensive debate. That it favored institutionalization seems unproblematic. In the prewar period, the primary threat to state control of industrial relations was a recalcitrant management that fought bitterly against both unions and the republican state in order to maintain proprietary control over the workplace. In the postwar period, the primary threat was posed by revolutionary communists, who sought to expose institutionalized trade unionism as a tool for obscuring the extraction of surplus value and hoped to induce the working class to overthrow capital and the state alike. The bugbears of each period, however, had their counterparts in the other: The prewar years had the antiparliamentarian left, and the postwar years had a small body of noncooperative employers. Steering was used to contain the communist left.^ Accommodation was used to punish those employers who continued to defy government procedures and attempted to crush unions unilaterally, rather than treat with them legalistically. Upon the achievement of state domination of the industrialrelations process, the government was able to control the substance of negotiated agreements within coal mining and often to set wages and prices directly. The state in many cases came to close working agreements with management and permitted the unilateral setting of wages by the companies through the cynically formalistic use of the legal mechanisms of conciliation. Thus, the full attainment of the institutionalization of industrial relations in French coal mining in the 1930s led to the reduction of actual bargaining to a symbolic show through the use of Edelman concessions. The survival of accommodation after World War I can still be seen in the small handful of firms that continued to defy public demands for the institutionalization of labor relations. The two most consistent instances of postwar managerial refusal to bargain with reformists were in Carmaux (Tarn) and Ronchamp (Saone-Haute). In Carmaux, labor relations had been consistently contentious since the workers' victory of 1892; unlike the rest of France, management there steadfastly refused to deal ''No word stronger than contain is appropriate in light of the continued unitaire mobilization of workers at the shopfloor level and the agitations leading to the explosions of the Popular Front.

188

Chapter 8

directly with the union. Even though wages and conditions continued to be set through de facto collective bargaining, management insisted on negotiating exclusively through intermediaries, usually the prefect of Tarn. The prefect or his assistants were given the job of transporting written communications back and forth between labor and management or carrying oral messages between two parties located in different rooms of the prefecture. This somewhat spiteful show of symbolic resistance was not well received by the local authorities, who did what they could to create face-to-face negotiations between the concerned parties (AN F7 13905; AN F7 13796). At the same time, the company did whatever it could to undercut traditional conciliation procedures. In 1920, it was one of the only companies that systematically resisted arbitration as a strategy for strike resolution. In 1921, it deliberately lowered wages without the 6 months' notice required by the contract. The union challenged the legality of this maneuver. The company was defeated twice in court, although it ultimately prevailed on a higher appeal. Although the company had few friends in the prefecture or the ministry of labor, it did have allies in the legislature and the ministry of justice. Following the protests of a local senator, the minister of justice asked for the resignation of the district attorney who argued the union's case. In 1922, the company tried to get its mechanics and electricians to sign away their legal rights to pensions and disability payments in return for a modest increase in wages. When the workers refused to sign, the company threatened to fire any employee who would not waive his pension rights. This whole initiative was blatantly illegal; the company appears to have dropped the matter at a later date (AN F7 13791). Tarn was distinctive not only for the combativeness of its managers, but also for the extent to which the prefect sided with labor in working for higher rates of pay. In wage negotiations in 1920, 1923, and 1925, the prefect explicitly espoused the union position, asking the companies to raise the workers' wages. He often encouraged the company to make concessions analogous to those made in the north or in the Loire, even though those areas were substantially more prosperous than Carmaux (AN F7 13791; AN F7 13793). These governmental interventions were probably effective: In the postwar period. Tarn's residual wages were among the highest in the nation. Though this may have been the result of exogenous factors, the continuous presence of a frustrated prefect who had every reason to support the union in industrial conflicts is likely to have had a significant impact. The same events occurred on a smaller scale in Ronchamp, with a

Schism and Steering

189

more minor series of managerial violations. Here the company was willing to meet directly with the union. The company director, however, skipped meetings he did not want to attend as a strategy for avoiding the making of concessions. The company was also caught violating one of its collective bargains by fining workers for low production, a relatively rare occurrence. 1° As was the case in Tarn, the Saone-Haute prefect reacted to these circumstances by becoming a relatively strong advocate of labor. Although residual wages did not reach the same level attained in Carmaux, the prefect pressed the company for increases in wages and generally supported the union in its struggles with management (AN F7 13795; Pas-de-Calais M 2382). These cases of managerial noncooperation were relatively rare. The more common pattern was for companies to manipulate rather than avoid mechanisms of conciliation, thus securing their economic needs while remaining in full compliance with the law. The techniques capitalists used to enlist state support for their positions have been documented in Chapter 6. The organized lobbying of the coal operators was often very successful, because both French economic growth and defense policy depended on high levels of production in the coal industry. As a result, collective bargaining in the postwar period became characterized by the widespread use of the Edelman concession. As defined in Chapter 7, an Edelman concession is a modified form of Boulwareism in which the company preannounces an adverse change in wages or working conditions, then immediately retracts it in collective bargaining in favor of a milder demand. At this point, the company refuses to make any further improvements, on the grounds that it has already made a generous concession. Both the initial posting and the concession in collective bargaining, however, have been carefully choreographed before negotiations. The company is in essence participating in last-offer-first bargaining, although symbolically it has complied with its obligation to both bargain with and make concessions to the union. Edelman concessions are more likely to appear in management than union offensives because in these cases, employers could be reasonably iiThere were several reasons why there were relatively few successful administrative prosecutions of mines for violating the job control terms of collective contracts. My personal suspicion was that the objective number of violations was low until the 1930s, when there was a dramatic increase in Taylorization. The number of successful prosecutions was also lowered by a decided promanagement bias among the state's inspecting engineers and by the minimal protection of workers' shopfloor rights provided for in the contracts themselves.

190

Chapter 8

expected to make the first move. In French coal mining, these concessions could be identified by the following features: 1. A public announcement of the initial harsh managerial demand, rather than waiting to offer the demand formally in collective bargaining. Customarily, the new adverse conditions were printed on circulars and conspicuously posted throughout the workplace a day or two before negotiations were to begin. In contrast, management responses to union-driven salary offensives were delivered exclusively in face-to-face collective bargaining, even when management had formulated its response in advance of the formal session. 2. Announcement of the second concession very early in the formal bargaining session. There was rarely enough time between the initial posting and the second concession to suggest that management was acting on significant new information. Even when concessions followed the initial union presentation, the union responses were sufficiently predictable that their contents could have been used to generate a first offer by management. 3. No significant concessions after the Edelman concession. In contrast, in worker offensives, the final contracts went through several iterations of concessions until they reached a form acceptable to the union. Edelinan concessions can be found in the national wage contracts of 1922, 1927, and 1931 (Pas-de-Calais M 2382; Loire 15 J 2449; Loire 15 J 2239; AN ¥7 13791). All of these contracts were negotiated in economic downturns, when management was capable of obtaining a reduction in the prime. The 1926 and 1929 negotiations were significantly more untidy; these occurred during periods of economic expansion and involved unitaire strikes (AN F7 13795; Pas^de-Calais M 2387). Another prerequisite for Edelman bargaining was the tacit approval of the state. To some extent, this was produced by extensive managerial consultation with high government officials. Formal negotiations were generally preceded by informal consultations between representatives of the Comite Central des Houilleres and the ministers of labor or public works. The seal of state approval for each particular Edelman concession, however, came from pattern bargaining initiated by the state. Wage negotiations across France tended to occur in a reasonably predictable regional sequence. Most salary revisions would first be negotiated in the Saar. Subsequently, there would be a settlement in the northern basin; there could be significant slippage between the terms of the Saar and Pas-de-Calais agreements. Subsequently the Loire would

Schism and Steering

191

settle in the light of the northern contracts, and thereafter the remaining central and southern basins would negotiate following the model of the Loire contract (AN F7 13792; AN F7 13793; AN F7 13796)." What is important here is that the pattern of settlements started in the Saar. The Saar coal mines were in the public rather than the private sector, being property of the French state. When the state itself was imposing harsh terms on its workers, it gave carte blanche to private employers to be equally severe. This did not mean that labor in the coal mines was completely crushed. To reiterate arguments made earlier, residual wages did not decline after the schism; they were only redistributed toward reformists. The state in some cases used its influence to raise wages globally, notably in 1925, 1926, and 1929. Furthermore, a number of important welfare-state concessions were made to the miners, particularly in the area of family allowances (Saposs 1931). However, the full institutionalization of labor relations gave the state the relative autonomy to raise or lower wages at will, imposing a nontrivial level of corporatist control. Steering benefits toward a set of prostatist reformists who foreswore taking independent strike action in favor of parliamentary reformism left the coal miners wholly dependent upon the goodwill of their governmental patron. Many workers profited from this arrangement; however, the cost of these benefits was a loss of control over the industrial relations process.

A BRIEF DIGRESSION ON ETHNIC DIVISIONS IN THE LABOR FORCE The discussion thus far has emphasized the organizational segmentation of the labor force: the division of the working class into rival, noncooperating unions. No discussion of divided labor movements can be complete, however, without some sort of discussion of the ethnic or demographic segmentation of the labor force. When sociologists customarily discuss the segmentation of the labor force, they usually invoke not organizational segmentation, but demographic segmentation: the division of the working class into rival noncooperating ethnic or gender groups. A strong argument would be that capitalists provoke racial hostilities among workers as a strategy for diverting the energies of workers "There were exceptions, such as in 1921 and 1925, when the Loire settled prematurely; rapid action occasionally led to disadvantageous terms (Pas-de-Calais M 2375; Pas-deCalais M 2382).

192

Chapter 8

away from class conflict and toward ethnic conflict (Reich 1981; Commons 1913). More moderate and widely accepted versions of this theory argue that ethnic divisions inhibit workers' mobilizations for collective action when residential segregation and linguistic barriers produce bifurcated social networks that minimize cross-group communication (Gordon et al. 1982; Form 1985). Furthermore, even if ethnic groups cooperate fully, the presence of disadvantaged groups can lower wages simply because the minority workers are paid lower salaries as a result of discrimination (Farley 1984). In theory, this should have been an important predictor of wages in French coal mining. Although the labor force was primarily native-born French before World War I, afterward an acute labor shortage in the mines induced the coal companies to import foreign labor. These workers came primarily from Poland, with smaller contingents from Italy, Germany, and North Africa. Thus, throughout most of the postschism period, the labor force in French coal mines was approximately 30% foreign. These workers showed all of the characteristics of a reserve labor army, including vulnerability to discrimination: In 1928, Polish skilled underground miners earned 24 francs a day, whereas a comparable French miner earned 40 francs (Cross 1983; Reid 1985b). Oddly enough, however, the presence of ethnic minorities in the labor force did not depress wages. As has been reiterated many times, wages did not decline after 1923, despite the massive importation of cheap foreign labor. Even more surprisingly, if one includes measures of the ethnic composition of the labor force in the Parks models of Chapter 3, the percentage of miners that are foreign born is positively, not negatively, related to wage rates. The relationship is statistically significant at the .001 level; including ethnicity reduces all of the class-conflict-related variables to insignificance. These equations are not presented in detail because the ethnicity finding is highly unreliable. The measures of the percentage of all miners who were foreign born come from the French censuses of 1926,1931, and 1936. As a result, ethnic equations include only the nine observations from 1926 to 1935 and are not strictly comparable to the results reported in the rest of the book. Data attribution problems are severe. Furthermore, unlike the equations in the rest of the book, two thirds of the data on the key variable, ethnicity, are estimated. This is likely to produce significant measurement error. The positive relationship between ethnicity and pay, even if it exists, is misleading. Foreign workers were generally employed in highly prosperous firms where the demand for new employees outstripped the capacity of the local labor market. Even though aliens were paid less

Schism and Steering

193

than native workers, the presence of immigrant labor was an indicator of the capacity of a coalfield to pay high wages to attract workers. In 1930, 37% of the labor force of Pas-de-Calais and 40% of the labor force of Nord were foreign; these were the highest-paying departments in France. In contrast, only 22% of the labor force of Vendee, 25% of the labor force of Loire-Haute, 25% of the labor force of Gard, and 18% of the labor force of AUier were foreign. A parallel consideration applied for women and juveniles. Because women and youths could only work aboveground, they only appeared in large numbers in firms with extensive surface facilities. These tended to be companies engaged in the systematic secondary processing of coal, which could involve the washing or refining of coal to produce a purer and more expensive product, the creation of coal by-products, coke or agglomerates, or the production of electricity. The profit margins on processed coal or coal by-products were substantially greater than that Ifrom raw coal. Thus, any mine that had the facilities to use women would also be a company that could afford to pay premium wages for all workers, whether above or below the ground. Although women and juveniles were paid significantly less than men, the presence of women and juveniles in the labor force is again positively correlated with residual wages. As before, Pas-de-Calais and Nord led the nation in the use of women and children, with the share of non-adult-male workers ranging from 10% to 15%. The comparable figure for a typical southern department would be from 3% to 8%. Although industrial composition was important, it was not the only reason that the presence of foreign workers did not lower aggregate wages. An additional factor was that immigrant workers received support from their home governments that prevented the French from engaging in wholesale exploitation.i^ Of these immigrants, the Italians were the most protected. Although Mussolini was not especially supportive of the rights of Italian workers in Italy, he was adamant about maintaining their rights abroad. The remittances of Italian workers to their families was an important source of foreign exchange for the Italian government; this, along with a strong sense of nationalism, gave the Fascists a good incentive to protect Italian miners in France. The Italian emigration office had the legal power to ban any emigration of local workers that was not in the national interest. They reviewed potential contracts between French coal mines and Italian workers, refusing those that were not sufficiently lucrative. i^The material in this discussion of government support of alien workers comes entirely from the excellent treatment of this subject by Cross (1983).

194

Chapter 8

Furthermore, in 1927, the Italiari government began to discourage emigration, in the interest of both building local industry and encouraging settlement in Italy's African colonies. French employers had to raise their bids in order to compete with these state-imposed labor priorities. Italy was a relatively minor source of labor supply for French coal mines. Poland was by far the most important source of such workers. It too, however, began to imitate Mussolini's policies after observing the material improvements received by Italian workers. In 1919, Poland had obtained modest legal rights to control the flow of native workers to France. Because of high Polish unemployment, this treaty was renegotiated in 1924, giving more powers to French labor recruiters. Furthermore, in the early and middle 1920s, Polish workers obtained supplementary assistance from Polish nationals who had followed the miners to new homes in France. Priests and businessmen who serviced the newly expanding Polish communities used their small-scale influence to protest against more overt forms of injustice. In 1928, however, Poland began to restrict the supply of workers allowed to emigrate. In 1929, for example, it gave the French less than two thirds of the requested allotment. This provided diplomatic incentives in the years before the Great Depression for French employers to treat their Polish workers well.i^ Furthermore, foreign miners, particularly Poles, were active unionists. In many cases, they were the mainstay of local unitaire organizations. Although the confederes ultimately reached out to foreign workers, in their early years they took a nativist position, seeking to negotiation limits on foreign workers with state authorities. In contrast, the communists in the 1920s were eager to work with ethnic minorities; the support of foreign workers reinforced the communist message of the dangers of nationalism and the underlying unity of the worldwide proletariat. The unitaires held special meetings in Polish. Most of their publications were printed in both French and Polish. They also made significant overtures to North African workers. Though no attempt was made to communicate in Arabic, the unitaires held several protests against French military action in Morocco and attempted to include Algerians and Moroccans in all union meetings (Reid 1985b; HardyHemery 1984; AN F7 13795; Loire M Sup 548; Pas-de-Calais M 2383). Ethnicity was thus prevented from becoming a significant divisive force in French coal mining. This is in contrast to the American experience, where racial and national differences among workers have been instrumental in reducing labor militancy and transforming classwide i^During the Great Depression, Polish and Italian workers became redundant. Foreign workers of all nationalities were sent home in droves (Reid 1985b; Pas-de-Calais M 2383).

Schism and Steering

195

concerns into intraclass disputes over relative access to opportunities. The divisions in the French mining force were organizational, rather than demographic, and they reflected union ideology rather than ascribed status. French employers reduced militancy by steering and promoting reformism rather than by making nativist appeals to French national identity.

CONCLUSION The reader has seen that French authorities did steer benefits to confedere unionists as part of an explicit strategy of labor control. Attempts to promote reformism produced latent accommodationism in the prewar period, where radicals actually benefited disproportionately from concessions made to moderate leaders. In the 1920s and 1930s, however, the allocation of benefits became more precise. Overt attempts to crush reformists who joined communists in mixed strikes were enormously successful. Moderates who participated in united fronts received conspicuously worse settlements that did more docile reformists; these outcomes were well known to workers, who were then reluctant to cooperate with communists in the future. State policynr\akers explicitly pressured employers to raise wages in order to weaken the power of the unitaires and maintain social peace within their districts. Steering provided one set of limits to the viability of the strike tactic. Even unified labor movements, however, cannot afford unlimited strike militancy. The next chapter tempers the arguments of the rest of the book by showing when strikes can become excessive.

9 The Limits of Militancy

The book thus far has argued that the optimal strategy for workers to maximize their wages is to strike as frequently as possible. It is not important whether they win; nor is it important whether the workers strike over pay. This does not mean, however, that workers can make unlimited amounts of money by engaging in a policy of unlimited striking. There are three constraints that may limit the overall returns from an aggressive strategy. The first is that whereas strikes did raise wages, the magnitude of this raise was very modest. Under the best conditions, the effect of a strike was to raise wages only about 3%. This is not a trivial amount. Most workers would notice a 3% difference in their pay packets, but it is not substantively overwhelming either. Admittedly, this effect could be cumulated both by increasing the number of strikes per year and by considering the impact of many years of strikes. To obtain major wage increases, though, precisely such a long-term policy would need to be implemented. The second constraint is that for a strategy of strike maximization to work, strikes need to be relatively short. Striking is not cost free to workers; every day of striking represents a day of lost wages. If workers forgo too many lost days of being paid, their conflict costs can swamp any possible benefits from increased wage rates. At the most extreme example, if workers struck 365 days a year, they would never be able to cash in on whatever levels of wages they were able to achieve. More realistically, it takes very great increases in wages per strike to amortize even a month of lost wages. If workers can obtain their wage increases with 1- or 2-day strikes, they will earn far more money than if they have to engage in all-out battles of attrition. A third issue is that wages continue to be determined primarily by market rather than conflictual forces. Neoclassical considerations do not 197

198

Chapter 9

fix wages down to the last penny. At any particular historical moment, there is sufficient ambiguity about what the market wage should be that managers and workers can argue over the setting of such a wage, with workers benefiting from obtaining the higher figure. This range of indeterminacy, however, is small. Wages can only deviate from market levels when firms are effectively buffered from the impact of neoclassical forces. French coal mining companies, like most other companies, were not buffered in the aggregate from long-term competitive pressures. French coal mines were not a glorified civil service that could set wages by political fiat; employers who paid too much risked losing sales to regional or international rivals. Market forces thus provided limits to what could be obtained by class conflict.

THE IMPACT OF STRIKES

How much of an impact did strikes actually have on wages? The average wage gain per strike is easily calculated from the coefficients for total strikes given in the core regressions. For 1890-1935, this figure can be found in Model III of Table 3.1; for the preschism and postschism periods, it can be found in the Unity and Schism equations of Table 3.2. Note that in these tables, the coefficients were multiplied by 100 for clarity, whereas for this analysis the actual magnitudes are required. Thus, B for 1890-1935 as a whole is .0049; for the preschism period, it is .138; and for the postschism period, it is -.0986. The total strike variable was coded as a categorical variable, with the respective values being 0,1, 2, 3 or 4, and 5 or more. As such, in most cases (but not all), one strike produces a change of 1 in the total strike variable. Using such a shorthand overestimates the effects of striking by giving too much weight to the effects of strikes in areas that have already received high strike activity. Because the goal of this discussion is to show that the impact of strikes is small, equating the effect of one strike with a change of 1 in the total strike variable produces a conservative bias in the analysis by overestimating the effect. An intuitive sense of the substantive meaning of the regression coefficients can be obtained by dividing the Bs by the average value of the dependent variable (real wages) for each period. This figure shows what proportion of mean wages could be increased by a single strike. For 1890-1935, this figure is .001; for the preschism period, it is .031; and for the postschism period, it is -.018. Thus for the period as a whole.

The Limits of Militancy

199

each strike raised wages approximately one tenth of 1 percent. Such an effect is trivial. As Chapter 3 showed, however, the regression coefficient for the global period underestimates the effect of striking by falsely combining periods of accommodation and of steering. During the accommodation period, each strike raised wages 3.1%; during the steering period, each lowered wages 1.8%. Neither of these numbers is particularly large. This estimate can be increased by considering the entire range of strike activity rather than calculating the effect of a single strike. Moving from minimum to maximum strike activity involved a change of 4 in the metric for total strikes. This would imply an increase of 0.4% for the entire period, a positive effect of 12.4% for the preschism period, and a negative effect of 7.2% for the postschism period. These numbers are not overwhelming, but they are substantially larger than the singlestrike estimates. Employees calculating the size of annual pay increases would notice the effect of a 3% raise; they would certainly take great notice of a 12% raise. It cannot be said, however, that successful strike strategy could take an impoverished worker and turn him or her into a labor aristocrat. Militant and nonmilitant miners made relatively similar amounts of money, with the most strike prone earning less than 15% more than the most passive. Using the core equations from Table 3.1 and 3.2 may underestimate the effects of strikes by not taking account of the additional impact of large national strike waves. These equations use the national wage rate of coal mining throughout France as a control variable. It is possible, though, that national strike mobilizations raised wages globally. If this occurred, part of the benefit of participating in a strike wave may appear in the national wage coefficient rather than in the local strike indicator itself. This can be tested by reestimating the effect of strike activity after dropping the national wage rate as a control. When this is done, the absolute values of the effect of strikes actually decline. For the entire period, the b is reduced to .0016 from .0049. For the prewar period, the b declines from .138 to .034, and for the postwar period, the equivalent figures are —.0027 instead of —.0986. The only case where using this methodology produces a greater effect is in the postwar period, when using either methodology the effect of strikes is negative, and "greater" implies moving toward zero. This suggests that at a zero-order level, national wage rates suppress the true stronger effects of strike activity that were reported correctly in the original equations. Nevertheless, these effects are small.

200

Chapter 9

WAGES, STRIKE COSTS, AND STRIKE DURATION Identifying optimal strike strategies cannot be done without taking the question of strike costs into account. Obtaining wage gains through striking can be self-defeating if the costs of striking outweigh the actual gains. In order to estimate the relative profitability of striking, one needs to calculate (a) the average wage gain per strike, (b) the average cost of striking in terms of lost workdays per strike, and (c) the extent to which longer durations increase the average wage gain per strike. The last section reported the average wage gain per strike. To contrast the costs and benefits of striking, it is simpler to concentrate on estimates for single strikes. The analysis from this point will concentrate on the maximum estimate for the positive effect of striking—the .031 figure for the accommodationist period. In the steering period, strikes lowered wages; the additional wages that were lost during the act of striking only exacerbated the economic harm to workers. In any analysis of the period as a whole, forgone wages would dominate any costbenefit analysis, because the effects of strikes on wages were relatively trivial. It is only for the accommodationist period that there are enough potential wage gains from striking to justify the calculation of transaction costs. The cost of striking can be estimated easily. The dependent variable for this analysis as a whole has been the real daily wages received per workday; in the accommodationist period, this figure was 4.39 francs a day. Thus, every day of striking cost a worker 4.39 francs. If each strike raised wages .138 francs, then a worker could recoup the cost of a day of striking in 31.8 workdays. This represents a little more than 5 workweeks, assuming a 6-day week. After these 32 days, all of the wage gain attained by the strike would be pure profit for the workers. Considering that the new wage scale would be held in perpetuity (or at least would be the baseline used in future negotiations), it would seem that workers could amortize the cost of a 1-day strike fairly easily. Of course, any increase in the number of days required to complete the strike added significantly to the cost of striking. A 2-day strike would require 63.6 days to work off; a 1-week strike would require 190.8 days; and a 1-month strike would require 773.2 days. If a miner wanted to break even over the course of a year, for example, he could strike for 9.5 days. If he struck for only 9 days, he would earn money by the end of the year. If he struck for a 10th day, he would have lost money by year's end, although he would start making money at the beginning of the next year. Clearly, this analysis implies that workers were better off having

The Limits of Militancy

201

short strikes rather thar\ longer ones. Such a strategy would only pay off, though, if short strikes were as effective as long strikes at raising wages. This was true empirically for French coal miners, and it is probably also true for workers in most other settings. One might argue speciously that strike duration should correlate with wages, because strike duration measures the amount of resources that workers have invested in coercing management. A 1-day strike would represent the mobilization of only enough resources to deprive employers of 1 day of production, whereas a 1-month strike represents the mobilization of enough resources to deprive employers of 30 days of production. Hicksian models show the fallacy of such thinking. Neoclassical strike theorists make the eminently reasonable assumption that both firms and unions can anticipate the result of strikes in advance of their conclusion. When the outcome of the strike is particularly easy to forecast, a negotiated settlement can be reached without an actual stoppage. In more ambiguous settings, the strike may have to run for a brief period—to allow both sides to obtain preliminary estimates of their relative strengths—before negotiations begin in earnest. Every day that the strike goes on, both parties learn about whether the union can obtain an acceptable turnout, whether the employer can obtain replacement labor, whether the government is likely to intervene on behalf of one of the parties, and the general capacity of the two sides to inflict costs on each other. These models predict, quite plausibly, that the strike should end soon after both sides amass enough data to project the end result (Hicks 1963; Gramm 1986; Hayes 1984; Tracy 1986). Such a model implies that strikes should be longest when unions and managements have roughly equal levels of strength. In these cases, very precise, nuanced information is needed to predict the strike outcome, and the strike may have to run virtually to its natural limit before both parties are ready to bargain. One-sided battles are easier to prejudge. Both overwhelming union victories or overwhelming union defeats can be forecast quite easily. One would expect, then, that the population of settings with short strikes represent a mixture of cases of workers with either very little strength or a great deal of strength. Long strikes represent cases of workers with intermediate strength. To the extent that wages are correlated with worker strength, this implies that wages should near the global mean in both cases; thus, strike duration should have no significant impact on strikes. Table 9.1 shows the results of OLS regressions of the the residuals from Model II on Table 3.1 on strike duration. Model II regresses wages

202

Chapter 9 Table 9.1 Regression Estimates of the Effect of Strike Duration on Residual Wages Period

B

t

Effect of extending strike 1 month

1890-1935 1890-1922 1923-1935

.00254 .00233 -.00052

2.32 2.03 -0.96

0.076 francs 0.070 francs -0.016 francs

Note; The Ns for the analyses are 322, 220, and 102, respectively.

on the core economic variables and union status. Strike duration is the simple unweighted average of the days lost in each strike for each department-year. The methodology used here is different from that used in the previous analyses; Parks models cannot be used for this estimation, because data is missing for all cases for which no strike occurred. It is not relevant to talk about strike duration for departments in which no strikes occurred. A parallel problem existed for discussion of strike victories and pay versus nonpay strikes. In those cases, it was possible to avoid dropping cases and abandoning a cross-section timeseries format by reconceptualizing each concept into variables that could be relevant for every observation in the data set. This not only maintained a higher N but avoided contaminating the analysis with selection bias from unmeasured variables associated with the presence of strikes. No such solutions present themselves for the analysis of duration; as a result, a simple regression of residual wages on duration is presented. Table 9.1 shows the regression of the residuals of Model II on strike duration for the whole period and for the two subperiods. At the most superficial level, duration would appear to have an impact on wages: The coefficient for strike days is positive and significant for both the 1890-1935 as a whole and the accommodationist subperiod. The significance tests, however, are very misleading in this regard. If one considers the substantive size of the regression coefficient, the impact of duration is trivial. Remember that residual wages are measured in the same metric as raw wages, francs per day. Unlike in the previous equations, no multipliers have been added for clarity. These coefficients can be interpreted simply and intuitively as the number of additional francs that could be obtained from each day of striking. Thus, in the first equation, the effect of 1 day of striking is to raise wages .00254 francs. This may be statistically significant from zero, but it is only V4 of 1 centime. There

The Limits of Militancy

203

was very little one could purchase in the France of 1890 for V4 of 1 centime. Obviously, using a 1-day criterion is too harsh. One assesses the impact of strike duration in terms of additional weeks, months, or quarters. The right column of the table therefore presents the effects of wages on each additional month (30 days) of striking. The effect of duration remains unimpressive: Each additional month of striking obtained less than 8 centimes; in contrast, every day of striking cost nearly 5 francs in lost wages. This analysis suggests that strikes needed to be short in order to be remunerative. How effective were French coal miners at reducing the duration of their strikes to the point of economic viability? The median duration of strikes in French coal mining from 1890 to 1935 was 2.2 days. In the accommodationist period, it was 3.4 days; in the schismatic period, it was 1.8 days. For virtually every period considered, the average strike was substantially shorter than the 9.5-day limit on the economic viability of strikes. Of course, the distribution of strikes was highly skewed. Though most strikes ended in less than 4 days, some lasted a month or longer. It might be tempting to argue that long strikes were "fundamental" strikes, all-out battles in which workers established that the balance of power was in their favor, whereas short strikes were "peripheral" strikes that cashed in on the strategic advantages labor had obtained in the more substantial disputes. If this was the case, then extending the length of strikes might be important in this critical subset of conflicts, if not in the population of strikes as a whole. There is some element of truth to this. The first great Montceau-lesMines strike lasted 25 days; however, the second lasted only 4 days. Both raised wages, even though the second was technically a loss. It is inconceivable to imagine the second strike being as effective as it was without the presence of the first and longer strike. The 1892 Carmaux strike was relatively short for a major strike, running only 14 days, including Sundays, when no work would have been done. Most major strikes were longer. The assorted strikes that constituted the 1902 General Strike ran between 1 and 2 months, and the various strikes in the 1906 Courrieres strike wave lasted between 5 and 6 weeks. Major strikes tended to be long during the accommodationist period because of the time-consuming nature of the mediation process that was used. After a strike commenced, the local prefect would organize negotiations. If these negotiations failed, then arbitration would have to be agreed to. Then there would be disputes over the selection of the arbitra-

204

Chapter 9

tors, and the arbitration itself could take time. Furthermore, long strikes often generated parliamentary debate. These debates could produce ministerial directives, derived in part from political pressure, for prefects to threaten management with the loss of their mineral concessions. These relatively attractive outcomes for workers could not be produced by a 2-day strike. By 1906, formal collective wage bargains became more widespread, and with them the presence of governmental supervision of the negotiating process as soon as a deadline for contract expiration approached. This accelerated the process of public mediation, as well as the process of national officials committing themselves to either labor or management's position. Consequently, the duration of strikes declined. It could be argued that in the early period, however, workers needed to participate in at least one long strike to generate credibility for their bargaining threat. Once this one long strike demonstrated their organizational capacity and the presence of public support for their position, though, further long strikes were counterproductive. Fortunately, as the duration statistics show, workers seem to have been aware of some of these concerns. In general, workers kept their strikes to 3 days or less, in essence guaranteeing that the strikes would be profitable.

THE MINOR ROLE OF INDUSTRIAL CONFLICT The third limit to industrial conflict was that wages were overwhelmingly determined by market forces. By the time neoclassical factors had taken their toll, there was simply very little variance left for strikes and unions to influence. This can be illustrated graphically by considering the goodness of fit of models that do or do not include conflict variables. Unfortunately, a major limit of the Parks models presented here is that they do not provide any well-accepted measures of goodness of fit. To obtain such statistics, it is necessary to revert to OLS regression and to report R-squares. Such R-squares are low estimates of the true explained variance of the equations. This is because OLS regression is inefficient in the presence of systematic regionally based errors or autocorrelation. The excessive error terms generated by this technique underestimates the standard error of the coefficients and, with this, the i?-squares. Table 9.2 shows OLS i?-squares for the core equations of the book for the period as a whole and for each of the two subperiods. In general, the predictive power of these models is outstanding, with no equation explaining less than 72% of the total variance. Most of this explanatory power, however, comes from the economic control variables. Models

The Limits of Militancy

205

Table 9.2 R - S q u a r e s for O L S Regression M o d e l s of Wages Model

1890-1935

1890-1922

1923-1935

Economic controls only Economic + Union Economic + Union + Strikes

.79 .80 .80

.72 .75 .75

.78 .78 .78

with no class-conflict terms explain between 72% and 79% of the variance. Once wage rates have been predicted with that level of specificity, there is simply not very much of anything left to explain.i Another strategy for assessing the impact of class conflict is to assess the effect of class-conflict variables relative to the neoclassical variables in the main equations of Chapter 3. Table 9.3 shows the ratio of the effects of class conflict variables, strikes, and union status to the effects of the economic controls. The effect of each variable is defined here as the absolute value of the Parks coefficient multiplied by one standard deviation of the independent variable itself. The absolute value is chosen to minimize distortions from negative coefficients when multiple variables are summed. Because what is of interest is not only the strength of individual variables but the relative impacts of two sets of variables, conflictual and economic. Table 9.3 also shows the ratio of the effect of each conflict variable to the effect of all economic controls, and that of the effect of all conflict variables to this same global denominator. 2 Table 9.3 shows that, in general, the effects of strikes and union status were very modest. For the 1890-1935 period as a whole, the effect of each conflict variable was less than the effect of any neoclassical variable. The best ratios show union status as being 30% of the effect of employment, 23% of the effect of year, and 13% of the effect of productivity. Every other conflict effect on the chart is less than 10% of the effect of the control variable; in many cases, the magnitudes are of the order of 1% or 2%. Overall, the combined impact of conflict variables is less than 2% of that of the combined market variables, and the effect of strikes is only V3 of 1% of that of the combined controls. Ht would be tempting to argue that adding conflict variables adds only adds between 0 to 3 points of explained variance, and that therefore the cumulative impact of union variables is trivial. The OLS K-squares, however, underestimate the addition in goodness of fit produced by union status and strike participation. ^The ratio of the effect of all conflict variables to each individual predictor can be readily obtained by adding the ratio for strikes and the ratio for union status. Those numbers are not reported here to avoid cluttering an already very full table.

206

Chapter 9

Table 9.3 Ratios of the Effects of Conflict Variables to the Effects of Market Variables 1890--1935

Ratio to effect of Productivity Employment Coal prices Wholesale prices Percentage nonagricultural National salaries Year All economic controls Ratio of effects of strikes and union status (combined) to effects of all economic controls

1890--1922

1923--1935

Strikes

Union

Strikes

Union

Strikes

Union

.032 .075 .020 .019 .021 .011 .058

.129 .301 .080 .075 .084 .043 .231

.097 .177 .009 .147 .071 .048 .256

.122 .223 .086 .185 .090 .061 .322

.041 .593 .132 .191 .028 .027 .114

.744 10.7 2.37 3.45 .501 .478 2.06

.003

.013

.013

.017

.008

.149

.016

.030

.157

Note: Economic effects are defined as Parks coefficients times one standard deviation of the predicting variable. All coefficients are reported as positive regardless of true sign. Equations used are Model III in Table 3.1 and the unity-schism equations of Table 3.2.

It is not surprising that conflict variables work poorly for the period as a whole. After all, the Parks coefficient for strikes in that period was insignificant. The situation improves when one considers the two subperiods separately, although the impact of both strikes and unions remains relatively small. In the accommodationist period, all of the conflict effects remain smaller than all of the market effects. The best performance by strikes is relative to year, where it barely exceeds 25% of the latter effect. The best performance by union status is also relative to year, but it fails to exceed 33% of the effect. Combined, all conflict variables still have only 3% of the effect of the combined market variables; strikes only have a little more than 1% of the effect of the market variables. The most favorable period for the relative effectiveness of conflict variables is the postschism period. This comes less from an increased magnitude of the strike and union variables than from the general weakness of such neoclassical predictors as employment and year. The union effect actually exceeds the effects of four of the market-related variables; strikes, however, continue to be smaller than every neoclassical predictor. Even in this "best" period, the effect of conflict variables is less than

The Limits of Militancy

207

16% of the effect of the combined economic controls. The effect of strikes is less than 1% of the effect of the combined economic controls. Contemporary observers were more than familiar with the substantial impact of market forces in shaping wages as a whole. Nearly every bargaining session on wages invoked market factors as justification for proposed changes. The variable discussed most commonly was coal prices; however, production, stockpiles, and employment were also germane. Price, production, and employment data were routinely available to employers, unionists, and state regulators through the very statistical source used in this study, the Statistiques de I'Industrie Minerale. These figures were also broadcast through employers' association publications, trade journals, and the business press. Data on competing wages were also generally available. Government regulators, unionists, and employers routinely shared information on collective bargains within their respective camps. Furthermore, the press published data on foreign wages, often in the same sections that published data on the price of foreign coal and the volume of imports. Coal negotiators were knowledgeable about their markets; as transcripts of their negotiations generally show, they used this information in the process of wage setting.3 Superficially, it might not be surprising that wages in the 1890-1935 period were shaped by market forces. It might even be suggested that in other periods, market forces would be weaker and union forces stronger. Writers in the regulation-economics tradition have suggested that there are regimes of production, and that in early capitalism, wages will be more subject to neoclassical influences than they will under the impact of modern corporatism (Boyer 1979). French coal miners, however, enjoyed all of the benefits of state-regulated industrial relations that German and Swedish metalworkers received in midcentury. The government was an active participant in their industrial relations, and it placed explicit pressures on employers to raise wages. The workers were fully unionized and organized, with the full capacity to create national strikes that could disrupt production. They were the beneficiaries of extensive legal protections providing such statutory supplements to their wages as family allowances. They also enjoyed the benefits of Keynesianism as beneficiaries of government-induced expansion of energy

^Citing every example of the use of a coal price or an employment figure in wage negotiations would be both exhausting and not useful. These references are omnipresent. Typical cites, however, can be found in Loire 15 J 2239 and Loire 15 J 2446. An outstanding analysis of the psychologies of managers and workers in calculating a proper wage can be found in Trempe (1971).

208

Chapter 9

production for military purposes. If neoclassical forces dwarfed the effects of class conflict for Third Republic French coal miners, it is hard to see who else could have effectively beaten the market in any period.* The relatively primacy of economic variables may sit uncomfortably with some sociologists. Recent sociological writings on wages have emphasized the primacy of internal labor markets over external markets, and of structural factors over market factors. Such arguments make a great deal of sense in the prediction of individual wages. Neoclassical wage theories are almost useless at explaining the wages of particular workers. Such arguments claim that wages are linked to marginal productivity, but the marginal productivity of individual workers is nearly unknowable. Output is very difficult to measure unless supervisors watch workers constantly. Michael Burawoy's classic Manufacturing Consent (1979) documents the arbitrary and somewhat fraudulent ways in which productivity gets reported to management. At AUis-Chalmers, the monitoring of output was relatively rigorous; in most other settings, the slippage between productivity and pay would be even greater. This is particularly the case when one takes into account the accounting difficulties associated with assessing the contributions of the various members of a tightly knit team, or measuring the added value of such amorphous auxiliary functions as advertising, clerical work, physical maintenance, or personnel administration. The sheer impossibility of matching pay with individual productivity produces the well-known phenomenon of salaries being determined by the rules of bureaucratic internal labor markets. Seniority, education, and somewhat arbitrary personnel evaluations get rewarded in lieu of actual productivity. Systems of annual raises and lateral job transfers motivate workers as a whole to increase their productivity while creating a disjuncture between the performance and pay of particular members of the organization (Edwards 1979; Lazear 1981). These objections lose their force, however, when considering such aggregates of workers as occupations, firms, industries, and nations. Although the productivity of any one worker is unknowable, the productivity of large groups of workers is easily assessable. Employers know how much steel is produced by a given factory, or whether tool and die makers produce more on average than apprentices. They may administer internal labor markets that produce idiosyncratic results for ••The tendency of the corporatist regimes of central Europe to roll back wages and deconstruct national bargaining arrangements during the recessions of the 1970s and 1980s is a further illustration of the relative fragility of political sheltering from market forces (Streeck 1984).

The Limits of Militancy

209

individuals, but they can tailor the rules of these markets so that overpaid workplaces evolve toward lower wages and underpaid ones can evolve toward higher wages. The manipulation of classification systems, the retitling of jobs, or the strategic use of gender and ethnicity in job assignment can have dramatic consequences for aggregate pay rates while leaving the basic rules of internal labor markets intact. Aggregates of workers can obtain wages that are above market level. If this were not the case, neither the union nor the strike variables in this study would have any impact at all. The conditions under which a premium wage can be obtained, however, are quite specific. The employer has to be buffered in some way from market forces so that the payment of an arbitrarily high wage will not produce an immediate siphoning of sales to lower-cost competitors. Buffering from wage competition derives from two sources— illiquidity of capital, and illiquidity of knowledge.^ Many firms can charge monopoly rents because of the significant time it would take rivals to set u p a competing business. The creation of alternative suppliers may involve the building or refitting of physical plants and the establishment of reliable systems of supply, distribution, and marketing. All of these require the raising and commitment of funds, which is not an instantaneous procedure. Furthermore, many industries have legal barriers to entry, ranging from the time required to obtain building permits to the elaborate political struggles required to enter highly regulated and rationed industries. While rivals wrestle with the complexities of issuing bonds, buying real estate, designing machines, hiring and training a labor force, and obtaining the approval of the relevant authorities, preexisting firms can charge monopoly rents and use those proceeds to finance premium salaries. This window of opportunity may last only a year; under the right circumstances, though, it can last decades. Firms can also be protected from competition by proprietary knowledge. The primacy of American automobile companies in the 1950s and of IBM in the 1960s and 1970s was based on monopolies of the relevant technology of their trades. Even if rivals could raise the capital to mount an offensive, they would lack the scientific capacity to make a comparable or superior product. Proprietary knowledge does not have to be

^Market forces can also be restricted by illiquidity of labor. It is well known that workers can be geographically immobile, unwilling to cross occupational lines, and unaware of high-paying opportunities in nearby firms. These forces, though, keep workers from taking higher-paying jobs, not lower-paying ones. They are relevant to a discussion of how exploitation can persist over time, but not to a discussion of the persistence of premium wages.

210

Chapter 9

associated with particular employers; it can just as easily be located with particular unions. Artisans possess closely held technical skills without which the production of high-quality goods is difficult or impossible. These skills allow for the creation of craft unions and the enforcement of a premium wage rate. These workers are protected from wage competition by the institutionally enforced nonexistence of other workers with comparable training and abilities. As long as knowledge does not diffuse, both the monopolist firm and the monopolist union can enjoy the benefits of nonmarket wages. Unfortunately, in the long run, capital is perfectly liquid, and all knowledge diffuses. Administrative delays in the implementing of production plans ultimately come to an end; capital can be raised; buildings can be equipped; and staff can be hired and trained. Political barriers are harder to overcome, but just as capitalists can use political influence to see that their rivals are excluded, the competition can use influence to see that its entry will be facilitated. Though in the short run some markets can be absolutely closed, in the long run the appropriate use of financial resources, exploitation of political cleavages, and, where necessary, provision for local participation can provide the prerequisites for market entry. The current globalization of world production provides a compelling illustration of this process. Knowledge is as difficult to contain as production. Aspiring newcomers have resorted to such devices as hiring the engineers of competitors, sending allies to work for old-master firms, financing the foreign education of their employees, and purchasing and analyzing technologically superior products in an attempt to unravel the secrets of proprietary technology. By using such devices, the French learned the arts of English textile machinery and railway construction, the Japanese learned the secrets of American automobile production, and the world learned how to build personal computers. Union knowledge is just as easy to deconstruct. Taylorism, Fordism, and the Bravermanian degradation of labor have all been successfully employed by capitalists to break workers' control over the labor process and ensure the transmission of occupational skills from artisans to managers (Braverman 1974; Zimbalist 1979). The class power of French coal miners was not destroyed during the time period of this book. Market forces constrained the size of the premium that could be paid, but they did not eliminate the premium altogether. Local firms were limited in the price they could pay by the presence of foreign competition. They received some shielding from foreign suppliers because of the specific requirements of particular customers for special grades of coal, as well as the high cost of transporting

The Limits of Militancy

211

coal over long distances. If premium wages created a significant long price disadvantage for French coal, this could subsidize both the transport costs of foreign rivals and the transition costs for employers to shift to technologies using different brands of coal. Because these costs were very substantial, however, French coal mines enjoyed the protection of modest barriers to the entry of competitors, which in turn permitted the payment of a premium to their unions. The market position of coal finally collapsed. The 1920s and 1930s saw an increase in the global production of coal. New mines in Poland, Czechoslovakia, and other eastern European countries added to the number of new competitors at the same time that technological progress increased the yields from such traditional suppliers as Britain, Germany, and the United States. The latent overproduction was exacerbated by a global reduction in the demand for coal. The Great Depression produced a short-term crisis in consumption as the economic collapse of the primary Western economies gutted the demand for energy throughout the world (Comite Central des Houilleres Frangaises 1935; Peyerimhoff 1931; Sauvy 1969). The crisis of the 1930s appeared to be cyclical in nature. The depression, however, masked a long-term and more fundamentally adverse threat to French coal mining. The economies of the 1930s, 1940s, and 1950s were moving away from coal as a primary source of energy toward electricity and petroleum. As railways became electrified, as automobiles became more prevalent, and as steam disappeared as the fuel of industrial machinery and home heating, coal came to represent an increasingly trivial share of energy consumption. In the more stringent environment of the 1940s and 1950s, the relative disadvantages of French coal became more pronounced. French coal veins were smaller, deeper, and more gaseous than were the deposits of France's primary competitors. Furthermore, because French mines were often older than those of the British or Germans, they were more exhausted, required more maintenance, and involved the development of increasingly more peripheral, secondary deposits. By the 1960s and the 1970s, many mines had become entirely uncompetitive and subject to closure. The French state downsized aggressively (Reid 1985a). By 1990, the French national mining company was producing so little coal that it had to purchase German coal from the Ruhr to pay retired French miners the traditional pensioner's allotment of free coal for heating. It is not known whether this economic collapse eliminated the premium wages of coal unionists. This study does not continue past 1936, and thus the nature of residual wages has not been examined for this late period. It is possible that unionists with political power maintained

212

Chapter 9

some sort of premium on top of a declining overall base. Successful miners may have been like sailors in the crow's nest of a sinking ship, temporarily sheltered from immediate disaster despite continually worsening absolute and relative positions. Militancy was powerful but not omnipotent. One strike could raise wages 3% a year. Three percent a year is a noticeable amount. Nevertheless, the benefits of strikes could easUy be consumed in the costs of class conflict. Strikes had to be kept strategically short if they were to be effective. Strikes also could not undo the effects of economic adversity; when markets and prices deteriorated, the union premium deflected only a part of this impact. Overall, even if workers used ideal strategy, their well-being would be determined primarily by market forces. After economic conditions were fully taken into account, there were not very many discretionary funds left to account for. When they engaged in their historical battles, French coal miners were truly fighting over the residuals.

10 Making Strikes Pay

This has been an analysis of the relationship between strike strategy and the wages of late Third Republic French coal miners. Late Third Republic French coal miners by themselves are only of interest to a small number of French historians; however, the question that motivated this study— what strategies of class conflict will maximize the economic well-being of the working class—is of very general interest indeed. The title of this book promised that French coal miners would have lessons in the strategy of class conflict from which other workers could learn. What are these lessons precisely?

MILITANCY MAKES SENSE The Importance of Frequent Striking The empirical analysis showed that, so long as the labor movement remained undivided, the optimal strategy was to strike frequently. This was true both for the French coal miners' movement as a whole before 1923 and for the areas dominated by unitaires after 1923. It did not matter whether these strikes won or lost, nor did it matter what issues the strikes were fought over. All that mattered was that a large number of strikes occurred. Militancy raised wages by convincing authorities that workers were interested in striking for its own sake. By doing so, the latter induced employers to provide them with extra wages as a peace payment above and beyond the amount that might have been expected from their relative cost of striking. These findings clearly support Piven and Cloward's (1977) contentions that only continuous social disruption by the poor ensures a redistribution of wealth from the upper class to the lower 213

214

Chapter 10

class. So long as the labor force remains undivided by dual unionism— and the labor movement has any power whatsoever to inflict losses on management—unions should be willing to strike and strike and strike again: multiple times in the same year, multiple times in the same place. The squeaky wheel gets the grease. The Advantages of Pursuing Broadly Based Political and Social Agendas Workers also demonstrated militancy by abandoning economism. A critical finding from the analysis is that the increase in wages from nonpay strikes was greater than that from strikes that contained a salary demand. Demanding control over the labor process or demanding major political reforms is a credible way of demonstrating ideological commitment to a radical program. This ideological posture convinced authorities of the seriousness of the union's commitment to social conflict, and to strikes as a means of pursuing larger social goals. This finding is a direct rebuke of the narrow business unionism favored by American unions. U.S. labor intellectuals, borrowing from the tradition of Commons and Perlman, have argued that unions should not pursue political agendas because ambitious programs of social change are far removed from the bread-and-butter concerns of average workmen. Like Marx and Lenin, they have claimed that workers' consciousness tends to be narrowly economistic; unlike Marx and Lenin, they have not argued that this is an ideological obstacle to be conquered. Instead, they have claimed that socialism is incompatible with effective unionism. Socialism was supposedly a grafting of the ideals of bourgeois intellectuals onto the organizations of the working class; such an irrelevant agenda could only weaken the effectiveness of unions by alienating workers and diverting resources away from monetary goals. Orthodox business unionists are right in one regard: Communist programs did alienate workers. Let us briefly consider the French experience in this ligtit. Amdur (1986) has compellingly shown for SaintEtienne and Limoges during the syndical schism that workers disapproved of both the social democratic passivity of the reformists and the revolutionary ideals of the communists, yearning for precisely the economic unionism that Commons and Perlman advocated.i The material on French coal miners supports this assertion. Turnout for the great communist general strikes was appallingly low. The miners did not sup^To be fair, Gallie (1983) has shown widespread internalization of socialist goals for a random sample of workers in contemporary France.

Making Strikes Pay

215

port the 1920 Railway Nationalization Strike and provided only tepid backing for such political actions as the 1926 Sympathy Strike for the English General Strike, or any of the strikes opposing the Moroccan war. Nevertheless, the Wisconsin school is wrong in believing that radical unions are limited to advocating noneconomistic programs. The unitaires built support for themselves by pursuing bread-and-butter issues as well. The communists dominated elections for miner safety delegates and pension board representatives, because they were more vigorous than the reformists in pursuing issues relating to job security, work pace, safety, and benefits (Hardy-Hemery 1984). Likewise, the communists were the union of choice for Polish workers because of the former's outspoken commitment to equal treatment for Frenchmen and foreigners (Reid 1985b; Cross 1983). Although Marxist theorists may have condemned economism, the unitaires pursued salary demands vigorously. Advocates of business unionism may have been right that many French workers were indifferent to calls for ending the extraction of surplus labor. However, they could accept "intellectual" demands as long as they were accompanied by short-term economic demands. More importantly, pursuing social goals did not vitiate the campaign for monetary goals. There was little evidence that workers who received more of one type of benefit received less of another. During the accommodationist period, French workers often received substantial economic benefits from strikes oriented toward obtaining political or legislative goals. The 8-hour-day strikes produced salary increases well before the legislature came around to actually approving a shorter workday. Workers with high levels of noneconomic benefits, such as short workdays or safe mines, had higher wages as well. Business unionists could argue in their defense that unitaire areas had generally lower wages than did reformists. But this was attributable to the divide-and-conquer tactics of authorities reacting to a split labor movement rather than any implicit failure of Marxist unionism itself. The findings are likely to have been quite different had the study considered some period where 100% of the labor force had been communist. Some support for this comes from considering the preschism period. The pre-World War I miners' union was hardly an American-style business union; it held both general and political strikes, and it supported socialist political candidates. It also obtained very significant wage gains for its workers, even as it piled up legislative victories. Writers in the Wisconsin school might argue more lamely that political unionism might work in France but not in the United States, where working-class ideology is more conservative. Such an argument could be legitimately grounded in the American-exceptionalism argument

216

Chapter 10

Perlman (1928) developed in his Theory of the Labor Movement. This hypothesis claims that American workers are particularly averse to socialism because of the ethnic segmentation of our labor force, the weakness of the federal state, and the widespread reverence for private property born in the robustness of the American economy and ready access to the ownership of land. Though this argument sounds plausible in the abstract, it is hard to believe that American workers would have opposed most of the demands of French socialist unionism. The FTSS used political tactics to campaign for such demands as an 8-hour day, social security, health benefits, and government assistance to the mining industry when it was in economic distress. These benefits have been longstanding features of the American welfare state that have been well accepted and legitimated by the American working class. To be sure, the unitaires also added demands concerning the revolutionary overthrow of the state. The substantial backing of American miners for such organizations as the IWW and the communist Mine-Mill, however, suggests that levels of U.S. support for radical unionism were comparable to the minority support the French gave to the unitaires. The lesson here would seem to be that political unionism can be extremely effective at producing monetary gains for workers. Workers do not suffer economic harm from the pursuit of social agendas; in fact, these agendas may increase wages by providing compelling evidence of militancy. Radical unionism becomes less viable if a competing moderate union emerges, or if the union becomes so preoccupied with its social program as to forswear any pursuit of economistic goals whatsoever. The first danger is real; the second is less substantive. The Benefits of Losing Workers demonstrated militancy by striking in the face of sure loss. Losing was actually beneficial for French workers; they obtained higher wage settlements from strikes that were lost than from those that were won. This extremely counterintuitive finding can only be explained by invoking the relationship between militancy and wages. Losing strikes are more likely to be over radical demands that bring up all-out managerial resistance. Furthermore, unions that engage in strikes that seem destined to lose convince authorities of their willingness to strike for striking's sake, which is compelling evidence of their commitment to militancy. In this regard, the unitaires' performance was exemplary. They struck after complete and total losses, after experiencing turnouts of 5% or 10% during recessions and depressions, and after being abandoned at

Making Strikes Pay

217

the last minute by their confedere rivals. Virtually no economic or political development could dissuade the communists from striking. We would normally think of such a mindlessly aggressive policy as an invitation to suicide. In areas where communists were statistically predominant, however, the strategy worked well. In communist districts, workers made more money when the unitaire unions struck frequently. The importance of being willing to absorb losses cannot be overestimated. One of the major weaknesses of the American labor movement is the tendency for major defeats to strangle future offensives and conflicts. The PATCO and Hormel strikes were disasters for workers not only because those particular conflicts were long and expensive losses but because they effectively terminated future conflicts in those firms, buying employers lasting labor peace. Employers can afford to commit maximum resources to crushing labor conflicts if a one-shot investment in repression can guarantee a 10- or 20-year period of peace. Companies would be less willing to endure the costs of an extended strike if these expenses had to be repeated on an annual or biannual basis. The air traffic controllers in the PATCO strike, with their ability to shut down the nation's air transport, had an intrinsically strong bargaining position. Would the Reagan administration have been willing to fire and rehire 50% of the labor force 3 times in 3 years? Admittedly, there are constraints that prevent American unions from participating in unlimited striking. American collective bargains are legally binding, raising the possibility of breach-of-contract suits for midcontract strikes. Furthermore, it is one thing for union leaders to call a strike, and another to induce the rank and file to go along. These admittedly real obstacles can be reduced by a policy of short strikes. Quick returns to work may make full prosecution of striking unions uneconomic; they also minimize the cost to each worker of engaging in conflict, making workers more willing to gamble on multiple actions. The Importance of Keeping Strikes Short One of the great weaknesses of the American labor movement is its propensity to engage in excessively long strikes. American strikes are among the longest in the world (Ross and Hartman 1960; Walsh 1983). It is not known whether long strikes correlate with higher or lower wage settlements. They do, however, correlate with losses; the greater the duration of a strike, the more likely it is that it will produce no concessions for workers (Knowles 1952). The logic behind long strikes is that the more labor workers withhold, the more production will be withheld, the more sales will be lost.

218

Chapter 10

and the greater will be the economic damage done to employers. This may be true in some settings. If an employer has an inventory of finished goods that must be eliminated before he or she loses sales, then strikes may have to outlast the inventory. In all too many settings, however, it is likely that the marginal economic damage done to employers declines as the strike progresses. In the first stages of a strike, the employer is caught more or less by surprise. He or she does not have alternative laborers or production arrangements. The withdrawal of strikers means a relative shutdown of production, and with it a loss of revenue. Once the strike has started, the union has done all it can do; there will be little further threat unless massive state intervention or third-party support is forthcoming. In contrast, the employer can start making alternative arrangements to protect himself or herself from what is now a well-understood withdrawal of labor. He or she can put managers to work on the floor, hire replacement workers, shift production to other facilities, or negotiate delays with vendors and customers. A strike is most likely to be effective when the employer is facing an urgent deadline. The optimal tactic is to strike right before the critical date, when it will be difficult for the company to make alternative arrangements to honor its obligations. Transport workers who strike on busy holiday weekends exploit this tactic well; the employer has little choice except to either settle or absorb the costs of making secondary arrangements with the ticket holders. In many cases, however, employer strike costs will be discrete rather than continuous, and the passage of time will incur relatively fewer costs on the company. Replacement labor, once hired, will remain in place, often working for less money than the original workers. Alternative supplies will have been bought in bulk. Internal restructuring of other plants to absorb extra work will have been accomplished and can be continued at virtually no cost. Many of these expenses will have been financed with savings on salaries, revenues that increase steadily with every day the strike continues. These savings on salaries continue to enrich employers even if the union has a stranglehold on the business and no alternative sources of production are possible. The employer offsets his or her lost revenues with reduced expenses for labor and supplies. Unless workers can find alternative employment, they do not receive parallel compensation. In the battle of attrition that follows, most employers have deeper pockets than do workers. Once the employer is no longer caught by surprise or forced to make emergency adjustments, the capacity of the union to create economic damage diminishes, sometimes to trivial levels. Workers will win more strikes by blitzkrieg attack than they will by siege.

Making Strikes Pay

219

Excessive duration also reduces the economic advantages obtained by strikes. For French coal miners, only strikes that lasted nine days or less were remunerative. A case can be made that workers needed one long strike to break the legacy of autocratic employer control in settings where managers had never been challenged previously by a strong, governmentally supported union. Once the credibility of workers' bargaining threat had been established, however, there were no further advantages to duration. Every day of striking resulted in lost wages for workers; excessive duration could consume the entire wage gains of a successful work stoppage. Short strikes have other advantages, which were discussed in the previous section. They involve less of a drain on the union budget, demand a less catastrophic commitment of resources by workers, and do not stake the credibility of union leadership on the results of one big strike. By diversifying strike actions, the union diversifies risk and increases the probability of having some form of tangible result come out of one of its stoppages. Workers, rather than viewing strikes as enormous undertakings that risk bankrupting the union and themselves for little or no gain, will see them as casual adventures, like bets in poker, that can hazard minor losses for benefits in the aggregate. This will improve rank and file morale, increase the commitment of members to the union, and give the union leadership a substantially more credible organizational weapon. The strategic choice facing workers can be easily illustrated as follows. Imagine a union leader who, for some reason, could only strike thirty days in a given year. He could have one long strike of thirty days, thirty 1-day strikes, ten 3-day strikes. If the union leader could identify eight times in a year in which a short strike would seriously inconvenience the company, he could have eight 3- to 4-day strikes in which each day of striking produces an acute cost for the employer, as opposed to thirty days of strikes with maybe five to six days of acute costs and twenty-five days of routine costs. Too often American workers have attempted to defeat employers with one decisive blow; a potential threat that is too often neglected is the death of one thousand well-chosen cuts. The Dangers of Centralization The centralized control of strikes by union bureaucracies risks gutting the effectiveness of organized labor. Unions benefit from maximizing their militancy. Any union structure that increases the strike rate should be encouraged; any structure that pacifies unions is more suspect. Theoretically, it is possible to imagine a highly centralized strike-

220

Chapter 10

prone union. Early communists working on the Leninist model may have believed that if a radical leadership maintained control over the leadership, then a vigorous program of centralized strike calls could shake the working class out of its complacency and artificially induce mobilizations that otherwise would not have occurred. The dangers of such a policy, however, were correctly perceived in an earlier era by Michels (1962) and Trotsky (1969) and, more recently, by Piven and Cloward (1977). Powerful union leaders would come to value comfortable working relationships with members of political and economic elites over engaging in conflict for its own sake. Negotiations with authorities would lead to the internalization of capitalist values by the union negotiators themselves. Leaders would also develop a fear of losing conflicts, because this could result in their own removal from power. These combined forces lead to union executives becoming averse to conflict, resulting in the tendency (noted by all of the above authors) for the leaders of working-class organizations to "sell out" to the capitalists and institute formal limits on strike activity. The definitive test of the relationship between union centralization and strike activity does not exist; however, there is substantial evidence supporting the above hypothesis. Net of the historical material these authors present themselves, one can invoke Korpi's (1978) history of Swedish metalworkers, where increasing union centralization was used to impose internal constitutional limits on striking on an internal communist opposition that was not at all reconciled to the theory of class compromise. There is also Walton and McKersie's (1965) analysis of the cooptation of union negotiators in collective bargaining sessions and Myron Roomkin's (1976) statistical correlation of strike frequency with constitutional decentralization in a sample of contemporary American unions. The only case that can be made for the centralized control of strike militancy is that strategically sophisticated union leaders may be able to prevent workers from engaging in ill-conceived conflicts that would produce sure losses. Such an argument becomes less supportable, though, if losses are not related to wage attainments. It is, in fact, to union leadership's advantage to have a set of militants within the organization that central negotiators cannot control. This gives the union representative the desirable option of being extremely conciliatory in formal negotiations, generating goodwill by being wholly sympathetic to management concerns, while at the same time, being organizationally incapable of making concessions because of exogenous pressure from radicals within the rank and file. It is extremely desirable in any negotiation to have a bargaining agent who is

Making Strikes Pay

221

relatively powerless, incapable of making independent deals, and subject to the review of an absent but implacable distant authority. The "obstinate owner" or "radical rump" can reject concessions and make late-breaking unanticipated demands. This allows the negotiator the opportunity to deal in bad faith and renege on his or her word while behaving with complete personal propriety. All one has to do is sincerely bemoan the unreasonable policies of one's absent master. Centralized union power, by the same logic, makes unions relatively less powerful by giving their leaders the capacity to make crippling concessions. Union anarchy caused by multiple competing, aggressive factions increases the power of labor by preventing management from obtaining easy deals with longtime bargaining partners who have come to share comfortable cooperative norms.

BONA FIDE STRATEGIC DILEMMAS Dual Unionism and Steering The relatively straightforward recommendations of the previous section become more complex during periods of dual unionism. If two unions exist and one becomes notably more moderate than the other, then authorities may steer benefits toward the docile union and away from the militant union as a strategy for destroying the second and bolstering the first. If a full-fledged steering policy has been implemented, the optimal strategy for both unions is to eschew strikes; strikers from both factions will be punished. The French reformists obtained significantly higher wages than did the urutaires because of the moderate strike-averse posture. Furthermore, they were never able to raise their wages by striking, whether in reformist, mixed, or unitaire districts. The communists were also better off not striking; they lost wages from striking in any district in which there was a credible reformist opposition. It is tempting to imagine that workers could overturn steering by concerted militancy. One might think that if both sides struck vigorously, they could create a whipsaw effect that would have been to their common advantage. This is not supported by the data. Joint strikes never raised wages; in fact, in reformist and mixed districts, such strikes lowered wages dramatically. Moderates who collaborated with revolutionaries were punished. There were two exceptions to this rule of avoiding strikes. First, in districts where the radicals were free from moderate opposition, the former theories of strike maximization applied. It is not possible to apply

222

Chapter 10

steering to a setting where the militant union has entrenched control over the majority of workers. Communists continued to be able to strike effectively in heavily communist regions. The second exception requires more judgment. Not all authorities respond to dual unionism with full-blown steering policies. When the labor movement divided in the Loire in 1903, the French state continued its former policies of accommodationism. As a result, workers in the militant breakaway union obtained higher residual wages than did those of the more moderate mainstream union. Steering was a learned response that occurred at a particular historical conjuncture. It is not always clear that employers will respond to a schism with steering. One of the ironies of French labor policy in the period of dual unionism is that employers were not able to lower overall residual wages significantly despite the division in their labor force. All steering accomplished was a redistribution of benefits from centers of militancy to centers of passivity; the overall wage bill remained the same. Steering may easily occur, however, in other settings with dual unionism. In the 1940s and 1950s, American employers overwhelmingly supported noncommunist unions when faced with a communist threat (Cochran 1977). Likewise, in the 1970s, California grape growers supported the Teamsters over the UFW. It is not known in either of these cases whether the moderate unions actually obtained more money than their competitors, let alone higher residual wages. Such a finding, though, would not be implausible given the general repression that radical unions experienced during these periods. Thus, there are no tidy recommendations for whether unions ought to strike during periods of dual unionism. The only sure recommendation would be a policy of trial and error: Observe the results of one's own strikes and those of the other union, as well as the reactions of authorities to the schism. Afterward, make a careful guess about whether accommodationism or steering is in effect. In the former case, strike; in the latter, desist. There is, however, one unambiguous lesson that does emerge from this discussion: Radical workers should avoid dual unionism. If the goal of militants is to increase the strike activity of the labor force, then they want to create a situation in which this militancy will always be rewarded and strikes will be to workers' advantage. This condition generally holds when there is a single union. Dual unionism opens up opportunities for authorities to pacify the labor force by implementing explicit policies of steering. It is hard to see any tactical advantage dual unionism might give radicals that would compensate for increasing workers' vulnerability to

Making Strikes Pay

223

this repressive tactic. The case for dual unionism is that bourgeois union oligarchs can stifle the militancy of their unions with internal constitutional limits on the use of the strike weapon. If militant workers have their own union, it is reasoned, then they can call strikes at will. This, however, is less of an advantage than it would seem. Internal constitutional limits are relatively helpless at stopping workers who are intent on striking from going out. Many great historical strikes were opposed by union leadership at first, then only reluctantly espoused by the bureaucracy after the workers were out and the strike was a fait accompli. The early CIO sitdowns in automobiles and the great French strike wave of 1968 are particularly well-known examples of this phenomenon (Piven and Cloward 1977; Crouch and Pizzorno 1978). Union leaders cannot prevent strikes; all they can do is label strikes they do not approve of as "wildcats." Thus, there is no reason why leftists cannot operate effectively as militant rumps within mainstream unions. This is, in fact, how the minoritaire communists operated in the French labor movement for the first few years after the Bolshevik Revolution. Within this organizational framework, they participated in a number of major national strikes and won significant victories. Even if they were excluded from significant bargaining roles, they had the potential through union politics and through their self-generated strike capacity to press mainstream union leadership for concessions. Admittedly, the radical leaders themselves would haive preferred the autonomy and freedom of action that comes from having their own union. It was better to stay in permanent opposition under conditions that favored militancy for both factions, however, than to obtain leadership of a personal organization and gut the incentives for militancy for both sides. Tactics in the Face of Economic Vulnerability All of the tactics that have been discussed in this book assume that labor has a realistic strike threat. If labor lacks the capacity to impose costs on management, then there is no credible strike threat and workers should desist from striking, and, if the employer presses the point, unionization as well. The labor movement in advanced capitalist nations faces unprecedented challenges that have seriously undermined the viability of the strike tactic. The global expansion of capitalism has introduced new competitors into the workplace with weak or nonexistent unions that allow them to enjoy significant price advantages over unionized producers in the developed core. This includes both late developers such as

224

Chapter 10

Japan and Third World countries making their initial forays into international manufacture. At the same time, the expansion of air travel, the development of global telecommunications, and the reductions of bulk associated with modern high-technology products have reduced the role of transportation costs in manufacturing. This has allowed core companies who once had been forced by freight charges to manufacture close to supplies of raw materials or major markets to relocate their facilities in remote areas where labor is significantly cheaper. The threat of regional deindustrialization, and the relocation of jobs out of the developed core poses severe market limits on the extent to which wages can be raised through industrial conflict. Intensified competition does not change the strike tactics of workers who can engage in bona fide collective action; it merely disqualifies certain workers from participating in industrial conflict at all. Any union that can claim a monopoly on skill or labor, or that can exploit an employer's long-standing vulnerability to deadlines, can continue to use the tactics suggested for Third Republic French coal miners, which are generally applicable for workers with independent bases of class power. The workers who cannot use these tactics are those who face severe competitive threats, or who risk immediate relocation of production facilities.2 The question here is whether cutthroat competition or disinvestment could occur in the relative short term. If economic sanctions are likely to be delayed until the long term, workers should use their labor power immediately to maximize immediate wages. Foreign competition expands inexorably; Third World wages will always be cheaper than those in the developed world, because of the absence of minimum-wage legislation and state-mandated benefits and legal protection. As soon as a Caribbean baseball factory comes on line, it will threaten the employment of baseball makers in New Jersey. The latter may as well maximize their wages during their 5-year grace period before internationalization destroys their jobs once and for all. Even under the catastrophic conditions that face the labor move^This discussion will bypass a significant third source of weakness, the presence of government repression. The economics of French coal were sufficiently comparable to those of modern industries that by considering the range of variations in both coal mining economies and modem economies, it is possible to make educated guesses as to what types of tactics are financially feasible. Both late Third Republic France and modern advanced capitalist states provide basic legal protections for unionized workers. How strikes might work in the absence of such protections requires the analysis of labor activity in more authoritarian regimes.

Making Strikes Pay

225

ment today, there is generally a subset of workers who are relatively immune from the forces of foreign competition and job relocation. These are workers in industries where service must be provided locally. Government is one such sector; the Internal Revenue Service is unlikely to turn tax collection over to the Japanese government no matter how efficiently the latter could perform these duties. Other industries that are forced to supply service in geographically predetermined areas include construction, transportation, medical care, hotels and restaurants, utilities, education, and retail distribution. Most of these jobs are in the service sector. It is not true that all service industries provide geographically fixed employment; financial services and entertainment, for example, are extremely mobile. Many services, however, have the same geographic limits on competition that would allow for the development of a credible union threat. Although the present crisis endangers unionism in many other industries, workers in these sectors should be relatively unaffected. This does not mean, though, that all of them could start striking tomorrow. Workers in any industry have to assess whether they employ meaningful monopolies of skill, whether their employers are vulnerable to pressure from deadlines, and whether they can mobilize enough workers to make mass rehirings prohibitively expensive. Furthermore, some industries, such as construction and transportation, are cyclical, and workers in these can only take action during favorable conditions. Many service workers who are generally unorganized today, however, would fit all of the above criteria. It is hard to explain, given the general shortage of nurses in this country, the extraordinary training required to create replacements, and the nondeferable nature of medical services, why nursing does not have one of the most militant and successful unions in the United States. It is also hard to see why the air traffic controllers were crushed after only one strike. These workers continue to maintain the potential to shut down the entire nation; furthermore, training requirements limit management's access to replacement labor. Similarly, it is hard to see why strikes by teachers, police, and fire fighters are not more common, particularly in big cities where the number of individuals to be replaced would make large-scale firings extremely impractical. Though laws may exist restricting the strike activity of public officials, these laws are difficult to enforce on a massive basis. Could one really incarcerate every policeman in New Orleans? This present call for more militant service unionism does not imply an abandonment of traditional manufacturing unions. Many smokestack workers continue to have credible strike threats. Admittedly, workers in

226

Chapter 10

outdated Rust Belt facilities, making inferior products under antiquated conditions, cannot afford to strike. Job termination for them is imminent, and minor changes in the economic calculus could lead to the immediate elimination of their plant. But while Chrysler workers in Kenosha might not dare strike, does this same limitation apply to General Motors workers in the Saturn plant? Admittedly, the latter were carefully screened for antiunion attitudes and located in a region with few competing economic opportunities (although this is changing very rapidly with the widespread relocation of manufacturing to the states of Kentucky and Tennessee). General Motors, however, cannot afford significant delays in Saturn production. Labor action by a properly motivated work force could be substantially more effective here than it would be elsewhere at General Motors or in U.S. auto production.

CASHING IN ON THE GOOD YEARS Striking can be effective, but it is not a panacea for workers' problems. Strikes need to be kept short, and the effect they have on wages is modest. Ultimately, employers learn to sidestep unions through technological change and disinvestment, destroying their capacity to obtain premium wages. Because the analysis of residual wages is a new science, little is known about the effectiveness of national labor movements in creating economistic gains for workers. Careful analyses of other populations besides French coal miners may uncover a more general rule, the tendency for workers to evolve from a state of earning market wages to a period of high militancy combined with premium wages, followed by a long-term withering of the strike weapon and the slow reduction of residual wages back toward market level. Even during the middle period, residual wages would probably only be somewhat higher than the market wage; union power correlates with market factors being at their most favorable level. Labor movements do vary, however, in their capacity to cash in on the good years. Just because premium wages ultimately disappear, and just because the size of these premiums is modest, there is no reason for workers not to maximize their benefits while they are sheltered from competition. Institutional commitments to labor peace, the avoidance of noneconomic issues, and reduction of strike militancy in the issue of tactical conservatism all do workers an economic disservice. Workers who fight over the residuals do have the potential for making more money. To get these residuals, though, their organizations really will have to buckle down and fight.

References

This study makes extensive use of archival sources. Most of these materials are identified in the body of the text with an abbreviation indicating the location of the materials, followed by the document classification number one would use at the archive to access the material. The abbreviations are as follows: AN or Archives Rationales: The French national archives in Paris. PC or Pas-de-Calais: The Pas-de-Calais departmental archives in Arras. L or Loire: The Loire departmental archives in Saint-Etienne. Saint-Etienne: The Saint-Etienne municipal archives. M or Moselle: The Moselle departmental archives in Metz. Note that the letter M frequently appears as a document classification for materials involving mining or public works. Only archival references that start with M or Moselle as opposed to merely containing the letter M can be found in Metz. SG or Statistiques des Grives: Material from either the quantitative records or supplementary narratives in the Statistiques des Greves. The SG materials used in this study were graciously provided by Charles Tilly from his own personal holdings. SG materials can be found in both American and French research libraries.

SECONDARY SOURCES Abadie, M. 1927. " Lutte Contre les Feux des Mines: Methodes Employees, Resultats Obtenus aux Mines de Decazeville." Revue de I'lndustrie Minerals 20:65-103. Amdur, Kathryn. 1986. Syndicalist Legacy: Trade Union and Politics in Two French Cities in the Era of World War L Urbana: University of Illinois Press. 1987. "Tradition Revolutionaire entre Syndicalisme et Communisme dans la France de I'Entre-Deux-Guerres." Mouvement Social 139:27-50. Aronowitz, Stanley. 1973. False Promises: Shaping of American Working Class Consciousness. New York: McGraw-Hill. 227

228

References

Ashenfelter, Orley, and George Johnson. 1969. "Bargaining Theory, Trade Unions and Industrial Strike Activity." American Economic Review 40:35-49. Audibert, Paul. 1933. "Generalites sur I'Organisation Rationnelle du Travail dans les Mines." Revue de I'lndustrie Minerale 26:355-385. Averitt, Robert. 1968. Dual Economy: Dynamics of American Industrial Growth. New York: Norton. Axelrod, Robert. 1984. Evolution of Cooperation. New York: Basic Books. Barbier, Marcel. 1929. "Les Houilleres Frangaises et Sarroises et le Mouvement en Faveur de I'Organisation Scientifique." Revue de I'lndustrie Minerale 22: 454-520. Beaubernard, R. 1981. Montceau-les-Mines: Un Laboratoire Sociale au XIXe Steele. Paris: Civry. Becker, Brian, and Craig Olson. 1986. "Impact of Strikes on Shareholder Equity." Industrial and Labor Relations Review 39:425-438. Berelson, Bernard, Paul Lazarsfeld, and William McPhee. 1954. Voting: Study of Opinion Formation in a Presidential Campaign. Chicago: University of Chicago Press. Bibb, Robert, and William Form. 1977. "The Effect of Industrial, Occupational and Sex Stratification on Wages in Blue Collar Markets." Social Forces 55:974996. Blackburn, Robert. 1967. Union Character and Social Class: Study of White Collar Unionism. London: Batsford. Blau, Peter, and Otis Dudley Duncan. 1967. American Occupational Structure. New York: Wiley Block, Fred. 1977. "Ruling Class Does Not Rule: Notes on the Marxist Theory of the State." Socialist Revolution (now Socialist Review) 33:6-28. Boncour, Paul. 1945. Entre Deux Guerres: Souvenirs sur la Troisieme R&publique. Paris: Plon. Bonnefous, Georges. 1956. Histoire Politique de la Troisieme Republique. Paris: Presses Universitaires de France. Boulin, Pierre. 1927. l!Organisation du Travail dans la Region Envahie de la France pendant I'Occupation. Histoire Economique et Sociale de la Guerre Mondiale. New Haven: Yale University Press. Bourgeaud, M. 1938. Salaire des Ouvriers des Mines de Charbon en France Depuis 1900. Paris: Presse Universitaire Frangaise. Boyer, Robert. 1979. "Wage Formation in Historical Perspective: French Experience." Cambridge fournal of Economics 3:99-118. Braverman, Harry. 1974. Labor and Monopoly Capital. New York: Monthly Review Press. Brecher, Jeremy. 1977. Strike! Boston: South End Press, Brogan, D. W. 1967. Development of Modern France, 1870-1939. London: Hamish Hamilton. Buder, Stanley. 1967. Pullman: An Experiment in Industrial Order and Community Planning. New York: Oxford University Press. Burawoy, Michael. 1979. Manufacturing Consent. Chicago: University of Chicago Press.

References

229

1983. "Factory Regimes under Advanced Capitalism." American Sociological Review 48:587-605. Cameron, David. 1984. "Social Democracy, Corporatism, Labor Acquiesence and the Representation of Economic Interest." In John Goldthorpe (ed.), Order and Conflict in Contemporary Capitalism: Studies in the Political Economy of Western European Nations, pp. 143-178. New York: Clarendon, Campbell, Peter. 1958. French Electoral Systems and Elections 1789-1957. New York: Praeger. Caron, Frangois. 1979. Economic History of Modern France. New York: Columbia University Press. 1957. Les Salaires dans I'lndustrie Frangaise (Les Charbonrmges). Paris: Colin. Chabert, Alexandre, 1957. Les Salaires dans L'lndustrie Frangaise (Les Charbonnages). Paris: Colin. Chaloner, W. H. 1973. Social and Economic Development of Crewe: 1780-1923. Manchester: Manchester University Press. Chapman, Brian. 1955. Prefects and Provincial France. London: Allen and Unwin. Clark, Marjorie. 1930. History of French Labor Movement 1910-28. University of California Publications in Economics, vol. 8, no. 1. Berkeley: University of California Press. Clegg, H. A. 1970. System of Industrial Relations in Great Britain. Oxford: Basil Blackwell. Cochran, Bert. 1977. Labor and Communism: Conflict That Shaped American Union. Princeton: Princeton University Press. Coleman, James. 1990. Foundations of Social Theory. Cambridge: Harvard University Press. Comite Central des HouiUeres Frangaises. Villa. Legislation Miniere et Legislation Ouvriere: Repertoire Methodique des Lois, Decrets, Circulaires, et Autres Documents Officiels Interessant les Mines et Leur Personnel. Paris: Comity Central des Houilleres Frangaises. 1921b. Rapport Annuel du President. Paris: Comite Central des Houilleres Frangaises. 1925. Rapport Annuel du President. Paris: Comite Central des Houilleres Frangaises. 1930. Rapport Annuel du President. Paris: Comite Central des Houilleres Frangaises. _. 1935. Rapport Annuel du President. Paris: Comite Central des Houilleres Frangaises. Comite des Houilleres de la Loire. 1925. Conventions et Accords sur les Conditions de Travail dans les Mines du Bassin de Saint-Etienne depuis le ler Janvier 1900. Saint-fitienne: Theolier. Commons, John. 1913. Labor and Administration. New York: Sentry. Cousineau, Jean-Michel, and Robert Lacroix. 1986. "Imperfect Information and Strikes: An Analysis of Canadian Experience, 1967-82." Industrial and Labor Relations Review 39:377-87. Cronin, James. 1979. Industrial Conflict in Modern Britain. Totowa, N.J.: Rowman and Littlefield.

230

References

Cross, Gary. 1983. Immigrant Workers in Industrial France: Making of a New Laboring Class. Philadelphia: Temple. Crouch, Colin, and Alessandro Pizzomo (eds.). 1978. Resurgence of Class Conflict in Western Europe since 1968. London: Macmillan. De la Taille, M. 1939. Consequences de la Reduction de la Durie du Travail dans les Mines du Nord et du Pas-de-Calais. Paris: Librairie Social et Economique. Derber, Milton. 1960. Local Union-Management Relationships. Urbana: University of Illinois Press. Dolleans, Edouard. 1953. Histoire du Mouvement Ouvrier. Paris: Colin. Domhoff, G. William. 1983. Who Rules America Now?: A View for the 80's. Englewood Cliffs, N.J.: Prentice-Hall. Duchemin, Rene. 1940. Organisation Syndicale Patronale en France. Paris: Bureau d'Editions. Dulot, Charles. 1923. "Present Position in the French Trade Union Movement." International Labor Review 3:695-716. Duncan, Greg, and Bertil Holmlund. 1983. "Was Adam Smith Right After All? Another Test of Compensating Wage Differentials." Journal of Labor Economics 1:366-379. 1966. Wage Determination under Trade Unions. New York: Kelley. Eckstein, Susan. 1977. Poverty of Revolution: State and Urban Poor in Mexico. Princeton: Princeton University Press. Edelman, Murray. 1969. "Conservative Political Consequences of Labor Conflict." Essays in Industrial Relations Theory. Ames: Iowa State University Press. Edwards, P. K. 1981. Strikes in the United States. Oxford: Basil Blackwell. Edwards, Richard. 1979. Contested Terrain. New York: Basic Books. Elwitt, Sanford. 1986. Third Republic Defended: Bourgeois Reform in France, 18801914. Baton Rouge: Louisiana State University Press. Farber, Henry. 1978. "Bargaining Theory, Wage Outcomes and the Occurrence of Strikes." American Economic Review 68:262-271. Farley, Reynolds. 1984. Blacks and Whites: Narrowing the Gap? Cambridge: Harvard University Press. Faur6, Petrus. 1956. Histoire du Mouvement Ouvrier dans le Department de la Loire. Saint-Etienne: Dumas. F6d#ration Nationale des Travailleurs du Sous-sol et Similaires. 1931. Congres National. Paris: Confederation Generale du Travail. 1923. Congres National. Paris: Confederation Generale du Travail Unitaire. 1924. Congres National. Paris: Confederation G6n4rale du Travail Unitaire. 1925. Congres National. Paris: Confederation Generale du Travail Unitaire. 1926. Congres National. Paris: Confederation Generale du Travail Unitaire. 1928. Congres National. Paris: Confederation Generale du Travail Unitaire. Fine, Martin. 1971. Toward Corporatism: Movement for Capital-Labor Collaboration in France 1914-36. Ph.D. dissertation. University of Wisconsin-Madison.

References

231

Foner, Philip. 1964. History of the Labor Movement in the United States. 4 vols. New York: International. Form, William. 1985. Divided We Stand: Working Class Stratification in America. Urbana: University of Illinois Press. France. Administration des Mines. 1937. "Le Marche Charbonnier." Annates des Mines (13e Serie) 11:1-433. France. Bureau de Documentation Miniere. 1900-1935. Statistiques de I'Industrie Minerale. Paris: R^publique Frangaise. France. Institut National de la Statistique et des Etudes Economiques. 1966. Annuaire Statistique de la France: Resume Ritrospectif. Paris: Republique Frangaise. France. Ministfere de Commerce, Office du Travail. 1890-1935. Statistiques des Greves. Paris: Republique Frangaise. 1894. Salaires et Durie du Travail dans I'Industrie Frangaise. Paris: Republique Frangaise. 1900-1912. Annuaire des Syndicats Professionels. Paris: Republique Frangaise. Freeman, Richard, and James Medoff. 1984. What Do Unions Do? New York: Basic Books. Friedland, Roger, and Jimmy Sanders. 1986. "Private and Social Wage Expansion in the Advanced Market Economies." Theory and Society 15:193-222. Friedman, Gerald. 1988. "Strike Success and Union Ideology, the United States and France, 1880-1914." Journal of Economic History 32:1-25. 1990. "Capitalism, Republicanism, Socialism and the State: France 18711914." Social Science History 14:151-174. 1991. "Decline of Paternalism and the Rise of Employers' Associations in France, 1880-1914." In Sanford Jacoby (ed.). Masters into Managers, 153-172. New York: Columbia University Press. Gaillard, Michel. 1976. "Le ler mai 1890 dans le Bassin Houiller du Card." Mouvement Sociale 94:59-76. Galand, Pierre. 1937. Les Mines du Nord et du Pas-de-Calais depuis la Guerre. Paris: Librarie Technique et Economique. Gallie, Duncan. 1983. Social Inequality and Class Radicalism in France and Britain. New York: Cambridge University Press. Garmy, Rene. 1934. Histoire du Mouvement Syndical en France. 2 vols. Paris: Bureau d'Editions. Georges, M. 1921. "UEffort des Mines du Pas-du-Calais Non Envahi de 19141919." Revue de I'Industrie Minerale 14:339-369. Gignoux, Claude. 1942. L'Economie Frangaise entre les Deux Guerres. Paris: Jouve. Gillet, Marcel. 1957. "Aux Origines de la Premiere Convention d'Arras: Bassin Houiller du Nord et du Pas-de-Calais de 1880 a 1891." Revue du Nord 39:111124. 1973. Les Charbonnages du Nord de la France au XIXe Steele. Paris: Mouton. Gordon, David, Richard Edwards, and Michael Reich. 1982. Segmented Work, Divided Workers. New York: Cambridge University Press. Gramm, Cynthia. 1986. "Determinants of Strike Incidence and Severity: A Microlevel Study." Industrial and Labor Relations Review 39:361-376.

232

References

Gras, L. J. 1922. Histoire Economique Ginerale des Mines de la Loire. Saint-Etienne: Theolier. Greene, William. 1990. Econometric Analysis. New York: Macmillan. Griffin, Larry, Michael Wallace, and Beth Rubin. 1986. "Capitalist Resistance to the Organization of Labor before the New Deal: Why? How? Success?" American Sociological Review 51:147-167. Gruner, Edouard. 1920. "Les Depredations Systematiques des Armees Allemandes dans les Houill^res du Nord de la France." Revue de VIndustrie Minerale 13:65-81. Hainsworth, Raymond. 1976. "Les Greves du Front Populaire de mai et juin 1936: Une Nouvelle Analyses Fondee sur I'Etude de ces Greves dans le Bassin Houiller du Nord et du Pas-de-Calais." Mouvement Social 96:3-30. Hamermesh, Daniel, and John Wolfe. 1990. "Compensating Wage Differentials and the Duration of Wage Loss." Journal of Labor Economics 8:S175S197. Hanagan, Michael. 1980. Logic of Solidarity: Artisans and Industrial Workers in Three French Towns 1871-1914. Urbana: University of Illinois Press. 1989. Nascent Proletarians: Class Formation in Revolutionary France. Cambridge: Basil Blackwell. Hardy-Hemery, Odette. 1970. "Rationalisation Technique et Rationalisation du Travail a la Compagnie des Mines d'Anzin, 1927-38." Mouvement Social 72:3-48. 1984. De la Croissance d la Desindustrialisation: Un Siecle dans le Valenciennois. Paris: Presses de la Fondation Nationale des Sciences Politiques. Hatzfeld, Henri. 1971. Du Pauperisme a la Securite Sociale: Essai sur les Origines de la Securite Sociale en France 1850-1940. Paris: Armand Colin. Hayes, Beth. 1984. "Unions and Strikes with Asymmetric Information." fournal of Labor Economics 2:57-83. Hechter, Michael. 1983. Microfoundations of Macrosociology. Philadelphia: Temple. Herding, Richard. 1972. Job Control and Union Structure. Rotterdam: Rotterdam University Press. Hicks, John. 1963. Theory of Wages. 2nd ed. New York: St. Martin's. Hill, Stephen. 1981. Competition and Control in the Workplace. Cambridge: MIT Press. Irvine, William. 1989. Boulanger Affair Reconsidered: Royalism, Boulangism and the Origins of the Radical Right in France. New York: Oxford University Press. Isaacs, Larry, and Larry Griffin. 1989. "Ahistoricism in Time-Series Analyses of Historical Process: Critique, Redirection and Illustrations from U.S. Labor History." American Sociological Review 54:873-890. Jencks, Christopher, Marshall Smith, Henry Acland, Mary Jo Bane, David Cohen, Herbert Gintis, Barbara Heyns, and Stephan Michelson. 1972. Inequality: A Reassessment of the Effect of Family and Schooling in America. New York: Basic Books. Kalleberg, Arne, Michael Wallace, and Robert Althauser. 1981. "Economic Segmentation, Worker Power and Job Inequality." American Journal of Sociology 87:651-683.

References

233

Kelley, Jonathan, and Herbert Klein. 1981. Revolution and the Rebirth of Inequality: A Theory Applied to the Natiorwl Revolution in Bolivia. Berkeley: University of California Press. Kerr, Clark, and Abraham Siegel. 1954. "The Interindustry Propensity to Strike—an International Comparison." In Arthur Kornhauser (ed.). Industrial Conflict, pp. 189-212. New York: McGraw-Hill. Kmenta, Jan. 1971. Elements of Econometrics. New York: Macmillan. Knowles, Kenneth Guy Jack Charles. 1952. Strikes: A Study in Industrial Conflict. Oxford: Basil Blackwell. Kochan, Thomas, Harry Katz, and Robert McKersie. 1986. Transformation of American Industrial Relations. New York: Basic Books. Kornhauser, William. 1959. Politics of Mass Society. Glencoe, IL: Free Press. Korpi, Walter, 1978. Working Class and Welfare Capitalism. London: Routledge and Kegan Paul. Kriegel, Annie. 1964. Aux Origines du Communisme Frangaise. 2 vols. Paris: Mouton. Kuisel, Richard. 1981. Capitalism and the State in Modern France. London: Cambridge University Press. Labi, Maurice. 1964. La Grande Division des Travailleurs: La Premiere Scission de la CGT 1914-21. Paris: Payot. Lafitte-Laplace, Jean. 1933. L'Economie Charbonni&re de la France. Ph.D. dissertation. University of Paris. Paris: Jouve. Laligant, Georges. 1919. "Du Taylorisme dans la Houillere." Revue de I'lndustrie Minerale 12:181-206. Langrogne, E. 1919. "Quelques Questions d'Avenir dans I'lndustrie Miniere." Revue de I'lndustrie Minerale 12:103-180. Lazear, Edward. 1981. "Agency, Earnings Profiles, Productivity and Hours Restrictions." American Economic Review 71:606-620. LeFranc, Georges. 1953. Le Syndicalisme en France. Paris: Presse Universitaire Frangaise. 1963. Le Mouvement Socialiste sous la Troisieme Republique. Paris: Payot. 1965. Histoire du Front Populaire 1934-8. Paris: Payot. 1967. Le Mouvement Syndical sous la Troisieme R6puhlique. Paris: Payot. 1976. Les Organisations Patrormles en France. Paris: Payot. Lenin, V. I. 1939. Imperialism: Highest Stage of Capitalism. New York: International. 1977. "On Strikes." In Tom Clarke and Laurie Clements (eds.). Trade Unions under Capitalism, pp. 57-63. London: Fontana. Lens, Sidney. 1959. Crisis of American Labor. New York: Barnes. Lester, Richard. 1958. As Unions Mature: An Analysis of the Evolution of American Unionism. Princeton: Princeton University Press. Levard, Georges. 1955. Chances et Perils du Syndicalisme Chretien. Paris: Librairie Artheme Fayard. Leveque, M. 1916. "Historique des Forges de Decazeville." Revue de I'lndustrie Minerale 9:5-97, 141-236. Lewis, H. Gregg. 1963. Unionism and Relative Wages in the United States. Chicago: University of Chicago Press.

234

References

Lorwin, Val. 1954. French Labor Movement. Cambridge: Harvard University Press. Loze, Edouard. 1891. La Greve de 1891 dans les Bassins Houillers du Nord et du Pasde-Calais. Arras. Machlup, Fritz. 1946. "Marginal Analysis and Empirical Research." American Economic Review 36:519-554. Macqueron, Pierre. 1904. L'Oeuvre du Syndicat des Mineurs du Pas-de-Calais. Lille: Robbe. Maier, Charles. 1975. Recasting Bourgeois Europe: Stabilization in France, Germany and Italy in the Decade after World War L Princeton: Princeton University Press. Mangez, Maurice. 1925. De I'Evolution des Salaires des Mineurs dans les Houilleres du Nord de la France et des Avantages CompUmentaires Accordes. Lille: Robbe. Mann, Michael. 1973. Consciousness and Action among the Western Working Class. London: Macmillan. Marks, Gary. 1989. Unions in Politics: Britain, Germany and the United States in the Nineteenth and Early Twentieth Centuries. Princeton: Princeton University Press. Marsden, David. 1986. End of Economic Man? Custom and Competition in Labour Markets. New York: St. Martin's. Martin-Saint-Leon, Etienne. 1923. Les Deux C.G.T.: Syndicalisme et Communisme. Paris: Plon. Marx, Karl. 1968. "Wages Prices and Profit." In Karl Marx and Friedrich Engels, Selected Works, pp. 224-229. London: Lawrence and Wishart. 1976. Poverty of Philosophy. London: Lawrence and Wishart. Mayeur, Jean-Marie, and Madeleine Reberioux. 1984. Third Republic from its Origins to the Great War, 1870-1914. New York: Cambridge University Press. Mazgaj, Paul. 1979. Action Frangaise and Revolutionary Syndicalism. Chapel Hill: University of North Carolina Press. Michel, Joel. 1974. "SyndicaUsme Minier et Politique dans le Nord et dans le Pasde-Calais: Le Cas Basly (1880-1914)." Mouvement Social 87:9-34. Michels, Robert. 1962. Political Parties. New York: Free Press. Miliband, Ralph. 1968. State in Capitalist Society. New York: Basic Books. Mishel, Lawrence. 1986. "Structural Determinants of Union Bargaining Power." Industrial and Labor Relations Review 40:90-104. Moore, Henry LudweU. 1911. Law of Wages: Essay in Statistical Economics. New York: Macmillan. Moutet, Amy. 1986. "La Rationalisation dans les Mines du Nord a I'Epreuve du Front Populaire." Mouvement Social 135:63-100. Oberschall, Anthony. 1973. Social Conflict and Social Movements. Englewood Cliffs, N.J.: Prenttce-Hall. O'Connor, James. 1973. Fiscal Crisis of the State. New York, St. Martin's. Offe, Claus, and Volker Ronge. 1975. "Theses on the Theory of the State." New German Critique 6:139-147. Okun, Arthur. 1981. Prices and Quantities: A Macroeconomic Analysis. Washington, D.C.: Brookings Institution. Oliver, M. 1922. Politique du Charbon 1914-21. Paris: Alcan.

References

235

Olson, Craig. 1981. "An Analysis of Wage Differentials Received by Workers on Dangerous Jobs." Journal of Human Resources 16:167-185. Olson, Mancur. 1965. Logic of Collective Action. Cambridge: Harvard University Press. Paige, Jeffery. 1975. Agrarian Revolution: Social Movements and Export Agriculture in the Third World. Berkeley: University of California Press. Parks, R. W. 1967. "Efficient Estimation of Regression Equations When Disturbances Are Both Serially and Contemporaneously Correlated." Journal of the American Statistical Association 62:500-509. Parsley, Clifford. 1980. "Labor Union Effects on Wage Gains: A Survey of Recent Economic Literature." Journal of Economic Literature 18:1-31. Perlman, Selig. 1928. Theory of the Labor Movement. New York: Macmillan. Perrot, Michelle. 1974. Les Ouvriers en Greve. 2 vols. Paris: Moutet. Peyerimhoff, Henri. 1925. "Le Probleme Houiller." Revue de Paris 22:22-24. 1931. "Le Charbon." Revue des Deux Mondes: 835-861. Phelps-Brown, E. H., and Margaret Browne. 1968. Century of Pay: Course of Pay and Production in France, Germany, Sweden, the United Kingdom, and the United States of America 1860-1960. London: St. Martin's. Piven, Frances Fox, and Richard Cloward. 1977. Poor People's Movements: How They Succeed, How They Fail. New York: Pantheon. Podgursky, Michael. 1986. "Unions, Establishment Size, and Intra-Industry Wage Effects." Industrial and Labor Relations Review 39:277-284. Pollins, Harold. 1971. British Railways: An Industrial History. Totowa, N.J.: Rowman and Littlefield. Poulantzas, Nico. 1973. Political Power and Social Classes. London: New Left Books. Prechel, Harland. 1990. "Steel and the State: Industry Politics and Business Policy Formation, 1940-89." American Sociological Review 55:648-668. Prost, Antoine. 1964. La CGT a I'Epoque du Front Populaire, 1934-1939. Paris: Colin. Przeworski, Adam. 1980. "Material Bases of Consent: Economic and Politics in a Hegemonic System." Political Power and Social Theory 1:21-66. Ragin, Charles, Shelley Coverman, and Mark Hayward. 1982. "Major Labor Disputes in Britain 1902-38: Relationship between Resource Expenditure and Outcome." American Sociological Review 47:238-252. Reder, Melvin, and George Neumann. 1980. "Conflict and Contract: The Case of Strikes." Journal of Political Economy 88:867-886. Reich, Michael. 1981. Racial Inequality. Princeton: Princeton University Press. 1985a. Miners of Decazeville: Genealogy of Deindustrialization. Cambridge: Harvard University Press. 1985b. "Limits of Paternalism: Immigrant Coal Miners' Communities in France." European History Quarterly 15:99-118. Robert, Jean-Louis. 1980. La Scission Syndicale de 1921: Essai de Reconnaisance des Formes. Paris: Sorbonne. Roomkin, Myron. 1976. "Union Structure, Internal Control and Strike Activity." Industrial and Labor Relations Review 29:198-217.

236

References

Rosen, Sherwin. 1969. "Trade Union Power, Threat Effects and the Extent of Organization." Review of Economic Studies 36:185-196. Ross, Arthur, and Paul Hartman. 1960. Changing Patterns of Industrial Conflict. New York: Wiley. Rubin, Beth. 1986. "Class Struggle American Style: Unions, Strikes and Wages." American Sociological Review 51:618-633. Saposs, David. 1931. Labor Movement in Postwar France. New York: Columbia University Press. Sauvy, Alfred. 1969. Histoire Economique de la France entre les Deux Guerres. 2 vols. Paris: Sauret. Schneider, Eugene. 1945. Charbon, son Histoire, son Destin. Paris: Plon. Sewell, William, and Robert Hauser. 1975. Education, Occupation, and Earnings. New York: Academic Press. Shorter, Edward, and Charles Tilly. 1974. Strikes in France. New York: Cambridge University Press. Siegler, Jean. 1909. "Sur le Developpement de I'Exploitation des Mines dans le Bassin de la Loire de 1855 a 1908." Revue de I'Industrie Minerale 2:5-29. Simard, Marc. 1983. "Situation Economique de I'Enterprise et Rapports de Production: Cas de la Compagnie des Mines d'Anzin (1860-1894)." Revue du Nord 65:581-597. Simiand, Francois. 1904. Salaire des Ouvriers des Mines en France. Paris: Soci6te Nouvelle de Librairie et d'Edition. Skocpol, Theda. 1979. States and Social Revolutions. Cambridge: Cambridge University Press. 1980. "Political Response to Capitalist Crisis: Neo-Marxist Theories of the State and the Case of the New Deal." Politics and Society 10:115-201. Smelser, Neil. 1959. Social Change in the Industrial Revolution. Chicago: University of Chicago Press. Smith, Robert. 1979. "Compensating Wage Differentials and Public Policy: A Review." Industrial and Labor Relations Review 32:339-351. Snyder, David, and William Kelly. 1976. "Industrial Violence in Italy 1878-1903." American Journal of Sociology 82:131-162. Spaeth, Joe. 1976. "Characteristics of the Work Setting and the Job as Determinants of Income." In William Sewell et al. (eds.). Schooling and Achievement in American Society, pp. 161-178. New York: Academic Press. Stearns, Peter. 1968. "Against the Strike Threat: Employer Policy toward Labor Agitation in France, 1900-1914." Journal of Modern History 40:474-500. Stephens, Evelyn, and John Stephens. 1982. "Labor Movement, Political Power, and Worker Parhcipation in Western Europe." Political Power and Social Theory 3:215-249. Sternhell, Zev. 1986. Neither Right nor Left: Fascist Ideology in France. Berkeley: University of California Press. Stinchcombe, Arthur. 1968. Constructing Social Theories. New York: Harcourt Brace, Jovanovich. Stone, Judith, 1985. Search for Social Peace: Reform Legislation in France 1890-1914. Albany: State University of New York Press. Streeck, Wolfgang. 1984. "Neo-Corporatist Industrial Relations and the Econom-

References

237

ic Crisis in West Germany." In John Goldthorpe (ed.). Order and Conflict in Contemporary Capitalism: Studies in the Political Economy of Western European Nations, pp. 291-314. Oxford: Clarendon. 1966. White Collar Trade Unions: Contemporary Developments in Industrialized Societies. Urbana: University of Illinois Press. 1972. Comparative Labor Movements: Ideological Roots and Institutional Development. Belmont, Calif.: Wadsworth. Thaler, Richard, and Sherwin Rosen. 1976. "Value on Saving a Life: Evidence from the Labor Market." In Nestor Terleckyj (ed.). Household Production and Consumption, pp. 265-297. National Bureau of Economic Research. New York: Columbia University Press. Tilly, Charles. 1978. From Mobilization to Revolution. Reading, Mass.: AddisonWesley. Tracy, Joseph. 1986. "An Investigation into the Determinants of U.S. Strike Activity." American Economic Review 76:423-436. Trempe, Rolande. 1968. "Le Reformisme des Mineurs Frangais a la Fin du XIXe Siecle." Mouvement Social 65:93-107. 1971. Les Mineurs de Carmaux. Paris: Editions Ouvrieres. 1984. "Struggles of French Miners for the Creation of Retirement Funds in the Nineteenth Century." In Anne-Marie Guillemard (ed.). Old Age and the Welfare State, pp. 101-113. New York: Russell Sage. Trotsky, Leon. 1969. Leon Trotsky on the Trade Unions. London: Pathfinder. United Kingdom. Royal Commission on Trade Unions and Employers Associations. 1968. Report of the Royal Commission on Trade Unions and Employers Associations. Command Paper 3623. London: Her Majesty's Stationery Office. Valois, Georges. 1927. L'Etat Syndical et la Representation Corporative. Paris: Valois. Viaud, M. 1929. "La Rationalisation dans 1'Exploitation des Mines." Revue de ITndustrie Minerale 22:367-382. Villain, Frangois. 1918. "La Prevoyance Sociale dans les Mines et la M^tallurgie." Revue de ITndustrie Minerale 11:85-108. Viscusi, W. Kip. 1979. Employment Hazards: Investigation of Market Performance. Cambridge: Harvard University Press. Viscusi, W. Kip, and Charles O'Connor. 1987. "Hazard Warnings for Workplace Risks: Effects on Risk Perceptions, Wage Rates and Turnover." In W. Kip Viscusi and Wesley Magats (eds.). Learning about Risk: Consumer and Worker Responses to Hazard Information, pp. 98-124. Cambridge: Harvard University Press. Wallace, Michael, and Arne Kalleberg. 1981. "Economic Organization of Firms and Labor Market Consequences: Toward a Specification of Dual Economy Theory." In Ivar Berg (ed.). Sociological Perspectives on Labor Markets, pp. 77118. New York: Academic Press. Walsh, Kenneth. 1983. Strikes in the United States and Europe: Measurement and Incidence. New York: St. Martin's. Walton, Robert, and Robert McKersie. 1965. Behavioral Theory of Labor Negotiations. New York: McGraw-Hill. Webb, Sydney, and Beatrice Webb. 1897. Industrial Democracy. London: Kelley.

238

References

Whyte, William Foote. 1955. Street Corner Society: Social Structure of an Italian Slum. Chicago: University Chicago Press. Wright, Erik Olin. 1979. Class Structure and Income Determination. New York: Academic Press. Zevaes, Alexandre. 1947. Histoire du Socialisme et du Communisme en France de 1871 a 1947. Paris; France-Empire. Zimbalist, Andrew. 1979. Case Studies in the Labor Process. New York: Monthly Review Press.

Index

Abadie, M., 98 Accident rate, 112. See also Fatality rate in French coal mining Accommodation, 137-162, 164, 165, 170172, 173-175, 186, 187-189, 203, 215 defined, 137 see also Militancy, benefits of Aircraft construction, low vulnerability to strikes, 109 Albi, 118, 142 AUier, 22-26, 73, 176-177, 193 General Strike of 1902, 151, 153 Alpes-Hautes,. 22-26 Alsace-Lorraine, 15. See also Moselle Althauser, Robert, 29 Amdur, Kathryn, x, 12, 156, 214 American unionism. See Contemporary American unionism Anarchists (direct action syndicalists), 141, 144, 156, 166-175, 187 opposition to state industrial conciliation procedures, 130, 166, 173, 174, 174n regional base, 156, 166, 168 Annales des Mines, 94, 148 Annuaire des Syndicats, 48-49, 165 Antistatists. See Anarchists Arbitration, 142-143, 152-154, 159, 159n, 167 Ardeche, 22-26 Aronowitz, Stanley, 69 Arnaud, Pierre, 159, 160-162 Arras, Convention d'. See Convention of Arras Ashenfelter, Orley, 34

Associational life. See French coal miners, associational life Asymmetry of managerial and union bargaining logic, 32-34 Audibert, Paul, 73 Autocorrelation in wages, 51-52 Averitt, Robert, 29 Aveyron. See Decazeville Avion, 182 Axelrod, Robert, 34, 38

Barbier, Marcel, 73 Bargaining, vi, 66 Bartuel, Casimir, 176 Basly, Emile, 124, 126, 139, 140, 148, 166 Beaubernard, R,, 13, 14, 124, 143-146 Becker, Brian, 31 Benoit, 173 Bethune, Cie. de, 108, 116, 138 Bibb, Robert, 3 Blackburn, Robert, 31 Blanzy, Cie. de. See Montceau-les-Mines Blau, Peter, 3 Block, Fred, 120, 124 Boncour, Paul, 123 Bonnefous, Georges, 123, 128 Bouchard, H., 173 Boulangerism, 127, 129-130, 139 Boulin, Pierre, 16 Bourgeaud, M., 19 Bourgeois reform. See Reformism; Solidarisme; Welfare state Boyer, Robert, 207 239

240

Index

Boys' employment and conditions, 170-172, 193 Bracketing, 176n, 182 Braverman, Harry, 210 Brecher, Jeremy, 1, 27 Britain, Tensions over coal imports, 123 British Unionism, 9, 35, 92, 114 Brogan, D. W., 15, 123, 129, 130 Browne, Margaret, 46 Bruay, 119, 138 Buder, Stanley, 14 Buffering from wage competition, 209-212 from illiquidity of capital, 209, 210 from illiquidity of knowledge, 209-210 from illiquidity of labor, 209n reduction for French coal miners, 210-212 Burawoy, Michael, 3, 92, 208 Business unionism, 6. See also Economism; Reformism; Wisconsin school Buxiere-les-Mines. See Allier

Charter of Amiens, 166 Chasse des renards. See Strikes; Violence Chomage. See Labor scarcity; Unemployment Christian unionism, 31, 145, 164-166, 179180 limited impact in coal mining, 165, 165n Centralization of unions, dangers of, 219221 reduces beneficial factionahsm in bargaining, 220-221 reduces militancy, 219-220 reduces probability of losing, 220 CIO sitdown strikes, 223 Clark, Marjorie, l l n , 130, 177 Clegg, H. A., 2, 35 Clemenceau, Georges, 128, 143 Cloward, Richard, 1, 27, 88, 213, 220, 223 Coal importance to national economy, 114-

Calvignac, 142 Cameron, David, 4 Campbell, Peter, 127, 128, 130 Cantal, 19, 22-26, 155, 176-177 Capital intensity, 89n Carmaux, 18, 19-20, 22-26, 74, 93n, 110, 117, 124, 129, 141-143, 144, 151, 154, 187-188, 203 1892 strikes, 141-143, 187 1902 General Strike, 151, 154 1923 General Strikes, 110 postwar managerial offensives, 187-188 Caron, Frangois, 15 Cartel des gauches, 127-128 Case selection, 6-10, 13, 15, 43-44 Catastrophes as source of distortion in compensating

116, 122-123, 211 miUtary asset, 123 perishability, 115 Coal market, French, 12-13, 15-17, 109110, 111, 114, 115-117, 122, 123, 132133, 139, 144, 145, 146, 147-148, 161, 167-168, 211 capacity for shortfalls to be met by imports, 113, 114-115, 116, 123, 211 regional differences, 17-19, 22-25 see also Prices, coal Coal miners, favorable status internationally, 15 Cochran, Bert, 2, 31, 38 Coleman, James, 4 Collective bargaining, tripartite, 13, 14,

differentials tests, 98 as stimulus to strikes, 80 see also Courrieres disaster Center-left political alliances, 126-127. See also State support for coal miners Centrist parties. See Center-left political alliances; Deputies, Chamber of, party composition of Chabert, Andre, 7 Chaloner, William, 14 Chamber of deputies. See Deputies, Chamber of

129-130 Combeplaine, 159-160 Comfort/courtesy disputes, 71, 76-77, 78, 100-101 Comite Central des Houilleres Frangaises, 17, 95, 108, 111, 114, 116, 120, 121-122, 123, 124-125, 134, 149, 156, 161, 162, 178, 189, 190 de facto management bargaining agent, 121-122, 125 origins, 124-125 state participation as member, 121-122 Commons, John, 2, 92, 192, 214

Index Communism. See Communists, U.S.; Marxism; Unitaires Communists, U.S., 38, 168, 222 Company control of mining towns, 13-14, 74, 124, 139, 141-145. See also Paternalism Compensating differentials, 87-88, 89-91 emphasis on fatalities as basis for compensation, 96-97 inconsistencies between arguments concerning older workers and workers in general, 91 inconsistency with neoclassical theory of wages, 90 inverse relation between wages and mining fatalities after 1923, 98 lack of evidence for, 89 lack of correlation wages and mining fatalities, 96-99 positive correlation wages and length of working day, 94-95 see also Dangerous work; Trade-offs between monetary and nonmonetary benefits Compromise. See Success and compromise combined Conell, Carol, x Confederation Francaise des Travailleurs Chretiens, 165 Confederation Generale des Travailleurs (CGT), 18, 31, 166, 168, 177 CGT of 32 (faction), 182 see also Confederes Confederation Generale des Travailleurs Unitaire (CGTU). See Unitaires Confederes, 6, 11 Effect of confedere strikes. See Steering measurement of affiliation to, 59-60 quietness of strikes, 119 regional base, 18-19, 26, 168, 168n strike averseness, 175 toleration of strikes under Arnaud, 161 see also Cooptation of union leadership Contemporary American unionism, 4, 9, 33, 35, 72, 92, 103, 136, 214-219, 222, 223-226 class power of workers in government and services, 225 air traffic controllers, 225 nurses, 225

241 Contempoary American unionism {cont.) class power of workers in government and services {cont.) police and fire fighters, 225 teachers, 225 economic challenges to, 223-224 excessive length of strikes, 217-219 factors inhibiting strike-proneness, 217 methodological problems in studying, ix need to strike before deindustrialization takes hold, 224-226 vulnerability to one-shot decisive defeats, 217 willingness of workers to tolerate political demands, 215-216 Contract length, 154-155 Convention of Arras, 14, 138-140, 141 Cooptation of union leadership, 12n, 220 Corporatism in French coal mines, 14-15, 207-208. See also Collective bargaining, tripartite; State participation in industrial relations Corporatist class compromise, ix, 29, 208n. See also Social democratic theory Cost of demands. See Demands, cost of Cost theory, 68, 71-75, 76, 77-79 Cotte, Jean, 167 Courri^res disaster, 98, 149 Coverman, Shelley, 3, 27, 65 Cransac, 118 Creuse, 19, 22-26, 133 Cronin, James, 3, 4, 10 Cross, Randy 13, 133, 192, 193n, 215 Cross-national studies, viability of, 8 Cross-tabulation, appropriate use of, 60 Crouch, Colin, 223 Cousineau, Jean-Michel, 3, 27, 30n, 31 Dangerous work eliminatable through market competition, 90 lowers productivity and, 90 raised legal costs and, 90-1 see also Compensating differentials; Fatality rates in coal mining; Safety concerns of workers Data attrition, 60, 192 Data availability to contemporaries, 207, 207n Data sources, 7-9, 43, 46, 47-49, 67, 80, 95, 97, 99, 109n, 116, 127-128, 138n, 141n, 143n, 146n, 192

242 Days lost in strikes. See Strikes, Mandays lost Days worked as a denominator in rates, 45, 133n Decazeville, 17, 18, 22-26, 93n, 110, 117, 124, 176, 178, 180 1923 General Strike, 110 Deindustrialization, 224 elimination of some workers from union activity by, 224 need to strike before deindustrialization takes hold, 224-226 De la Taille, M., 94, 147 Demand. See Coal market, French; Labor scarcity Demands, 65-104 ambitious demands cause ambitious settlements, 66 censoring of unwinnable, 102 cost of, 30-31, 31n data on, 50, 67, 99 see also Economism; Pay strikes Demographic segmentation, 191-195 Departments, selection of. See Case selection Deputies, Chamber of, 126-129, 134, 139, 142, 145, 149, 167, 188 party composition of, 127-129 Design. See Research design Direct action syndicalists, 166-175. See also Anarchists Direct representation of capitalist interests. See Instrumentalism Documentary evidence, 108 DoUeans, Edouard, 11, 165 Domhoff, William, 124 Donovan report. See United Kingdom Royal Commission on Trade Unions and Employers Associations Doumer, Paul, 128n Dourges, Cie. de, 139 Dreyfus case, 144 Dual sector theory, 29 Dual unionism, 2, 31, 221-223 bad for radical workers, 222-223 effect on strike effectiveness, 28, 34-40, 54-58 effect negated when both unions act identically, 34-36 moderates outperform radicals when present, 38, 59-61

Index Dual unionism {cont.) probability of being associated with ideological differentiation among unions, 36 see also Single union representation, presence of; Steering; Whipsawing Dulot, Charles, l l n Duncan, Greg, 88 Duncan, Otis Dudley, 3 Dunlop, John, 51 Durafour, Antoine, 159, 159n Duration. See Strike, duration Dynamic measures of demand moderation, 68, 69-70, 80-83 Eastern Airlines, 33 Easy grievances, 76-77 Eaton, Adrienne, x Eckstein, Susan, 3 Economic planning, 122-123, 123n, 186-187 Economism, vi, 2, 19, 88-89, 92, 103, 214216. See also Trade-offs, monetary and nonmonetary benefits; Union choice; Wages, importance of; Wisconsin school Economy, French. See French economy Edelman, Murray. See Edelman concession Edelman concession, 150-151, 152, 164, 187, 189-190 frequency in downturns, 190 Edwards, P. K., 3, 4, 34, 65 Edwards, Richard, 2, 192, 208. Effects of strikes vs. causes of strikes. See Strikes, Effects of vs. causes of Eight-hour day, 74, 94-95, 126, 144, 146, 147, 148, 159, 172-175, 178 Law of 1905, 172, 173 see also Length of workday; Overtime; Work hours disputes Electoral strategy, 119, 128, 128n, 130 Electrical power, 17 Elwitt, Sanford, 129 Employer indifference to strikes. See Managerial indifference to strikes Employers' associations 138-139. See also Comite Central Hes Houilleres Frangaises Employer interest in low wages. See Managerial interest in low wages Employer noncooperation with industrial conciliation. See Managerial noncooperation with industrial conciliation

Index Employer opposition to republican democracy. See Managerial opposition to republican democracy Employer offensives, forms of, 33 Employers, national differences in preoccupation with job control, 69n, 108 Employment, 16, 24, 45, 52-59, 80-81, 84, 96, 97, 103, 112, 115-116, 132-135, 133n, 145, 206, 207. See also Coal market, French English unionism. See British unionism Error components models, 51, 52n Escalier, 173 Ethnic exclusion, 71, 74n, 100-101 Ethnic divisions. See Foreign miners Exchange rates excluded from analysis, 4647 External validity. See Generalizability

Family allocations, 131-132, 178, 191 Farber, Henry, 27, 68 Farley, Reynolds, 192 Fatality rates in French coal mining, 75, 8081, 149 effect on wages, 96-99 measurement of, 97, 97n period effects in, 98 see also Accident rate; Courrieres disaster Faure, Petrus, 49, 124 Federation des Travailleurs du Sous-sol, 11, 49, 138-155 independence of CGT, prewar, 166 recognition of, 14-15 strike averseness among prewar leadership, 140, 148, 151, 151n support for state industrial conciliation procedures, 130, 139, 140, 149, 152-153 See also Confederes Federation des Travailleurs du Sous-sol Unitaire, 49. See also Unitaires Fine, Martin, 12 Fire supervisor disputes, 71, 74, 74n, 76-77, 78-79, 78n, 100-101, 142, 157-159. see also Weak managerial control Fire worker disputes, 71, 74, 74n, 76-77, 78-79, 100-101. See also Ethnic exclusion Firminy, 93n, 133, 155-156, 157, 157n, 168172 Foner, Phihp, 2, 31, 168

243 Foreign miners, 13, 44n, 110, 133, 157-159, 176, 182, 191-195 effect on wages, 192-193 extent of use of, 192 home government protection of, 193-194 indicator of firm prosperity, 192-193 lower mihtancy of working class, 191-192 support from local co-nationals, 194 suffer wage discrimination, 192 unionization of, 157-159, 194 Form, William, 3, 192. Fractional strikes, 35-36 France, Administration des Mines, 17 Free riders, v. See also Olson Mancur, Prisoner's Dilemma Freeman, Richard, 1, 40, 53, 132n French coal market. See Coal market, French French coal mines geological limits on, 211 effect of World War I on. See World War I nationalization of. See Nationalization of French coal mines French coal miners, 6-10 associational life, 13 buffering from market pressure, 207-208, 210-212 decline of, 211-212 historical overview, 6, 10-20, 210-212 militancy, relative, 10 past studies of their wages, 19-20 power of, 12-14, 40, 109-113, 207-208, 210-212 capacity to cut off coal supplies, 110111 similarity to midcentury Swedish workers, 207-208 see also State regulation of coal mining; Strikes, capacity to harm employers; Wage levels in French coal mining French economy, 15. See also Great Depression French Mine Service, 109, 109n, 110-111, 114, 123, 131, 144, 158, 159-160 French union data, low quality of, 47 French unionism, 4, 14-15, 31, 47, 72, 72n, 169 Friedland, Roger, 4 Friedman, Gerry, x, 13, 47, 65, 129 Fringe benefits, 44, 93, 93n, 130-132, 132n, 134-135, 156

244 Fringe benefits {cont.) correlation of wages with, 93n, 132n see also Family allocations; Paternalism; Political concessions; Unemployment, paid employer relief; Pension/sick pay disputes; Welfare state benefits for miners Front unique. See Joint strikes; Unitaires, attempts to form front unique Gaillard, Michel, 140n Galand, Pierre, 4, 16, 134n Gallie, Duncan, 47, 55, 129, 132, 177, 178, 214n Card, 22-26, 178, 180-181, 193 Strike of 1890, 140-141 General Strikes of 1923, 176, 179 Strike of 1928-1929, 118, 181 Garment industry, vulnerability to strikes, 109 Garmy, Rene, 2, l l n , 31, 130, 165 Gender. See Women's employment and conditions General English Sympathy Strike of 1926, 124n General prices. See National prices General Railway Strike of 1920, 55, 177178, 215 General strikes disadvantages, 113 prewar as a whole, 146, 176 General Strike of 1902, 117-118, 146-155, 167, 203 General Strike of 1906, 146, 149, 168-172, 203 General Strike of 1919, 176, 177 General Strike of 1921, 176, 178 General Strikes of 1923, 109-112, 121, 124n, 157, 157n, 176, 179-180 General Strike of 1926, 160, 215 General Strike of 1928-1929, 160 General Strike of 1931, 160 General Strike of 1968, 223 Generalizability, vi-vii, x, 6, 10 Georges, M., 16, 116, 132 German unionism, 9, 35n, 114 Gignoux, Claude, 13, 47 Gillet, Marcel, 11, 14, 18, 124, 138 Gordon, David, 2, 192 Gramm, Cynthia, 201

Index Gras, L. ]., 18 Great Depression, 17, 132-135, 161, 194n Greene, WiUiam, 52 Grievance, occasional superiority to strike as unit of analysis, 70 Grievance data, 50, 67 Griffin, Larry, 3, 54 Gruner, Edouard, 16 Guttman scale, 48

Hainsworth, Raymond, 14, 73, 183 Hamermesh, Daniel, 88 Hanagan, Michael, 156, 168 Hard grievances, 76-77 Hardy-Hemery, Odette, 16, 18, 73, 111, 122, 134n, 183, 194, 215 Hartman, Paul, 6, 32, 217 Hatzfeld, Henri, 131 Hauser, Robert, 3 Haute-Cappe, 168 Hayward, Mark, 3, 27, 65 Hayes, Beth, 31n, 201 Heat disputes, 70, 71, 74-75, 76-77, 78 Hechter, Michael, 4 H^rault, 22-26 Herding, Richard, 69, 88 Hicks, John, v, 1, 27, 28, 29-34, 40, 45, 51, 53, 66, 68, 145, 201 Hicksian models. See Neoclassical theories of strikes Hill, Steven, 69 Historicism in unions, 38 Holmlund, Bertil, 88 Hormel, 33, 217 Hours disputes. See Eight-hour day. Work hours disputes Humanity, 157n, 158, 176 Human capital, 74 Human relations theory, 68-69 Ideological reversals, cost of, 38 Imperfect information as cause of strikes, 29-30 Implicit contract theory, 145 Income equalization disputes, 70-71, 72, 76-77, 78, 100-101 Industry level national bargaining (contemporary), 171

Index Inflation (but not prices) excluded from analysis, 46-47 Informational striking, 30. See also Neoclassical strike theory Institutionalization of labor relations, 186191. See also State regulation of industrial relations Intermediate grievances (in difficulty), 7677 International Workers of the World (IWW), 31 Isaacs, Larry, 3, 54 Isere, 22-26, 73 Instrumentalism, 107, 124-126 Irvine, William, 127, 130 Italian miners. See Foreign miners Jaunes. See Christian unionism Jaures, Jean, 124, 128 Jencks, Christopher, 3 Job control disputes. See Work assignment/rules disputes Johnson, George, 34 Joint coordination during prewar schisms, 168-169 Joint strikes, 119, 179-180 common decision-making during, llOn, 179-180 effects of, 34-35, 39, 61-64 Juvenile employment. See Boys' employment and conditions Kalleberg, Ame, 29, 96 Katz, Harry, 33 Kelley, Jonathan, 3 Kelley, William, 3 Kerr, Clark, 10, 13 Klein, Herbert, 3 Kmenta, Jan, 51n Knowles, Kenneth Guy Jack Charles, 3, 217 Kochan, Thomas, 33 Kornhauser, William, 4 Korpi, Walter, 4, 11, 168, 220 Kourchid, Olivier, x Kriegel, Annie, 2, l l n , 177 Kuisel, Richard, 122, 123n Labi, Maurice, l l n Labor aristocracy, 168 Labor exchanges, 133

245 Labor process concerns. See Work assignment/rules disputes Labor scarcity, 13, 15-17, 110, 112, 116, 132-135, 145. See also Employment; Unemployment Lacroix, Robert., 3, 27, 30n, 31 Lafitte-Laplace, Jean, 12, 114, 116, 123, 130, 131, 132 Laligant, Georges, 73 Lamendin, A., 124, 126 Langrogne, E , 73 La Peronniere, 168 Large sample analysis, ix Late pay/pay timing disputes, 70, 71-72, 72n, 75-77, 78 Latent bargaining strength, grievances with, 75-76, 77, 78 Layoffs, disputes over. See Punishment/layoff disputes Lazear, Edward, 208 LeFranc, Georges, 4, l l n , 14, 31, 128, 130, 166 Left wing parties, strength of. See Deputies, Chamber of, party composition of Legislative process, French, slowness of, 148 Length of workday, 94-96, 111-112 inversely correlated with wages, 95-96 see also Eight-hour day; Overtime; Work hours disputes Lenin, V. I., 4, 31, 88, 168, 214 Lens, 17, 119, 138-139 Lens, Sydney, 1, 12n Lester, Richard, 12n, 32 Levard, Georges, 31 Leveque, M., 18 Lewis, H. Gregg, 1, 40 Liability costs, 90-91, 131, 156 Ligny-Ies-Aire, 138n Limoges, 214 Logistic regression, 81 Loire, 18, 19, 22-26, 93n, 108, 116-117, 118, 133, 155-162, 178, 188, 191, 191n aftermath of 1893 strike, 48-49 Arnaud purge and schism, 160-162 coal supply survey of 1908, 116-117 Combeplaine strike of 1926, 159-160 correlation of strikes in wages postschism, 155 departmental strike of 1900, 116, 167, 172

246 Loire {cont.) eight-hour day disputes in, 172-175 General Strike of 1902, 116, 117-118, 148149, 151, 152-153, 167-168, 172 General Strike of 1906, 116, 168-172 salary settlement produced, 170-172 absolute increase same for strikers and non-strikers, 171 idiosyncratic categories, 171n nonstriking firms paid more, 171 relative increase higher for strikers, 171 General Strikes of 1923, 157, 157n, 176 General Strike of 1926, 160 General Strike of 1928-1929, 160 General Strike of 1931, 160-161 railway nationalization strike of 1920, 177 regional distribution of Confedere vs. Unitaire affiliation, 155-156 regional distribution of employment in, 156n Schism of 1903-1911, 166-175 cooperation between factions during, 166, 168-169 weakness of antistatists, 169-170 ViUeboeuf strike of 1925, 157-159 See afeo Tirminy; Haute-Cappe; La Peronniere; Ricamarie; Roche-la-Moliere; Saint-Etienne; Villeboeuf Loire-Haute, 22-26, 193 General Strike of 1902, 151, 153, 154 Loire, Mines de la (corporation), 170-172 Lordstown, 92 Lorwin, Val, 2, 31, 47, 69n, 72n, 141 Losing, effects of. See Winning, effects of Loze, Edouard, 139

Machlup, Fritz, 1, 27 Maier, Charles, 177 Maintenance of mines during strikes, 118 Majoritaires, 176, 177 Management, incapacity to truly force strikes, 33-34 Managerial indifference to strikes, 106, 107113, 135-136 documentary evidence for, 107-108 consistency with anti-unionism, 108 Managerial interest in low wages, 106, 119120, 135-136

Index Managerial noncooperation with industrial conciliation, 130, 139-140, 147, 149154, 164, 167, 178, 187-189 consequences of higher wages, 147, 149154, 164, 188 Managerial opposition to republican democracy, 129-130, 139, 140, 143-145 Mainstream unionism, radicals more effective within, 223 Mandays lost. See Strikes, mandays lost Mangez, Maurice, 44, 130, 149, 154 Mann, Michael, 3 Marks, Gary, 15 Marsden, David, 51 Martin-Saint-Lton, Etienne, l l n Marx, Kari, 27n, 88, 214 Marxism, 3, 4, 27, 31, 69, 124, 214 consistency with worker consciousnessness, 214-215, 214n See also Marxist functionalism; Union choice; Theory of the State; Unitaires Marxist functionalism, 107 Materialist theories of conflict. See Rationality in conflict Mayeur, Jean-Marie and Madeleine Reberioux, 15, 126-127, 130 Mazgaj, Paul, 127 McKersie, Robert, 12n, 31n, 33, 66, 220 Medoff, James, 1, 40, 53, 132n Merchant marine. Relations with coal companies, 115n, 120, 122, 133-134 Michel, Joel, 11, 18, 126, 130 Michels, Robert, 12n, 32, 220 MiHband, Ralph, 124 Militancy, benefits of, 1, 27, 28, 32, 34, 40, 54-59, 62-63, 64, 66, 85, 97, 102-103, 133, 137-162, 170-175, 198-189, 204207, 205n, 212, 213-221 functional substitute for long contracts in recessions, 147, 154-155 Militancy, causes of, 168. See also Strikes, causes of Militancy, limits of, 197-212 Millerand, Alexandre, 128, 143, 178 Miner deputies, 13, 124, 126, 131, 139 Miner safety delegates, 158 Minimum wage, 147, 152n Ministry of Foriegn Affairs, 121, 124n Ministry of Interior, 147, 149, 153 Ministry of Justice, 188

Index Ministry of Labor, 108, 121, 160, 173^174, 188 Ministry of Public Works, 121, 139, 144, 145 Minoritaires, 176, 177 Mishel, Laurence, 1 Mixed representation districts, 59-60, 64. See also Steering, benefits workers in mixed representation districts Moderation of demands, measurement of, 67-79 Monarchism, 129-130 Montceau-les-Mines, 18, 19, 22-26, 74, 93n, 124, 143-146, 165-166, 165n, 203 Montrambert et la B^raudiere, Cie. de, 108, 118, 170-172 Moore, Henry Ludwell, 19 Moroccan miners. See Foriegn miners Mortality in French mines. See Fatality rate in French coal mines Moselle, 22, 43, 44n, 178 General Strike of 1923, 109-112, 179-180 see also General Strikes of 1923 Moutet, Amy, 73, 111, 134n, 183 Multiple short strikes, advantages of, 219 Multipliers for variables, 53, 57, 58, 81, 96 Multi-regional strikes, processing of, 50 Mussolini, Benito, 193, 194

National coal wages. See Wage levels in French coal mining Nationalization of French coal mines, 14, 211 Neoclassical theories of strikes, v, 1, 27, 2934, 40, 53, 66, 201 how altered if strikes themselves have utility, 32 underlying assumption of symmetric bargaining logic, 32-33 Neoclassical theories of wages, 19, 45, 64, 204-212 actual use by contemporaries, 207, 207n applicability to aggregates, 208-209 opposed by sociologists, 208 poor at explaining wages of individual workers, 208 prime determinant of wages, 197-198, 204-212 high explained variance of market related predictors, 204-207

247

Neoclassical theories of wages {cont.) prime determinant of wages {cont.) ratios of conflict to neoclassical predictors, 205-207 see also Buffering from wage competition Neumann, George, 1, 27, 31, 32, 33 Nievre, 19, 22-26 Nonagricultural employment, 46, 52-59, 84, 97, 103, 206 Nonmonetary benefits, 93, See also Heat disputes; Paternalism; Pay strikes; Trade-offs between monetary and nonmonetary benefits Nonsubstitutability of different grades of coal, 114-115 Nord, 16, 18, 93n, 118, 119, 124 , 133, 138140, 150-152, 154-155, 168 168n, 177, 190, 193 General Strike of 1902, 150-152 Strike of 1893 and aftermath, 48-49, 118 see also Pas-de-Calais Northern basin. See Nord; Pas-de-Calais; Regional differences Oberschall, Anthony, 2, 108 O'Connor, Charles, 88 O'Connor, James, 3n, 107 Offe, Claus, 107 Office du Travail Survey of Length of Working Day 1892, 95-96 Okun, Arthur, 145 Olivier, M., 12, 17, 115, 116, 122, 123 Olson, Craig, 31, 88, 91 Olson, Mancur, v, 34. See also Prisoner's dilemma Ordinary least squares regression, 83, 96, 201-203, 204-207 appropriate use of, 51-52, 96, 202 goodness of fit measures, 204-205 non-additivity, 205n inefficiency, 204 see also Parks cross-section time-series regressions Ostricourt, 182 Overtime, 111-112. See also Eight-hour day; Length of workday Paganon,134 Paige, Jeffery, 4 Pariiamentary left, 11, 13, 126-129. See also Miner deputies

248 Parks cross-section time-series regression, 43, 51-52, 60, 84, 97, 102-103, 192 calculating meaningful interpretations of coefficient sizes, 198-199 constant versus region-specific rhos, 52, 204 goodness of fit in, 52 Parks, R. W., 51n Parsley, Clifford, 1, 19, 53 Part-time producers, data quality for, 44n Parties, political, French. See Political parties, French Pas-de-Calais, 16, 17, 18, 93n, 110, 111, 118, 119, 120, 124, 126, 133, 154-155, 168, 168n, 177, 181-186, 188, 190, 193 agitation for Convention of Arras, 138140 1893 strike and aftermath, 48-49, 118 1902 General Strike, 150-152 1923 General Strikes, 110, 176 prefectural views on industrial conflict in, 119, 181-186 see also Nord; Courrieres disaster PATCO strike, 217 Paternalism, 93, 93n, 141-145, 156. See also Company control of mining towns Pattern bargaining, 121, 190-191 Pay strikes as percentage of all strikes, 87, 87n effect of, 50, 53-54, 56-57, 87-104 less effective than non-pay strikes in raising wages, 93, 102-103, 214 measurement of, 50, 87 see also Economism; Wages, Importance of; Tradeoffs, Monetary and nonmonetary benefits Pay timing. See Late pay/pay timing Pension defaults, 131 Pension/sick benefit disputes, 70, 74, 75, 76-77, 78, 100-101, 126, 131, 146, 147148, 149, 178, 188 Percent nonagricultural. See Nonagricultural employment Period, choice of, 14, 43 Period effects, 54-59, 64, 84, 97, 198-199, 202-203, 205-207 constancy of wage attainment across periods, 57-59, 164 none concerning strike success, 83 Perlman, Selig, 2, 4, 92, 168, 214, 215-216

Index Perrot, Michele, 13, 117 Petite-Rosselle, Cie. de, 113 Peyerimhoff, Henri, 13, 17, 44 Phelps-Browne, E. H., 46 Pigeot, 108, 161 Piven, Frances, 1, 27, 88, 213, 220, 223 Pizzorno, Alessandro, 223 Podgursky, Michael, 1 Poincare, Henri, 122-123 Police reporters, 9, 140, 140n Policy recommendations, 213-226 Polish miners. See Foreign miners Political concessions, 11, 15, 94, 107, 124, 126, 129, 130-135, 146-148, 178, 191 tenuous linkage between issues and timing of political strikes, 146, 148-149 see also Eight-hour day; Pensions/sick pay disputes; Welfare state benefits for miners Political parties, French, inconsistent ideologies of, 127-128, 128n Political rights of workers, 74. See also Company control of mining towns PoUins, Harold, 115 Pooled cross-section time-series methods. See Parks cross-section time-series regression Popular Front, 14, 73, 182, 187 Postwar energy crisis, 16-17 Poulantzas, Nicos, 124 Prechel, Harland, 125 Prefects, 116, 119, 132, 133, 134, 139, 140141, 142, 149, 152-153, 155, 156, 158159, 173, 180-186, 187, 188, 189 Prices, coal, 45-46, 52-59, 84, 97, 103, 145, 206, 207. See also Coal market, French Prices, national, 46, 52-59, 84, 97, 103, 145n, 160n, 206 Prime (multiplier to base salaries), 148, 150154, 150n Prisoners' dilemma, 34, 37-38, 106 Productivity, 25, 45-46, 84, 97, 103, 111112, 134n, 206, 208 not lowered by past wage increases, 19 raises wages, 19, 52-59 role of geology in, 72n Prost, Antoine, 47 Przeworski, Adam, 4, 29 Public goods, theory of. See Prisoners' dilemma

Index Punishment/layoff disputes, 71, 73-75, 7 6 17, 78, 100-101, 142-143, 144-145, 157-159, 189 Puy-de-D6me, 22-26 General Strike of 1902, 150, 151, 153, 155 Radical party (centrist French political party), 126-128, 128n Ragin, Charles, 3, 27, 65 Railroads, relations with coal companies, 115, 115n, 121, 122, 133-134 Railway nationalization strike of 1920. See General railway strike of 1920 Rapaport, Anatole, 37 Rationality in conflict, 4-5 Reberioux, Madeleine, 15, 126-127, 130 Reder, Melvin, 1, 27, 31, 32, 33 Reformism. See Confeder^s, Solidarisme Regional contracts, 14, 138-141, 169-172 Regional differences. See Coal market, France, regional differences; Confederes, regional base; Unitaires, regional base Regulation economics, 207 Reich, Michael, 2, 192 Reid, Donald, 13, 14, 17, 73, 111, 117, 124, 192, 194, 211, 215 Remoteness of coal mines, 13, 156, 178-179 Repression, 13, 117-119, 133, 140, 177-178, 179-180, 214n capacity for altering the state intervention-high wage correlation, 136 moderation in use of coercive measures, 117 routine use of troops in strikes, 117 see also State support for coal miners Research design, ix, 6, 14, 15, 43-52, 76, 79-81, 83-84, 93-94, 99-100, 102. See also Parks cross-section time-series regression; Sample specification Residual wages, 5, 7-8 Results of strikes. See Strikes, determinants of winning; Winning, effects of Revolution effect of, 3 prospect for in France, 12, 129-130, 176177 Ricamarie, 118 Right wing political support for unions, 126-127

249 Robert, Jean-Louis, l l n Roche-la-Moliere, Cie de. See Firminy Roche-la-Moliere (town), 155-156, 168 Ronchamp, 187, 188-189 Ronge, Volker, 107 Roomkin, Myron, 32, 220 Rosen, Sherwin, 1, 88 Ross, Arthur, 2, 6, 32, 35, 217 Rubin, Beth, 3, 27 Ruhr, 110, 123, 124n Saar, 110, 121, 122, 124n, 190-191 French control of dominial mines, 121122, 191 Sarre-et-Moselle, Cie. de. 111, 113 Safety concerns of workers, 88, 149, 159160 lack of same, 98-99, 149 Safety disputes, 71, 76-77, 78-79, 100-101, 149, 159-160 Saint-Etienne, 18, 73, 93n, 118, 156, 157, 167, 168-172, 173-175, 214 Saint-Etienne, Houilleres de (corporation), 170-172, 173-175 Saint-Hilaire, Cie de, 153-154 Salary defenses, 70, 71, 74-75, 76-77, 78, 148-149, 161 Salary offensives, 70, 71, 74-75, 76-77, 78, 138-142, 142-143, 144-146, 149, 157, 160 Sample specification. See Case selection, research design Sanders, Jimy, 4 Saone et Loire. See Montceau-les-Mines Saone-Haute, 22-26, 187, 188-189 Saposs, David, l l n , 31, 47, 72, 95, 130, 131, 133, 141, 166, 178, 191 Sauvy, Alfred, 13, 47 Saturn workers, 225-226 Savoie, 22-26 Schism in French unionism, 6, 11 greater severity of 1923 schism, 164, 175176 prewar, 164-175 versus war as basis for periodizing differential strike effects, 54-57, 64 Schneider, Eugene, 12, 114 Scientific management. See Taylorism Secondary sector, 29, 224-226

250 Service workers, 225 Sewell, William, 3 Shadow unionization, 48-49 Shopfloor disputes. See Work assignment/rules disputes Shorter, Edward, vi, 2, 3, 4, 6, 9, 10, 34, 49, 72, 129, 141, 168 Siegel, Abraham, 10, 13 Siegler, Jean, 18 Significance tests, appropriateness of, 100 Simard, Marc, 18 Simiand, Francois, 19 Single industry studies, advantages of, 8 Single uruon representation, presence of, 9 Skocpol, Theda, 4, 124 Smelser, Neil, 4 Smith, Robert, 88 Snyder, David, 3 Social agendas. See Economism Social democrats. See Confed^r^s Social democratic theory, 3-4 Socialist Party, 124, 127-129, 142, 144. See also Parliamentary left Solidarisme, 126-127, 129-130 Sources, data. See Data sources Spaeth, Joe, 3 State autonomy, 107, 124-126, 129-130 diminished autonomy during economic crises, 125 State dependence on coal, 12-13, 15, 114, 115-117, 120-124, 178, 189 State investment in coal mines, 122 State interest in high wages, 105-106, 126136, 186 State interest in low wages, 105-106, 120126, 135-136 approval of Edelman concessions, 190 State interest in strike avoidance, 105-106, 113-119, 135-136, 180-186 political concerns, 119 State ownership of coal assets, 120-122. See also Saar, French control of dominial mines State participation in industrial relations, 13-15, 105-195, 203-204, 207-208 effect on strike duration, 203-204 state control increased by steering, 164, 187-191 support for collective bargaining to promote industrial peace, 129, 130, 147, 149-154, 155, 164, 178-179, 187-189

Index State participation in industrial relations (cont.) unemployment reduction, attempts at, 132-135, 162, 183-186 see also Repression; State support for coal miners State regulation of coal mining, 94, 105, 120-122, 130-132, 142, 144, 148, 186191, 207-208 State support for coal miners, 13, 55, 126135, 138-162, 173-175, 188, 189 link to strike success, 83, 129 withdrawal of support for unitaires, 155162, 164, 178 toleration of some unitaires, 155, 157159, 178-179 see also Repression; State interest in strike avoidance Static measures of demand moderation, 6 7 69 79-82. See also Cost theory. Strong managerial control. Weak managerial control Statistiques de I'lndustrie Min^rale, 7, 43, 80, 97, 207 Statistiques des Greves, 9, 14, 43, 49, 73, 99 Status attainment research, 3 Steams, Peter, 4, 108, 117 Steel companies, relations with coal companies, 115n, 116-117 Steering, 34, 36-37, 40, 163-196, 215, 221-223 accommodation potential, 164, 165, 172, 195, 222 benefits workers in mixed representation districts, 38, 59-61, 64, 186 mixed representation benefit not applicable when radicals in majority, 60n benefits moderate workers, 38, 59-61, 63-64, 163, 186, 195 passive moderates outeam strikeprone radicals, 63-64 desirability of having archival evidence to substantiate, 163 does not lower overall wage payments, 57-59, 164, 183, 186, 188, 191, 222 facilitated by employers' associations, 106 facilitated by government intervention, 106-107 facilitates state control over labor movement, 164 forces radicals to be majority for strikes to be effective, 39-40, 61-64, 155, 221

Index Steering {cont.) gut effects of moderate strikes, 39, 40n, 61-64, 163, 195, 221 moderate strikes ineffective in mixed and communist areas, 62-63 moderate strikes lower wages in moderate areas, 62, 163 joint strikes punished in moderate and mixed areas, 62-63, 163, 179-180, 221 learned response by authorities, 163-164, 222 prewar, 171-172, 174 public good from the point of view of individual employer, 106 radical strikes lower wages in moderate and mixed areas, 62-63 radical strikes raise wages in radical areas, 62-63, 155-162, 221-222 versus weakening as explanation of period effects, 57-64 see also Dual unionism; Joint strikes Stephens, Evelyn, 89 Stephens, John, 89 Sternhell, Zev, 127 Stinchcombe, Arthur, 38 Stone, Judith, 126-127, 129 Strategic errors, origins of, 5 Strategic studies, importance of, ix, 1, 5-6, 6n Strategy, optimal, 27-41, 213-226 frequent striking while buffered from capital mobility, 226 in one-union case, 28-34 in two-union case, 34-40 Strength requirements for striking, 28-29, 40, 223-226 Strike law of 1892, 72, 141 Strikes as a cost or benefit in their own right, 31, 32, 33, 213-214 capacity for harming employers, 109-113, 114, 217-218 disadvantages of general strikes, 113 greatest in first days of strike, 218 role of deadline pressure, 109, 113, 218 role of capacity to recoup revenues after strike, 109 role of resource differentials, 108-109, 112-113, 218 role of surprise, 114 see also French coal miners, power of

251 Strikes (cont.) causes of, 2 - 3 , 30n, 31 categorical transformation of, 50-51, 51n coding for grievance content, 79-80 costs of, 30-31, 200, 203, 218-219 cyclical theories of, 33 data, 8-9 days lost. See Strikes, mandays lost deterrrunants of winning, 3, 79-83 moderation of strike demands, 66, 8 1 83,85 no link to period, turnout, government intervention, 83 past concessions, 69-70, 81-83 traditional claims of importance of strength, 65-66, 80-82, 83 difficult for union leaders to stop, 223 duration, vi, 80, 142, 197, 200-204, 217 average, 203 of critical battles, 203 longest among contestants with equal strength, 201 longest among workers with intermediate strength, 201 meaningless for cases with no strikes, 202 risk increases with, 219 state mediation's effect on, 203-204 traditional claims of association with strength, 201 see also Strikes, effects of, by duration effects of, vi, 5, 27-42, 52-64, 97, 204212, 213-214 by confederes, 61-64 by duration, 201-203, 212, 217-219 by joint strike forces, 61-64 by unitaires, 61-64 compared to neoclassic predictors, 204207, 212 importance of estimating long-term wage effects for, 86 importance of estimating strength of opposition, 85-86 magnitude, 197, 198-199, 200 by day, 202-203 by month, 203 removing effect of strike waves, 199 using full range of strike variance, 199 misestimated by labor historians, 8 5 86

252 Strikes {cont.) misestimated by labor historians {cont.) no effect unless period taken into account, 52-54, 64 ratio of strike effects to neoclassical effects, 205-207 effects of vs. causes of, vi 2-3 frequency, measurement of, 49-51 frequency of as bargaining data, 30-32 in confedere departments, 61-64 in mixed departments, 61-64 in unitaire departments, 61-64, 155-162, 216-217 mandays lost, 80-81 only startable by workers, 33-34 rareness of, 5 violence, 117-119, 141, 142, 143, 144 chasse des renards, 118, 158 see also Dual unionism, effect on strike effectiveness; Duration; Imperfect information as cause of strikes; Strength requirements for striking; Winning, effects of Strong managerial control, 68-69, 75, 76, 77-78 Sturmthal, Adolph, 2, 31, 35, 69n, 72, 92, 130, 171 Success and compromise combined, 76, 76n, 80-81, 99-100 Surcharge disputes, 70, 71, 76-77, 78 Swedish unionism, 9, 11, 35n, 114, 168, 220 Syndicalism, 6, 31, 166-175. See also Anarchists

Tarn. See Carmaux Taxation of coal, 120-121, 134 Taylorism, 17, 73, 134n, 183, 189n, 210 resistance to, 73-74 Teamsters, 222 Thaler, Richard, 88 Theory of the state, 124 Tilly, Charles, v-vii, x, 2, 3, 4, 6, 9, 10, 34, 49, 72, 108, 129, 141, 168 Time as control variable, 46, 52-59, 84, 94, 97, 103, 206 Timing, pay. See Late pay/pay timing Tit for tat, efficacy of, 37-38 Tournier, Alain, 176 Tracy, Joseph, 31, 33, 68, 201

Index Trade-offs, monetary and nonmonetary benefits, 87-104, 214-216 complementarity of monetary and nonmonetary benefits, 92, 214, 215 empirical correlation of monetary and nonmonetary benefits, 92-99, 215 empirical correlation of nonmonetary demands with wages, 102-103, 215 only pay-nonpay tradeoffs considered, 99-100 rarity of same, 100-101 theoretical potential of trade-offs, 87-89 see also Compensating differentials; Pay strikes, effect of; Union choice Training times, 74 Trempe, Rolande, 11, 13, 14, 18, 19-20, 74, 94, 117, 118, 124, 126, 129, 131-133, 146, 147, 154, 207 Trend. See Time as control variable Tripartite collective bargaining. See Collective bargaining, tripartite Trotsky, Leon, 3, 88, 220 Turnout, 110 157n, 160, 176, 177 exceeds union membership in France, 47 supposed link to strike effectiveness, 36, 83 Turnover, 74 Unemployment, 17, 45n, 132-135, 183-185 Marches against, 119, 133, 162 Paid employer relief, 134-135, 183-185 Paid state relief, 133 Political consequences, 133-135, 185 see also Labor scarcity Union affiliation, measurement of, 59-60 Union choice, 88-89, 91-93 among grievances empirically rare, 93, 99-102 lack of evidence for, 89 less important than variance in worker strength, 91-92 noneconomic demands raise rather than lower wages, 92-93, 102-104 relevant only to workers of intermediate strength, 91 see also Economism; Marxists; Trade-offs between monetary and nonmonetary benefits Union decentralization as cause of militancy, 31-32, 35, 47 legal incentives for in France, 47

Index Union membership dual membership common in pre-World War I France, 165 measurement of, 47-49, 59-60 guttman scale relation between striking and union membership, 48 measures intentionally incorporate unspecified class conflict effects, 48 weak link to actual militancy in France, 47, 72, 72n see also Union wage effect Union procedural disputes, 71, 72, 72n, 7677, 78, 100-101, 138-140, 142-143, 144-146 see also French unionism; Union membership, weak link to actual militancy in France Union recognition. See Federation des Travailleurs du Sous-sol, recognition of; French unionism; Union membership, weak link to actual militancy in France Union wage effect, ix, 1, 19, 27, 40, 52-59, 64, 84, 97, 103, 145, 204-207 ratio of union wage effect to neoclassical effects, 205-207 Unionism. See British unionism; Christian unionism; Contemporary American unionism; British unionism; French unionism; German unionism; Swedish unionism; White collar unionism Unitaires, 6, 11, 12 attempts to form front unique, 119, 157, 160, 161, 175 effect of strikes. See Steering effectiveness as tacticians, 55 hostility to parliamentary socialists, 159n international coordination of strikes, 110 measurement of affiliation to, 59-60 nonparticipation in collective bargaining, 155, 156, 158-159, 178-179 pension and disability representatives, 181, 182, 184, 215 political strikes, 92, 124n, 158-159, 160 regional base, 18-19, 26, 155-156, 168, 168n relations with Christian unionists, 180 shop floor disputes, 18, 73-74, 157-160 strike militancy, 12, 160, 161 support for foriegn workers, 157-159, 194, 215

253

Unitaires {cont.) unemployment marches, 119, 133, 162 wages, pursuit of, 12, 157, 215 see also General Strikes of 1923; Work assignment/rule disputes United Farm Workers, 222 United Kingdom Royal Commission on Trade Unions and Employers Associations, 2, 35, 93 U.S. unionism. See Contemporary American unionism Valois, Georges, 130 Vendee, 19, 22-26, 178-179, 193 Verpilleux, 173-175 Viaud, M., 73 Villain, Francois, 130 Villeboeuf, 118, 156, 157-9 Violence. See Strikes, violence Viscusi, W. Kip, 88, 9, 98 Viviani, Rene, 173-174 Vuillemin, 139 Wage change as indicator both of radicalism and of bargaining strength, 81-82 Wages, general differences between actual payments and published rates, 44-45, 139n, 167 importance of, 4, 19, 87, 87n rareness of data on, 7 sticky downward, 51-52, 98, 145 Wage levels in French coal mining, 16, 23, 134, 143, 145, 162, 170-172, 178, 188, 190, 191, 211-212 measurement of, 44-5, 46, 132 national coal wages as determinants of, local wages, 46, 52-59, 84, 97, 103, 206 Wages past, 80-83 Waldeck-Rousseau, Pierre, 128, 147 Wallace and Hussein models, 51, 52n Wallace, Michael, 3, 29, 96 Walsh, Kenneth, 4, 217 Walton, Richard, 12n, 31n, 32, 66, 220 Watrin defenestration, 117 Weak managerial control, 69, 75, 76, 77, 7 8 79. See also Fire supervisor disputes Webb, Beatrice, 29 Webb, Sydney, 29 Welfare State Benefits for Miners, 129, 130135, 178, 191. See also Polifical conces-

254 Whipsawing, 35-36, 173-175 White collar unionism, 31 Winning, effects of, vi, 1-2, 65-86, 216-217 direct estimation of effect of wins versus effect of losses, 83-85 distortion potential from observations with no strikes, 83 lack of empirical information on, 65 mutual cancellation of strength and militancy effects, 66-67 why losers should earn more than winners, 65-67, 216 why winners should earn more than losers, 65 see also Strikes, determinants of winning Wisconsin School, 2, 4, 92, 103, 214-216. See also Economism Women's employment and conditions, 170, 193 Wolfe, John, 88

Index Work assignment/rules disputes, 2, 71, 7274, 76-77, 78, 79n, 100-101, 144-145, 159-160, 167, 189n. See also Taylorism Workdays as denominator. See Days worked as denominator in rates Work hour disputes, 71, 74-75, 76-77, 78, 100-101, 144-146, 147, 172-175. See also Eight-hour day; Length of workday Working class representation in government. See Parliamentary left World War I, 16, 43, 115-116, 122, 131-132. See also Schism versus war as basis for periodizing strike effects Wright, Erik, 3n Year. See Time as control variable Yellow unionism. See Christian unionism Zevaes, Alexandre, 128, 129 Zimbalist, Andrew, 210