Participation in the Age of Globalization and Information, Volume 9 (Advances in the Economic Analysis of Participatory & Labor-Managed Firms)

PARTICIPATION IN THE AGE OF GLOBALIZATION AND INFORMATION i ADVANCES IN THE ECONOMIC ANALYSIS OF PARTICIPATORY AND L...

51 downloads 932 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

PARTICIPATION IN THE AGE OF GLOBALIZATION AND INFORMATION

i

ADVANCES IN THE ECONOMIC ANALYSIS OF PARTICIPATORY AND LABOR-MANAGED FIRMS Series Editor: Derek C. Jones

ii

ADVANCES IN THE ECONOMIC ANALYSIS OF PARTICIPATORY AND LABOR-MANAGED FIRMS VOLUME 9

PARTICIPATION IN THE AGE OF GLOBALIZATION AND INFORMATION EDITED BY

PANU KALMI Helsinki School of Economics, Finland

MARK KLINEDINST University of Southern Mississippi, USA

Amsterdam – Boston – Heidelberg – London – New York – Oxford Paris – San Diego – San Francisco – Singapore – Sydney – Tokyo iii

ELSEVIER B.V. Radarweg 29 P.O. Box 211 1000 AE Amsterdam, The Netherlands

ELSEVIER Inc. 525 B Street, Suite 1900 San Diego CA 92101-4495 USA

ELSEVIER Ltd The Boulevard, Langford Lane, Kidlington Oxford OX5 1GB UK

ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK

r 2006 Elsevier Ltd. All rights reserved. This work is protected under copyright by Elsevier Ltd, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK: phone (+44) 1865 843830, fax (+44) 1865 853333, e-mail: [email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2006 British Library Cataloguing in Publication Data A catalogue record is available from the British Library. ISBN-10: 0-7623-1278-5 ISBN-13: 978-0-7623-1278-8 ISSN: 0885-3339 (Series)

∞ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

iv

CONTENTS LIST OF CONTRIBUTORS

ix

FOREWORD

xi

INTRODUCTION Panu Kalmi and Mark Klinedinst

xiii

THEME I: EMPLOYEE PARTICIPATION AND FIRM PERFORMANCE WHAT TYPES OF ORGANIZATIONS BENEFIT FROM TEAM PRODUCTION, AND HOW DO THEY BENEFIT? Jed DeVaro and Fidan Ana Kurtulus

3

THE NATURE, SCOPE AND EFFECTS OF JOINT LABOR-MANAGEMENT COMMITTEES IN JAPAN Takao Kato

55

EMPLOYEE OWNERSHIP AS A JOINT MANAGEMENT-LABOR DRIVE TOWARD CARING AND SHARING Erik Maaloe

81

EMPLOYEE PARTICIPATION RIGHTS IN CORPORATE GOVERNANCE: AN ECONOMIC RATIONALE, A TEST OF A LEADING THEORY, AND SOME INITIAL POLICY PROPOSALS Stephen C. Smith v

105

vi

CONTENTS

THEME II: MANAGERIAL COMPENSATION AND CONTROL PERFORMANCE-RELATED PAY IN DANISH COOPERATIVE CREAMERIES Morten Hviid

149

CORPORATE GOVERNANCE AND EXECUTIVE COMPENSATION IN BULGARIA AFTER MASS PRIVATIZATION: EVIDENCE FROM NEW PANEL DATA Derek C. Jones and Mark Klinedinst

177

THEME III: CO-OPERATIVES AND SOCIAL ENTERPRISES WORKERS COOPERATIVES AS AN ALTERNATIVE COMPETITIVE ORGANIZATIONAL FORM Morris Altman

213

CO-OPERATIVES IN GLOBALIZATION: THE ADVANTAGES OF NETWORKING Isabelle Halary

237

POSITIVE EXTERNALITIES OF ORGANISATION CULTURE: THE SOCIAL INTEGRATION OF WORKING ADULTS WITH LEARNING DISABILITIES Abraham Mukolo, Robert Briscoe and Agus Salim

265

THEME IV: PARTICIPATION: PAST, PRESENT, FUTURE GUILD SOCIALISM REVISITED, AND ITS YUGOSLAV COUNTERPART Mario Ferrero

299

Contents

vii

WHITHER SELF-MANAGEMENT? FINDING NEW PATHS TO WORKPLACE DEMOCRACY David Ellerman

321

THE FUTURE OF BROAD-BASED OPTIONS: WHAT RESEARCH TELLS US Corey Rosen

357

This page intentionally left blank

viii

LIST OF CONTRIBUTORS Morris Altman

Department of Economics, University of Saskatchewan, Canada

Robert Briscoe

Centre for Co-operative Studies, University College Cork, Cork, Ireland

Jed DeVaro

Department of Labor Economics, Cornell University, NY, USA

David Ellerman

Economics Department, University of California at Riverside, CA, USA

Mario Ferrero

Department of Public Policy and Public Choice (POLIS), University of Eastern Piedmont, Alessandria, Italy

Isabelle Halary

University of Reims, France

Morten Hviid

The Norwich Law School and Centre for Competition Policy, University of East Anglia, UK

Derek C. Jones

Department of Economics, Hamilton College, NY, USA

Panu Kalmi

Department of Economics, Helsinki School of Economics, Helsinki, Finland

Takao Kato

Department of Economics, Colgate University, NY, USA, IZA Bonn, Germany and Center on Japanese Economy and Business, Columbia Business School and TCER, NY, USA

Mark Klinedinst

Department of Economics, University of Southern Mississippi, MS, USA

Fidan Ana Kurtulus

Department of Economics, Cornell University, NY, USA

Erik Maaloe

Aarhus School of Business, Denmark ix

x

LIST OF CONTRIBUTORS

Abraham Mukolo

Centre for Co-operative Studies, University College Cork, Cork, Ireland

Corey Rosen

Executive Director, National Center for Employee Ownership, Oakland, CA, USA

Agus Salim

Department of Statistics, University College Cork, Ireland

Stephen C. Smith

Department of Economics, George Washington University, Washington, DC, USA

FOREWORD This volume of Advances in the Economic Analysis of Participatory and Labor-Managed Firms marks the third volume in this series to be produced by Elsevier. The two previous volumes, Employee Participation, Firm Performance and Survival, edited by Virginie Perotin and Andrew Robinson, published in 2004, and The Determinants of the Incidence and the Effects of Participatory Organizations, edited by Takao Kato and Jeffrey Pliskin and published in 2003, marked the re-launching of the series. (The series began in 1985. Six volumes appeared during 1985–1995 when the series was published by JAI and was co-edited by Jan Svejnar and Derek C. Jones.) A key aim of the re-launched series is to publish Advances on a regular and annual basis. Reflecting a deepening pool of talent as the field of participation has grown during the last 20 years or so, another change, as is evident in the present volume, is to make frequent use of guest editors for issues of Advances. As series editor, I welcome suggestions and proposals from readers for particular issues. Other changes concerning Advances will be modest. Advances will continue to act as a forum for high-quality original theoretical and empirical research in the broad area of participatory and labor managed organizations. The original rationale for the series was the observation that while general and specialized journals publish work in this field, many do so only occasionally. There continues to be a need for an annual periodical that presents some of the best papers in a single volume. While the focus will continue to be on economic issues, analytical studies on closely related areas are also welcome. Advances will also continue to serve as an outlet for high-quality pieces that regular journals often consider to be too long. The broad area of participation and labor management has changed much since the inception of the series in 1985. The tragic disintegration of the Former Republic of Yugoslavia also meant the disappearance of the principal systemic example of self-management. But the collapse of the former USSR has also triggered widespread experimentation with diverse forms of participation in many transition economies, notably many firms with large degrees of employee ownership. Among firms in western economies we also witness xi

xii

FOREWORD

the continued growth of diverse institutional arrangements that provide for participation by employees in decision-making as well as in enterprise results. Many economies continue to be characterized by flourishing cooperative sectors, though the scope, nature and form of these sectors varies markedly across countries and over time. Of special note perhaps is the fact that several important examples of worker cooperatives continue to thrive, with the Mondragon Cooperative Consortium now representing the seventh largest consortium in Spain. Against this institutional backdrop much new and innovative theoretical and empirical work in the broad field has appeared. The key aim of the Advances series continues to be to make it a broad-based periodical within which is presented both new theoretical results and fresh evidence on the performance of participatory firms and sectors. The intent is to maintain high quality and to place this periodical among other successful Elsevier series. I hope you will be informed and stimulated by this volume and that you will consider contributing to it and conveying information about Advances to other interested colleagues. Derek C. Jones Series Editor

INTRODUCTION This volume of Advances in the Economic Analysis of Participatory and Labor-Managed Firms consists of 12 original and innovative articles. The first four papers relate to the growing literature on employee participation and firm performance. The second group of papers looks at the impact of ownership structures on managerial compensation and control. The third set of papers analyzes the role of co-operatives in the changing economic environment. The three papers in the final section range from the historical perspective on participation to the role different forms of participation may play in the future. The first article in the first part of the book, Employee Participation and Firm Performance, is by Jed DeVaro and Fidan Ana Kurtulus on ‘‘What Types of Organizations Benefit from Team Production, and How Do They Benefit?’’ This study uses a cross section of British companies from 1998 to estimate structural equations that treat as endogenous the decisions to have team production and autonomy. By using three different measures of firm performance they analyze how teams benefit the organization. A finding that is echoed in later papers of the book is the complementarity hypothesis that firms with higher levels of financial participation are more often those who benefit from team production. A number of other results are analyzed here concerning unionization, firm size, industry, age of the firm, etc. They also find that the typical establishment benefits from team production in higher financial performance, product quality and higher labor productivity. Takao Kato’s insightful article ‘‘The Nature, Scope and Effects of Joint Labor-Management Committees in Japan’’ pulls together the theoretical literature on high-performance human resource management and an extensive data set on Japanese firms. The sample data come from a survey of some of the largest firms in the country and detail the practice of participatory committees, unionization, employee stock ownership plans and profit sharing. The diffusion and extent of these practices in Japan are analyzed, a reminder to us that some of these programs are fairly new but have made substantial headway. This data is combined with outcome measures to determine if these various schemes have an effect. A number of different specifications are estimated in an attempt to single out among all xiii

xiv

INTRODUCTION

the participatory practices used, which have the most powerful impact on the firm. The results consistently show the strength of information sharing through joint labor-management committees. Of particular note in his research are complementarities that arise among participative practices, especially those that extend to the lowest levels of the firm’s organizational structures, e.g., unions and shop floor committees. Erik Maaloe’s article ‘‘Employee Ownership as a Joint ManagementLabor Drive Towards Caring and Sharing’’ is based on extensive fieldwork in six employee-owned companies in Ohio. In his article, Maaloe describes the difficult process from the transfer of property rights to the emergence of the culture of employee ownership. He highlights the role of different actors in this process, and the difficulties management and employees face in revising their attitudes, beliefs and behavior during the process. Maaloe stresses the role of experience, mutual trust and honesty in the formation of co-operative culture. Revision of one’s beliefs, respect to the other party and transmission of information are found to be key elements promoting these processes. Maaloe’s article is not only scientifically important, but it has also important practical implications for setting up employee-owned firms and developing a co-operative culture. Stephen C. Smith’s article on ‘‘Employee Participation Rights in Corporate Governance: An Economic Rationale, a Test of a Leading Theory, and Some Initial Policy Proposals’’ covers a broad range of interests in the literature from theoretical and empirical to concrete proposals for encouraging the broader ownership in the United States. Smith stresses the importance of Employee Participation Rights (EPRs) in that they provide for legal guarantees to the design of the workplace rather than just employer whim. A variety of EPRs are considered from a number of different countries. Evidence is presented for 723 German companies on the positive impact from training with works councils. The experiences with EPRs from different parts of the world are then argued to be an important foundation of a stable, equitable and efficient democratic market system. Policy proposals are then developed for the US that are practical and minimize government interference. The second theme is Managerial Compensation and Control. Morten Hviid’s paper ‘‘Performance Related Pay in Danish Cooperative Creameries’’ studies the use of managerial incentive schemes among the Danish creamery co-operatives in the late 19th and early 20th century. He uses a unique data set based on archival research. As such, it is the first study on managerial incentive problems among co-operatives, and also an important contribution to the literature on managerial compensation.

Introduction

xv

Hviid’s findings show that the creameries were aware of the potential incentive problems. They varied over time in response to changes in the external environment, also within individual creameries. This finding goes against some earlier findings on within-firm contract uniformity. Another interesting finding is that the contracts were often quite complex. The author suggests that the simple production process might have made complex contracts feasible. However, a number of co-operatives also used an interesting simple solution to the incentive problem: they gave the manager property rights to a particular amount of milk, thus aligning the incentives of the manager with the producer-members. Derek C. Jones and Mark Klinedinst use a new data set to look at ‘‘Corporate Governance and Executive Compensation in Bulgaria after Mass Privatization: Evidence from New Panel Data.’’ This data set covers some of the more dynamic years of the transition in Bulgaria, the period from 1997 to 2001. Using a number of performance measures and controls they look at new firms and some firms that have been tracked by the authors since before the transition began. Managers seem to be coming out on top of the privatization with more control and higher pay in both absolute and relative terms. Although the spread in pay rates has grown over time, it is still well below that found in other older market economies. CEO pay is sometimes determined by performance measures, but often the size and ownership are the critical factors. The third part of the book, Co-operatives and Social Enterprises, begins with Morris Altman’s paper on ‘‘Workers Cooperatives as an Alternative Competitive Organizational Form.’’ Altman shows that when the preferences of workers are taken into account and the costs that might be found from developing participative firms, then it is possible to find more efficient participative firms that pay their workers well competing on a fairly level playing field with traditional hierarchical firms. The continuing x-inefficiency of typical firms allows the well paying participative firms to elicit greater effort and to continue to compete successfully. A number of graphs are used to show how this might play out given variations in effort and technology. The preferences of workers to be a part of a participative firm that is more efficient and pays better is often stymied by the lack of information about organizational forms, preferences of the owners and the lack of ready access to financial markets. Isabelle Halary’s article ‘‘Co-operatives in Globalization: The Advantages of Networking’’ relates to the question of the sustainability of co-operatives. The author argues that membership in a network is important in order to survive in global competition. While this applies to both conventional firms

xvi

INTRODUCTION

and co-operatives, co-operatives are likely to have an advantage in this respect, because they have the common culture, social skills and shared values that are necessary to build up networks. Co-operative form is also insulated from the pressures from financial markets and therefore well suited for establishing long-lasting relationships. Halary evaluates these arguments by studying empirically three wellknown groups of co-operative enterprises: Mondragon Corporation Co-operative in the Basque country in Spain, the industrial districts of Emilia-Romagna, and the French SCOP Entreprises. The networks of these three forms of co-operatives share some similarities, especially in horizontal and financial networks. However, the Mondragon co-operatives as well the co-operatives Emilia Romagna also have more vertical ties and a stronger territorial element than the French co-operatives. Overall, her paper is an important contribution on the nature and significance of networks among co-operatives. Abraham Mukolo, Robert Briscoe and Agus Salim’s paper on ‘‘Positive Externalities of Organisational Culture: The Social Integration of Working Adults with Learning Disabilities’’ shows how important workplace participation is determining a person’s overall attitudes (and probably aptitudes). The more input that people with disabilities have in decision-making at their workplace the more likely they are to have more positive interactions in their private lives outside of work. A number of different types of work settings and characteristics, as well as controls for individuals such as age and degree of disability are considered when making comparisons. Attempts to integrate adults with disabilities in society would be guided by this research to investigate the possibilities of encouraging these adults by setting up employment for them that gives them the chance to have their voice be heard. The final section of this book contains three essays on Participation: Past, Present and Future. Mario Ferrero’s article ‘‘Guild Socialism Revisited, and its Yugoslav Counterpart’’ describes an early 20th century socialist blueprint, Guild Socialism. It was the vision of a corporatist worker-governed socialism, based on multiparty negotiations of incomes, prices and quantities. Ferrero presents in his article an enlightening discussion on the central tenets of this almost-forgotten theoretical construct. Differently from pure syndicalist models, the Guild Socialist blueprint involved the use of elaborate negotiating and bargaining institutions to check the power of the producers. The Guild Socialists also departed from the collectivists, who wanted to delegate the management of the industries to state representatives. In contrast, some Guild Socialists, notably G.D.H. Cole, wanted to

Introduction

xvii

replace the machinery of the state by a system of functional representation through the Guilds. Ferrero goes on to compare Guild Socialism with the actual Yugoslav experiment of worker self-management. He shows that these two systems actually had a number of similarities, in respects to the decentralization of enterprise governance, but centralization of many basic economic decisions; suppression of political and economic competition; multi-party negotiations at all various stages; and political representation. Ferrero presents convincingly the case that the Yugoslav experiment was in many respects closer to the Guild Socialist model than to its allegedly theoretical representation, the Labor-Managed Firm – model. David Ellerman contributed the chapter titled ‘‘Whither Self-Management? Finding New Paths to Workplace Democracy.’’ This paper was given as the keynote address to the 12th Conference of the International Association for the Economics of Participation meeting in Halifax, Canada in 2004. This paper searches out the roots theoretically in the literature on democracy on why firms should be organized participatively in a truly democratic society. The inalienable rights arguments that have helped stop such abuses as slavery are shown to have application to the theory of the firm and the law. His arguments clearly show the fraud of equating the consensual contracts for ‘‘things’’ as wholly different than the supposedly non-coercive self-sale involved in an employment contract. The law certainly realizes that a person who makes a criminal decision while employed is held accountable, but firms generally do not allow the employee the ability to participate in the decision-making that runs key aspects of the firm and often treats the employee as a ‘‘thing,’’ alienated from the choices that have to be made. He cites a conservative Englishman Eustice Percy, the German codetermination laws and the Japanese post-war experience in pointing out possible routes on the path to workplace democracy. The National Center for Employee Ownership’s Director, Corey Rosen contributed ‘‘The Future of Broad-Based Options: What Research Tells Us.’’ This paper comes from the rarely found perspective of someone who combines the talents of an entrepreneur, a leader in the effort to make employee ownership more widespread and who also keeps abreast of research in the area as well. This broad background gives credence to his overview of employee ownership and related research in the US. The complementarities found in some of the other papers of this volume are given further support in this paper. Companies that combine broad-based employee ownership with open-book management and a high-involvement

xviii

INTRODUCTION

organization of work are the best companies in terms wages, wealth accumulation and other measures of corporate performance. The spate of scandals in corporate governance (e.g., Enron and WorldCom) has brought increased scrutiny on the accounting of the awarding of stocks and options. Corey Rosen looks at the impact these rule changes may bring on broad-based employee ownership. Although realizing that the public’s understandable concern with irrational executive compensation, that may include stock, may lead some companies to not offer employee ownership at the same level they have in the past, this paper points out that the large number of employees now who own stock and the proven benefits indicate that broad-based employee ownership in the US will likely remain strong and possibly grow. Together these papers draw on the expertise of a number of leading thinkers in the area and promise the reader a challenge to help deliver practical policies to transform our work world and society. Panu Kalmi and Mark Klinedinst Guest Editors

THEME I: EMPLOYEE PARTICIPATION AND FIRM PERFORMANCE

1

This page intentionally left blank

2

WHAT TYPES OF ORGANIZATIONS BENEFIT FROM TEAM PRODUCTION, AND HOW DO THEY BENEFIT? Jed DeVaro and Fidan Ana Kurtulus ABSTRACT Using data from a large cross-section of British establishments, we ask how different firm characteristics are associated with the predicted benefits to organizational performance from using team production. To compute the predicted benefits from using team production, we estimate structural models for financial performance, labor productivity, and product quality, treating the firm’s choices of whether or not to use teams and whether or not to grant teams autonomy as endogenous. One of the main results is that many firm characteristics are associated with larger predicted benefits from teams to labor productivity and product quality but smaller predicted benefits to financial performance. For example, this is true for union recognition as measured by the number of recognized unions in an establishment. Similarly, when a particular firm characteristic is associated with lower benefits from teams to labor productivity or product quality, the same characteristic is frequently associated with higher predicted benefits to financial performance. This is true for the degree of financial participation and employee ownership and also for establishment size and a Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 3–54 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09001-0

3

4

JED DEVARO AND FIDAN ANA KURTULUS

number of industries. These results highlight the advantages of analyzing broader measures of organizational performance that are more inclusive of the wide spectrum of benefits and costs associated with teams than the labor productivity measures frequently studied in the teams literature.

1. INTRODUCTION Arguments suggesting that innovative systems for organizing and managing employees generate improved employee achievement and organizational performance abound. Recent research on workplace practices transferring power to employees, described collectively as ‘‘high-performance practices,’’ has identified employee participation as a key element of sustained competitive advantage. Employee participation in the form of team-based work structures, often ‘‘self-managed teams’’ conferred with considerable autonomy, figures prominently as one dimension of high-performance work systems and is the focus of this paper. Specifically, our goal is to shed light on the question of what types of organizations benefit from team production and how they benefit (e.g., through higher financial performance, labor productivity, or product quality). Much of the vast literature on teams and other high-performance practices focuses on the question of whether and how organizations benefit from use of these practices, as opposed to the question of what types of organizations benefit. A popular approach is the case study, examining one or a relatively small number of firms, usually over time (e.g., Bartel, 2004; Batt, 1999, 2001, 2004; Hamilton, Nickerson, & Owan, 2003; Batt & Appelbaum, 1995). While the case study approach is useful for answering questions about whether and how organizations benefit from teams, it is somewhat less useful for answering questions about what types of organizations benefit, since by their very nature case studies involve little or no variation in firm characteristics. Introducing variation in firm characteristics usually requires comparisons of different case studies, as in meta analyses.1 The difficulty with this approach is that studies vary widely in their data, measurement of variables, methodology, sampling period, geographic region of analysis, research questions, and in a variety of other respects. If the goal is to identify how a particular firm characteristic, say firm size, is associated with the predicted benefits to organizational performance from using teams, this question is unlikely to be answered convincingly through comparisons across case studies. Such a task would involve assembling a relatively small number of case studies that vary by the size of the organizations studied; inevitably these organizations will vary in a multitude of

What Types of Organizations Benefit from Team Production

5

dimensions other than the firm size, making it impossible to control for these other factors. Even if it were possible to make controlled comparisons, it would not be clear what weights should be assigned to the individual studies being surveyed. An alternative to the case study approach, and the one taken in this paper, is the use of broader cross-sections or panels of organizations. Examples of studies taking this approach include Black and Lynch (2001, 2004), DeVaro (2004a, b), Eriksson (2003), Kato and Morishima (2002), Cappelli and Neumark (2001), and Caroli and van Reenen (2001). Like case studies, broader samples of firms can shed light on the question of whether and how organizations benefit from teams. In addition, these data are somewhat more conducive to answering questions about what types of organizations benefit from teams, due to their inherent variation across organizational types. Of course, these broader data sets suffer a number of disadvantages relative to case studies. First, since the samples are more heterogeneous, the definitions of variables (for instance, the meaning of ‘‘team production’’) are not as obviously comparable as they would be across observations in a single case study. Second, significant heterogeneity in these samples increases the threat that unobserved heterogeneity may bias the estimated effect of teams on organizational performance. Other things equal, panel data are always preferable to cross sectional data in that they can accommodate individual effects to mitigate concerns about unobserved heterogeneity. However, exploiting such panel data in studies of team production invariably involves compromising either on the breadth of the sample or on the richness of the information available in the data or both. Our study uses a large, nationally representative cross section of British establishments in 1998. Our sample has variation not only in whether or not team production is used, but on the type of team production used, in particular whether or not team members are granted autonomy. In addition to detailed firm characteristics for use as controls, the sample also includes multiple measures of organizational performance (financial performance, labor productivity, and product quality). To our knowledge, there are no large, nationally representative panel data sets available that contain information on the types of teams used (autonomous or non-autonomous), firm characteristics, and multiple measures of organizational performance. Our research strategy is therefore to exploit the unique and extensive information contained in our cross-sectional data, while estimating structural models to address concerns about unobserved heterogeneity biases. There are three main distinguishing features of our work. First, we estimate structural models that treat the choices of team production and

6

JED DEVARO AND FIDAN ANA KURTULUS

whether to grant teams autonomy as endogenous variables, as opposed to the typical approach that treats these variables as exogenous right-hand-side variables in a regression. Second, our models interact the ‘‘teams’’ treatment with all other firm characteristics (both observed and unobserved). Allowing the teams treatment effect to vary with organizational characteristics allows us to make statements about what types of organizations are predicted to benefit from team production. Third, we compare three measures of organizational performance (financial performance, labor productivity, and product quality) in the same analysis, allowing us to make statements about how teams benefit organizational performance. Though there are some exceptions, such as DeVaro (2004a) and Huselid (1995), financial performance is rarely seen as the outcome measure in studies of the effects of high-performance practices on organizational performance. This variable is of particular interest as an overall measure of firm performance, since as a measure of profit it is more inclusive than other outcome measures of the wide array of benefits and costs associated with human resource practices.2 There are three stages to our empirical analysis. In the first stage, we estimate structural models for our three measures of organizational performance. In the second stage, we use the parameter estimates from the first stage to compute the ‘‘predicted benefit to organizational performance from using team production’’ for each establishment in our sample. In the third stage, we assess how a particular firm characteristic is associated with the predicted benefits to organizational performance from using team production, holding other firm characteristics constant. A progressive example is useful for motivating our structural approach. Consider the following two regressions, where Y is a continuous measure of firm performance, TEAMS is a dummy variable equaling 1 if the firm uses team production and 0 otherwise, X is a firm characteristic such as firm size, and e a disturbance term uncorrelated with TEAMS and firm size: Y ¼ b0 þ b1 TEAMS þ b2 X þ

(1)

Y ¼ b0 þ b1 TEAMS þ b2 X þ b3 ðTEAMS X Þ þ

(2)

(In practice, of course, X would be a vector containing many characteristics other than firm size.) If the goal is to ask what the effect of teams is on firm performance, either regression would be informative. The teams ‘‘treatment effect’’ of interest would be b1 in Model 1 and b1+b3X in Model 2. If, on the other hand, the goal is to ask (as in this paper) whether the benefits from teams are greater in larger firms than in smaller firms or vice versa,

What Types of Organizations Benefit from Team Production

7

only Model 2 is informative. If the estimated b3 is found to be negative, we would conclude that the benefits to organizational performance from using teams decrease with firm size. In sum, Model 2 provides a simple way of answering our research question of what types of firms (or in this example what size firms) benefit from team production. Model 2 assumes that the unobserved determinants of firm performance are the same whether team production is used or not. Allowing these unobserved determinants of firm performance to differ by whether or not teams are used, or interacting TEAMS with e, yields the following generalization of Model 2: Y ¼ b0 þ b1 X þ 1 ¼ b2 þ b3 X þ 0

ifTEAMS ¼ 1 ifTEAMS ¼ 0

ð3Þ

The teams treatment effect in Model 3 is (b0 – b2)+(b1 – b3)X. This is computationally identical to the teams treatment effect in Model 2. Models 1–3 can be criticized on the grounds that they assume that the disturbances are uncorrelated with TEAMS, a case that is quite difficult to make. Even if X were to include a detailed set of observable firm characteristics, there would inevitably be some inherently unobservable factors (such as managerial talent or the degree of congeniality and cooperation among the workers) that would influence both firm performance and the tendency of the firm to engage in team production. The consequence is that in all three regressions the estimated teams treatment effect is biased, yielding misleading answers to the question of whether firms benefit from team production and to the question of whether large firms experience different benefits from teams than do small firms. To address this endogeneity problem, one can specify an additional equation that determines TEAMS. Letting TEAMS* denote a continuous latent index that can be thought of as the firm’s propensity to engage in team production, consider the following Model 4: Y ¼ b0 þ b1 X þ 1

ifTEAMS ¼ 1

¼ b2 þ b 3 X þ 0 TEAMSn ¼ a0 þ a1 Z þ 2

ifTEAMS ¼ 0

TEAMS ¼ 1 ¼0

if TEAMSn 40 if TEAMSn 0

ð4Þ

Assuming multivariate normality of the disturbances, estimation of this model yields consistent estimates of the teams’ treatment effect of interest.

8

JED DEVARO AND FIDAN ANA KURTULUS

Whereas in Model 3 we incorrectly imposed (implicitly) the assumptions covð1 ; 2 Þ ¼ 0 and covð0 ; 2 Þ ¼ 0; in Model 4 these covariances are unrestricted parameters to be estimated. The structural models we estimate in this paper are only slightly more complicated than Model 4, the two main differences being that our measures of Y are discrete rather than continuous and that we introduce autonomy into the model and treat it as endogenous in addition to teams. These models were proposed and estimated in a pair of recent related papers (DeVaro, 2004a, b). The first of these considered financial performance (interpreted as profit) as the measure of organizational performance, and the second considered labor productivity and product quality. Both papers addressed the question of whether the team production affects organizational performance but did not consider how these effects differ in different types of firms. In the present paper, we consider all the three measures of organizational performance and ask how the predicted benefits of teams vary with organizational characteristics. We now turn to a discussion of the theoretical background underlying teams research and previous evidence on what types of organizations are most likely to benefit from team structures.

2. PREVIOUS RESEARCH ON FIRM CHARACTERISTICS AND THE BENEFITS OF TEAMS TO ORGANIZATIONAL PERFORMANCE Theoretical models in the economics literature on the effects of team production on organizational performance involve a comparison of the benefits and costs to team production (Alchian & Demsetz, 1972). Some of the main benefits of team production accrue through productive information sharing among workers, when potential team members have knowledge that is nonduplicative and also relevant to the production process (Lazear, 1995, 1998). The potential costs of team production include costs associated with regular team meetings and training, and shirking and free-riding among team members (Alchian & Demsetz, 1972; Holmstrom, 1982; Rasmusen, 1987; Itoh, 1991, 1992; McAfee & McMillan, 1991; Legros & Matthews, 1993). Kandel and Lazear (1992) argue that teams alleviate costly monitoring of workers in the presence of asymmetric information by relying on monitoring of workers through peer pressure. In empirical work, team structures are sometimes the central focus of the analysis (e.g., DeVaro, 2004a,b; Hamilton et al., 2003; Boning, Ichniowski, &

What Types of Organizations Benefit from Team Production

9

Shaw, 2003) and are sometimes one of a number of high performance practices that are analyzed together (e.g., Eriksson, 2003; Black & Lynch, 2001; Ichniowski, Shaw, & Prennushi, 1997). Frequently the teams under study are self-managed or autonomous, reflecting the fact that such teams are used more commonly than closely managed or non-autonomous teams.3 A common finding in the literature is that autonomous teams have positive effects on firm performance. For example, Eriksson (2003) finds a positive effect of self-managed teams on labor productivity in a cross section of establishments, and Hamilton et al. (2003) finds a 14% increase in labor productivity after the introduction of self-managed teams in a garment manufacturing plant. 2.1. What Types of Organizations Benefit from Team Production? While much has been written on the subject of whether and how teams confer benefits to organizational performance, less attention has been devoted to the question of what types of organizations benefit from teams. Our analysis addresses a large number of organizational characteristics, but the following four are of particular interest: union membership, firm size, financial participation and employee ownership, and industry. We now discuss what the previous literature has to say about the relationship between teams and each of these in turn. 2.2. Union Membership While some argue that the presence of unions at the workplace constrains the ability of management to redesign jobs to incorporate new work systems such as teams, others claim that unions can promote the introduction and continued existence of such systems by facilitating increased dialogue between workers and management. A meta-analysis conducted by Doucouliagos and Laroche (2003) concluded that unions have a negative impact overall on labor productivity in the United Kingdom.4 Empirical evidence in the previous literature on the relation between the presence of unions and teams has been mixed. Osterman (1994) found that the presence of unions in an organization is not an important determinant of the adoption of highperformance workplace practices, including self-managed teams, whereas McNabb and Whitfield (1997) found that recognized unions facilitate the introduction of teamwork in establishments. Apart from this issue of adoption of teams, the question of whether the benefits to firm performance from using teams vary with the presence of

10

JED DEVARO AND FIDAN ANA KURTULUS

unions has been addressed by Black and Lynch (2004). That study found that when union membership and self-managed teams are interacted in a cross section regression, the effect on labor productivity is negative, suggesting that unionized establishments that use self-managed teams tend to have lower labor productivity. When union membership and self-managed teams are interacted in a fixed effects regression, the effect on labor productivity is positive. However, neither of these results was statistically significant. McNabb and Whitfield (1997) used cross sectional data from the third wave of the Worker Industrial Relations Survey (the wave previous to the one used in the current paper) to show that the joint effect of union presence and teamwork on relative financial performance is positive. Our findings are closer to Black and Lynch’s cross-sectional results in that ours suggest negative relationships between financial performance and both the presence of unions in the establishment and the number of recognized unions in the establishment. 2.3. Firm Size To our knowledge, there has been no work done on the relationship between firm size and the predicted benefits from team production, though some attention has been devoted to the question of how the adoption of team production varies with firm size. For example, Osterman (1994) used data on 694 manufacturing establishments in the U.S. to find that smaller establishments are more likely to adopt innovative work practices including teams. McNabb and Whitfield (1997), on the other hand, found that the propensity to adopt teamwork programs is positively associated with firm size and being part of a large organization. The question of whether the benefits from team production vary by firm size is of interest if for no other reason than the fact that most labor market measures vary with firm size. One story that could give rise to a firm-size effect concerns the relative costs of alternative means of monitoring workers. It might be that in smaller organizations, internal monitoring by peers in the context of teams is more effective than in larger organizations. An alternative story that would suggest the opposite result is that larger organizations face greater problems of coordination and information sharing than do smaller organizations, so that the benefits from teams would be increasing in the scale of the organization. 2.4. Financial Participation and Employee Ownership In our study, worker participation in firm decision-making is considered at the group level and is reflected in the granting of autonomy to team members

What Types of Organizations Benefit from Team Production

11

(giving team members the latitude to jointly decide how the work is to be done). Since over half of the teams in our sample are autonomous by this definition, the decision to organize production in teams very often coincides with a decision to increase the degree of worker participation in firm decision making at the group level. Recent research has shown that such employee involvement in decision making is an effective means of enhancing firm performance when implemented along with employee ownership schemes such as profit sharing and share ownership (Kruse et al., 2004; Kruse, 2002; Freeman & Dube, 2000; Kruse & Blasi, 1997; Ichniowski et al., 1997; Ben-Ner & Jones, 1995). Eriksson (2003) argued that new work practices and new pay practices such as teams bonus, individual bonus, stock or stock options, and profit sharing are complementary in the sense that, if a firm that has adopted new work practices introduces performance-related pay schemes, this enhances productivity further, though this complementarity is found to be more on the level of individual compensation schemes, rather than group-based pay incentives. Ben-Ner and Jones (1995) argued that both ownership without participation, and participation without ownership, can actually decrease firm performance by increasing worker–firm conflict. Adams (2003) found the existence of complementarities between the use of profit sharing and the delegation of decision-making power on an individual basis to production line workers. Other studies have argued that financial participation is conducive to aligning the goals of the employees with the goals of the firm by directly linking the workers’ pay to firm performance. However, the goal-alignment process needs to be supported both by financial participation and employee ownership, or what is called ‘‘direct participation,’’ and employee involvement in decision making at all levels of the firm hierarchy, or what is called ‘‘indirect participation’’ (Kato & Morishima, 2002). Kruse (2002) has explained the intuition as follows: ‘‘Employee ownership may improve firm performance by decreasing labor-management conflict and serving as a collective incentive to improve workplace cooperation, information-sharing, and organizational citizenship behavior. This may be limited by the freerider problem when rewards are shared with co-workers, direct incentives for better work becomes weak as the number of coworkers expands. To counteract this problem and encourage higher performance, firms may combine employee ownership with employee participation in decision-making and other human resource policies to encourage a sense of ownership, draw more fully on worker skills and information, and create company spirit and work norms.’’ (p. 71). In this paper, we present evidence corroborating this complementarity hypothesis; our analysis shows that establishments whose

12

JED DEVARO AND FIDAN ANA KURTULUS

employees engage in higher levels of financial participation are also more likely to benefit from team production through higher financial performance.

2.5. Industry According to the cross sectional analysis of McNabb and Whitfield (1997), the presence of teamwork programs is most common in the wholesale and retail sectors, and least likely in production industries and non-metal manufacture, whereas establishments in banking, insurance, and finance are generally more likely to adopt teams. Empirical statements about which industries experience the largest predicted benefits from teams and other workplace practices must be made primarily on the basis of comparisons of different case studies or different industry-specific data sets. The vast majority of past studies have focused on the manufacturing industry. Comparatively less is known about the effects of teams on firm performance in other industries, though there are some results. In services, an example is Batt (1999), which found that the use of self-managed teams among customer service and sales workers yields a statistically significant improvement in self-reported service quality and sales per employee (the measure of labor productivity in sales occupations). When combined with new technology usage, teams boost sales by an even greater magnitude. Batt and Appelbaum (1995) looked at two industries, namely telecommunications (where they considered customer service staff as well as ‘‘network crafts’’ occupations such as installation and repair crews) and apparel manufacturing (where they considered sewing machine operators) and found that teams have a significant and positive impact on workers’ perceptions of the quality of work done by their work groups.5 Batt’s (2001) empirical results suggest that there are no significant differences in labor productivity and service quality when field technicians work in teams rather than independently. In manufacturing, the effect of teams and other high-performance work practices has generally been found to be positive. Hamilton et al. (2003) found a 14% increase in labor productivity after a switch to self-managed teams in a garment-manufacturing plant in Northern California. Also in the apparel manufacturing industry, Berg, Appelbaum, Bailey, and Kalleberg (1996) found that team production improved such outcomes as quality, costs, and responsiveness to retailers via better coordination among team members as a result of their ability to self-regulate work, eliminate bottlenecks, resolve conflicts, help one another to solve problems, and make improvements to the production process.

What Types of Organizations Benefit from Team Production

13

Boning et al. (2003) used data from the steel-manufacturing sector to find that production lines that adopt problem-solving teams experience large gains in productivity. However, this is true only for production lines that undertake complex production processes and products – in less complex environments, there is no benefit to using teams. Ichniowski et al. (1997) studied the effects of high-performance workplace systems in a specific production process in steel manufacturing, namely steel finishing lines. They considered teams (specifically the existence and prevalence of formal work teams for the purposes of problem-solving activities, and worker membership in multiple problem-solving teams) as one of a number of human resource management (HRM) practices used, though their focus was on HRM systems more generally rather than teams specifically. Their results indicate that finishing lines, which utilize a set of innovative work practices have higher levels of worker productivity than lines that use more traditional practices and that there exist complementarities between certain high-performance practices. Black and Lynch (2001) found that U.S. firms in manufacturing that use high-performance workplace practices, such as regular group meetings, benchmarking, self-managed teams and profit sharing have higher productivity and wages than other firms. Black and Lynch (2004) suggested that the adoption of such innovations was an important factor contributing to the jump in multifactor productivity of the U.S. economy in the second half of the 1990s. On the other hand, Cappelli and Neumark (2001) studied the effects of highperformance work practices including benchmarking, regularly scheduled group meetings to discuss work-related problems, job rotation, self-managed teams, pay for skill and profit sharing in a panel of manufacturing firms. Their results show that these practices may raise productivity, though with little statistical significance, and that the effects on overall labor efficiency are small.

3. DATA: WORKPLACE EMPLOYEE RELATIONS SURVEY (WERS) 1998 The data are from the management questionnaire in the 1998 wave of the British Workplace Employee Relations Survey (WERS) (Department of Trade and Industry and Advisory, 2001), jointly sponsored by the Department of Trade and Industry, ACAS, the Economic and Social Research Council, and the Policy Studies Institute. Distributed via the UK Data Archive, the WERS data are a nationally representative stratified random sample covering British workplaces with at least 10 employees except for those in the following 1992 Standard Industrial Classification (SIC) divisions: agriculture,

14

JED DEVARO AND FIDAN ANA KURTULUS

hunting, and forestry; fishing; mining and quarrying; private households with employed persons; and extra-territorial organizations. Some of the 3,192 workplaces targeted were found to be out of scope, and the final sample size of 2,191 implies a net response rate of 80.4% (Cully, Woodland, O’Reilly, & Dix, 1999) after excluding out-of-scope cases. Data were collected between October 1997 and June 1998 via face-to-face interviews, and the respondent was usually the most senior manager at the workplace with responsibility for employment relations. Table 1 displays the industry composition of the sample, using 12 industry categories in the 1992 SIC.

Table 1.

Distribution of Workplaces by Industry and Largest Occupational Group. Number of Establishments

Percent of Total

299 80 112 322 127 136 101 227 183 244 249 111

13.7 3.7 5.1 14.7 5.8 6.2 4.6 10.4 8.4 11.1 11.4 5.1

Distribution by industry Manufacturing Electricity, gas, and water Construction Wholesale and retail Hotels and restaurants Transport and communication Financial services Other business services Public administration Education Health Other community services Total

2191

100

Distribution by largest occupational group at workplace Managers and administrators Professional occupations Associate professional and technical occupations Clerical & secretarial occupations Craft & related occupations Personal and protective service occupations Sales occupations Plant and machine operatives Other occupations Total

15 309 180 390 231 314 237 278 237 2191

0.7 14.1 8.2 17.8 10.5 14.3 10.8 12.7 10.8 100

15

What Types of Organizations Benefit from Team Production

Our measures of organizational performance are discrete responses to three survey questions concerning the establishment’s current financial performance, labor productivity, and quality of product or service, relative to other establishments in the same industry. Responses include: ‘‘A lot better than average’’, ‘‘Better than average’’, ‘‘About average for industry’’, ‘‘Below average’’, ‘‘A lot below average’’, and ‘‘No comparison possible’’. Few establishments report below-average performance for any of the three measures. While this might indicate reporting error in the dependent variables, such errors need not have consequences for our analysis unless a respondent’s likelihood of overstating performance is systematically related to the choices of teams and autonomy.6 Furthermore, an establishment’s inclusion in the sample is conditional on its being operational, and length-biased sampling arises when operational establishments are sampled at a point in time. High-performing establishments have long durations of operation and are more likely to be sampled than low-performing establishments with low durations of operation. The pronounced asymmetry in reported performance that is observed in the data is therefore not surprising.7 The 1998 wave of the WERS includes a follow-up question asking respondents their interpretation of the term ‘‘financial performance’’. The frequency of responses is as follows:8 Interpretation of ‘‘financial performance’’

Number of Firms

% of Firms

Profit or value added Sales, fees, budget Costs or expenditure Stock market indicators (e.g., share price) Other specific answer

952 374 389 54 31

52.9 20.7 21.6 3.0 1.7

1,800

100.0

Total

Studies using earlier waves of these data did not have access to this follow-up question and were forced to pool disparate interpretations of the dependent variable in the same analysis (Machin & Stewart, 1990, 1996; McNabb & Whitfield, 1997). As shown in DeVaro (2004a), the estimated effect of teams on financial performance is very sensitive to the interpretation of financial performance. When estimating the model for financial performance, we use only those establishments interpreting financial performance as profit or value added.

16

JED DEVARO AND FIDAN ANA KURTULUS

A skeptic might argue that the three subjective measures of organizational performance are measuring essentially the same thing. If that were the case, we would expect extremely high correlations among the three measures. This is not so, as revealed by the following correlation matrix: Correlation Matrix for the Measures of Organizational Performance.

Financial performance Labor productivity Product quality

Financial Performance

Labor Productivity

Product Quality

1.000 0.507 0.328

1.000 0.352

1.000

The correlations are all positive and statistically significant at the 1% level, but their average value is only 0.40. For each establishment, the data contain information about the proportion of employees in the largest occupational group that works in formally designated teams. Responses are in the following discrete categories: ‘‘All 100%’’, ‘‘Almost all 80–99%’’, ‘‘Most 60–79%’’, ‘‘Around half 40–59%’’, ‘‘Some 20–39%’’, ‘‘Just a few 1–19%’’, ‘‘None 0%’’. An advantage of this survey question is that it specifically refers to ‘‘formally designated’’ teams. This precise wording of the question directs the respondent’s attention to situations of true joint production and should reduce the respondent’s likelihood of reporting the use of teamwork simply on the basis of a cooperative or friendly atmosphere of ‘‘team spirit’’ at the workplace. A drawback of the survey question is that it is restricted to the largest occupational group at the establishment. The sample may contain establishments in which team production is heavily used in occupational groups other than the largest, yet the response to this question might be ‘‘None 0%’’. This measurement issue is one limitation of the study. The survey also contains a measure of team autonomy that corresponds well to the notions of autonomy discussed in the theoretical literatures of economics and organizational behavior (Aghion & Tirole, 1997; Hackman, 1987).9 For establishments that report the use of formally designated teams in the largest occupational group, the respondent manager is asked to respond ‘‘Yes’’ or ‘‘No’’ to the following statement: ‘‘Team members jointly decide how the work is to be done.’’ Since both the team and autonomy variables are defined in terms of the establishment’s largest occupational group, we provide the distribution of the sample by largest occupational group in the lower panel of Table 1. All observations in the WERS are coded according to the UK Office of Population, Censuses and Surveys

What Types of Organizations Benefit from Team Production

17

Standard Occupational Classification (SOC) codes, and we aggregated these to produce nine one-digit categories.

4. METHODOLOGY There are three stages to our methodology. First, we estimate structural models for each of our three measures of organizational performance (financial performance, labor productivity, and product quality). Second, we use the estimated parameters from the first stage to compute for each establishment in the sample a ‘‘predicted benefit to organizational performance from using team production’’. These first two stages closely follow DeVaro (2004a,b), so our treatment here is brief, and we relegate the technical details to the appendix. Third, we ask how the ‘‘predicted benefit’’ computed in our second stage varies as a function of the firm characteristics in the model. This sheds light on what types of firms are likely to benefit from team production and in what areas (financial performance, labor productivity, and product quality) they are likely to benefit. 4.1. Stage 1: Estimating Structural Models of Teams, Autonomy, and Organizational Performance The model has the same structure for each measure of organizational performance. Since the three endogenous variables (organizational performance, teams, and autonomy) in each model are observed as discrete responses, the structural model specifies probabilities for all possible outcomes. To make the analysis tractable, some aggregation is needed to reduce the number of outcomes. Since relatively few respondents report below-average performance, we aggregate the lowest three categories for each of the organizational performance measures as follows: 1 ¼ ‘‘About average for industry’’ or below; 2 ¼ ‘‘Better than average’’; 3 ¼ ‘‘A lot better than average’’. Furthermore, we consider only whether team production is used or not in the largest occupational group, rather than focusing on the fraction of that group that engages in team production. That is, we aggregate the teams variable as follows: TEAMS

¼ ¼

1 if positive fraction of workers in the largest occupational group is in teams 0 otherwise

The sequence of the model is as follows. First, the establishment decides whether or not to use teams (TEAMS ¼ 0 or 1). Given that teams are

18

JED DEVARO AND FIDAN ANA KURTULUS

chosen, the establishment then decides whether to grant the teams autonomy (AUTO ¼ 0 or 1). Finally, these choices of teams and autonomy affect organizational performance. Let Yi* denote a latent indicator of organizational performance for the ith establishment, relative to the industry average, and let Yi denote its ordered, discrete realization, taking values of 1, 2, or 3. The four-equation structural model has the following form: Yin ¼ aAUTOi þ X 1i d1 þ 1i if TEAMSi ¼ 1 ¼ X 1i d2 þ 0i if TEAMSi ¼ 0 TEAMSi n ¼ X 2i b þ 2i AUTOi n ¼ X 3i c þ 3i if TEAMSi ¼ 1 Y i ¼ 1 ifY ni o0 ¼2 ¼3

if0 Y ni oc where c40 ifY ni c

TEAMSi ¼ 1 if TEAMSi n40 ¼ 0 otherwise AUTOi ¼ 1 if AUTOi n40 and TEAMSi n40 ¼ 0 if AUTOi n 0 and TEAMSi n40 We assume multivariate normality of the disturbances, ð0 ; 1 ; 2 ; 3 Þ Nð0; RÞ; and estimate the four equations jointly by maximum likelihood. The vector of parameters to be estimated, h, includes a, d1, d2, c, b, c, s02, s12, s13, and s23, where the notation sij means cov(ei, ej). By treating organizational performance as a switching regression, we allow for a full set of interactions between teams and all observed and unobserved determinants of performance. Autonomy, on the other hand, enters only as a dummy-endogenous variable on the right-hand-side of the organizational performance equation when teams are used. In principle, it would be possible to allow for a full set of interactions of autonomy with the determinants of performance, but that analysis would not be feasible with our sample size.10 4.2. Stage 2: Constructing Predicted Benefits of Teams to Organizational Performance The estimation results from Stage 1 allow us to construct measures of the ‘‘predicted benefits from using teams’’ for each establishment in the sample. Recall that the measures of organizational performance assume values of

What Types of Organizations Benefit from Team Production

19

1, 2, or 3 according to whether the establishment’s recent performance relative to that of others in the industry is average or below, better than average, or a lot better than average. The effect of teams on organizational performance is the change in the probabilities that performance is in each of these three categories when team production is used in the largest occupational group compared to when it is not used. Letting Yi denote the discrete measure of organizational performance (taking values 1 ¼ ‘‘average or below’’, 2 ¼ ‘‘above average’’, 3 ¼ ‘‘a lot above average’’), the following three measures give the effect of team production on performance for establishment i:11 ðEffectA1Þi ¼ ProbðY i ¼ 1jTEAMSi ¼ 1Þ

ProbðY i ¼ 1jTEAMSi ¼ 0Þ

ðEffectA2Þi ¼ ProbðY i ¼ 2jTEAMSi ¼ 1Þ ðEffectA3Þi ¼ ProbðY i ¼ 3jTEAMSi ¼ 1Þ

ProbðY i ¼ 2jTEAMSi ¼ 0Þ ProbðY i ¼ 3jTEAMSi ¼ 0Þ

Since for an individual establishment these three effects must sum to 0, any two of them contain all of the information about the effect of team production on that establishment’s performance. We therefore focus on Effects A1 and A3. 4.3. Stage 3: How Do the Predicted Benefits from Teams Vary with Firm Characteristics? We next ask how the predicted benefit from using team production that we compute in Stage 2 varies as a function of the covariates in the model. That is, we are interested in seeing how the functions EffectA3 and EffectA1 vary with changes in a particular firm characteristic, holding the other firm characteristics constant. We do this computation slightly differently according to whether the covariate in question is a single dummy variable, one dummy variable from a group of related dummies, or a ‘‘continuous’’ variable.12 When the covariate of interest is a dummy variable, X, we compute ðEffectA3jX ¼ 1Þ ðEffectA3jX ¼ 0Þ for each workplace, evaluating each of the other covariates at their actual establishment values. Then we take the average of these differences across all workplaces in the sample to obtain a summary measure of how EffectA3 varies with changes in X. We follow the analogous approach for EffectA1. If the particular X is one dummy in a group of related dummies the computation is the same as the one just described, except that the other dummies in the same group are evaluated at 0 rather than at their observed values for each workplace. The three groups of dummies that are considered in this way are the industry, occupation, and ownership variables. The reference group for each of these categories is the

20

JED DEVARO AND FIDAN ANA KURTULUS

wholesale and retail industry, clerical and secretarial occupations, and public sector workplaces. Finally, when the covariate of interest is ‘‘continuous’’, the differences we compute are not between X ¼ 1 and X ¼ 0 but rather between the 0.75 and 0.25 quantile of X, again evaluating all of the other covariates at their individual values for each workplace. More precisely, consider the following definitions: xji xkaj;i xkaj;i

xqj

covariate j for establishment i vector of covariates for establishment i excluding covariate xji vector of covariates for establishment i excluding covariate xji when xji is one dummy variable in a multiple-dummy group, setting to 0 all other dummies in the multiple-dummy group to which xji belongs quantile q of covariate xj

Then the following expressions illustrate how EffectsA3 and A1 change, on average, with a particular covariate xj : Case 1: xj is binary but not part of a multiple-dummy group X 1 N EffectA3 xji ¼ 1; xkaj;i DEffectA3 ¼ N i¼1

EffectA3 xji ¼ 0; xkaj;i

X 1 N EffectA1 xji ¼ 1; xkaj;i N i¼1

EffectA1 xji ¼ 0; xkaj;i

DEffectA1 ¼

Case 2: xj is binary and part of a multiple-dummy group (i.e., industry, occupation, ownership) X h i 1 N EffectA3 xji ¼ 1; xkaj;i N i¼1

h i EffectA3 xji ¼ 0; xkaj;i

X h i 1 N EffectA1 xji ¼ 1; xkaj;i DEffectA1 ¼ N i¼1

h i EffectA1 xji ¼ 0; xkaj;i

DEffectA3 ¼

What Types of Organizations Benefit from Team Production

21

Case 3: xj is continuous DEffectA3 ¼

X h i 1 N EffectA3 xji ¼ x0:75 ; xkaj;i j N i¼1

h i EffectA3 xji ¼ x0:25 ; xkaj;i j

DEffectA1 ¼

X h i 1 N EffectA1 xji ¼ x0:75 ; xkaj;i j N i¼1

h i EffectA1 xji ¼ x0:25 ; xkaj;i j

The preceding definitions refer to the quantiles 0.25 and 0.75. Note, however, that the estimation samples differ for the three measures of organizational performance (N ¼ 889 for financial performance, N ¼ 1; 660 for labor productivity, N ¼ 1; 839 for product quality) and therefore the quantiles may differ as well. Since most of the covariates are binary, however, the quantiles of interest are frequently identical across the three estimation samples. The only exceptions are establishment size, number of part time workers, financial participation, multi-skilling, and number of recognized unions, and for these variables the quantiles differ only slightly across the three samples. For consistency, we define all quantiles using the largest sample of the three, namely the product quality sample. We also tried estimating all three models on the smallest sample of 889 observations, so that the quantiles exactly coincided for each model, and the results were similar to those we report here. Our preference, however, is to report results based on the largest estimation sample possible for each of the three measures of organizational performance.

5. EMPIRICAL RESULTS Table 2 displays means and standard deviations of the variables in the model. We present the parameter estimates from the structural models (Stage 1 of our three-stage methodology outlined in the previous section) in Appendix Tables A1–A3. Since the structural models are non-linear, these parameter estimates lack straightforward interpretations. For our purposes, they are useful mainly for computing EffectsA3 and A1 in Stage 2 for each of the three measures of organizational performance. The difference between EffectA3 and EffectA1 provides a univariate index of the predicted benefits of team production on organizational performance and facilitates a simple presentation of the distributions of these benefits.13 Figs. 1–3 display kernel density estimates of (EffectA3 EffectA1) for financial performance, labor productivity, and product quality. These distributions have a slight

22

JED DEVARO AND FIDAN ANA KURTULUS

Table 2.

Means and Standard Deviations. Mean

Dependent variables Financial performance Labor productivity Product quality Teams Autonomy

Standard Deviation

1.801 1.632 1.942 0.870 0.554

0.728 0.681 0.701 0.337 0.497

0.217 294.308 0.258 0.380 0.439 0.241 0.659 0.633 0.096 0.104 0.890 3.000 1.538 0.844 3.941 0.296 2.729 2.310 3.742 5.369

0.412 857.423 0.280 0.486 0.496 0.428 0.474 0.482 0.294 0.305 0.313 1.852 2.006 0.363 2.040 0.457 1.086 0.846 0.980 0.930

Firm ownership Private sector franchise Private sector non-franchise Alternative private sector franchise Alternative private Sector non-franchise Public sector

0.011 0.372 0.011 0.300 0.309

0.106 0.483 0.104 0.458 0.462

Industry Manufacturing Electricity, gas, and water Construction Wholesale and retail Hotels and restaurants Transport and communication Financial services Other business services

0.136 0.037 0.051 0.147 0.058 0.062 0.046 0.104

0.343 0.188 0.220 0.354 0.234 0.241 0.210 0.305

General firm characteristics Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Just-in-time production Information Incentive alignment Decisions Work at home

23

What Types of Organizations Benefit from Team Production

Table 2. (Continued ) Mean Public administration Education Health Other community services Largest occupational group at workplace Managers and administrators Professional occupations Associate professional and technical occupations Clerical and secretarial occupations Craft and related occupations Personal and protective service occupations Sales occupations Plant and machine operatives Other occupations

Standard Deviation

0.084 0.111 0.114 0.051

0.277 0.315 0.317 0.219

0.007 0.141 0.082 0.178 0.105 0.143 0.108 0.127 0.108

0.082 0.348 0.275 0.383 0.307 0.350 0.311 0.333 0.311

Note: Statistics for financial performance are computed using the subsample of 952 establishments for which financial performance is interpreted to mean profit or value added. For all other variables, statistics are based on the full sample.

negative skew for financial performance and labor productivity and a slight positive skew for product quality. As the graphs illustrate, all three distributions peak at a positive number, indicating that the typical establishment is predicted to benefit from team production through higher labor productivity, product quality, and financial performance.14 Our main results are from Stage 3 of our analysis, and these are displayed in Tables 3–5 for financial performance, labor productivity, and product quality. The first and second columns of each table illustrate how EffectsA3 and A1 vary with a particular firm characteristic, holding the other characteristics constant. The third column displays the difference between the first two columns, thus answering the question of how (EffectA32EffectA1Þ varies with a particular covariate, holding the others constant. Table 6 summarizes the qualitative results from Tables 3–5 by listing firm characteristics along with the sign of their predicted effect on organizational performance (i.e., the sign of DðEffectA3 EffectA1Þ). Before commenting on the results, we note that a limitation of our analysis is that many of the parameters underlying the predicted benefits of teams are estimated with low precision. As a result, the confidence bands associated with the differences we report in Tables 3–5 would be fairly wide. Our results should therefore be viewed as suggestive, and definitive statements will require corroboration in future work with new data sets.

24

JED DEVARO AND FIDAN ANA KURTULUS

Density

1

0.5

0 -2

-1

0 Effect A3 – Effect A1

1

2

Fig. 1. (EffectA3 EffectA1) for the Financial Performance Model. Note: (EffectA3 EffectA1) is a Measure of the Degree to which a Firm Benefits in Terms of Its Financial Performance from Team Production. EffectA3 is the Effect of Team Production on the Probability that a Firm’s Financial Performance is a Lot Above Industry Average, and EffectA1 is the Effect of Team Production on the Probability that a Firm’s Financial Performance is Average or Below. The Bandwidth of 0.083 was Chosen to Minimize the Mean-integrated Squared Error if the Data were Gaussian and a Gaussian Kernel were Used.

The results concerning unions, firm size, industry, and financial participation and employee ownership are of particular interest. As seen in Table 6, workplaces at which some employees belong to a union are predicted to lose in all three dimensions from using teams. Having some unionized workers is associated with a decrease of more than 7 percentage points in the probability that labor productivity is a lot above the industry average and more than a 5 percentage point decrease in the probability that product quality is a lot above the industry average. These changes are accompanied by increases in the probability that performance is at or below the industry average of more than 15 percentage points for labor productivity and nearly 7 percentage points for product quality. The predicted benefit of teams to financial performance is also lower in the presence of unionized workers. While the probability that financial performance is a lot above the industry average is actually slightly higher in unionized settings, the probability that financial performance is average or below also increases (and by a greater amount).

25

What Types of Organizations Benefit from Team Production 2

Density

1.5

1

0.5

0 -1

-0.5

0 Effect A3 – Effect A1

0.5

1

Fig. 2. (EffectA3 EffectA1) for the Labor Productivity Model. Note: (EffectA3 EffectA1) is a Measure of the Degree to which a Firm Benefits in Terms of its Labor Productivity from Team Production. EffectA3 is the Effect of Team Production on the Probability that a Firm’s Labor Productivity is a Lot Above Industry Average, and EffectA1 is the Effect of Team Production on the Probability that a Firm’s Labor Productivity is Average or Below. The Bandwidth of 0.043 was Chosen to Minimize the Mean-integrated Squared error if the Data were Gaussian and a Gaussian Kernel were Used.

While a greater degree of union recognition (measured by the total number of recognized unions at the workplace) is associated with benefits to labor productivity and product quality from using teams, the reverse is true for financial performance. This result highlights the type of misleading inference that might be drawn by focusing solely on outcome measures, such as labor productivity and product quality (both of which are common in the empirical teams literature) rather than on broader measures such as financial performance that are more inclusive of various benefits and costs. While labor productivity and product quality might indeed be helped by teams in settings with high union recognition, it might be that the costs of organizing team members are prohibitively large and overshadow the benefits. These results concerning unions, along with the finding from Black and Lynch (2004) that the interaction of union membership and self-managed

26

JED DEVARO AND FIDAN ANA KURTULUS

2

Density

1.5

1

0.5

0 -0.5

0

0.5 Effect A3 – Effect A1

1

Fig. 3. (EffectA3 EffectA1) for the Product Quality Model. Note: (EffectA3 EffectA1) is a Measure of the Degree to which a Firm Benefits in Terms of its Product Quality from Team Production. EffectA3 is the Effect of Team Production on the Probability that a Firm’s Product Quality is a Lot Above Industry Average, and EffectA1 is the Effect of Team Production on the Probability that a Firm’s Product Quality is Average or Below. The Bandwidth of 0.042 was Chosen to Minimize the Mean-integrated Squared Error if the data were Gaussian and a Gaussian Kernel were Used.

teams was insignificant in regressions of labor productivity, cast doubt on the hypothesis that team production is more likely to enhance firm performance in unionized settings. On the other hand, McNabb and Whitfield (1997) found that the joint effect of union presence and teamwork on relative financial performance is positive in their analysis of the 1990 Workplace Industrial Relations Survey (WIRS), the predecessor of WERS. While the McNabb and Whitfield result is counter to ours, there are some important differences between the two studies that might account for the discrepancy. First, the studies differ methodologically in that McNabb and Whitfield estimate a binomial logit for financial performance whereas we estimate a structural ordered probit in which teams and autonomy are endogenous. Second, since the questions on teams that we use are new to the 1998 WERS, McNabb and Whitfield used a different measure of teams. Third, we estimate our financial performance model only on those establishments for

27

What Types of Organizations Benefit from Team Production

Table 3.

How the Predicted Benefit of Teams on Financial Performance Varies with Firm Characteristics. WEffectA3

Basic firm characteristics Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers Under one year Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Just-in-time production Information Incentive alignment Decisions Work at home Firm Ownership Private sector franchise Private sector nonfranchise Alternative private sector franchise Alternative private sector non-franchise Industry Manufacturing Electricity, gas, and water Construction Hotels and restaurants Transport and communication

WEffectA1

W(EffectA3

EffectA1)

0.036 0.630 0.066

0.073 0.494 0.126

0.108 1.124 0.192

0.009 0.054

0.030 0.118

0.039 0.172

0.069

0.149

0.217

0.007 0.053 0.085 0.004 0.053 0.029 0.077

0.017 0.068 0.148 0.061 0.126 0.050 0.114

0.010 0.121 0.233 0.065 0.179 0.080 0.191

0.064 0.014 0.049 0.029 0.040 0.004 0.010

0.127 0.017 0.105 0.038 0.059 0.005 0.014

0.191 0.032 0.154 0.067 0.098 0.009 0.024

0.136 0.025

0.446 0.013

0.582 0.012

0.171

0.334

0.505

0.134

0.285

0.420

0.038 0.128 0.087 0.079 0.006

0.054 0.311 0.185 0.124 0.024

0.091 0.439 0.272 0.203 0.030

28

JED DEVARO AND FIDAN ANA KURTULUS

Table 3. (Continued ) WEffectA3 Financial services Other business services Public administration Education Health Other community services

0.032 0.023 0.051 0.114 0.028 0.128

Largest occupational group at workplace Managers and 0.037 administrators Professional occupations 0.086 Associate professional and 0.069 technical occupations Craft and related 0.007 pccupations Personal and protective 0.016 service occupations Sales occupations 0.018 Plant and machine 0.058 operatives Other occupations 0.022

WEffectA1

W(EffectA3

EffectA1)

0.061 0.047 0.088 0.126 0.031 0.162

0.093 0.070 0.139 0.240 0.060 0.290

0.060

0.096

0.188 0.273

0.274 0.342

0.024

0.031

0.037

0.052

0.006 0.102

0.024 0.160

0.040

0.062

Note: The omitted categories for firm ownership, industry, and largest occupational group are, respectively, public sector, wholesale and retail, and clerical and secretarial. (EffectA3)i for establishment i is defined as ðEffectA3Þi ¼ ProbðY i ¼ 3jTEAMS i ¼ 1Þ2ProbðY i ¼ 3jTEAMS i ¼ 0Þ; which is the effect of team production on the probability that establishment i’s financial performance is a lot above industry average. Cell entries under ‘‘WEffect A3’’ indicate how this effect varies, on average, by the firm characteristic indicated by the row title. This is computed slightly differently according to whether the characteristic is a single dummy variable, one dummy variable from a group of related dummies, or a continuous variable. When characteristic X is a single dummy variable, we computed for each establishment ðEffectA3jX i ¼ 1Þ ðEffectA3jX i ¼ 0Þ holding all of the other characteristics at their actual establishment levels. Then we took the average of these differences across all establishments in the sample. When the X in question is one dummy in a group of related dummies, the computation was the same as the one just described, except that the other dummies in the same group were evaluated at 0 rather than at their observed values for each establishment. The three groups of dummies that are considered in this way are the industry, occupation, and ownership variables. When the X in question is continuous, the differences we computed were not between X ¼ 1 and X ¼ 0 but rather between the 0.75 and the 0.25 quantile of X, again holding all the characteristics of the establishment at their individual values. (EffectA1)i for establishment i is defined as ðEffectA1Þi ¼ ProbðY i ¼ 1jTEAMSi ¼ 1Þ2ProbðY i ¼ 1jTEAMSi ¼ 0Þ; which is the effect of team production on the probability that establishment i’s financial performance is average or below. The ‘‘WEffectA1’’ column was computed analogously. The ‘‘4ðEffectA32EffectA1Þ’’ column is simply the difference between the first and second columns and is our measure of the degree to which the predicted benefit from teams to financial performance varies with the firm characteristic indicated by the row title. N ¼ 889:

29

What Types of Organizations Benefit from Team Production

Table 4.

How the Predicted Benefit of Teams on Labor Productivity Varies with Firm Characteristics. WEffectA3

Basic firm characteristics Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Just-in-time production Information Incentive alignment Decisions Work at home Firm ownership Private sector franchise Private sector nonfranchise Alternative private sector franchise Alternative private sector non-franchise Industry Manufacturing Electricity, gas, and water Construction Hotels and restaurants Transport and communication

WEffectA1

W(EffectA3 – EffectA1)

0.006 0.346 0.012

0.013 0.808 0.025

0.019 1.154 0.037

0.003 0.005

0.007 0.010

0.010 0.015

0.071

0.163

0.234

0.073 0.039 0.013 0.013 0.009 0.002 0.020

0.152 0.089 0.024 0.024 0.019 0.006 0.041

0.225 0.129 0.037 0.037 0.028 0.008 0.061

0.027 0.047 0.006 0.014 0.012 0.004 0.004

0.054 0.095 0.011 0.027 0.025 0.008 0.008

0.081 0.143 0.017 0.041 0.037 0.011 0.012

0.158 0.006

0.486 0.024

0.644 0.030

0.049

0.160

0.209

0.078

0.163

0.242

0.004 0.002 0.006 0.005 0.027

0.008 0.006 0.014 0.010 0.054

0.012 0.008 0.020 0.014 0.080

30

JED DEVARO AND FIDAN ANA KURTULUS

Table 4. (Continued ) WEffectA3 Financial services Other business services Public administration Education Health Other community services

0.008 0.010 0.011 0.009 0.000 0.001

Largest occupational group at workplace Managers and 0.006 administrators Professional occupations 0.019 Associate professional and 0.007 technical occupations Craft and related 0.013 occupations Personal and protective 0.001 service occupations Sales occupations 0.016 Plant and machine 0.011 operatives Other occupations 0.010

WEffectA1

W(EffectA3 – EffectA1)

0.018 0.022 0.025 0.018 0.001 0.002

0.026 0.033 0.036 0.027 0.001 0.003

0.012

0.018

0.034 0.015

0.053 0.023

0.025

0.037

0.001

0.002

0.034 0.025

0.050 0.036

0.021

0.031

Note: The omitted categories for firm ownership, industry, and largest occupational group are, respectively, public sector, wholesale and retail, and clerical and secretarial. (EffectA3)i for establishment i is defined as ðEffectA3Þi ¼ ProbðY i ¼ 3jTEAMSi ¼ 1Þ2ProbðY i ¼ 3jTEAMS i ¼ 0Þ; which is the effect of team production on the probability that establishment i’s labor productivity is a lot above industry average. Cell entries under ‘‘WEffectA3’’ indicate how this effect varies, on average, by the firm characteristic indicated by the row title. This is computed slightly differently according to whether the characteristic is a single dummy variable, one dummy variable from a group of related dummies, or a continuous variable. When characteristic X is a single dummy variable, we computed for each establishment ðEffectA3jX i ¼ 1Þ ðEffectA3jX i ¼ 0Þ holding all of the other characteristics at their actual establishment levels. Then we took the average of these differences across all establishments in the sample. When the X in question is one dummy in a group of related dummies, the computation was the same as the one just described, except that the other dummies in the same group were evaluated at 0 rather than at their observed values for each establishment. The three groups of dummies that are considered in this way are the industry, occupation, and ownership variables. When the X in question is continuous, the differences we computed were not between X ¼ 1 and X ¼ 0 but rather between the 0.75 and the 0.25 quantile of X, again holding all the characteristics of the establishment at their individual values. (EffectA1)i for establishment i is defined as ðEffectA1Þi ¼ ProbðY i ¼ 1jTEAMSi ¼ 1Þ2ProbðY i ¼ 1jTEAMSi ¼ 0Þ; which is the effect of team production on the probability that establishment i’s labor productivity is average or below. The ‘‘WEffectA1’’ column was computed analogously. The ‘‘4ðEffectA32EffectA1Þ’’ column is simply the difference between the first and second columns and is our measure of the degree to which the predicted benefit from teams to labor productivity varies with the firm characteristic indicated by the row title. N ¼ 1; 660:

31

What Types of Organizations Benefit from Team Production

Table 5.

How the Predicted Benefit of Teams on Products Quality Varies with Firm Characteristics. WEffectA3

Basic firm characteristics Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Just-in-time production Information Incentive alignment Decisions Work at home Firm ownership Private sector franchise Private sector nonfranchise Alternative private sector franchise Alternative private sector non-franchise Industry Manufacturing Electricity, gas, and water Construction Hotels and restaurants Transport and communication Financial services

WEffectA1

W(EffectA3 – EffectA1)

0.020 0.230 0.014

0.024 0.531 0.016

0.045 0.761 0.030

0.048 0.059

0.067 0.075

0.115 0.133

0.036

0.039

0.075

0.051 0.019 0.018 0.107 0.089 0.016 0.008

0.068 0.022 0.005 0.149 0.153 0.024 0.017

0.120 0.040 0.023 0.256 0.242 0.040 0.024

0.012 0.022 0.050 0.025 0.036 0.010 0.012

0.021 0.023 0.070 0.026 0.039 0.011 0.013

0.033 0.045 0.120 0.051 0.074 0.021 0.025

0.164 0.023

0.207 0.073

0.370 0.096

0.026

0.045

0.071

0.042

0.090

0.132

0.001 0.110 0.045 0.027 0.081

0.019 0.138 0.076 0.027 0.138

0.020 0.248 0.120 0.054 0.219

0.081

0.117

0.198

32

JED DEVARO AND FIDAN ANA KURTULUS

Table 5. (Continued ) WEffectA3 Other business services Public administration Education Health Other community services

0.047 0.097 0.011 0.039 0.003

Largest occupational group at workplace Managers and 0.022 administrators Professional occupations 0.004 Associate professional and 0.057 technical occupations Craft and related 0.072 occupations Personal and protective 0.025 service occupations Sales occupations 0.039 Plant and machine 0.026 operatives Other occupations 0.029

WEffectA1

W(EffectA3 – EffectA1)

0.074 0.134 0.004 0.071 0.006

0.121 0.231 0.007 0.110 0.009

0.019

0.041

0.018 0.067

0.014 0.124

0.104

0.176

0.048

0.073

0.048 0.020

0.087 0.046

0.031

0.060

Note: The omitted categories for firm ownership, industry, and largest occupational group are, respectively, public sector, wholesale and retail, and clerical and secretarial. (EffectA3)i for establishment i is defined as ðEffectA3Þi ¼ ProbðY i ¼ 3jTEAMSi ¼ 1Þ2ProbðY i ¼ 3jTEAMS i ¼ 0Þ; which is the effect of team production on the probability that establishment i’s product quality is a lot above industry average. Cell entries under ‘‘WEffectA3’’ indicate how this effect varies, on average, by the firm characteristic indicated by the row title. This is computed slightly differently according to whether the characteristic is a single dummy variable, one dummy variable from a group of related dummies, or a continuous variable. When characteristic X is a single dummy variable, we computed for each establishment ðEffectA3jX i ¼ 1Þ ðEffectA3jX i ¼ 0Þ holding all of the other characteristics at their actual establishment levels. Then we took the average of these differences across all establishments in the sample. When the X in question is one dummy in a group of related dummies, the computation was the same as the one just described, except that the other dummies in the same group were evaluated at 0 rather than at their observed values for each establishment. The three groups of dummies that are considered in this way are the industry, occupation, and ownership variables. When the X in question is continuous, the differences we computed were not between X ¼ 1 and X ¼ 0 but rather between the 0.75 and the 0.25 quantile of X, again holding all the characteristics of the establishment at their individual values. (EffectA1)i for establishment i is defined as ðEffectA1Þi ¼ ProbðYi ¼ 1jTEAMSi ¼ 1Þ2ProbðYi ¼ 1jTEAMSi ¼ 0Þ; which is the effect of team production on the probability that establishment i’s product quality is average or below. The ‘‘WEffectA1’’ column was computed analogously. The ‘‘4ðEffectA32EffectA1Þ’’ column is simply the difference between the first and second columns and is our measure of the degree to which the predicted benefit from teams to product quality varies with the firm characteristic indicated by the row title. N ¼ 1; 839:

33

What Types of Organizations Benefit from Team Production

Table 6. Firm Characteristics by Sign of Association with ‘‘Predicted Benefit of Team Production on Organizational Performance’’.

Basic firm characteristics Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed Term workers under one year Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Just-in-time production Information Incentive alignment Decisions Work at home Firm ownership Private sector franchise Private sector non-franchise Alternative private sector franchise Alternative private sector non-franchise Industry Manufacturing Electricity, gas, and water Construction Hotels and restaurants Transport and communication Financial services Other business services

Financial Performance

Labor Productivity

Product Quality

+ + +

+

+

+ + +

+

+ +

+

+

+ + +

+

+ +

+

+

+

+ + + +

+

+

+

+

+

+

+ +

+

+ + +

+

+

+ +

+ +

34

JED DEVARO AND FIDAN ANA KURTULUS

Table 6. (Continued )

Public administration Education Health Other community services

Financial Performance

Labor Productivity

Product Quality

+ + +

+

+

+

+

Largest occupational group at workplace Managers and administrators Professional occupations + Associate professional and + technical occupations Craft and related occupations Personal and protective service occupations Sales occupations + Plant and machine operatives Other occupations

+ +

+

+ +

Note: +/ represents the signs from Tables 3–5 of DðEffecA32EffectA1Þ; which is a measure of the degree to which the predicted benefit, in terms of financial performance, labor productivity or product quality, from teams varies with firm characteristics.

which the respondent interprets the term financial performance as synonymous with profit or value added. This restriction was not possible in the McNabb and Whitfield analysis, since the 1990 WIRS did not ask about the respondent’s interpretation of financial performance. This difference is potentially important. For example, DeVaro (2004a) found that the estimated effect of teams on financial performance is quite sensitive to the respondent’s interpretation of financial performance. Regarding firm size, we see that larger establishments experience larger predicted benefits from teams to financial performance. However, smaller establishments experience larger predicted benefits from teams to both labor productivity and product quality. One interpretation of these results is that the nature of the monitoring problem changes with the scale of the organization, such that internal monitoring by peers in the team context is relatively more effective when the organization is small (so that labor productivity is enhanced more by teams in small establishments). But in larger establishments there is greater scope for specialization of tasks and

What Types of Organizations Benefit from Team Production

35

also more diversity in worker information sets, implying that team members’ information sets are more likely to be non-duplicative (so that the benefits of teams are higher). Such benefits from teams in large establishments might outweigh the potentially higher costs of monitoring, so that on net the benefits of teams to financial performance are increasing with establishment size. The result concerning establishment size further highlights the dangers of drawing inferences about the benefits of teams based on measures of organizational outcomes that do not capture the full spectrum of benefits and costs from teams. Our results for financial participation suggest that establishments in which workers receive compensation through various types of variable-pay schemes, including profit-related pay and share ownership schemes, experience higher predicted benefits from teams to financial performance than do establishments with less participation. This result is consistent with the complementarity hypothesis discussed earlier. While financial participation is associated with higher predicted benefits of teams to financial performance, it is also associated with lower predicted benefits of teams to labor productivity and product quality. As was the case for establishment size, this result suggests that standard outcome measures such as labor productivity are not inclusive enough of the various benefits and costs of teams to identify the positive organizational benefits from teams that accrue in firms with a high degree of financial participation. All of the industry results should be interpreted relative to the reference group of wholesale and retail. Relative to wholesale and retail, the financial services and other business service industries have lower predicted benefits of teams to all three measures of organizational performance, and the education industry has higher predicted benefits across all three measures. Predicted benefits to financial performance and labor productivity in manufacturing are higher than in wholesale in retail, which is consistent with a significant volume of previous work in the manufacturing sector (focusing heavily on measures of labor productivity as an organizational outcome) that is generally supportive of positive effects of team production on firm performance. An interesting result is that for all of the other industries, the sign of the predicted benefits of teams is common for labor productivity and product quality but is the opposite of that for financial performance. For example, industries for which the predicted benefits of teams to labor productivity and product quality are negative but the predicted benefits to financial performance are positive include health, public administration, construction, and electricity, gas, and water. Industries for which the predicted benefits to labor productivity and product quality are positive but the

36

JED DEVARO AND FIDAN ANA KURTULUS

predicted benefits to financial performance are negative include hotels and restaurants, transport and communication, and other community services. In addition to the results on unions, firm size, financial participation, and industry, some further interesting results emerge concerning the other covariates. We conclude this section by noting a few of these. First, the result concerning the age of the establishment is interesting. The binary variable ‘‘operation over five years’’ is associated with lower predicted benefits of teams to all three measures of organizational performance, suggesting that younger firms experience larger benefits from teams. Second, the results concerning the variables ‘‘incentive alignment’’ (i.e., employees are fully committed to the values of the firm) and ‘‘decisions’’ (i.e., most decisions at this workplace are made without consulting employees) are as we would expect. That is, a greater degree of disagreement with the statement about incentive alignment and a greater degree of agreement with the statement about decisions is associated with lower predicted benefits of teams to all three measures of organizational performance. Third, the results concerning ‘‘work at home’’ are intuitive as well: as the fraction of time spent working at home during normal business hours increases, the predicted benefits of teams to all measures of organizational performance decrease (by over 2, 1, and 2 percentage points for financial performance, labor productivity, and product quality, respectively), which is to be expected since frequent interaction and collaboration with other team members is necessarily reduced when an employee does a significant amount of work from home.

6. CONCLUDING REMARKS While a significant volume of research addresses the question of how teams and other high-performance work practices affect organizational performance, comparatively little attention has been devoted to the question of what types of organizations benefit from teams. Our paper has aimed to partially address this gap. Although we report results for a large set of firm characteristics, our focus has been on firm size, unions, financial participation, and industry. A distinguishing feature of our approach is the use of structural models in which teams and autonomy are treated as endogenous determinants of organizational performance. This is important because accurate measures of the extent to which the benefits of teams to organizational performance vary with organizational characteristics require accurate measures of the predicted benefits of teams to organizational performance. If the firm’s choices of teams and autonomy are treated as

What Types of Organizations Benefit from Team Production

37

exogenous, as is common in the teams literature, then the resulting estimates of the effects of teams and autonomy will be biased as a result of correlations among the unobserved determinants of teams, autonomy, and organizational performance. As shown in DeVaro (2004b) such biases can be quite substantial, particularly in the case of product quality. An interesting general pattern of results that emerges from our analysis is that while a particular firm characteristic might be associated with large predicted benefits of teams to labor productivity and product quality, the same firm characteristic is often associated with lower predicted benefits of teams to financial performance. Similarly, even when a characteristic is associated with lower benefits of teams to labor productivity or product quality, it often is associated with higher benefits to financial performance. This finding is of particular interest since financial performance is much less frequently seen as an outcome variable in this literature than measures like labor productivity or product quality. Since financial performance, as a measure of profit, is more inclusive than these other measures of the full spectrum of benefits and costs induced by teams, our results suggest that studies that focus only on labor productivity and product quality might produce an incomplete picture of the total effect of teams on firm performance. In future research, it would be interesting to implement our structural approach using other data sets. One possibility is to make use of the fifth wave of the WERS when it is released. We close our discussion with a final recommendation. While our structural models interact the ‘‘teams’’ treatment with all of the firm characteristics, autonomy enters our models only as an intercept shift. In future work with larger data sets, it would be interesting to generalize the model by interacting autonomy with all of the firm characteristics as well. This would allow two separate regressions to be run in Stage 3, one for the ‘‘predicted benefits to organizational performance from using autonomous teams’’ and another for the ‘‘predicted benefits to organizational performance from using non-autonomous teams.’’ Then one could discern what types of firms would benefit more from autonomous team production than from closely managed team production, and vice versa.

NOTES 1. It is, of course, possible to obtain variation in characteristics within the context of a single case study if, for example, one exploits variation across multiple establishments in the same firm. Ultimately, however, general inferences cannot be drawn from analysis of only one or a handful of observations from a specialized production setting.

38

JED DEVARO AND FIDAN ANA KURTULUS

2. A more commonly used outcome variable in this literature is labor productivity (DeVaro, 2004b; Eriksson, 2003; Kato & Morishima, 2002; Black & Lynch, 2001; Ichniowski et al., 1997; Banker, Field, Schroeder, & Sinha, 1996; Ichniowski, 1990). Other outcome variables have also been studied, such as innovation and R&D (Michie & Sheehan, 1999), turnover (Huselid, 1995), worker well-being and wages (Caroli & van Reenen, 2001, Bauer & Bender, 2001), product quality (DeVaro, 2004b; Ichniowski & Shaw, 1999; Ichniowski et al., 1997; Banker et al., 1996), worker satisfaction (Batt, 2004; Batt & Appelbaum, 1995; Godard, 2001), worker absenteeism (Askenazy, Caroli, & Marcus, 2001) and firms’ layoff rates (Osterman, 2000). An extensive survey of the empirical literature outside the discipline of economics concerning the effects of team production can be found in Cohen and Bailey (1997). 3. However, recent evidence from a large, nationally representative sample of British establishments suggests that both types of teams confer benefits of similar magnitude to financial performance, labor productivity, and product quality (DeVaro, 2004a,b). 4. This result does not necessarily generalize across countries. For example, Doucouliagos and Laroche (2003) show in their meta-analysis that there is a positive association between unions and labor productivity in the United States. 5. Batt and Appelbaum (2003) also looked at employee job satisfaction and organizational commitment as dependent variables. They found that for workers in the network crafts and sewing machine operators’ occupations, teams significantly improve job satisfaction and organizational commitment. On the other hand, for workers in customer service occupations, this is not the case. 6. This kind of misreporting could arise if managers who adopt innovative work practices want to believe that their organizations are doing better, thereby rationalizing their decision to employ those practices. 7. Consistent with this line of argument, Machin and Stewart have shown the subjective financial performance measure in the WIRS to be a good predictor of workplace closure. 8. The responses sum to 1,800 instead of the full 2,191 because some respondents reported that no comparison was possible or that the relevant data were not available. 9. See DeVaro (2004a) for a discussion of the theoretical rationale for granting teams autonomy. 10. The statistical model requires some standard identifying restrictions on the disturbance covariance matrix. All of its diagonal elements are normalized to one. Furthermore, both the disturbance covariances s01 and s03 are assumed to be 0. The restrictions on the covariance matrix are weaker than those that are imposed in the simpler ‘‘non-structural’’ approaches common to the teams’ literature. These approaches also impose (implicitly) s01 ¼ 0 and s03 ¼ 0: In addition, however, they impose the restrictions s02 ¼ 0; s12 ¼ 0; s13 ¼ 0; and s23 ¼ 0; whereas our models treat these as unrestricted parameters to be estimated. This issue is not discussed in the teams literature, because with the standard approach of using a single equation for organizational performance (with only one disturbance term in the model), treating teams and/or autonomy as exogenous variables on the right-hand-side, all of these assumptions are implicit rather than explicit as they are in our models. Apart from the non-linearities introduced by the distributional assumptions, identification of the model is facilitated by a set of exclusion restrictions on the covariates in each equation. The specification for each equation is detailed in the appendix. More

What Types of Organizations Benefit from Team Production

39

discussion of the exclusion restrictions and their justification on the basis of theory, previous empirical work, and independent tests confirming that the variables appear unimportant in the equations from which they are excluded, can be found in DeVaro (2004a,b). 11. In the appendix, we state the formulae for computing these three effects. The notation for these Effects ‘‘A1’’, ‘‘A2’’, and ‘‘A3’’ was introduced in DeVaro (2004a,b). In those studies, the role of the ‘‘A’’ in the notation was to distinguish the effects from the analogous ‘‘predicted benefits from using autonomous teams’’ (Effects B1, B2, and B3) and the ‘‘predicted benefits from using non-autonomous teams’’ (Effects C1, C2, and C3). We do not discuss the ‘‘B’’ and ‘‘C’’ effects in the present paper, because here we are interested in the relationships of the firm characteristics to the predicted benefits from teams. In principle, we could analyze either the ‘‘B’’ effects or the ‘‘C’’ effects like we do the ‘‘A’’ effects, though this would yield no insights beyond those we present in the paper. The reason is that while the structural model interacts the ‘‘teams treatment’’ with all of the covariates, the ‘‘autonomy treatment’’ simply shifts the intercept. Hence, any relationships between the covariates and the ‘‘B’’ or ‘‘C’’ effects arises because of interactions of the covariates with ‘‘teams’’ rather than with ‘‘autonomy’’. 12. The only covariates in the model that are ‘‘continuous’’, in the usual sense, are establishment size and the fraction of employees who work part time. However, the model also includes several variables recorded in the survey as ordered discrete categories. These include information, incentive alignment, and decisions (5 categories each); work at home (6 categories); multi-skilling and off-site training (7 categories each); and number of recognized unions (11 categories). To economize on the number of parameters to be estimated, we treat each of these variables as continuous indexes rather than creating multiple dummies for each category. In unreported sensitivity checks, we found the same qualitative results in models that include these variables as multiple dummies rather than as ‘‘continuous’’ indexes. 13. While a univariate measure of the predicted benefits from teams eases the presentation of results, collapsing EffectA3 and A1 into (EffectA32EffectA1) necessarily involves some loss of information, since neither EffectA3 nor EffectA1 can be inferred from their difference. For example, suppose that one workplace has EffectA3 ¼ 0:10 and EffectA1 ¼ 0:07; whereas another has EffectA3 ¼ 0:07 and EffectA1 ¼ 0:10: Both are ranked the same by the criterion EffectA3 EffectA1: However, this lost information is not particularly useful for our purposes. To see why, consider an alternative to (EffectA3 EffectA1) that assigns greater weight to increases in EffectA3 than to decreases in EffectA1, or vice versa. Using such an alternative criterion, one of the two workplaces in the example could clearly be ranked above the other in terms of predicted benefits from teams. Absent any information about the establishments’ loss functions, however, there is no clear basis for weighting increases in EffectA3 differently from decreases in EffectA1. 14. More detailed information on the overall effects of teams (as opposed to the interactions we focus on in this paper) can be found in DeVaro, 2004a, b. Those analyses report magnitudes at the 0.25, 0.50, and 0.75 quantiles of (EffectA3 EffectA1) as well as the analogous information for autonomous teams and non-autonomous teams (as opposed to teams in general). Those studies also report all of these results from models that treat teams and autonomy as exogenous, to determine the nature of the bias that arises when the endogeneity of teams and autonomy is ignored.

40

JED DEVARO AND FIDAN ANA KURTULUS

ACKNOWLEDGMENT Thanks to The UK Data Archive for access to the data and to David Rosenblum, Dana Samuelson, Emma Stephens, and an anonymous referee for helpful comments.

REFERENCES Adams, C. P. (2003). The use of profit sharing when workers make decisions: Evidence from a survey of manufacturing workers. In: T. Kato & J. Pliskin (Eds), Determinants of the incidence and the effects of participatory organizations: Advances in the economic analysis of participatory and labor-managed firms, (Vol. 7, pp. 173–209). Amsterdam: Elsevier/ JAI. Aghion, P., & Tirole, J. (1997). Formal and real authority in organizations. Journal of Political Economy, 105, 1–29. Alchian, A. A., & Demsetz, H. (1972). Production, information costs, and economic organization. American Economic Review, 62, 777–795. Askenazy, P., Caroli, E., & Marcus, V. (2001). New organizational practices and working conditions: Evidence from France in the 1990. CEPREMAP Working Papers (Couverture Orange) 0106, CEPREMAP. Banker, R. E., Field, J. M., Schroeder, R. G., & Sinha, K. K. (1996). Impact of work teams on manufacturing performance: A longitudinal field study. Academy of Management Journal, 39(4), 867–890. Bartel, A. P. (2004). Human resource management and organizational performance: Evidence from retail banking. Industrial and Labor Relations Review, 57(2), 181–203. Batt, R. (1999). Work organization, technology, and performance in customer service and sales. Industrial and Labor Relations Review, 52(4), 539–564. Batt, R. (2001). The economics of teams among technicians. British Journal of Industrial Relations, 39(1), 1–24. Batt, R. (2004). Who benefits from teams? Comparing workers, supervisors, and managers. Industrial Relations, 43(1), 183–212. Batt, R., & Appelbaum, E. (1995). Worker participation in diverse settings: Does the form affect the outcome, and if so, who benefits? British Journal of Industrial Relations, 33(3), 354–378. Bauer, T. K., & Bender. S. (2001). Flexible work systems and the structure of wages: Evidence from matched employer-employee data. IZA Discussion Paper No. 353, August. Ben-Ner, A., & Jones, D. C. (1995). Employee participation, ownership, and productivity: A theoretical framework. Industrial Relations, 34(4), 532–554. Berg, P., Appelbaum, E., Bailey, T., & Kalleberg, A. L. (1996). The performance effects of modular production in the apparel industry. Industrial Relations, 35(5), 356–373. Black, S. E., & Lynch, L. M. (2001). How to compete: The impact of workplace practices and information technology on productivity. The Review of Economics and Statistics, 83(3), 434–445. Black, S. E., & Lynch, L. M. (2004). What’s driving the new economy?: The benefits of workplace innovation. The Economic Journal, 114(February 2003), F97–F116.

What Types of Organizations Benefit from Team Production

41

Boning, B., Ichniowski, C., & Shaw, K. (2003). Opportunity counts: Teams and the effectiveness of production incentives. NBER Working Paper #8306, May. Bryson, A, & Wilkinson, D. (2002). Collective bargaining and workplace performance: An investigation using the workplace employee relations survey 1998. Department of Trade and Industry Employment Relations Research Series No. 12. Cappelli, P., & Neumark, D. (2001). Do ‘high-performance’ work practices improve establishment-level outcomes? Industrial and Labor Relations Review, 54(4), 737–775. Caroli, E., & van Reenen, J. (2001). Skill-biased organizational change? Evidence from a panel of British and French establishments. The Quarterly Journal of Economics, 116(4), 1449–1492. Cohen, S., & Bailey, D. (1997). What makes teams work: Group effectiveness research from the shop floor to the executive suite. Journal of Management, 23(3), 239–290. Collier, W., Green, F., Peirson, J., & Wilkinson, D. (2004). Training and establishment survival. Working paper. University of Kent, Canterbury. Cully, M., Woodland, S., O’Reilly, A., & Dix, G. (1999). Britain at work: As depicted by the 1998 workplace employee relations survey. London: Routledge. Department of Trade and Industry and Advisory. (2001). Conciliation and arbitration service, Workplace Employee Relations Survey: Cross-Section, 1998 [computer file] (6th ed.). Colchester, Essex: UK Data Archive [distributor] January, SN: 3955. DeVaro, J. (2004a). Teams, autonomy, and the financial performance of firms. Industrial Relations, forthcoming. DeVaro, J. (2004b). The effects of self-managed and closely managed teams on labor productivity and product quality: An empirical analysis of a cross section of establishments. Working paper. Cornell University. Doucouliagos, C., & Laroche, P. (2003). What do unions do to productivity? A meta-analysis. Industrial Relations, 42(4), 650–691. Eriksson, T. (2003). The effects of new work practices – Evidence from employer-employee data. In: T. Kato & J. Pliskin (Eds), Determinants of the incidence and the effects of participatory organizations: Advances in the economic analysis of participatory and labormanaged firms, (Vol. 7, pp. 3–30). Amsterdam: Elsevier/JAI. Freeman, R., & Dube, A. (2000). Shared compensation systems and decision-making in the US job market. Incomes and productivity in North America, Papers from the 2000 Seminar. Washington, DC: Secretariat of the Commission for Labor Cooperation. Godard, J. (2001). High performance and the transformation of work? The implications of alternative work practices for the experience and outcomes of work. Industrial and Labor Relations Review, 54, 776–805. Hackman, J. R. (1987). The design of work teams. In: J. Lorsch (Ed.), Handbook of organizational behavior. Englewood Cliffs, NJ: Prentice-Hall. Hamilton, B. H., Nickerson, J. A., & Owan, H. (2003). Team incentives and worker heterogeneity: An empirical analysis of the impact of teams on productivity and participation. Journal of Political Economy, 111(3), 465–497. Holmstrom, B. (1982). Moral hazard in teams. Bell Journal of Economics, 13(Autumn), 324–340. Huselid, M. A. (1995). The impact of human resource management practices on turnover, productivity, and corporate financial performance. Academy of Management Journal, 38(3), 635–872. Ichniowski, C. (1990). Human resource management systems and the performance of US manufacturing businesses. NBER Working Paper No. 3449, September.

42

JED DEVARO AND FIDAN ANA KURTULUS

Ichniowski, C., & Shaw, K. (1999). The effects of human resource management systems on economic performance: An international comparison of US and Japanese plants. Management Science, 45(5), 704–721. Ichniowski, C., Shaw, K., & Prennushi, G. (1997). The effects of human resource management practices on productivity: A study of steel finishing lines. American Economic Review, 87(3), 291–313. Itoh, H. (1991). Incentives to help in multi-agent situations. Econometrica, 59(May), 611–636. Itoh, H. (1992). Cooperation in hierarchical organizations: An incentive perspective. Journal of Law, Economics, and Organization, 8(April), 321–345. Kandel, E., & Lazear, E. P. (1992). Peer pressure and partnerships. Journal of Political Economy, 100(August), 801–817. Kato, T., & Morishima, M. (2002). The productivity effects of participatory employment practices: Evidence from new Japanese panel data. Industrial Relations, 41(October), 487–520. Kruse, D. (2002). Research evidence on the prevalence and effects of employee ownership. Journal of Employee Ownership Law and Finance, 14(4), 65–90. Kruse, D., & Blasi, J. (1997). Employee ownership, employee attitudes, and firm performance: A review of evidence. In: D. Lewin, D. J. B. Mitchell & M. A. Zaidi (Eds), The human resources management handbook, Part 1. Greenwich, CT: JAI Press. Kruse, D., Freeman, R., Blasi, J., Buchele, R., Scharf, A., Rodgers, L., & Mackin, C. (2004). Motivating employee-owners in ESOP firms: Human resource policies and company performance. In: V. Perotin & A. Robinson (Eds), Advances in the economic analysis of participatory and labor-managed firms, (Vol. 8, pp. 101–128). Oxford, UK: Elsevier Science Ltd. Lazear, E. P. (1995). Personnel economics. Cambridge, MA: MIT Press. Lazear, E. P. (1998). Personnel economics for managers. New York: Wiley. Legros, P., & Matthews, S. A. (1993). Efficient and nearly efficient partnerships. Review of Economic Studies, 60(July), 599–611. Machin, S. J., & Stewart, M. B. (1990). Unions and the financial performance of British private sector establishments. Journal of Applied Econometrics, 5(4), 327–350. Machin, S. J., & Stewart, M. B. (1996). Trade unions and financial performance. Oxford Economic Papers, 48, 213–241. McAfee, R. P., & McMillan, J. (1991). Optimal contracts for teams. International Economic Review, 32(August), 561–577. McNabb, R., & Whitfield, K. (1997). Unions, flexibility, team working and financial performance. Organization Studies, 18(5), 821–838. McNabb, R., & Whitfield, K. (1998). The impact of financial participation and employee involvement on financial performance. Scottish Journal of Political Economy, 45(2), 171–187. Michie, J., & Sheehan, M. (1999). HRM practices, R&D expenditure and innovative investment: Evidence from the UK’s 1990 workplace industrial relations survey (WIRS). Industrial and Corporate Change, 8(2), 211–234. Osterman, P. (1994). How common is workplace transformation and who adopts it? Industrial and Labor Relations Review, 47(2), 173–188. Osterman, P. (2000). Work reorganization in an era of restructuring: Trends in diffusion and effect on employee welfare. Industrial and Labor Relations Review, 53, 179–196. Rasmusen, E. (1987). Moral hazard in risk-averse teams. Rand Journal of Economics, 18(Autumn), 428–435.

43

What Types of Organizations Benefit from Team Production

APPENDIX The statistical model specifies probabilities for each of the nine possible outcomes that a workplace might realize. Letting i index workplaces, these potential outcomes and their probabilities are as follows: Probability

Yi ¼

TEAMSi ¼

AUTOi ¼

1 1 2 2 3 3 1 2 3

1 1 1 1 1 1 0 0 0

1 0 1 0 1 0 Unobserved Unobserved Unobserved

P1i(h) P2i(h) P3i(h) P4i(h) P5i(h) P6i(h) P7i(h) P8i(h) P9i(h)

Let Zij ¼ 1 ¼0

if Workplace i experiences the jth outcome otherwise; for i ¼ 1; 2; . . . ; N and j ¼ 1; 2; . . . ; 9

Then the likelihood function, L*, is L ¼

N Y 9 Y

Z

Pji ji

i¼1 j¼1

and the log-likelihood function, L, is L¼

N X 9 X

Z ji log Pji

i¼1 j¼1

Since each of the endogenous variables is observed only discretely, each probability of the form Pji(h) is a multiple integral of the joint density f(e0i, e1i, e2i, e3i). Suppressing all subscripts i, the expression for P1(h) is as follows: P 1 ð yÞ ¼

Z

ðaþX 1 d1 Þ 1

Z

1 X 2b

Z

1

f ð1 ; 2 ; 3 Þ d3 d2 d1 X 3g

Probabilities P2(h) to P9(h) are similarly defined.

44

JED DEVARO AND FIDAN ANA KURTULUS

Formulae for Computing Effects A1, A2, A3 Each effect is a function of the Pj(h) and is computed as follows, evaluating the expressions for Pj at the estimated values of h: 1i þP2i ðEffectA1Þi ¼ PP 6

Pji

j¼1

3i þP4i ðEffectA2Þi ¼ PP 6

Pji

j¼1

5i þP6i ðEffectA3Þi ¼ PP 6

Pji

j¼1

P7i 9 P

Pji

j¼7

P8i

9 P

Pji

j¼7

P9i

9 P

Pji

j¼7

Control Variables We include the following common set of control variables in each of X1, X2, and X3. Single-establishment firm: dummy variable that equals 1 if the establishment is either a single independent establishment not belonging to another body, or the sole UK establishment of a foreign organization and equals 0 if the establishment is one of a number of different establishments within a larger organization. Establishment size: total number of full time, part time, and temporary workers at the establishment. Fraction of part time workers: number of part time workers at the establishment as a fraction of establishment size. Temporary workers: dummy variable that equals 1 if there are temporary agency employees working at the establishment at the time of the survey and equals 0 otherwise. Fixed term workers under one year: dummy variable that equals 1 if there are employees who are working on a temporary basis or have fixed-term contracts for less than one year and equals 0 otherwise. Fixed term workers over one year: dummy variable that equals 1 if there are employees who have fixed term contracts for one year or more and equals 0 otherwise. Union workers: dummy variable that equals 1 if any of the workers at the establishment belong to a union and equals 0 otherwise. Private sector franchise: dummy variable that equals 1 if the establishment is a private sector company and a franchise and equals 0 otherwise. Private sector non-franchise: dummy variable that equals 1 if the establishment is a private sector company but not a franchise and equals 0 otherwise.

What Types of Organizations Benefit from Team Production

45

Alternative private sector franchise: dummy variable that equals 1 if the establishment is an alternative private sector firm and a franchise and equals 0 otherwise. Alternative private sector non-franchise: dummy variable that equals 1 if the establishment is an alternative private sector firm but not a franchise and equals 0 otherwise. The following additional control variables were included in the financial performance equation in DeVaro (2004a), and we include them in our labor productivity and product quality models as well so that our specification is identical across the three measures of organizational performance. A number of these additional variables have been found to be significantly associated with financial performance in earlier analyses of the WIRS/WERS data, for example, union activity, (Bryson & Wilkinson, 2002); training (Collier et al., 2004); and financial participation (McNabb & Whitfield, 1998). Financial participation: dummy variable that equals 1 if any employees at the workplace receive payments or dividends from any of the following variable pay schemes (profit-related payments or bonuses, deferred profit sharing schemes, employee share ownership schemes, individual or group performance-related schemes, other cash bonus). This variable was included in the financial performance equation in McNabb and Whitfield (1998) and found to be significant; their study used data from an earlier wave of the survey, so their definition of this variable differed slightly from ours. Owner manager: dummy variable that equals 1 if any of the controlling owners of the workplace are actively involved in day-to-day management on a full-time basis, and 0 otherwise. This question was only asked of private sector workplaces for which a single individual or family has controlling interest (meaning at least 50% ownership) over the company. Foreign owned: dummy variable that equals 1 if workplace reports that either of the following two statements best describes the ownership of the workplace (predominantly foreign owned, meaning 51% or more; foreign owned/controlled) and 0 if any of the following three statements is chosen (UK owned/ controlled, predominantly UK owned, meaning 51% or more; UK and foreign owned). This question was asked only of private sector workplaces. Operation over five years: dummy variable that equals 1 if the workplace has been operating at its present address for five years or more, and 0 otherwise Multi-skilling: degree of multi-skilling at the workplace in the largest occupational group. Question asks what fraction of these employees are formally trained to be able to do jobs other than their own. Responses are: ‘‘None 0%’’ (1); ‘‘Just a few 1–19%’’ (2); ‘‘Some 20–39%’’ (3); ‘‘Around half 40–59%’’ (4); ‘‘Most 60–79%’’ (5); ‘‘Almost all 80–99%’’ (6); ‘‘All 100%’’ (7).

46

JED DEVARO AND FIDAN ANA KURTULUS

Number of recognized unions: total number of recognized unions at the workplace. Induction training: dummy variable that equals 1 if there is a standard induction program designed to introduce new employees (in the largest occupational group) to the workplace, and 0 otherwise. Off-site training: discrete variable measuring the proportion of experienced employees (in the largest occupational group) that have had formal off-the-job training (away from the normal place of work, but either on or off the premises) over the last 12 months. Responses are: ‘‘None 0%’’ (1); ‘‘Just a few 1–19%’’ (2); ‘‘Some 20–39%’’ (3); ‘‘Around half 40–59%’’ (4); ‘‘Most 60–79%’’ (5); ‘‘Almost all 80–99%’’ (6); ‘‘All 100%’’ (7). In the TEAMS equation, we include a dummy variable indicating whether a just-in-time system is in operation at the establishment. Specifically, the employer is asked ‘‘Does this workplace operate a system designed to minimize inventories, supplies or work-in progress? This is sometimes known as Just-in-Time.’’ Responses are coded as 1 for yes and 0 for no. In the autonomy equation, we include a set of four proxies for the organizational and informational structure of the establishment, the alignment of incentives between workers and owners, and the importance to the firm of monitoring inputs. The first three of these are qualitative measures of managerial opinion. The respondent manager is asked to comment on each of a list of statements, responding with ‘‘Strongly agree’’ (1), ‘‘Agree’’ (2), ‘‘Neither agree nor disagree’’ (3), ‘‘Disagree’’ (4), or ‘‘Strongly disagree’’ (5). The questions of interest as determinants of team autonomy are as follows: Information: ‘‘Those at the top are best placed to make decisions about this workplace.’’ Incentive alignment: ‘‘Employees here are fully committed to the values of this organization.’’ Decisions: ‘‘Most decisions at this workplace are made without consulting employees.’’ In addition to these managerial opinion variables, as a proxy for the importance the employer places on monitoring worker inputs, we include a discrete variable measuring the proportion of workers at the establishment that ever work from home during normal working hours. Responses include: ‘‘Half or more 50%+’’, ‘‘A quarter up to a half 25–49%’’, ‘‘Up to a quarter 10–24%’’, ‘‘A small proportion 5–9%’’, ‘‘Hardly any (less than 5%)’’, or ‘‘None 0%’’. The full specification of the exogenous variables in X1, X2, and X3 is summarized in the following tables.

47

What Types of Organizations Benefit from Team Production

Exogenous Variables Included in Structural Models. FINPERi X1 Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year Fixed term workers over one year Union workers Private sector franchise Private sector nonfranchise Alternative private sector franchise Alternative private sector nonfranchise Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Just-in-time production Information Incentive alignment Decisions Work at home Industry controls (12) Occupation controls (10)

LABPRODi QUALITYi X1 X1

TEAMSi X2

AUTOi X3

Yes

Yes

Yes

Yes

Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes

Yes

Yes

Yes

Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes Yes Yes

Yes Yes Yes

Yes Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes Yes

Yes Yes

Yes Yes Yes Yes Yes Yes

48

JED DEVARO AND FIDAN ANA KURTULUS

Table A1.

AUTO Single-establishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Private sector franchise

Estimates from the Structural Model for Financial Performance. FINPER (TEAMS ¼ 1)

FINPER (TEAMS ¼ 0)

TEAMS

AUTO

1.103 (0.205) 0.122 (0.107) 0.006 (0.010) 0.231 (0.152) 0.165 (0.095) 0.016 (0.073) 0.052 (0.119) 0.103 (0.122) 0.207 (0.084) 0.126 (0.108) 0.166 (0.099) 0.013 (0.103) 0.023 (0.019) 0.090 (0.033) 0.088 (0.109) 0.013 (0.015) 0.384 (0.364)

0.337 (0.253) 0.047 (0.059) 0.136 (0.405) 0.244 (0.181) 0.123 (0.261) 0.172 (0.371) 0.203 (0.267) 0.068 (0.235) 0.176 (0.296) 0.286 (0.346) 0.238 (0.285) 0.027 (0.047) 0.027 (0.112) 0.334 (0.232) 0.004 (0.034) 2.743 (0.007)

0.349 (0.160) 0.028 (0.028) 0.479 (0.329) 0.251 (0.129) 0.164 (0.135) 0.126 (0.256) 0.078 (0.137)

0.261 (0.119) 0.009 (0.010) 0.248 (0.229) 0.291 (0.132) 0.017 (0.104) 0.232 (0.167) 0.057 (0.117)

0.644 (1.663)

0.247 (0.508)

49

What Types of Organizations Benefit from Team Production

Table A1. (Continued )

Private sector nonfranchise Alternative private sector franchise Alternative private sector nonfranchise Just-in-time

FINPER (TEAMS ¼ 1)

FINPER (TEAMS ¼ 0)

TEAMS

AUTO

0.070 (0.145) 0.100 (0.380) 0.263 (0.148)

0.293 (0.369) 0.742 (1.648) 0.660 (0.357)

0.180 (0.254) 0.147 (1.091) 0.260 (0.260)

0.235 (0.243) 0.222 (0.573) 0.514 (0.261)

0.350 (0.135)

Information Incentive alignment Decisions Work at home Constant c s

02

s

12

s

13

s

23

0.746 (0.265) 0.916 (0.081) 0.843 (17.149) 0.618 (5.656) 0.804 (0.001) 0.533 (0.038)

1.050 (0.553)

0.691 (0.410)

0.092 (0.042) 0.261 (0.071) 0.023 (0.048) 0.065 (0.060) 0.559 (0.526)

Note: Standard errors from the parametric bootstrap are in parentheses and are based on 75 bootstrap replications. The omitted firm ownership category is public sector. N ¼ 889: Indicates significance at the 10% level. Indicates significance at the 5% level.

50

Table A2.

JED DEVARO AND FIDAN ANA KURTULUS

Estimates from the Structural Model for Labor Productivity. LABPROD (TEAMS ¼ 1)

LABPROD (TEAMS ¼ 0)

TEAMS

AUTO

0.768 (0.267) 0.054 (0.091)

0.026 (0.298)

0.140 (0.122)

0.190 (0.107)

0.003 (0.005) 0.277 (0.127) 0.043 (0.073) 0.010 (0.081)

0.039 (0.073) 0.237 (0.328) 0.190 (0.288) 0.059 (0.217)

0.052 (0.024) 0.301 (0.219) 0.110 (0.219) 0.220 (0.103)

0.004 (0.004) 0.040 (0.181) 0.277 (0.075) 0.043 (0.080)

0.066 (0.082)

0.420 (0.384)

0.251 (0.163)

0.004 (0.093)

0.020 (0.079) 0.098 (0.065)

0.388 (0.233) 0.378 (0.223)

0.146 (0.103)

0.051 (0.087)

Owner manager

0.089 (0.117)

0.039 (0.322)

Foreign owned

0.079 (0.102)

0.165 (0.300)

Operation over five years

0.050 (0.098)

0.118 (0.298)

Multi-skilling

0.044 (0.015)

0.062 (0.040)

AUTO Singleestablishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year Fixed term workers over one year Union workers Financial participation

51

What Types of Organizations Benefit from Team Production

Table A2. (Continued ) LABPROD (TEAMS ¼ 1)

LABPROD (TEAMS ¼ 0)

TEAMS

AUTO

Number of recognized unions

0.009 (0.022)

0.069 (0.123)

Induction training

0.047 (0.070)

0.099 (0.248)

Off-site training

0.022 (0.016)

0.097 (0.054)

Private sector franchise Private sector non-franchise Alternative private sector franchise Alternative private sector non-franchise Just-in-time

0.557 (0.278) 0.144 (0.099) 0.557 (0.373)

2.751 (0.454) 0.217 (0.347) 0.282 (1.980)

0.670 (1.379) 0.069 (0.175) 0.152 (0.414)

0.596 (0.287) 0.094 (0.137) 0.274 (0.373)

0.052 (0.106)

0.457 (0.431)

0.170 (0.183)

0.103 (0.131)

0.289 (0.118)

Information Incentive alignment Decisions Work at home Constant c

0.944 (0.239) 1.168 (0.064)

1.278 (0.769)

0.762 (0.284)

0.105 (0.032) 0.190 (0.041) 0.059 (0.037) 0.060 (0.046) 0.685 (0.436)

52

JED DEVARO AND FIDAN ANA KURTULUS

Table A2. (Continued ) LABPROD (TEAMS ¼ 1) s

02

s

12

s

13

s

23

LABPROD (TEAMS ¼ 0)

TEAMS

AUTO

0.046 (0.361) 0.386 (1.415) 0.496 (0.018) 0.693 (0.009)

Note: Standard errors from the parametric bootstrap are in parentheses (75 bootstrap replications). Dummies for industry and largest occupational group are also included in the TEAMS* and AUTO* equations. The omitted firm ownership category is public sector. Sample size is 1,660. Indicates significance at the 10% level. Indicates significance at the 5% level.

Table A3.

Estimates from the Structural Model for Quality of Product or Service.

AUTO Singleestablishment firm Establishment size Fraction of part time workers Temporary workers Fixed term workers under one year

QUALITY (TEAMS ¼ 1)

QUALITY (TEAMS ¼ 0)

TEAMS

AUTO

1.234 (0.123) 0.181 (0.090)

0.144 (0.173)

0.183 (0.123)

0.155 (0.105)

0.002 (0.003) 0.043 (0.103) 0.075 (0.073) 0.092 (0.057)

0.080 (0.047) 0.246 (0.243) 0.224 (0.162) 0.030 (0.171)

0.048 (0.020) 0.204 (0.189) 0.067 (0.095) 0.205 (0.091)

0.004 (0.005) 0.001 (0.161) 0.208 (0.081) 0.021 (0.064)

53

What Types of Organizations Benefit from Team Production

Table A3. (Continued )

Fixed term workers over one year Union workers Financial participation Owner manager Foreign owned Operation over five years Multi-skilling Number of recognized unions Induction training Off-site training Private sector franchise Private sector nonfranchise Alternative private sector franchise Alternative private sector nonfranchise Just-in-time Information Incentive alignment

QUALITY (TEAMS ¼ 1)

QUALITY (TEAMS ¼ 0)

TEAMS

AUTO

0.014 (0.071) 0.059 (0.077) 0.039 (0.054) 0.206 (0.103) 0.079 (0.072) 0.037 (0.078) 0.020 (0.018) 0.047 (0.016)

0.224 (0.229) 0.003 (0.168) 0.016 (0.147) 0.179 (0.186) 0.283 (0.265) 0.388 (0.213) 0.049 (0.031) 0.064 (0.070)

0.242 (0.141) 0.189 (0.099)

0.059 (0.099) 0.007 (0.094)

0.067 (0.066) 0.018 (0.013) 0.517 (0.276) 0.373 (0.077) 0.068 (0.279) 0.282 (0.094)

0.116 (0.160) 0.003 (0.034) 0.261 (1.206) 0.492 (0.259) 0.007 (0.858) 0.547 (0.281)

0.840 (1.384) 0.031 (0.152) 0.070 (0.347) 0.126 (0.162)

0.578 (0.356) 0.093 (0.114) 0.402 (0.345) 0.004 (0.111)

0.303 (0.086) 0.071 (0.027) 0.202 (0.038)

54

JED DEVARO AND FIDAN ANA KURTULUS

Table A3. (Continued ) QUALITY (TEAMS ¼ 1)

QUALITY (TEAMS ¼ 0)

TEAMS

AUTO

0.748 (0.207)

0.057 (0.032) 0.071 (0.035) 0.808 (0.292)

Decisions Work at home Constant c s

02

s

12

s

13

s

23

0.586 (0.185) 1.177 (0.069) 0.719 (0.003) 0.071 (0.391) 0.741 (0.002) 0.489 (0.045)

1.490 (0.367)

Note: Standard errors from the parametric bootstrap are in parentheses and are based on 75 bootstrap replications. Controls for industry and establishment’s largest occupational group are also included in the TEAMS and AUTO equations. The omitted firm ownership category is public sector. Sample size is 1,839. Indicates significance at the 10% level. Indicates significance at the 5% level.

THE NATURE, SCOPE AND EFFECTS OF JOINT LABOR-MANAGEMENT COMMITTEES IN JAPAN Takao Kato ABSTRACT In this paper we begin with documenting the nature and scope of a popular participatory employment practice in postwar Japan, or Joint Labor Management Committees (JLMCs), using unique survey data. We combine the survey data with corporate proxy statement data and produce firm-level micro data for over 200 Japanese firms with JLMCs which contain detailed information on various characteristics of their JLMCs. The resulting micro data are then used to estimate the effects on productivity and other outcome measures (such as profitability and labor cost) of varying attributes of Japanese JLMCs. For diverse specifications, consistently we find the positive and highly significant productivity effect of the breadth of information sharing via JLMCs. Specifically, sharing information obtained during JLMC meetings with ALL employees on one additional issue will be associated with a 3% increase in productivity. As such, our findings point to the importance of disseminating information shared via JLMCs widely to all employees. We further identify two important mechanisms through which such information Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 55–80 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09002-2

55

56

TAKAO KATO

dissemination can occur effectively, i.e., unions and employee participation/involvement at the grassroots level such as Shop Floor Committees. This constitutes yet another example of the complementary role of unions in participatory employment practices and synergy between various participatory employment practices.

1. INTRODUCTION This paper contributes to the growing literature on Human Resource Management Practices (HRMPs) by providing new evidence on the scope, nature and effects of an important feature of Japanese HRMPs – Joint LaborManagement Committees (JLMCs).1 Especially, by focusing on representative participation such as JLMCs, this paper complements the existing literature that tends to pay most attention to shopfloor-level practices (such as self-directed teams) and financial participation (such as profit sharing and ESOPs) and often neglects representative participation.2 This paper uses a unique survey of publicly traded firms in Japan, the HRM Survey of Japanese Firms and provides new evidence on the nature, scope and effects of Japanese JLMCs. The survey was administered in collaboration with Motohiro Morishima at Keio University’s Keio Economic Observatory during the summer of 1993. The sample universe of the HRM Survey of Japanese firms was the Toyo Keizai Kaisha Shiki Ho that provides a list of all firms listed in Japan’s three major stock exchanges, Tokyo, Osaka and Nagoya. In 1993 there were 2,127 firms listed in those three exchanges.3 The survey itself was preceded by a pilot phase in which an earlier version of the instrument was tested on human resource managers of several firms as well as on researchers of the Japan Institute of Labor, the Japan Productivity Center and the Japan Securities Research Institute who conducted similar yet smaller surveys separately in the past. On the basis of what we learned from this, the questionnaire was revised. The final version of the questionnaires was mailed to Director of HR/Personnel (Jinji Bucho) of all 2,127 firms using a list of addresses from the Toyo Keizai Kaisha Shiki Ho in August of 1993. The survey response rate of 17% (20% for manufacturing) is comparable to most prior surveys of similar nature in Japan. We did attempt to prompt the non-respondents by phone but our effort was largely unsuccessful.4 In the next section, we begin with documenting the nature, evolution and significance of Japanese JLMCs, using our unique survey data. We then combine the survey data with corporate proxy statement data and create enterpriselevel micro data for over 200 Japanese firms with JLMCs which contain

The Nature, Scope and Effects of Joint Labor-Management Committees

57

detailed information on the attributes of their JLMCs. The resulting micro data are used to estimate the effects on productivity and other outcome measures (such as profitability and labor cost) of varying attributes of Japanese JLMCs. A closer look at the postwar Japanese experience of participatory employment practices and their effects appears to be of particular public policy interest for many countries considering participatory employment practices a way to improve their performance and thus competitiveness. First, as Levine and Tyson (1990) suggest, relatively higher job security (often ensured by intra-firm transfers and transfers to related firms of workers) and strong group cohesiveness (supported by compression of wage and status differentials) of Japanese workers in large manufacturing firms in the postwar era point to an industrial relations system favorable to successful participation. Furthermore, stable growth, lower unemployment and stable financial corporate grouping (banks and institutional shareholders as stable, long-term suppliers of capital) in the postwar period point to an external environment conducive to flourishing participation. The economic slowdown in the 1990s (in particular the recent banking crisis) and a rapidly aging workforce have allegedly been eroding these participationfriendly environments. However, as Kato (2001, 2003) show, Japanese employment practices, such as ‘‘lifetime employment’’ and participatory employment system, appear to be much more enduring than popular rhetoric suggests. The paper is organized as follows. In the next section, we document the nature, evolution and significance of Japanese JLMCs. In Section 3 we present our regression results on the effects of varying attributes of JLMCs, followed by a concluding section.

2. THE SCOPE, NATURE AND DIFFUSION OF JLMCS IN JAPAN One of the core mechanisms for industrial relations of large Japanese firms is JLMCs. Established at the top level (corporate and/or establishment level) and involving both management and labor representatives, JLMCs serve as a mechanism for employee participation and involvement at the top level on a large variety of issues ranging from basic business policies to working conditions. Many observers attribute the peaceful enterprise-level labormanagement relations observed in Japanese companies to the establishment of JLMCs (Shimada, 1992; Inagami, 1988; Kato & Morishima, 2002). According to the survey, in 1950 only about 20% (30% for manufacturing) of publicly traded firms in Japan had JLMCs. During the next two

58

TAKAO KATO

decades, the institution diffused rapidly. Thus, by 1970 the figure had risen to close to 60% (70% for manufacturing). For the next two decades the institution diffused steadily, and, as of 1993, fully 80% of all firms (nearly 90% for manufacturing) reported to have standing JLMCs. Table 1 compares the incidence of JLMCs between firms with varying characteristics. First, JLMCs are found to be more popular among manufacturing firms than non-manufacturing (90% vs. 67%). Second, JLMCs are far less popular among non-union firms (only 20% of non-union firms have JLMCs). These differences are found to be statistically significant at the 1% level. To see if JLMCs tend to be adopted along with other participatory employment practices, we calculate the proportion firms with JLMCs for those with and without such practices. We consider the following four practices: (i) Shop-Floor Committees (SFCs) in which supervisors and employees on shop floor discuss issues such as shop-floor operations and shop-floor environments; (ii) Small Group Activities (SGA) or activities such as quality control (QC) circles and Zero Defects in which small groups at the workplace level Table 1.

Incidence of JLMCs in 1993. Number of Firms

% Firms with JLMCs

All

349

81.66

Manufacturing Non-manufacturing

218 131

90.37 67.18

With unions Without unions

293 55

93.52 20.00

With shop-floor committees Without shop-floor committees

163 184

88.34 75.54

With small group activities Without small group activities

255 92

89.02 60.87

With employee stock ownership plans Without employee stock ownership plans

338 11

81.95 72.73

With profit sharing plans Without profit sharing plans

92 253

80.43 82.21

Source: HRM Survey of Japanese Firms. Statistically significant at the 1% level.

59

The Nature, Scope and Effects of Joint Labor-Management Committees

voluntarily set plans and goals concerning operations and work together toward accomplishing these plans and goals; (iii) Employee Stock Ownership Plans (ESOPs) or a policy through which the firm forms an ESOP trust (Mochikabu Kai) consisting of its non-executive employees and promotes ownership of its own shares by the trust; and (iv) Profit Sharing Plans (PSPs) or a pay system in which the total amount of bonuses are linked to a measure of firm performance, such as profit.5 The table shows that JLMCs are more prevalent among firms with SFCs and SGAs (or employee participation at the grass roots) than other firms. Both differences are statistically significant at the 1% level. The table also shows that JLMCs are more popular among firms with ESOPs than other firms although the difference turns out to be not statistically significant at the 10% level. Lastly, we find no statistically significant difference in the incidence of JLMCs between firms with and without PSPs. In Table 2, for indicators such as average labor force, value added,6 capital/labor ratio, labor productivity, ratio of labor cost to value added, per capita labor cost, Return On Asset (ROA) and profit margin, we have calculated averages for firms with and without JLMCs for the accounting year 1992. An interesting profile of the average firm with a JLMC emerges.

Table 2.

Characteristics of Firms with and without JLMCs.

Number of firms Labor( ¼ number of employees) Value added, in millions of yen Capital/labor ratio( ¼ fixed assets, in millions of yen/labor) Productivity( ¼ value added/labor) Ratio of labor cost to value added ( ¼ Total labor cost, in millions of yen/value added) Per capita labor cost ( ¼ Total labor cost, in millions of yen/labor) ROA( ¼ after-tax profit, in millions of yen/total asssets, in millions of yen) MARGIN( ¼ after-tax profit, In millions of yen/sales, in millions of yen)

Firms with JLMCs

Firms without JLMCs

285 3076.1 27401.93 19.64

64 1186.06 9514.66 37.82

8.14 0.49

12.62 0.31

3.90

2.58

0.02

0.02

0.02

0.03

Source: HRM Survey of Japanese Firms and Nikkei Needs. Statistically significant at the 1% level.

60

TAKAO KATO

In comparison with non-JLMC firms, JLMC firms have larger labor forces (3076: 1186), higher value added (nearly three times higher), higher ratio of labor cost to value added (50%: 30%), and greater per capita labor cost (by 50%). These differences are statistically significant at the 1% level. On the other hand, we failed to find any statistically significant differences in capital/labor ratio, labor productivity, ROA and profit margin between firms with and without JLMCs. Within JLMCs, a number of issues are discussed, ranging from basic business policies to working conditions to social and athletic activities sponsored by the firm. Table 3 shows the proportion of firms with JLMCs that discuss each of the 16 issues. As expected, over 90% of JLMCs discuss directly labor-related issues such as layoff and workforce reduction; promotion/ demotion and other changes in employee status; working hours/holidays/time off; safety and health issues; wage and bonus; and fringe benefits. Somewhat surprisingly only about 60% of JLMCs discuss hiring and staffing plans; and training and development. Among more business-related issues (D1–D4), corporate restructuring is the most often discussed item (82%). About 75% of JLMCs discuss basic business strategies; and sales and production plans. 66% of JLMCs discuss introduction of new technology/equipment. Such businessrelated issues discussed in JLMCs can be confidential. According to field research conducted at a variety of Japanese firms, labor representatives to JLMCs believe that some of the information they receive from top management can be considered ‘‘insider information.’’7 Table 3 also shows that JLMCs in manufacturing are more likely to discuss various issues, in particular business-related issues (D1–D4) than those in non-manufacturing, and the differences are statistically significant at least at the 5% level. Table 4 shows the proportion of firms with JLMCs that share information on various issues with ALL employees. Information on corporate restructuring turns out to be the most widely shared. Thus, over 80% of firms with JLMCs share information on corporate restructuring with all employees.8 Major HRM changes, such as wages, training, promotion and sales/profit projections are the second most widely shared items. Plans for transfers and layoffs of employees turn out to be the least widely shared. The table also shows that sales/profit predictions and plans for transfers and layoffs are more widely shared in manufacturing firms than other firms, and the differences are statistically significant at least at the 10% level. Table 5 shows summary characteristics of JLMCs. The average JLMC meets 10.7 times a year; discusses 12.7 issues out of 16 issues; and shares information on 4.7 issues out of 7 issues with ALL employees. There is one statistically significant difference in the summary attributes between

The Nature, Scope and Effects of Joint Labor-Management Committees

Table 3.

61

Proportion of Firms with JLMCs that Discuss Various Issues. % Firms with JLMCs that Discuss each of the following 16 issues

D1: Basic business strategies D2: Sales and production plans D3: Corporate restructuring D4: Introduction of new technology/ equipment D5: Hiring and staffing plans D6: Transfers of employees D7: Layoff and workforce reduction D8: Promotion/ demotion and other changes in employee status D9: Working hours/ holidays/time off D10: Safety and health issues D11: Mandatory retirement D12: Wage and bonus D13: Separation pay/ pension plans D14: Training and development D15: Fringe benefits D16: Cultural activities/sports

All

N

Manufacturing

N

Nonmanufacturing

N

71.06

273

74.87

191

62.2

82

69.85

272

76.32

190

54.88

82

77.54

276

82.11

190

67.44

86

62.41

266

66.49

185

53.09

81

53.48

273

57.37

190

44.58

83

71.01

276

75.66

189

60.92

87

88.81

268

91.4

186

82.93

82

92.34

274

92.59

189

91.76

85

93.45

275

94.21

190

91.76

85

87.50

272

91.53

189

78.31

83

89.61

279

88.6

193

91.86

86

90.68 89.89

279 277

91.24 89.53

194 191

89.41 90.7

85 86

59.41

271

61.05

190

55.56

81

91.51 69.66

271 267

93.65 73.4

189 188

86.59 60.76

82 79

Source: HRM Survey of Japanese Firms. Statistically significant at the 1% level. Statistically significant at the 5% level. Statistically significant at the 10% level.

62

TAKAO KATO

Table 4.

Proportion of Firms with JLMCs that Share Information on Various Issues with All Employees. % Firms with JLMCs that Share Information on One of the following 7 Issues with all Employees All

S1: Sales/profit predictions S2: Basic business Strategies S3: Sales and production plans (for example, introduction of new products) S4: Corporate restructuring S5: Introduction of new technology/ equipment S6: Plans for transfers and layoffs of employees S7: Major HRM changes, (wages, training, promotion)

N

Manufacturing

N

Non-manufacturing

N

191

57.83

83

65.69

274

69.11

68.00

275

65.97

191

72.62

84

64.04

267

61,.38

189

70.51

78

81.02

274

82.54

189

77.65

85

65.66

265

63.44

186

70.89

79

52.38

273

57.67

189

40.48

84

69.63

270

70.59

187

67.47

83

Source: HRM Survey of Japanese Firms. Statistically significant at the 1% level. Statistically significant at the 10% level.

manufacturing and non-manufacturing. The average JLMC in manufacturing discusses 13.1 issues whereas the one in non-manufacturing discusses only 11.8 issues.

3. THE PRODUCTIVITY EFFECTS OF VARYING ATTRIBUTES OF JLMCS On the whole, economic theory is ambiguous as to the predicted effect of HRMPs on enterprise performance, such as productivity. (For reviews, see

63

The Nature, Scope and Effects of Joint Labor-Management Committees

Table 5.

Summary Attributes of JLMCs. All Mean S.D. N

Manufacturing Mean

Non-manufacturing

S.D. N

Mean

S.D.

N

FREQ ¼ Number of JLMC meetings per year

10.68 9.46 281 11.08

9.67 193

9.81

8.98

88

DISCUSS ¼ Total number of items (D1–D16) discussed in JLMCs

12.71 3.56 238 13.05 3.34 170

11.84

3.95

68

4.51

2.07

74

SHARE ¼ Total number of items (S1–S7) that the firm shares with all employees

4.65 2.16 250

4.71

2.21 176

Source: HRM Survey of Japanese Firms. Statistically significant at the 5% level

the essays in Blinder, 1990; Gibbons, 1997). Focusing on individual motivation and performance, however, there are two major hypotheses that predict positive performance effects, the goal alignment hypothesis and the human capital hypothesis. First, JLMCs are expected to reduce information asymmetry between labor and management and, consequently, prevent the development of adversarial labor-management relations. In industrial relations, employers tend to have more information about the status of the company and business strategies. Employees, under typical collective bargaining arrangements, have no means of obtaining such information except for resorting to hard bargaining often accompanied with the threat of strikes (Tracy, 1992). Such behavior on the part of the unions and employees may lead to adversarial labor-management relations, which may, in turn, have negative consequences for productivity. Voluntary sharing of information by management, through JLMCs, is apt to lower the cost of such information asymmetry and is likely to have positive productivity effects. Second, management may voluntarily share information with labor to enhance worker loyalty and commitment (Kleiner & Bouillon, 1991). Worker cooperation may also be obtained through higher worker commitment and loyalty. Enhanced worker loyalty and cooperative behavior are all expected to have positive productivity effects. In other words, sharing information on private information which has been heretofore restricted to owners and top management is likely to result in goal alignment and trust between labor and management. Better informed via JLMCs, workers, while

64

TAKAO KATO

still striving for their own benefit, may be more likely to be persuaded that it is in their interests to cooperate with management and improve firm performance. They may see more clearly the link between their own behavior and enlargement of the benefits through firm prosperity. In addition, information sharing via JLMCs is likely to restrict management’s opportunistic behavior and enhance the level of trust that workers have for management. In most business organizations, the economic game is repeated. In such repeated game where the interdependence between labor and management is apt to continue in the future, provision of private, business information is likely to enable labor to detect management’s deception and limit opportunistic behavior. Furthermore, workers are more apt to develop trust in management who shares information voluntarily. In short, by preventing the negative consequences of management’s moral hazard and raising the positive effects of labor’s cooperative behavior, JLMCs are likely to have favorable effects on productivity. Second, according to the human capital hypothesis, JLMCs may play an important role of providing labor with an effective voice in the firm and hence reduce the costs of exit from the firm, saving specific human capital.9 In the absence of unions, JLMCs may well be the sole voice mechanism, while in the presence of unions they may supplement the direct voice mechanism of unions. To discern the productivity effects of varying attributes of JLMCs, we estimate: lnðProductivityÞ ¼ a þ bH þ g lnðKLRÞ þ dEMPAGE þ industry dummy variables þ u

ð1Þ

where Productivity is average productivity of labor (value added/employees), H a vector of variables representing the effects of varying attributes of JLMCs; KLR the capital/labor ratio, EMPAGE the average age of employees, and a an intercept, and b (vector) and g the slope coefficients. We use a constant returns, OLS specification since prior studies which estimated production functions using cross-sectional data of Japanese firms tend to point to the appropriateness of the constant returns specification10 and OLS estimates are generally more robust against specification errors than simultaneous equation methods. For the disturbance term, u, we assume u NIDð0; s2 Þ: For H we consider three attributes of JLMCs: (i) the frequency of JLMC meetings; (ii) the scope of issues discussed in JLMCs; and (iii) the breadth of information sharing (or how broadly information is shared among employees). These three attributes can be considered potentially important ingredients for successful and high-impact JLMCs which, as explained above,

The Nature, Scope and Effects of Joint Labor-Management Committees

65

have greater positive goal alignment and human capital effects on firm performance. For example, when JLMCs meet frequently, discuss a wide variety of issues ranging from directly labor-related issues to more businessrelated issues, and are broad-based (information shared via JLMCs is widely disseminated among regular employees), information asymmetry between labor and management may be more effectively reduced. Specifically, we use FREQ (number of JLMC meetings per year) to measure the first attribute of JLMCs. For the second attribute (the scope of issues discussed in JLMCs), as shown in Table 6, we consider our summary index, DISCUSS (total number of items discussed in JLMCs) as well as individual dummy variables (DIS1–DIS16). Likewise, for the third attribute (the breadth of information sharing), we use our summary index, SHARE (total number of issues on which information is shared with all employees) as well as individual dummy variables (SHA1–SHA7). In addition to ln(KLR) which is a standard independent variable in this sort of productivity function estimation, we include average age of employees EMPAGE to control for possible differences in labor quality between firms. We expect firms with more experienced labor force to be more productive than firms with less-experienced labor force. Furthermore, to control for industry-specific differences in productivity, we include 17 industry dummy variables. Ideally, we want to include more controls to account for other firm characteristics that might affect productivity, such as managerial ability. Unfortunately, the data do not include such information. Alternatively, one could collect the data for multiple years and estimate fixed effect models to account for unobserved and time-invariant firm heterogeneity. Since data on varying attributes of JLMCs are available only for 1992, such fixed effect models cannot be estimated. We estimate Eq. (1) by using new firm-level micro data containing over 200 publicly traded firms in Japan which have JLMCs and provide detailed information on the afore-mentioned three attributes of their JLMCs. This data set was assembled by merging two data bases. The data on JLMCs were from our HRM Survey of Japanese Firms. The data on Q, K and L were compiled from the Nikkei financial data tapes, Nikkei Needs. Tables 7a and b summarize the OLS results. For all specifications, the estimated coefficients on ln(KLR) and EMPAGE are positive and significant at the 1% level, confirming our prior expectations (firms with higher capital/ labor rations and more experienced labor force have higher labor productivities). In Specification (1) we consider FREQ as well as two summary indices, SHARE and DISCUSS for H. The estimated coefficient on SHARE is positive and significant at the 5% level while the estimated coefficients on

66

TAKAO KATO

Table 6.

Summary Statistics.

Variable

Mean

S.D.

ln(Productivity) ROA MARGIN Ratio of labor cost to value added ¼ RLCVA Per capita labor cost ln (capital/labor ratio) ¼ ln(KLR) EMPAGE ¼ average age of employees FREQ DIS1 ¼ 1 if JLMC discusses D1, 0 otherwise DIS2 ¼ 1 if JLMC discusses D2, 0 otherwise DIS3 ¼ 1 if JLMC discusses D3, 0 otherwise DIS4 ¼ 1 if JLMC discusses D4, 0 otherwise DIS5 ¼ 1 if JLMC discusses D5, 0 otherwise DIS6 ¼ 1 if JLMC discusses D6, 0 otherwise DIS7 ¼ 1 if JLMC discusses D7, 0 otherwise DIS8 ¼ 1 if JLMC discusses D8, 0 otherwise DIS9 ¼ 1 if JLMC discusses D9, 0 otherwise DIS10 ¼ 1 if JLMC discusses D10, 0 otherwise DIS11 ¼ 1 if JLMC discusses D11, 0 otherwise DIS12 ¼ 1 if JLMC discusses D12, 0 otherwise DIS13 ¼ 1 if JLMC discusses D13, 0 otherwise DIS14 ¼ 1 if JLMC discusses D14, 0 otherwise DIS15 ¼ 1 if JLMC discusses D15, 0 otherwise DIS16 ¼ 1 if JLMC discusses D16, 0 otherwise SHA1 ¼ 1 if the firm shares S1 with all employees, 0 otherwise SHA2 ¼ 1 if the firm shares S2 with all employees, 0 otherwise SHA3 ¼ 1 if the firm shares S3 with all employees, 0 otherwise SHA4 ¼ 1 if the firm shares S4 with all employees, 0 otherwise SHA5 ¼ 1 if the firm shares S5 with all employees, 0 otherwise SHA6 ¼ 1 if the firm shares S6 with all employees, 0 otherwise SHA7 ¼ 1 if the firm shares S7 with all employees, 0 otherwise DISCUSS ¼ DIS1+DIS2+DIS3+DIS4+DIS5+?+DIS14+DIS15+DIS16 SHARE ¼ SHA1+SHA2+SHA3+SHA4+SHA5+SHA6+SHA7 FISH ¼ 1 if the firm is in fishing/mining, 0 otherwise CONS ¼ 1 if the firm is in construction, 0 otherwise FOOD ¼ 1 if the firm is in food, 0 otherwise TEXT ¼ 1 if the firm is in textile, 0 otherwise CHEM ¼ 1 if the firm is in chemicals, 0 otherwise. STEE ¼ 1 if the firm is in steel, 0 otherwise META ¼ 1 if the firm is in metal, 0 otherwise NONF ¼ 1 if the firm is in nonferrous metal, 0 otherwise STON ¼ 1 if the firm is in stone, 0 otherwise MACH ¼ 1 if the firm is in machinery, 0 otherwise ELEC ¼ 1 if the firm is in elec. machinery, 0 otherwise TRAN ¼ 1 if the firm is in transp. equip., 0 otherwise PREC ¼ 1 if the firm is in precision, 0 otherwise OTHM ¼ 1 if the firm is in other manufacturing industries, 0 otherwise TRAD ¼ 1 if the firm is in retail and wholesale trade, 0 otherwise UTIL ¼ 1 if the firm is in electric/gas, communication and transportation, 0 otherwise SERV ¼ 1 if the firm is in service, 0 otherwise OTHN ¼ 1 if the firm is in other non-manufacturing industries, 0 otherwise

1.94 0.02 0.02 0.49 3.90 2.55 37.73 10.68 0.71 0.70 0.78 0.62 0.53 0.71 0.89 0.92 0.93 0.88 0.90 0.91 0.90 0.59 0.92 0.70 0.66 0.68 0.64 0.81 0.66 0.52 0.70 12.71 4.65 0.01 0.08 0.04 0.04 0.16 0.02 0.04 0.02 0.03 0.09 0.13 0.06 0.02 0.04 0.10 0.06 0.06 0.01

0.69 0.03 0.04 0.35 2.33 0.83 3.39 9.46 0.45 0.46 0.42 0.49 0.50 0.45 0.32 0.27 0.25 0.33 0.31 0.29 0.30 0.49 0.28 0.46 0.48 0.47 0.48 0.39 0.48 0.50 0.46 3.56 2.16 0.08 0.27 0.20 0.20 0.37 0.14 0.18 0.16 0.18 0.29 0.33 0.23 0.16 0.19 0.30 0.24 0.23 0.08

Source: HRM Survey of Japanese Firms and Nikkei Needs.

OLS Estimates on the Effects on Productivity of Varying Attributes of JLMCs (Dependent Variable: ln(Productivity)).

(1) Independent Parameter Variable Estimate

(2) Parameter Estimate

(3) Parameter Estimate

(4) Parameter Estimate

(5) Parameter Estimate

0.288 (5.58) 0.033 (2.69) 0.006 (1.61) 0.036 (1.95) 0.004 (0.39)

0.293 (5.86) 0.044 (3.79) 0.006 (1.74) 0.039 (2.34)

0.299 (5.99) 0.045 (3.87) 0.006 (1.66) 0.038 (2.25)

0.283 (5.74) 0.044 (3.78) 0.006 (1.54) 0.041 (2.47)

0.278 (5.64) 0.036 (3.08) 0.007 (1.90) 0.039 (2.31)

ln(KLR) EMPAGE FREQ SHARE DISCUSS DIS1 DIS2 DIS3 DIS4 DIS5

(6) Parameter Estimate 0.264 (5.47) 0.038 (3.23) 0.005 (1.49) 0.044 (2.73)

0.094 (1.21) 0.103 (1.31) 0.134 (1.58)

(7) Parameter Estimate

(8) Parameter Estimate

(9) Parameter Estimate

(10) Parameter Estimate

(11) Parameter Estimate

0.280 (5.64) 0.043 (3.70) 0.007 (1.84) 0.037 (2.25)

0.269 (5.56) 0.037 (3.19) 0.007 (1.98) 0.044 (2.67)

0.275 (5.62) 0.037 (3.18) 0.006 (1.59) 0.042 (2.61)

0.272 (5.61) 0.037 (3.21) 0.006 (1.62) 0.043 (2.65)

0.272 (5.52) 0.041 (3.53) 0.007 (1.78) 0.037 (2.27)

(12) Parameter Estimate 0.273 (5.68) 0.037 (3.25) 0,006 (1.61) 0.043 (2.72)

The Nature, Scope and Effects of Joint Labor-Management Committees

Table 7a.

0.031 (0.42)

67

0.099 (140)

(1) Independent Parameter Variable Estimate

(2) Parameter Estimate

(3) Parameter Estimate

(4) Parameter Estimate

(5) Parameter Estimate

(6) Parameter Estimate

DIS6

(7) Parameter Estimate

68

Table 7a. (Continued ) (8) Parameter Estimate

(9) Parameter Estimate

(10) Parameter Estimate

(11) Parameter Estimate

0.016 (0.20)

DIS7

0.053 (0.49)

DIS8

0.174 (1.29)

DIS9

0.149 (1.01)

DIS10

0.000 0.00

DIS11 R-Square Sample size

(12) Parameter Estimate

0.505 211

0.477 235

0.485 234

0.474 238

0.480 230

0.480 236

0.468 236

0.478 234

0.479 235

0.478 235

0.470 237

0.138 (1.17) 0.480 239

Source: HRM Survey of Japanese Firms and Nikkei Needs. Note: Absolute values of t statistics are in parentheses. All models include a contant term and industry dummy variables listed in Table 6 except for OTHM which serves as a reference industry. Statistically significant at the 10% level. Statistically significant at the 5% level. Statistically significant at the 1% level.

TAKAO KATO

OLS Estimates on the Effects on Productivity of Varying Attributes of JLMCs (Dependent Variable: ln(Productivity)).

(13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) Independent Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Variable Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate ln(KLR) EMPAGE FREQ SHARE DIS12 DIS13 DIS14 DIS15 DIS16

0.273 (5.68) 0.038 (3.34) 0.006 (1.67) 0.043 (2.69) 0.128 (1.06)

0.272 (5.65) 0.038 (3.27) 0.006 (1.62) 0.043 (2.69)

0.272 (5.59) 0.037 (3.25) 0.006 (1.63) 0.043 (2.65)

0.273 (5.57) 0.037 (3.18) 0.008 (2.08) 0.039 (2.41)

0.261 (5.35) 0.034 (2.89) 0.006 (1.66) 0.037 (2.25)

0.111 (0.95) 0.070 (1.00)

0.287 (6.22) 0.045 (4.28) 0.003 (0.95)

0.308 (6.35) 0.048 (4.45) 0.004 (1.08)

0.284 (5.79) 0.040 (3.69) 0.004 (1.21)

0.312 (6.55) 0.047 (4.34) 0.003 (0.89)

0.298 (6.30) 0.048 (4.49) 0.005 (1.56)

0.305 (6.51) 0.048 (4.57) 0.004 (1.28)

0.283 (5.70) 0.045 (4.08) 0.004 (1.15)

The Nature, Scope and Effects of Joint Labor-Management Committees

Table 7b.

0.000 0.00

69

0.030 (0.41)

70

Table 7b. (Continued ) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) Independent Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Parameter Variable Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate Estimate 0.188 (2.75)

SHA1

0.152 (2.12)

SHA2 SHA3

0.032 (0.46)

SHA4

0.081 (0.94) 0.148 (2.09)

SHA5

0.109 (1.65)

SHA6 SHA7 R-Square 0.479 Sample size 240

0.478 238

0.479 236

0.478 234

0.487 232

0.479 267

0.470 268

0.462 260

0.462 267

0.482 258

0.479 266

0.079 (1.08) 0.464 262

Note See footnote of Table 7a Statistically significant at the 10% level. Statistically significant at the 5% level. Statistically significant at the 1% level.

TAKAO KATO

The Nature, Scope and Effects of Joint Labor-Management Committees

71

FREQ and DISCUSS are not statistically significant at the 10% level. The sign and size of the estimated coefficient on SHARE suggest that sharing information with all employees on one more issue will result in a 3% increase in labor productivity.11 Our summary index of the scope of issues discussed in JLMCs, DISCUSS turns out to have no statistically significant link to productivity. It is still possible, however, that productivity may be related to whether JLMCs discuss a specific issue. To see if this is the case, we consider individual dummy variables, DIS1–DIS16 as an alternative to DISCUSS. The results are summarized in Specifications (2)–(17). As shown in Tables 7a and b, we find no statistically significant effect on productivity of any of those individual dummy variables. On the other hand, the estimated coefficients on SHARE in all of these specifications are positive and significant at least at the 5% level and often at the 1% level. The size of the estimated coefficients on SHARE vary from 0.3 to 0.4, implying a 3–4% productivity gain from sharing information widely with all employees on one more issue. In Specifications (18)–(24) in Table 7b, we consider individual dummy variables, SHA1–SHA7 as an alternative to SHARE to see if it is particularly beneficial to share information widely with all employees on a specific issue. We find that the estimated coefficients on SHA1, SHA2 and SHA5 are positive and statistically significant at least at the 5% level (at the 1% level for SHA1), suggesting a particular benefit of sharing information on sales/ profit predictions, basic business strategies and the introduction of new technology and new equipment widely with all employees. The size of the estimates imply that firms sharing information on sales/profit predictions widely with all employees enjoy over 19% higher productivity than other firms and that firms sharing information on basic business strategies as well as on the introduction of new technology and new equipment widely with all employees enjoy about 15% higher productivity than other firms, respectively. Note that we enter these seven individual dummy variables one at a time and do not consider them together simultaneously. Therefore, for example, we ought not to interpret the estimated coefficient on SHA1 that the firm sharing no information with all employees on any issue will gain a 17% productivity gain by simply sharing information with all employees on one issue (sales/profit predictions). It turns out that over 80% of firms sharing information on sales/profit predictions with all employees are also sharing information with all employees on S2, S3 and S4. The estimated coefficients on FREQ are always negative, small and often statistically insignificant. Specifically, for 14 out of 24 specifications, they are not statistically significant at the 10% level and even for the remaining

72

TAKAO KATO

10 specifications, they are often significant only at the 10% level. These results on FREQ are perhaps not too surprising. Once the scope of issues discussed in JLMCs and the breadth of information sharing are controlled for, more JLMC meetings will add very little to the reduction of information asymmetry and the enhancement of worker loyalty. The direct and indirect cost of holding more JLMC meetings (such as the opportunity cost of representatives to JLMCs) may even cause productivity to fall somewhat. In short, we find the breadth of information sharing to be an important ingredient for successful and high-impact JLMCs with significant positive effects on productivity. To investigate further what makes JLMCs more broad-based and hence successful, we contrast mean SHARE between unionized and non-union firms; and between firms with and without other participatory employment practices. The average unionized firm turns out to share information on 4.7 issues with all employees whereas the average non-union firm turn out to share information only on 2.7 issues with all employees (the difference is statistically significant at the 1% level). Firms with SFCs (Shop-Floor Committees: supervisors and employees on shop floor discuss issues such as shop-floor operations and shop-floor environments), on average, are found to share information on 5.08 issues with all employees while firms without SFCs, on average are found to share information only on 4.24 issues with all employees (again the difference is statistically significant at the 1% level). We also conducted similar comparison for firms with and without SGAs (Small Group Activities); firms with and without ESOPs (Employee Stock Ownership Plans); and firms with and without PSPs (Profit Sharing Plans). We find no statistically significant difference. Field research also points to the complementary role of unions and SFCs in making information sharing via JLMCs more broad-based. First, unions use formal (e.g., union newsletters, union meetings) and informal mechanisms to disseminate information shared during JLMC meetings widely to the general membership (which means all regular full-time employees in large Japanese firms). Union representatives to JLMC are sometimes asked even by middle-level managers to share information obtained via JLMC with them. We ought not to underestimate the importance of unions as an effective mechanism to disseminate information to regular employees. Second, SFCs of Japanese firms facilitate sharing of information concerning not only standard shop-floor issues but business and corporate strategic plans which are discussed during the JLMC meetings. It is not too surprising that firms equipped with both JLMCS and SFCs are able to share more information widely with all employees than firms with only JLMCs.12

The Nature, Scope and Effects of Joint Labor-Management Committees

73

Our findings shed light on the recent literature on the effects of HRMPs that often finds evidence for organizational complementarity, i.e., some HRMPs may be more effective when used in combination with other HRMPs.13 Specifically, Kato and Morishima (2002) find evidence for a complementarity between JLMCs and SFCs in Japan. Our evidence suggests a specific mechanism through which this complementarity arises. Or one of several key functions of SFCs is, as Kato (2003) suggests, to disseminate information shared with a small group of labor representatives during JLMC meetings widely to regular employees. According to our findings, SFCs tend to make JLMCs more effective and successful, and therefore more productivity-enhancing by facilitating the wide dissemination of information shared during JLMC meetings to the rank and file.

4. OTHER OUTCOME EFFECTS To see if varying attributes of JLMCs have impact on other outcomes, we run similar regressions with alternative outcome variables. Specifically, we consider: (i) ROA (profit/total asset); (ii) profit margin (profit/sales); (iii) ratio of labor cost to value added (total labor cost/value added); and (iv) per capita labor cost (labor cost/employees). ROA and profit margin (MARGIN) are standard accounting measures of profitability and ratio of labor cost to value added (RLCVA) is used by prior studies on the effects of participatory employment practices in Japan.14 Most recently, some researchers are shifting their focus from the effects on firm performance of participatory employment practices to the effects on worker outcomes (such as job satisfaction and compensation) of such practices.15 Our data allow us to use Per capita labor cost as a proxy for total compensation for the average worker. Table 8 summarizes the OLS estimates for the effects on other outcome measures. It turns out that we fail to find any statistically significant effect on these outcomes of the three attributes of JLMCs.

5. CONCLUDING REMARKS In this paper, we have documented the nature, evolution and significance of Japanese JLMCs, using our unique survey data. We have combined the survey data with corporate proxy statement data and have produced firmlevel micro data for over 200 Japanese firms with JLMCs which contain detailed information on a variety of characteristics of their JLMCs. The

74

Table 8.

OLS Estimates on the Effects on Other Outcomes of Varying Attributes of JLMCs.

Independent Variable

ln(KLR) EMPAGE FREQ SHARE DISCUSS R Square Sample size

Dependent Variable ROA*100

ROA*100

MARGIN*100

MARGIN*100

RLCVA*100

RLCVA*100

Per Capita Labor Cost

Per Capita Labor Cost

Parameter Estimate

Parameter Estimate

Parameter Estimate

Parameter Estimate

Parameter Estimate

Parameter Estimate

Parameter Estimate

Parameter Estimate

8.533 (2.61) 1.122 (1.44) 0.381 (1.52) 0.876 (0.76) 0.366 (0.53) 0.320 213

8.992 (3.11) 1.243 (1.85) 0.278 (1.31) 0.649 (0.69)

0.316 (1.95) 0.189 (4.91) 0.005 (0.37) 0.055 (0.97) 0.034 (0.98) 0.475 213

0.212 (1.20) 0.224 (5.44) 0.012 (0.96) 0.072 (1.25)

0.375 (1.39) 0.101 (1.58) 0.001 (0.04) 0.096 (1.01) 0.014 (0.24) 0.146 213

0.397 (1.62) 0.080 (1.39) 0.008 (0.44) 0.069 (0.86)

0.130 246

0.181 (0.52) 0.106 (1.28) 0.009 (0.35) 0.130 (1.05) 0.035 (0.47) 0.293 213

0.275 246

0.329 246

0.402 246

TAKAO KATO

Source: HRM Survey of Japanese Firms and Nikkei Needs. Statistically significant at the 10% level. Statistically significant at the 5% level. Statistically significant at the 1% level.

0.208 (0.66) 0.096 (1.29) 0.003 (0.13) 0.090 (0.87)

The Nature, Scope and Effects of Joint Labor-Management Committees

75

resulting micro data have been used to estimate the effects on productivity and other outcome measures of varying attributes of Japanese JLMCs. The literature on the effects of HRMPs tends to focus on the effects of the incidence of such practices and research on the effects of varying characteristics of such practices is limited. We have contributed to the literature on participatory employment practices by providing new systematic evidence on the effects of varying attributes of Japanese JLMCs. First, for all specifications we have found statistically significant impact on productivity of the breadth of information sharing. Specifically, sharing information with all employees on one additional issue will result in a 3% increase in productivity. Second, for all specifications we have found no statistically significant effect on productivity of the scope of issues discussed in JLMCs. Third, we have found a relatively weak link between productivity and the frequency of JLMC meetings. Overall, the results point to the importance of disseminating information shared via JLMCs to all employees. We have identified two important mechanisms through which such information dissemination can occur effectively, i.e., unions and employee participation at the shopfloor level (e.g., SFCs). As such, this constitutes yet another example of the complementary role of unions in participatory employment practices and synergy among various participatory employment practices. Lastly, we have found no evidence for any statistically significant effect of varying attributes of JLMCs on other outcome measures (profitability and labor cost). In recent years, with Japan’s prolonged economic slowdown, the popular rhetoric within Japan as well as outside of Japan has been shifting and has become less positive about traditional Japanese management with particular emphasis on employee participation and involvement (some even suggest the replacement of the participatory system with the Anglo-American-style shareholder model of corporate governance and active external labor market). Our findings point to the overall positive productivity effects of JLMCs (a key ingredient element of the participatory employment system of Japanese firms), and thus provide a caution against any hasty attempts to dismantle Japan’s once celebrated system of employee participation and involvement.

NOTES 1. We are currently witnessing an impressive accumulation of systematic evidence on the outcome effects of such practices in the U.S. Earlier studies focused on their effects on firm performance (including productivity and profitability). See, for

76

TAKAO KATO

example, in the economics literature, Ichniowski, Shaw, and Prennushi (1997), Helper (1998), Batt (1999), Freeman and Kleiner (2000), Freeman, Kleiner, and Ostroff (2000), Bartel (2004), Black and Lynch (2001, 2004), Cappelli and Neumark (2001), and articles featured in a special issue of Industrial Relations edited by Ichniowski, Casey; Thomas A. Kochan, David I. Levine, Craig Olson and George Strauss (Vol. 35, July 1996). Most recently the literature also considers their effects on outcomes for workers (such as job satisfaction, wages and motivation). See, for example, Appelbaum, Bailey, Berg, and Kalleberg (2000), Blasi and Kruse (1995), Freeman et al. (2000), Bailey, Berg, and Sandy (2001), Hamilton, Nickerson, and Owan (2003), Bauer and Bender (2001), Whitfield (2000), Osterman (2000) and articles featured in a special issue of Industrial Relations edited by Michael J. Handel and David I. Levine (Vol. 43, January 2004). However, such evidence is still relatively limited elsewhere. See, for example, Jones and Kato (1995) and Kato and Morishima (2002) for Japan; Leoni, Cristini, Labory, and Gaj (2001) for Italy; Addison and Belfield (2000) and Conyon and Freeman (2001) for the U.K.; Eriksson (2003) for Denmark; and Bayo-Moriones, Galilea-Salvatierra, Merino-Dı´ az de Cerio (2003) for Spain. 2. Notable exceptions are Morishima (1991a, b) that use firm-level micro data to find the statistically significant positive correlations between the extent of JLMCs and productivity, and the statistically significant correlations between stronger JLMCs and shorter and smoother wage negotiation. This paper will build on Morishima’s earlier work and use more recent and larger data which provide important additional information on JLMCs, i.e., how broadly information is shared. 3. Our sample universe is virtually all listed firms in Japan. The only listed firms not included in the sample universe are a very small number of firms listed only in other local stock exchanges (about three dozens of relatively small local firms). 4. We examined the representativeness of our sample by comparing the incidence of practices between our survey and other surveys and found no particular bias. For more detailed information on the survey, see Kato and Morishima (2002). 5. For Japanese SFCs and SGAs, see for example Kato (2003); for Japanese PSPs, see for example Freeman and Weitzman (1987), Ohashi (1989), Hashimoto (1990), Nakamura and Nakamura (1991), Brunello (1991), Hart and Kawasaki (1995), Ohkusa and Ohtake (1997); and for Japanese ESOPs, see for example Jones and Kato (1995). 6. The data on value added as a proxy for Q were created by adding labor costs and depreciation to operational profit. A similar procedure was used to calculate value added for earlier years by the Oriental Economist (Toyokeizai Shinpo Sha) whose value added data were widely used by both scholars and practitioners (Jones and Kato, 1995 used these value added data for their study of Japanese ESOPs during earlier years). 7. When there is a union, labor-side representatives are almost always union representatives, while even in the absence of unions, the majority of labor-side JLMC members are elected by employee vote. See Kato, (2003) for details. 8. According to the afore-mentioned field research, such information on corporate restructuring is often shared with labor representatives to JLMCs prior to public announcement (Kato, 2003). 9. In the context of trade unions, the argument was first developed by Freeman (1976).

The Nature, Scope and Effects of Joint Labor-Management Committees

77

10. For example, Jones and Kato (1993)’s cross-sectional estimates of CobbDouglass production functions using Japanese data suggest that Japanese firms operate with approximately constant returns to scale. 11. It is possible that the estimated 3% productivity gain may be capturing the productivity effects of unobserved firm heterogeneity. A standard practice to account for such omitted variable bias is to estimate fixed effects models. Unfortunately, information on the three attributes of JLMCs is available for only one year and we are unable to estimate such fixed effects models. 12. See Kato (2003) for details on the field research. 13. See, for instance, Fitzroy and Kraft (1987), Weitzman and Kruse (1990), Levine and Tyson (1990), Jones and Pliskin (1991), Ben-Ner and Jones (1995), Kandel and Lazear (1992), Kruse (1993), Holmstrom and Milgrom (1994), Baker, Gibbons, and Murphy (1994), Milgrom and Roberts (1995), Ichniowski et al. (1997), Helper (1998), Black and Lynch (2001, 2004), Kato and Morishima (2002) and articles featured in a special issue of Industrial Relations (Vol. 35, July 1996). 14. See, for instance, Morishima (1991a). 15. For the effects on worker outcomes of participatory employment practices, see for example Appelbaum et al. (2000), Freeman and Kleiner (2000), and Jones, Kato, and Weinberg (2003).

ACKNOWLEDGMENT This research was assisted by grants from the Abe Fellowship Program of the Social Sciences Research Council and the American Council of Learned Societies with funds provided by the Japan Foundation Center for Global Partnership. The HRM Survey of Japanese Firms was conducted at Keio Economic Observatory with support from the Picker Fellowship Program and Asian Studies Program of Colgate University, the Jerome Levy Economics Institute, Keio University, JIL, JPC-SED, the Japan Securities Research Institute and Motohiro Morishima. I am grateful to the Guest Editor, Mark Klinedinst for helpful comments. I also benefited from participating in the International Conference on Organizational Design, Management Styles and Firm Performance, University of Bergamo, Bergamo, Italy, June 22–23, 2001.

REFERENCES Addison, J. T., & Belfield, C. R. (2000). The impact of financial participation and employee involvement on financial performance: A re-estimation using the 1998 WERS. Scottish Journal of Political Economy, 47(November), 571–583.

78

TAKAO KATO

Appelbaum, E., Bailey, T., Berg, P., & Kalleberg, A. L. (2000). Manufacturing advantage: Why high-performance work systems pay off. Ithaca and London: Cornell University Press, ILR Press. Bailey, T., Berg, P., & Sandy, C. (2001). The effect of high-performance work practices on employee earnings in the steel, apparel, and medical electronics and imaging industries. Industrial and Labor Relations Review, 54(March), 525–543. Baker, G., Gibbons, R., & Murphy, K. J. (1994). Subjective performance measures in optimal incentive contracts. Quarterly Journal of Economics, 109(November), 1125–1156. Bartel, A. P. (2004). Human resource management and organizational performance: Evidence from retail banking. Industrial and Labor Relations Review, 57(January), 181–203. Batt, R. (1999). Work organization, technology, and performance in customer service and sales. Industrial and Labor Relations Review, 52(July), 539–564. Bauer, T. K., & Bender, S. (2001). Flexible wages systems and the structure of wages: Evidence from matched employer-employee data. Working Paper No. 2980. London: Centre for Economic Policy Research Bayo-Moriones, J. A., & Galilea-Salvatierra, P. J. (2003). Merino-Dı´ az de Cerio, J., Participation, cooperatives and performance: An analysis of spanish manufacturing firms. In: T. Kato & J. Pliskin (Eds), Determinants of the incidence and the effects of participatory organizations: Advances in the economic analysis of participatory and labor-managed firms (Vol. 7, pp. 31–56). Amsterdam: Elsevier/JAI. Ben-Ner, A., & Jones, D. C. (1995). Employee participation, ownership, and productivity: A theoretical framework. Industrial Relations, 34(October), 532–554. Black, S. E., & Lynch, L. M. (2001). How to compete: The impact of workplace practices and information technology on productivity. Review of Economics and Statistics, 83(August), 434–445. Black, S. E., & Lynch, L. M. (2004). What’s driving the new economy? The benefits of workplace innovation. Economic Journal, 114 (February), F97–116. Blasi, J., & Kruse, D. (1995). Employee ownership, employee attitudes, and firm performance. Working Paper No. 5277. Cambridge: National Bureau of Economic Research. Blinder, A. S. (Ed.) (1990). Paying for productivity: A look at the evidence.Washington, DC: Brookings Institution. Brunello, G. (1991). Bonuses, wages and performance in Japan: Evidence from micro data. Ricerche Economiche, 45(April–September.), 377–395. Cappelli, P., & Neumark, D. (2001). Do ‘‘high-performance’’ work practices improve establishment-level outcomes? Industrial and Labor Relations Review, 54(July), 737–775. Conyon, M. J., & Freeman, R. B. (2001). Shared modes of compensation and firm performance: UK evidence. Working Paper No. 8448. Cambridge: National Bureau of Economic Research. Eriksson, T. (2003). The effects of new work practices – evidence from employer–employee data. In: T. Kato & J. Pliskin (Eds), Determinants of the incidence and the effects of participatory organizations: Advances in the economic analysis of participatory and labormanaged firms (Vol. 7, pp. 3–30). Amsterdam: Elsevier/JAI. FitzRoy, F. R., & Kraft, K. (1987). Cooperation, productivity, and profit sharing. Quarterly Journal of Economics, 102(February), 23–35. Freeman, R. B. (1976). Individual mobility and union voice in the labor market. American Economic Review, 66(May), 361–368. Freeman, R. B., & Kleiner, M. M. (2000). Who benefits most from employee involvement: Firms or workers? American Economic Review, 90(May), 219–223.

The Nature, Scope and Effects of Joint Labor-Management Committees

79

Freeman, R. B., Kleiner, M. M., & Ostroff, C. (2000). The anatomy of employee involvement and its effects on firms and workers. Working Paper No. 8050. Cambridge: National Bureau of Economic Research. Freeman, R. B., & Weitzman, M. L. (1987). Bonuses and employment in Japan. Journal of the Japanese and International Economies, 1(June), 168–194. Gibbons, R. (1997) Incentives and careers in organizations. In: D. M. Kreps & K. F. Wallis (Eds), Advances in economics and econometrics: Theory and applications: Seventh world congress (Vol. 2, pp. 1–37). Cambridge: Cambridge University Press. Hamilton, B. H., Nickerson, J. A., & Owan, H. (2003). Team incentives and worker heterogeneity: An empirical analysis of the impact of teams on productivity and participation. Journal of Political Economy, 111(June), 465–498. Hart, R. A., & Kawasaki, S. (1995). The Japanese bonus system and human capital. Journal of the Japanese and International Economies, 9(September), 225–244. Hashimoto, M. (1990). The Japanese labor market in a comparative perspective with the United States: A transaction-cost interpretation. Kalamazoo, MI.: Upjohn Institute for Employment Research. Helper, S. (1998). Complementarity and cost reduction: Evidence from the auto supply industry. Working Paper No. 6033 (revised). Cambridge, MA: National Bureau of Economic Research. Holmstrom, B., & Milgrom, P. (1994). The firm as an incentive system. American Economic Review, 84(September), 972–991. Ichniowski, C., Shaw, K., & Prennushi, G. (1997). The effects of human resource management practices on productivity: A study of steel finishing lines. American Economic Review, 87(June), 291–313. Inagami, T. (1988). Japanese workplace industrial relations. Industrial Relations Series, No. 14. Japanese Workplace Industrial Relations. Tokyo: The Japan Institute of Labour. Jones, D. C., & Pliskin, J. (1991). The effects of worker participation, employee ownership and profit sharing on economic performance: A partial review. In: R. Russell & V. Rus (Eds), Ownership and participation: International handbook of participation in organizations (Vol. 2, pp. 43–63). Oxford: Oxford University Press. Jones, D. C., & Kato, T. (1993). Employee stock ownership plans and productivity in Japanese manufacturing firms. British Journal of Industrial Relations, 31, 331–346. Jones, D. C., & Kato, T. (1995). The productivity effects of employee stock-ownership plans and bonuses: Evidence from Japanese panel data. American Economic Review, 85(June), 391–414. Jones, D. C., Kato, T., & Weinberg, A. (2003). Changing employment practices and the quality of jobs in Central New York: Evidence from case studies of medium-sized manufacturing establishments in Central New York. In: E. Appelbaum, A. Bernhardt & R. Murnane (Eds), Low-wage America: How employers are reshaping opportunity in the workplace (pp. 479–526). New York: Russell Sage Foundation. Kandel, E., & Lazear, E. P. (1992). Peer pressure and partnerships. Journal of Political Economy, 100(August), 801–817. Kato, T. (2000). The recent transformation of participatory employment practices in Japan. Working Paper No. 7965. Cambridge: National Bureau of Economic Research. Kato, T. (2001). The end of lifetime employment in Japan? Evidence from national surveys and field research. Journal of the Japanese and International Economies, 15(December), 489–514.

80

TAKAO KATO

Kato, T. (2003). The recent transformation of participatory employment practices. In: S. Ogura, T. Tachibanaki & D. Wise (Eds), NBER conference report labor markets and firm benefit policies in Japan and the United States (pp. 39–80). Chicago: University of Chicago Press. Kato, T., & Morishima, M. (2002). The productivity effects of participatory employment practices: Evidence from New Japanese panel data. Industrial Relations, 41(October), 487–520. Kleiner, M. M., & Bouillon, M. L. (1991). Information sharing of sensitive business data with employees. Industrial Relations, 30(Fall), 480–491. Kruse, D. L. (1993). Profit sharing: Does it make a difference? Kalamazoo, MI: W.E. Upjohn Institute for Employment Research. Leoni, R., Cristini, A., Labory, S., & Gaj, A. (2001). New work practices in Italy – Adoption and performance effects. Paper presented at the International Conference on Organizational Design, Management Styles and Firm Performance, University of Bergamo, Bergamo, Italy. Levine, D. I., & Tyson, L. D. (1990). Participation, productivity and the firm’s environment. In: A. S. Blinder (Ed.), Paying for productivity (pp. 183–236). Washington, DC: Brookings Institution. Milgrom, P., & Roberts, J. (1995). Complementarities and fit: Strategy, structure, and organizational change in manufacturing. Journal of Accounting and Economics, 19(April), 179–208. Morishima, M. (1991a). Information sharing and firm performance in Japan. Industrial Relations, 30(Winter), 37–61. Morishima, M. (1991b). Information sharing and collective bargaining in Japan: Effects on wage negotiation. Industrial and Labor Relations Review, 44(April), 469–485. Nakamura, M., & Nakamura, A. (1991). Risk behavior and determinants of bonus versus regular pay in Japan. Journal of the Japanese and International Economies, 5(June), 140–159. Ohashi, I. (1989). On the determinants of bonuses and basic wages in large Japanese firms. Journal of the Japanese and International Economies, 3(December), 451–479. Ohkusa, Y., & Ohtake, F. (1997). The productivity effects of information sharing, profit sharing, and ESOPs. Journal of the Japanese and International Economies, 11, 385–402. Osterman, P. (2000). Work reorganization in an era of restructuring: Trends in diffusion and effects on employee welfare. Industrial and Labor Relations Review, 53(January), 179–196. Shimada, H. (1992). Japan’s industrial culture and labor management relations. In: S. Kumon & H. Rosovsky (Eds), The political economy of Japan. Cultural and social dynamics (Vol. 3, pp. 267–291). Stanford: Stanford University Press. Tracy, J. S. (1992). An Investigation into the Determinants of US Strike Activity. In: P. Garonna, P. Mori & P. Tedeschi (Eds), Economic Models of Trade Unions (pp. 213–231). London: Chapman & Hall. Weitzman, M. L., & Kruse, D. L. (1990). Profit sharing and productivity. In: A. S. Blinder (Ed.), Paying for productivity (pp. 95–140). Washington, DC: Brookings Institution. Whitfield, K. (2000). High-performance workplaces, training, and the distribution of skills. Industrial Relations, 39(January), 1–25.

EMPLOYEE OWNERSHIP AS A JOINT MANAGEMENT-LABOR DRIVE TOWARD CARING AND SHARING Erik Maaloe ABSTRACT Organizational change is generally looked at as a planned process. Yet, when the employees buy their own company they embark on a venture that eventually emerges as a self-growing culture. The article outlines major challenges and findings, including some of the emotional consequences. At first both management and workers expect each other to change. But gradually, a minority of technical activists sets an inspiring example for the rest to follow. As management and labor hereafter begin to listen to and recognize the need to remedy the concerns of each other, a new cooperative spirit of sharing, caring and honesty slowly emerges. The findings are based upon a four-year-long cross-comparative study of six, mainly 100 percent employee-owned manufacturing companies in the US.

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 81–103 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09003-4

81

82

ERIK MAALOE The success of ESOP companies is mainly due to management making themselves open to be watched and informed about their performance by those who – better than anybody else – are able to evaluate their deeds from day to day, and as much as themselves want them to perform at their best: their fellow employee owners.

1. INTRODUCTION Employee ownership is catching on. More than 25 million American employees now own $10,000 worth of shares in the companies they work for (NCEO, 2004). And with such a success that US ESOP companies1 are setting an example for the rest of the world.2 The impact varies with the degree of ownership. Two percent ownership is nice and as such appreciated as any other employee benefit. Five percent is a further step up. The price for a workstation in industry is around US $150,000. Thus, 5 percent ownership is a major investment and an impetus for a more conscientious behavior. With more than 25 percent, employee ownership becomes a real challenge as the workers and members of staff now have reasons to feel ‘‘they really own the place’’. So, they have to face how really to share and make it work – as this article exemplifies, based as it is on the experiences from 100 percent ownership ESOPs – and they will eventually.3 However, it certainly does not happen with the same ease or pace from place to place. Even though the challenges business wise are often the same, it may be an exasperating affair. The worker–management relations of the past, the personalities involved as well as the story behind the establishment of an ESOP all have their hold on the people involved. Thus, every situation has distinctive traits of its own. The growth of a high-percentage ESOP is a story – like many others – of facing and struggling with challenges. But it is also unique, as it becomes a shared employee-management venture. Without any previous experience, the newly hatched owners have to learn as they go along. And surely you can get inspiration from others, not least by visiting or reading internal case stories from other ESOPs (Maaloe, 1998). But fundamentally, shared ownership is such a new venture that every company has to pave its own way. The company has to be re-constructed neither bottom up, nor top down but simply from the inside. Thus, it does not help just to implement any structural solutions known from other places before the time is ripe.

Employee Ownership as a Joint Management-Labor Drive

83

Past attitudes, beliefs and even ideologies are ingrained in our bodies and cannot just be washed away over night. The past has to wither away as people gain new experiences of working together for what now – in a real sense – slowly emerges to be their company. Thus becoming employee owners for real is a tale that may enlighten us about what participation really means, set as it is within the confines of a self-growing culture.

1.1. About the Study The article condenses what happened through a four-year period at one partially and five 100 percent employee-owned ESOPs in Ohio into a single, ideal-typical (Weber, 1968 [1904]) story line. Thus, the results to be reported are radical in the sense that they refer to ESOPs with a higher degree of employee-ownership than most ESOPs. They accordingly set a standard for what may be achieved. Apart from Weirton Steel, with its, at that time, 8,500 employee owners, all sites selected were manufacturing companies, ranging from 100 to 300 employee owners, where skilled workers produced series or even a one-item variety of products. The sites were selected primarily according to (a) how long they had been ESOPs and/or (b) whether they were unionized. The study at Weirton Steel was limited to an elucidation of the education and practices of problemsolving groups, to which the present article contains no direct references. Year after year, day after day, I was allowed to walk the floor, sit in meetings, interview and observe for a period of two months. During the field study, which in total lasted nearly six months, more than 120 interviews were taped alongside numerous informal ones, juxtaposed and integrated with various series of observation and collections of written company documents. Taken together, the findings allow us to pass on some reflections on organizational behavior in general, and in particular on the potential impact of minority groups on the formation of a self-growing industrial culture of extended ownership. The findings here may thus be seen as exemplary in the sense that most employee-owned companies are only partially so, even though the internal flow of events of the one who was only partially owned, Reuther Mold, did follow the same course of events as the rest. Alongside the study, a new approach to case studies was developed: explorative integration, a cyclic approach based on continuous exploration, comparison and juxtaposition of field data with the integration of theory (Maaloe, 1996, 2002), based upon analytical generalization (Maaloe, 2004).

84

ERIK MAALOE

2. A PROCESS OF SEVERAL PHASED CHALLENGES 2.1. Getting Started As soon as the ESOP – not least a 100 percent employee-owned one – is established, both workers and management immediately feel relieved. Having been up for sale ‘‘the company now is ours, it cannot be sold without we being asked’’. What a relief! And what a wonderful weekend after having received your certificate as owner the day before! Yet looking around the shop the first Monday morning as an owner, it is as if nothing had changed. Everything around you looks the same as when you left it the Friday before. Yet inside people know that something has indeed changed: they themselves have changed. Consequently, they start – as owners – to be a little more conscientious with their daily routines, which the supervisors, to their delight, often are the first to notice, see Quote 1. Quote 1. Jim Cooper, Quality Supervisor at Reuther Mold, about how his fellow employees changed the very first week they had become owners: JC: Ah, no I think, ah, the operators, the fellows doing the machine work, I think they’re a lot more cautious. They’ll double check, they’ll check either with me or the foreman. They’re, they’re a lot more cautious. They o say, they’ve got a little more in it. So whatever they can do to improve it.4

But there was more to come, and soon a new reality dawned, and front-line supervisors were among the first to be hit by unforeseen challenges, as we shall see. 2.2. Unshackling the Grip of the Past Whatever the potential, whatever you want and may have hoped for – in this case to become an owner – the imprint of an adversarial culture of the past remains embedded in our minds. It rests in the movements of the body and makes its ‘‘voice’’ heard through the initial thought patterns and emotional reactions of both workers and managers. And this is not just theory, as many a supervisor learns the first day he leaves his cage as usual to confront a worker and suddenly realizes – when he looks around – that the person in front of him is now one of the owners (see Quote 2): Quote 2, What Jim Cooper said about himself, within the first week as an ESOP: JC: I think, ah, I think I caught myself trying to be a little more tolerant and trying to stay aware the.., maybe give in a little .. be a little more understanding of myself in the

Employee Ownership as a Joint Management-Labor Drive

85

other guys’ shoes. Don’t be too quickly judge. A lot of us has thirty year habits, oh thirty, thirty-five years. We got, we got to loosen them up, and I think myself I’m loosening them up, so in some ways, maybe other ways I’ve tightened them some. EM: Aha. Which ways have you tightened them? JC: Oh, maybe, some of the work quality, try to make the quality a little better. But in personality, in trying to get along with the guys, be more understanding. And maybe stop some of the waves, not create too many waves. I’ve loosened up with the blacks a little too. When I say something, it better be what I say. Because I hate to have to take it back, if it be proven wrong, you know. So I, I’m a little more cautious about being right, first time.

And whatever the initial joy, the employee owners soon realize that they bought their company at a price. They have become responsible for a huge loan. And – as they see it – in order to pay it back ‘‘we’’ all have to cooperate and save. In order to do what can be done ‘‘here and now’’, a minority group of ‘‘technical activists’’ just think ‘‘we’’ should go ahead and save where we can. So they begin to re-structure their own workplace, making improvements that they have had in mind for years but never instigated. Why? Because if they had, they themselves would not have benefited, while the other party – often defined as the management or, even worse, some absentee owner – would. But now ‘‘as the company is ours, everyone stands to benefit’’. So, they just move ahead. 2.3. Hesitating, Waiting for the ‘‘Others’’ to Change The activists are easily spotted both for their obvious initiatives and their outspoken philosophy of ‘‘now we can do something both for ourselves and the company’’. Yet the activists will, for a while, be a minority of less than 20 percent, while the majority of the employees – the ‘‘complaisant’’ ones – hold their breath in order to see how management will react. Middle managers, for their part, now expect the employees to accommodate themselves to their ideas. They are not all that ready to support the spontaneous taking charge on the floor. Some even fight it politically, for instance by relating all sorts of stories about how stupid the workers really are, how little they understand about doing business, how little they know about running a plant, etc. The only question that eschews their mind is of course: Why has the workforce not been told? Furthermore management has always, from time to time, argued for improvements most of which came to nothing. So the complaisant ones have

86

ERIK MAALOE

one more reason just to wait and see whether management might change and start listening to them as ‘‘we now are going to work together as owners’’. Upper management, though, as they begin to stroll through the workshops, is all too pleased. They enjoy being on the floor as workers rush to them with all sorts of suggestions. Yet, they soon grow impatient. What the employees seem to want is ‘‘small stuff’’. They pressure middle management in order to promote more substantial changes. Yet, the workers now less than ever just want to do as they are told. They want to be informed and asked in due time in order to forestall premature, half-baked decisions. This may result in a stalemate for months, even half a year. They all want change, yet both the management and the workforce expect the other party to change first. ‘‘They’’ have to prove that they respect the fact that ‘‘we now all own the place’’. The turning point comes when more and more managers are led to realize that they have to listen, explain and teach – not just in theory, but also in reality. 2.4. Making it Real However, as already stated, while the majority still stays put a minority on the floor or in the office – the activists – want to prove to themselves and their comrades that ‘‘we’’ in fact have become owners. Without asking, they implicitly expect to be allowed to do ‘‘the small things’’ around their own working stations. And they act and do what can be done here and now. And they are serious. They soon realize that they will probably do more harm than good if they embark on projects that they have neither the conceptual means to grasp nor the sufficient information to evaluate vis-a`-vis other potential possibilities. So they proceed, step by step, and improve what is possible here and now. However, they soon reckon ‘‘there is only so much we can do on our own’’. Thus, having started more or less individually and later in small groups in order to ease the flow of jobs, the activists, as Step One, ask the management to inform them about the cost structure of jobs – so we can take action to prevent losses; cross-train them in different tasks – which will ‘‘benefit me as a worker as well as increase the internal flexibility of the firm’’, and to be allowed to replace more or less outworn machinery with more durable and accurate; enlighten them about the business in general. For more details refer to Fig. 1.

Employee Ownership as a Joint Management-Labor Drive

87

Management opening up to the call for participation Management support ? time Employees talking about what could be done if ..

Fig. 1.

Workerowners first efforts and call for help

Getting Started as a Worker Owner and Testing whether it is for Real.

2.5. First Crisis? The reaction of management – or, rather, the individual managers as well as the stewards5 – to the pressure for ‘‘freedom to do and demand for insight’’ from below determines the outcome of this first obstacle to organizational growth. Some managers do sense the potential. Accordingly, they will with great diligence, often, even at day One, do their outmost to release and support the potential for immediate small-scale technical change. Other managers – still thinking along the lines of the need to control – are too occupied with how and whether it could be possible to get workers to police each other now. A third group – a tedious group of middle managers – fights the aroused activism for at least two reasons. First, they see the floor-level activism as a threat to their traditional prerogatives as managers to make decisions. Secondly, they may even choose to see long-due improvements, as they now occur on the floor, as a sign, which the upper management might perceive as a lack of social and/or technical skills on their part as managers. How would you as a sales manager, for instance, react if a worker in shipping came to your door with the results of a ‘‘customer survey’’ he had instigated on his own initiative by slipping a questionnaire into each delivery he had packed? Most managers, though, will – along with the majority of workers – take a more complaisant ‘‘let’s wait and see attitude’’. Although soon enough

88

ERIK MAALOE

they will have to take a stand in the ensuing tensions between the technical activists and the more tedious managers. Thus, we reach the point that determines whether or not the company will enter into its first crisis as an ESOP. And note that both parties involved may have the very best reasons for the position they take. Workers for not getting sufficient support for the improvements they have suggested; while management at large feels threatened by the aroused technical activism on the floor and/or disgusted by the, at times, all too evident lack of foresight and coordination of actions taken further down in the company. If you are emotionally attached to believing ‘‘the others’’ to be stupid, it is not difficult to find incidents to prove it or behave in ways that will eventually make them appear so. Of course, there is only one right way of loosening this knot: learning to work together!

2.6. The CEO Seeing the Light One could argue that employees as stock owners should not have a say beyond what stock owners in general might air at the bi-annual general meeting. But the employees want to be owners, not just investors. So, any initially hesitant top manager has to find a way to tackle the challenges from below. Harry Featherstone, for one, did and recalls with joy how he came to see the light, refer to Quote 3 given below. Quote 3. Harry Featherstone, CEO at Will Burt, Ohio on how he certainly came to realize the implication of people wanting a say. HF: It didn’t dawn on to me until this year, where I was going wrong.. It kept aggravating me that people said. ‘‘Harry, I want to make decisions’’. And I would say: ‘‘What decisions’’. And they’d say: ‘‘We don’t know’’. OK ? And I’d say : ‘‘Well, how do you want to make them’’. And they’d say: ‘‘Well, we don’t know’’. I’d say: ‘‘OK, do you want to sit in on my job?’’ And they said: ‘‘Oh no..’’. I said: ‘‘Well what do you want to do’’. Ah, one of our people went to a program in April and came back and said: ‘‘Harry, think of this, everybody must be...have control and ownership in the sphere of their influence’’.

Employee Ownership as a Joint Management-Labor Drive

89

That kicked me over to say what was wrong. And then I said: ‘‘If Jack is the very best skilled person and through all of his training and does his job in his sphere of influence.. then all of us in each sphere will be the best company in the world’’ And that’s what we have to train for. EM:Yeah. The idea is that the individual operator shall own his own work area and e.g. the maintenance people will see the operators as their customers..as if you have your own maintenance shop you will make sure that your customers are satisfied?

As a consequence, Will Burt – with the help of the local college – launched a major training program for its employee owners that later was to gain national recognition. And, of course, without the genuine commitment and personal example set by top management, the ESOP will not fully blossom.

2.7. ‘‘Learning to Work Together’’ The ease with which this captivating phrase is put on paper belies how difficult it is to achieve! No company would survive for long if people did not somehow work together. Thus, the phrase has far deeper implications than may at first meet the eye. What one has to learn is not just to work together as managers and workers but as owners; turn out to be potential equals, as everyone does his very best to improve his own performance as well as that of the co-owners. It cannot be taught nor just copied. It has to be an experience to be created from within. Or, as I came to see it from the outside: through a ‘‘selfgrowing’’ culture. The way it is done is as simple as it is beautiful: First, create a shared language for the identification and analysis of areas in need of upgrading. Then try, by a shared effort, to solve the technical and administrative challenges at hand (see Fig. 2).

2.8. Making the Company and the Actions of Everybody More Transparent In the past, when two parties – workers and management – met to solve a dispute without sharing the same language, the scene was set for making compromises. Half-bad, half-good decisions are not the way to solve problems so long as the underlying issue remains untouched. But it is a way to learn to live with them politically. Now, in order to move forward rather than being stuck in the jaws of the adversarial scripts of the past, one has to

90

ERIK MAALOE

Management support

Management disappointed Joint management-employee facilitated education ?

Yes No demotivated, disgusted employees

Fig. 2.

Does Management Really Act to Make the Employees into Owners and thus Train and Educate them Alongside Themselves?

learn and accept to sit down together and develop a shared language-to-be as owners. And, as people want to start where they are and face what is immediately most promising or troublesome, most ESOPs – despite the high hopes of some managers – do abstain from any grand-scale improvements at first. They will come later on. First, a lot of small operational challenges, which the workers themselves have deplored for years, have to be solved. Thus the first potential crisis passes as workers and management jointly embark on a quest for troubleshooting in order to verify the depth, breadth and consequences of technical, administrative and social challenges and to develop and – what they soon learn – coordinate alternatives. Thus, upper management has to install techniques for doing so commit themselves and support it by their personal example (see Fig. 3). This calls for techniques such as fish-bone-diagramming based on explorative brainstorm for identification of possible causes of unwanted situations. Yet the consequences of fact-finding for the management within employeeowned companies may be harder to face than in the industry at large, as there is no limit to what the workers want to investigate and correct. And remember that within not at least a 100% ESOPs as those in the sample referred to here they – the worker-owners as a group – will represent the majority of votes. So, just as the management watches the performance of the operators to help them, the employees now want to acquire the means to evaluate the performance of management in order to learn as well as to be taught.

Employee Ownership as a Joint Management-Labor Drive

91

Managers being open to criticism and able to change their own ways and means ?

Joint management-employee facilitated education ?

Yes No

Disappointed employees

Fig. 3.

Setting up Committees for Joint Problem Solving.

2.9. A Potential Second Crises as Management Comes Under Increased Pressure So employee owners have to learn, and who should be better at training them than the management? Yet by becoming owners themselves, they have already so much on their hands that they may quite easily be frustrated by having to deal with a flow – as they see it – of minor details or even absurd suggestions for changes. Yet, the only way to move forward is to teach all and everybody to clean their place and thus the entire house bottom up. This has a lot of practical implications: It is not enough just to ‘‘open the books’’. They have to be made readable. It is not enough just to ask for suggestions for improvement. They have to be discussed. And worse, managers have to learn not just to identify and quantify trouble spots, but to listen and rely on others in order to do so too. Furthermore, top management and stewards have to establish a supplementary structure for the coordination of problem-solving groups in order not to swamp the board with operational details. Sharing of concerns has to be voiced at face-to-face meetings rather than by mail to ensure not only proper informational feedback but also that the messages conveyed are relevant and sufficiently qualified to be understood by the addressee. This is in congruence with what the best-driven western companies try to accomplish as well. Yet, there is a difference. In ESOPs the workers have voting power, and thus the internal drive here toward sharing has more clout and thus, in turn, depth. Thus, as training and active collaboration on

92

ERIK MAALOE

solving practical problems increase, the company gradually moves toward a second crisis, as elaborated in Fig. 4. Thus, the management is challenged to teach but also – as a powerful minority group – to be ready to listen, evaluate and change their own inborn ways and exclusive strategies for decision making. And if not, the board has to find some structural means to make them do it. At the employee-owned steel mill at Weirton, they decided only to promote managers with a good track order for promoting participation and setting up problem-solving teams, with obvious and measurable results.

2.10. Caring: The Complaisant Becoming Placid Supporters It is not so much a question of control, although some stewards can initially be quoted for stressing it. In an employee-owned company, people come to care, for their job, for each other and even for management. Thus managers, too, must expect to be appraised for their performance by the hundreds of people who – far better than any absent and often minor owner of stock – watch what they are doing from day to day. Slowly, as results appear, the former idea – institutionalized by union membership – of ‘‘mutual protection against the whims of management’’ withers away as the activist slowly begins to be seen as a role model for all

Managers being open to criticism and able to change their own ways and means ? Continued success

Yes No Employee wrath Failure a lesson learned and an idea withering away

Fig. 4.

Is Management Really Open to Suggestions that Might Improve their Performance?

93

Employee Ownership as a Joint Management-Labor Drive

the employees. In turn the complaisant may, even though they probably never will become as active as the activists want them to be, eventually begin to serve them in two important ways. First by feeding them ideas of their own and secondly more important as serving as a ‘‘conservative’’ sounding board for testing perhaps, at times, the too high-flown ideas for change. For an overview of all phases of challenge, refer to Fig. 5. 2.11. The CEO will have to Lead the Way CEOs may enter the ESOP with as much eagerness and devotion as the most ardent technical activist or as hesitant as the more complaisant workers. But one thing is certain: the ESOP will not be developed before the commitment of the CEO is obvious to everybody. He, first of all, has to show that he supports the training programs by being part of them. And, in some of the cases, in which the CEO or Section Head in the past was a worker himself, he may even be the driver from the Day One. So, the real issue here is whether and when management begins to identify the workers as their partners and themselves as just owners a par with the other employees. ‘‘We’’ have different jobs, yes, and different positions, yes, Managers being open to Management Management criticism and able to change opening up to the disappointed call for participation their own ways and means ? Joint Management management-employee support ? facilitated education ? Continued success

Employees talking about what could be done if ..

Worker owners first efforts and call for help

Yes No

Yes No

Demotivated, disgusted employees

Disappointed employees

Yes No

Employee wrath

Failure a lesson learned and an idea withering away

Fig. 5.

Full Participation Time Line.

94

ERIK MAALOE

but ‘‘we’’ have to share in order to not only survive but also to grow. And they do, and they do it well as far as the numbers tell us. So, let us take a look at the process as a social-dynamic venture.

3. SOCIAL-DYNAMIC INTRICACIES OF THE OBSERVED PROCESS OF CHANGE 3.1. The Situation as it was Most people want to see themselves as being in charge. This is not always possible. But we have several options. We may instead seek to improve some selected, more partial skills in order to be able to do more, e.g. by the avenue of training, job enrichment, etc., try to fight those who we perceive stand in our way – either directly, through political countermeasures, or indirectly, through withdrawal of help – and/or by and inventing grand schemes and/or seeking the company of others who will help us prove to ourselves that we are in some sense masters. In the old industrial culture, all these strategies were used. Operators mainly stayed put, while aspiring managers opted for training, labor unions for defensive political strategies and everybody was more or less prone to nourish thoughts that could prove that they knew best. However positively defined within their own frame of reference, reactionary and defensive modes carried the day. The only chance of survival was either passively to lie low or actively to outmaneuver somebody else.

3.2. Situational Structures Shape Feelings and, in turn, Perceptions Most of us associate ownership with ‘‘being in charge’’! Thus, the emergence of or push for technical activism on the floor may seem to be a natural outcome of the change of ownership. And in many companies, across industries, and regardless of ownership, skilled workers are increasingly in charge. Yet, to achieve ownership is indeed a wonder. It eliminates the power difference and thus shatters the foundations on which workers have organized themselves on the floor vis-a`-vis their – as Marxists will have it – exploiting adversaries.

Employee Ownership as a Joint Management-Labor Drive

95

The older workers and not least the stewards would – in the past – immediately let any newly engaged apprentice know that ‘‘we’’ do not give management anything without getting what ‘‘we’’ deserve in return. Workers thus policed each other to ensure that no one would transcend the limits. And as management, for their part, thought that they had the right to set up a regime of their own, such a defensive attitude was understandable. Thus, the ‘‘good’’ worker was coerced into aligning himself with his peers. The few workers who tried to contribute ideas of their own ran the risk of being ignored by their supervisors, or, if heard, being more or less ostracized as a ‘‘management pet’’ by their mates on the floor. It was better to stay put according to a stoic dictum: If you cannot change it, do not bother. Compliance became the way to survive. And it still is, not least in companies in which you have to pay lip-service to the values that management publicly has chosen to endorse. A dynamic that Festinger (1962) generalized in his famous basic thesis: ‘‘When there is a dissonance – i.e. between what you would like to do and can do – persons will actively avoid situations which would likely increase the dissonance’’. With the advent of shared ownership, the situation has changed, at least in theory: ‘‘We’’ have become ‘‘they’’. Thus, the activists expect the former coercion among comrades to have evaporated. So they try to act as ‘‘they’’, the managers, do, i.e. by taking technical initiatives. Some managers make the same assumption – but with a different interpretation. They still believe coercion to be the name of the game. Thus they expect and call for workers to police each other as always, although now the group pressure should be turned around: not only should those on the floor work harder, they should pressure low performers to do so too. The workers may do so to some extent, but solidarity is still what matters the most. Instead, they expect management to listen and help them improve working conditions, planning and upgrade production. Apparently, it is easier to change the way one talks about the world, e.g. how one perceives it intellectually, than actually change schemata6 with their roots in the emotional and behavioral patterns of the past along with behavior. With these tensions abounding, the majority of workers at first remain as complaisant as ever. The activists are just a minority at the start. Yet, as time goes by, the results achieved by the activists create interest at first, then wonder and at last the tangible proof that ‘‘we are in it together’’. At Reuther Mold, activists for instance took charge and restructured the tool room. This meant that every new sharpened end-mill an operator thereafter took in his hand became a symbol and an implication of what employee ownership meant and the promise it might entail for all of them in the future.

96

ERIK MAALOE

Thus, gradually, the activists become the new role models, and most of their peers support them, both passively and actively, by offering them ideas to work with. For me, who in a bygone age was trained as an apprentice, it was certainly fascinating to witness how the change of ownership could compel a hitherto ‘‘suppressed’’ minority group to break the confines of a hitherto powerful majority-driven shop culture. 3.3. How can a Minority of Activists Sway a Culture Based on Group Pressure? While in the past, the former political activists worked within a reactive mode, the new type – the technical activists – is a pro-active one. And, in fact – to the surprise of many a supervisor – a few workers who in the past were seen as troublemakers now switch to become some of the most observant instigators of change. In the past, they just rocked the boat out of boredom. This, in turn, did induce some managers to favor some of the other troublemakers of the past, with disastrous results. It simply excited the wrath of the otherwise ‘‘good’’ employees – as they are apt to call themselves – who felt slanted. The occurrence as well as the fate of the activists is interesting. First, they are at no loss to defend what they do. Now that ‘‘we have become owners, whatever we do for the benefit of the company is a benefit to ourselves too’’. Second, what happens at least in part runs counter to the prevalent notion that change has to occur top down.7 We are now in a position to sum up our cases and indicate when and how a somewhat powerless minority may gain influence. It is possible when the minority members individually represent a less rigid attitude toward what is permissible; as they take charge, they display autonomy; the first tangible contributions achieved create a stir of interest among peers; the activists are able to argue that what they do is not just to their own immediate advantage, but beneficial for the group at large; and the results are recognized as valuable by at least some of their former adversaries. Or, in short, when the high self-regard expressed by the minority members can be measured against the results they achieve.

Employee Ownership as a Joint Management-Labor Drive

97

Due to group pressure, the majority of the workers on the floor used to establish defensive, complaisant norms for what kind of behavior was permissible. Now, the activists, not least in their own eyes, stand for a less constrained, more engaging and prosperous future. Thus, they emerge as an icon of what ownership is. Subsequently, as the employee ownership matures, they see the success of the company as a ‘rewarding appreciation’ for themselves.

3.4. The Emotional Dynamics of the Process of Change As activism begins to permeate the company, more and more people will let themselves be drawn into problem-identification-and-solving groups. Yet, this transformation process is not necessarily a smooth one. It may – particularly when the start has been slow and sullen – suddenly explode, as it did at a company where the labor–management relations in the past had been horrendous, (Maaloe, 1998, Chapter 2.2). The initial, apparently successful initiatives taken by the first and really modest activists led to a lot of spontaneous meetings on the floor, in which people – including the tedious ones – milled around trying hard to make decisions on behalf of everybody else. The management was not in a position to stop it. But the complaisant group – the ones who were hesitant about committing themselves – did: production had to have first priority, and nobody – including the stewards – was allowed to be absent from the work station for more than 1 hour. Yet, I am sure the event was useful. The oppressed ones simply needed to let out steam. An example of katharsis! Thereafter, the company learned to prioritize its effort, assign budgets to problem-solving groups as well as setting up a structure for doing so. However rational the process may look, bundles of emotions are apparently involved. And, certainly, the long-term change of attitudes and personal identities is one of the most fascinating findings of our study. A growing ownership culture has a profound impact on its participating members. In the beginning, the workers try to convince themselves that the world has changed. They look round the shop floor telling themselves: ‘‘This is all ours now’’. Apart from being a legal fact, this is just words. Nothing else has changed. Jobs, routines, delays, everything is the same and so are some of the managers, as we have seen. They, management, still expect ‘‘us’’ to change according to their self-proclaimed greater wisdom.

98

ERIK MAALOE

Yet, as ‘‘you’’ become absorbed in problem-solving tasks together with management representatives, this is what catches your attention. In the past, ‘‘you’’ devoted your surplus energies to rave about the lack of information and lack of managerial foresight, favoritism, etc. Now ‘‘you’’ become absorbed in the search for technical details, understanding the inter-phases between what is produced in your own department and the adjoining one, identifying alternative solutions and implementing the outcomes of a problem-solving quest on an at first tentative basis. Thus, a year later, ‘‘you’’ may suddenly realize that you have changed your attitude, not least toward management. ‘‘You’’ have come to know them – not just by sight but as useful fellow owners who are as devoted as yourself to making the company grow and prosper. And certainly one does not – as a participant observer – need a pad or a stopwatch to reckon that the interchange between the management and workers has increased considerably. Nevertheless, in the present context a quote will do (see Quote 4): Quote 4. The former shop steward, Phil Bowling, about management one and a half year after the company became an ESOP: PB: I’d say the majority of them (management) have really changed and really opened their eyes about, you know. Not that they hadn’t in the past, it’s just that now, it seems like they, they do care a lot more. And they look forward to everything. –: Care both for the job and for, for the shop? PB: Yeah. Because they realize that it’s all a, it’s all a main goal. Everybody is a part of it and it’s an integral system. It has to be done that way. Everybody has to help, you know. Because if they don’t, and you’ve got a problem at a break-down. –: Yeah, the chain and the weakest. PB: That’s the point, that’s it. So, I think everybody has really made an ‘‘about face’’. They’ve started coming on a heck of a lot better, especially in the last year, I’ve seen a lot of, lot of changes, you know..and attitudes..and the things they do, and the interest they show, you know. Even in the foremen’s end, you know. See it a little from their side, too, and as well as they have learned to see it from your side. I think it’s one of the best things that they ever did, you know (laughter). It’s really opened my eyes a lot, you know, it’s, it’s a hard process. –: It opened your eyes seeing the world with management eyes, too? PB: Aha, I know. –: Realizing that it’s, that it’s a tough job, too. PB: Yes, very tough.

Employee Ownership as a Joint Management-Labor Drive

99

–: Very tough, you say. Tougher than yours? PB: Yeah, I find mine quite easy (giggle), you know. And it’s, I do, I get along with people really well. I enjoy talking to people, you know. And, ah, I like to see things coming. –: What, in what sense do you think management’s job is more tough? PB: Pressure. –: Pressure on them, yeah. PB: We do our jobs, we go home, you know. Then we come back and we do our jobs and go home, you know. And we don’t think much about it when we leave here. I think they take a little more of their jobs home with them and they have to. They’re in constant contact with people all the time, that means profits or loss. Whereas we out there on the floor, we do the job, you know. But they have to get it for us.- If they don’t get it for us. And management, too. And the foremen’s, they have to route the jobs right, have to it on the right machines. They have to do everything right. If they don’t, we’re on them like ..flies on shit (laughter), you know.. Well, yeah, we’re not easy on them at all. We make it hard to them..

As the ESOP matures, defensive attitudes are worn away and replaced with a growing concern for flexibility and an experimental, honest search for identification of the causes and troubleshooting. As results accumulate, workers now talk with warmth, pride and gratitude about what they have achieved together with management. Just as individual managers now speak respectfully about the individual workers they have come to know. For a full illustration of the individual change process, refer to Fig. 6.

3.5. A Fairy Tale? Yes! So What Matters isy You do not believe me? You should not. Not everyone is won over the first year or two. Some are never going to. And certainly it takes a lot of fights between managers, and sometimes even between stewards, before a new culture based on warmth and mutual recognition emerges. It is a long way. It begins with sheer politeness, as the supervisor, on the first day, acknowledges that the workers now own the stock, and it ends with mutual care for the product, each other and the company, based on a continued effort to reduce costs, improve quality and expand production. It is all about learning not only to do your best but also to look at yourself with the eyes of the others to see what you really do.

100

ERIK MAALOE

3: Management support = Coaching

RECOGNITION I & my world has changed in reality

RECOGNITION

PERSONAL SPHERE

Becoming and owner and a potential contributor

2: my world has changed in principle

Emergence of new EXPERIENCE CHANGE OF ATTITUDE towards (organizational) life

Aroused greater EMPATHY for others

Imprint from a past of adversarial attitudes and agonies

ORGANIZATIONAL SPHERE

1: A new situation:

New and enriched K KNOWLEDGE NO & INSIGHT -technical and social Enriched UNDERSTANDING collectively as well as individually THE SPHERE OF EMOTIONS AND FEELINGS

Fig. 6.

The Attitude Change Circle of Social Change within an Employee-Owned Company.

An observant, caring and yet demanding culture! Everyone expects everyone to think before acting and ask if in doubt. The management can expect to be forgiven for having made a decision, which in hindsight proves to be a mistake, but only so far as they did not make it without exploring whether the people touched by it had information that ought to and could have been brought forward. Thus what counts is in the beginning, the will to save money, later the will to improve and expand, but a vigilant eye on the need to increase equity; the will to inform at all levels with diligent honesty

Employee Ownership as a Joint Management-Labor Drive

101

the readiness to inform about mistakes made and ask for help in order not to aggravate the consequences neither technically nor politically; self-observant behavior; or in short: care for the company, costs, products, processes and each other. This may seem as a fairy tale out of the Human Relations School. But my informants left me in no doubt: The more congenial atmosphere, the challenge of being part of it all, was nice. But the real drive was that of being an owner: to participate in sharing the profit and the growth of equity. 3.6. The Potential Emergence of New Cultures Finally, I would like to make some concluding remarks about what we may learn from the success of the companies with a substantial degree of shared ownership, as further outlined in Fig. 7. Because our study is presented as a narrative of individual change, the lesson we learn regarding ‘‘value management’’ is as general as it is a tough one: No one has any reason – apart from appearance’s sake – to change their attitude in accordance with whatever management thinks would be appropriate. Only new experiences and people’s emotions are solid enough to break through the old layers of complacency and pave the way for individually grounded new attitudes. New values are not just something that can be adopted; they have to be created through acquired practice – by and through your hands! Thus it is of paramount importance for management to provide room, time and training for the employees to make their own experiences. Trust me, most of us will change our attitudes in accordance with the experiences we make. Thus, it is best to let us expand! Through troubleshooting exercises, ‘‘we’’ learn that it is no use just blaming others. We are all part of it. Therefore the successful owners are not those who tried to outmaneuver their adversaries, but those who succeeded in creating an atmosphere of ‘‘let us see what we can do ourselves and together’’! Thus, the success of ESOP companies is not least due to management realizing that they have to make themselves open to be watched and informed about their performances by those who better than anybody else are able to evaluate, from day to day, what they do, and as much as themselves want them to perform at their best: Their fellow employee owners.

102

ERIK MAALOE CONFUSION 1 Getting started as an ESOP

3 Taking charge of one's own job and sharing tasks with others

2 Individual action looking at challenges close at hand

LEARNING BY DOING

BEING IN IT AS A THEORETICAL UNDERSTANDING

5 Evolvement of trust and mutual appreciation

4 Learning by working together in teams

NEW COGNITIVE STYLE AND EMOTIONAL STAND

Fig. 7.

6 Being in it together

NEW CULTURE 7 Sharing history & an identity of owner & workmanship

The Phase Model of Potential Social Change within an Employee-Owned Company.

NOTES 1. ESOP: A particular US program, ‘‘Employee Stock Ownership Plan’’, for establishing employee-owned companies, with certain credit benefits and an obligation to maintain the stock acquired until one leaves the company, typically at retirement. 2. For references see ‘‘European Federation for Employee Shareholders’’, http:// www.efesonline.org 3. For supplementary statistical evidence for all ESOPs in Ohio, refer to Logue and Yates (2002). The book contains statistics for size, growth and array of other measurements for economical performance of ESOPs in Ohio. 4. This and the following quotes are all derived directly from field notes and may – in order to convert spoken language into written – be slightly abbreviated by the wish and subsequent approval of the informants (see Maaloe, 1996, p. 10). 5. Stewards as well as a union play such an important role – as shown in Maaloe (1998) – that it would deserve an article in itself. Here it will suffice to say that while the union generally does play a significant role in establishing as well as promoting the ESOP, it may in some cases – particularly in larger companies – later turn out to be just as a restraining factor as the most tedious managers.

Employee Ownership as a Joint Management-Labor Drive

103

6. Schemata: Sets of more or less mutually dependent concepts that dominates the ways one perceives an assemblage of social or natural phenomena. A stock concept of Cognitive Psychology, dating back to Bartlett (1932). 7. Recently as an idiom most forcefully presented by the ‘‘Business Process Reenginerering’’ Movement.

REFERENCES Bartlett, F. (1932). Remembering, an experimental and social study. Cambridge: Cambridge University Press. Festinger, L. (1962). A theory of cognitive dissonance. Stanfrod, CA: Stanford University Press. Logue, J., & Yates, J. (2002). The real world of employee ownership. Ithaca, NY: Cornell University Press. Maaloe, E. (1996, 2002). Casestudier af og om mennesker i organisationer. Copenhagen: Akademisk Forlag. Maaloe, E. (1998). The employee owner – organizational and individual change within manufacturing companies as participation and sharing grow and expand. Copenhagen: Akademisk Forlag. Maaloe, E. (2004). In case of case research. Working Paper #9, rev. ed., Aarhus School of Business. NCEO. (2004). National Center for Employee Ownership Report #2, Vol XXIV. Weber, M. (1968). Gesammelte Aufsa¨tze zur Wissenschaftslehre. Ed. von Winckelheim, J C B Mohr (Paul Siebeck), Tu¨bingen, Germany. Originally published in 1904.

This page intentionally left blank

104

EMPLOYEE PARTICIPATION RIGHTS IN CORPORATE GOVERNANCE: AN ECONOMIC RATIONALE, A TEST OF A LEADING THEORY, AND SOME INITIAL POLICY PROPOSALS Stephen C. Smith ABSTRACT This paper examines roles of mandated employee participation rights (EPRs), such as works council legislation, in corporate governance. Links between employment and corporate relationships are stressed. Market failure arguments are developed, predicting that EPRs, and the interaction between EPRs and investments in skills, can positively impact productivity; preliminary evidence from German establishments is generally supportive. A qualitative appraisal concludes that EPRs have not harmed economies that adopt them. Policies to expand EPRs in the US are introduced, jointly encouraging skill development and employee decision-making participation, full rights for employee stock ownership plan

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 105–146 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09004-6

105

106

STEPHEN C. SMITH

(ESOP) participants, legal regulation of terms such as ‘‘participation,’’ and EPR extension services.

1. INTRODUCTION This paper examines the nature of employee participation rights (EPRs) and considers their role and potential impact in corporate governance. Works councils of the type that have become the norm in Europe are a key example of EPRs. The economic literature on EPRs is reviewed, with an emphasis on market failure analysis as to why EPRs might raise efficiency, and yet be underprovided by the market. Then, a simple test of a leading theory of their purpose and effect is presented: do EPRs enhance investment in skills at the workplace so that performance is improved? Theories of the role of EPRs in corporate governance, as developed by Smith (1991) and Blair (1995) among others, are argued to imply that EPRs, as well as the interaction between EPRs and investments in firm-specific human capital or skills more broadly, should have a significant positive impact on productivity. The theory is then tested with data from a sample of firms from Germany; though there are some qualifications, the predicted impacts are generally found. In particular, in small and medium enterprises with less than 100 workers, with controls for size and industry effects in place, and while also controlling for the individually positive productivity effects of works councils and training as stand-alone determinants of performance, the combination of training with works council is observed to raise net sales by a strikingly high 37%. Although the case for EPRs does not rest solely on productivity improvements, these results suggest that they may provide efficiency gains. The paper then presents a brief qualitative appraisal of whether countries with EPRs have been paying any economic price, or realizing any economic or social benefits from their presence. In particular, whether the EU performs worse than the US economically, and if so whether EPRs might be a factor. The conclusion is that there is no evidence of any net economic cost and some evidence of net social benefit, and possible accumulating long-run economic benefit, of EPRs. Finally, some modest policy proposals are offered to expand EPRs in the US economy. The proposals center around encouraging both skill development and genuine employee decision-making participation through what are termed Human Capital Partnership plans. Incentives would include preference criteria for government contracts. The proposals also address requiring full ownership rights for employee stock ownership plan (ESOP)

Employee Participation Rights in Corporate Governance

107

participants, legal regulation of the use of terms such as ‘‘participation,’’ and development of and public support for extension services to spread what is known about which types of participation institutions are most effective. The proposals are designed to get the largest positive impact with the least intrusive regulations. The policy path proposed in this paper is challenging, but some reorganization of the workplace is essential to ensure a more efficient and more democratic American capitalism. The paper begins by defining EPRs, providing examples of them, and then considering the economic case for mandating or providing incentives for establishing EPRs.

2. WHAT ARE EPRs? I define employee participation as the expectation that employees will have a measured say and stake in the quality and stability of their jobs. Employee participation rights (EPRs), as I define them, are more specific. They are legal or contractual guarantees of rights of employees to consultation, or veto power, and joint design and implementation of specific workplace decisions and practices. EPRs have legal and contractual meaning; the more general term ‘‘participation’’ does not. But neither term generally extends to bargaining over baseline remuneration. In any given economy, the level of EPRs may differ across firms, within any one firm from workplace to workplace, and across types of decisions – such as the size and arrangement of work cubicles; organization and speed of an assembly line; layoffs, dismissals and hiring; pollution, and occupational health and safety; investments, relocations and plant closings – and training. EPRs may result from privately negotiated contracts, from administrative regulations, or from legal mandates. They generally include individual rights, such as to due process in cases of dismissal, but are exercised collectively by some recognizably democratic process, usually through elected employee representatives. Their common feature is their ultimate enforceability in courts of law, such as in labor courts in the case of Germany.

2.1. Examples of EPRs The most important examples of EPRs have been found in the EU, where they have emerged primarily from legal mandates. Public policies aimed at

108

STEPHEN C. SMITH

requiring managers to share authority with employees were introduced in many European countries in the postwar period. In Germany, employees have the legal right to such participation through two legally mandated structures: the Betriebsrate, or works councils, which provide for participation in the immediate workplace, and Mitbestimmung, or codetermination (CD), which provides for employee representation on the board of directors.1 Under CD laws, half of the supervisory board members of large firms (over 2,000 employees) and one-third in medium size firms (500–2,000 employees) are elected by employees,2 while stockholders usually appoint the tie-breaking chairman in large firms. The German works council system was established under the Betriebsverfassungsgesetz, or Works Constitution Act (WCA); this is the form of EPRs examined empirically in this paper. Under the WCA, firms with more than 5 employees must establish management–employee works councils to arbitrate most aspects of workplace decision-making. In practice, the creation of councils depends on the initiative of the establishment’s employees; and the legal mandate is only enforced when employees request the councils and management is slow to establish them. Nationally, about 46% of all companies in Germany have a works council, but they cover 66–75% of all employees;3 thus they are more common in larger than in smaller firms. Members of the councils are elected for a term of 4 years.4 The councils are very popular in Germany, and seem almost universally accepted by owners, management, and unions, as well as employees. In fact, in 2001, the rights of works councils were strengthened over only limited opposition. Councils negotiate over a bundle of interrelated company policies. On some issues, they have the right to information and consultation, in others a veto power over management initiatives, in still others the right to co-equal participation in the design and implementation of policy. Their rights are strongest in social and personnel matters such as the introduction of new payment methods, the allocation of working hours and the introduction of technical devices designed to monitor employee performance. Works councils in Germany are institutionalized bodies of worker representation that have functions distinct from those of unions. They do not have the right to strike. Moreover, the WCA does not allow wage negotiations; the aim is to restrict distributional conflicts on the establishment level. Rather, works councils are designed to increase joint establishment surplus. They are tied to the obligation to cooperate with management ‘‘in a spirit of mutual trustyfor the good of the employees and of the establishment.’’ Norway, Denmark, Sweden, Austria, the Netherlands, and Luxembourg have analogous laws while Italy and most of the remaining European

Employee Participation Rights in Corporate Governance

109

OECD countries (excepting Britain) mandate councils for employee participation at the workplace, though not on the board. In France, employees also send non-voting representatives to the board5 (see Rogers & Streeck, 1995). In June 1994, the European Union reached agreement on a directive that requires works councils as part of its Social Charter. The directive required companies employing over 1,000 workers in any of its member countries, or over 150 workers in at least two countries, to establish works councils. The directive stated that councils must be consulted on relocation plans, job changes, investment and ‘‘any proposal creating serious consequences for workers’ interests.’’ After a two-year voluntary phase, starting in 1996, these councils became mandatory, though with some loopholes; after 1999, some of these loopholes were closed. Implementation was delayed in the UK until 2000, because it had previously opted out of the EU Social Charter. These multinational councils have less power and have so far not been as active as national councils in Germany and some other European countries. Surveys by Bain and Hester (2000) show that while over 500 European works councils had been established by 2000, almost three-quarters of such councils met only once a year. They also observed that at this early stage, employee representatives are still getting up to speed, and can be overwhelmed with large volumes of data presented at the last minute. However, the precedent for extending works councils to international activities of firms has been clearly set (Stirling & Tully, 2004). US and other non-European firms operating in Europe are affected by the rules – this is giving US firms growing experience with working with works councils (see, e.g., Marginson, 2004), and some are realizing that EPRs can bring significant benefits of a culture of workplace trust and cooperation.

3. WHAT IS THE ECONOMIC CASE FOR EPRs? The economic case for EPRs rests on the need to correct a set of market and organizational failures. The most widely accepted economic rational for EPRs is the protection of investments in firm-specific human capital. This argument was first advanced, although indirectly, by Dow (1987) and developed by Smith (1991) and Blair (1995, 1999). The general problem is that relationship-specific investments (Williamson, 1975, 1985) are subject to holdup. That is, whatever the understanding and agreement prior to making investments, bargaining power changes after one party has made commitments that are of value to one particular relationship,

110

STEPHEN C. SMITH

such as purchasing specialized machinery and otherwise gearing up to supply a firm as an outsourcer, a joint venture, or an employment contract. At that point, the other party has the incentive and the means to try to force a renegotiation. For example, once one party has built or paid for, installed, and moved up the learning curve with specialized machinery, that party is stuck with the investment; it is a sunk cost. Certainly, a specialized machine may have at least some value on the second-hand market, but generally this will be substantially below than the value of the full investment. Of course, to a degree, legal contracts can help mitigate such holdup problems, but such contracts can never foresee all contingencies, and in any case, legal enforcement is costly. Thus, this problem can cause otherwise valuable relationshipspecific investments never to be undertaken, or to be scaled-down. Although some investments may still be undertaken, these may be of lower quality, if effort is withheld or the parties fail to commit their best resources. Worse, the returns on these investments may be dissipated in legal battles. There are various institutional ways around such holdup problems; in the case of two firms, one of which supplies the other, an option is vertical integration, with the result that both firms are owned by the same investors (Grossman & Hart, 1986). Many acquisitions in the US economy continue to be of this type. At the same time, outsourcing is increasingly widespread. Clearly, technology has changed, but also, it would be valuable to know whether legal innovations have emerged that enable firms to better circumvent holdup problems in outsourcing in recent years. But reputation effects may be helpful here as well: usually, outsourcing with relationship-specific investments are done by large firms, which expect to be repeat buyers in the market for years if not decades to come. If a major buyer (an auto company, for example) becomes known as a firm that reneges on suppliers after they make dedicated investments, it may be difficult to convince other suppliers to make similar investments in future periods. An analogous problem emerges with firm-specific human capital (FSHK). By definition, capital is an accumulated flow of investment, which provides a stream of expected returns over an extended period. Human capital means education and skills paid for in time and money, and accumulated by employees and firms, which is also expected to provide a stream of income returns. Economists favor the term because of the formal analogy between these two types of investment. General human capital, like literacy and numeracy, increases the employee’s value at any firm, and so is generally paid for by the employee. Investment in firm-specific human capital, or skills of value to one firm but not others, is normally modeled as shared, through wage adjustments, by firm and employee. Clearly, there is no absolute line

Employee Participation Rights in Corporate Governance

111

between general and specific human capital; there are often shades of gray, measured by mobility costs, broadly construed. It may be helpful to consider some examples of firm-specific human capital: skills for operation of specially designed equipment, type of machinery proprietary or otherwise unique technical knowhow other skills for unique production processes skills for specialized management systems, software, accounting, etc. knowledge of corporate ‘‘culture’’ knowledge of who in organization has certain skills, contacts, experiences knowledge how to work best with specific company employees, or whom to best work with on a project knowledge how to best serve long-term customers with special needs or requirements knowledge of how to anticipate emerging needs of the company and its regular customers (reinforced as other companies have increased FSHK)

Analogous to relationship-specific capital investments, the problem is that after employees invest in such FSHK, management has an incentive to renege and not pay employees the return they were implicitly promised on their human capital investments. While firms usually pay for training, workers accept lower-than-market wages until their FSHK has been developed. The workers then expect higher wages in the future, but generally have no power to insist on this. The problem is that by definition, FSHK is useful in that one firm, but not in any others. Clearly, this problem cannot be solved by vertical integration, because one cannot purchase labor due to the prohibitions on involuntary servitude. Moreover, the possibilities for solving this problem through the impact on reputation are far more limited in the case of FSHK. The problem was addressed by Dow (1987, p. 24): ‘‘who monitors those in positions of authority, in order to ensure that their self-interest does not threaten collective interests?’’ In the presence of firm-specific human capital and other transaction costs, exit or its threat is insufficient. Dow concluded that ‘‘the unaided market cannot accomplish’’ the creation of managerial monitoring by employees ‘‘in part because asset idiosyncrasy is often substantial on both sides of the authority relationship, and in part because the relevant information is unlikely to pass easily across organizational boundaries (Teece, 1982), thus disabling reputational protections which might emerge via the managerial labor market.’’

112

STEPHEN C. SMITH

Blair (1995) doubts that workers will believe that reputation matters to firms. She draws on Shleifer and Summers (1988) to argue that when a market for corporate control is active, return on FSHK may be shifted to raiders. Clearly, this prospect would severely limit the effectiveness of reputation costs as a guarantor that management will hold to its implicit agreements with workers. Worse, I would argue that the threat is endogenous: the more workers in a firm are realizing a return on FSHK, the more raiders will want to take over the firm. In part, as a result, there is no reason to believe that this problem is limited to ‘‘declining’’ industries. Finally, it may be pointed out that there is active turnover of management across firms; the manager may be in another company when an agreement she made is reneged upon; I return to this time-horizon problem shortly. Thus, workers and unions would prefer in general to develop general human rather than specific human capital. This is in part because they are more protected from opportunism; but in part because it raises bargaining power. If workers’ skills can be used in other firms, this raises their opportunity cost, and thus makes a higher ‘‘threat point’’ more credible in bargaining with employers. Employers, similarly, like to undermine workers’ bargaining power, such as by finding alternative ways of organizing production, including technical innovations designed for this purpose (Dow, 1993). These problems also give workers an incentive to look for opportunities for what I have termed quasi-specific human capital (Smith, 1988). This is human capital likely useful only in one firm at the time it is created, but able to be brought out of the firm in an entrepreneurial fashion as a start-up, that may compete with the original firm. The firm, in its turn, wishes to devise constraints to keep this from happening. The result can be considerable distortion in the organization of the firm, including strategies to prevent employees from gaining skills that could reveal company secrets. Finding ways to make FSHK work better is potentially in everyone’s interest, but getting there involves solving a collective action problem, as described later in this section. Blair (1995) makes the important point that workers with FSHK are in effect among the residual claimants of the enterprise. Since their FSHK is useless anywhere else, if the firm goes bankrupt, workers share last place in line with the company owners to recoup their investments, which is their FSHK, useful only in that one firm and only if it is operating at all. If after all parties who are owed tangible bills by the company are paid, and then the net worth of the company is negative, employees lose all of their FSHK. In fact, it might be further argued that workers are even behind the owners in line, because the owners might recoup part of their investment (such as

Employee Participation Rights in Corporate Governance

113

through the sale of plant and equipment) even as the firm shuts down. Moreover, the owners, who make the decision about whether to stay in business, count returns to FSHK as part of labor costs, rather than as part of the firm’s economic rents. As a result, as Blair points out, workers with FSHK are subject to the risk of the residual claimant just as shareholders are, perhaps more so. Indeed, a significant fraction of Fortune 500 firms disappeared from the economic landscape in the past decade through bankruptcies and various takeovers that can put FSHK at risk. Thus, there is as much of a case for granting decision-making rights to workers with FSHK as there is for investors, from the point of view of designing law to maximize the creation of net social wealth.6 The legal protection of EPRs, by decreasing the risk that management will renege on agreements with employees, increases employee confidence in their investments in the firm. It would be difficult to achieve this employee confidence without such legal intervention (Streeck, 1984, p. 417). This would in turn promote incentives for employees to take a more long-term view of the firm’s interests, decreasing their incentive to opportunistically seek out short-run benefits. It may also be asked, if EPRs have benefits, why do firms not set up such a structure on their own without an external legal framework? The transaction costs for an individual company to set up credible guarantees without an overall legal framework may simply be too high; by analogy, consider the lower value of voluntary disclosure of financial information without government auditing and sanction authority. With a clear legal and standardized framework already in place, legal costs for establishing a credible agreement are kept low. The analysis in the literature has stressed firm-specific human capital, but as a practical matter, the arguments also largely apply to investments in general human capital.7 This is for two interrelated reasons: (1) general and specific human capital investments are often efficiently bundled together and (2) vocational learning that has an experiential component – such as on-thejob training – tends to be more effective than out-of-context classroom learning. The advantage for experiential learning will apply to more general as well as to more specific skills. And if learning firm-specific skills also entails learning general skills, it will often be advantageous to learn both onthe-job, whether we are considering further training within a particular job or formal job upgrading associated with a certificate of skill. This is because of the general pedagogical value of combining theoretical education with experiential content and the particular experiential value in that the skills to be learned are to be applied at that workplace. There are also advantages to

114

STEPHEN C. SMITH

be gained from returns to scale in the learning process. Learning always entails ‘‘set up costs,’’ and it is more efficient to learn related subjects at the same time to avoid having to employ separate instructors, each using their own method, to independently set up the context for similar learning exercises.8 In US practice, training inside the firm usually has some general human capital content. In the 1990s, there were regular reports in the business press of firms engaging in remedial education alongside training for specific job skills for new workers. The need for EPRs is magnified because of the role of management in a modern corporation. There is an incentive for managers to take expanded authority in the firm not just because and only when it is efficient, but when they have the power to do so and it benefits the manager. It is well known that there are incentives and opportunities for managerial opportunism toward shareholders, as the industrial organization, management, and economic behavior literatures examine in detail (Williamson, 1975, 1985). But there are also incentives and opportunities for managerial opportunism toward employees. There are at least four major incentives for management to distort the organization of work away from that which would maximize profits. These distortions would decrease the level and quality of training and lower the returns on FSHK and quite likely the returns on equity as well. However, there is a potentially offsetting role for EPRs to blunt the impact of these incentives for opportunism (the following draws on Smith, 1991): (i) Credit-taking Opportunism. Managers may often find that to maximize their own rewards in an organization, they should design and implement an organizational structure that does not maximize company profits. Management has an incentive to make it appear that innovative ideas originate with the managers rather than their employees; and that they are otherwise more productive than their subordinates. This impression justifies high salary as well as promotions and bonuses. The ability of managers to create the impression that they are more often the source of productivity and innovation will depend on the organization of the workplace, which is designed by managers who have an incentive to consider their own private benefits and costs before those of the enterprise as a whole. Considerable distortions will be possible before the manager’s private costs resulting from any overall damage to the firm equal or exceed his private benefits from a more efficient organization, which might, among other things, credit the manager less for its successes.9 For example, employees often complain that they are ‘‘forced to train their bosses,’’ who may be a management trainee or

Employee Participation Rights in Corporate Governance

115

recent transfer. The manager faces obvious disincentives against pointing out the subordinates responsible for his success. The circumstances may lead employees to withhold valuable training or other information, useful for innovation or organizational development, from their supervisors. Thus, too little training of both higher and lower level employees may result; EPRs might reasonably lessen the impact of such organization failures.10 The results of a distorted organizational structure may include reduced incentive for employees to develop and put forward innovative ideas11; and false information, or ‘‘signals,’’ to higher authorities about the distribution of abilities in the organization. EPRs provide employees with a regular grievance channel to higher-level managers or ultimately to owners; even if not used often, their existence may deter management opportunism. (ii) Time horizon opportunism. In the US, management careers typically involve working sequentially for a number of companies. Further, senior executives tend to be relatively close to retirement age, and thus the time horizon of management can be shorter than either that of owners or other employees. Managers’ rewards will be based on perceived performance during their tenure. Since, in many cases, the median age is lower and expected tenure is higher for nonmanagerial employees, one valuable ‘‘balance’’ that EPRs can offer is to reduce management time horizon opportunism.12 Managerial actions taken with a short-term view to increase management rewards at the expense of the long-run viability of the company violate implicit labor contracts as well as managers’ contract with shareholders. By the time that information on the ability or actions of top management becomes clear to the market in the absence of EPRs, penalties for management may be otherwise outweighed by private benefits of their opportunistic behavior. Knowledge of these structural incentives may reduce investors’ willingness to make financial investments, but particularly reduce employees’ willingness to make specific human capital investments. The latter disincentive is apparent in the US downsizing movement, which may have gone much farther than optimal from a long-run perspective, lowering investments in R&D and in employee training.13 EPRs have slowed downsizing in Europe, and this has some inefficient features. But these firms continue to invest in training and have been seeking other methods of adapting to increasing international competition that may place them in a better position for the long run. Managers are inhibited from taking a longer-time perspective by the need to keep share prices high enough to avoid a hostile takeover (Shleifer & Summers, 1988). Some hostile takeovers may

116

STEPHEN C. SMITH

have been motivated to essentially strip workers of their rights to returns from their FSHK. While there is a role for the market for corporate control, excessive ease of takeovers may reduce incentives for specific human capital investment. So it is a two-edged sword. Employees do not necessarily benefit from either incumbent management or outside contenders. EPRs would reduce but not eliminate these problems. (iii) Information flow opportunism. Management has an incentive to provide a less than optimal flow of information, among employees and between owners and employees, for example. Since bargaining power is generally correlated with information, centralization of information, without access to it when necessary, can lead to its hoarding and misuse. EPRs carry some risks of slower decision-making (or informational inefficiency). But in a second-best world in which agents may have the incentive to behave opportunistically, its potentially slower decision-making may produce a superior average result by offering an internal quality control mechanism over management decisions.14 The firm may not introduce EPRs voluntarily because they may be costly to individual managers, who make decisions about internal organization. Although managers may benefit from ensuring that employees receive credible information, they may fear that other functions of EPRs, including quality control over management decisions, represent a threat to their jobs. Board membership for employee representatives, for example, is a two-way information channel; the reverse channel from employees to owners may be the one management is concerned about. Thus, the presence of local workplace decision-making requirements, and even non-voting employee board members deprived of full information, could lead to improved efficiency. Employees and owners, as well as managers, have incentives to withhold certain information. But EPRs are likely to reduce these problems by revealing some information and suggesting to perceptive participants where other information is being distorted or withheld. (iv) Authority-hoarding opportunism. In part because of the incentive to create opportunities for management opportunism, and in part due to the direct managerial utility from authority, EPRs are preferred ‘‘workplace characteristics’’ in systematic undersupply, adding to the firm’s compensation costs.15 Research findings show that the incidence of stress is highest among workers who are rated as having little control over their jobs, other things held constant. Medical studies show that men whose jobs combine high psychological demands with little control over their work face heart attack risk two to three times as great as other male workers, other things held constant (Karasek et al., 1988).

Employee Participation Rights in Corporate Governance

117

A survey of 600 US workers16 found that 34% of respondents considered quitting due to job stress, and 14% did so; 46% of workers called their jobs highly stressful. The study concluded that lack of control over one’s job is the central cause of job stress, which in turn reduces productivity and causes more turnover, absenteeism and illness that was believed previously. In the UK, the Health and Safety at Work Act requires firms to minimize job stress, and in 2003 a legal action was initiated against a hospital that failed to adequately address workplace stress.17 As jobs become more complex, greater control over one’s work may be necessary to reduce stress – another source of the EPR-productivity linkage – and promote effective training. Conversely, increasing required participation levels of employees, for example through teams, without accompanying this with genuine authority, training and other resources, is likely to increase employees’ levels of stress (cf. Batt, 1993). This is another reason why merely creating teams in companies, without providing real EPRs, is unlikely to be as productive as when credible EPRs are in place. Key managers will often resist EPRs even when they know they would benefit the firm. Klein (1984) found that even first-line supervisors are strongly resistant to ‘‘employee involvement’’ programs: ‘‘most revealing, perhaps, is the finding that although nearly three-quarters (72%) of the supervisors view these programs as being good for their companies and more than half (60%) see them as good for employees, less than a third (31%) view them as beneficial to themselves.’’ Among other causes, Klein’s surveys point to supervisors’ fear of loss of status and power in the workplace. Any disadvantageous firms with EPRs might face in competing for entrepreneurial talent would be limited if they were legislated for all firms (Putterman, 1982, p. 157). And there is some evidence (Wever, 1994) that German managers now feel they benefit from codetermination. Further, coordination failures may be present (Smith, 2003). For example, if EPRs are distributed throughout the economy, employees will encounter similarly empowered counterparts in joint ventures, sales, or other market activities, maximizing the benefits of such decentralized authority.18 Moreover, some process innovations may fit with the organizational comparative advantage of firms with EPRs, such as non-Tayloristic operations utilizing knowledge and skills impacted in the work team, that is unobservable to management. Such innovations would be selected against when workers have an incentive to shirk. But if workers represented by works councils may overcome this problem through appropriate and verifiable bargaining, that

118

STEPHEN C. SMITH

includes a combination of direct financial incentives of ownership, augmented by the incentive for mutual peer monitoring, then these innovations may be efficiently used in firms with EPRs; the more such firms present, the greater the incentive to invest in such innovations.19 In all of this, note that some performance-related pay, which provides incentives to share and use new knowledge and innovation acquired by employees, seems to go hand in hand with EPRs. There might be a concern that employees with EPRs could exercise them in a way that would lead the firm to behave in an excessively risk-averse manner. While this is theoretically possible, I would argue that in fact US firms (which lack EPRs) have been acting in an excessively (socially costly) risk-loving manner in the last two decades. It has been widely noted that US stock markets have shown a tendency to speculate on corporate cash flows (see e.g. Shiller, 2000). Moreover, incentive stock option schemes carry some moral hazard risk; with little of their own paid-in capital at risk, executives have an incentive to enter the corporation into highly speculative activities in the hope of windfall capital gains. EPRs could play a constructive role by partially counterbalancing this tendency to risk-loving behavior. However, the impact of attitude toward risk would in any case be rather muted, because works councils are generally exercised at the workplace level, while corporate strategy is designed and implemented at a higher level. Other market failures that EPRs may help to address include problems in occupational health and safety, and environmental protection. For example, when workers have EPRs, which should facilitate tradeoffs between various types of benefits, the firm may pollute less or cause fewer negative externalities, because workers live in the community near the plant (Vanek, 1971; Askildsen, Jirjahn, & Smith, forthcoming). Evidence from Germany not only demonstrates a robust relationship between works councils and investments in pollution abatement, but also indicates that these investments tend to be employee-led.20

4. SOME EVIDENCE ON THE IMPACT OF EPRs AND SKILL DEVELOPMENT The theoretical framework developed in the previous section predicts a positive effect of EPRs on productivity. The theory also stresses the relationship between EPRs on the one hand, and development of FSHK on the other; these two factors interactively are predicted from the theory to have a positive effect on firm performance, other things equal. In this section, a simple and

Employee Participation Rights in Corporate Governance

119

preliminary test of this theory is presented, examining the relationship between EPRs, training and efficiency in a production function analysis. There is growing evidence that works councils lead to increased productivity at the establishment level (see e.g. Hubler & Jirjahn, 2003; Zwick, 2004; Schank, Schnabel, & Wagner, 2004). But these gains are not necessarily universal or robustly measured, as some studies show no (positive or negative) impact (see e.g. Schank et al., 2004). One reason for the ambiguity may be the failure to account for interactive effects with skill development; the approach used in this section represents a first step toward addressing this. The firm survey data used in the regressions cover 723 manufacturing firms in North Rhine-Westphalia, Germany’s most populous and most heavily industrialized state. These firms were divided into nine major industry groups as reported in Table 1.21 The data set includes information on our three variables of interest: works councils, training, and firm performance. Analysis of these data shows a statistically significant positive correlation between training, EPRs, and net sales (or alternatively sales per worker). A range of controls was used to ensure that results were not driven by firm size. The results in the following regressions may be interpreted as a production function analysis with industry dummy and interactive variables controlling for capital stock and variations in missing material use as a proportion of net sales. In Table 2, firm output in the full data set of 695 observations is examined.22 The dependent variable is log of net sales. The log of the labor force, as well as the log of labor force squared as an extra control for firm size, are introduced as independent variables; both are statistically significant, the former with a t-ratio of about 4.5 and the latter at about 1.9. To control for industry-specific effects, intercept dummies for eight of the nine industry groups are introduced, most of which are highly significant and all of which are at least marginally significant.23 Then, to allow for industry-specific Table 1. IND1 IND2 IND3 IND4 IND5 IND6 IND7 IND8 IND9

Definitions of Industry Groups.

Chemical, refining Plastics, rubber Stone and clay Metal fabricating and metalworking Steel, machinery, motor vehicle manufacturing Electrotechnical, precision engineering Pulp, paper, printing Leather, textiles, clothing Food, tobacco products

120

STEPHEN C. SMITH

Table 2. Variable

Dependent Variable: Log of Net Sales.

Parameter Estimate

INTERCEPT IND1 IND2 IND3 IND4 IND5 IND6 IND7 IND8 BETRIEBS TRAINING LOGL SLOPEL1 SLOPEL2 SLOPEL3 SLOPEL4 SLOPEL5 SLOPEL6 SLOPEL7 SLOPEL8 INTERACT LOGLABSQ N ¼ 695 Adjusted R2: 0.8241

0.225465 0.877647 0.969965 0.936180 1.121792 1.039577 0.990431 1.461102 1.260835 0.146723 0.008985 0.615697 0.203031 0.144407 0.201358 0.164825 0.144157 0.114449 0.249691 0.231317 0.076824 0.023280

Standard Error 0.38077919 0.45717008 0.43103727 0.57751612 0.36000613 0.34738559 0.38303838 0.37450942 0.52701987 0.07341364 0.10271169 0.13557326 0.10069849 0.10032918 0.14749181 0.08443990 0.08191450 0.08984239 0.08968667 0.12250512 0.13322036 0.01219481

T for H0: Parameter ¼ 0 0.592 1.920 2.250 1.621 3.116 2.993 2.586 3.901 2.392 1.999 0.087 4.541 2.016 1.439 1.365 1.952 1.760 1.274 2.784 1.888 0.577 1.909

Prob4|T| 0.5540 0.0553 0.0248 0.1055 0.0019 0.0029 0.0099 0.0001 0.0170 0.0461 0.9303 0.0001 0.0442 0.1505 0.1726 0.0514 0.0789 0.2031 0.0055 0.0594 0.5644 0.0567

labor elasticities, slope dummies are introduced, termed SLOPE1 through SLOPE8 in Table 2, most of which are statistically significant. Again, each of these two types of industry variables is also intended to help control for the missing data on capital use and further control for firm-size effects. Then, the labor variables may be interpreted as inputs in a translog production function, with industry data and other controls partially acting as proxy for unobserved capital data. With these several controls, BETRIEBS, the existence of a Betriebsrate, or works council, is positive and significant at the 5% level with a t-statistic of about 2. Net sales are some 15% higher in such firms, other things equal. Thus, even controlling for firm size, industrial sector, and industry-specific production processes, and training, works councils have positive effects on productivity, at least in this data set. A limitation of the data set is that no quantitative information on time in, and financial expenditures on, training is available. But the data set does

Employee Participation Rights in Corporate Governance

121

contain some information that serves as a proxy for significant training as defined in Section 2. In Table 2, the variable TRAINING refers to the existence of significant training activities in the enterprise and is constructed as follows. Managers were asked if the firm conducts worker training of duration longer than one day. If they indicated yes, they were then asked to indicate whether this training was to train employees to ‘‘perform a higher function within the enterprise’’ or just to function better at the same level. The former is chosen as the available variable most closely conforming to the predictions of the theory of the relationship between FSHK and EPRs set out in the previous section. The data limitations lead to some measurement error, but if anything lead to attenuation bias, making it more unlikely that the hypothesized relationships will be discerned in the data. The 19% of firms in the sample that both offered longer duration training, and then indicated the ‘‘higher function’’ intent of training, were then assigned a training value of 1; all others were assigned the value of 0. In the regression as reported in Table 2, this variable is positive but not statistically significant. The term labeled ‘‘INTERACT’’ captures the interaction between the presence of a works council and the presence of significant training as defined above; this interaction applies to some 11% of the firms in the sample. The coefficient is also positive but not statistically significant;24 but as we will see, results are far sharper when only smaller firms are considered. In the sample, 55% of firms had a works council, while 45% did not. Of 443 firms in our sample with less than 100 employees, just 136 have a works council, while amongst the 274 firms with more than 100 workers, 251 have a works council. Probit analysis in our sample on the incidence of a works council confirms that the probability that a works council is present is a statistically significantly increasing function of firm size, measured by labor force size and sales.25 Thus, there may be some remaining concern that what is being picked up is a size effect, in which larger firms are more likely to be productive because of scale economies, but are also more likely to have both councils and training. It is at least possible that results could remain biased for this reason, despite the fact that the (log of) labor force size, its square, and several industry and interaction variables correlated with size appear in the regressions. As a final control for this possibility, the sample was split between large and small firms. Focusing on the smaller firms also has the benefit of providing greater variability in the presence of works councils.26 In Table 3, firms with 100 or more workers are omitted and the effect of the presence of a works council, training, and the interaction of these two variables are considered only among the smaller firms. All controls for size and industry effects are retained in this new analysis; the coefficient on log

122

Table 3. Variable

STEPHEN C. SMITH

Dependent Variable: Log of Net Sales Firms with Less Than 100 Employees. Parameter Estimate

INTERCEPT IND1 IND2 IND3 IND4 IND5 IND6 IND7 IND8 BETRIEBS TRAINING LOGL SLOPEL1 SLOPEL2 SLOPEL3 SLOPEL4 SLOPEL5 SLOPEL6 SLOPEL7 SLOPEL8 INTERACT LOGLABSQ N ¼ 427 Adjusted R2: 0.4744

0.218365 1.681894 0.333548 1.053234 1.526940 1.461165 1.141280 1.224034 1.263729 0.131468 0.011840 0.436322 0.448555 0.042417 0.257235 0.294503 0.288268 0.177492 0.178093 0.242210 0.366275 0.031558

Standard Error 1.07904043 0.89129628 0.80146585 1.22545703 0.71486053 0.68011147 0.71972503 0.71601586 0.97275764 0.08754576 0.10859217 0.58938209 0.25741650 0.23089781 0.37331669 0.20935645 0.19820715 0.20834627 0.21221423 0.27104186 0.19148100 0.08418410

T for H0: Parameter ¼ 0 0.202 1.887 0.416 0.859 2.136 2.148 1.586 1.710 1.299 1.502 0.109 0.740 1.743 0.184 0.689 1.407 1.454 0.852 0.839 0.894 1.913 0.375

Prob4|T| 0.8397 0.0599 0.6775 0.3906 0.0333 0.0323 0.1136 0.0881 0.1946 0.1340 0.9132 0.4595 0.0822 0.8543 0.4912 0.1603 0.1466 0.3948 0.4018 0.3721 0.0565 0.7080

of labor squared becomes insignificant in this smaller and more homogeneous size range. The effect of works council presence remains positive, and at least marginally significant with a t-ratio of 1.5, with a predicted increase in net sales of some 13%. Training remains positive but not statistically significant. But now, the interaction effect between training and the presence of a works council becomes positive and statistically significant at just above the 0.05 level. More important, it is economically highly significant, with the coefficient estimate indicating that, with the full range of control variables in place, and already controlling for councils and training as stand-alone determinants of performance, the combination of training with works council raises net sales by a strikingly high 37%. In sum, links between firm performance and EPRs – by itself and interactively with training – are at least tentatively confirmed by these data. Although the data set has important limitations, and available evidence is

Employee Participation Rights in Corporate Governance

123

still insufficient to draw firm conclusions about the efficiency of EPRs in fostering further training in Germany, or the precise effect of this training on productivity, the direction of the findings is clear. The next step is to apply this type of analysis to data sets that examine firms over more than one year, and provide a wider range of observations on the firms, including more precise training, capital, and value added measures.

5. AN APPRAISAL OF RECENT TRENDS IN EPRS Do EPRs help or hurt economies that make them mandatory? Earlier in the paper it was noted that there is growing evidence that works councils lead to increased, or at least do not lead to decreased, productivity.27 But could their presence have negative effects not captured at the firm level? This is a large issue, which cannot be answered in full at this stage, let alone in a short article. But in this section, I will raise some questions and offer some hypotheses for further analysis. 5.1. Do Economies with EPRs Perform Differently? Worker participation, particularly legally mandated EPRs, may seem under assault throughout the world, battered by twin forces of globalization and technological change. Though often underestimated, in most parts of the developed world, substantial worker participation was well established a quarter century ago, and continued to evolve in the 1980s as new forms, notably employee stock ownership, rose to greater prominence. Forms of worker participation varied around the world but included the lifetime job in Japan, union membership in the US, and works councils and codetermination in Western Europe. The emergence of increasingly broad-based stock options in high-tech and other innovative firms in the US in the mid-1980s offered another potential channel for employees to gain decision-making as well as financial participation in the operations of the firm. Through economic forces, collective bargaining, and government regulation, the role of worker participation and its development as EPRs appeared to be on an inexorable rise. In recent years, however, unions have been in decline; in the wake of Enron and similar scandals, employee stock ownership and options has come to be perceived as carrying more risk and less upside potential. The lifetime job is turning out to be of shorter than expected duration in Japan and in companies in other countries where it was practiced, with rhetoric of

124

STEPHEN C. SMITH

the ‘‘virtual corporation’’ and the realities of outsourcing seeming to signal a secular rise in bargaining power of capital, and at least contributing to a defensive posture for labor. However, in the EU, works councils seem, if anything to be growing in prominence as the EU-wide works councils have been mandated and steadily implemented. Unions are much less in decline in some European countries than in the US. Although the newly established European councils have less power than some of the national councils, particularly in the case of Germany, still the trend appears different in Europe than North America or Japan, where EPRs were never well established, and works councils and codetermination in particular were never mandated or widespread. What are the underlying causes of these shifts, and what are their longrun economic and social consequences? This paper cannot fully address this large research agenda. But in this section, I will address in a preliminary way the question of whether the evidence warrants the conclusion that the US and other economies without EPRs are doing better than those with such rights, and if so, whether EPRs could be playing a negative role. Japan has been thought of in the literature as a key economy with considerable participation of employees in company affairs. After decades of rapid growth, since 1990, Japan has endured an extended period of relative stagnation. But for the arguments advanced in this paper, the critical point is that Japanese participation, to whatever degree it is or is not found in typical firms, remains informal; it carries no legal rights. The lifetime job had at best extended through the third of the economy that is the core, leading sector. The remaining two-thirds of jobs have been characterized as a kind of second-class economic citizenship. Unions are sometimes quite weak, company-union type organizations (Aoki, 1984). More recently, the dramatic decline of the concept of the lifetime job in Japan has not seemed to help the country to recover from its ongoing economic problems. Other rising East Asian economies also lack EPRs. But all of these economies are still growing by means of adapting technology from abroad. The theory of the role of EPRs in FSHK suggests that mandated works councils should be especially effective in companies whose competitive advantage rests in continuous innovation. Japan is doing relatively poorly without EPRs, and the US is (arguably) doing relatively well without them. How is the EU doing comparatively, given its relatively strong EPRs? The US has far less formal worker participation than the EU. The prevailing view is that in the dozen years since the single common market was forged, the EU – particularly the countries in the Eurozone – has barely grown while the US has forged ahead. The free market economy of the US,

Employee Participation Rights in Corporate Governance

125

with its ease of hiring-and-firing, is often claimed, for example in editorials in the Wall Street Journal, to explain the difference. But this view is far too simplistic, as a recent analysis by Kevin Daly shows.28 It is true that the Eurozone countries have a lower income per person than the US in purchasing power parity terms – approximately 30% lower. But this gap has not widened in the 1993–2003 period. Although US GDP has grown some 1% faster than that of the EU eurozone countries, the difference is explained by labor force growth largely resulting from immigration to the US and faster population growth. GDP per head has grown about 1.8% per year in the US, and by a nearly identical rate of 1.7% per year in the Eurozone. In addition, the labor productivity gap has actually been closing between the US and the Eurozone in this period; and that is the measure most relevant to the theories examined in this paper. Output per hour worked is now just 4% lower in the Eurozone than in the US; and in some leading economies with EPRs such as France, per hour productivity is higher than in the US. Based on this productivity change, Daly suggests that asset prices should grow no more slowly in Europe and the US, and in fact they have moved quite closely over the last 10 years. Thus, the presence of EPRs has not hurt the return on capital investments in the aggregate. The difference between the per-worker income and the per-population income is explained first and foremost by the fact that US employees work longer hours than their European counterparts. Eurozone employees worked an average of 15% fewer hours than US workers. The remaining 28%-lower labor utilization in the Eurozone may be explained by the smaller share of the adult population in employment. Some of this difference is explained by higher payroll taxes and other indirect costs in the EU than the US; but the per-worker cost of works councils is a tiny fraction of these costs. In recent years, the German economy has struggled, but it has several other special macroeconomic problems and dislocations to contend with, particularly the extremely large costs of adapting to unification. This does not seem to have anything to do with EPRs. In fact, works councils and codetermination may plausibly have facilitated the needed social and economic adjustments. Moreover, there is a growing incidence of employee voice in the US through ESOPs that feature employee voting rights and other forms of participation that begin to take on the character of genuine EPRs. The incidence, determinants, and impacts of this phenomenon needs further research, but nascent EPRs via participatory ESOPs could have already made a modest positive contribution to recent US productivity growth.29 Moreover, recall that although EPRs are generally mandated in Europe, they are not uniformly observed or applied. But where works councils are

126

STEPHEN C. SMITH

observed, higher, or at least not lower, productivity is observed ceteris paribus, as results in the previous section indicate. There are some other differences between the US and the EU that, while they have not manifested themselves in economic performance, have clear social implications and may yet have measurable economic effects. The EU countries have far fewer layoffs, either temporary or permanent. This means less loss in FSHK. They have fewer plant closings, which in addition to smaller loss of FSHK also means greater protection of social capital. The fact that people work shorter hours hardly seems a harsh indictment. Given the high level of wealth of both the EU and the US, taking some of that wealth in the form of higher leisure does not seem an unreasonable choice. At the same time, in part because leisure is not observable or valued in markets, the quest for positional goods and status may lead to too much work and too little leisure without regulation from the social point of view (for a review of some of the key arguments see Frank, 1987). Although a higher overall rate of labor force participation is probably desirable, it is unclear how a reform of EPRs would result in such a change. Further, in light of the analysis of time-horizon opportunism presented in Section 2, any apparently superior US investment performance may well be fundamentally short-term in nature. Moreover, since 2001 the US labor participation rate has fallen significantly. Finally, there is no expectation that the specific forms of EPRs are at this stage optimally designed in Germany or elsewhere in Europe. It is plausible that some problems with rigidity and uniformity of the structure and regulation of EPRs in these countries exist and could be improved upon. This is taken into consideration in discussing policies below. In sum, there is no decisive evidence that economies with EPRs perform better than those without. But neither is there any evidence that these economies are hurt or perform less well, and in a number of characteristics, European economies and individual companies with EPRs are performing well on economic and social indicators. 5.2. Why are EPRs Particularly Relevant to the Emerging Economy of the 21st Century? The organization of the firm and its employment relationships is likely to play a central role in determining productivity growth in the coming decades. In the emerging economy a more democratic capitalism should strongly benefit both efficiency, and social cohesion and equity. High trust and a mutual commitment to long-term cooperation between owners, managers, and employees is an integral component not only of competitiveness

Employee Participation Rights in Corporate Governance

127

today, but also of the capacity for innovation and productivity growth to build competitiveness for the future.30 And with the importance of firmspecific human capital, the very distinction between labor and capital in the enterprise is already breaking down, while both share a role as residual claimants. It might be countered that these observations typically apply to highly educated, skilled employees. But perhaps for this very reason they may be indicative of future trends in advanced economies. It is also plausible to argue that the US has more informal participation, made possible by increasingly flexible organizational structures; and that employee ownership is far more widespread in the US. Cheap telecommunications and computing have enabled a much more flexible and decentralized work organizations in many sectors, which almost by definition has transferred more decision-making autonomy to workers, while eliminating layers of bureaucracy. If worker participation can be measured in part by how few layers of authority lie between a worker and the top decision maker, worker participation may be judged to have increased. Of course, having their work downsized and outsourced is little comfort to the millions of factory workers and cubicle-dwellers who have learned overnight that they have lost their jobs after years of work for the same employer. White-collar outsourcing to India and elsewhere has caused understandable concerns.31 But despite the growing opportunities for productivity gains from flexible teams not just within corporations but also across corporate and national boundaries, it is a safe prediction that a majority of the economically active labor force will continue to work for corporations. Independent contractors and consultants typically accept such positions only after a layoff as a transitional arrangement before returning to conventional employment. Efficient firm boundaries and workplace employment levels may be changing, but is not plausible that they will disappear, and this implies that the efficiency case for some forms of EPRs will endure. Indeed, EPRs may help make needed ongoing adjustments in tasks in establishments without losing FSHK that continue to be relevant.

6. POLICY: CONSIDERING FLEXIBLE INCENTIVES TO CREATE EPRs THROUGH ‘‘PARTNERSHIP WORKPLACES’’ IN THE US In this section, I develop some policy proposals to promote EPRs that could raise the productivity of American workers, in a way that extends beyond the elite to the mainstream workforce, and improve the long-run performance of

128

STEPHEN C. SMITH

American firms. Such changes could help to reverse the dramatic increases in inequality of recent years and help to ensure a more democratic American capitalism. Despite such potential benefits, such changes are emerging slowly if at all, and policies to encourage them have not been followed.32 The reasons for this slow emergence include market and organizational failures, the short-time horizon of American companies in recent years (which itself results from several interlinked causes), a threat to management prerogatives, lack of information, and simple organizational inertia. The case for a government role may be summarized in seven categories: (1) The creation of basic knowledge resources is a key element of the plans outlined in this paper; the government generally has a significant role to play in creating such public goods. The most immediate knowledge resource is information about participation plans and what it takes to make them work effectively. There are also knowledge resource benefits that would be created by the more innovative economy, with the natural attendant cross-industry spillovers that these programs are designed to engender. (2) The government has a key role in promoting education; the Human Capital Partnership Plan (HCP Plan) is concerned directly (and the other plans indirectly) with education and skill levels of the public. (3) Society’s attitude toward income risk is likely to differ from the individual’s. Although the economy as a whole might be made more productive by the proposals in this paper, individuals’ concern for income stability may lead to failure to accept them unless there is some compensating incentive. Thus, government support for the package can help bring about what is in essence a public good. (4) Government action can be efficacious because the programs are strictly analogous to other efforts in which government has made critical contributions, especially agricultural extension programs. (5) Even though all firms might be made more profitable in the long run if each made investments in achieving ‘‘internal cooperative solutions,’’ these investments are not equally valuable at each point in time. If one firm discontinues these investments during periods of rapid growth, when they are less needed, it may be able to earn higher short-term profits and take market share from firms that continue them. This may doom the investing firm in the competitive marketplace before the value of its investments is revealed during a time of adversity. This is a classic prisoner’s dilemma; government incentives may help to ensure cooperative solutions.

Employee Participation Rights in Corporate Governance

129

(6) Decision-making participation may be threatening to managerial power and perquisites. Klein found that even first-line supervisors are strongly resistant to ‘‘employee involvement’’ programs: 72% of supervisors view these programs as being good for their companies and 60% see them as good for employees, but only 31% view them as beneficial to themselves.’’ Top management may be reluctant to risk supervisors’ alienation even to gain participation that offers a ‘‘favorable bet’’ that productivity will rise. In addition, unchecked authority of management creates inherent problems of opportunism, as described in detail in Section 3. The problem is structural: if managers are replaced, management in general retains the motive and the means to organize production in an opportunistic way. (7) Workers with FSHK are in effect among the residual claimants of the firm, as Margaret Blair has argued. If the firm goes bankrupt, workers share last place in line with the company owners to recoup their FSHK investments, which are useful only in that one firm and only if it is operating at all. If the net financial worth of the company is negative, employees lose all of their FSHK. In fact, it might be further argued that workers are last in line, because owners might recoup part of their investment (such as through the sale of plant and equipment) even as the firm shuts down. Moreover, owners count returns to FSHK as part of labor costs, rather than as part of the firm’s economic rents. As a result, workers with FSHK are subject to the risk of the residual claimant just as shareholders are – and perhaps more so. Indeed a significant fraction of Fortune 500 firms disappeared from the economic landscape in the past decade, through bankruptcies and various takeovers that can put FSHK at risk. Thus, there is a case for sharing decision-making rights with workers with FSHK from the point of view of designing law to maximize net social wealth. Each policy developed in this section has been designed to minimize government interference in firms. Thus, each policy is designed to ensure a cost-effective return in public benefits from the publicly supported programs consistent with enabling participating firms to retain substantive freedom of choice over the details of partnership workplace plans. These policies are specifically designed to respond to the analysis earlier in the paper, and structured to improve workplace training and organizational development through influencing firms to treat more employees as a long-term resource and effective partner in the firm – and employees to think of the long-term requirements of their companies.

130

STEPHEN C. SMITH

Profit and Gain Sharing Programs. First, the incentive to use the rights to participate in decisions, and to use them productively will be greater if there is explicit sharing of the resulting gains with employees. The economic literature shows that profit- and gain-sharing programs offer several positive workplace incentives; and the empirical results are almost unanimous in their confirmation.33 Perhaps the most important of these is the incentive for employers and employees to make investments – primarily in specific skills – necessary for more rapid productivity growth and the long-term health of individual enterprises (the building blocks of industry competitiveness). In this sphere the need for new mandates or tax breaks is not clear, especially because so little is known about which types of incentive programs are most effective. Thus, new forms of ‘‘extension services’’ to promote the growth of knowledge and disseminate information about what is and is not effective as it becomes available is probably the most important policy at this time. Employee Stock Ownership Plans. Second, ESOPs have a role to play in providing ordinary employees with a stake in their firms and in the American system. But most plans that were established during the ESOP boom were not designed primarily for purposes of organizational development. Instead, they have mostly been used as a cheap source of capital. But the evidence is that ESOPs only achieve productivity objectives when combined with an effective system of employee participation in decision-making.34 Moreover, when employee-owners do not have the rights to information and decision-making that other major owners could expect, and when they lack EPRs more generally, ESOPs have been subject to considerable abuses. The case of Enron was only the most widely reported debacle for employees who had counted on their ESOP shares for retirement or other needs. This experience shows that passive or even partially voting ESOPs are not enough; there is a need for legally mandated EPRs to be attached to ESOPs, particularly those that have been established with the benefit of tax incentives. Thus, firms could be required to commit themselves to introduce at least some decision-sharing features to receive any new benefits under the weakened but still available tax breaks now in force. Moreover, ESOPs might not be the only benefiting form of employee ownership under more flexible guidelines giving more discretion to firms while ensuring that public policy objectives are met. The conditioning of future tax benefits on genuine employee participation in decision-making could be designed in various ways. This would have to be verified for firms to be eligible for public assistance such as tax breaks; but an important objective in designing the policy would be to minimize

Employee Participation Rights in Corporate Governance

131

paperwork and other costs to firms. Three measures that are clearly beneficial and easy to establish and verify are (1) direct employee voting rights on stock, (2) employee board membership, and (3) workplace joint decisionmaking committees with clear authority in relevant areas. There may well be other effective measures that cannot be fully anticipated. Other proposals by companies about ways to achieve this goal suited to their particular circumstances could be given reasonable consideration. A HCP Plan, with its provisions for decision-sharing (discussed further below) would help qualify the firm on grounds (2) and (3). Firms might, for example, be expected to institute two of the three possibilities in an easily verifiable manner in order to receive additional tax benefits or other incentives. Each represents an aspect of what are generally understood to be the basic entitlements of ownership; ESOP firms will have earned their tax breaks and accomplished the public purpose when they show evidence of having extended full rights of ownership to their employees. Though the ESOP lobby may resist this change, in the end it will probably be necessary to protect any tax benefit for employee ownership. A strong advantage of these eligibility criteria is that they may be verified with minimal intrusion and paperwork. Initial qualification would only require a simple, one-time documentation that (1) employees have direct voting rights on stock (as indicated at the ESOP charter), (2) there is employee voting board membership (given in the board bylaws), and/or (3) there are workplace joint decision-making committees with clear authority in relevant areas (look at the appropriate employment contract agreement, committee charters or, where relevant, the HCP Plan). This could be done without any government official ever setting foot in a plant. No more paperwork would be involved than photocopying the relevant documents, such as explanations for employees about how the programs work, which would likely have been prepared in any event by any firm serious about employee partnership. After that, enforcement could be done on the basis of employee complaints. There would be no need for surprise or mandatory on-site inspection. But if such employee complaints were received that met certain credibility tests (such as numbers of complaints), the company would have to agree to a special review or administrative hearing, perhaps binding arbitration. Possibly a stronger version of arbitration would be required for a firm to keep its special status and benefits. The ESOP changes are necessary to make sure that tax breaks do lead to an increase in real, effective employee partnership. To the degree that these changes reduce the number of firms that can take advantage of tax breaks or other incentives, any savings could support extension of credits to other

132

STEPHEN C. SMITH

firms which might choose to introduce genuinely participatory forms of employee ownership appropriate to their own organizational characteristics but not corresponding to the tax supported ESOP form. Or, they could help fund extension services on other aspects of EPRs. More flexible guidelines would give firms more discretion in creating the system of employee ownership most suitable for their internal environment, while in the participation dimension, guidelines would be ‘‘tightened’’ to ensure that firms were complying with the real public policy objectives in return for public benefits. For some firms, for example, voting ESOPs will not be as appropriate as direct ownership, broad-based stock options plus employee seats on the board, or long-term profit-sharing (perhaps through simulated shares35 with less downside risk) plus management–employee workplace partnership committees. Some firms may wish to propose their own unique forms of participatory ESOPs. Unlike previous ESOP legislation, requirements for participation would become a condition for receiving tax or other benefits. But also unlike current ESOP legislation, benefits would not be limited to too narrow a choice of specific organizational forms. The fact that the ESOP is today the most common form of employee ownership is in large measure due to the cumulative incentive of the tax breaks they have conferred, especially in previous periods. Employee ownership is in itself a worthy goal of democratic capitalism, provided it does not result in employees putting too many of their assets in one ‘‘basket.’’ The policies are intended to promote a longer-term American corporate perspective, high total productivity, innovativeness, quality orientation, better and more flexible industrial relations, and centrally, to help get an urgently needed upgrading of Americans’ workplace skills underway. Among the options for accomplishing these objectives, this strategy has the worthy feature of helping to promote a more democratic capitalism. Human Capital Partnership Plans. Third, it is proposed that one of the criteria for selecting government contractors be whether a firm has developed what is termed here a HCP Plan. The intent of the plan is to provide a stake in company decisions to employees who develop skills relevant to building the firm’s innovativeness and capacity for high quality, or develop other abilities needed for high total productivity of that firm. The details would be left to the individual firm, but the plan would normally include one of several mechanisms for partnership decision-making at the workplace level or seats on the board of directors of the firm elected by these participating ‘‘human capitalists.’’ ‘‘Participation’’ would be defined legally, just as terms such as ‘‘business opportunity’’ are. When a company claims to

Employee Participation Rights in Corporate Governance

133

offer employees the chance to participate in decisions, this claim would be required to have a concrete definition, primarily to encourage ‘‘investors’ (i.e., the employees’) confidence’’ in the development of specific skills. Benefiting firms would also be expected to develop a plan, suited to their own special circumstances, to help less skilled workers ‘‘graduate’’ to ‘‘human capitalist’’ status; there might be separate incentives for doing so. The plan would offer a distinctly American version of works councils and codetermination, aimed at maximizing the productivity benefits while minimizing the potential drawbacks of that system. It would focus on one of the aspects of codetermination and/or works councils for which there is a clear economic rationale, the building of firm-specific skills and assets. The plan would encourage all workers to become more highly skilled and all firms to have an interest in training them. In these respects, the plans can help to reverse the dramatic increases in inequality between high- and low-skilled workers that have opened up in the last quarter century. After experience had been gained with alternative arrangements, research may determine which types had proven most effective, and incentives for selecting these forms could be considered. Firms that have implemented serious, ongoing training programs, with improved job security and genuine decision-making partnership, have typically improved their productivity.36 An integrated strategy involving these three elements is increasingly vital to company success. For example, longterm employment relationships are widely considered to be a prerequisite for participative employer–employee relations. Yet, despite the presence of a few heralded examples, relatively few American companies have followed this path, and on a rather limited scale; perhaps these changes are perceived as too risky and threatening to the status quo. If anything, employment security in particular has become even more tenuous in recent years. Public inducements and guidelines have an important role to play in a general, effective campaign to transform the American workplace. The HCP Plans would be designed to encourage employees to develop skills specific to the firms in which they work, to in effect make investments in the long-term health of their employers. These investments might include some form of employee ownership or other financial participation as well as human capital specific to the needs of the firm. In return, firms would offer a long-term employment understanding; the new partnership and understandings would be validated and supported by the firm sharing specific decision-making rights with their new ‘‘human capitalist’’ partners. The role of employees with FSHK as de facto residual claimants would be formally recognized.

134

STEPHEN C. SMITH

Some form of incentive for employees to work toward the long-term competitiveness and profitability of the firm, plus the participation that will provide the channels for employee ideas to be put into practice, are natural complements, and appropriate incentives can encourage firms to create this mix. On the other hand, these are guiding principles rather than sharply defined specific policies that clearly dominate all similar policies. Companies’ own judgments and the marketplace can help sort out which specific combination of programs make sense in which industries. If government were to be involved with implicit micromanagement of the details of this choice, the program could fail through firm or union resentment or resistance, if not outright policy errors on the part of government. And at the same time, a claim by management that employees will participate in decision-making must mean something concrete if it is to secure the firmspecific investments it is intended to achieve. Thus, the HCP Plan would have to meet broad guidelines to receive public benefits or incentives (or at some point possibly mandates), but these would not include highly detailed, specific procedures; details would be left up to the firm (a framework similar to that proposed above for ESOPs would apply). Moreover, firms would be given the leeway to evolve the plans in consultation with the appropriate employee representatives as they acquired more experience with them. It is legitimate to ask how employment security may be enhanced at a time when flexibility has become so critically important to companies, and competition has grown ever sharper, including from companies that outsource from India. So far, only a small fraction of jobs have been outsourced. But manufacturing jobs have been decimated, with more than two million lost in the US since 2000 alone. Moreover, although white-collar outsourcing, particularly to India, has been modest so far, it may accelerate. First, employment security need not be offered to all workers, or offered upon employment, but may be limited to those who satisfy certain criteria, and may be of more limited scope than traditional tenure. Professors know that some colleagues awarded tenure respond by working harder, such as in tackling far longer-term and risky research programs with potentially high payoffs; and by making more relationship-specific investments, such as by directing unique academic programs. There is also a potential for abuse, which is why the tenure decision is made so carefully and only after an extended period of observation and professional development, and there are limits – tenure is revocable for incompetence or contractual violations. Some qualitatively similar arrangements could be established in business if some basic organizational designs were available. Such arrangements have long

Employee Participation Rights in Corporate Governance

135

been common in the professions; and even though partnerships are in decline as a form, this does not mean that forms of job tenure should be unattainable. This model becomes more relevant as more jobs take on a professional character. More generally, there could be an explicit agreement that when outsourcing is under discussion, the works council or other representative body is consulted first, and given a reasonable time to find alternative strategies that would lead to comparable savings or quality, and/or to work together strategically to find alternative niches for which the workforce has a continued or emerging comparative advantage. There are substantial benefits to be gained from encouraging such arrangements, and these can be achieved without mandating highly specified procedures. Through developing alternative forms of human capital partnership programs, each company can become increasingly like a small laboratory seeking improvements in competitiveness. This in turn may lead to an accelerated adoption of other improvements in workplace organization and productivity complementary to high skills and employee involvement. Over time, favorable practices would spread to other firms in the economy, and inevitably, some of these improvements would also get transferred abroad. Continuous improvement will of necessity become a fundamental part of the way business is done in advanced economies. Kanter (1997, pp. 133, 190–194) suggests that part of the answer is ‘‘employability security.’’ She is sober about its limits, but proposes reasonable remedies to them. As she argues: Employability security has become a catchphrase at some companies. Sometimes it is used to mask brutal intentions, such as preserving the right to cut employment at any time and with little notice. But done correctly, with the right values and intentions, employability security can actually enhance long-term loyalty. By offering ongoing learning to upgrade skills and by spreading the power to innovate, companies help employees to perform better, which allows the companies to perform better, which in turn preserves and expands jobs. Such practices also ensure that current employees continue to be important contributors who are valued by their employers and offered opportunities to continue to grow.

Kanter has also proposed an ‘‘Employability security contract’’ [in ‘‘A New Human Resources Agenda’’ on pages 190–194.] Her model contract reads as follows: Our company faces competitive world markets and rapidly changing technology. We need the flexibility to add or delete products, open or close facilities, and redeploy the work force. Although we cannot guarantee tenure in any particular job or even future

136

STEPHEN C. SMITH

employment, we will work to ensure that all our people are fully employable-sought out for new jobs here and elsewhere. We promise to: – recruit for the potential to increase in competence, not simply for narrow skills to fill today’s slots. – offer ample learning opportunities, from formal training to lunchtime seminars – the equivalent of a month a year. – provide challenging jobs and rotating assignments that allow growth in skills even without promotion to ‘‘higher’’ jobs. – measure performance beyond accounting numbers and share the data to allow learning by doing and continuous improvement. – retrain employees as soon as jobs become obsolete. – recognize individual and team achievements thereby building external reputations and offering tangible indicators of value. – provide three-month educational sabbaticals or external internships every five years. – find job opportunities in our network of suppliers, customers, and venture partners. – tap our people’s ideas to develop innovations that lower costs, serve customers, and create markets – the best foundation for business growth and continuing employment.

This is one possible strategy that could be made to fit within the HCP concept. The issue once again is whether it will have contractual force, and how it will be enforced. A HCP committee, such as a works council or similar arrangement, could be charged with codetermining its implementation. The agreement may be made part of the employment contract and constituted in such a way that it would be enforceable legally, but at the same time, set up in such a way as to facilitate and encourage internal resolution of any disputes in the overwhelming majority of cases. And it would need to spell out the most important practical issues. For example, if the company is offering security, under the HCP the company may be expected to agree to pay and help the employee to actually get the new job that is supposed to be securely available. But the employability security notion is one starting approach to adapting these ideas to the changing realities of the workplace. Clearly, this can either be done in a superficial way that only creates cynicism and lack of genuine workplace partnerships or with real empowerment that has substantive impact. In this regard, one modest but potentially highly significant and broadly applicable public step would be to restrict the use of terms like employee participation to clearly understand substantial meanings. Indeed, such terms will need contractual weight if they are to elicit the full competitiveness benefits expected from the partnership workplace. As these concepts continue to catch on among businesses and employees, their importance as a

Employee Participation Rights in Corporate Governance

137

selling point in company recruitment efforts is likely to grow. To cite one relevant precedent, in more than half of the states, the term ‘‘business opportunity’’ has been defined in a legal statute, and its use is regulated by state agencies. Thus, in order to take advantage of whatever incentives are offered for providing participation plans, a firm would be required to commit itself to a legal concept of what participation will mean, which gives appropriate security, including information disclosure, to affected employees. Tax benefits may be used to encourage HCP plans. However, it may be more appropriate to use preference criteria in procurement. Like other preference criteria for government contracts, such as generally offered for veterans or for firms with minority participation in FCC licensing decisions, the presence of an HCP Plan would be made only a part of the decision (and so would not be required). This approach would minimize the intrusion and paperwork demands for firms for which such a plan might prove of less immediate benefit. One advantage of using preference criteria instead of tax breaks or other fiscal inducements is that it would not reduce government revenues or increase expenditures in face of the recent record budget deficits. It would thus reduce the risk that firms would find a way to take advantage of government largesse without adhering to the public policy purpose. And we do not wish to undermine needed macroeconomic adjustment, as the twin deficits have returned with a vengeance, while encouraging the needed organizational adjustment. On the other hand, a strictly mandatory requirement would probably be premature if not unworkable in the American context. But this policy strategy may be expected to build up a growing sector of the economy experimenting with different versions of the plan, from which the most effective form(s) of the plan could later be determined. At that time, tax incentives might be added. Finally, even persuasion and exhortation by appropriate public officials as part of a national competitiveness and democratic capitalism campaign might be enough to tip the balance in some cases. In sum, the policy goal is to give a stake in company decisions to those employees who build skills relevant to the firm’s innovativeness and capacity for high quality, or acquire other abilities needed for high total productivity of that firm. Firms would also be expected to develop a plan to help less skilled workers ‘‘graduate’’ to ‘‘human capitalist’’ status; separate incentives for doing so are envisioned, as part of an integrated national human capital strategy. The details could be left to the individual firm and the collective bargaining process where applicable, but unless the firm proposes a reasonable alternative, it would normally include one or more of the following: seats

138

STEPHEN C. SMITH

on the board of directors of the firm elected by these participating ‘‘human capitalists,’’ mechanisms for joint decision-making at the workplace level, and/or expanded and well-understood company information flow procedures. The HCP Plan would represent an American version of European codetermination and works councils, without its potential drawbacks; it would focus on the specific human capital aspect of codetermination and works council law, the aspect for which there is the clearest economic rationale, and now, given the results in this paper, some supporting evidence. This approach would work well in the American environment, offering a more flexible and voluntaristic as well as more finely honed policy instrument. The slow emergence of such arrangements is not for lack of employees wanting such a plan. Even in the 1980s, A US Chamber of Commercecommissioned poll showed that 84% of the American workforce would like the chance to participate in management decisions, while a survey by Peter Hart Associates showed that two-thirds of Americans would prefer to work in a firm with employee ownership (Jones, 1987, p. 493). A 1994 EBRI/ Gallup national poll found employees more likely to prefer an ownership share in company than having a larger paycheck; and 80% said employers should be allowed to contribute company stock to fund the retirement plan (Kruse & Blasi, 1999). Although the Enron debacle may lead to greater caution on employee ownership, the desire for a say in decision-making to protect firm-specific human capital as well as employee ownership shares may be stronger than ever. But new polling could help to design such instruments in ways that would be most acceptable and effective.

7. CONCLUDING REMARKS This paper has examined employee participation rights – legal or contractual guarantees of rights of employees to consultation, or veto power, and/ or joint design and implementation of specific workplace decisions and practices; German works councils are a key example. It was shown that in theory EPRs might raise workplace efficiency by enhancing the quantity and quality of investment in skills at the workplace. Among other empirical findings, in Lower Saxony, Germany, among small and medium enterprises with less than 100 workers, the interaction of training with works council presence raises net sales by a statistically and economically significant 37%. The case for EPRs does not rest solely on productivity improvements, and there are data quality limitations for the current study, but these results suggest that EPRs may provide efficiency gains. The paper proposed modest

Employee Participation Rights in Corporate Governance

139

incentives, such as preference criteria for government contracts, for US firms to provide specialized EPR agreements (termed Human Capital Partnership plans). Other policies would require full ownership rights for ESOP participants, legally regulate the use of terms such as ‘‘participation,’’ and develop workplace ‘‘extension services’’ regarding the state of knowledge about EPRs and related institutions. The policy path proposed in this paper is challenging, but some reorganization of the workplace is essential to ensure an efficient and dynamic, and more democratic, American economy. New policies are needed to reduce the exposure of employees to investment risks – particularly firm-specific human capital – over which they have no control. They are also needed to address the growing job stress facing American workers. The result of the policies, particularly the human capital partnership programs, would be a real stake for employees, particularly those who continuously upgrade their skills. There may be resistance from managers who feel their prerogatives threatened by these changes, unions that have long resisted partnership in responsibility for the firm, employees who are not used to continuously upgrading their skills, and stockholders who fear expanded employee participation or profit sharing could slow dividend growth. The HCP proposal is targeted at the point where competitiveness can be most affected: the building of firm-specific human capital skills. This is where partnership with employees and sharing management decision-making in the company is needed both for efficiency gains and for a more democratic capitalism. The package of labor market adjustments discussed in the paper, most notably human capital development compacts, gain sharing, greater employee ownership, employment security for human capitalists, shared decision-making, and probably a guaranteed minimum income to counterbalance increased volatility of labor market income, are among those most critical to the efficient and fair functioning of the emerging economy. Jack Orsburn and his colleagues37 draw from wide experience with implementing participation programs in a familiar pattern that they call the Algernon syndrome, after the Flowers for Algernon (‘‘Charley’’) story about a retarded man who participates in an experiment raising his intelligence and creativity to the highest levels, but later sinks back to his earlier state. ‘‘Like this unfortunate man, mature teams will regress – feeling betrayed, frustrated and hostile – if management fails to meet their needs out of neglect or because of some unadvertised motive for implementing teams (to reduce headcount, say, or to excise managers, cut costs, or create ‘‘happier employees’’)yulterior management motives or lack of tangible support usually surface and often precipitate a steep decline in team productivity’’ (Orsburn

140

STEPHEN C. SMITH

et al., 1990, pp. 143, 157). This helps explain the temporary nature of gains in some of the experiments. The problem, seen in the lens of the analysis of this paper, is that participation has lacked real basis in a permanent and verifiable company commitment or genuine employee participation rights. These unfortunate experiences underscore the need for serious, long-term policies like those outlined in this paper. Employee investments of all kinds will need at least the same protections that financial securities receive, or underinvestment, of human capital, and complementary physical capital will be the likely result. These policies will be needed to ensure both the economy’s efficiency and its fairness, as the need for workers and workplaces to steadily improve skills in the developing knowledge economy. In sum, the partnership workplace plan can help raise long-run productivity and incomes, while helping to build a more equitable American society.

NOTES 1. The two components are often collectively termed codetermination, though this is not strictly accurate from a legal standpoint. Note also that German company law provides for a ‘‘management board’’ composed of professional managers who run day-to-day operations, and a ‘‘supervisory board’’ composed of elected representatives of shareholders and, under the codetermination system, of elected representatives of employees. The supervisory board appoints and dismisses the management board and supervises important strategic company decisions. 2. This may be done indirectly through a kind of electoral college. 3. See Buchtemann, 1989, cited in Sadowski and Decker, 1993, and Mu¨llerJentsch (1995), in Rogers and Streeck (1995). 4. A full account is found in Ministry of Labor and Social Affairs (1991, 1992). 5. For details see Rogers and Streeck (1995). Earlier discussions are found in Summers (1980, 1982), Kuhne (1980), Aoki (1984), Kolvenback (1978, 1982), CarbyHall (1977), and IDE (1981). 6. For an intriguing and potentially applicable alternative starting point that might also be applied to the analysis of EPRs in the context of protection of FSHK, see Rajan and Zingales (1998, 2001). 7. General training can be identified with programs that provide general certification schemes, covering general rather than firm-specific qualification gradations. In theory, the comparative advantage in managing general training would seem to lie in an educational institution or perhaps a company specializing in training, rather than on-the-job training in a firm with EPRs; but this conclusion may be reversed when there are economies of scope between the two types of training coupled with experiential learning advantages. I would like to thank Dieter Sadowski for helpful discussions on the dimensions of bundling general and specific training. 8. While I make these claims on the basis of subjective experience as a university professor rather than a vocational skills instructor, and they would need to be tested

Employee Participation Rights in Corporate Governance

141

empirically, these principles should if anything operate all the more strongly in vocational education. 9. Since the same set of incentives hold for all managers, the owners of the firm may be able to do little about the problem merely by dismissing management (Dow, 1987, p. 24), and in any case micro-monitoring of managers by shareholders is impractical (Coffee, 1986). 10. Other opportunistic promotion and organizational design incentives are developed in Smith (1991). 11. See Smith (1988, 1994) for further discussion of incentives for employee contributions to innovation. 12. Jensen and Meckling (1979) and Furubotn (1985) have implicitly assumed a first best world in which management is the perfect agent of capital when they argued that codetermination might have negative time-horizon implications by involving employees with finite horizons in decisions. Moreover, the time horizon of stock markets in the US has become increasingly viewed as surprisingly short, from the perspective of traditional efficient market theory. 13. For a review of the arguments see Steven Pearlstein, ‘‘As Firms Go ‘Lean and Mean,’ Benefits, and Dangers, Emerge. Economists Fear Restructurings in ‘93 May Have Gone Too Far,’’ Washington Post, Sun., Jan 2, 1994, pp. H1, H16. 14. For a full discussion of first and second best authority structures in the presence of transaction costs and uncertainty, see Smith (1991). 15. This is a market failure for which policy may be a necessary corrective. For a general equilibrium analysis see Dreze and Hagen (1978) and Dreze (1976). 16. The study was conducted by Northwestern National Life Insurance, reported in the Wall Street Journal, May 7, 1991, p. 1. This finding has been confirmed in subsequent research, some of which is reported at http://www.hazards.org/workedtodeath, website accessed March 2, 2004. 17. See for example the BBC reports at http://news.bbc.co.uk/1/hi/health/ 2993116.stm and http://news.bbc.co.uk/1/hi/health/3124783.stm, websites accessed March 2, 2004 18. For a detailed discussion, in the context of high-performance workplaces, see Levine (1995). 19. For an applicable formal model, see Dow (1993). For an examination of a variety of related issues, in the somewhat different context of cooperative firms, see Smith (2003). 20. See Askildsen, Jirjahn, and Smith (In press). 21. The survey was undertaken by the Institut Arbeit und Technik (IAT, or Work and Technology Institute) in Gelsenkirchen in 1991. Industry groupings were formulated by IAT researchers. I am indebted to the IAT, particularly Matthias Knuth, for making the survey available to me in electronic form for this analysis, and for their hospitality during my visit there. 22. This is the number of firms remaining after deleting the firms with relevant missing variables. 23. Of course, only 8 of 9 industries can be introduced into the regression to avoid perfect multicollinearity. The last industry dummy, dropped to prevent multicollinearity, is that of the food processing industry (see Table 1); it also turns out to be the most productive of the industry groups by this measure (hence the many reported negative coefficients on the industry dummies).

142

STEPHEN C. SMITH

24. Training, like other investment activities, is an inherently intertemporal investment, and our data offers only a one-time ‘‘snapshot’’ of the firm. Because of the lack of time series data, it is impossible to know if these were investments that later paid off more substantially. 25. With a full range of controls, the w2 significance levels are 0.0015 for log of labor and 0.0278 for log of sales. The full probit analysis results are available from the author. 26. This raises the question of whether the presence of councils should be treated as an endogenous variable. However, since works councils generally address a far broader set of decision areas beyond training matters, a request for a works council by employees is generally separate from subsequent bargaining over the degree of training. Demands that a council be established come from a small number of workers or union representatives, but the councils themselves are elected by secret ballot by all covered employees. Although it is still possible that council presence is endogenous in a way that affects the findings, no plausible instruments are available for council presence. Even in this case, I would argue that it is a valuable and new piece of information to know the conditional expectations even if council presence, for example, is endogenous. For an exploration of determinants of council presence using a different data set with a set of variables well suited for that purpose, see Jirjahn and Smith (in press). 27. For example, Addison, Schnabel, and Wagner (2001) Addison et al 2001, Hubler and Jirjahn (2003), Zwick (2004), and Schank et al. (2004). The first of these papers observes higher wages and lower profits with works councils, but these estimates, while interesting, are less convincing. 28. These observations are based on OECD data, and are nicely summarized and analyzed in Daly (2004). 29. For relevant evidence, see e.g. Gates (1998); Blasi, Conte, and Kruse (1996); Blasi and Kruse (1991); and Jones and Pliskin (1991). 30. This may be thought of a cooperative game outcome (Aoki, 1984; Smith, 1991); sociologists such as Wolfgang Streeck have considered the same issues and reached analogous conclusions from an industrial sociology viewpoint in using the term ‘‘high trust.’’ 31. Commentators in leading newspapers have downplayed the significance of white-collar outsourcing: for example Robert Samuelson in the Washington Post, Jan. 14, 2004, and Christopher Caldwell in the Financial Times, Feb. 7/8, 2004. Arguably, a significant fraction of the jobs going to India are the type of positions on their way to being sufficiently routinized to be fully replaced with computers, such as through web-based applications and artificial intelligence (Levy & Murnane, 2004). Of course, this does not apply in the case of some higher skill software and research applications now increasingly done in India (Friedman, 2005). Nor is additional education always the answer – some of the jobs being outsourced to India require professional or masters degrees. However, there are major retraining and market repositioning issues, and EPRs may help with the needed adjustments. 32. I developed an early version of some of these proposals in an unpublished report for the Progressive Policy Institute in 1991 (revised, 1992), which in turn served as a background paper for the Democratic Leadership Council Cleveland Proclamation, specifically its resolution on ‘‘meeting the global challenge’’ by supporting ‘‘workers efforts to become more competitive’’ (http://www.ndol.org/documents/ cleveland_proclamation.pdf). Although the proposals were not taken up by the

Employee Participation Rights in Corporate Governance

143

incoming Clinton administration, now may be a good time to revive and develop these proposals for public debate, given the scandals over employee ownership such as Enron, new concerns about job outsourcing to developing countries – now extending to some of the very types of employment thought in 1992 to be the skilled jobs of the future – and the growing concern over steadily increasing inequality in American society. Meanwhile, works councils in Europe have proven themselves over the past dozen years to be remarkably durable and effective institutions even under heightened competition from the US and developing countries, and the considerable stresses of EU enlargement. 33. See e.g. Kruse and Blasi (1997), and several contributions in previous volumes in this Series. 34. See Jones and Pliskin (1991), Levine and Tyson (1990), and Kruse and Blasi (1997). 35. Simulated shares are not tradable claims to ownership but are instead accounting devices which offer gain or profit sharing to employees based on the performance of the company’s stock rather than some other measure such as current declared profits. See for example Meade (1986). 36. Levine (1995). Note that in a period in which job stress and hours of work (even to the extent of ‘‘voluntary’’ give-backs of vacation days) are increasing, the healthimproving and stress-reducing characteristics of employee empowerment are also potentially significantly welfare enhancing, but are beyond the scope of this study. 37. Orsburn et al. (1990). The authors detail an 8-stage process for a conventionally managed company to reach fully self-directed work team organization, starting with information sharing, dialogue, several stages of group problem solving and expanding self-direction.

ACKNOWLEDGMENTS This paper was presented at the conference on ‘‘Corporations, Markets, and the State: A Multidisciplinary Inquiry into the Future of Modern Global Business,’’ George Washington University, March 4–6, 2004. I would like to thank conference participants and an anonymous referee for useful suggestions on the paper. I have also greatly benefited from discussions with coauthors on another paper on works councils in Germany, Jan Erik Askildsen and Uwe Jirjahn. In addition, this article draws on two earlier discussion papers, prepared for Economic Policy Institute and Progressive Policy Institute, respectively. A version of the EPI discussion paper appeared in Spanish in 1995 as ‘‘Derechos de Participacion, Formacion y Eficiencia de los Trabajadores: Hipotesis y Evidencia Empirica para Alemania,’’ in Economiaz. I would like to thank Eileen Appelbaum, Peter Berg, David Levine, Matthias Knuth, Will Marshall, Lee Price, Dieter Sadowski, Rob Shapiro, Wolfgang Streeck, and Kirsten Wever for valuable discussions while preparing those earlier documents.

144

STEPHEN C. SMITH

REFERENCES Addison, J. T., Schnabel, C., & Wagner, J. (2001). Work Councils in Germany: Their effects on establishment performance. Oxford Economic Papers, 53, 659–694. Aoki, M. (1984). The cooperative game theory of the firm. Oxford: Oxford University Press. Askildsen, J. E., Jirjahn, U., & Smith, S. C. (In Press). Works councils and environmental investment: Theory and evidence from German panel data. Journal of Economic Behavior and Organization. Bain, T., & Hester, L. K. (2000). Similarities and differences in European Works Council agreements. Proceedings of the Industrial Relations Research Association (pp. 45–51) USA, MA: Boston. Batt, R. (1993). Work reorganization and labor relations in telecommunications services: A case study of BellSouth corporation, MIT Ind. Perf. Ctr. Working Paper 004-93WP, August. Blair, M. M. (1995). Ownership and control. Rethinking corporate governance for the twenty-first century. Washington: Brookings. Blair, M. M. (1999). Firm-specific human capital and theories of the firm. In: M. M. Blair & M. J. Roe (Eds), Employees and corporate governance. Washington: Brookings. Blasi, J., Conte, M., & Kruse, D. (1996). Employee ownership and corporate performance among public corporations. Industrial and Labor Relations Review, 50, 60–79. Blasi, J., & Kruse, D. (1991). The new owners: The mass emergence of employee ownership in public companies and what it means to American business. New York: HarperBusiness. Carby-Hall, J. R. (1977). Worker participation in Europe. London: Croom Helm. Coffee, J. C. (1986). Shareholders vs. Managers: The Strain in the corporate web. Michigan Law Review, 85, 1–109. Daly, K. (2004). Euroland’s secret success story. Goldman Sachs Global Economics Paper No. 102. Dow, G. K. (1987). The function of authority in transaction cost economics. Journal of Economic Behavior and Organization, 8, 813–838. Dow, G. K. (1993). Why capital hires labor: A bargaining perspective. American Economic Review, 83, 118–134. Dreze, J. (1976). Some theory of labor management and participation. Econometrica, 44, 1125–1140. Dreze, J., & Hagen, K. P. (1978). Choice of product quality: Equilibrium and efficiency. Econometrica, 46, 493–513. Frank, R. H. (1987). Choosing the right pond: Human behaviour and the quest for status. New York: Oxford University Press. Friedman, T. (2005). The world is flat: A brief history of the twenty-first century. New York: Farrar, Straus and Giroux. Furubotn, E. G. (1985). Codetermination, productivity gains and the economics of the firm, Oxford Economic Papers, 22–39. Gates, J. (1998). The ownership solution: Toward a shared capitalism for the 21st century. Portland, OR: Perseus Press. Grossman, S. J., & Hart, O. D. (1986). The costs and benefits of ownership: A theory of vertical and lateral integration. Journal of Political Economy, 94, 691–719. Hubler, O., & Jirjahn, U. (2003). Works councils and collective bargaining in Germany: The impact on productivity and wages. Scottish Journal of Political Economy, 50(4), 471–491. IDE Research Group. (1981). Industrial democracy in Europe. Oxford: Clarendon.

Employee Participation Rights in Corporate Governance

145

Jensen, M. C., & Meckling, W. H. (1979). Rights and production functions: An application to labor-managed firms and codetermination. Journal of Business, 52(4), 469–506. Jirjahn, U., & Smith, S. (in press). What factors lead management to support or oppose employee participation – with and without works councils? Hypotheses and evidence from Germany. Industrial Relations, (Forthcoming). Jones, D. (1987). Alternative sharing arrangements: A review of their effects and some policy implications for the US. Economic and Industrial Democracy, 8, 224–249. Jones, D. C., & Pliskin, J. (1991). The effects of worker participation, employee ownership, and profit-sharing on economic performance: A partial review. In: R. Russell & V. Rus (Eds), Ownership and participation international handbook of participation in organizations (Vol. 2, pp. 43–63). New York: Oxford University Press. Kanter, R. M. (1997). On the frontiers of management. Boston: Harvard Business Review Book. Karasek, R. A., et al. (1988). Job characteristics in relation to the prevalence of myocardial infarction in the US health examination survey and the health and nutrition examination survey. American Journal of Public Health, 78, 910–918. Klein, J. A. (1984). Why supervisors resist employee involvement. Harvard Business Review, 87–95. Kolvenback, W. (1978). Employee councils in European companies. Deventer: Kluwer. Kolvenbach, W. (1982). Cooperation between management and labor. A Survey of co-determination mechanisms in European corporations. Deventer: Kluwer. Kruse, D., & Blasi, J. (1997). Employee ownership, employee attitudes, and firm performance: A review of the evidence. In: D. Lewin, D. J. B. Mitchell & M. A. Zaidi (Eds), The Human Resources Management Handbook, Part 1. Greenwich, CT: JAI Press. Kruse, D., & Blasi, J. (1999). Public opinion polls on employee ownership and profit sharing. Journal of Employee Ownership Law and Finance, 11, 3–25. Kuhne, R. J. (1980). Co-determination in business. Workers’ representatives in the boardroom. New York: Praeger. Levine, D. I. (1995). Reinventing the workplace: How business and employees can both win. Washington: Brookings. Levine, D. I., & Tyson, L. D. (1990). Participation, productivity and the firm’s environment. In: A. S. Blinder (Ed.), Paying for productivity: A look at the evidence. Washington: Brookings. Levy, F., & Murnane, R. J. (2004). The new division of labor: How computers are creating the next job market. Princeton: Princeton University Press. Marginson, P. (2004). The impact of European Works Councils on management decisionmaking in UK and US-based multinationals: A case study comparison. British Journal of Industrial Relations, 42, 209–233. Meade, J. (1986). Alternative systems of business organization and of workers’ remuneration. London: Allen and Unwin. Ministry of Labor and Social Affairs. (1991). Codetermination in the Federal Republic of Germany, Bonn. Ministry of Labor and Social Affairs. (1992). Labor law in Germany: An overview, Bonn. Mu¨ller-Jentsch, W. (1995). Germany: From collective voice to co-management. In: J. Rogers & W. Streeck (Eds), Works Councils – consultation, representation, and cooperation in industrial relations (pp. 53–78). Chicago: University of Chicago Press. Orsburn, J. D., et al. (1990). Self-directed work teams: The new American challenge. Homewood, IL: Business One Irwin. Putterman, L. (1982). Some behavioral perspectives on the dominance of hierarchical over democratic forms of enterprise. Journal of Economic Behavior and Organization, 3, 139–160.

146

STEPHEN C. SMITH

Rajan, R. G., & Zingales, L. (1998). Power in a theory of the firm. Quarterly Journal of Economics, 113, 387–432. Rajan, R. G., & Zingales, L. (2001). The firm as a dedicated hierarchy: A theory of the origins and growth of firms. Quarterly Journal of Economics, 116, 805–851. Rogers, J., & Streeck, W. (Eds) (1995). Works councils: Consultation, representation, and cooperation in industrial relations. Chicago: University of Chicago Press for National Bureau of Economic Research. Sadowski, D., & Decker, S. (1993). Contractual regulations on continuing vocational education and training in Germany. Trier, Germany: Institut fur Arbeitsrecht und Arbeitsbeziehungen in der EG, for the European Commission. Schank, T., Schnabel, C., & Wagner, J. (2004). Works councils – sand or grease in the operation of German firms. Applied Economics Letters, 11, 159–161. Shiller, R. J. (2000). Irrational exuberance. Princeton: Princeton University Press. Shleifer, A., & Summers, L. (1988). Breach of trust in hostile takeovers. In: A. Auerbach (Ed.), Corporate takeovers: Causes and consequences. Chicago: University of Chicago Press. Smith, S. C. (1988). On the incidence of profit and equity sharing: Theory and an application to the high tech sector. Journal of Economic Behavior and Organization, 9, 45–58. Smith, S. C. (1991). On the economic rationale for codetermination law. Journal of Economic Behavior and Organization, 16, 261–281. Smith, S. C. (1994). Innovation and market strategy in Italian industrial cooperatives: Econometric evidence on organizational comparative advantage. Journal of Economic Behavior and Organization, 23, 303–321. Smith, S. C. (2003). Network externalities and cooperative networks: Stylized facts and theory. In: L. Sun (Ed.), Ownership and governance of enterprises, UN-WIDER (pp. 181–241). Basingstoke, Hampshire, England: Palgrave/Macmillan. Stirling, J., & Tully, B. (2004). Power, process, and practice: Communications in European Works Councils. European Journal of Industrial Relations, 10, 73–89. Streeck, W. (1984). Codetermination, the fourth decade. In: B. Wilpert & A. Sorge (Eds), International perspectives on organizational democracy. New York: Wiley. Summers, C. W. (1980). Worker participation in the U.S. and West Germany: A comparative study from an American perspective. The American Journal of Comparative Law, 28, 367–392. Summers, C. W. (1982). Codetermination in the United States: A projection of problems and potentials. Journal of Comparative Law and Securities Regulation, 4, 155–191. Teece, D. (1982). Towards an economic theory of the multiproduct firm. Journal of Economic Behavior and Organization, 3, 39–63. Vanek, J. (1971). The participatory economy. Ithaca, NY: Cornell. Wever, K. S. (1994). Learning from works councils: Five unspectacular cases from Germany. Industrial Relations, 33, 467–481. Williamson, O. (1975). Markets and hierarchies. New York: Free Press. Williamson, O. (1985). The economic institutions of capitalism. New York: Free Press. Zwick, T. (2004). Employee participation and productivity. Labour Economics, 11, 715–740.

THEME II: MANAGERIAL COMPENSATION AND CONTROL

147

This page intentionally left blank

148

PERFORMANCE-RELATED PAY IN DANISH COOPERATIVE CREAMERIES Morten Hviid ABSTRACT Following the formation of the first cooperative creamery in Denmark in 1882, the number of creameries adopting this particular organisational form grew rapidly over the next 20 years. These cooperatives had to delegate the running of the creamery to a professional manager leading to potential incentive problems. This paper documents the extensive use of performance related pay. Three findings stand out. Firstly, the creameries’ awareness of the incentive problems. Secondly, the incentive contracts used were often quite complex and sometimes non-linear. Thirdly, contract terms often varied over time, even within individual creameries.

1. INTRODUCTION The invention of the continuous cream separator in 1878 changed the way in which butter was produced in Denmark and made it one of the leading butter exporters of the time. The Danish farming industry at that time

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 149–176 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09005-8

149

150

MORTEN HVIID

consisted of a small number of big (typically manor) farms, and a very large number of small farms with an average herd size somewhere between six and fourteen cows (Henriksen, 1999, p. 61). Prior to 1878, because of their size the latter faced the particular problem that their individual output of butter was too low to fill a barrel of fresh butter for the export market. The price fetched by a barrel consisting of butter from a mixture of sources would at best reflect the average quality. Those farmers who produced butter of very high quality consequently lost out. Collecting milk and processing it centrally to get large homogeneous batches was not economically viable because of poor road surfaces, which resulted in the milk being shaken during transportation. This reduced the amount of cream, which could be extracted through conventional skimming. The continuous cream separator, using centrifugal forces, overcame this and in addition usually outperformed hand-skimming by skilled dairymaids. The downside of the new technology was that a creamery would need to process raw milk from around 400 cows to achieve minimum efficient scale. Except from the very large estates, which had their own creameries, commercial production of butter would require regular supplies from a large number of farmers, and because of the associated incentive problems related to the quality and security of supply, a new governance structure. The initial solution was a private creamery owned by the manager; but from 1885 the dominant organisational form became the cooperative creamery owned by (a subset of) the raw milk suppliers (farmers) and managed by a dairyman employed by the cooperative. The success of cooperatives was both relative and absolute. While the total number of creameries grew, the number of private creameries fell, see Table 1.

Table 1. Year Mid-1880s 1888 1894 1898 1901 1905 1909 Source: Bjørn (1982).

Development in the Number of Creameries. Co-operatives

Private

o 50 388 907 1,013 1,067 1,087 1,163

4500 468 215 260 209 207 255

Performance-Related Pay in Danish Cooperative Creameries

151

Unlike the private creamery, where the manager was the residual claimant, the cooperatives introduced a separation between ownership and control, with the associated familiar agency problems. To be successful, the cooperatives would have needed to find ways to overcome these, either through monitoring or direct incentive provision. Using a unique data set of contracts between cooperatives and managers extracted from the minutes of the Board and General Meetings of 214 cooperative creameries, the aim of this paper is to further our understanding of how they did so. We focus on the early period of these cooperatives, from their formation in the mid1880s until the First World War. While cooperatives have played a significant role in most areas of Danish agriculture, the creameries were the first producer cooperatives and hence the pioneers in dealing with the problem of motivating the manager. Little has been written about the managerial incentive problems in cooperatives.1 Existing theoretical literature on cooperatives has focused on other sources of strengths and weaknesses of this organisational form such as the asymmetry of members (Banerjee, Mookherjee, Munshi, & Ray, 2001; Hart & Moore, 1998), the commitment of members (Rey & Tirole, 2000) and the value of democratic voting rules (Albæk & Schultz, 1997). This is to my knowledge the first paper to focus on managerial incentive problems in cooperatives. Both the minutes and other contemporary sources show that the cooperative creameries were aware of the incentive problems and structured the contract with their manager to account for these. This paper describes the various methods and terms used and their evolution. There are clearly a number of alternative ways to motivate the manager. If the Board, or its Chair, has the time and expertise to monitor the manager, a series of shortterm contracts may be sufficient. Where reliable comparators exist, the manager’s fear of losing the job might provide adequate motivation. Alternatively, (s)he could be given hard incentives by linking pay to performance. As the relative merits of these different ways of providing incentives depend on the ability of the principal (the chairman of the Board), there is no reason why they should not co-exist, nor any reason to think that one would lead to a better outcome for the cooperative than another. Three findings dominate the results. One is the creameries’ awareness of the incentive problems and how the solution to these changes over time. The second is that the incentive contracts used are often quite complex and in some cases non-linear. The third is that contract terms often varied over time, even within individual creameries.

152

MORTEN HVIID

The remainder of the paper is organised as follows. Section 2 argues that the dominance of the cooperative governance structure indicates that they must have been able to solve the incentive problems, relating to the manager in a manner which was efficient, relative to private ownership. Section 3 discusses the incentive problems in the production of butter. Section 4 describes the main details of the managerial contracts. Section 5 concludes.

2. PRIVATE VS. COOPERATIVE CREAMERIES The success of the cooperatives may arise from their superiority at solving other problems associated with producing butter from the supply of a large number of farmers. It is often argued that cooperatives, with their more ready access to non-legal sanctions and personal monitoring of members, would be well suited to ensure that each farmer kept to their promise to supply the creamery with unadulterated raw milk. As argued in Henriksen and Hviid (2004), the evidence does not support this view of cooperatives for Danish creameries. Instead they tended to rely on contractual obligations and penalties together with technical rather than personal monitoring of the milk. It may be that the cooperative organisational form came to dominate for two reasons, which are unrelated to managerial performance. One argument is that the Danish farmers simply preferred to be in control and were, if necessary, willing to sacrifice profits to ensure this. Another argument relates to financial constraints. The farmers might have had easy access to credit because they could put up their farm as collateral, while experts in butter making wishing to establish a private creamery might have faced financial constraints. There seems to be broad agreement that the farmers aimed to maximise the return on the raw milk, and hence the main driving force behind the cooperative was to make money, see Henriksen (1999, p. 59, note 3). We find further support for this in our data. For example, the cooperative creameries made extensive use of competitive tendering both with members and outsiders, thus showing no favouritism or social preference. Similarly, the cooperatives showed no special solidarity with other creameries. When a large savings bank (Bondestandens Sparekasse) in which many cooperatives were financially involved, went bankrupt, leaving a large number of the cooperative creameries with a loss, those who had not used the failing bank showed no inclination to bail the others out. While cooperatives were

Performance-Related Pay in Danish Cooperative Creameries

153

prepared to process the milk of other creameries when their machinery broke down, they did so for a fee, which did not discriminate between cooperative and private creameries. There is considerable evidence that raising finance was not a general problem.2 Firstly, we find several cases where the manager of a cooperative lends the cooperative substantial sums of money, in some cases up to 20% of the cost of building a creamery, which suggests that there were skilled individuals who would have been able to put up a substantial part of the capital themselves. Secondly, in 33 of the creameries we investigated, the manager was required to provide an up-front bond as surety for his handling of the assets. Finally, although the dairymen who managed the cooperatives had little or no track record initially, they would have built this up over time. During the period in question, the successful managers of cooperatives should have been able to convince savings banks to lend them money either to buy out the suppliers,3 or start up on their own. Thus there is no compelling evidence that private individuals, skilled in running a creamery, were unable to raise the necessary capital to finance a restructuring to a supposedly more efficient organisational form. If one accepts the arguments that both organisational forms were feasible at least financially and in terms of managerial talent and that the farmers who supplied the raw milk were driven mainly by concerns about the return they would receive, then the cooperative organisational form must have been able to perform at least as well as the private.4 The success of the cooperatives demonstrates that they dealt successfully with the incentive problems relating to the motivation of the manager, and that their solutions did not put them at a significant disadvantage relative to the private creameries.

3. INCENTIVE PROBLEMS IN BUTTER MAKING The raw milk was produced by the farmers and transported to the creamery where the dairyman transformed this into butter and occasionally cheese using labour and other inputs. The creamery consisted of the manager and staff, for example an assistant dairyman, a dairymaid, a trainee and possibly a stoker. If the staff were all hired and paid by the cooperative, the manager merely became the supervisor of the work force rather than their direct principal. This can give rise to the potential problem of collusion between manager and staff, where the manager protects the staff from the consequence of poor performance, see Tirole (1986). For the creameries, there

154

MORTEN HVIID

was a very simple solution to this problem, which was adopted in the vast majority of cases. The manager paid the staff out of his own pocket; he also hired and fired the staff, which was seen as giving him more authority over them. Thus, the cooperative creameries solved one of the incentive problems by putting the manager in a similar relationship to the staff as an owner would have. A manager could influence the outcome of the creamery in three main ways. He could affect all the main costs other than raw milk and transport, namely energy, cleaning and packaging material, capital in the shape of buildings and machinery and labour, either through direct efforts or through monitoring the work force. Secondly, the manager could affect the quantity of butter obtained from the raw milk, typically measured by the ratio of butter produced to raw milk used, by monitoring the work force and ensuring that the machinery was well maintained. Finally, the manager could affect the quality of the butter and hence its price. Table 2 shows summary statistics for three of the performance measures from the 1904 official creamery statistics from 523 creameries. Some of the variations in performance across the creameries is determined by matters at least partially out of the control of the manager, such as the location of the creamery, the age of its capital equipment and the quality of the herds. Indeed, if the creameries are all efficient, the variation should be unrelated to shortcomings in managerial efforts. However, to get a feel for what a manager might achieve if properly incentivised, we have computed the effect on surplus of reducing costs by 1 standard deviation (2,586 Kroner5), reducing the milk per butter ratio by 1 standard deviation (3,947 Kroner) and increasing the butter price by one standard deviation (1,592 Kroner). While one has to keep in mind that these are at best very imprecise measures of what could have been achieved by incentivising a manager, when compared with the average managerial salary of 3,200 Kroner, it would suggest that doing so might have been worthwhile.

Table 2.

Performance Measures for 1903–04.

Average costs (excluding raw milk and transportation): øre/1,000 lb milk Milk/Butter: lb/lb Price of Butter: øre/lb

Average

S.D.

230.7

54.61

25.87 91.7

Source: Mejeri-Drifts-Statistikken for 1903–1904.

0.592 0.798

Min 112 23.9 88.0

Max 1005 28.0 94.6

Performance-Related Pay in Danish Cooperative Creameries

155

The challenge for the Boards of the creameries was to ensure that the manager balanced the three dimensions (quantity, quality and costs) to maximise the surplus for the cooperative.6 There were a number of different ways to achieve this. They could rely on the threat of firing combined with either day-to-day monitoring or annual comparisons with other creameries. They could use the fact that contracts were customarily annual and make renewal dependent on monitoring and comparisons. Finally, they could have provided direct incentives for the manager by relating at least some of his pay to one or more of the measures of performance. There are as we shall see a number of measures, which could be used. These involve different degrees of risk and in some cases cover only some of the tasks undertaken by the manager, for example by focusing on the quantity produced rather than quality or cost. In such cases, the manager may misallocate his time unless other measures of performance are included to counteract any distortions. Motivating the manager almost always involves shifting some risk to the manager and so requires some form of compensation where the manager is risk averse. The Board would need to keep an eye on the participation constraint, not just to ensure that a good manager would accept the wage offer, but also to ensure that they did not overpay. An added complication with respect to the participation constraint was that there is some evidence of increasing returns to managing a creamery in the sense that the required salary was increasing and concave in the size of the creamery as measured by the amount of raw milk produced. The reason for the concavity lies partially in a greater scope for specialisation with a larger creamery with more staff. Staffing costs which were met by the manager were increasing but at a decreasing rate. It is in addition also likely that the managers themselves obtained a positive utility from managing a bigger creamery so that a doubling in size would not require a doubling in salary. In a cooperative where the growth in its size within a year might be large, the simplest way to keep the manager at the level of his reservation utility would be to link part of his salary to the size of the creamery, but in a non-linear way. One example would be a fixed rate per pound of milk for the first two million pounds, and slightly less for any amount of milk thereafter.

4. MANAGERIAL INCENTIVES The main data come from minutes of general and board meetings of 214 cooperative creameries.7 Of these 197 include information on at least one managerial contract. In total, we have 3,422 observations of individual

156

MORTEN HVIID

Table 3. Period Whole period Before 1894 1894–1902 1903–1907 1908–1915

Number of Performance Measures.

None (%)

One (%)

Two (%)

Three (%)

Contracts

25.9 9.6 27.1 29.7 31.3

60.3 74.0 56.6 57.8 58.1

11.7 14.7 14.8 8.6 8.7

2.1 1.7 1.6 3.8 1.9

3,422 585 1,149 683 1,005

contracts. The level of detail varies considerably, from just providing the main components to transcripts of the actual contract, in one case a fourpage document. We have not extrapolated to cover missing years where there was any doubt that the contract had remained the same. Because of the importance to the creamery of the book of minutes as a record of decisions, the information they contain is likely to be accurate. What we cannot know is what has been left out. The data are summarised in Table 3. It demonstrates widespread use of performance related pay and that at least some of the contracts were complex in that they included several performance measures. From the total of 3,422 contracts of different vintages, about 75% used one or more performance related elements in the managerial wage contract. Across time, there is a great variety both in terms of whether performance related pay is implemented and how many measures are used. After 1894, about 30% of contracts include no explicit performance related pay element, which could be explained by an improved ability to monitor and hence to use other means such as firing to provide incentives, as well as better information about what a reasonable level of salary was.8 After 1907, we also see a drop in the number of incentive measures used within the falling number of contracts which contained any incentives. This may also be due to increased understanding of the production process. Behind the numbers in Table 3 is a more complicated story, which is explained below by focussing in greater detail on the performance measures used in incentive contracts.

4.1. Firing as Incentive We first look at the potential power and use of non-performance-related measures such as firing or non-renewal of short-term contracts. Such measures require more extensive monitoring.

Performance-Related Pay in Danish Cooperative Creameries

157

4.1.1. Monitoring The highest authority of the cooperative was the General Meeting (Generalforsamling) to which all the members of the cooperative belonged and which took all long-term decisions such as investment. The General Meeting elected the Board, usually with seven or nine members, which elected a Chairman from its midst. One of the most important roles of the Board, and particularly of the Chairman, was to monitor the performance of the manager. How heavily the Board or its chairman got involved in the day-to-day running of the creamery varied considerably from cases where the manager had almost total freedom to run the creamery to instances where the manager was little more than an administrator. Although the Board members and the Chairman would be in the creamery regularly, they were not usually experts in butter making, making direct monitoring difficult. Thus monitoring was never likely to be perfect and the abilities and costs of monitoring may have varied considerably across cooperatives. As members would each only receive some of the additional surplus generated by their monitoring, free riding might be a concern. This was to some degree mitigated by the inclusion on Boards in the early period of a disproportionate number of farmers with big herds, Bjørn (1982). As surplus was shared according to the amount of milk delivered, they would gain more than average from an improved performance. More important to ensure that the Board members carried out its duties might have been the effect of their status in the local community and the associated likelihood of being elected to join the Board of other cooperative ventures such as bacon factories and savings banks.

4.1.2. Efficiency Wages The cooperatives were, particularly early on, accused by observers such as Professor Bøggild, of trying to lower managers’ wages, with some cooperatives going as far as putting the job of manager up for tender.9 The cooperatives generally appeared keen to keep the manager close to or at his reservation utility where he is just indifferent between accepting the job or leaving. This is not compatible with an efficiency wage argument where the manager is paid above the going rate in order to make firing or non-renewal more costly. As noted above the minimum salary necessary to hire a manager was likely to be (increasing and) concave in the size of the creamery. Creameries, which were unsure about what would happen to their size during the contract period of the manager, might want to take this into consideration by making the salary an increasing, concave function of size, such

158

MORTEN HVIID

as the one described at the end of Section 2. In our data, we find that 313 of 3,422 of the contracts in our data set (9%) do so. The remuneration of the manager included free use of a house and garden and one might argue that the fear of losing these if he lost his job might have a similar effect to an efficiency wage. However, where a manager had to leave his house and garden before collecting a harvest from his garden, he was typically compensated and any investment made by the manager, which had a life beyond the tenancy would typically be repaid. As the manager normally had three months notice to seek other employment and accommodation, the potential penalty imposed by sacking him from the loss of a house and garden was probably modest.

4.2. Short-Term Contracts Short-term contracts were customary and managers were typically hired for 1 year at a time with their contract renegotiated annually. The fear of not having ones contract renewed obviously creates incentives, at least to perform well in the short-run, but requires either direct monitoring or comparison with the performance of other creameries. As just noted, monitoring was costly and imperfect. Cross dairy comparisons were also problematic. Though creameries published annual results in various weekly publications such as Mælkeritidende (and after 1897 in a national collection of statistical information, Mejeri-Drifts-Statistikken) and relevant data were available, comparisons across creameries were difficult because of significant heterogeneities between even neighbouring creameries.10 The threat of non-renewal of the employment contract was credible, as advertisements for jobs attracted a large number of qualified applicants, changing the manager was not costless to the creamery.11 In 1895, Johansen (1896) found 121 changes of manager over 10 years in 96 dairies, more than half because of dissatisfaction with the manager or disagreements over pay. The dairies in the data set described in this paper had a much lower, but still significant level of managerial turnover. One-year contracts may create long-term problems if the manager becomes too focused on short-term performance. He often appears to be the driving force behind innovation and investment in the rapidly improved technology of cream separation as well as equipment maintenance. Shortrun incentives might lead to both poor maintenance and under investment by the cooperatives, resulting in poor long-term results. The manager was also either the sole or chief monitor of the quality of the supply of milk. The

Performance-Related Pay in Danish Cooperative Creameries

159

fear of making enemies, something that could ultimately cost him his job, might make the manager less vigilant when carrying out this task.

4.3. Common Performance-Related Incentive Systems As Table 3 shows, about 75% of the managers wage contracts include one or more performance related elements, usually supplementing a component, which is either fixed or related to exogenous variables such as the size of the creamery. However, there is a great variety over time both in terms of the extent and complexity of performance related pay. To explore this further, the performance related measures are divided into four broad categories. 1. Relative performance evaluation. Two of the most common forms of relative performance evaluation are: RPE1 Payment relative to a target price, typically a fraction of: ðPB Pn Þ B: RPE2 Payment relative to a target level of productivity: ðB M=mn Þ PB : where PB is the realised price of butter, Pn a target price of butter, M the amount of raw milk processed, B the amount of butter produced, m M=B the milk to butter ratio and mn the standard industry norm for m; 2. Piece Rate – PR. Payment is fixed per pound of butter produced. Occasionally it is restricted to a particular quality such as first class butter or butter which met the expectation of the wholesaler. 3. Profit or Revenue Sharing. The three most common measures are: PS1 A percentage of the net surplus of the creamery. PS2 The property rights to the returns to a particular amount of milk. Typically involves paying the average price for the raw milk and a share in the surplus according to that share of the milk. RS1 A percentage of the revenue from the sale of butter. 4. Objective bonus and penalty scheme. The bonus schemes were typically fixed amounts either for achieving a pre-set target; or for winning a prize at an exhibition; or for scoring a particular level at some public quality assessment. Penalties were imposed for missing targets.

160

MORTEN HVIID

The development over time of these measures and the extent to which the contract imposed a share of the costs on the manager are shown in Table 4. Table 4 demonstrates that all the measures were popular at some point in time, but that none had enduring popularity. Simultaneous use of relative performance evaluation, PR and profit sharing in the same industry is itself unusual and interesting. Since these data refer to the early phase of a new industry, they reveal extensive experimentation with the ideal performance measures in incentive contracts. A better understanding of the data can be obtained from a closer examination of the dynamics of the four most popular measures, RPE1, PR, PS1 and PS2, which changes most over time. This is followed by more investigation of which incentive measures were used singly and in combination with others. 4.3.1. Relative Performance Evaluation Only 12% of all the 3,422 contracts include RPE1, and their use was not distributed evenly over time (Table 4). About two-thirds of the contracts relating to the period up to 1894 include RPE1, when RPE1 was the incentive measure of choice; after 1894, it went completely out of fashion. The detail of its construction helps to explain this pattern. The target or reference price was based on the weekly price quotation provided by the Copenhagen Butter Merchants, known as ‘‘The Copenhagen Quotation’’ or the ‘‘The Danish Butter Quotation’’. Before 1894, the quoted price for first class butter, commonly referred to as the ‘‘top quotation’’, was rather oddly below the highest achieved prices of first class butter. The typical RPE1 term offered the manager 25% of the difference between what his butter fetched and the top quotation. Proposals to reform the Copenhagen Quotation were discussed during the early 1890s; and in November 1894; the quotation was reformed to ensure that the top quoTable 4. Period

Use of Performance Measures.

RPE1 (%)

RPE2 (%)

PR (%)

PS1 (%)

PS2 (%)

RS1 (%)

Bonus (%)

Penalty (%)

Costs (%)

Pre-1894 1894–1902 1903–1907 1908–1915

64.6 4.6 1.2 0.0

3.1 3.9 1.3 0.0

7.7 42.5 36.6 21.2

29.4 19.8 9.8 6.4

0.0 5.0 22.7 33.8

0.7 5.2 4.2 3.2

1.4 8.8 10.0 15.6

6.0 11.7 8.9 6.9

14.9 42.8 49.6 42.7

Total

12.8

2.1

29.1

15.5

16.2

3.7

9.8

8.7

39.4

Note: Any given contract can contain more than one of these measures at once and hence the rows sum to more than 100%.

161

Performance-Related Pay in Danish Cooperative Creameries

tation really was the highest price anyone could hope to achieve.12 This revision apparently shattered confidence both in the particular reference price and in the measure in general. This is illustrated by Table 5, which also shows numbers of contracts, which included these incentive measures immediately before and after the reform. The fraction of contracts, which include RPE1 based on the Copenhagen Quotation as a measure stay roughly at two-thirds from 1887 to 1892. This is followed by a small drop in 1893, a large drop in 1894, and then virtual abandonment in 1895,13 from which it never recovers. The creameries seem unwilling to find another reference price or to work with a target below the reformed Copenhagen Quotation. For 85 of the 104 creameries who had used RPE1 at some time or other, we have information about the contract terms, which immediately replaced RPE1. Table 6 shows the dates when creameries abandoned RPE1, the vast majority 67%, in 1893 and 1894. The link of a large part of the manager’s salary to a target price, which might be revised adversely clearly increased the risk for the managers, which may partly explain why some creameries reformed their payment system before the Copenhagen Quotation was reformed. Many of the creameries that did retain RPE1 until the reform date included a term guaranteeing the manager a minimum salary; others included contingent terms, specifying what would happen in case a reform took place. RPE1 had the advantage of rewarding both quantity and (relative) quality, reflected in the price, which the butter fetched, but it had two potential downsides. Firstly it focused only on the revenue side, a factor often balanced by making the manager liable for a fraction of a pre-specified part of the costs (see also Table 8). Secondly, the gap between quotation and reality was increasing up to 1894, providing managers with an annual windfall. The

Table 5.

Contacts with RPE1.

Year

RPE1

Total

Fraction (%)

1887 1888 1889 1890 1891 1892 1893 1894 1895

26 45 55 59 60 66 59 23 2

38 70 78 85 90 100 108 120 122

68.4 64.3 70.5 69.4 66.7 60.0 55.6 19.2 1.6

162

MORTEN HVIID

Table 6.

Year of Abandonment of RPE1. Observations

Before 1892 1892 1893 1894 After 1894

12 9 38 19 7

Total

85

cooperatives were keen to claw back this windfall, leading to disagreement between the manager and the Board in many creameries, and in some cases culminating in a change of manager. Improved information about what butter could fetch from the wholesalers may account for the drift away from RPE1 before 1894. The reform may have accelerated the inevitable demise of a performance measure, which had lost its relevance. 4.3.2. The Impact of the Reform The reform of the butter quotation triggered a lively debate about pay schemes. In a meeting of the Jutland Farmers Association in September 1894, the question of manager’s pay was debated extensively. One suggestion14 proposed that the manager was paid: (a) a fixed salary of 1,200 Kroner, (b) a payment contingent on the amount of butter, (c) a percentage of the price of butter, (d) a percentage of the net surplus, and (e) 5% of revenue from sale of cheese. The manager would also receive payments in the form of house, garden, firewood, lighting and skimmed milk, cheese, butter and cream for daily household consumption (including staff). Table 4, indicates that many creameries implemented at least part of this proposal by increasing use of measures such as PR (b) and RS1 (d). The debate about manager pay culminated in the offer of a prize (25 Kroner) for the best new pay scheme in the industry magazine, Mælkeritidende. Of the many suggestions, which were received, two received prizes. The first prize was awarded to the following scheme (where it is method rather than actual payments which are important): 1. Fixed salary 600 Kroner. 2. 2.5% of revenue from butter up to a revenue of 100,000 Kroner and 1.5% of revenue above 100,000 Kroner (the manager pays for salt, colour and ‘‘renkultur’’).

Performance-Related Pay in Danish Cooperative Creameries

163

3. 6% of revenue from Cheese (the manager pays for salt, cumin, colour and rennet). 4. At the end of each year, the average consumption of coal, oil, soda and other polishing and cleaning goods per 100-pound butter for the dairy association is computed. Managers of creameries with consumption above this average pay 50% of the excess, whereas those who use less than the average are paid 50% of the saving. 5. Each week, the average price for butter for all dairies in the Association is computed. In creameries where the price was above average, the manager is paid 10% of the amount above the average. In creameries where the price was below average, the manager pays 10% of the difference to the creamery. 6. If the milk is paid according to fat content, the manager is given an extra amount a year, depending on the size of the creamery, 100–200 Kroner. Clearly this is an intricate system, but the author offers arguments for the importance of each point. From his discussion, it appears that he was well aware of the various incentive problems, including those raised by multitasking, which arise in the cooperative dairy sector. He pointed out that some of the components favour the larger dairies, but that this is counteracted by other components, which penalise size. By contrast, the suggestion that won second prize is extremely simple. 1. The manager is paid with a certain number pounds of milk from the milk delivered to the creamery. The beauty of this proposal is that it aligns the manager’s incentives with those of an average member. To assess whether these proposals had any impact, the schemes that replaced RPE1 are shown in Table 7. More than 75% include some incentive element the year after RPE1 was abandoned. This confirm the minuted debates, which indicate no move to abandon all incentive pay from either the employer or employee side. The shift is mostly to PR, which is not surprising given the similarity to RPE1. Before the reform, many creameries achieved a price of 4 øre ‘‘above top’’, and the manager received 25% of the difference, i.e. 1 øre per pound of butter. 4.3.3. Piece rate (PR) 996 out of 3,422 contracts in our data set include PR. Fig. 1 shows the development of the performance measure. It demonstrates how PR became the new alternative to RPE1, going from less than 10% of the contracts

164

Table 7.

MORTEN HVIID

Performance Measure in First Contract after Abandonment of RPE1. Observations

Percentage

No measure

20

23.3

One measure Of which: RPE2 PR PS1 PS2 RS

58 0 39 14 1 4

67.4 0.0 45.3 16.3 1.2 4.7

8 3 3 1 1

9.3 3.5 3.5 1.2 1.2

86

100.0

Two measures Of which: RPE2 & PR PR and PS1 PR and RS PS1 and PR Total

pre-1894 to almost 35% in 1894, and 45% in 1895. It maintains this level of popularity until about 1902, when a slow decline sets in, flattening out at around 20% by 1910. This durability is best explained by its simplicity as a measure. However, while simple to calculate, PR is almost totally focused on quantity. The Board could have rectified this either by monitoring quality and costs directly, or by adding other measures to reward these aspects. We explore such combinations in more detail below. 4.3.4. Profit Sharing The most natural performance measure might be profits, the prime objective of the cooperative, and this featured in both the prize-winning suggestions in the 1894 competition. The development of PS1 and PS2 is charted in Fig. 2. The profit sharing measure, PS1, though very popular in the period 1890– 1894, falls dramatically after 1894, apparently alongside the relative performance measure RPE1, following the reform. However, schemes based on property rights to milk did not grow significantly in popularity until about 1902. This coincides with the period when the use of PR starts to decline. Comparison with Fig. 1 shows that PS2 overtook PR around 1907–1908. Why would PS2 replace PS1? A manager with property rights to 2% of the milk of the creamery would receive the average price for the milk, determined exogenously, and a 2% share of the surplus of the creamery. A PS1 offering 2% of the surplus would typically be combined with an

45.00% 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00%

1915

1914

1913

1912

1911

1910

1909

1908

1907

1906

1905

1904

1903

1902

1901

1900

1899

1898

1897

1896

1895

1894

before 1894

0.00%

Performance-Related Pay in Danish Cooperative Creameries

50.00%

Year

Fig. 1.

The Development of Piece Rate over Time.

165

40.00%

35.00%

30.00%

25.00%

20.00%

15.00%

10.00%

5.00%

0.00% 1887 1888 1889 1890 1891 1892 1893 1894

Fig. 2.

1895 1896 1897

1899 1900

Year

Profit-Sharing Measures.

1898

1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914

166

PS1

PS2

1915

MORTEN HVIID

Performance-Related Pay in Danish Cooperative Creameries

167

annual wage. PS1 thus appears to offer the same as PS2, but with lower risk to the manager. As contemporaries note, using percentages of net surplus is problematic because of the way the surplus arose. Suppliers (members) were paid for their deliveries of raw milk according to an agreed, but essentially arbitrary price, which represented a cost to the cooperative. Suppliers were committed to take back a certain fraction of the skimmed and butter milk for which they were charged an agreed, but again largely arbitrary price. This provided an income to the cooperative alongside that obtained for the butter and cheese. The surplus could be increased by members paying themselves a lower price for raw milk and/or charging a higher price for the returned skimmed milk. Since the surplus was distributed to each supplier (member) according to their share of total raw milk supplied, the suppliers (members) would get the same total return on their milk (although they would have to wait longer to get their share of the surplus).15 Apart from any issues of time preference, the member should be unaffected by this. While a representative member would therefore be indifferent to the size of the surplus, the manager would not. If the method was PS1, the lower the surplus the lower the contribution to his total wage, the cooperative would therefore have to state ex ante the relevant transfer prices for raw milk and skimmed milk, and be committed to them for the duration of the managerial contract. In contrast, PS2 treats the manager like a representative member, paying him for the notional amount of milk charging him for the returned notional skimmed milk and then giving him a share of the surplus. PS2 is thus considerably simpler, and for the cooperative has the additional virtue of making the manager ‘‘one of them’’. This also helps to explain why PS1 fell out of favour in 1894 because the transfer price for raw milk was typically determined relative to the Copenhagen Quotation for butter, which we have seen had itself lost credibility. 4.3.5. Summary of the Evolution of Performance Related Pay From the tables and figures three phases can be identified. Up to 1894, the managerial contracts are dominated by two measures, RPE1 and PS1. Following the reform in November 1894, both fall out of favour and their place is taken by a PR measure. Then, from a slow start in 1894 when first publicly proposed, another profit sharing measure PS2 gradually takes over and dominates from 1907/1908 onwards where the third phase begins. While the revision to the Copenhagen Quotation in 1894 can be identified as the cause of the first ‘‘shock’’ to the payment systems, the cause of the second transition date is harder to pin down. The only ‘‘exogenous’’ event is an attempt to negotiate a national scale for managerial salaries which

168

MORTEN HVIID

started around 1906 and which culminated in 1908 with an agreed pay scale. Although only specifying the level of the total salary, the method, PS2, was clearly in the background. It did not step into the foreground until the revision to the national pay scale in 1918 from when PS2 was explicitly adopted. The pay scales between creameries and managers were first negotiated locally in the County based Dairy Associations. From Fig. 2, we see that the adoption of PS2 occur much more gradually and this fits well with the payment form being talked about at first local and then national levels. Through this, an increasing number of Chairmen and Boards would become aware of the potential benefits of PS2.

4.4. Variety of Measures within a Contract We find that 13.8% of contracts specify more than one incentive measure (Table 3). An article in Mælkeritidende in 1890, discussing theoretical and practical issues of running a creamery, drew up a suggested model contract, of which paragraph seven is relevant for the pay of the manager. In translation it states: In return for his services, he receives: in fixed annual salary of Kronery, 25% of what is achieved above the top quotation for the butter, 5% of the sales value of the cheese, 5%, 10% or 15% (depending on the size of the fixed annual salary and how the creamery pay for milk) of the net surplus when these are based on paying the milk according to the top quotation (for butter) per 28 pound milk and charging 3 or 4 Øre per kilo of the returned (skimmed) milk. In addition he gets in all cases a free abode (with garden) for himself, his family and the staff, firewood, lighting, one pægl (a measure) of cream daily, 3 pounds of butter a week and the necessary cheese and skimmed milk for the housekeeping Mælkeritidende (1890, pp. 377–378).

The winning entry from the 1894 competition also included several measures. From contemporary sources, especially the debates in Mælkeritidende it is clear that the industry understood the incentive problems, which would arise when the manager had multiple tasks and at least in theory tried to construct the wage contracts to ameliorate these. The cooperative’s aim for the manager is clearly to generate as much income for the members as possible by producing the profit maximising quantity–quality combination. While some measures, such as PS1 and PS2, relate directly to this aim, others relate only to parts of it. For example, PR focuses only on quantity, RPE1 focuses both quantity and quality, whereas RPE2 focuses solely on efficiency in production. Measures that focus only

169

Performance-Related Pay in Danish Cooperative Creameries

on parts of the overall aim should ideally be combined appropriately with other measures. Table 8 enables us to see which measures are typically used on their own and which are used in combination with others. As expected, contracts that include PS2 (87.5%) typically contain no other performance measures,16 nor any requirement for the manager to share in paying for the costs. Surprisingly, PS1, which might be expected to look like PS2, is used in combination with other measures in a third of the cases. Except for RPE2, which is almost never used alone, the remaining performance measures are used on their own between two-thirds and threequarters of the time. Making the manager share in costs is as expected more common for measures such as RPE2, PR and RS1 which give no direct incentives to cost minimisation. The same should have been true for RPE1, but is not, possibly because the contracts all relate to the early period where the data are weaker on detail. Seventy three contracts from 13 creameries use three measures. Of these, 40 contracts from four creameries contain: (PR, PS1 and Bonus). As PR focuses solely on the quantity of butter and bonus solely on the quality, adding a profit sharing measure includes a concern for costs and only one of the four creameries includes a direct requirement of costs sharing. In contracts with two measures, PR is often combined with another measure, typically either PS1, RPE2 or Bonus. The expectation that in particular PR could not stand alone is at least to some extent confirmed by their use in combination with other measures. From our sample, not only are many of the contracts complex in that they include several performance measures and/or non-linearities, they also vary Table 8.

Fraction of Contracts with a Single Performance Measure or Share in Costs.

Measure

RPE1–overtop RPE2–M/B PR–piece rate PS1–% surplus PS2–milk RS1–% revenue Objective bonus Other

Contracts with One or more Measures

Fraction Containing a Single Measure (%)

Fraction Sharing in Costs (%)

439 72 996 531 553 125 334 35

73.1 5.6 68.6 66.5 87.5 72.8 30.2 71.4

23.5 79.2 61.4 26.0 37.6 67.2 52.1 57.1

170

MORTEN HVIID

over time within each creamery. While some of these non-linearities relate to one or more of the performance measures, others relate to measures exogenous to the manager such as size. Unlike Lafontaine and Shaw (1999), who for franchising find that contract terms within a franchise rarely change, many of the creameries in our sample used the annual renegotiation of the contract to vary terms. Moreover, contract terms are not just altered when a new manager is employed, existing managers may either be offered a contract with different measures, or a contract with the same measures, but different magnitudes. Although we cannot test for this, it is as if the creameries are trying to keep the manager to his participation constraint every year. Our findings of a high degree of variety in contract terms is also at odds with the literature on retail contracting surveyed by Lafontaine and Slade (1997), who highlight the large degree of within-firm contract uniformity found in the existing empirical literature, see also Lyons (1996).

4.5. Risk The use of incentive contract transfers risk from the farmers to the manager. The creameries did occasionally take measures to ensure that the manager was partially insured by specifying a minimum total wage. Equally they occasionally specified a maximum wage to prevent the manager from extracting too much rent. The incidence of this is illustrated in Fig. 3, where we see the two measures following the same trend. 653 contracts specifying a minimum and 598 specifying a maximum, with 461 specify both, implying that there are a non-trivial number of contracts which specify either just a minimum or a maximum. From Fig. 3, the inclusion of a minimum salary starts growing about 1892, the time when there are rumours about a reform of the Quotation, which if implemented could reduce wages considerably. It peaks in 1896/1897, a period where the creameries are finding new ways to incentivise the manager and then steadily falls to about 15% of contracts in 1915. Looking more closely at the data, insurance is more likely with the performance measure PS1 (33.3%) and PR (26.7%) and less likely with RPE1 (13.9%) and PS2 (16.8%). This is not surprising, since following the revision to the quotation, the price of butter was uncertain. This has an indirect effect on a manager with a share in the surplus, because the internal transfer price for milk delivered to the creamery typically followed the butter Quotation so that if it was too high, the farmers would be paid too much for their milk and hence the net surplus would be correspondingly smaller.

25.00%

20.00% Mi n 15.00% Ma x 10.00%

5.00%

Performance-Related Pay in Danish Cooperative Creameries

30.00%

1915

1913

1911

1909

1907

1905

1903

1901

1899

1897

1895

1893

1891

1889

1887

1885

0.00%

Year Fig. 3.

Contracts Specifying a Minimum and Maximum Salary. 171

172

MORTEN HVIID

5. CONCLUSION The Danish cooperatives made extensive use of performance related pay in their contracts with their managers, something that must have contributed significantly to their success. The period we have focused on was one in which the cooperatives were learning how to run a business and where they would have had very little to compare with or learn from. The separation of ownership and control was in that period likely to affect mainly cooperatives. We know that performance-related pay was used in the bacon factories (Just, 1989) but these were formed after the majority of the creameries and in grocery retail (Just, 1984, pp. 120–134), which was at least contemporary, but also quite different in nature. Studying the way in which the cooperatives solved their incentive problems thus gives us a chance to see learning in action. From the minutes of Board and General Meetings, as well as the contemporary debate, it is clear that the participants in the cooperatives had a good understanding of incentive problems and could and would use complex contractual solutions where necessary. They fit the standard principal agent models surprisingly well. We see them solving the principal–supervisor–agent problem by delegating the payment of the agents to the supervisor. We see them use contingent contracts and risk sharing through minimum pay when the environment gets more risky. We see them worry about multi-tasking distorting the incentives of the manager when using performance measures related to sub-goals such as quantity rather than profits. In the early phase, the members of the cooperatives would have been relatively uncertain about what could be achieved and what constituted a good or even an adequate performance by the manager. Likewise, managers would have been unsure of their own abilities. Using performance related pay might then place a lot of risk on the manager, who might be unwilling to take up the post. Almost all managers were paid a monthly salary in addition to any incentive element. Basing pay partly on an exogenous standard (RPE1) would remove any fear that the cooperative would behave opportunistically, while leaving the members reassured that their manager would try to be among the best. The problem with this type of incentive is that it does not focus directly on what the members care about, profits. In the later phase, when both members and managers had a much clearer idea about what could be achieved in terms of quantity, quality and costs, using a profit related measure is clearly more efficient and powerful. It is therefore

Performance-Related Pay in Danish Cooperative Creameries

173

interesting to see that the property right to milk becomes increasingly dominant as a measure. Moreover from 1918 until today it has become the only incentive measure included in the standard contracts. When looking at the incentive schemes used during the period we have studied, one is struck by a number of features. The schemes are often complex containing several performance measures, they may be non-linear17 and even state contingent. Moreover, the terms of the incentive schemes varied considerably not just across cooperatives, but also over time within cooperatives. Any convergence to what one might term a national scheme occurs very late and as a result of central negotiations between associations of cooperatives and dairy men. This is somewhat surprising since there is throughout the period a lively debate about payment methods with several prominent proposals of best practice. Thus, the variety is not caused by ignorance alone. A future aim is to link incentive schemes to measures of performance to test whether these actually worked as intended. While the findings that the cooperatives used relatively complex and nonlinear incentive schemes may go against some of the current empirical literature, it is worth pointing out that butter making is a relatively simple business typically producing a single product. It may therefore be that using the complex measures is only really feasible where the production process is simple and hence one has to be careful about drawing too strong conclusions about the use of performance-related pay in modern manufacturing.

NOTES 1. For a survey of incentive provision in firms, see Prendergast (1999) and for specific empirical analyses of whether incentives work, see e.g. Lazear (2000) and Paarsch and Shearer (2000). 2. Both Bjørn (1982) and Buch (1960) suggest that the cooperative creameries were never constrained in terms of raising capital. 3. A small number of cooperatives did convert to private ownership or leased the creamery to the manager after their initial period, but the vast majority did not. 4. Henriksen and Hviid (2004) identify one potentially significant difference in what a private and a cooperative creamery can do using contracts. Cooperatives relied on long-term contracts, which were enforceable in Denmark. The value in these long-term contracts is their powers of commitment and the corresponding cost of the loss of flexibility. Unlike the private creameries, cooperatives could renegotiate the supply contracts collectively, with the decision of the majority binding upon the minority. Thus due to their structure, the contracts used by the cooperatives had a degree of flexibility not available to the private creameries. 5. 1 krone ¼ 100 øre.

174

MORTEN HVIID

6. Thus they have to solve a multi-task principal-agent problem as described in Holmstrom and Milgrom (1991). 7. While many of these early books of minutes have been lost, a considerable number has survived and the selection was largely governed by availability. The surviving books of minutes are stored either in The Danish National Business Archives or in local history archives. The majority can be tracked using a database provided by Danpa (Danmarks Nationale Privatarkivdatabase). Having tracked down the location of the archival material, the information used in this paper was extracted through a critical reading of these books. 8. There may also have been a growing awareness of multi-task problems in the incentive provision, which would militate against the use of performance-related elements. 9. There is a particularly heated debate in Mælkeritidende, the leading magazine for the dairy trade, during the early 1890s concerning this practise. 10. The distribution and quality of the cows as milkers as well as the distance the milk had to travel and the state of the local roads are among the most important reason for heterogeneity, making even local comparisons potentially misleading. As we shall see, this does not imply that the creameries never used relative performance measures. 11. Hiring a new manager typically involved the majority of the members meeting the short-listed candidates. 12. An example of its effect on managerial pay may help. Assume that the wholesaler paid the cooperative 94 øre per pound in both September and November 1894, that the quotation was 90 øre in September and the revised quotation 98 in November. A manager who got 25% of the price above the quotation would get 1 øre per pound in September, and if unlucky enough that the contract was symmetric, had to pay a penalty of 1 øre per pound in November, otherwise nothing. 13. Focusing on the number of contracts rather than creameries may over – or understates the popularity of this measure. A total of 104 creameries at some point include RPE1 in their contract. This is 69.5% of the 148 creameries in our sample who were founded by 1894 and where we have minute books from 1894 and earlier. This confirms the result that the use of RPE1 was widespread before the reform of the butter quotation. 14. Based on discussions between the South Jutland Creamery Association and the Vejle County Dairymen Association. 15. To illustrate this, consider a reduction of the price paid for the milk of 1 unit. If member i supplies mi pounds of milk, his loss would be mi. If the total amount of raw milk delivered is M, the increase in the surplus would be M and member i’s share of this increase in surplus would be (mi/m)M ¼ mi. The calculation for skimmed milk is roughly the same except that skimmed milk is a (fixed) proportion of the raw milk. 16. Where a second measure is used, it is almost always a bonus for getting a prize at a butter exhibition. 17. While the observed non-linearities in one or more of the performance measures are not unusual, the occasional non-linearity, which the manager cannot influence, is. This can only really be explained by a desire to keep the manager to his reservation utility in a way, which avoids too frequent renegotiations.

Performance-Related Pay in Danish Cooperative Creameries

175

ACKNOWLEDGMENTS An earlier version of this paper was presented at the EARIE conference in Dublin 2001 and at seminars in East Anglia, Kent and Newcastle. I am very grateful to participants, to Bruce Lyons, Michael Raith, Bob Sugden, Catherine Waddams and particularly Ingrid Henriksen, the editor Panu Kalmi and an anonymous referee for comments and suggestions and to Martin Bork and Susanne Krogh Jensen for very conscientious research assistance. This paper is part of a project with Ingrid Henriksen, funded by the Danish Social Research Council and their financial support is gratefully acknowledged. Part of this work was carried out while visiting the Haas School of Business, University of California Berkeley and their hospitality is gratefully acknowledged.

REFERENCES Albæk, S., & Schultz, C. (1997). One cow – one vote. Scandinavian Journal of Economics, 99, 597–615. Banerjee, A., Mookherjee, D., Munshi, K., & Ray, D. (2001). Inequality, control rights, and rent seeking: Sugar cooperatives in Maharashtra. Journal of Political Economy, 109, 138–190. Bjørn, C. (1982). Dansk Mejeribrug 1882–1914 (Danish dairies 1882–1914). In: C. Bjørn (Ed.), Dansk Mejeribrug 1882–2000. København: De Danske Mejeriers Fællesorganisation. Buch, L. (1960). Investeringer inden for Mejerierne og Slagterierne (Investment within dairies and bacon factories). Nationaløkonomisk Tidsskrift, 98, 131–152. Hart, O., & Moore, J. H. (1998). Cooperatives vs. outside ownership. NBER Working Paper no. 6421. Henriksen, I. (1999). Avoiding lock-in: Co-operative creameries in Denmark, 1882–1903. European Review of Economic History, 3, 57–78. Henriksen, I., & Hviid, M. (2004). Contracts in the governance of the early Danish dairy sector. Mimeo: University of East Anglia. Holmstrom, B., & Milgrom, P. (1991). Multi-task principal agent problems: Incentive contracts, asset ownership, and job design. Journal of Law, Economics and Organization, 7(Special Issue), 24–52. Johansen, J. (1896). Oversigt over forholdene paa jydske Mejerier, særligt vedrørende Mejeristernes Lønninger m.v. Ugeblad for de samvirkende jydske Mejeriforeninger [weekly magazine for the dairy associations of Jutland], vol. 2. Just, F. (1984). Brugsforeningsbevægelsen 1866–1920 (the cooperative movement 1866–1920). Esbjerg: Sydjysk Universitetsforlag. Just, F. (1989). Lederrekruttering til de første andelsslagterier (Recruitment of managers to the first cooperative bacon factories), Bol og By 2, 154–173.

176

MORTEN HVIID

Lafontaine, F., & Shaw, K. L. (1999). The dynamics of franchise contracting: Evidence from panel data. Journal of Political Economy, 107, 1041–1080. Lafontaine, F., & Slade, M. (1997). Retail contracting: Theory and practice. Journal of Industrial Economics, 45, 1–25. Lazear, E. P. (2000). Performance pay and productivity. American Economic Review, 90, 1346–1361. Lyons, B. R. (1996). Empirical relevance of efficient contract theory: Inter-firm contracts. Oxford Review of Economic Policy, 12(4), 27–52. Paarsch, H. J., & Shearer, B. (2000). Piece rates, fixed wages, and incentive effects: Statistical evidence from payroll records. International Economic Review, 41, 59–92. Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic Literature, 37, 7–63. Rey, P., & Tirole, J. (2000). Loyalty and investments in cooperatives. Mimeo: IDEI. Tirole, J. (1986). Hierarchies and bureaucracies: On the role of collusion in organizations. Journal of Law, Economics and Organization, 2, 181–214.

CORPORATE GOVERNANCE AND EXECUTIVE COMPENSATION IN BULGARIA AFTER MASS PRIVATIZATION: EVIDENCE FROM NEW PANEL DATA Derek C. Jones and Mark Klinedinst ABSTRACT By using new panel data for a sample of Bulgarian firms that comprises both state-owned and privatized firms (including new private firms), evidence is presented on the potential impact of ownership and age of the firm on diverse issues concerning corporate governance and executive compensation during 1997–2001. Privatization status and whether firms are de novo or not is found to be associated with differences in many areas including: the size and composition of company boards; the size of CEO pay; internal wage differences; the incidence of performance-based compensation (PBC); firm objectives; and patterns of decision-making influence. To investigate the determinants of executive compensation we first estimate standard CEO specifications. These baseline regressions reveal that CEO pay is: (i) positively related to size (ii) positively related to Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 177–209 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09006-X

177

178

DEREK C. JONES AND MARK KLINEDINST

performance; (iii) significantly affected by ownership; and (iv) influenced by whether a firm is de novo or not. These findings and the fact that both size and performance elasticities are much larger than those estimated before the start of mass privatization provide more general support than previously for the view that privatization has imposed strong discipline on the level of CEO compensation. In a series of additional regressions we proceed beyond standard specifications and examine the impact on CEO pay on other aspects of corporate governance. We find CEO pay is associated with: decision-making influence; whether the contract provides for PBC; whether the firm belongs to an employer’s federation; the extent of employee and managerial ownership. However some dimensions of corporate governance are not systematically associated with CEO pay. Chief amongst these is board structure. Many of these findings provide support for the view that managerial influence (rather than agency relationships) plays a key role in corporate governance in Bulgarian firms.

1. INTRODUCTION The broad topic of ‘‘corporate governance’’ has attracted great attention in the transition economics literature. Particular attention has been paid to selected issues notably econometric assessments of the impact of new forms of ownership on firm behavior and outcomes such as corporate performance.1 At the same time, in the main much less attention has been paid either to assembling econometric evidence on other related issues, such as the determination of executive pay, or to documenting stylized facts on the nature and evolution of key dimensions of corporate governance, such as board size and board structure and to incorporating such variables into econometric work.2 In this paper we make use of new panel data to make contributions in these areas. Our data are for a sample of Bulgarian firms during 1997–2001. This is an interesting period to examine for Bulgaria since it coincides with the introduction of economic reforms such as the adoption of widespread privatization and the introduction of a currency board. Before 1997 economic reforms in Bulgaria tended to be rather limited; after 1997 the reform process accelerated. During the period after 1997 the available evidence indicates that Bulgaria seems to be doing rather better macroeconomically than was the case before 1997.3 However there is a general paucity of empirically grounded microeconomic studies during this period. In addition,

Corporate Governance and Executive Compensation in Bulgaria

179

our data are particularly interesting because they are for firms that survive as functioning economic entities from what was originally a random sample of state-owned firms that was generated in 1989.4 By following our original sample during this reform period we are able to examine the impact both of privatization as well as firm origins on important dimensions of corporate governance. The structure of the paper is as follows. In the next section we discuss conceptual matters. Since ours is not a theoretical contribution this review is quite brief. To provide institutional context for our empirical work we then review key aspects of the transition in Bulgaria. This is followed by a discussion of our unusual data and then a section in which we report some basic descriptive statistics emerging from our new data for key aspects of corporate governance in Bulgaria during this period. In the next and main part of the paper we report our findings on the factors that appear to determine CEO pay in Bulgaria. Our strategy is first to use our panel data to estimate some baseline regressions that highlight those factors that previous theoretical and empirical work have highlighted – size, performance, ownership and firm origins. In more exploratory work we augment these specifications with other variables that are designed to capture other aspects of corporate governance that other work indicates are potentially important in the pay determination process but which do not yet appear to have been widely employed in econometric studies of firms in transition economies.

2. CONCEPTUAL FRAMEWORK Recent years have witnessed an explosion of literature that largely investigates the broad area of corporate governance for western firms. While diverse issues have been investigated, themes that have received particular attention include the nature and role of corporate boards, links between formal structures and decision-making patterns and the nature and determinants of executive pay.5 While the transition literature has not been immune from these influences, in the main the issues surrounding corporate governance that have attracted most attention in empirical work on transition have been more restricted. Our discussion concentrates on some of these themes that have been stressed in the transition literature, for example matters surrounding the managerial labor market during transition, especially the roles of different forms of enterprise ownership and the importance of firm origins. In addition, and recognizing that space limits do not permit a comprehensive discussion, we will also briefly highlight the

180

DEREK C. JONES AND MARK KLINEDINST

potential importance of other aspects of corporate governance that have tended to be comparatively neglected in empirical work on transition. When we then turn to examine the specific issue of the determinants of executive compensation we will provide a more integrated discussion of these potentially important but neglected dimensions of corporate governance. In addition, to provide context for our subsequent empirical work, from time to time we also briefly discuss findings from previous empirical work in this area for transition countries including Bulgaria during 1997–2001. One legacy of arrangements in Soviet-type economies was the existence of managerial reward systems in which pay was mainly a base wage and the pay of top managers was a low multiple of the average wage. Theorists have pointed out how these arrangements would be expected to result in acute incentive and motivational problems for managers (e.g. Bonin, 1976; Weitzman, 1976), produce extensive managerial slack (e.g. Ickes & Samuelson, 1987; Litwack, 1991) and, in turn, lead to several systemic inefficiencies including diverse pathologies of production (e.g. Putterman, 1993). Hence, in order to facilitate successful overall reform during early transition, many have stressed the crucial importance of reforming incentive systems (e.g. Aghion, Blanchard, & Burgess, 1994). For instance, when executive compensation is structured so as to provide pecuniary incentives for managers to pursue profitability, then arguably more market-oriented managerial behavior would be encouraged. In the context of early transition, downsizing of overstaffed state-owned firms and productivity increases appear to be key ingredients of successful reform. Arguably such adjustments will be facilitated when executive compensation is structured so as to reward managers for rational downsizing and productivity increases. Hence it is important to examine the nature and structure of CEO compensation contracts. Have pay relativities (CEO pay relative to averages and the low paid) changed during transition? Are these arrangements different according to firm ownership and whether the firm is new or not and are key features dependent on whether the manager is a new appointment or not? In addition, theorists have noted the existence of several problems in labor markets, especially those of moral hazard and adverse selection. Such problems are especially evident when information is asymmetric – for example when one party to the employment relation (usually the principal) has imperfect information about the other party (the agent) and obtaining reliable information is costly. In fast changing environments facing transition economies – though environments often faced by much institutional inertia (North, 1990) these issues are expected to be particularly important and to lead to diverse hypotheses. In particular, compared to state-owned firms,

Corporate Governance and Executive Compensation in Bulgaria

181

mainstream theorists hypothesize that potentially there will be acute differences in principal–agent relationships in privatized firms. In turn this will produce variation both in the structure of executive compensation and in the effects of the structure of executive compensation on firm performance. Thus, some forms of private ownership can be expected to lead to stronger discipline on CEO compensation and to reduce the rent associated with state-owned firms.6 While the issues of the links between ownership and CEO pay as well as pay relativities have been examined in empirical work for transition economies (for a review of some of this work see Djankov & Murrell, 2002), other aspects of corporate governance that have tended to be stressed in recent work for western firms do not seem to have appeared much on the radar yet in work on transition economies. For example, recent work by Bebchuk and Fried (2004) argues that executive compensation is not most usefully viewed through an arms-length contracting approach whereby CEO compensation is structured in order to overcome agency issues arising from the relationships among shareholders, directors and executives. Instead they argue that the most fruitful way to approach these issues is through the lens of managerial power that stresses the often-considerable influence that executives exert over directors. It would seem that this theme is especially useful to examine in the context of transition economies where many have argued that insiders, notably managers, have accumulated enormous power (e.g. Aghion et al., 1994). Perhaps of most interest in transition economies with often volatile ownership configurations and important legacies on company behavior from previous eras, is whether there is expected to be a close link between ownership, formal board structure and patterns of influence within firms? Thus are different board structures (the extent to which different board members are assumed to represent different stakeholders) expected to be closely related to ownership structures (or rather instead to reflect concerns of different stakeholders)? Or concerning decision-making, it is important, for example, to see whether the state is perceived by key economic agents such as CEOs as exerting influence on certain matters even when the state is not an owner. And since many have argued that many firms in transitional economies are effectively run by non-managerial insiders, what is the evidence concerning the scope and extent of employee influence on decision-making, including in those firms in which they have significant ownership stakes? Another important issue concerning corporate governance that has attracted a lot of attention recently in the west is that of board size. Some have conjectured that for reasons such as attempting to appease conflicting

182

DEREK C. JONES AND MARK KLINEDINST

constituencies, company boards in western companies might sometimes become too large. In the context of transition economies, where often company boards represent an innovation, there is a parallel interest in board size and understanding whether there is an optimal size for the company board. A related issue is whether there are good reasons to expect that the situation might be different in firms that are de novo (compared to privatized firms that were formerly state-owned). For example, for reasons such as smaller average size, both company boards and CEO pay might be expected to be smaller on average in de novo firms than in long-established firms. The transition has also witnessed the emergence of non-state institutions outside of the firm that potentially impact the nature and effects of aspects of corporate governance. One dimension of this is membership in an employers’ federation. Does membership in such bodies act to enhance or moderate CEO pay? Is the information that membership in such a body provides expected to lead to different effects on CEO pay depending on whether the firm is state-owned or private? Another issue is that of the speed at which ‘‘market mentality’’ is being disseminated and whether this is leading to profound changes in the firm’s objectives as perceived by key agents. For example, do firms’ CEOs perceive that their raison d’etre is profit maximization or do other potential goals such as employment security and sales growth play prominent roles? Again, there are good reasons to believe that the situation might be different in firms that have started afresh and/or are located in the private sector and thus not as hamstrung from pursuing certain objectives as state-owned firms might be. Turning to empirical work, up to now very little detailed evidence has been furnished on many of these matters as to what is actually happening in managerial labor markets during transition. Several studies have highlighted the role of managerial power during transition. Most studies (e.g. as reviewed by Estrin & Wright, 1999) find that even where there has been mass privatization, managers are very powerful and exercise greater power than is usually held to be appropriate for the proper functioning of a market economy (Nuti, 1997). However, often these studies are based on small and unrepresentative samples of firms. In addition, there are some empirical studies which point to the potentially important role of differences in management behavior in accounting for at least some of the differences in firm adjustment during early transition (e.g. Pinto, Belka, & Krajewski, 1993 for Poland). Arguably such differences at least in part reflect differences in management quality that, in turn, are linked with differences in the structure of executive compensation. Moreover, the work by Groves, Yongmiao, McMillan, and Naughton (e.g. 1995) points to the role that corporatization

Corporate Governance and Executive Compensation in Bulgaria

183

(rather than privatization) plays in producing better motivated management (and, in turn, enhanced enterprise performance). For Bulgaria, there has been work on CEO pay determination. While these studies do find a role for firm size and enterprise performance, the relevant elasticities have been found to be quite low (which is difficult to explain only though an agency view). In addition, previous empirical work for Bulgaria has found that one of the most important factors both economically as well as statistically of CEO pay is ownership, with managers of state firms earning considerable rents. As yet, however, there do not appear to be any studies that investigate the impact of other dimensions of corporate governance such as membership of an employers federation, board structure or patterns of influence.7 In addition, there are no studies that examine the period since the introduction of mass privatization in Bulgaria.

3. TRANSITION IN BULGARIA Turning to transition in Bulgaria it is clear that the Bulgarian transition continues to be a most difficult one.8 During the early years of transition, compared to many other transition countries (especially potential EU accession countries), Bulgaria faced unfavorable initial conditions. These included the impact of United Nations’ sanctions against Yugoslavia and Iraq that hurt Bulgaria more than many countries and high external indebtedness. But there were also substantial policy errors (including limited restructuring and high levels of corruption), which combined to produce high inflation, little foreign direct investment, a collapse of traditional markets and dampen economic progress. Not until 1994 was there positive economic growth, and even that was short-lived culminating in a massive financial crisis in 1996–1997.9 Characterizing the Bulgarian transition strategy from 1990 to 1997 is not straightforward because, in some ways the Bulgarian approach apparently was quite radical – including rapid price liberalization, a new competition policy and extensive and swift small-scale privatization – and thus is reminiscent of ‘‘big bang’’ experiences elsewhere.10 At the same time, the pace of change in other areas was excruciatingly slow and there were major policy blunders (e.g. Bristow, 1996). While formal policy changes have been heterodox, both the implementation of policy measures and the receptiveness of economic agents to changes seem to have been quite uneven. Consequently the de facto pace of institutional change often turns out to be far less than a casual observer might expect

184

DEREK C. JONES AND MARK KLINEDINST

based on de jure changes and arguably (Jones & Miller, 1997) the pace of change, in fact, has been far less dramatic than the indicators in studies such as E.B.R.D. (1996) would suggest. In particular, the fear of accelerating unemployment paralyzed the political process. The move to privatization of large state-owned firms was quite limited (at least until the beginning of 1997). Hence during this early period a focus of reform efforts was on corporatization rather than wholesale privatization. In this respect the situation confronting Bulgarian managers was apparently more similar to that confronting Chinese managers than managers of privatized firms in many transition economies in the former Soviet Union. However, there are important differences. For one thing, evidence from surveys (Jones, Klinedinst, & Rock, 1998) points to Bulgarian managers, absent ownership changes, exercising considerable influence in the typical firm. Furthermore, while privatization is not necessarily a requirement for restructuring, various special interest groups have managed to retard attempts to restructure (Pamouktchiev, Parvulov, & Petranov, 1997) rather than look for ways to improve state enterprise efficiency by improving governance structures, managerial positions have become part of the political spoils system. Perhaps most important, because of the failure to deal with the problem of bad debts, the context within which Bulgarian managers operated often was characterized by continuing soft budget constraints. The financial implosion of 1996–1997 was followed by the establishment of a Currency Board and significant structural changes and institutional reforms. These initiatives have led to visible and sustained improvements in some important dimensions of the macroeconomic context.11 Thus the last few years have seen positive economic growth and a much lower inflation rate. However, the level of real GDP is still only about 80% of the 1989 level. Moreover, the labor market is beset with many profound problems including an official unemployment rate that currently is close to 20% and which during the last several years has hovered officially in the mid- to upper teens.12 Since our econometric work hinges on differences in ownership structures, it is important to consider key aspects of the legal institutional changes that have occurred in Bulgaria in recent years. Compared to other cases (including Russia and also those of Poland, Hungary and the Czech Republic), developments in transition economies such as Bulgaria in general received much less attention, and also much less is known about them. Throughout the first half of the 1990s, and largely reflecting an unstable political environment with large and frequent shifts in government policies,

Corporate Governance and Executive Compensation in Bulgaria

185

privatization in Bulgaria proceeded very slowly. Indeed during 1992–1997 it is estimated that at most 7% of state assets were privatized (Miller & Petranov, 2000). Thereafter a series of privatization initiatives were introduced. These included a program of mass privatization that closely followed the Czech scheme and included provisions for the establishment of investment funds and privatization auctions. A quickening of the pace of privatization was also shown by the use of other privatization vehicles including cash and insider privatization. Insider privatization followed a similar pattern as overall privatization, but with a slightly more sustained pattern. The upshot of these initiatives was to dramatically accelerate the pace at which privatization proceeded and the extent to which the economy became privately owned – more than 60% by the end of the millennium. In addition, there was dramatic evidence of the emergence of significant concentration of ownership in individual Bulgarian firms. Furthermore, there is evidence of the emergence of significant diversity in the patterns of ownership in Bulgarian firms. This is exemplified by the emergence of privatized firms in which insiders as well as different types of outsiders (including privatization funds and individuals who did not work in the firm) hold majority ownership, sometimes in the same industry.

4. THE DATA AND BASIC FINDINGS ON ASPECTS OF CORPORATE GOVERNANCE The data were collected from several different sources. Our latest data came from a survey of firms administered in the spring and fall of 2002 and from the National Statistical Institute of Bulgaria. The firms were originally selected back in 1989 to represent a random sample of industrial firms across regions and industry. Out of our original sample of 490 companies, 344 are still operating, 26 are either drafting a restructuring plan or proceeding with one, while another 72 have been or are in the process of being liquidated.13 The data we use essentially represent new waves of data collection from several sources as used and described in earlier studies (e.g. Jones & Kato, 1996 on CEO pay determination). The main sources are the Bulgarian Management Survey (BMS), the Bulgarian Economic Survey (BES) and the Bulgarian Labor Flexibility Survey (BLFS). The BLFS was a project sponsored by the ILO to assess microeconomic changes in labor practices in Bulgarian industry. The BLFS involved 490 establishments, selected to ensure a nationally and sectorally representative sample. Specifically, the

186

DEREK C. JONES AND MARK KLINEDINST

population was defined as all state-owned (in 1989) Bulgarian manufacturing organizations (SOEs) that operated on a for-profit basis and had more than 80 employees in 1992, the year of the first wave of data collection.14 Subsequent waves of data collection took place in 1994 and in 1996. For this study we collected data on all surviving firms from that sample for the period 1997–2001. The original BMS collected survey data from the chief executive officers in the same 490 Bulgarian firms. A wide variety of questions were asked including information about chief executives, including pay and the method and terms of appointment. Data were also gathered concerning some firm characteristics, for example the form of enterprise ownership. For this study a new survey was designed (though there was considerable overlap in questionnaire design from earlier waves) and administered during 2002. The data that we use in this paper are essentially drawn from the current waves of data collection. The nature of these data inevitably influences the empirical strategy that we are able to undertake at this time. Importantly, our data constitute a panel and are also quite rich in important respects. For example, we are able to construct diverse measures of interesting variables such as the extent of influence of major groups on different decisions. All-in-all this enables us to undertake a number of different specifications in our quest to rigorously analyze the impact of privatization and business performance on the determination of CEO pay.15 But first we draw on our new data to report a number of stylized facts on what is happening to important dimensions of corporate governance in Bulgaria.16 We do this in large part because, as noted earlier, while there has been much conjecturing on diverse issues in the broad area of corporate governance, the underlying empirical evidence is often thin. In this section of the paper we use our new data to provide some new facts and also, so far as possible, we draw comparisons from other work. Our findings are presented in a number of tables. In Table 1 we present information on board size and board structure.17 From Part (A) we see that in all years the median board size category is about 1–3. In addition, this compact modal class is becoming more commonplace – median board size is falling. Perhaps surprisingly from Part (C) we see that there are far more firms with more than three directors in de novo firms while both state-owned and privatized firms that were formerly state-owned, on average, tend to have tiny boards. Part (B) details the formal influence of managers as reflected in managerial presence on company boards. Perhaps the most conspicuous example of managerial power that these data reveal is that in more than one in three

187

Corporate Governance and Executive Compensation in Bulgaria

Table 1.

Size and Composition of the Company Board. 1997

1999

2001

(A) Size of company board 1–3 4–6 7–9 49

47 39 12 2

52 35 10 2

60% 30 8 2

(B) Fraction of the board that represents managers 0% 40%o20% 420%o40% o40%o60% o60%o80% o80%o99% 100%

14 13 22 8 7 1 35

15 12 19 9 9 0 36

14 11 22 6 9 0 38

54 46

53 47

54 46

53 41 6

56 38 6

53 41 6

25 30 33 12

23 29 29 9

25 30 33 12

(C) By ownership and origin State 1–3 4–6 Old private 1–3 4–6 7–9 New private 1–3 4–6 7–9 49

Notes: (1) All entries are percentages of total respondents. (2) Entries may not sum to 100% because of rounding.

firms all directors are managers. In more than half of sample firms, managers’ representatives held at least 40% of seats on the board. In Table 2 we present information on executive pay both in absolute terms and relative to the earnings received by the average worker and the lowest-paid worker. From Part (A) it is clear that, with the exception of 1997, the year following the crisis, real CEO pay has tended to grow. Also while pay relativities have tended to widen they still remain quite narrow with the average CEO receiving less than four times the average worker in 2001. The cramped nature of internal wage systems is also revealed in the ratio of executive pay to the earnings of the lowest worker. Reflecting policy

188

Table 2.

(A) All firms Defceopay Ceo/avpay Ceo/lowpay

1997

1998

1999

2000

2001

All years

758.9 (560.3) 3.31 (2.36) 6.63 (7.80)

677.8 (464.8) 3.30 (1.98) 6.33 (4.46)

753.4 (593.5) 3.47 (2.32) 6.66 (6.02)

792 (5844) 3.62 (2.17) 6.73 (4.88)

810 (601.14) 3.77 (2.58) 6.66 (4.67)

759.5 (565.1) 3.5 (2.30) 6.61 (5.68)

741 (428) 2.74 (1.06) 5.20 (2.83)

710 (319) 3.05 (1.18) 5.85 (3.38)

774 (354) 3.16 (1.11) 5.58 (2.68)

860 (400) 3.51 (1.44) 6.28 (3.49)

878 (422) 3.54 (1,48) 6.15 (2.88)

746 (421) 3.50 (2.31) 5.91 (4.91)

863 (597) 3.64 (2.49) 7.59 (9.27)

765 (510) 3.54 (2.04) 6.78 (4.29)

871 (679) 3.85 (2.65) 7.46 (6.64)

910 (650) 3.98 (2.29) 7.51 (5.21)

919 (645) 4.07 (2.33) 7.45 (4.94)

865 (616) 3.82 (2.38) 7.33 (6.32)

482 (440) 2.79 (2.54) 4.91 (4.65)

418 (307) 2.76 (2.14) 5.36 (5.55)

439 (309) 2.67 (1.73) 5.26 (5.77)

466 (362) 2.82 (2.05) 5.03 (4.42)

516 (486) 3.19 (3.47) 5.07 (4.57)

467 (397) 2.88 (2.55) 5.18 (5.11)

Note: All value entries are in real 2001 leva.

DEREK C. JONES AND MARK KLINEDINST

(B) By ownership and origin State-owned Defceopay Ceo/avpay Ceo/lowpay Old private Defceopay Ceo/avpay Ceo/lowpay New private Defceopay Ceo/avpay Ceo/lowpay

Chief Executive Pay: Means (Standard Deviations).

Corporate Governance and Executive Compensation in Bulgaria

189

makers’ raising minimum wage rates the ratio has hovered around 6.6 during the five-year-study period. When similar data are assembled for firms with different ownership and for new private firms, some interesting patterns emerge. Most apparent is the relatively low pay of CEOs in new private firms. In 2001 on average they earned 40% less than their counterparts in both state and privatized firms.18 While CEOs in private firms that were formerly state-owned firms are the best paid they are only marginally better paid than is the average CEO in a state firm. The observed difference in CEO pay changes over the 1997–2001 period contrast sharply with findings for earlier periods. For example during 1992–1995, managers in state firms earned considerably more than their counterparts in corporatized and privatized firms. The elimination of this difference during the current period tends to point to the absence of rent associated with state-owned firms although the final verdict will need to wait for multiple regression analysis in which firm performance is controlled for. Perhaps reflecting smaller average size, internal wage structures within new private firms are also found to be more compressed than in other firms. Part (A) of Table 3 reveals that about 36% of firms have contracts for CEOs that provide for PBC; there is some indication that the incidence of such contracts is falling. More noticeable and surprising are the differences by type of ownership. Performance pay is more than twice as commonplace in state-owned firms as in privatized firms and more than three and one-half times more frequent then in new private firms (although the pay in new and mainly smaller firms may not capture potential future equity value owned by the CEOs). The data reported in Part (B) of Table 3 reflect CEOs being asked to assign a total of 10 points to 3 possible firm objectives. Perhaps the more interesting finding to emerge from this exercise is that when the data are disaggregated by ownership and firm origins, no strong and clear differences are apparent in the emphasis placed on these different objectives for firms that are either state-owned, private or de novo. In Table 4 we report some illustrative findings on patterns of the distribution of influence within firms. These data represent the views of CEOs who used a 4-point Likert scale to gauge the influence on several different issues of six different groups – managers, employees, foreign owners, owners in general, banks and financial institutions and the state. Issues included strategic issues (such as the development of long-term plans and employment reduction) as well as matters with more of a workplace focus (such as safety and health). These data clearly show that on most issues in about two-thirds of the cases managers believed that they exercised considerable influence. This is in

(A) Performance-based pay All firms State Old private New private

190

Table 3.

Form of the Contract and Firm Objectives.

1997

1999

2001

41 61 40 24

40 67 37 22

36 76 32 21

0% 8 7 3

0%o20% 29 34 7

420o40% 51 45 54

440o60% 9 10 29

460o80% 1 2 5

480o100% 0 0 0

100% 1 2 2

9 9 8

31 26 35

46 54 49

14 9 8

0 1 1

0 0 0

0 1 0

12 7 4

28 40 26

46 40 53

10 10 12

4 1 3

0 0 0

0 2 1

0 5 1

9 4 12

53 53 57

31 30 26

4 6 4

0 0 0

3 2 0

(B) Firm objectives (1) All firms

Notes: (1) Entries in Part (A) represent the percentages of firms where CEOs have contracts with PBC. (2) Entries in Part (B) represent the percentage of respondents who gave that objective the indicated weight (respondents were forced to choose answers that summed to 100). (3) Entries may not sum to 100% because of rounding.

DEREK C. JONES AND MARK KLINEDINST

Profit maximization Employment security Sales growth (2) By ownership and firm origin Profit maximization State Old private New private Employment security State Old private New private Sales growth State Old private New private

191

Corporate Governance and Executive Compensation in Bulgaria

Table 4.

The Distribution of Influence on Selected Issues. Long-Term Plans (LTP)

Employment Reduction (LRED)

Wage Setting (WAG)

Safety and Health (SH)

(A) Management High Moderate Low None

67 18 12 3

65 13 3 20

60 15 4 20

66 11 5 18

(B) State High Moderate Low None

12 6 11 71

9 5 8 78

10 5 8 78

14 5 10 70

(C) Employees High Moderate Low None

3 12 18 67

6 14 18 62

6 20 23 50

13 30 18 38

Notes: (1) All entries are percentages of total respondents. (2) Entries may not sum up to 100% because of rounding. (3) Respondents were asked to assess the influence of different groups on various issues using a 4-point Likert scale.

sharp contrast to assessments for the influence exercised by other groups. Indeed the degree of power attained by all other groups tended to attain a ‘‘high’’ level in about only one in 10 cases and then for selected issues (and not across the board for all issues). In other words, on average CEOs believe that they are very much in charge and they are not being hamstrung by ‘‘excessive’’ influence from other quarters whether this arises from owners, the state or other employees. Moreover, in unreported findings, when data on patterns of influence are assembled by enterprise ownership and according to type of privatization, no pronounced differences are apparent. In other words, on average, Bulgarian managers exert considerable power irrespective of firm ownership. Finally we look at measures of size and performance (Table 5). We consider two alternative size measures: (i) Emp (total employment); (ii) Defsales (real sales in 2001 leva). Table 5 shows that there was significant downsizing during the period 1997–2001 as average employment in sample firms fell from 300 to 192. As such this continued a process that was apparent during earlier periods for these firms (see, e.g. Jones et al., 1998). More surprisingly perhaps, real sales fell on average by about a similar amount – just over a

192

DEREK C. JONES AND MARK KLINEDINST

Table 5.

All firms Emp Defsales Rvaddprod Rsalprod State Emp Defsales Rsalprod Rvaddprod Oldprivate Emp Defsales Rsalprod Rvaddprod Newfirm Emp Defsales Rsalprod Rvaddprod

Sample Firms: Size and Performance, 1997–2001: Means (Standard Deviations). 1997

1998

1999

2000

2001

All years

300 (472) 9933 (33251) 9.3 (27.1) 22.75 (61.3)

300 (539) 7308 (21620) 8.7 (10.8) 17.8 (21.5)

233 (387) 6203 (16911) 10.3 (17.7) 18.7 (28.6)

200 (293) 5833 (16963) 12.2 (20.4) 21.1 (30.2)

192 (295) 6133 (18259) 14.3 (35.6) 23.6 (42.9)

243 (409) 7082 (22290) 11.1 (24.0) 20.8 (35.6)

447 (871) 24549 (68757) 21.1 (30.6) 8.8 (11.4)

569 (1147) 15133 (43069) 20.1 (27.9) 8.5 (12.2)

326 (668) 10843 (25172) 19.7 (30.0) 11.0 (12.7)

262 (343) 8193 (16637) 21.0 (30.4) 11.1 (10.9)

206 (240) 8996 (21119) 36.9 (88.3) 26.5 (81.7)

370 (749) 13953 (40719) 24.4 (49.4) 13.7 (40.2)

298 (369) 9473 (26213) 27.2 (75.6) 10.5 (33.8)

281 (356) 7384 (17924) 19.6 (21.5) 9.3 (10.8)

232 (319) 6731 (18058) 20.4 (30.3) 11.1 (20.1)

217 (316) 6890 (20716) 22.8 (27.8) 12.2 (17.4)

212 (327) 6959 (21321) 23.8 (30.5) 12.8 (18.1)

250 (346) 7446 (21462) 22.7 (42.0) 11.2 (21.4)

168 (179) 2872 (7633) 10.8 (20.0) 5.6 (8.6)

150 (150) 2605 (7020) 11.24 (17.9) 7.1 (10.3)

122 (125) 1827 (4876) 9.28 (16.9) 5.01 (6.7)

102 (100) 1543 (3961) 10.1 (15.8) 6.0 (6.8)

105 (118) 1377 (3606) 8.2 (9.0) 5.4 (5.3)

124 (131) 1816 (5075) 9.5 (14.8) 5.7 (7.0)

Corporate Governance and Executive Compensation in Bulgaria

193

third. However, when these indicators of size are examined by ownership and firm origins it is apparent that the downward trends are strongest for firms that remain in the state sector. For example, for the average stateowned firm employment fell by about 241 while real sales contracted by almost two-thirds. This compares with drops in employment and sales in old privatized firms of less than 30%.19 For performance the available data enable us to examine two measures of productivity, namely real sales per worker (Rsalprod) and real value added per worker (Rvaddprod). For all firms while there is not much movement for the average firm in average sales per worker, value-added productivity has increased sharply – by more than 54% in five years. Moreover, when these measures are examined by the type of ownership there are some surprising differences. By far the biggest gains have been made by state-owned firms. Real value added per worker in new private firms has barely changed during the five-year period. As such these trends contrast with data from earlier periods – before the onset of mass privatization, the comparable figures for firms in this sample show labor productivity falling during the four years preceding 1996.

5. DETERMINANTS OF EXECUTIVE COMPENSATION In this section we turn to examine one key issue in the broad area of corporate governance and executive compensation in more depth, namely the determinants of executive compensation. We focus on this matter in large part because this is an issue for which previous work exists in Bulgaria (e.g. Jones & Kato, 1996). Also, to date econometric work has been restricted to fairly parsimonious specifications (focusing only on size, performance and ownership) and during periods when there was limited privatization. In view of our previous discussion about the potential importance for executive compensation of other factors such as board structure, patterns of influence and firm origins this is an area that lends itself naturally to more extensive empirical investigation. Moreover, it is an interesting exercise to see if those factors that under previous economic conditions were found to exert varying influence on executive pay, are found to play similar roles in the pay determination process after the spread of private ownership and other reforms. To study the determinants of the level of CEO pay, we estimate a standard CEO pay log–log model using a panel of data during 1997–2001. Two

194

DEREK C. JONES AND MARK KLINEDINST

sets of regressions are estimated – our ‘‘baseline’’ regressions and a series of more ‘‘exploratory augmented regressions’’. In the first set of baseline regressions we augment a standard chief executive compensation equation20 with dummy variables indicating whether the firm was state-owned or not and, if privatized, whether the firm was de novo or not as well as controls for region, industry and time. That is, lnPayit ¼ ait þ bnlnðSIZEit Þ þ cnlnðPERFORMANCEit Þ þ d nðSTATEit Þ þ enOLDPRIVit þ f nREGi þ gnINDi þ hnTIMEt þ uit ð1Þ where Payit ¼ chief executive pay of firm i in year t; SIZEit ¼ size of firm i in year t; PERFORMANCEit ¼ standard firm performance measures such as various productivity measures of firm i in year t; STATEit ¼ 1 when a firm is state-owned, 0 otherwise; OLDPRIVit ¼ 1 if a privatized firm was originally a state firm, with 0 indicating the firm is de novo; INDi ¼ a series of nine industry dummies; REGi ¼ a series of five regional dummies; and TIMEt ¼ year effects. The disturbance term, uit, is assumed to be distributed NID(0, s2) and we estimate random effects models.21 PERFORMANCE and SIZE are standard categories that have been included in prior empirical studies of executive compensation in the U.S., the U.K. and other advanced market economies and in the limited work on transitional economies. In the western literature, the application of principal-agent theory to the design of executive contracts in general predicts a positive correlation between managerial pay and some observable measure of firm performance (which eventually translates into improved well-being for shareholders).22 To adequately measure PERFORMANCE, the debate in the western literature has usually centered on the respective merits of measures of stock market returns compared to various accounting measures such as ROA (the return on assets). However, in a context of embryonic capital markets as in Bulgaria, this debate is moot since stock market measures are highly suspect. Moreover, many have argued (e.g. Pohl, Anderson, Claessens, & Djankov, 1997; Earle & Estrin, 1995) that the key performance measure may be labor productivity. Thus, in our empirical work we focus on different measures of labor productivity as an alternative firm performance measure. Both performance measures are entered as logarithms. The inclusion of a measure of SIZE in empirical work in western literature usually follows from theories which stress the importance of factors such as spans of control in determining CEO pay (e.g. Klinedinst, 1991). For transition economies, another consideration is that under communism being a chief executive of a larger firm with many employees often translated into

195

Corporate Governance and Executive Compensation in Bulgaria

more political power and thus an improved ability to obtain higher pay. To see if this force persists, our data allow us to use two alternative SIZE measures: (i) lnemp (the log of the number of workers); (ii) lndefsales (the log of real sales – in 2001 real leva).Year dummy variables are included to capture technological change and other shocks common to all firms as well as possible measurement errors of inflation. In addition we include controls for industry (IND) and regional (REG) effects. We estimate Eq. (1) by merging data from the aforementioned three surveys (the BMS, the BLFS and the BES). We successfully assembled an unbalanced panel of up to 215 firms that provide data necessary for our regression analysis from 1997 to 2001. We estimated four specifications of Eq. (1) depending on the selection of the size and performance measures.23 Importantly, for all specifications reported in Table 6, separate F tests refute the joint exclusion of year, region and industry dummy variables at the 1% level. The first and third columns of Table 6 show the results when the log of employment is used as a size measure, whereas columns 2 and 4 report findings using the log of real sales as a measure of size. For each size measure, results are reported when the two different performance measures are used (for example, real value added per worker, Rvaddprod in columns 1 and 2). Finally, all models include

Table 6.

Determination of CEO Pay 1997–2001 Baseline Regressions (Standard Errors in Parentheses). (1)

Lnemp

(2)

0.1705

1nrvaddprod

0.158 (0.0221)

(0.0271) 0.1838 (0.0226) 0.0297 (0.0236)

Lnrsalprod State Oldprivate

0.2606 (0.140) 0.342 (0.121)

(4)

0.159

(0.0218) Lndefsales

(3)

0.2065 (0.1388) 0.2572 (0.120)

0.1593 (0.0271)

0.2264 (0.025) 0.2072 (0.135) 0.2510 (0.121)

0.06698 (0.033) 0.2072 (0.1393) 0.2510 (0.1210)

Notes: (1) All estimates include controls for time, region (5) and industry (9). (2) The number of estimations was always 842. (3) All variables are defined in the appendix. Significant at 1%. Significant at 5%.

196

DEREK C. JONES AND MARK KLINEDINST

dummy variables for whether or not the firm was state-owned or a privatized firm that was formerly state-owned (as opposed to a new private firm). We begin by examining the impact of ownership and firm origin on CEO pay. From Table 6 we see that the ‘‘OldPrivate’’ coefficients are statistically significant in all four cases (5% or better). This is strong evidence that whether a firm is a new firm (the base case) matters much in the determination of CEO pay. Specifically we see that CEOs in firms that are privatized and are former state-owned earn higher pay then their private counterparts that are new firms, even after controlling for size and performance. These positive and significant estimates suggest that, after controlling for firm performance, the average CEO working for a former state-owned and privatized firm receives, depending upon specification, 25–34% more pay compared to his/her counterpart in newly privatized firms. There is also some, though much weaker, evidence that managers of state-owned firms earn rents. In one of four cases chief executives in state-owned firms earn 26% more pay than do their counterparts in new private firms, other things equal. Hence, there is some weak evidence that the rent earned by state firm CEOs seems to be significant not only statistically but also economically. Turning to the relationship between executive pay and size, no matter which measure of size is used, evidence is found of a positive relationship and all four size coefficients are statistically significant at the 1% level. While results are not sensitive to the choice of the performance measure, the estimated pay elasticities of size are in the range of 0.15–0.18. For example, as sales increase by 10%, CEO pay increases by 1.5% to 1.8%; for employment, the comparable effects are similar. These elasticities are slightly lower than those obtained in other studies. For example, Rosen (1990) in reviewing various western studies on the estimated elasticity of pay with respect to scale finds a typical value of 0.25 while Jones and Kato (1996) report elasticities of size of 0.2–0.4 for Bulgaria. Finally, there is also strong support for the existence of a relationship between CEO pay and performance. For both measures of productivity, statistically significant relationships are found at the 1% level. Interestingly these elasticities are sometimes larger than those obtained in other studies. Thus, Rosen (1990) finds that the estimated sensitivity of pay to accounting measures are in the 1.0–1.2 range, whereas our estimated elasticities for Bulgaria for productivity (measured as real value added or lnvaddprod) range from 0.03% to 2.3% depending on the size measure. Moreover these findings contrast with earlier findings for Bulgaria when only a weak link with productivity was previously found. This finding of a strong link between pay and performance contrasts sharply with earlier findings.

Corporate Governance and Executive Compensation in Bulgaria

197

It suggests that conditions that better facilitate managers being able to give improved attention to comparative business performance have now been provided, thus representing a major change from earlier periods. Increasingly managers are being forced to devote more attention to performance. In the remaining part of this section we report findings from a second and more exploratory set of regressions. These findings are reported in Tables 7–9. In these regressions we draw on some of the findings reported earlier on aspects of corporate governance to extend the basic framework that is indicated by Eq. (1) and which we have reported in our baseline regressions. For example, in Table 7 we augment the basic specification to include variables that reflect three important dimensions of corporate governance and executive compensation that are highlighted in our survey data. The three additional variables are: (i) the fraction of the firm’s labor force that Table 7. CEO Pay Determination: Baseline Regressions Augmented by the Incidence of Performance-Based Compensation (PBC), the Extent of Managerial Ownership (MANOWN) and Membership of an Employers’ Association (BIA). (1) lnemp lnrsal

0.182 (0.037) 0.147 (0.034)

(2)

0.035 (0.05)

lnrsalprod

Oldpriv BIA PBC MANOWN

0.258 (0.160) 0.206 (0.118) 0.127 (0.104) 0.180 (0.074) 0.574 (0.208)

(4) 0.214 (0.037)

lnrvaddprod

State

(3)

0.258 (0.160) 0.206 (0.118) 0.127 (0.104) 0.180 (0.074) 0.574 (0.208)

0.003 (0.043) 0.182 (0.037) 0.245 (0.161) 0.197 (0.117) 0.131 (0.106) 0.183 (0.075) 0.501 (0.204)

0.148 0.034) 0.162 (0.029) 0.255 (0.160) 0.232 (0.115) 0.135 (0.105) 0.163 (0.074) 0.576 (0.208)

Notes: (1) Standard errors are in parentheses. (2) All estimates include controls for time, region (5) and industry (9). (3) All variables are defined in the appendix. Significant at 1%. Significant at 5%. Significant at 10%.

(1) Lnemp

lnvaddprod

The Determination of CEO Pay: Models Including Management Influence. (2)

0.1777 (0.0309)

Lndefsale 0.1214 (0.024)

Dmanltp1 Dmanltp2 Dmanltp3

(4)

0.187 (0.024) 0.134 (0.265)

0.1919 (0.1631) 0.2586 (0.143) 0.1437 (0.0181) 0.0337 (0.000) 0.2349 (0.1234)

0.170 (0.0298) 0.113 (0.023) 0.1884 (0.026) 0.192 (0.163) 0.2571 (0.122) 0.1435 (0.078) 0.0345 (0.091) 0.235 (0.123)

0.0182 (0.3661) 0.192 (0.164) 0.257 (0.143) 0.144 (0.0787) 0.0346 (0.091) 0.235 (0.1234)

Dmanlred2 Dmanlred3 708

708

(6)

0.175 (0.031)

Dmanlred1

N

(5)

708

0.3042 (0.167) 0.3615 (0.146)

0.193 (0.186) 0.233 (0.101) 0.191 (0.156) 714

(7)

(8)

0.169 (0.030) 0.183 (0.024) 0.012 (0.025)

0.254 (0.165) 0.2881 (0.144)

0.194 (0.085) 0.249 (0.010) 0.193 (0.153) 714

0.169 (0.030)

0.183 (0.206) 0.255 (0.165) 0.280 (0.145)

0.195 (0.085) 0.2481 (0.099) 0.194 (0.153) 714

0.013 (0.036) 0.255 (0.165) 0.280 (0.145)

0.195 (0.085) 0.248 (0.010) 0.194 (0.153) 714

Notes: (1) Standard deviation in parentheses. (2) The number of observations was always 842. (3) All variables are defined in the appendix. Significant at 1%. Significant at 5%. Significant at 10%.

DEREK C. JONES AND MARK KLINEDINST

OldPriv

0.2364 (0.168) 0.3336 (0.144) 0.1418 (0.0799) 0.01188 (0.093) 0.2476 (0.1256)

(3) 0.1702 (0.0298)

lnsalprod State

198

Table 8a.

Corporate Governance and Executive Compensation in Bulgaria

Table 8b.

The Determination of CEO Pay: The Role of Managerial Influence (Continued). (1)

Lnemp

(2)

0.180

0.176

(0.031)

(0.030)

0.120 (0.023)

Lnrsalprod State Oldpriv Dmwage1 Dmwage2 Dmwage3 N

0.286 (0.164) 0.348 (0.144) 0.250 (0.082) 0.215 (0.097) 0.329 (0.167) 718

(3)

0.190 (0.024) 0.010 (0.025)

Lndefsale Lnvaddprod

199

0.194 (0.026) 0.231 (0.162) 0.259 (0.144) 0.253 (0.082) 0.227 (0.096) 0.328 (0.164) 718

0.230 (0.162) 0.261 (0.144) 0.252 (0.081) 0.226 (0.096) 0.330 (0.164) 718

(4)

0.174 (0.030)

0.020 (0.039) 0.231 (0.163) 0.259 (0.144) 0.253 (0.082) 0.227 (0.096) 0.328 (0.164) 718

Note: All estimates include controls for time, region (5) and industry (9). See notes in Table 8a. Significant at 1%. Significant at 5%. Significant at 10%.

are managers as well as owners (MANOWN); (ii) whether the firm is a member of the main employers’ association, the Bulgarian Industrial Association (BIA); and (iii) whether the chief executive’s contract provides for PBC. While to date these do not appear to have been used in econometric investigations of executive pay in transition economies, some of these variables have appeared in studies elsewhere. Our first finding is that, in the main, the addition of these explanatory variables does not alter the basic story that has already been reported. That is, reassuringly, the coefficients for most of the variables in our core model – reflecting size, performance, ownership and whether a firm is new or not – are essentially unchanged from those reported earlier. The only exception is that CEO pay determination is no longer as significantly related to whether or not a firm is de novo.24 Consistent with those who hypothesize that managerial ownership can be used to help to resolve agency problems (but also consistent with those who

200

Table 9.

DEREK C. JONES AND MARK KLINEDINST

The Determination of CEO Pay: The Role of State Influence. (5)

Lnemp

(6)

(7)

0.171

0.161 (0.032)

(0.0329) Lndefsale Lnvaddprod

0.131 (0.025)

0.186 (0.026) 0.002 (0.028)

Lnsalprod State Oldpriv Dstalred1 Dstalred2 Dstalred3 N

0.306 (0.182) 0.362 (0.162) 0.202 (0.086) 0.218 (0.129) 0.041 (0.106) 644

(8)

0.252 (0.182) 0.276 (0.162) 0.190 (0.085) 0.203 (0.127) 0.058 (0.105) 644

0.161 (0.032)

0.057 (0.058) 0.254 (0.182) 0.273 (0.163) 0.196 (0.085) 0.197 (0.126) 0.059 (0.105) 644

0.198 (0.028) 0.254 (0.182) 0.273 (0.163) 0.197 (0.085) 0.196 (0.127) 0.059 (0.104) 644

Notes: (1) Standard deviation in parentheses. (2) All estimates include controls for time, region (5) and industry (9). (3) The number of observations was always 842. (4) All variables are defined in the appendix. Significant at 1%. Significant at 5%. Significant at 10%.

argue that more managerial ownership leads to enhanced managerial power), CEOs in firms in which ownership by managers has a stronger presence earn higher pay, other things equal. Also, there is weak evidence that membership in an employers’ federation is seen as a force that tends to curb CEO pay. Lastly, we observe that CEO pay is lower by about 18% in firms in which CEOs pay is tied to enterprise performance. This result does not provide support for agency theorists but does offer comfort for those who articulate views based on growing managerial power. In Tables 8 and 9 we turn to examine the potential impact of the distribution of decision-making influence on executive pay. We are interested in investigating whether it is ownership or influence that matters most in determining CEO pay. Amongst potential decision-makers, reflecting the attention that has been devoted to management, in these exploratory regressions we concentrate on the role of managerial power. In these

Corporate Governance and Executive Compensation in Bulgaria

201

regressions, managerial power is measured by a series of dummy variables and this procedure is followed for a number of different types of issues. For example, managerial influence on long-term plans ranges from ‘‘high’’ influence (DMANLTP1) to zero (DMANLTP4). For the other issues for which we report illustrative results, namely employment reduction (LRED), wage determination (WAG) and safety and health (SH), we use four identical categories. Analogous procedures are used to examine the potential impact on CEO pay of varying degrees of power that other groups might possess. Thus a high degree of influence by the state on long-term plans is represented by ‘‘DSTALTP1’’ whereas a low degree of power by employees on employment reduction is captured by DEELRED4. In all of the reported regressions, a single set of dummy variables that captures the extent of influence by a particular economic agent is used to augment the basic specification that was discussed earlier. While we will first discuss the findings for the role of managerial power, we note that in most reported regressions the findings that emerge from the core-regression model are mainly unaffected by the inclusion of any of the separate vectors of influence variables.25 That is in most of the findings reported in Tables 8 and 9 (as well as in unreported findings) normally both size and performance continue to be statistically significantly associated with CEO pay and CEOs in new firms typically earn less that their counterparts in the private sector, other things equal. Also, the size of these core coefficients is not affected by the inclusion of sets of influence variables. In addition, we find weak evidence that managerial power has a strong and positive influence on CEO pay. For some categories of managerial influence this relationship is always found. Thus on all issues when managers perceive that they have ‘‘high’’ influence (D ¼ 1), and relative to a managers who perceive that they have no influence, CEO pay is higher by between 14% and 25%. Equally the particular pattern of effects tends to vary depending on the particular issue that is being used to proxy managerial power. Thus for the issue of employment reduction only high and moderate levels of managerial power are seen to impact CEO pay, whereas for wage determination all levels of managerial influence exert upward pressure on managerial pay and for long-term planning moderate managerial influence has no statistically significant effects. In Table 9 we report regressions that include a vector of dummies for state (rather than managerial) influence. When state power is measured for the issue of long-term plans we find that state influence does not appear to play a significant role in influencing CEO pay. But the picture is different when we turn to some other issues. This is clearly shown from the findings for

202

DEREK C. JONES AND MARK KLINEDINST

state influence on employment reduction. On this matter, always when state influence is perceived to be high (DSTALRED ¼ 1), and sometimes when it is moderate, CEO pay is adversely affected. All-in-all these selected regressions that purport to capture the role of influence by different agents on CEO pay point to influence playing a bigger role than formal ownership. Thus in most regressions ownership (as proxied by State) continues, as in our baseline regressions, not to be statistically significant. This contrasts sharply with findings for the role of ‘‘influence’’ particularly managerial influence.

6. CONCLUSIONS The results here show a number of important trends in executive compensation and corporate governance in Bulgaria. The size of corporate boards was falling over the period under investigation, 1997–2001, with only new firms having an average board size over three. Managerial influence is considerable, fully one-third of boards are all managers. This influence is reflected in the fact that real income for executives grew over this period as well as the ratio of CEO pay to the average and lowest employee. The pay ratios are still relatively narrow (6.7 highest to lowest) relative to other market economies, especially relative the U.S., where it is not uncommon to find ratios in the hundreds (Klinedinst, 1991; Kaplan, 1994). Moreover, from Table 2, where we report summary statistics when the sample is divided according to ownership, we see that in new private firms, that also have larger boards, fewer employees and sales pay their CEOs 40% less than their counterparts in state and private former state firms. These new firms also, somewhat surprisingly, are much less likely to offer performancebased pay for their CEOs than other firms (in 2001, 21% versus the overall average of 36%), even though the firm objectives are stated to be similar in the three different ownership groups. On important strategic issues (longterm plans, employment, etc.) two-third of managers typically thought they exercised considerable influence relative to the state or employees. For most firms, teamwork and a truly participative environment look to be years away as firms are created or redefine themselves apart from the state’s heavy hand. The tough choices firms have had to make during these years, especially within a macro environment that became dominated by the influence of the newly formed Currency Board, have resulted in real value added per worker going up from 1997 to 2001, particularly in the state sector.

Corporate Governance and Executive Compensation in Bulgaria

203

Size matters for CEO pay, a finding that is consistent with studies from other market economies. This correlation with size may represent the additional administrative duties that come with more employees and sales, but it could also represent the strong control that managers have relative to other stakeholders. Performance measures for the firm are also found to be strongly correlated with executive compensation and show an increased awareness of market discipline. This restriction on CEO pay is also evidenced in the lower compensation found for CEOs in firms that are members of the Bulgarian Employers’ Federation and that have PBC schemes, some that included ownership as well. Even though Bulgarian firms are influenced by these professional organizations, market performance indicators and in some cases the state, the strong influence that managers have over many strategic decisions make their pay become more often a reflection of their internal power in the firm, a finding similar to those for other market economies (e.g. Bebchuk & Fried, 2004; Klinedinst, 1991; Rosen, 1990). The reforms have brought great market discipline, but CEO pay, even if it is limited to the fraction that can be measured, needs greater performance linkage to ensure that Bulgarians firms remain competitive.

NOTES 1. For recent reviews for transition economies, see Djankov and Murrell, 2002; Bevan, Estrin, and Schaffer (1998) and Jones (2004). These reviews reveal that there is far less agreement than seems to be the case at first blush. On the impact of ownership and other aspects of corporate governance on firm performance for the case of Bulgaria, see Jones, Klinedinst, and Rock (1998) and, more recently, Jones and Klinedinst (2003). 2. Of course, there are some exceptions. For example, for an earlier study of executive compensation in Bulgaria, see Jones and Kato (1996). For a recent study of wage determination in Bulgaria, see Jones and Simon (2005). 3. For accounts and evaluation, see E.B.R.D. (2003). Note, however, that there are important exceptions to an overall favorable assessment of macro performance – for example, unemployment remains stubbornly high at close to 20%. 4. Studies of earlier waves of data for this sample include, Jones, Klinedinst, and Rock (1998), Jones, Kato, and Avramov (1995) and several essays in Jones and Miller (1997). 5. See, for example, the papers in the Journal of Economic Perspectives (vol. 17, 2003) on executive compensation. 6. For a discussion of the implications of differences in the principal–agency relationship in state-owned versus privately owned firms and the implications for CEO pay, see Jones and Mygind (2003).

204

DEREK C. JONES AND MARK KLINEDINST

7. While ‘‘Employers Federations’’ existed during the command era, as with other institutions that span the command and transition eras (such as trade unions), the nature and functions of such organizations in a market environment are substantially different than in the past. An important role of employers’ federations today is to collect information on CEO pay. 8. For general accounts of early transition, see Bristow (1996) and the essays in Jones and Miller (1997). 9. During this early transition era, the Bulgarian performance arguably lay somewhere in the middle range for transforming economies – poorer than the Visegrad group and the Baltic Republics but better than many other former communist countries, including Ukraine, Belarus and arguably Russia. Most noticeably, in Bulgaria declines in output and average real income are much greater and unemployment and inflation much higher than in the former group, though often better than in the latter (e.g. Blanchard, 1996; Murrell, 1996). In addition, Bulgaria has been exceptional in its growth pattern. Most transition economies did not begin to grow until inflation was under 50%. Fischer, Sahay, and Vegh (1996) Fischer et al. (1996) report that only Bulgaria and Romania began to grow before inflation dropped below 50%. 10. For accounts of reform in Bulgaria, see Bristow (1996) and Jones and Miller (1997). 11. For a review, see Stoev (2002). 12. For a review, see Rutkowski (2002). 13. In part because of the extensive changes in record keeping in a fast changing environment, we are unable to determine what has happened to 48 of the original firms. 14. The sampling design for enterprises operated at two levels. First, five groupings of the 320 municipal districts in Bulgaria were selected on the bases of geographic and urban variability, reproducing aggregate country-wide industry distributions and minimizing data collection costs (Sofia, Pernik, Pleven, Burgas and Plovdiv). Second, within each of the five regions, population enumeration lists of SOEs were compiled by the Central Statistical Bureau. The number of sampled firms per region was set to reproduce the population proportions of firms per region in 1989 (the first year for which data were gathered). The five regions contained a population of 727 SOEs. Within each region, within major industry categories, firms were ordered by size and the approximate two-thirds largest were selected up to the desired sample size of about 500. Thus the sample contains 69% of the population of firms, but selected to reproduce population distributions by region and industry. In terms of employment, the sample SOEs contain about 95% of all SOE employees in the five regions in 1989. 15. Some sample firms are labeled ‘‘de novo’’. These are firms that are new legal entities but which have arisen from the remnants of firms in the original sample that were liquidated. 16. While the data we present in this paper are for unbalanced panels, preliminary investigations using a balance panel indicate trends that are broadly similar. 17. According to our data, the typical Bulgarian firm has a unitary board structure. Only one in six firms had a two-tier board, with a supervisory board accompanying the main company board.

Corporate Governance and Executive Compensation in Bulgaria

205

18. These differences in compensation are not all accounted for by differences in size but do relate to differences in performance (see on for more discussion on this point). 19. Employment and real sales in de novo firms have also fallen considerably – by more than 33% (Table 5.) Presumably this contraction reflects many de novo firms being required to assume initial obligations as part of privatization deals that involved substantial excess capacity. 20. See, for instance, Murphy (1999) for a discussion of standard CEO compensation equations. 21. Therefore this treats the ai as a random variable. The choice of random effects over fixed effects specification is always supported by Hausman tests using a critical level of 0.10. 22. In the main, the executive compensation literature (which is overwhelmingly undertaken for studies of developed economies) does not measure performance relative to an industry mean, even though such measures often are publically available in the west. In the case of transition economies such as Estonia, the use of performance measures relative to industry means is difficult to implement since such data are not publically available and, the relatively few firms in most Estonia industries would mean that even if such a procedure were feasible it would be especially sensitive to outliers. 23. As indicated earlier, our estimates are for unbalanced panels. If the identical specifications are estimated using a balanced panel for 169 firms, the findings are essentially unaltered from those reported in the text. 24. However, in one case, lnsales is no longer statistically significant. 25. However, we again note that in some regressions sales per worker is no longer statistically significant.

ACKNOWLEDGMENTS Both authors acknowledge support from the National Council for Eurasian and East European Research and the Davidson Institute at the University of Michigan. The paper has benefited from comments by Jeffrey Pliskin and an anonymous referee.

REFERENCES Aghion, P., Blanchard, O., & Burgess, S. (1994). The behavior of state firms in Eastern Europe, pre privatization. European Economic Review, 38, 1327–1349. Bebchuk, L., & Fried, J. (2004). Pay without performance. Cambridge: Harvard University Press. Bevan, A. A., Estrin, S., & Schaffer, M. E. (1998). Determinants of enterprise performance during transition. Mimeo, Center for Economic Reform and Transformation, HeriotWatt University, Edinburgh.

206

DEREK C. JONES AND MARK KLINEDINST

Blanchard, O. (1996). Theoretical aspects of transition. American Economic Review, 86(2), 117–122. Bonin, J. (1976). On the design of managerial incentive systems in a decentralized planning environment. American Economic Review, 66(4), 682–687. Bristow, J. (1996). The Bulgarian economy in transition. Cheltenham: Elgar. Djankov, S., & Murrell, P. (2002). Enterprise restructuring in transition economies: A quantitative survey. Journal of Economic Literature, 40(September), 739–792. Earle, J. S., & Estrin, S. (1995). Alternative ownership forms: The impact on restructuring. Economics of Transition, 3(1), 111–115. Estrin, S., & Wright, M. (1999). Corporate governance in the former Soviet Union: An overview. Journal of Comparative Economics, 27(3), 398–421. European Bank for Reconstruction and Development (E.B.R.D.) (Various years). Transition Report. E.B.R.D., 1997–2004, London. Fischer, S., Sahay, R., & Vegh, C. A. (1996). Stabilization and growth in transition economies: The early experience. Journal of Economic Perspectives, 10(2), 45–66. Groves, T., Yongmiao, H., McMillan, J., & Naughton, B. (August 1995). China’s evolving managerial labor market. Journal of Political Economy, 103(4), 873–892. Ickes, B., & Samuelson, L. (1987). Job transfers and incentives in complex organizations: Thwarting the Ratchet effect. Rand Journal of Economics, 18, 275–286. Jones, D. C. (2004). Ownership and participation: A review of empirical evidence for transition economies. In: V. Pe´rotin & A. Robinson (Eds), Employee participation, firm performance and survival, advances in the economic analysis of participatory and labour managed firms (Vol. 8, pp. 171–209). Jones, D. C., & Ilayperuma Simon, K. (2005). Wage determination under plan and early transition: Bulgarian evidence using matched employer–employee data. Journal of Comparative Economics, (June), 227–243. Jones, D. C., & Kato, T. (1996). The determinants of chief executive compensation in transitional economies: Evidence form Bulgaria. Labour Economics, 3(3), 319–336. Jones, D. C., Kato, T., & Avramov, S. (1995). Managerial labour markets in transitional economies: Evidence from Bulgaria. International Journal of Manpower, 16(10), 14–24. Jones, D. C., & Klinedinst, M. (2003). Bulgarian brownfield privatization and firm performance. Mimeo, Hamilton College. Jones, D. C., Klinedinst, M., & Rock, C. (1998). Productive efficiency during transition: Evidence from Bulgarian panel data. Journal of Comparative Economics, 26, 446–464. Jones, D. C., & Miller, J. (Eds) (1997). The Bulgarian economy: Lessons from reform during early transition. Brookfield, Vermont: Ashgate. Jones, D. C., & Mygind, N. (2003). Ownership and executive compensation in Estonia. Working paper, Hamilton College. Kaplan, S. (1994). Top executive rewards and firm performance: A comparison of Japan and the U.S.. Journal of Political Economy, 102, 510–546. Klinedinst, M. (1991). Are CEOs paid their marginal products? An empirical analysis of executive compensation and corporate performance. Australian Bulletin of Labor, 17(2), 118–131. Litwack, J. (1991). Discretionary behavior and soviet economic reform. Soviet Studies, 43, 255–279. Miller, J., & Petranov, S. (2000). The first wave of mass privatization in Bulgaria: An immmediate aftermath. Economics of Transition, 8(1), 225–250.

Corporate Governance and Executive Compensation in Bulgaria

207

Murphy, K. J. (1999). Executive compensation. In: O. Ashenfelter & D. Card (Eds), Handbook of labor economics, Vol. 3. Amsterdam: North Holland. Murrell, P. (1996). How far has the transition progressed? Journal of Economic Perspectives, 10(2), 25–44. North, D. (1990). Institutions, institutional change and economic performance. Cambridge: Cambridge University Press. Nuti, M. (1997). Employeeism: Corporate governance and employee ownership in transitional economies. In: M. I. Blejer & M. Skreb (Eds), Macroeconomic stabilization in transition economies (pp. 126–154). Budapest–London–New York: Central European University Press. Pamouktchiev, H., Parvulov, S., & Petranov, S. (1997). Process of privatization in Bulgaria. In: D. C. Jones & J. Miller (Eds), The Bulgarian economy: Lessons from reform during early trasition (pp. 205–230). Brookfield, Vermont: Ashgate. Pinto, B., Belka, M., & Krajewski, S. (1993). Transforming state enterprises in Poland. Brookings Papers on Economic Activity, 1, 213–270. Pohl, G., Anderson, R. E., Claessens, S., & Djankov, S. (1997). Privatization and restructuring in central and Eastern Europe: Evidence and policy options. World Bank Technical paper 368, Washington, DC. Putterman, L. (1993). Division of labor and welfare: An introduction to economic systems. New York: Oxford University Press. Rosen, S. (1990). Contracts and the market for executives. NBER Working Paper, No. 3542. Rutkowski, J. (2002). Why is unemployment so high in Bulgaria? Mimeo, World Bank. Stoev, G. (2002). Macroeconomic developments in Bulgaria after the 1996–1997 crisis: A labor market persepctive. Mimeo, Institute for market economics, Sofia. Weitzman, M. (1976). The new soviet incentive model. Bell Journal of Economics, 7, 252–257.

APPENDIX : DEFINITIONS OF VARIABLES COBD ¼ Size of the company board MAN/COBD ¼ fraction of the board that represents managers EE/COBD ¼ fraction of the board that represents employees State ¼ 1 if the firm is still mainly state-owned, 0 otherwise Privold1 if a privatized firm is a former state-owned firm, 0 otherwise (de novo) DefCEOPAY ¼ real monthly salary of CEO (2001 leva) (excluding bonuses and nonmonetary compensation) CEO/AvwageRATIO ¼ (CEOPAY/average monthly salary of all employees) CEO/lowpay ¼ ceopay/lowest monthly salary PBC ¼ 1 if the manager has a contract that provides for PBC, 0 otherwise MANOWN ¼ fraction of the workforce that are owners and managers BIA ¼ 1 if the firm is a member of the Bulgarian Industrial Association, 0 otherwise

208

DEREK C. JONES AND MARK KLINEDINST

SIZE ¼ A measure of size: EMP ¼ employment; lnemp ¼ log of employment DefSALES ¼ real sales (thousands 2001 leva); lndefsales ¼ log of real sales PERFORMANCE ¼ A measure of enterprise performance: RSALPROD ¼ real sales per worker; lnrsalprod ¼ log real sales per worker RVADDPROD ¼ real value added per worker; lnrvaddprod ¼ log real value added per worker STATE ¼ 1 if majority owner is state, 0 is otherwise PRIVATE ¼ 1 where private owners own a majority of the voting equity and firm is formerly state-owned NEWFIRM ¼ 1 where private owners own a majority of the voting equity and firm is de novo DMANLTP1 ¼ managers have high degree of influence on long-term plans (using a 4-point Likert scale), 0 otherwise DMANLTP2 ¼ managers have moderate degree of influence on longterm plans (using a 4-point Likert scale), 0 otherwise DMANLTP3 ¼ managers have low degree of influence on long-term plans (using a 4-point Likert scale), 0 otherwise DMANLTP4 ¼ managers have no degree of influence on long-term plans (using a 4-point Likert scale), 0 otherwise DMANLRED1 ¼ managers have high degree of influence on decision to reduce employment (using a 4-point Likert scale), 0 otherwise DMANLRED2 ¼ managers have moderate degree of influence on decision to reduce employment (using a 4-point Likert scale), 0 otherwise DMANLRED3 ¼ managers have low degree of influence on decision to reduce employment (using a 4-point Likert scale), 0 otherwise DMANLRED4 ¼ managers have no degree of influence on decision to reduce employment (using a 4-point Likert scale), 0 otherwise DMANWAG1 ¼ managers have high degree of influence on wage setting (using a 4-point Likert scale), 0 otherwise DMANWAG2 ¼ managers have moderate degree of influence on wage setting (using a 4-point Likert scale), 0 otherwise DMANWAG3 ¼ managers have low degree of influence on wage setting (using a 4-point Likert scale), 0 otherwise DMANWAG4 ¼ managers have no degree of influence on wage setting (using a 4-point Likert scale), 0 otherwise DMANSHi ¼ four dummy variables for the extent of managerial influence on decisions concerning safety and health

Corporate Governance and Executive Compensation in Bulgaria

209

DSTALTP1 ¼ State has high degree of influence on long-term plans (using a 4-point Likert scale), 0 otherwise DSTALTP2 ¼ State has moderate degree of influence on long-term plans (using a 4-point Likert scale), 0 otherwise DSTALTP3 ¼ State has low degree of influence on long-term plans (using a 4-point Likert scale), 0 otherwise DSTALTP4 ¼ State has no degree of influence on LTP (using a 4-point Likert scale), 0 otherwise DSTALREDi ¼ four dummy variables for the extent of state influence on the decision to reduce employment DEESHi ¼ four dummy variables for the extent of employee influence on decisions concerning safety and health

This page intentionally left blank

210

THEME III: CO-OPERATIVES AND SOCIAL ENTERPRISES

211

This page intentionally left blank

212

WORKERS COOPERATIVES AS AN ALTERNATIVE COMPETITIVE ORGANIZATIONAL FORM Morris Altman ABSTRACT Developing an alternative and more realistic modeling of the firm, the key point of this paper is that workers cooperatives represent a form of corporate governance, which is a subset of the participatory organizational form, that constitutes a competitive alternative to the typical relatively hierarchical and narrowly controlled firms. An important component of the cooperative advantage lies in its capacity to increase the quantity and quality of effort inputs into the ‘production process.’ However, to do so incurs economic costs. Thus, cooperatives can yield competitive outcomes without driving out of the market non-cooperative organizational forms. To some extent, whether cooperative or other participatory solutions are adopted depends upon the preferences of economic agents since cooperatives are shown to be competitive even in an extremely competitive environment. However, dominant or not, the cooperative solution can yield higher social–economic welfare levels to members.

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 213–235 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09007-1

213

214

MORRIS ALTMAN

1. INTRODUCTION The fact that the workers’ cooperative is a far from dominant organizational form has continuously served to fuel the debate as to capacity of cooperatives to survive in competitive markets. However, it is well documented that, although workers’ cooperatives are far from dominant, those that are established have tended to be competitive, relatively profitable and productive as compared to their ‘narrowly’ and non-cooperatively owned and operated counterparts (for example, Bartlett, Cable, Estrin, Jones, & Smith, 1992); Bonin, Jones, & Putterman, 1993; Cosgel & Murray, 1998; Craig & Pencavel, 1992; Craig & Pencavel, 1995; Dow, 2003; Gulati, Isaac, & Klein, 2002; Johnson & Whyte, 1977; Jones, 1985; Jones & Svejnar, 1985; Merrett & Walzer, 2004; Park, Kruse, & Sesil, 2004). Still, the question remains, flowing from conventional (neoclassical) theory, if worker cooperatives are economically sustainable and viable why are they not dominant? Is not their lack of dominance and their overall marginal role in the economy ‘proof’ of the economic failure of cooperatives wherein some cooperatives survive for extra-economic or exceptional reasons? (Putterman, 1984). Are the empirics presented in support of the relative efficiency and competitiveness of worker cooperatives, therefore, fundamentally flawed and misleading, being without a theoretical foundation? A key point articulated here is that the workers cooperative represents a form of corporate governance that constitutes a competitive alternative to the conventional privately controlled firms. But this does not imply competitive superiority. Therefore, there exists neither a cooperative or noncooperative imperative, even in a competitive market economy. Moreover, that the worker cooperative does not dominate is not a product of the relative inefficiency of the cooperative form of governance. Thus, one must seek other reasons for the fact that cooperatives are typically not major players in the realm of production in market economies. In the modeling presented below, consistent with empirical evidence, the workers cooperative represents only one ‘extreme’ alternative to hierarchical organizational forms. Privately owned participatory firms (which allows for some workers’ ownership of firm assets) represent another alternative, one where workers need not risk their capital nor bear the risks entailed in ownership. Also, they need not invest the time and effort, which might be required at the management and corporate decision-making level in a worker cooperative. The workers cooperative can be categorized as a subset of a spectrum of types of the participatory firm all of which can be modeled as relatively efficient cost-competitive alternatives to the conventional hierarchical firm.

Workers Cooperatives as a Competitive Alternative

215

That the workers cooperative can be a viable alterative to the traditional hierarchical organizational form as can the more generalized participatory firm, which need not be worker owned, is assumed away in the conventional economic modeling of the firm because of some of the fundamental and largely unrealistic behavioral and institutional assumptions of the model. For this reason, alternative assumptions are introduced here, amending the conventional model of the firm, to help explain the evident actual and potential viability of the workers cooperative. This revision, based on Altman (1992, 1996, 1998, 1999, 2000, 2001, 2002), supplements the literature which suggests that the workers cooperative need not pose a competitive disadvantage.1 It is argued in this paper that the workers cooperative or the relatively non-hierarchically managed firm can serve both as a competitive alternative to the traditional firm and generate superior levels of material welfare to its workforce. The cooperative advantage lies in the capacity of participatory firms to increase the quantity and quality of effort inputs into the ‘production process,’ thereby producing higher levels of output and a superior quality of output. However, to do so incurs economic costs. Thus, the workers cooperative can yield competitive outcomes without driving noncooperative or non-participatory organizational forms out of the market. Overall, the cooperative solution can produce higher social-economic welfare levels to members while also overcoming significant market failures. Market forces, no matter how powerful, do not require non-cooperative solutions to socio-economic problems. Market forces, do not generate such an imperative. Competitive markets and cooperative organizational forms are all quite consistent. Given that competitive markets do not yield a noncooperative imperative, one need explain why cooperatives are not dominant even if they can be competitive. I argue that components of such explanations include institutional constraints, preferences of workers, managers, and owners, bargaining power considerations, as well as participatory alternatives to the workers cooperative.

2. WHAT IS A COOPERATIVE? Prior to addressing the issue of the economic viability of the workers cooperative and their economic and welfare consequences it is important to note what a cooperative is. A fundamental distinguishing organizational characteristic of a cooperative is that it is owned and controlled by the people who produce the organization’s output or use its products, supplies, or services. Thus, cooperatives are designed, in theory, to meet the both the

216

MORRIS ALTMAN

current preferences of the cooperatives’ members and to adjust to meet their changing preferences. In this sense cooperatives are constructed to better match the needs of a relatively much broader base of individuals than alternative and more narrowly owned organizations. For example, a privately owned hierarchical firm need not meet the preferences of its workers as compared to a workers cooperative. The preferences of members of the firm hierarchy tend to dominate in the non-participatory firm. Different types of cooperatives, structured to meet the preferences of particular groups of individuals include, workers cooperatives (worker controlled firms), agricultural cooperatives (producer cooperatives (PCs) as well as cooperatives which focus upon the joint marketing, bargaining, processing, and purchasing supplies and services), business cooperatives (where small enterprises engage in joint purchasing and marketing), retail cooperatives (member-controlled organization that sells consumer goods to members and non-members), and credit unions (member-controlled non-profit organization designed to provide financial services to its members). Any organization which refers to itself as a cooperative but lacks the structure and practice of democratic governance by members is not a cooperative. The modeling presented in this paper only refers to organizations, which actually fit the definition of a cooperative since only such an organization can lend itself to the predicted welfare improving and competitive effects elaborated upon below. Thus, a self-identified cooperative which is uncompetitive and not welfare improving but also is not, ‘‘owned and controlled by the people who produce the organization’s output or use its products, supplies or services,’’ does not constitute a case in point refutation or challenge to the theory of cooperative presented below (Centre for the Study of Cooperatives, 2002).

3. A BEHAVIORAL MODEL OF THE FIRM, WORKER COOPERATIVES, AND THE PARTICIPATORY FIRM A fundamental assumption of the conventional model of the firm is that economic agents are rational utility maximizers wherein profit maximizing is assumed to be a critical argument of each agent’s utility function. For simplicity and to be consistent with the neoclassical frame of reference which we are critiquing, we assume behavior which is maximizing, consistent, at least somewhat forward-looking, and rational in the sense of being intelligent and purposeful (Becker, 1996, p. 22). Thus, we assume that in the workers cooperative all agents are rational utility maximizers. But the conventional

Workers Cooperatives as a Competitive Alternative

217

model assumes that rationality implies not only that agents are utility maximizers dominated by a profit maximization argument in their objective function, but also effort maximizers in both its quantity and quality dimension irrespective of the competitive state of the market or of the industrial or human relations configurations of the firm (Altman, 1999; Friedman, 1953; Reder, 1982). All agents thus have the same goal or objective function with regards to effort and profit maximization and they are assumed to do the best they can from the perspective of the firm (profit maximization given maximum effort inputs) from the get go. A behavioral model of the firm, which I have developed in some details elsewhere (for example, Altman, 1992, 2001, 2002) introduces the assumption that wages and working conditions affect effort levels and technical change, an argument buttressed, for example, in the industrial relation and related literature (Altman, 2002; Ichniowski, Kochan, Levine, Olson, & Strauss, 1996). Thus effort is introduced into the short-run production function as a variable. This modeling of the firm extends Leibenstein’s (1966, 1979) x-efficiency theory. The possibility of x-inefficiency being of some importance to an understanding the workers cooperative is raised by McCain (1973). X-efficiency refers to a scenario where effort is maximized so that firms are operating along its production possibility frontier. X-inefficiency obtains when effort inputs, in either its quantity or quality dimension, are not maximized. The standard assumption in microeconomics is that effort is invariant to changes in wages and working conditions. Therefore, all members of the firm are assumed to be maximizing utility only when their effort inputs are maximized. More recently, in the efficiency wage literature, effort variability is introduced to provide an explanation for rational downward nominal wage inflexibility (for example, Akerlof & Yellen, 1986). A simplified, conventional production function is given by Q ¼ L; K; T; E (1) where Q is the output, L the labor (adjusted for quality and human capital), and K the capital stock inclusive of land. In the conventional model, in the short run, (T) technology is held constant. Moreover, (E¯) the quantity and quality of effort per unit of labor time is assumed to be a constant in both the short- and long-run at some utility maximizing and cost minimizing level. Short-run output is simply a function of the basic factor inputs, labor, and capital. Flowing from the basic production function, the essence of the conventional model, from the perspective of many critical public policy issues, is

218

MORRIS ALTMAN

given by Eq. (2), where one assumes for simplicity one factor input (labor). AC ¼

wL w ¼ Q Q=L

(2)

Marginal or average costs of production

AC is the average cost, w the wage rate, L the labor inputs, and Q the output. If effort is fixed at some maximum level, any increase in the wage or related benefits to labor results in higher average costs. Moreover, since effort is assumed to be at a maximum, organizational culture or human relations systems are assumed to be optimal, otherwise improvements in organizational culture would generate higher levels and quality of effort inputs and thus lower production costs. Ceteris paribus, with lower costs firms can either charge lower prices or earn greater profits. If effort is not maximized, one would have to explain how firms with non-maximizing agents could survive on the market and why rational agents would sacrifice all this cash lying on the sidewalk; cash easily realizable were effort is increased and maintained at optimal levels. Some of the implications of the conventional model can be illustrated in Diagram 1. Line T illustrates the conventional rendering of the relationship between labor costs and average and marginal production costs. By definition any increase in such costs yields higher average and marginal costs as

T CO*

G0

CO0

W0

Diagram 1.

W* Wage rate

W1

Wages and Efficiency.

CO

Workers Cooperatives as a Competitive Alternative

219

per Eq. (1). In a world where labor is the only cost of production, there is a one-to-one relationship between the percentage increase in labor costs and the percentage increase in unit production costs. In reality, however, this extreme relationship does not hold since the extent of the unit cost increase generated by a percentage increase in labor costs critically depends upon the share of labor costs in total costs. The percentage increase in unit costs which flow from a given percentage increase in labor costs diminishes with increases in the percentage share in total costs by non-labor inputs. This point is expressed as dAC dW W L ¼ (3) AC W W L þ NLC AC is the average costs, dQ the change in output, dW the change in the wage rate, and NLC the non-labor costs. Thus, if labor costs comprise only 10 percent of total costs and labor costs increase by 20 percent, unit production costs will increase by 2 percent. If labor costs comprise 100 percent of total costs the same 20 percent increase in labor costs yields a 20 percent increase in unit costs. Nevertheless, in the conventional model, labor cost increases yield increases in unit costs; by how much depends on the share of labor costs in total costs. The argument presented in this paper hinges on the assumption that where conflict within the firm predominates among economic agents; where information asymmetries exist as well as monitoring and enforcements costs with regards to the quantity and quality of efforts inputs; and where economic agents’ objective function include non-pecuniary arguments; x-inefficiency will exist, particularly when pressure is slack. Under these assumptions, higher wage rates, and improved working conditions and industrial relations can serve to reduce the level x-inefficiency. Thus, participatory firms, of which cooperatives are a subset, foster higher rates of productivity when effort is a choice variable in the production function. In this scenario, the conventional firm can be expected to be relatively xinefficient and therefore characterized by lower levels of productivity. Underlying the behavioral model presented here is the assumption that firms are not necessarily producing efficiently (Altman, 1992, 2001, 2002; see also Simon, 1991). In this case, effort need not be maximized per unit of input and the firm is operating inside of the production possibility frontier and thus x-inefficiently in production. Since the quantity and quality of effort are discretionary variables labor productivity is not simply a function of capital per worker or technological change. Once effort is discretionary, efficient production requires certain conditions be met internal to the firm

220

MORRIS ALTMAN

that are independent of external market forces. Specifically, indirect incentives must be developed to ensure that the quantity and quality of effort inputs are maximized – utility maximization takes place at x-efficient levels of effort inputs only under particular configurations of in-firm incentives. Moreover, market forces, no matter how competitive, need not impose such conditions upon the firm. There is a growing literature which suggests that only in more trusting cooperative-based forms of industrial organization is effort per unit of labor input maximized (for example, Altman, 2001, 2002; Gordon, 1996, 1998; Ichniowski et al., 1996; Tomer, 2001). Under such a system of industrial relations, labor has a direct and meaningful input into the decision-making process, transparency and accountability are built into the decision-making process of the firm allowing for the fostering of reciprocity between firms members within and between different levels of the firm’s decision-making and production structure, labor compensation is related in part to productivity, there is relative job security, and there is a relatively less hierarchical administrative structure. In this case, unlike what is predicted in the conventional model, labor productivity can be expected to be below the maximum unless the firm adopts a more cooperative system of industrial relations generating the incentives for utility maximizing agents to worker smarter and harder. Labor compensation represents only one dimension of such incentives. The same literature that argues for a positive relationship between productivity and work culture also posits that establishing a relatively cooperative work environment is costly. To become more productive or x-efficient requires the firm investing in what Tomer (1987) refers to as organizational capital. This includes, improved screening and training of employees and managers and redesigning and reconfiguring the shop floor for a more participatory work environment. Members of the firm hierarchy would also have to work harder and smarter. In addition to organizational capital the firm would also have to invest in higher wages and/or improved working conditions. In the behavioral model of the firm, therefore, there are two intimately related dimensions to becoming more x-efficient. One dimension speaks to increasing productivity whereas the other speaks to increasing costs. One should also note that when effort inputs are referred to, one is not simply speaking to economic agents working harder. A critically important dimension of increasing effort inputs is working smarter yielding increases in productivity. The worker controlled firm goes a few steps beyond the participatory firms, since here workers own the firm and incentives are typically more

Workers Cooperatives as a Competitive Alternative

221

oriented to group as opposed to individual productivity. Just as in the more generalized participatory firms, some form of profit-sharing in the context of effective participation in the decision-making process, are critical to enhancing productivity Bonin et al (1993, p. 1316). It is also true that within a cooperative there is no conflict between workers and management per se in so far that the firm’s objective function is determined by all firm members. There is more of an incentive for firm members to cooperate toward increasing the quality and quantity of effort inputs toward some maximum. Moreover, mutual monitoring and peer pressure toward the realization of the collective end is of importance. This also reduces monitoring and enforcement costs. Some of this would characterize privately owned participatory firms as well, but not necessarily to the same extent. Although worker cooperatives tend to be more egalitarian than participatory firms (an Israeli Kibbutz would be a good example), group incentives and related behavioral interdependence exist to enhance productivity as opposed to the more individualized incentives of privately owned participatory firms. The extent of group-based incentive structure and a flatter hierarchical structure also tends to be greater in the worker cooperative than in the privately owned participatory firm (Abramitzky, 2004; Bonin et al., 1993; Craig & Pencavel, 1995; Guttman & Schnytzer, 1989; Putterman, 1983). Bonus pay or profit sharing, active and substantive worker participation in micro and macro aspects of decision-making and related to this, transparency in the decision-making process, and ownership (especially of course in worker cooperatives), serve to at a minimum neutralize moral hazard problems expected to arise in more egalitarian organizational forms. As with the more generalized participatory firm, labor compensation represents only one dimension of the incentives required to make the firm more productive. It possible for workers to trade-off income for other valued benefits in the course of deciding to increase effort inputs. Thus effort related productivity increases can take place in the absence of increases in labor compensation. However, the evidence suggests a strong positive empirical relationship between productivity and labor compensation especially when contextualized in an environment where workers have effective voice in the firm’s decisionmaking process. Finally, although the workers cooperative is an organizational form, which positively affects productivity, it is not a necessary condition for such increases. It is not the asset ownership by workers per se which generates productivity increases, it is rather the participatory organizational form (Gordon, 1998).2 In the original Leibenstein formulation of x-efficiency theory, any slack in effort input, ceteris paribus, results in higher unit production costs as per

222

MORRIS ALTMAN

Eq. (2). X-inefficiency persists in this scenario as a consequence of firms being afforded protection through monopolistic markets, tariffs, subsidies, and the like. Remove these barriers and x-inefficiency in production is likely to be shocked out of the system. In the formulation of x-efficiency theory presented here such protection is not a necessary condition for the existence of x-inefficiency nor is perfect product market competition or highly competitive products markets a sufficient condition to eliminate x-efficiency and to somehow maximize effort inputs in both its quantity and quality dimension. Given that the conventional wisdom typically assumes competitive pressures is in the long-run sufficient to generate rational expectations competitive results in terms of market clearing and price formation; that these pressures do not suffice to generate efficiency in production is of importance to any discourse on potential x-inefficiency or productivity slack in the system. X-inefficient firms remain cost competitive in an unprotected competitive environment if low levels of x-efficiency are related to low rates of labor compensation and high wage firms become competitive by becoming relatively x-efficient. Related to this, low-wage firms cannot be expected to drive out of the market high-wage firms when low-wage rates are causally related to low levels of x-efficiency and high wages are causally related to relatively high levels of x-efficiency. Low-wage competitors need not drive economies toward a low wage road of competitiveness. High-wage and lowwage firms and x-efficient and x-inefficient firms can survive simultaneously in competitive equilibrium if they produce at the same level of average cost – if they operate at offsetting differential levels of x-efficiency. In terms of Eq. (2), as long as any change in (w) is offset by a corresponding change in (Q/ L), which is affected by variation in effort inputs, average costs do not change as production costs and levels of x-inefficiency vary. Moreover, lower levels of x-efficiency need not result in higher unit production costs and higher levels of x-efficiency need not yield lower unit production costs. It would be possible for there to exist in competitive equilibrium an array of firms characterized by different wage rates, related production costs, and levels of working conditions, producing at an identical unit cost, if these differentials in costs are compensated for by differentials in the level of x-efficiency. In Diagram 1, an array of wage costs, a proxy of various production costs, up to W1, is associated with a unique unit cost of production, CO0. Given technology, once one goes beyond W1, effort increases hit the wall of diminishing returns, increasing at a rate that no longer suffices to compensate for increases in production costs (w). Thus, some linearity is assumed

Workers Cooperatives as a Competitive Alternative

223

between production costs (w), effort inputs and productivity, up to W1, when diminishing return set in. Unlike in the conventional wisdom, given by curve T, there is no unique relationship between w and unit production costs (CO). In the behavioral model there is only a fuzzy set relationship given by the multiple equilibria of production costs relative to unit production costs, which are sustainable even in highly competitive product markets. Moreover, there exists a fuzzy set relationship between production costs and systems of industrial relations where the more participatory firm is related to higher rates of labor compensation and the more hierarchical firm to lower rates of labor compensation. Finally, in this particular modeling of the firm it is possible and likely for there to be various combinations of routines or practices for any particular type of firm (participatory versus hierarchical, for example), which yields the same level of average costs. Critical to the argument presented here is that even in the most competitive of product market environments there can exists a multiple equilibria of labor costs and forms of firm governance consistent competitive production costs. Moreover, in this behavioral scenario, there is no incentive for firm profit maximizing managers or owners (the decision makers of the typical firm) to make their firms relatively x-efficient in that increasing efficiency has no impact on unit costs, profits, or managerial-owner income. They can be expected to be indifferent between x-efficient and x-inefficient forms of production from the perspective of unit costs since there is no advantage to be gained from producing x-efficiently unless owners and managers gain utility from improving firm efficiency and the material and social well-being of their employees. Market forces, per se, do not suffice to guarantee x-efficiency in production. Bargaining power within the firm can play a significant role in determining the extent to which the firm is more or less participatory and more or less x-efficient in production. The behavior model of the firm presented here can be extended to incorporate technological change, a detailed discussion of which is presented in Altman (1996, Chap. 6, 1998, 2001). In this case, technological change can be induced by improvements to labor’s pecuniary and non-pecuniary benefits. Technological change is endogenous as opposed to exogenous as it is the conventional modeling and it is adopted, in the first instance, to prevent unit costs from increasing, illustrated in Diagram 1 by a shift in the cost curve from CO to COT. In effect, there is an induced inward shift in the production isoquant as a consequence of the increased cost of a factor input. Induced technological change yields higher productivity but not necessarily lower unit production costs to affected firms as compared to firms where the conditions for induced technical changed are not present. Thus firms need

224

MORRIS ALTMAN

not necessarily adopt the most productive technologies – much depends upon variables, which induce the adoption and development of new technologies. This modeling of technical change provides another degree of freedom to firms affected by wage increases or by increases in other production costs and predicts that, ceteris paribus, the higher wage firms, for example, would be more technologically intensive than the relatively lowwage firms. Thus in workers cooperatives, as in any other type of firm, to allow for desired increases in benefits, above what is possible by increasing the quantity and quality of effort inputs, while remaining competitive, requires the development and adoption of more advanced technology.

4. SOME IMPLICATIONS OF THE BEHAVIORAL MODEL Once effort discretion is introduced as a choice variable into the modeling of the firm, the empirical findings suggesting that worker cooperatives are no less and can even be more productive than hierarchical firms makes much sense and should not be paradoxical even from the perspective of economic modeling characterized by rational utility maximizing economic agents in relatively competitive product markets. The incentives contained within the cooperative organizational form appear to be more conducive to higher productivity in terms of higher levels of x-efficiency and with regards to induced technological change. Ceteris paribus, one would expect that the worker cooperative should be more productive than the traditional hierarchical firm. But it is not at all clear that the workers cooperative can be expected to be more productive than the participatory, but privately owned firm. However, given that cooperatives might be more expensive to operate, especially with regard to pecuniary and non-pecuniary benefits to workers, the cooperative organizational form need not have any advantage over the privately owned hierarchical firm – the same holds true with regards to the privately owned participatory firm. On the other hand, one should not expect that the hierarchical firm should dominate the cooperative firm from a unit cost-competitive perspective. Economic theory, per se, once one accounts for effort variability as a choice variable, cannot predict either the dominance or demise of cooperative firms within the context of competitive markets. One can have an array of firms all operating at the same unit costs, which represent a variety of organizational forms, with some firms being relatively more x-efficient and technological advanced than others. But the

Workers Cooperatives as a Competitive Alternative

225

composition of organizational types cannot be determined by market forces alone since both the conventional hierarchical firm, the privately owned participatory firm, and the workers cooperative can be equally cost competitive while being characterized differential levels of x-efficiencies and technologies being employed. Therefore, one cannot explain the relatively small role played by either privately owned participatory firm or workers cooperative by standard efficiency reasoning. With respect to production costs such firms can do as well as the conventional hierarchical firm, while being even more efficient. That effort can be a variable input into the production function affected by the incentive structure of the firm allows for the possibility of workers cooperatives being competitive with non-participatory firms even when the cooperative employs old vintage capital stock. This helps explain how seemingly technological backward cooperatives can survive in a competitive market. If the objective function of workers is dominated by the survival of the firm and the worker–owners’ share in income is based on the relative prosperity of the firm, workers in the cooperative have the incentive to work smarter and harder than they would in the conventional hierarchical firm. If the two types of firms were x-efficient, the firm using the newest vintage capital stock would be more productive. However, if the newest vintage firm is hierarchical in structure and is, therefore, x-inefficient, while the ‘backward’ cooperative is x-efficient, it is possible for the latter to be cost competitive with the former, as long as the newest vintage capital does not yield a productivity advantage which cannot be compensated for by more effort inputs in the workers cooperative. This scenario is illustrated in Diagram 2 where all isoquants yield the same level of output. Therefore, ceteris paribus, the firm producing along the isoquant closest to the original is producing at the lowest unit cost whereas the firm producing along the isoquant furthest removed from the origin is producing at the highest unit cost. Isoquant QAXE is the x-efficient production isoquant of the firm using the newest vintage technology. QBXI is the x-inefficient isoquant of the firm using the relatively old vintage technology. If the newest vintage technology firm is x-inefficient due to its hierarchical organizational form it could still produce output at a lower unit cost, along isoquant QBXE1 for example, than the backward technology firm producing x-inefficiently. If the latter firm is transformed into a workers cooperative and effort inputs are increased as a consequence so as to keep the firm in business and this shifts this firm’s isoquant to QBXE1, the cooperative would be cost competitive with the x-inefficient newest vintage

226

MORRIS ALTMAN

1

Capital

C

2 B

QBXI QBXE1 QBXE2 QAXE 0

A Labor

Diagram 2.

X-Efficiency and the Technological Backwardness.

technology hierarchical firm. If the latter becomes more x-efficient shifting its isoquant inward, perhaps as a result of this firm transforming its organizational form into a more participatory one, the cooperative can compete on the basis of lowering labor income as long as effort inputs are not affected – workers agree to maintain effort inputs say as to keep their technological backward firm in business. In this case, the relative price of labor is reduced from AC to AB, shifting the firm’s isoquant to QBXE2. As the newly formed high-tech participatory firm becomes increasing xefficient, the low-tech workers cooperative survives only by further cutting wages. The alternative to surviving on the basis of low wages is for the cooperative to adopt the newest vintage technology. The implications of the above behavioral model for an understanding of the cooperative firms can be further assessed through Diagram 3. In the

227

Workers Cooperatives as a Competitive Alternative

C

A

H

D W1

B W0

G

Diagram 3.

F

Employment and X-Efficiency.

conventional model, any increase labor benefits cause the marginal cost curve (the wage rate inclusive of all costed benefits) to the firm to increase. This results in firm employment diminishing. If all firms experience the same marginal cost increases, ceteris paribus, employment declines throughout the economy. In Diagram 3, an increase in wages from W0 to W1 yields a fall in employment from F to G, given by the firm’s marginal revenue product curve. Thus, workers cooperatives, which value employment, should not rationally engage in any policy which would increase labor costs. However, if an increase in labor costs has an x-efficiency effect in production, this shifts the marginal revenue product curve outward, offsetting the negative employment effect which increasing labor costs would otherwise have. It is quite possible that the marginal revenue product curve shifts from AB to CD, thereby entirely offsetting the potential negative supply side employment effect. The latter is a plausible occurrence given the evidence suggesting only

228

MORRIS ALTMAN

marginal if any negative employment effects flowing from increases in minimum wages or the introduction of unions – increasing labor costs have not had the predicted negative effect upon employment (Card & Krueger, 1995; Freeman & Medoff, 1984). As long as the demand for output is sufficient to absorb the increased output, employment need not diminish as labor costs increase if there is a sufficient x-efficiency effect induced by such increases.

5. WHY WORKERS COOPERATIVES DON’T DOMINATE THE MARKET PLACE In a comprehensive survey of the literature on producer cooperatives (PCs), Bonin et al. (1993, p. 1316) concludes that: ‘‘The weight of the theoretical reasoning and evidence surveyed convinces us that the explanation of the relative scarcity of PCs lies in the nexus between decision-making and financial support. Worker control requires (at least partial) worker ownership for incentive reasons but the latter conflicts with the worker’s desire to hold a relatively low-risk, diversified portfolio. External financiers with no direct control of company governance will not commit significant funds without receiving a substantial premium to reflect the risk involved. Hence, workercontrolled PCs have difficulty finding internal sources and competing with CFs (conventional firms) for investment funds.’’ This argument is consistent with the theoretical arguments put forth in this paper, wherein there is no reason to expect worker cooperatives not to be competitive, yielding in addition high levels of material welfare plus additional non-pecuniary benefits. In terms of competitiveness, the behavioral model presented here suggests that, ceteris paribus, worker cooperatives could produce all output. The above argument suggests that the workers cooperative would be more dominant if financial resources were more readily available and if workers’ would be less risk adverse, although, the relative importance of these two factors is not specified.3 Moreover, cooperatives are often structured such that individual merit plays a marginal role in the determination of individual pecuniary and non-pecuniary benefits. Such egalitarianism might not appeal to a large segment of the workforce. What is typically ignored in the literature is the role which information imperfections play with regards to what worker cooperatives entail, in determining the demand for workers cooperatives. An important related empirical question is whether with such information imperfections corrected, would workers’ preferences change favorably toward workers cooperatives? In addition, the privately owned

Workers Cooperatives as a Competitive Alternative

229

participatory firm provides an alternative to workers cooperatives for workers with preferences for effective participation. Such a firm is not characterized by the risk, time-investment and, extreme egalitarianism, which might reduce the demand for cooperative organizational forms by workers themselves. Workers can secure significant pecuniary and nonpecuniary benefits here without incurring the costs entailed by becoming a member of a workers cooperative. Thus constraints to the formation of workers cooperatives include, financial-institutional, informational, preferences of workers, and the existence of preferred participatory alternatives to the workers cooperative. Given workers’ preferences, even if workers cooperatives were easy to establish and workers and prospective managers were fully informative about the nature of such an organizational form, it is in no ways clear that the market would soon be dominated by such firms. However, neither workers cooperatives nor privately owned participatory firms dominate the market place even though theory and evidence point to their capacity to meet the competitive challenge. It is also true that a large proportion of workers are very supportive of more participatory organizational forms (Chuma, Kato, & Ohashi, 2004; Freeman & Rogers, 1999). In the case of the privately owned participatory firm, firm owners and managers, often oppose this organizational form which would reduce the relative power, prestige, and pay of members of the firm hierarchy. Thus, one reason for the failure of this organizational form to dominate the competitive market place is the dominance of contrary preferences of firm owners and managers given the bargaining power of workers (Altman, 2002). Thus workers’ preferences are being frustrated. But preferences for effective participation are not the same thing as preferences for worker cooperatives. This argument is illustrated in Diagram 4. For any particular industry, total output is given at F. Assuming that the opportunity cost of production is identical across all firms at A, all output can be produced by either worker cooperatives, privately owned participatory firms, and privately owned CFs. The evidence and theory suggests that workers cooperatives would never dominate the market even if the demand for them could be met, given by DC at C. The output supplied by worker cooperatives could be at J, below what could be produced by cooperatives if workers preferences were realized (C). The cooperative gap, JC, is given here by the structure of the financial market. But a critical factor with regards to the relative unimportance of workers cooperative to total output is workers’ preferences. The demand for participatory firms is much greater than for workers cooperatives and is given by DP at D. The output produced by such firms is less than D and is given by K. The participatory firm gap, KD, is determined by the preferences

MORRIS ALTMAN

Cost

230

Dc

I

B

H

A DP

J KC

G

D

F

Output

Diagram 4.

Demand and Supply of Worker Cooperatives.

of members of the firm hierarchy opposing the reconfiguration of privately owned firms to a more participatory organizational form linked with bonuses and profit sharing. Increasing the importance of the participatory firms, given the preferences of workers, is most clearly related to providing workers with the capacity to realize their preferences in the bargaining process which would serve to increase the participatory nature of privately owned firms. Improving the financial market relative to workers cooperative could also have some effect; however, this effect is limited by the preferences of workers for participatory alternatives to workers cooperatives. More research is required to determine worker preferences for cooperatives relative to other participatory organizational forms and the extent to which information stocks and flows on different organizational forms have upon the formation of their preferences. Given that participatory firms tend to be more productive than CFs and that such firms do not play the role in the market place which they could given their relative competitiveness, per capita output is less than it could otherwise be. Moreover, given that participatory firms tend to be relatively more rewarding to workers, their level of material welfare is less than it could otherwise be. And given the expressed preferences of a large percentage of

Workers Cooperatives as a Competitive Alternative

231

workers for participatory organizational forms, the utility of workers is also lower than it need be, given that their preferences are being frustrated.

6. CONCLUSION When effort variability is introduced into the modeling of the firm, one can better explain the cost competitiveness of participatory firms, inclusive of workers cooperatives. The incentive structure within the participatory firms tends to make them more productive in terms of x-efficiency and technological change. However, such a modeling of the firm also contributes to explaining why such competitive organizational forms need not dominate. Participatory firms tend to be more costly to operate. Thus, to the extent that productivity gains match increased costs; the participatory firm has no competitive advantage over the conventional hierarchical firms. But neither would the conventional firm necessarily have an advantage over the participatory firm. Supply side economic theory, per se, cannot explain why participatory remains a relatively small player in the market. Indeed, from the perspective of competitiveness, one would expect that participatory firms to be relatively more important than they are. From the perspective of the behavioral model of the firm presented here, neither the workers cooperative nor the privately owned participatory firm is at competitive disadvantage with the conventional hierarchical firm. Moreover, the workers cooperative contains the incentives, which allow for technologically backward firms to be cost competitive against relatively high-tech hierarchical firms in terms of differential levels of x-inefficiency. Thus, a technologically backward firm that would face bankruptcy as an x-inefficient hierarchical firm could remain cost competitive as a cooperative. The relatively small role played by participatory firms appears to be closely related to preferences and financial markets, and perhaps also to information stocks and flows. Workers preferences for workers cooperatives are not dominant. Unless the nature of the workers cooperative is modified to better match the preferences of more workers, while maintaining workers’ control over the firm, there is little reason to expect that this variant of the participatory firm will become dominant in spite of its capacity to compete in a competitive market place (for example, see Zeuli, 2004). However, ceteris paribus, workers cooperatives would be of greater importance if financial markets better facilitated the development of cooperatives (Bowles & Gintis, 1994) and if workers were better informed about the nature of the workers cooperative. The relative unimportance of privately owned participatory

232

MORRIS ALTMAN

firms, however, is most closely related the dominance of the preferences of members of the firm hierarchy over that of workers. The growth of this variant of the participatory firm can be affected by institutional mechanisms which provide workers with more effective means to give voice to their preference for participatory organizational forms. The evidence suggests that this would have a significant positive impact upon economic efficiency and upon the material welfare of affected workers. To the extent that workers preferences for more participatory organizational forms are not met, theory and evidence suggests that society is worse off materially as would be prospective workers given the incentive structures of participatory firms. From this perspective the paucity of participatory firms is sub-optimal with respect to the capacity of society to maximize its material welfare and increasing the relative importance of such firms would be welfare enhancing, given the preferences of workers. Workers cooperatives and privately owned participatory firms can serve to improve economic efficiency and overall productivity within the context of a competitive market and increase the material well-being of workers and can achieve the latter by increasing the size of the economic pie as opposed to income redistribution. The relative unimportance of such organizational forms cannot be conveniently explained away in theory in terms of their competitive disadvantage. Instead, institutions, bargaining power, and preferences play a determining in the distribution of organizational forms in market economies.

NOTES 1. See Ward (1958) for the classic paper exploring the allocative efficiency effects workers cooperatives. 2. Gordon (1998, p. 200) argues, based on comparative international evidence: ‘‘It is unlikely that the operations or comparative advantages of the production systems of the cooperative (participatory organizational form) economies depend fundamentally on an egalitarian distribution of the means of productiony They are all capitalist economies and the ownership of productive assets is held fairly narrowly.’’ Rather, Gordon finds that the relative success of participatory economies is most strongly related to direct wage incentives to production workers tied to productivity increases, the provision of substantial employment security to workers, and effective voice to workers. And these factors are what most workers tend to value most; not the ownership of assets. Thus a radical redistribution in the ownership of assets is not necessary to realize many of the desired attributes of a participatory economy. 3. Dow (2003) makes the case that workers cooperatives can be designed so that the risk of being an owner are mitigated. Bowles and Gintis (1994) discuss the role capital markets in constraining the growth of workers cooperatives.

Workers Cooperatives as a Competitive Alternative

233

ACKNOWLEDGMENT This paper was presented at the conference, Mapping Co-operative Studies in the New Millennium, University of Victoria, Victoria, British Columbia, Canada, May 28–31, 2003. It is also a contribution to the Social Science and Humanities Research Strategic Themes program, ‘‘Co-operative Membership and Globalization: Creating Social Cohesion through Market Relations.’’ The author thanks Louise Lamontagne and an anonymous referee for their insightful comments.

REFERENCES Abramitzky, R. (2004). The limits of equality: An economic analysis of the Israeli Kibbutz. Working Paper, Northwestern University. Chicago. Akerlof, G. A., & Yellen, Y. L. (Eds) (1986). Efficiency wage models of the labor market. Cambridge, England: Cambridge University Press. Altman, M. (1992). The economics of exogenous increases in wage rates in a behavioral/xefficiency model of the firm. Review of Social Economy, 50, 163–192. Altman, M. (1996). Human agency and material welfare: Revisions in microeconomics and their implications for public policy. Boston: Kluwer Academic Publishers. Altman, M. (1998). High path to economic growth and development. Challenge: The Magazine of Economic Affairs, 41, 91–104. Altman, M. (1999). The methodology of economics and the survivor principle revisited and revised: Some welfare and public policy implications of modeling the economic agent. Review of Social Economics, 57(4), 427–449. Altman, M. (2000). Labor rights and labor power and welfare maximization in a market economy: Revising the conventional wisdom. International Journal of Social Economics, 27, 1252–1269. Altman, M. (2001). Worker satisfaction and economic performance. Armonk, NY: M.E. Sharpe Publishers. Altman, M. (2002). Economic theory, public policy and the challenge of innovative work practices. Economic and Industrial Democracy: An International Journal, 23, 271–290. Bartlett, W., Cable, J., Estrin, S., Jones, D. C., & Smith, S. C. (1992). Labor-managed cooperatives and private firms in north central Italy: An empirical comparison. Industrial and Labor Relations Review, 46, 103–118. Becker, G. S. (1996). Accounting of tastes. Cambridge, Mass: Harvard University Press. Bonin, J. P., Jones, D. C., & Putterman, L. (1993). Theoretical and empirical studies of producer cooperatives: Will ever the Twain meet? Journal of Economic Literature, 31, 1290–1320. Bowles, S., & Gintis, H. (1994). Credit market imperfections and the incidence of worker-owned firms. Metroeconomica, 45, 209–223. Card, D., & Krueger, A. B. (1995). Myth and measurement: The new economics of the minimum wage. Princeton: Princeton University Press.

234

MORRIS ALTMAN

Centre for the Study of Cooperatives. (2002). What is a co-op? Centre for the study of cooperatives, University of California-Davis, http://www.cooperatives.ucdavis.edu/what/ types.html Chuma, H., Kato, T., & Ohashi, I. (2004). What Japanese workers want: Evidence from the Japanese worker representation and participation survey. RIETI Discussion Paper Series 04-E-019, Japan. Cosgel, M., & Murray, J. (1998). Productivity of a commune: The shakers 1850–1880. Journal of Economic History, 58, 494–510. Craig, B., & Pencavel, J. (1992). The behavior of worker cooperatives: The plywood companies of the Pacific Northwest. American Economic Review, 82, 1083–1105. Craig, B., & Pencavel, J. (1995). Participation and productivity: A comparison of worker cooperatives and conventional firms in the plywood industry. Brookings papers on economic activity. Microeconomics, 121–174. Dow, G. (2003). Governing the firm: Workers’ control in theory and practice. Cambridge: Cambridge University Press. Freeman, R. B., & Medoff, J. L. (1984). What do unions do? New York: Basic Books. Freeman, R. B., & Rogers, J. (1999). What workers want. Ithaca: Cornell University Press, ILR Press. Friedman, M. (1953). The methodology of positive economics. In: M. Friedman (Ed.), Essays in positive economics (pp. 3–43). Chicago: University of Chicago Press. Gordon, D. M. (1996). Fat and mean: The corporate squeeze of working Americans and the myth of managerial ‘‘downsizing’’. New York: Free Press. Gordon, D. M. (1998). Conflict and cooperation: An empirical glimpse of the imperatives of efficiency and redistribution. In: E. O. Wright (Ed.), Recasting egalitarianism: New rules for communities, states and markets (pp. 181–207). London: Verso. Gulati, G. M., Isaac, T. M., & Klein, W. A. (2002). When a workers’ cooperative works: The case of Kerala Dinesh beedi. UCLA Law Review, 49, 1417–1454. Guttman, J. M., & Schnytzer, A. (1989). Strategic work interaction and the Kibbutz–Kolkhoz paradox. Economic Journal, 99, 686–699. Ichniowski, C., Kochan, T. A., Levine, D., Olson, C., & Strauss, G. (1996). What works at work: Overview and assessment. Industrial Relations, 35, 299–333. Johnson, A. G., & Whyte, W. F. (1977). The mondragon system of worker production cooperatives. Industrial and Labor Relations Review, 31, 18–30. Jones, D. C. (1985). The economic performance of producer co-operatives within command economies: Evidence for the case of Poland. Cambridge Journal of Economics, 9, 111–126. Jones, D. C., & Svejnar, J. (1985). Participation, profit sharing, worker ownership and efficiency in Italian producer cooperatives. Economica, 52, 449–465. Leibenstein, H. (1966). Allocative efficiency vs. ‘x-efficiency’. American Economic Review, 56(3), 392–415. Leibenstein, H. (1979). A branch of economics is missing: Micro–micro theory. Journal of Economic Literature, 17(2), 477–502. McCain, R. (1973). Critical note on Illyrian economics. Kyklos, 26, 380–386. Merrett, C. D., & Walzer, N. (2004). Introduction: Cooperative theory and its applications for the twenty-first century. In: C. D. Merritt & N. Walzer (Eds), Cooperatives and local development: Theory and applications for the 21st century (pp. 3–19). Armonk: M.E. Sharpe.

Workers Cooperatives as a Competitive Alternative

235

Park, R., Kruse, D., & Sesil, J. (2004). Does employee ownership enhance firm survival? In: V. Pe´rotin & A. Robinson (Eds), Employee participation, firm performance and survival: Advances in the economic analysis of participatory and labor-managed firms (Vol. 8, pp. 3–33). Amsterdam: Elsevier. Putterman, L. (1983). Incentives and the kibbutz: Toward an economics of communal work motivation. Zeitschrift fur Nationalokonomie, 43, 157–188. Putterman, L. (1984). On some recent explanations of why capital hires labor. Economic Inquiry, 22, 171–187. Reder, M. W. (1982). Chicago economics: Permanence and change. Journal of Economic Literature, 20, 1–38. Simon, H. A. (1991). Organizations and markets. Journal of Economic Perspectives, 5, 25–44. Tomer, J. F. (1987). Organizational capital – The path to higher productivity and well-being. New York: Praeger Publishers. Tomer, J. F. (2001). Understanding high-performance work systems: The joint contribution of economics and human resource management. Journal of Socio-Economics, 30, 63–73. Ward, B. (1958). The firm in Illyria: Market syndicalism. American Economic Review, 48, 566–589. Zeuli, K. (2004). The evolution of the cooperative model. In: C. D. Merritt & N. Walzer (Eds), Cooperatives and local development: Theory and applications for the 21st century (pp. 52–69). Armonk: M.E. Sharpe.

This page intentionally left blank

236

CO-OPERATIVES IN GLOBALIZATION: THE ADVANTAGES OF NETWORKING Isabelle Halary ABSTRACT In the last two decades, a new form of organization has progressively become predominant on many global markets: networks. Very few worker co-operatives have adopted such a pattern though, despite the fact that, as the theoretical literature shows, the advantages of network industrial structures are numerous and networking can be considered a necessity in the context of globalization. After introducing a new framework for analyzing networks, we argue that combining several dimensions of integration has been an important factor of efficiency in three case studies: Mondragon Corporacion Cooperativa, the industrial districts of Emilia-Romagna, and Scop Entreprises.

1. INTRODUCTION In the last two decades, a new form of organization has progressively become predominant on many global markets: networks. However, whereas many conventional firms have rapidly shifted from vertical integration to a Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 237–264 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09008-3

237

238

ISABELLE HALARY

reticular structure, most co-operatives still have to take the next step. For those co-ops which have tried it, networking has appeared as an efficient way to deal with globalization and technical change, while introducing a new kind of corporate governance and giving a chance to economic democracy. According to Bernard Guillon, a network is an endogenous form of organization enabling its members to coordinate their activities by creating an environment which generates dynamic externalities, complementarities, and cumulative phenomena, notably in regard to competencies (Guillon, 1992). Through globalization, this concept has actually taken very different forms such as supply networks relying on subcontracting or vertical partnership (Kinder, 2003; Kranton & Minehart, 2000), networks of firms derived from strategic alliances or cooperation agreements (Mouline, 1999), industrial districts (Dei Ottati, 2003; Saba, 1998), clusters, global production networks (Erst & Kim, 2002), and hollow corporations (Veltz, 2000, p. 174). Theoretically speaking, a network does not have to imply cooperation, i.e. carrying out of a common activity, harmonization of practices, and mutualizing of resources among its members. Coordination in a network could be limited to a simple dialog and exchange of information about ends and means. But as it happens, networks often rely on a vast number of cooperation agreements, which can either have a vertical or horizontal dimension. Actually, the largest networks combine supply chain coordination and horizontal alliances. Reticular structures have numerous advantages and, in the global competition, networking has become a necessity for investors’ enterprises as well as co-operatives, as we will see in Section 2. But the new corporate governance established since the 1980s in investor-owned firms, along with the short-termism and instability that were induced by financial globalization, may forbid the cooperation and trust that ground network relationships. In Section 3, we will explain why worker co-operatives appear to be better equipped for building a network organization. In Section 4, we will introduce a framework for analyzing networks through all of their dimensions and will use it in three case studies: Mondragon Corporacion Cooperativa (MCC), the industrial districts of Emilia-Romagna, and Scop Entreprises.

2. WHY WORKER CO-OPERATIVES NEED A NETWORK STRUCTURE The cellular pattern encompasses a multitude of productive morphologies among which one finds the large, integrated firm surrounded by a nebula of suppliers, the network of small- and medium-sized enterprises forming a

Co-operatives in Globalization

239

value chain, the hollow corporation, which coordinates mainly external manufacturing and commercial operations, the network resulting from technological or commercial alliances between large rival corporations, the economic interest group of small- and medium-sized businesses wishing to share tools and resources, or even the network of independent professionals pooling their supplies when bidding on large contracts (Veltz, 2000, p. 174). These different forms interpenetrate in the industrial reality. The diversity of the network phenomenon can be explained by the various advantages this type of structure produces for any kind of business and for worker co-operatives more specifically.

2.1. The Advantages of Networking for both Conventional and Co-operative Enterprises: Facing Innovations If its main function is to coordinate the activities of legally independent businesses, the network also enables them to take benefits from diverse and complementary competencies. Networking often follows deverticalization or technical disintegration in a context where informational integration is possible and necessary (Guillon, 1992). Deverticalization with cooperation allows sharing sunk costs, notably those concerning research and development, and mutualizing risks between firms, which have shortened their learning process by refocusing their activities on core competencies. Given the acceleration of products’ evolution and the race for innovation, a single enterprise – even when very large – cannot find in itself all the competencies needed to produce and improve its products. Networking provides the firm with an access to external knowledge and skills indispensable to keep up its rank in the global competition. Moreover, the ongoing relationship between different competencies within the network enables a reciprocal fertilization that generates new competencies (Cohendet, Kern, Mehmanpazir, & Munier, 1999; Antonelli, 1995). In the last decades, innovation has accelerated, thus becoming a major factor of uncertainty. A new product or process can be wiped out in only several months by a new technical generation. Some products that had gone through their development phase in the consumer electronics industry never went to market. From an evolutionary point of view, one could assume that at a given time in the industrial development, several technological paths are open. Therefore, a firm wishing to invest in a definite research path faces a high risk that a solution developed by its competitor becomes a standard and wins the market. Technological alliances have multiplied in the last two

240

ISABELLE HALARY

decades because they answer this problem quite well. They rally firms that are either rivals on the same market segment or operate in different but converging industries. These alliances form networks that produce at least three kinds of advantage. The first one deals with combining necessary knowledge coming from various fields. The second one is the large number of businesses involved and the amount of resources they can gather, thus giving the new product or process the viability and durability that is necessary in its take off stage. The third relies on the ability of networked oligopolies to intervene beforehand on the market in order to define its standards. This ex ante normalization ‘‘also establishes a set of rules on competition that bring stability in a field that is still under construction and therefore shows a high level of uncertainty’’ (Delapierre & Mytelka, 2003, p. 250). Because of accelerating technical change, new products are more sophisticated, and their production implies more specific tangible and human assets (Williamson, 1985; Coelho & Rastoin, 2002). This ‘‘radically changes the nature of relationships between actors: the transaction can no longer be anonymous nor purely instantaneous; a link of durable personal dependency is established between the parties’’ (Coriat & Weinstein, 1995, p. 57). Networks rely on a chain of relationships that generate social benefits such as the accumulation of tacit knowledge or informal know-how (Hakansson & Johanson, 1993). The reticular pattern and the co-operative linkage generate a specific advantage that one can refer to as a relational surplus. Networks must also be considered for their strategic goal; this kind of organization is deliberately used by some firms in order to obtain a competitive advantage. But, according to Jarrillo (1988), this can be achieved only with the impulse of a leading corporation, which is able to shape the whole structure to its own advantage. 2.2. The Advantages of Networking: Facing Globalization On financial markets, the tendency to short-termism is a characteristic feature of investors’ behavior and is passed on to firms’ strategies through corporate governance. Satisfying shareholders by respecting the international norm of return to equity requires a refocus on operations chosen according to the firm’s specific competencies and to externalize any other task. Thus, the industrial fabric has gone through deep restructuring that implies vertical disintegration and scattering of operative units. As a result, the dispersed businesses resulting from this movement have been subjected to a deeper uncertainty of outlets, for some of them, and of supply of raw products or components for others. A solution to this problem would be to

Co-operatives in Globalization

241

add informational integration to technical disintegration, which seems to be the case when firms choose to establish co-operative linkages and to adopt a reticular structure (Guillon, 1992). Cooperation offers the possibility to deal efficiently with uncertainty because this type of inter-firm relationship relies on an ongoing exchange of technical and market information (Allegret & Dulbecco, 1998). For Richardson (1972), firms should go as far as seeing that the volume of rival investments not be higher than a critical limit, regarding the expected demand for a given product, and that complementary investments reach the minimum necessary. Here ends Richardson’s idea of cooperation, but if firms expanded such a behavior at the network scale, would they not stand closer to collective planning than to market economy? For large multinational corporations, networking at a worldwide scale also answers financial motives. In a global production network, the flagship may be a brand leader such as Cisco, General Electric or IBM, or a contract manufacturer such as Solectron of Flextronics. The latter supplies the former so that the two networks are also interlinked. Such a structure at a global scale enables brand leaders to keep out of low margin manufacturing operations, thus increasing shareholder’s return (Erst & Kim, 2002). This explains the many divesting strategies recently pursued by large corporations that have sold out a large number of plants to contract manufacturers. As we have seen, networking productive operations has many advantages for any kind of enterprise, but co-operatives also have more specific reasons for adopting a reticular structure. 2.3. The Advantages of Networking for Worker Co-operatives As Stephen Smith (2004a, p. 181) shows, ‘‘co-operatives may benefit from being in a region with other cooperatives, or in a sector in which there are many cooperatives, or within a supply chain (that is, having significant forward or backward linkages) in which cooperatives are common. In other words there are network externalities or complementarities of organizational type, at least when it comes to organizational form.’’ On a territory where coops are numerous, banks and market actors can get used to dealing with this specific kind of organization, thus lowering transactions costs. This gives grounds to believe that there are strong externalities in regard to co-op structure and that networking helps with internalizing some of these externalities. In France, 25% of small- and medium-sized enterprises belonged to a group in 1998 (INSEE, 2003), and this proportion has strongly increased in the last decade. This figure can be explained by two parallel evolutions. One

242

ISABELLE HALARY

is the growing integration of small and medium businesses in the groups created by large firms that are widening their own networks. Another is the constitution of micro-groups – that is to say, groups made of small and medium businesses that established industrial and financial links among themselves. More than 6,000 micro-groups of less than 500 employees were created in France between 1990 and 1998. Facing this evolution, an isolated co-operative has very few chances to survive. The attitude of banks toward small- and medium-sized businesses in general and toward co-operatives in particular still lacks optimism: ‘‘Banks do not lend enough money to allow worker co-operatives development but the industrial environment pushed us into creating a group,’’ explains Jean-Michel Graciet, technical director of the Cema, a French co-operative that produces tools for the aeronautic industry (Gazsi, 2004, p. 6). Networks could be an answer to globalization; such a structure may help co-operatives in building internal markets for intermediary goods, finance, and technology, and help to ease their internationalization. Globalization has accelerated the merging of businesses, thus increasing the size of rivals in the worldwide competition. Enterprises have been forced to internationalize their operations and set up plants abroad. Those that succeeded in doing so have shown high profitable growth prospects and therefore have increased their market value. Thus, they have efficiently attracted most of the available financial resources, which in turn has allowed them to develop. Internationalization is also an answer to the expectations of clients who themselves have invested on foreign markets and looked for global solutions and services. The impact of all this on co-operatives can be very negative. Multinational groups have had huge resources allocated to communication, research and development and, more generally, investment. Therefore, their notoriety has increased to the detriment of co-operatives, whose clients may be drawn away. Worker co-operatives must face tougher competition with limited means, and the narrowness of their local or national market is a weakness in such an environment. Therefore, they are threatened either by decline or by the temptation to adopt a more conventional behavior that would rely on external growth and market finance. Thus, the transformation of co-operatives into joint-stock companies could be on its way (Fitoussi, 2001). Networking, though, could give co-operatives the possibility to target larger markets, pool financial resources, and strengthen in regard to global competition, without losing their specificity. The acceleration of capital mobility has also increased competition between workers of different geographic areas. Since co-operatives have high social

Co-operatives in Globalization

243

goals, they tend to pay higher wages and social benefits to their worker members, especially to the low-skilled ones, than conventional enterprises of comparable size do. The increased competition between workforces is then a real threat to co-operatives, which could be tempted either to give up their social goals in their homeland or to establish affiliates in low-wage countries without giving their new workers abroad the possibility to become co-operative members. Networking here again is a solution: vertical cooperation would help co-operatives to avoid working as subcontractors for large corporations and, as a group, start instead producing added-value products on a large scale using high technology, which in turn would allow them both to be competitive and maintain high social goals. Knowing that investors’ businesses are networking, there is an urgent necessity for worker co-operatives to do so. In many industries, global production networks structure the market. Multinational firms have established well-known brands and distribution networks, thus securing their access to well-paying customers. When they do not build a network among themselves, workers’ co-operatives have to enter in the supply network of large investors’ firms. But then their position is very weak, and they are dominated. In such a configuration, they cannot prevent their worker members from being exploited by capitalist firms. There is no solution out of network on the global market – it is either the capitalist network or the network of co-operatives. As Jose-Maria Arizmendi-Arrieta said, ‘‘a co-operative cannot survive alone’’ (cited in Servy, 1981, p. 57). Networks are increasingly popular among both large corporations and small businesses because they constitute flexible and efficient means of coordination when facing innovation and globalization. For co-operatives more specifically, networking has become an urgent necessity in global competition. A reticular structure can strengthen them by giving them a viable access to broader markets without questioning their co-operative identity. Moreover, as we will see in the next section, worker co-operatives foster all the values and competencies used in networking.

3. WHY WORKERS CO-OPERATIVES ARE BETTER EQUIPPED FOR NETWORKING 3.1. What a Network Organization Requires Cooperation is a relationship that firms build upon in the long term in order to share a limited set of diverse resources (financial capital, equipment,

244

ISABELLE HALARY

technology, etc.) without questioning the autonomy of the firms involved. In such co-operative arrangements, ‘‘the parties to them accept some degree of obligation, and therefore give some degree of assurance, with respect to their future conduct’’ (Richardson, 1972, p. 886). This definition suggests that co-operating in a network involves behaviors and competencies that are quite different from the ones that would be relevant in the case of market transactions. The network escapes both the logic of personal interest dear to mainstream economics and opportunism, as defined by transaction costs economists (Billand, 1998). Within a network, trust and embeddedness are grounded on frequent and durable relationships, reciprocal obligations, the feeling to belong to a social group, all of which being typical of well-established social links. This explains how networks and cooperation stand at the exact opposite of market transactions, which are supposed to be anonymous and instantaneous. The network structure relies on non-market interdependence that exists or can be created between the individuals of a society or organization (trust, social cohesion, and proximity reduce transaction costs), between businesses in an industry (industrial clusters and new value chain), and between businesses and their environment (innovative surroundings, industrial districts which foster learning processes, and collective goods) (Levesque, 2001). This non-market interdependence is the result of a social process. The new kind of partnership involved in building networks is ‘‘a social procedure of building and sharing that lie within the framework of long term relationships’’ and that implies ‘‘setting up routines and procedures linking the parties in forms of coordination and arbitration which are essentially nonmarket ones’’ (Coriat, 1997, p. 259). These specific forms of coordination cannot be set up unless the different actors have some kind of common ground. Cooperating in a network does not only mean coordinating in time the investments of different firms, but also implies establishing within the network personal relationships based upon trust, geographic or organizational proximity, common norms and values, common goals, and shared representations that will ground the quality of their cooperation. A network gathers a ‘‘set of actors linked by a number of explicit or tacit relationships that go from simple mutual acquaintance to cooperation, through the regular exchange of products, reciprocity, mutual trust or loyalty, and which constitute a basis for the collective learning of values, norms, behaviors or rules approved by consent’’ (Rizopoulos, 1997, p. 379). These rules, which are often informal, determine the direction and nature of material or informational flows between the actors.

Co-operatives in Globalization

245

This common culture is part of social capital, which according to Jean-Pierre Worms (2001), encompasses the resources that individuals can use to cooperate in a common undertaking: networks of social relationships, common representations, common norms, and common values. These values, representations, and norms ease the use of networks and give them efficiency. Among them, two are of prime importance: trust and reciprocity. For Worms, co-operatives make an intensive use of these resources. In an enterprise of the social economy, people associate in order to do something that matters not only to themselves or their members, but to the whole social body. Finally, building up networks may not go along with strained relations. The new forms of cooperation most often consist in contractual agreements between enterprises or organizations in order to carry out common projects while the parties keep their own identity and their autonomy. In this kind of relationship, competition does not disappear but has to coexist with forms of cooperation based on long-term commitment and reciprocity that might go beyond what the initial contract prescribed. Since they rely on the cooperation of actors that may be rivals in the global competition, flexible systems such as networks require that conflicts between employer and employees be reduced to the minimum and that relationships between suppliers and clients be steady in the long term (Hollingsworth & Boyer, 1997). The flexibility and integration that the new technologies of information and communication allow cannot be fully exploited without the cooperation of workers and subcontractors.

3.2. Why Conventional Firms should Experience Difficulties in this Task In order to succeed in supply specialization and cooperate in networks, firms must gather in industrial communities. Their members must share common norms and values; explicit rules must be adopted in order to reduce competition within the network, the goal being to eliminate all kinds of competition that do not favor innovation (Piore & Sabel, 1984). However, investor-owned enterprises have their own history; their culture is often due to their founder’s personality and, therefore, there is no reason why firms that try to cooperate would have a common culture. The industrial reality is full of examples of mergers failing notably because the managers of the two firms did not share the same values. Of course, in any conventional business, there are some common basic references such as market economy, liberalism, and above all competition. But under competition, firms adopt an aggressive behavior toward each other. Therefore, competition does not

246

ISABELLE HALARY

qualify as a value they could gather on. Firms should reduce competition in order to be able to cooperate. In the wake of globalization, competition has been intensified and its culture has deeply impregnated businesses, making it harder for them to take the first step and lay down their arms. Along with financial globalization, a new corporate governance has been set up. The pressure of investors and financial markets upon firms has been stronger. Higher returns are demanded in short time. Acquisitions, restructuring, downsizing, and layoffs have been frequent, which can suddenly break up the links and solidarities that have been built within a firm and between a firm and its environment. Employees’ and subcontractors’ positions have been more unstable and constantly jeopardized by competition. All this has created a tendency to short-termism and defensive selfishness that may prevent the right conditions for trust and cooperation from emerging. Being more educated and skilled in an economy that now relies on intangible resources and products, employees have tended to assert their autonomy, especially through productive cooperation. Then, as Carlo Vercellone (2004) points out, the tension between knowledge and corporate power has been so strong that it may help explain why management has resorted to making employment unstable and individualizing the wage relation, despite the fact that such a policy is highly contradictory to the needs of an economy based upon knowledge. This kind of economy requires a fostering of cooperation and the building up of networks, which as we have seen rely on long-term commitment, trust, collective thinking, and practices.

3.3. Why Co-operatives are Especially Well-Equipped for Building up Networks In a co-operative, cooperation is preferred to internal competition, sharing knowledge is common practice, and values are more or less the same across organizations because building a co-operative rather than an investorowned enterprise is not a neutral choice in the first place. For all these reasons, worker co-operatives have already got part of the common culture that so many businesses lack when they try to operate together. The values that usually guide the functioning of a co-operative are equality, solidarity, dignity on the job, and participation (Romero & Perez, 2003). Because of their legal status and their general or collective interest mission, enterprises of the public, non-profit, and cooperative sectors may provide sounder bases for partnership and cooperation than capitalist businesses do (Levesque, 2001). In co-operatives, there has been a long tradition of

Co-operatives in Globalization

247

nation-wide and international inter-cooperation. The co-op is especially well gifted to practice reciprocal arrangements. Because of the high level of trust that exists within co-ops, the transaction costs resulting from contracting among co-ops are much lower than those facing a conventional corporation. In joint-stock companies, business goals must conform to one criterion: maximizing profits and increasing the firm’s market value. Financial gain is the only target. The goals of co-operative enterprises are different and broader. If co-operatives must maintain financial viability, they also have social ambitions such as: maintaining steady quality jobs with wages and social benefits that allow a decent living; promoting their members by a fair access to training, knowledge, and culture; producing only goods and services that are appropriate and safe; supporting long-term sustainability in the communities in which they operate; promoting economic democracy and participation; providing ordinary people with the means of empowerment and protection against the domination of monopolies or financial power; and offering marginalized populations opportunities for an equitable reintegration (World Summit for Social Development, 1995). As a matter of fact, co-operatives give shape to the ideals of several generations of citizens, even though disagreements are frequent among co-operatives’ members. To the opposite, conventional firms try very hard to build up, quite artificially, a ‘‘corporate culture,’’ and whenever they succeed to do so, the values it contains are rather meaningless to workers who have internalized 20 years of social insecurity. Worker co-operatives are not to be closed down very easily, nor to be sold or relocated, since their investors are their workers. Their stability is a competitive advantage in itself: one can depend on them and believe in longlasting business relationships with them. And so can their workers, who therefore will be willing to invest their money, time, and efforts in building cooperative networks. Worker co-operatives’ values, goals, practices, legal status, and ownership make them natural network founders. We will see in the next section that some co-ops, which have tried networking on a large scale, have succeeded very well.

4. NETWORKS OF CO-OPERATIVES: FIRST RESULTS OF OBSERVATION Stephen Smith (2004b) compared the Mondragon and La Lega networks with respect to the way they address 10 issues, namely co-op entry, co-op exit,

248

ISABELLE HALARY

relations with government, decision making, second level co-ops, joint ventures and alliances, innovation, finance, risk, and employment. He finds that co-op density and network externalities play a central role, and that networking is the most efficient response co-ops can give to these externalities. While these results are consistent with the arguments presented here, we take a somewhat different approach to analyzing networks. Since the notion of ‘‘network’’ encompasses many kinds of arrangements, we want to know what networking actually means. Therefore, the dimensions of this phenomenon that we explore here derive more directly from industrial organization notions. We are interested in the nature of the linkages that were forged between firms that belong to a network, in the way in which they are connected. Studying Mondragon, the industrial districts of Emilia-Romagna, and Scop Entreprises, our goal is to determine which dimensions of networking they have actually been using and how much. We also want to examine which ones they have not used yet and why, as far as at it is possible to know. We want to explore the hypothesis that the use of different types of connections and how the different dimensions interact are critical for co-ops. We first have to introduce criteria that can be used in analyzing the structure and assessing the reality of networks. This first stage will give us a framework that we will use in three case studies: MCC, the industrial districts of Emilia-Romagna, and Scop Entreprises.

4.1. Three Criteria to Analyze Networks When analyzing networks of co-operatives, researchers can use some of the criteria that have already been used with respect to networks of conventional firms. Their degree of connectivity, the dimensions of their integration, and their degree of hierarchy or centralization are elements that can help with understanding the nature of any network. The degree of connectivity (Ricaboni & Pammoli, 2002) refers to the number of ongoing relationships that exist between the members of a given network, the frequency of the linkages, and the amounts involved in the internal transactions relative to the outward transactions. Dealing with this criterion is tantamount to answering the question of how important the network organization is to the businesses that are part of it. Are the partners strongly related? How much does the network affect their business and behavior? Are the commitments between them strong and steady? The dimensions of integration are diverse. A business relationship can be horizontal in many ways, including: the pooling of resources, the

Co-operatives in Globalization

249

mutualizing of costs and risks, the pooling of supply in order to answer the demand of very large clients, and the pooling of raw products purchase to reduce costs. Technological alliances can be either horizontal or cross-industries. The dimension of the network can also be vertical, as in the case of global production networks or when a complete set of partnerships between small and medium enterprises is established. Another dimension to take into account is embeddedness in a territory, as in the case of industrial districts or clusters. The network can also be financial when it includes inter-business credit or an internal capital market. Then we might have to consider the technical and physical nature of a network, when a computer network is set up to complement economic cooperation. Another dimension of this term is the phenomenon of network externalities: the more people, firms, clients, use a given technology, the more productive it becomes, which gives bases to establish a standard. Thus, a network of firms can use its own size and connectivity to generate network externalities that will help it to force its technology on the whole market, as it often happens in the electronics and computing industries. Finally, a few organizations consider employment solidarity as a social dimension of networks. In most studies, authors tend to separate these different dimensions; those who do research on vertical cooperation do not focus their attention on technological alliances, for instance. On the contrary, we chose here to analyze networks though all their possible dimensions, assuming that combining several of them in the same organization brings a higher degree of connectivity and, therefore, is more efficient than using only one of them. The third criterion we can take into account when analyzing networks is their degree of hierarchy. Some networks are strongly centralized and dominated by their leader, whereas others are more democratic. Guillon (1992) establishes a distinction between networks of firms that designate horizontal agreements between enterprises of comparable weight and network-firms where a leading corporation coordinates the operations of a multitude of small and medium businesses. In the former, the decision process can be rather democratic, whereas in the latter, asymmetry and hierarchy are often important. Taking these different criteria into account, we will now analyze three types of network among co-operatives. A table supplied in the appendix synthesizes our results concerning the dimensions of integration. 4.2. A Strongly Integrated and Hierarchical Network: MCC Mondragon Corporacion Cooperativa (MCC) is a co-operative group located in the Basque country in Spain. Organized in a network, it contains

250

ISABELLE HALARY

over 150 industrial enterprises (workers co-operatives and their affiliated companies that can be conventional businesses). But MCC can also depend on its powerful financial group made of two units: a very large bank, Caja Laboral Popular, and a dedicated social security system, Lagun-Aro. The MCC network is completed by a distribution group which notably holds the Eroski consumer co-operative, where over 32,000 people work today. MCC also operates three R&D centers and supports the Mondragon University, where about 4,000 students are enrolled. Mondragon is the first industrial group of the Basque Country and sixth industrial group of Spain, with total sales of h 9,232 millions and an operating profit of h 380 millions in 2002. MCC’s workforce has impressively increased in the last 10 years to reach 68,625 employees in 2003 against 25,317 in 1993, and still grew by 10.5% in 2002 despite the international recession. But only half of this workforce is made of co-operative members, although co-operatives keep on joining the group every year – three did so in 2002. Much of the recent growth in employment and reduction in membership rates are due to the acquisitions abroad. From the day its first co-operative was founded in 1956, Mondragon has always been a very centralized and hierarchical business. Today, each cooperative belongs to its sector subgroup; several subgroups constitute a division, and the seven industrial divisions, the financial group and distribution group are coordinated by corporate governing bodies, notably the General Council and the Co-operative Congress (structure and figures are available on http://www.mondragon.mcc.es/). According to many authors, the main reason for this highly hierarchical structure is efficiency: ‘‘The ceding of decision making power is made with the aim of obtaining an efficient unification of direction and strategic control. Any co-operative becoming a member of the MCC must accept the regulations of the Congress and has to coordinate its development with the rest of the co-operatives within the same sector sub-group. Some authority is shared with the corporate governing bodies; therefore, co-operatives are subject to strategic plans decided by the corporate group’’ (Bakaikoa, Errasti, & Begiristain, 2004, p. 15). MCC is also a highly integrated network. Nearly all the dimensions of integration are functioning in this group. In the horizontal dimension, MCC’s co-operatives do share resources and services, such as a general administration or technology. They mutualize costs and risks when they all contribute to finance their research centers, their university, and even their own social security system. Some of the co-operatives also form technical alliances such as Modutek, a technology unit specializing in car module

Co-operatives in Globalization

251

design and development created by Maier, Batz, and Cikautxo, three members of the automobile division, with the help of Fundacion MCC (participation of 50%). The network also includes very strong long-term vertical linkages, the entire operations being coordinated by MCC’s governing bodies. In the Components Division for instance, the co-operative Consonni produces heating elements for washing appliances such as dishwashers, washing machines, and washer-dryers. Meanwhile, the co-operative Copreci manufactures controls and electric drain pumps for washing machines, washer-dryers, and dishwashers and makes electric thermostats and electronic controls for electric and gas cooking. Then, in the household division, the co-operative Fagor manufactures electrical home products such as washing machines, dishwashers, and ovens. The same kind of backward and forward linkages have been established between the electronics subgroup of the components division and the co-operatives of the automotive and machine tools divisions, for instance. Very early, a financial dimension was given to this network: a co-operative bank, Caja Laboral Popular, was created in 1959, not only for carrying out banking activities but also for the centralization of the co-operatives accounts, and the financial and economic control of their activities (Servy, 1981). An actual financial solidarity within this network also shows up when a cooperative experiences difficulties; inter-enterprise loans and cross participations are part of the heavy commitments each new business signs up for when it becomes a member of MCC. This network is also deeply embedded in its territory: it shows a strong loyalty to the Basque Country. Until now it has claimed that its goal is to create and protect quality jobs for the Basque people who in return massively deposit their savings in Caja Laboral Popular. Some of the classes in Mondragon University are taught in Basque language, and the links between MCC co-operatives and the local population are strong. The geographic proximity of the different businesses allows them to share corporate services such as collective transportation or meals. Technically speaking, the co-operatives are inter-linked through a computer network, and a corporate purchase portal called Ategi has been functioning for several years. This business-to-business web site has improved the purchasing process of member co-operatives concerning varied products such as communications, office equipment, hardware, tooling, translations, graphic services, energy, steel, and so on. Finally, employment solidarity has been an original dimension of MCC. It constitutes another of the strong commitments of its co-operatives: when serious difficulties are experienced by one of them, the other co-operatives within a reasonable distance have to accept some of the redundant workers, who thus will be reallocated rather

252

ISABELLE HALARY

than dismissed. The only network dimension that seems not to exist in MCC concerns network externalities, as defined here above. If they do exist there, at least we have had no element of verification hitherto. The hierarchical nature and the high degree of integration of the Mondragon network come from its history. Inspired by the philosophy and example of Father Jose-Maria Arizmendi-Arrieta, five students of the Professional School of Mondragon, an institution created by the priest, incorporated the first cooperative, Ulgor, in 1956, at that time manufacturing paraffin stoves and heaters. When Ulgor’s operations developed impressively, its founders chose to spin-off some of its units and use vertical cooperation rather than big size integration as a means of coordination. Thus, Ulgor’s foundry was incorporated as Ederlan in 1963, its mechanic components unit became Copreci the same year, its electronic components unit became Fagor Electronica in 1966, and its hotel equipments unit was incorporated as Fagor Industrial in 1974 (Servy, 1981). The same idea of dissociation was applied concerning financial services, engineering, or commercial services. Caja Laboral Popular was created in 1959 and immediately played a leading role in the network. By the end of the 1960s, the total number of co-operatives had risen to 41. Some of these were spin-offs from Fagor, others had their origins in the Business Division of Caja Laboral, and yet others were formed by autonomous groups or as a result of the transformation of public limited companies (MCC, 2001). The strategy of spin-offs and cooperation adopted by the group’s founders as well as the initial leading role of Caja Laboral can partly explain the high degree of integration and hierarchical nature of the Mondragon network.

4.3. A Decentralized Network: The Industrial Districts of Emilia-Romagna In the industrial districts of Emilia-Romagna (Italy), co-operatives and conventional small and medium enterprises cooperate with each other. In this region, there are about 8,000 co-operatives that have built up their own organizations and solidarities in such a way that one might say they have formed networks of co-operatives within the Emilia-Romagna clusters. Consequently, it is difficult to isolate the network of co-operatives from the rest of the local reticular structures. Small- and medium-sized enterprises, both co-operative and conventional, have specialized and have formed industrial districts. One can find a knitwear district, a clothes district, and a ceramic tiles district in Modena; an automatic machinery district, a packaging machinery district, and an agricultural machinery one in Bologna; a wood-working machine tools district

Co-operatives in Globalization

253

in Carpi; a tomato canning district, a ham district, and a food-processing machinery one in Parma, etc. (Brusco, 1982). Most of the co-operatives in this region belong to one of the three national co-operative movements: the ‘‘Lega’’ on the political left (40,000 employees in Emilia-Romagna in 2003), the ‘‘Conf Coop’’ on the catholic center-right (1,858 co-ops and 40,000 employees), and the ‘‘Associazione’’ on the center left (470 co-ops). In addition, 3,000 non-affiliated co-operatives operate in this region (Thompson, 2003). The networking there involves all kinds of cooperatives: industrial worker co-operatives, as in the ceramic machinery industry; retail co-operatives, such as IperCo-op stores (food superstores of the Lega); or social co-operatives operating in health care, child care, senior care, or providing support for the disabled. This region has a long history of cooperation. The first co-operatives operating there were created in the 1860s. By 1921, there were 3,600 consumer co-operatives and 2,700 production co-operatives in the region. Between 1922 and 1944, all of them were taken over by the fascists. At the liberation, the Resistance movement took care of rebuilding a co-operative sector. Since WWII, the region has been continuously governed by a coalition of leftist parties. The regional government created policies and programs that favored both co-operatives and small- and medium-sized enterprises (Thompson, 2003). Cooperation between firms has been a highly regarded factor in the selection of projects to get local government support. In the clusters of Emilia-Romagna, nearly all dimensions of networking are used. Many horizontal linkages have been established in the industrial districts. Small enterprises and co-operatives often pool their offers and subcontract among themselves in order to bid on large projects or contracts and then share the work among the members of their industrial district. Within many sectors, the co-ops developed specific secondary co-operative – some for instance has grouped the 700 co-ops that make Parmegiano Cheese. Another meets the needs of co-operatives involved in the ceramics industry (Thompson, 2003). The Lega has also created structures that help the co-operatives with developing, financing, planning, and communicating. The vertical dimension is also present in these networks. Initially, the small firms of Emilia-Romagna were subcontractors of large enterprises and very dependent on them. Therefore, the situation was not satisfactory (Herman, 1988). This is why, with the help of their region, they established co-operative relations among themselves in order to create value-added products, linking many producers and creating local production systems that secure global markets. In this system, small firms are highly specialized; only a small part of them market finished goods. The others work as subcontractors (Brusco, 1982).

254

ISABELLE HALARY

Cooperation here clearly replaces vertical integration as a means of coordination. Co-operatives have added other forms of vertical networking to this general regional organization. The IperCoop stores, for instance, which are part of the Lega Coop consumer cooperative system, sell many of the co-operatives’ products such as the Parmesan Cheese. There is also a vertical solidarity among co-operatives that consist in a strong ongoing commitment to purchasing goods and services from within the co-operative sector or within the region (Thompson, 2003). A strong territorial dimension is a particularity of this decentralized organization. The industrial districts of Emilia-Romagna are deeply rooted in the local population. There have been both co-operative and craftsmanship traditions in this region where many citizens are actively involved in local associations and co-operatives of all sorts (Putnam, 1993). For political reasons, the regional government of Emilia-Romagna has set up structures to help small businesses and co-operatives and encourage networking among them. ERVET, for instance, is a public–private partnership created by the regional government with the support of the banks and industrial associations to promote economic development. This organization has set up a network of industry-specific service centers located within the industrial districts in the various towns (Brusco, 1982). These agencies provide small businesses and co-ops with R&D consultations, technical services, business development, and marketing and training. The network of co-operatives also gave itself a financial dimension through the Co-op Fund, which supplies capital to co-operatives and is funded by 3% of all the profits in the sector. Besides, many secondary cooperatives formed lending circles guaranteeing each other loans (Brusco, 1982). The Lega consumer co-operatives, which gather over a million members, have also helped financing the sector by collecting its members’ savings on accounts opened at their stores. These savings accounts pay a higher interest rate than the local banks do, and the co-ops use the capital for their development (Thompson, 2003). Employment solidarity is another dimension of this network; La Lega co-ops have an informal commitment to hiring workers from other co-operatives in a difficult situation, when it is possible for them to do so. Again, the decentralized structure and high connectivity of these networks largely result from regional history. A decentralized network model was deliberately chosen because it could allow ownership diversity, which was part of a local political compromise. Then, the high connectivity of the network has been encouraged by the regional government because it was the only way to make it work in a context of global competition.

Co-operatives in Globalization

255

4.4. The Scop Entreprises Network: Links but no Ties The Scop Entreprises network gathers 1,577 worker co-operatives located on the French territory. Over 35,000 people work there. Thirty percent of the co-ops operate in the building industry, 5% in the graphic and printing industry, 19% make goods, and 46% produce services (figures available on http://www.Scop.co-op/chiffres.php). Between 1993 and 2003, the number of French worker co-operatives has risen from 1,292 to 1,577, and the number of workers enrolled there has increased by 20% since 1993. While creating jobs, the co-ops have also improved their labor productivity; the value added per worker has risen from h 34,000 to h 40,000 between 1992 and 2002. Economic and financial results of the French co-ops have improved in the last 10 years; the sales per worker ratio has grown from h 75,000 to h 92,000, and between 75% and 80% of the co-ops have shown profits. In the year 2003, the cumulated profit of the network was h 110 millions, for sales of roughly 3 billion. The increase in jobs and units has not weakened the financial structure of the businesses; the average equity per worker has risen from h 4,000 to h 6,000, and net assets per worker (assets minus debts) have risen from h 18,000 to h 30,000 in the last decade. A confederation, the CG Scop organization, is at the core of the Scop Entreprises network. It provides a number of services and tools that help its members and fosters co-operation among them. However, the co-operatives remain very independent from one another, and therefore, establishing strong linkages between them mainly depends on their will to invest in this direction. The CG Scop confederation has willfully developed its network in two dimensions: horizontally and financially. In the horizontal dimension, CG Scop has established institutions in order to foster the exchange of information and co-operation between the co-ops on a territorial basis by way of 12 regional unions. Each of them provides advice and consulting services in accounting, law, and finance to all co-ops of its territory. Auditing is compulsory once a year. One of the most helpful actions of the regional unions is the support they provide in launching new co-ops, whether they are new ventures or are derived from workers’ buy out of conventional firms. In any case, all legal and financial matters are handled by the relevant regional union of CG Scop, which will provide some training in management as well when necessary. CG Scop has also tried to foster co-operation within each industry. Three federations were created: a building and civil engineering federation, a communication federation, and an electronics and metallurgy federation (Temis).

256

ISABELLE HALARY

Purchase pooling is another interesting device that is present in this network. Two of the three federations, i.e. the communication federation and the electronics and metallurgy federation, have set up groupings that buy materials in bulk or equipment for their members. Grouping and purchase pooling is present among the co-ops which operate in the business of building up electric lines as well. CG Scop is also discussing a contract with Sage in order to pool the co-ops’ demand in computing services and products. At a smaller level, horizontal linkages between co-ops can be established spontaneously. In the Rhone-Alpes region for instance, 11 research consultancy coops have gathered and built up a second level co-operative called Quadriplus Groupe, whose main purpose is supply pooling. There are cross-capital participations between each consultant and Quadriplus groupe. A lot of supply pooling has also been observed among the co-ops of the printing industry (Dhoquois, 2002). An informal kind of co-operation has been in use in the Temis federation (metallurgy) among co-ops, which make molds for foundries. Without any capital participation across them, Nor-Meca-Moul, Sgol, and Olaberria share information, monitor technological and economic development for each other, subcontract for one another and show a high level of solidarity in hard times, which goes as far as lending machines after a fire destroyed a plant (Merlant, 2002). The financial dimension is also very strong in the Scop Entreprises network. CG Scop has set up three financial institutions to fund co-ops mainly at their launching, but also in difficult times. Socoden provides business loans to co-ops and individual loans to co-op founders or new members who need help buying their shares. SofiScop brings its guarantee (sustained by the reputation of the whole network) to co-ops that seek loans outside of the network. SPOT issues equity to new co-op ventures or to established co-ops that want to finance their growth. Besides its own funding institutions, the network can rely on the support of the large co-operative banking group, Cre´dit Coope´ratif, whose specialty is to back up the foundation of new coops as well as investment in the co-operative sector. Some foundations, such as ESFIN-IDES or France Active, also help the third sector on a regular basis and work with the Scop Entreprises network. Although not very frequent, there are some cases of capital participation across the co-ops. In the East region for instance, Germat (Etablissements Franc- ois) owns part of Cussenot, Mader SA owns more than 66% of STPM and a small share of SNBR, and La Fonderie de la Bruche owns parts of both Colibru and Plastibru near Nancy. With the new European co-operative statute that will come into effect by 2006, the prospect of opening divisions or affiliates in other European countries

Co-operatives in Globalization

257

is also contemplated by some of the French co-operatives. More locally, we find many cases of spontaneous financial solidarity in this network; at the launching of a new co-op, or whenever it experiences difficulties, another member of the network will lend some funds; as Mrs Marie-Madeleine Maucourt from the regional union of Nancy explains: ‘‘Very often, the amount put forward is not huge but is sufficient to send a positive signal to both the banks and the commercial court in charge of the case’’ (interview with the author, March 4, 2005). Therefore, for example, Mader SA helped SNBR when they were facing hard times, and 30 co-ops added their financial support to the funds coming from Credit Cooperatif and several regional unions in order to save Le Travail in Limoges. In the vertical dimension, we find very few linkages that result solely from the initiatives of the co-operatives involved. Those that forged such links did so because they would rather spin-off their divisions than grow endlessly; they chose to create affiliates with whom they maintain forward and backward linkages. That has been the case for the co-op La Fonderie de la Bruche in Schirmeck (Eastern Regional Union, Metallurgy Federation), for instance, from which Colibru and Plastibru, two co-ops of the same region and federation, were derived. La Fonderie designs and makes component parts in aluminum and zinc for the automobile, domestic appliance, and heating and tool industries. Colibru applies surface treatment on aluminum, zinc, and other metal parts. Plastibru makes parts and tubes out of plastic or PVC. We see here a case of vertical co-operation among co-ops with financial links, but such a case is quite unusual in a network whose members prefer to operate independently. What makes the vertical dimension so weak in the Scop Entreprises network is first that the density of co-ops in any of its regions is too low, and, second, that there is absolutely no rule of priority toward inter-co-op transactions. If there is a commitment to the co-operative movement, it shows in difficult times by some kind of financial aid between the members of the network, but not continuously in daily commercial and industrial decisions. The territorial dimension of the network is a reality for co-ops of the building industry and for those that operate in training. Otherwise, the density of co-ops on any part of the French territory appears to be too low for them to take advantage of geographic proximity. There is no actual cluster of co-ops to speak of in any area of the country. In the building industry though, co-ops that are located close enough to one another know and help each other in getting new contracts; tend to work together on large deals; choose their new workers and members locally; participate in the training of local workforce; and have established strong

258

ISABELLE HALARY

relationships with the regional political institutions. In this industry, some co-ops of complementary activities operate together on a regular basis, taking advantage of both their geographic and economic proximity, as do Mader and STPM, two co-ops of the East region (Goebviller) – the former in the building industry, the latter in civil engineering. Near Troyes, SNBR Baˆtiment Re´gional (building industry) often works with Aubelec (electricity) and Les Mac- ons de Troyes (building industry). These three co-ops belong to the East Regional Union and to the building and civil engineering Federation. But as Michel Longe`re from SNBR explains: ‘‘We work with them because we trust them and know they are good at what they do, not because they are co-ops’’ (interview with the author, January 27, 2005). In fact, SNBR, which restores listed historic buildings, works regularly with a network of 14 businesses in different specialties, of which only these two are co-ops. A critical advantage in this line of business is to be well connected with the local people and institutions. SNBR has had good relationships with the public architect and the mayor of Troyes, the chamber of commerce and industry as well as the guild house; they also welcome students from the technical college for training periods, give the town’s foreign official guests a tour of the local historical buildings and so on. In the training sector also, the links between co-ops and local institutions are strong, since most contracts involve public funding or official accreditation to some extent. A technical dimension will soon be given to this network. CG Scop is working on its Agir project, which is going to establish both an internet and intranet between its members in order to foster communication and business-to-business transactions among co-ops. More locally, two co-ops, Germat (Etablissements Franc- ois) and Cussenot, which both sell building materials, have already set up their intranet which connects eight sites and headquarters in order to manage and coordinate their operations. Finally, no specific action in order to generate network externalities has been undertaken in the network, and no employment solidarity has formally been required from its members. To sum up the first results from this research on the French co-operative network, we may say that it is made up of a rather loose set of relationships. Very determined to remain independent, the French co-ops have established institutions that do provide critical services but exert few constraints and little influence on their economic behavior. To compare this network with the two others previously described in this study, it is the least centralized and hierarchical one, and also the least political one.

Co-operatives in Globalization

259

5. CONCLUSION With globalization, competition has intensified. In the face of multinational networked firms, isolated co-operatives have little chance to survive. For them, the issue is not to choose between networking and facing competition alone; nowadays, each worker co-operative has to take a stand between two alternative strategies: one is to join the global production network of a large corporation, and the other consists in building up a network with other cooperatives. If a co-op chooses to work with large investor-owned firms, it may have to give up the social goals and values that are at the core of its identity. Therefore, networking among co-operatives seems to be the only viable solution. Unlike investor-owned firms, co-operatives already have the common culture, social skills and values necessary to build up networks. Furthermore, their status preserves them from the pressure of financial markets. This gives them the stability they need to establish long-lasting relationships. When they do organize in networks, co-operatives are successful. MCC and the industrial districts of Emilia-Romagna are good examples of co-operative efficiency. Scop Entreprise has experienced significant growth and progress in the last decade, but the network is still not dense nor large enough to draw any persuasive conclusion in this respect. Hierarchical reticular structures and decentralized networks seem to succeed as well in the case of co-operatives. On the other hand, the degree of connectivity and the dimensions of integration must be taken into account when it comes to explain the efficiency of a network. Both MCC and the industrial districts of Emilia-Romagna are highly connected and use nearly all the possible dimensions of integration, especially the vertical dimension that has seemed to be lacking in Scop Entreprises. This difference can be explained by the history of these three groupings. The French one stems from the gathering of existing co-ops scattered on a large territory. In the Basque and Italian cases, some kind of center, pursuing a strong political goal, built up the network by creating new co-ops. Nowadays, there is another explanation for the lack of five dimensions in the Scop Entreprises network: the low density of co-operatives in any given region of the French territory has impeded co-ops from forging forward and backward linkages as well as from building up clusters locally. Consistently with the results of Smith, (2004 a, b), the shortage of network relationships has hampered co-op entry, which explains the low co-op density. Nonetheless, there has been some progress regarding both co-op entry and networking in

260

ISABELLE HALARY

Scop Entreprises during the last decade, and it will be worth following this grouping up in the near future. Our examples show that networking co-operatives are efficient. In all three cases, the networks have grown in terms of sales, exports, and jobs, making value added products using high technology. But then, another debate could be brought up: what sort of efficiency are we talking about? In networks of joint-stock companies, what counts is only economic and financial efficiency. In networks of worker co-operatives, seeking social efficiency and preserving the co-operative nature of the group must also be considered.

ACKNOWLEDGMENTS I would like to thank Panu Kalmi (editor) and Christopher Gunn (reviewer) for their helpful comments and suggestions. My thanks are also to the Union Re´gionale des Scops of Nancy for providing a lot of information.

REFERENCES Allegret, J. P., & Dulbecco, P. (1998). Le comportement de la firme innovante: Structure de gouvernance et mode de financement. Revue d’e´conomie Industrielle, 84, (2nd quarter), 7–26. Antonelli, C. (1995). Economie des re´seaux: varie´te´ et comple´mentarite´. In: A. Rallet & A. Torre (Eds), Economie re´gionale et e´conomie spatiale (pp. 253–272). Paris: Economica. Bakaikoa, B., Errasti, A., & Begiristain, A. (2004). Governance of the Mondragon Corporacion Cooperativa. Annals of Public and Co-operative Economics, 75(1), 61–87. Billand, P. (1998). The´orie des couˆts de transaction versus the´orie des re´seaux: quelques enseignements de la confrontation aux re´alite´s industrielles. Economies et Socie´te´s, Se´rie W, 4(July), 163–182. Brusco, S. (1982). The Emilian model: Productive decentralization and social integration. Cambridge Journal of Economics, 6, 167–184. Coelho, A., & Rastoin, J.-L. (2002). L’e´mergence de l’organisation en re´seau dans l’industrie: Le cas des FMN du secteur des boissons alcoolise´es. Economies et Socie´te´s, 9–10(September–October), 1503. Cohendet, P., Kern, F., Mehmanpazir, B., & Munier, F. (1999). Knowledge coordination, competence creation and integrated network in globalized firms. Cambridge Journal of Economics, 23, 225–241. Coriat, B. (1997). Globalization, variety and mass products: The metamorphosis of mass production in the new competitive age. In: R. J. Hollingsworth & R. Boyer (Eds), Contemporary capitalism, the embeddedness of institutions (pp. 240–264). Cambridge: Cambridge University Press.

Co-operatives in Globalization

261

Coriat, B., & Weinstein, O. (1995). Les nouvelles the´ories de l’entreprise. Librairie Ge´ne´rale Franc- aise, coll. Paris: Le Livre de Poche. Dei Ottati, G. (2003). Exit, voice and the evolution of industrial districts: The case of the postWord War II economic development of Prato. Cambridge Journal of Economics, 27, 501–522. Delapierre, M., & Mytelka, L. K. (2003). Cascades d’innovations et nouvelles strate´gies oligopolistiques. Revue d’e´conomie industrielle, 103(2–3), 233–252. Dhoquois, A. (2002). Co-operation inter-Scop: l’humain fait la force. Participer, 596(November–December). Erst, D., & Kim, L. (2002). Global production networks, knowledge diffusion and local capability formation. Research Policy, 31(8–9), 1417–1429. Fitoussi, J. P. (2001). Intervention at the XXIst national conference of the Groupe Cre´dit Coope´ratif: Les entreprises de l’e´conomie sociale rele`vent le de´fi de la mondialisation. Lettre du Cre´dit Co-ope´ratif 72, December. Gazsi, M. (2004). S’unir ou mourir? Partenariats Scop, mode d’emploi. Participer, 603(January–February), 15–19. Guillon, B. (1992). Technologie, organisation et performances – Le cas de la firme-re´seau. Revue d’e´conomie politique, 102(4), 563–592. Hakansson, H., & Johanson, J. (1993). The networks as a governance structure: Interfirm cooperation beyond markets and hierarchies. In: G. Grabher (Ed.), The embedded firm (pp. 35–51). London: Routledge. Herman, B. (1988). Industrial development: Targeting new and basic industries. National Council for Urban Economic Development, October. Hollingsworth, R. J., & Boyer, R. (1997). Coordination of economic actors and social systems of production. In: R. J. Hollingsworth & R. Boyer (Eds), Contemporary capitalism, the embeddedness of institutions (pp. 1–47). Cambridge: Cambridge University Press. Institut National des Statistiques et Etudes Economiques [INSEE]. (2003). Images e´conomiques des enterprises et des groupes au 1er janvier 2001. Insee Re´sultats-Economie, 6(March), 1–34. Jarrillo, J. C. (1988). On strategic networks. Strategic Management Journal, 9, 31–41. Kinder, T. (2003). Go with the flow: A conceptual framework for supply relations in the era of the extended enterprise. Research Policy, 32, 503–523. Kranton, R. E., & Minehart, D. (2000). Networks versus vertical integration. The Rand Journal of Economics, 31(3), 570–601. Levesque, B. (2001). Le partenariat: une tendance lourde de la nouvelle gouvernance a` l’e`re de la mondialisation: Enjeux et de´fis pour les entreprises publiques et d’e´conomie sociale. Annals of Public and Co-operative Economics, 72(3), 323–338. Merlant, P. (2002). Les fe´de´rations professionnelles cre´ent des liens entre Scop du meˆme me´tier. Participer, 596(November–December). Mondragon Corporacion Cooperativa [MCC], (2001). Mondragon Corporacion Cooperativa, The history of an experience, Pdf electronic file, February, Available at http:// www.mondragon.mcc.es Mouline, A. (1999). Les accords de co-ope´ration inter-entreprises dans les technologies de l’information: nouveaux apports quantitatifs et qualitatifs. Revue d’e´conomie industrielle, 89, 85–106. Piore, M. J., & Sabel, C. F. (1984). The second industrial divide. New York: Basic Books.

262

ISABELLE HALARY

Putnam, R. D. (1993). Making democracy work; civic traditions in modern Italy. Princeton, NJ.: Princeton University Press. Ricaboni, M., & Pammoli, F. (2002). On firm growth in networks. Research Policy, 31(8–9), 1405–1416. Richardson, G. B. (1972). The organization of industry. The Economic Journal, 82(327), 883–896. Rizopoulos, Y. (1997). Marche´s, organisations, re´seaux: alternatives ou comple´mentarite´s. In: C. Pallois & Y. Rizopoulos (Eds), Firmes et e´conomie industrielle (pp. 369–387). Paris: L’Harmattan. Romero, A., & Perez, M. (2003). Organizational culture, individual differences and the participation system in co-operativism of associated workers in Andalousia, Spain. Annals of Public and Co-operative Economics, 74(2), 283–320. Saba, A. (1998). The Italian industrial flexibility and globalization. Economie Applique´e, 51(1), 159–172. Servy, P. (1981). Les co-operatives de Mondragon. Socie´te´ Interprofessions Services, Bayonne (France). Smith, S. C. (2004a). Network externalities and co-operative networks: Stylized facts and theory. In: L. Sun (Ed.), Ownership and governance of enterprises: Recent innovative developments (pp. 181–201). New York: Palgrave/Macmillan. Smith, S. C. (2004b). Network externalities and co-operative networks: A comparative case study of Mondrago´n and La Lega with implications for developing and transitional countries. In: L. Sun (Ed.), Ownership and governance of enterprises: Recent innovative developments (pp. 202–242). New York: Palgrave/Macmillan. Thompson, D. J. (2003). Italy’s Emilia Romagna: Clustering co-op development. Co-operative grocer [Online], 109(November–December). Veltz, P. (2000). Le nouveau monde industriel. Paris: Gallimard. Vercellone, C. (2004). Le sens et les enjeux de la transition vers le capitalisme cognitif: Une mise en perspective historique. Communication for the international conference: The transformations of contemporary capitalism: Facts and theories. Current state of affairs and prospects, LAME Third study days, University of Reims, March 31–April 2, unpublished. Williamson, O. (1985). The economic institutions of capitalism. New York: The Free Press. World Summit for Social Development. (1995). Report, 6–12 March (A/CONF.166/9), Chap. I, resolution 1, annex II, paras. 31, 34, 88 and 91, Copenhagen. Worms, J. P. (2001). Intervention at the XXIst national conference of the Groupe Cre´dit Coope´ratif: Les entreprises de l’e´conomie sociale rele`vent le de´fi de la mondialisation. Lettre du Cre´dit Co-ope´ratif, 72(December), 4–16.

Dimension 1. Horizontal

2. Vertical

3. Territorial

Mondragon

Industrial districts of Emilia–Romagna

Scop Entreprises

Sharing general administration Purchase pooling Social security system: LagunAro Three R&D centers Mondragon University Technical alliances: e.g. Modutek Backward and forward linkages Cooperation between the components and household divisions: Copreci-Fagor Consonni-Fagor The Basque people is given first priority (jobs, schools, Universityy The Basque people deposit their savings in Caja Laboral The co-ops pool services locally

Supply pooling and work sharing Purchase pooling Secondary co-ops (Parmegiano cheese, Ceramics industry) Co-operative structures of the Lega (financing, planning, communicating)

Cooperative structures of CG Scop Unions, 3 federations Secondary co-ops: Quadriplus Supply pooling and work sharing in some local groupings Purchase pooling Very few backward and forward linkages

A little bit of local embeddedness in the building and training industries

263

From manufacturing to distribution (IperCo-op Stores) Cooperation and subcontracting Purchasing policy: priority to cooperative sector products and local products Regional embeddedness Regional co-operative and craftsmanship traditions Strong involvement of the regional government: public service centers within the industrial districts

Co-operatives in Globalization

APPENDIX: THE DIMENSIONS OF INTEGRATION

264

APPENDIX. 4. Financial

Group Bank: Caja Laboral Popular Inter-enterprise loans Cross participations

5. Technical

Computer network B2B site: Ategi

(Continued )

Financial institution: Co-op Fond Lending circles for loan guarantee Savings accounts of the Lega consumer co-ops

Financial institutions: Socoden, SofiScop and SPOT Support from Cre´dit Coope´ratif Cases of cross participations Voluntary financial solidarity in hard times is frequent among co-ops Soon to come: a computer network in project

6. Network externalities 7. Employment solidarity

Workers are relocated within MCC rather than dismissed

Informal commitment to hire workers from other co-ops in a critical situation when possible

ISABELLE HALARY

POSITIVE EXTERNALITIES OF ORGANISATION CULTURE: THE SOCIAL INTEGRATION OF WORKING ADULTS WITH LEARNING DISABILITIES Abraham Mukolo, Robert Briscoe and Agus Salim ABSTRACT Workers live two overlapping lives, at work and outside work. The spillover of favourable workplace experiences into non-work domains of life means that the workplace can be a means by which organisational members who experience network poverty arising from adverse social factors can overcome social exclusion. Social acceptance and interaction data from 105 adults with mild to moderate learning disabilities working in eight social enterprises in the UK and Ireland is examined to establish the link between organisation culture and workplace social integration. In this study organisation cultures in which user/worker-involvement in management and control decision-making is emphasised seem to engender a positive influence on the social interaction experiences of members with learning disabilities in work and non-work domains of life, having regard to difference in demographic factors, employment characteristics, country Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 265–296 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09009-5

265

266

ABRAHAM MUKOLO ET AL.

of residence, and level of disability. The study accentuates the importance of workplace democracy in enhancing the quality of life of working adults with learning disabilities, who might otherwise be disenfranchised in numerous areas of life.

1. INTRODUCTION In organisation analysis, notions of ‘fitting in’ and ‘the way things are done here’ are central to methods of assessing organisation culture and/or behaviour (Bailey, Schermerhorn, Hunt, & Osborn, 1986; Chatman & Jehn, 1994; Chatman, Polzer, Barsade, & Neale, 1998; Hofstede, 1998; Davies, Nutley, & Russell, 2000). These concepts are often viewed from the perspective of their impact on productivity and the economic performance of the organisation, but their impact on an individual’s experiences of life outside work is rarely scrutinised. For example, several studies have shown that interpersonal relationships have implications on worker motivation, job satisfaction, overall performance, and many other aspects of organisational life (Herzberg, 1968; Bailey et al., 1986; Cohen, 1997; Hart, 1999; Knight & Westbrook, 1999; Blau, 1999; Leornard, Beauvais, & Scholl 1999; Lou, 1999). The same concepts are central to understanding the organisational life of organisation members, particularly the extent of connectedness that a member enjoys in the organisation. Within the field of human relations there has been an interest, lately, to combine theory about organisations with psychodynamics theory in order to ‘‘better specify how organisational cultures condition the emotional experiences of their members on a day to day basis’’ (van Buskirk & McGrath, 1999). Furthermore, researchers concerned with the science of administration have also explored the issue of control in organisational life, in particular the extent to which non-managerial organisation members exert considerable control over work organisations (Jermier, 1998; Warhurst, 1998). Since it is a well-established view that workers live two overlapping lives, i.e., at work and outside work (Bailey et al., 1986; Cohen & Vigoda, 1998; Aryee, Fields, & Luk, 1999; Hart, 1999), connectedness in the workplace might have implications on connectedness in non-work domains of life. For organisational members, who experience network poverty (Putnam, 2002) because of constraints caused by unfavourable social values, norms, and practices, the spill-over of favourable workplace experiences into non-work domains of life, means that the workplace can be a means by which the socially excluded can overcome social disadvantage.

Positive Externalities of Organisation Culture

267

Research demonstrates that people with learning disabilities are disenfranchised in many areas of life, including workplace democracy (Kazdin & Matson, 1981; Arns & Linney, 1993; Heyman et al., 1997; Mank, Cioffi, & Yovanoff, 1997). However, significant efforts have been made in recent years, particularly in advanced industrialised countries, to facilitate the inclusion of people with learning disabilities in the mainstream of society via the labour market. These labour market-based responses to the needs of people with disabilities in general seem to be underpinned by the belief that social integration (as reflected by degree of social acceptance and interaction,1 and by removal of disability-related discrimination and social barriers) is enhanced by access to work/employment.2 Traditionally, efforts to provide employment to people with disabilities have been based on vocational rehabilitation models delivered by state agencies, charitable organisations, and other community groups providing sheltered work and recently, supported employment in the open market. Cooperative organisations have also emerged as means through which people with disabilities access employment and therefore support the social integration of people with disabilities.3 Faced with changes in employment legislation and labour market forces, private sector employers in western industrialised countries (Ireland and UK included) are increasingly supporting the workforce integration of people with disabilities. However, doubt is raised in current research about the extent to which economic organisations adequately support the social integration of the most marginalised in society such as people with disabilities. Therefore, there is need to establish precisely what options exist in organisational life for a person with learning disabilities for example. Furthermore, labour market insertion does not guarantee curtailment of disability-related social exclusion, in particular exclusion from active participation in workplace democracy. Therefore, there is need to establish how organisations of varying structure and orientation help people to tackle social network poverty. Examining economic democracy based on the experiences of atypical workers such as workers with learning disabilities uniquely enriches our current understanding of the positive externalities of participation in the workplace onto an individual’s experiences of life outside work. We examined social acceptance and interaction data from 105 working adults with mild to moderate learning disabilities, to establish the link between organisation culture and workplace social integration. Results of this study suggest that the social integration of organisation members, in particular working adults with learning disabilities, might be significantly influenced by difference in organisation culture. This implies that organisations

268

ABRAHAM MUKOLO ET AL.

can be structured to empower their members (particularly those experiencing social disadvantage) for integration in the wider society.

2. TRENDS IN WORK-BASED SOCIAL INTEGRATION RESEARCH Most contemporary research into work-based social integration of people with learning disabilities seems to prioritise strategies for ensuring that people with learning disabilities access and maintain employment in the open labour market, with some level of on-the-job support provided by service agencies (Kazdin et al., 1981; Shafer, Kregel, Banks, & Hill, 1988; Ferguson, McDonnell, & Drew, 1993; Rusch, Wilson, Hughes, & Heal, 1995; Kilsby & Beyer, 1996; Chadsey-Rusch, Linneman, & Raylance, 1997; Hughes, Kim, & Hwang, 1998). For example, Kilsby and Beyer’s (1996) research is typical of approaches to services provision that seek to encourage the development of marketable competencies among people with disabilities. Their study focuses on the extent to which workers with disabilities master motor skills associated with a job in the workplace and the extent to which they succeed in social encounters with non-disabled co-workers and supervisors. That is, how well workers with disabilities are able to meet the demands of the workplace independent of support and how well they are able to function in the social contexts that exist within these workplaces. However, it has been noted in some studies that supported employment jobs, which are accessed by people with learning disabilities in the open labour market, tend to be largely entry-level or menial-type jobs that offer little, if any, decision-making latitude to incumbents (Mank et al., 1997; Kilsby & Beyer, 1996). Since most of these workers with disabilities would have progressed from sheltered work settings that traditionally tend to engage participants in menial-type activities, the type of work they tend to get in open work settings therefore, seems to differ only marginally from that offered in sheltered workplaces. Perhaps, the challenges that people with learning disabilities encounter in open work settings could be related more to coping with social encounters with non-disabled co-workers and adjusting to the commercial orientation of those organisations, than the type of work they do. This might explain the concentration of research efforts on workplace social integration, particularly on adults with disabilities developing ‘appropriate’ social skills and adapting to the work environment. A substantial body of research on workplace social integration has demonstrated the significance of social skills to the social integration (in

Positive Externalities of Organisation Culture

269

particular, employability) of people with learning difficulties in the open labour market, regardless of the type of jobs accessed (Ferguson et al., 1993; Kilsby & Beyer, 1996; Mank et al., 1997; Hughes, Kim, & Hwang, 1998). However, the decision-making latitude of each job does not seem to have attracted much research interest in the field of the employment of people with learning difficulties, yet most organisation behaviour and human resources management literature suggests that widening the decision-making latitude of employees (e.g., through job enrichment, multi-skilling, selfmanaging work teams, employee participation, de-layering of organisation structures, worker control or democracy) is of strategic importance both for the improved responsiveness of organisations to market pressures (e.g., the pressure to sustain innovation, competitiveness, maintain market leadership, etc.) and also for worker motivation and productivity (Nightingale, 1982). It is therefore surprising that such concerns have not been widely seen as relevant for research into the employment of people with learning difficulties. Perhaps it has been considered premature for research to focus on these aspects of the employment of people with learning disabilities when there are still major challenges to accessing and maintaining even menialtype jobs.4 Nevertheless, self-determination and self-advocacy values increasingly underpin current approaches to services provision in the area of learning disabilities in western industrialised nations (Williams & Schoultz, 1984; Kroese, 1997; Chadsey-Rusch et al., 1997; Chappell, 2000; Chappell et al., 2001). This indicates that engagement in decision-making, by people with learning disabilities, about issues central to their lives is increasingly being prioritised in the learning disabilities sector. Since work organisations are microcosms of larger society, the changes in the learning disabilities sector could be regarded as reflecting changes in the values and principles cherished by the larger society, more so in western nations, where the wider society has lost confidence in the ability of public institutions such as disability support organisations and government agencies to meet public expectations (Nightingale, 1982). Reflecting on the rate of adoption of societal values by organisations David Nightingale (1982, p. 9) observes, ‘The work organisation cannot stand outside the realm of values, principles, and practices followed in other areas of society. Instead, the workplace should reinforce the values of society’. In western industrialised nations that have embraced political democracy the societal values that one would expect to be reflected in work organisations are those that in Nightingales’ (1982, p. 9) words ‘‘broaden and deepen democratic principles, and reinforce practices and beliefs about fair play, due process, and human equality’’. However, it has been noted that undemocratic practices continue to dominate work organisations in

270

ABRAHAM MUKOLO ET AL.

the private sectors of western industrialised nations under the guise of economic and productive efficiency (Nightingale, 1982; Jermier, 1998).5 It would appear that most disability support organisations that are pursuing the integration of adults with learning disabilities in the labour-force in these nations endeavour to ensure the integration of their clients even in workplaces that do not reflect democratic principles that society has increasingly compelled them to embrace. Conversely, there seems to be little demand from the learning disabilities sector in western industrialised nations that workplaces in the mainstream should of necessity reflect societal values of self-determination and self-advocacy. Workbased integration supports are therefore about drafting adults with learning disabilities into existing work-organisations with the latter not transforming significantly in harmony with the wider society. Therefore, the adult with a learning disability, like her/his non-disabled co-worker, is likely to ‘‘experience a vague and imperfectly articulated sense that a contradiction exists between our social values of individual expression, freedom, and initiative6 and the values of obedience and subordination in the workplace’’ (Nightingale, 1982, p. 6). Nonetheless, one notes that within segregated forms of employment (mainly enterprise structures that are either worker owned and controlled or are subsidiaries of disability services provider organisations) opportunities are increasingly being created for workers with learning disabilities not only to engage in paid and meaningful employment but also to secure roles that are endowed with a substantial decision-making latitude in accordance with values and principles that the wider society has embraced (Neufeldt & Albright, 1995; Neath & Schriner, 1998). This social integration study therefore extends beyond the boundaries of contemporary research in the learning disabilities sector and in workplace democracy in general by examining the extent to which involvement in organisational decision-making enhances workplace social integration and how this workplace experience spills over into social integration in nonwork domains of life.

3. METHODS It is generally acknowledged that organisation culture is reflected in the most cherished values, symbols, rituals, and other artefacts of organisations and in assumptions made about the way things are or aught to be (Davies et al., 2000; Kabanoff, Waldersee, & Cohen, 1995; Hofstede, 1994, 1998). Cultures of eight organisations that provide employment to 105 adults with learning disabilities in Ireland and the UK were estimated from an

Positive Externalities of Organisation Culture

271

analysis of mission statements, and from participants’ perceptions of values and practices (attitudes) of leaders in these organisations. Variable clustering was used to generate 10 clusters of organisation values espoused in 37 mission statements, and six clusters of leaders’ values and practices reported by adults with learning disabilities. Mission statement analysis was used primarily to test the reliability of organisation cultures that were derived from perceptions held by adults with learning disabilities.7 We then examined the association between organisation orientation and organisation status. By status we refer to the legal status of an organisation. Four types of organisation status were observed among the eight organisations from which participants were drawn, i.e., workers co-operatives, community associations, foundation/trusts, and organisations affiliated to religious institutions. These tend to be the main types of organisations that provide supports to adults with learning disabilities in Ireland and the UK. In the Irish context in particular, religion has been a key element in shaping the cultures of disability support organisations but in recent years its role has diminished with the secularisation of the Irish state. Therefore, it was important to distinguish social enterprises that are affiliated to religious institutions from those that are not. Social integration is often perceived as a function of social acceptance and interaction (Calvez, 1993; Blanc, 1998). For example, social acceptance and interaction feature in studies of workplace social integration of people with learning disabilities (Ferguson et al., 1993; Chadsey-Rusch & Linneman, 1997; Chadsey-Rusch et al., 1998; Hughes et al., 1998), and in diseasecure and therapy-motivated research (Thernlund & Samuelsson, 1993; OrthGomer, Rosengen, & Wilhelsen, 1993; Villeneuve, Lebel, & Lambert, 1992; Redaelli et al., 1992; Montero-Perez, Perula-de Torres, Martinez-de la Iglesia, & Jimenez, 1991; Kroese, Dagnan, & Loumidis, 1997). Social integration scores were estimated from self-reported levels of social acceptance and interaction obtained from 105 adults with mild to moderate learning disabilities. Multi-predictor analysis was employed to examine the relationship between social integration and organisation culture, having regard to personspecific characteristics of participants such as age, sex, partner status, level of disability, and independent living status. We also controlled for duration in employment, number of hours worked per week, country of residence, and in some cases the industry sector that the organisation is located. This study considers three main outcomes, workplace social integration (WI), non-work social integration (nWI) and overall social integration (OSI). OSI ¼ WI þ nWI: WI ¼ SAw þ SIw; and nWI ¼ SA þ SI

272

ABRAHAM MUKOLO ET AL.

SAw stands for social acceptance at work and SA stands for social acceptance outside work. SIw stands for social interaction at work and SI represents social interaction outside work. SAw, SIw, SA, and SI were each generated by aggregating binary variables derived from a transformation of constituent ordinal (ordered categorical) variables. The means, standards deviation, and correlation coefficients of these variables are listed in the appendix (Tables A1–A3). Therefore, WI, nWI, and OSI variables are discrete interval data. These variables were analysed as if they are continuous variables (Woodward, 1999). Some components of either social acceptance or social interaction (work and non-work related) are discrete interval data too, since they are aggregates of several binary variables. For example, work–talk involvement (an aspect of workplace social interaction) reflects the proportion of issues that each participant discusses with others in the workplace out of nine issues listed in the social integration study (SIS) questionnaire. For each issue, each participant either scores 0 (not discussed) or 1 (discussed). The maximum possible score for work–talk involvement is therefore nine and minimum zero. Work–talk involvement information is therefore treated as discrete interval data and is analysed in the same manner as a continuous variable. On the other hand, predictor variables are a combination of: (i) Binary variables such as sex, partner status, and country of residence; (ii) Categorical (nominative) variables such as age-group (ordinal), organisation status (ordinal, since they reflect the continuum in service provision models from traditional ¼ 4 to workers co-op ¼ 1), independent living status (ordinal, since they reflect degree of independence in living status from institutional ¼ 1 to independent ¼ 4), and employment type (Sheltered employment ¼ 1 and open employment ¼ 3); and (iii) Continuous variables such as age, years employed, weekly work hours, and level of disability. Disability data are best defined as discrete interval data. One-way ANOVA (univariate analysis) was used to examine the impact of difference in organisation status (the main explanatory variable) on each of the three main outcome variables (WI, nWI, and OSI). Multi-predictor analysis was then employed to adjust for other influences than organisation status and orientation. Post hoc analysis using the Least Significant Difference (LSD) test was employed to examine the significance of differences between mean scores of each pair of the organisation status variables. The test is also used with other predictor variables of interest.8

Positive Externalities of Organisation Culture

273

All predictor variables in our model were treated as non-normal distributions in the entire analysis and therefore Spearman’s correlations were employed to examine associations among and between them as well as between them and outcome variables. When outcome variables did not satisfy the normality assumptions that underpin either linear regression or analysis of variance methods, the non-parametric equivalent of the one-way ANOVA (i.e., the Kruskal–Wallis test) was employed. This was applied mainly in the analysis of workplace social integration (WI) variables (which have an ordinal level of measurement) such as work-group membership, socialisation with workmates etc. The Kruskall–Wallis test examines whether several independent samples are from the same population. In our case, the Kruskal–Wallis test ranks observations per variable under four organisation status categories and then computes medians of this ranked data per organisation status category, referred to as mean ranks. The Kruskal–Wallis test enabled us to establish the correlation between organisation status and each of the outcome variables that did not follow a normal distribution. When using the Kruskall–Wallis test we were concerned with establishing a tendency rather than the influence of specific organisation status categories. Since all the three main outcome variables in this study (WI, nWI, and OSI) follow a normal distribution (tested using QQ plots of residuals from linear regression analysis), we have been able to test not only the tendency but also the influence of specific organisation status categories using the LSD test.9 The latter has enabled us to demonstrate a continuum of service provision models as mirrored in the pairwise comparison of difference in mean scores for each of the three main social integration variables (pairs being organisational status groups).

4. RESULTS 4.1. Social Integration in the Workplace The Kruskal–Wallis test was used initially to select components of social acceptance and interaction in both work and non-work settings that vary significantly with organisation status (Table 1). Participants in the workers co-op had the highest mean rankings in five out of seven workplace social interaction and social acceptance measures shown in Table 1. Respondents in the community association category had the highest mean rankings for satisfaction with being in charge. In one of these community associations, adults

274

ABRAHAM MUKOLO ET AL.

Table 1.

Social Integration Variable Means Ranked by Organisation Status Using Kruskall–Wallis Test.

Social Integration Variables

Organisation Status (Mean Rank)

Work talk involvement Work group membership Social events attendance Work time interaction Socialisation with workmates Workplace social interaction Confidence to engage Being in charge satisfaction Social support opportunity Work support satisfaction Socialisation opportunity Work interaction opportunity Work friendship closeness Work friendship satisfaction Workplace social acceptance Workplace social integration

p-Value

Worker coop

Community association

Foundation/ Trust

Religious institution

92.31 96.75 49.94 45.06 58.88

74.79 61.61 52.75 56.21 53.25

45.88 68.29 54.69 50.19 44.84

45.35 36.07 52.61 54.85 56.44

0.000 0.000 0.810 0.545 0.219

91.13

71.64

49.24

44.54

0.000

66.88 54.25

31.50 73.00

49.00 42.48

58.67 53.28

0.000 0.005

58.94

43.00

56.45

52.86

0.260

44.81

45.75

52.38

56.43

0.265

68.44 46.81

63.75 51.79

46.03 52.05

51.67 54.74

0.060 0.636

42.81

55.00

48.02

56.67

0.130

43.81

56.00

45.40

57.67

0.042

56.94

45.96

44.76

58.67

0.168

86.06

63.07

47.07

48.68

0.004

with disabilities are actively engaged in staff-recruitment, are represented on boards of management, and have an active workers committee. This might explain why they have scored high on satisfaction with being in charge. Respondents from social enterprises that are affiliated to religious institutions had the highest score for satisfaction with the quality of friendship at work. As indicated in Table 1, what makes the most difference in inter-organisational patterns of social integration within the workplace are work–talk involvement (p-valueo0.001); work-group membership (p-valueo0.001); workplace social interaction (p-valueo0.001); confidence to engage with others in the workplace (p-valueo0.001); satisfaction with chances to be in charge (p-valueo0.01); work friendship satisfaction (p-valueo0.05); and workplace social integration (p-valueo0.01). Combining work–talk involvement with

275

Positive Externalities of Organisation Culture

work-group membership gives an index of participation in workplace decision-making (Fig. 1). Workplace social integration is significantly correlated with organisation status (p-valueo0.05) when all possible mediating factors, such as age, sex, number of hours worked per week, total number of years employed, employment type, country of residence, industry sector, and level of disability, are considered (Table 2). Furthermore, when all factors are considered, difference in organisation status has a significant impact on participants’ integration in work domains of life mainly with regards to work–talk involvement (p-valueo0.01) and aggregate social interaction (p-valueo0.01) as shown in Table 2. LSD test (not shown) indicates, that of the four organisation status categories, the most significant difference in workplace social interaction scores are between workers co-ops and religious institutions. The difference between community associations and workers co-ops was of marginal significance in all aspects of social interaction. Participants in Foundation/ trust organisations had significantly lower scores than those from workers co-ops on most aspects of social interaction. Although foundation/trust organisation members in the study had lower scores than community association members, these differences were generally of marginal significance. Participation in Decision Making by Organisation Status

Mean Participation Level

15

10

5

0 0

1

2 Organisation Status

3

4

Fig. 1. Means of Participation in Decision-Making for each of the Four Organisation Types. Participation is the Sum of the Number of Decision-Making Groups each Participant Belongs to and the Number of Issues/Concerns that each Participant Discusses with Others in the Workplace as Part of the Decision-Making Process.

276

ABRAHAM MUKOLO ET AL.

Table 2.

Adjusted Means of Workplace Interaction and Social Integration by Organisation Status.

Organisation Status

Mean

95% CI

7.88 6.41 3.77 2.28

5.26–10.49 4.87–7.95 1.39–6.16 0.60–3.95

11.37 7.62 5.00 2.52

8.65–14.08 6.02–9.23 2.52–7.47 0.78–4.26

15.35 11.48 8.79 7.16

12.08–18.63 9.55–13.41 5.80–11.77 5.07–9.26

28.51 22.59 20.27 18.89

23.14–33.87 19.43–25.75 15.39–25.16 15.46–22.33

25.84 23.49 19.92 19.76

21.78–29.89 20.66–26.32 17.30–22.54 17.34–22.18

25.23 23.40 19.96 19.93

22.38–28.07 20.98–25.82 18.28–21.65 18.73–21.12

a

Work talk involvement Worker co-op Community association Foundation/Trust Religious institution Participation in decision-makinga Worker co-op Community association Foundation/Trust Religious institution Workplace social interactiona Worker co-op Community association Foundation/Trust Religious institution Workplace social integrationa Worker co-op Community association Foundation/Trust Religious institution Workplace social integrationb Worker co-op Community association Foundation/Trust Religious institution Workplace social integrationc Worker co-op Community association Foundation/Trust Religious institution

p-Value 0.001

0.000

0.001

0.046

0.043

0.001

a

Adjusted for age, sex, disability, country of residence, industry sector, employment type, hours worked per week, and years in employment. b Adjusted for age, sex, disability, country of residence, employment type, hours worked per week, and years in employment. c Adjusted for age, sex, and disability.

Conversely, the impact of organisation status on social acceptance in work domains of life (Table 2) appears to be attenuated by personal characteristics of participants, in particular disability level, since the proxy for ‘level of disability’ is the only factor that was marginally correlated with workplace social acceptance (p-value ¼ 0.075) independent of organisation

Positive Externalities of Organisation Culture

277

status (results not shown). This implies that perhaps ‘learning disability’ has the effect of impairing participants’ perceptions of how they are accepted by others, regardless of the nature of work organisation they are associated with. 4.2. Social Integration Outside Work Organisation status and non-work social integration are positively correlated (Table 3). However, this correlation seems to be marginally significant (p-value ¼ 0.076, unadjusted and p-value ¼ 0.165 adjusted). Non-work social integration seems to be more strongly associated with disability level (p-valueo0.05), and independent living status (p-valueo0.05) than organisation status. 4.3. Overall Social Integration Multi-predictor analysis was also conducted to examine the effect of difference in organisation status on overall social integration (i.e., nWI+WI), taking other factors into consideration. Results shown in Table 3 indicate that organisation status correlates significantly with overall social integration when all relevant personal and organisation-specific factors are considered. The LSD test shows that whether or not additional predictors are considered, overall social integration scores obtained by worker co-op members are significantly higher than those obtained by users in foundation/trusts and religious institutions, but differ marginally from those obtained by community association members. In contrast, difference in scores among non-co-op social enterprises is of marginal significance. We found that disability level is also strongly correlated with overall social integration (p-valueo0.05). However, independent living status is marginally associated with overall social integration (p-value ¼ 0.07).

4.4. Organisation Orientation/Values Rather than relying on organisation status (i.e., legal status of organisations), differences in organisation orientations (values) could be more enlightening, as they reflect aspects of the work setting that workers with learning disabilities have to be mindful of. These workplace characteristics might influence the extent to which these workers interact with and feel accepted by others in the workplace.

278

ABRAHAM MUKOLO ET AL.

Table 3. Organisation Status Non-work social integration Univariate Worker co-op Community association Foundation/Trust Religious institution Multi-predictora Worker co-op Community association Foundation/Trust Religious institution Multi-predictorb Worker co-op Community association Foundation/Trust Religiousinstitution Multi-predictor c Worker co-op Community association Foundation/Trust Religious institution Overall social integration Univariate Worker co-op Community association Foundation/Trust Religious institution Multi-predictora Worker co-op Community association Foundation/Trust Religious institution Multi-predictorb Worker co-op Community association Foundation/Trust Religious institution Multi-predictor c Worker co-op Community association Foundation/Trust Religious institution a

Non-Work and Overall Social Integration by Organisation Status. Marginal Means

95% CI

18.50 15.00 13.79 15.24

15.19–21.82 12.83–17.17 12.29–15.30 14.14–16.35

18.60 15.75 14.57 15.17

15.53–21.66 13.66–17.84 13.10–16.05 14.06–16.28

19.35 15.51 15.53 16.25

16.07–22.62 13.25–17.78 13.59–17.48 14.74–17.76

20.03 15.09 14.95 15.59

16.30–23.75 12.57–17.61 12.49–17.41 13.29–17.88

46.17 36.79 32.55 34.82

40.49–51.85 33.07–40.51 29.97–35.14 32.92–36.71

46.05 38.24 34.38 34.84

41.05–51.05 34.83–41.65 31.98–36.78 33.03–36.63

47.32 38.50 34.18 35.07

40.94–53.69 34.10–42.89 29.89–38.47 31.17–38.96

52.29 38.59 35.30 34.44

43.32–61.26 33.39–43.80 27.54–43.06 29.11–39.76

p Value

0.076

0.126

0.142

0.165

0.00

0.00

0.01

0.02

Adjusted for age, sex, and disability level. Adjusted for age, sex, disability, partner status, independent living status, and country. c Adjusted for age, sex, partner status, independent living status, country, employment type, weekly work hours, years employed, and industry sector. b

Positive Externalities of Organisation Culture

279

Mission statements of 37 disability support organisations affiliated to the Federation of Voluntary Bodies (FVB) were obtained from this organisation’s website.10 Key words were identified from each of the 37 mission statements and/or statements outlining principal objectives. A weighting system was used to generate scores for each key word. Weighted scores were obtained by dividing the number of times each word/term appeared in the text by the total number of words in the whole statement. These figures were then converted into percentage values by magnifying them a hundred-fold, it was felt that percentage values would be large enough to be analysed statistically.11 A ‘cluster variable’ analysis (using the MINITAB 13 for Windows statistical package) yielded the following 10 organisation orientations by clustering 35 variables (key words) extracted from an analysis of these mission statements (Table 4).12 A dendogram is also given in Fig. 2 to illustrate the manner in which key words cluster. Short versions of key words are given in the dendogram for convenience purposes. Recognising the validity of the world view held by adults with learning disabilities, all the 105 respondents were asked to state the level of importance that they feel senior management in their respective organisations gives to a number of organisation orientation factors. These factors included among others ‘making sure every person is treated the same as others’, ‘following the rules of the organisation/workshop’, ‘making sure things are produced on time all the time’, ‘making sure everyone has a say in what goes on around here’, etc.13 Respondents were asked to rank these factors according to whether the item is ‘Very Important’, ‘A little bit Important’, and ‘Not Important’. There was also a provision for respondents to state that they were not sure or they did not know enough to grade the attitudes of their superiors. For most participants these questions proved to be the most difficult of all questions in the SIS Questionnaire, as anticipated. In most cases, what was being sought after was explained to participants as the extent to which superiors fuss about these issues or make a point to emphasise them on a regular basis. In most cases, not being able to remember after ample explanations had been given was the basis for ranking the item as not fussed about and therefore not important. These responses were then assigned numeric values, ‘2’ for ‘Very Important’, ‘1’ for ‘A Little Bit Important’ and ‘0’ for ‘Not Important’. Based on the above logic, all the ‘I Don’t Know’ responses were also assigned the value ‘0’. In order to avoid bias, responses were analysed using a computer package. Having tried a number of cluster levels ranging between 2 and 10,

280

ABRAHAM MUKOLO ET AL.

Table 4. Cluster Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 Cluster 7 Cluster 8 Cluster 9

Cluster 10

Organisation Orientation in the Irish Learning Disabilities Sectora. Key Factorsb

Acceptance, creativity, ethics, community and loyalty to traditions Accountability, comprehensiveness, and innovativeness Advocacy, individualism, dignity of person, rights, and health Caring, diversity, serving the family, and responsiveness Choice, Friendship, loving relationships, and social ties Commitment, service independence, and needs driven Contribution, volunteering and selfsacrifice, respect Equality, independence of clients, and justice Competency (organisation and staff), honesty and currency (up-to-date with knowledge and research developments) Excellence or excelling, high standards of quality sharing.

Orientation Traditional and community (normative) Organisation growth focused Individual rights and well-being Family care Amity Service autonomy (flexibility and pragmatism) Productivityc Justice and equality Knowledge (emphasises professionalism) Quality (competitive)

a Crossed items in Table 4 do not, in the opinion of the authors, fit within the categories that the computer has assigned them to. Kabanoff et al. (1995) observed a tendency by organisations to include among espoused values certain values that compensate for a presumed shortcoming in an established orientation or mission. Therefore, the crossed items are considered to be compensatory values per given orientation. For example, family care orientation has its roots in antiquity and generally falls within orientations that have recently been criticised for not being responsive to the changing needs of clients or for prioritising the family over the adult with a disability. It therefore makes logical sense that the term ‘diversity’ would communicate new thinking within this established orientation and thereby compensate for this orientation’s shortcomings. The same could be said about choice within an amity-oriented social enterprise, and sharing (communicating values of non-competition) within organizations that prioritise excelling above all others in terms of service quality (competitive values). b 35 Cluster Variables: ‘ACCEPT’ (c8)-‘NEEDRIVE’ (c43). c Probably upholds the ‘dignity of labour’ philosophy.

six meaningful clusters of variables were obtained from Variable Clustering using the MINITAB 13 for Windows program as shown in Table 5.14 These clusters were based on between and within variable variations, i.e., the extent to which respondents’ scores varied per variable and the variation in mean scores across all 13 variables.

281

Positive Externalities of Organisation Culture

-41.28

Similarity

5.81

52.91

100.00 PT AL OY ITY ITY ICE IES HIP ING ING ITY OR SE CY TY CY NT ION HE NT EP IVE NC ITY ING CY ITY TH ISM TS IBU CT OL LTY ICE EP CE HIC DIL TIV UN HO CIT DS OV AR RS ILY ON ET ES EN OU AT RE ME IND DR LE AL AR CA GN EAL IL IGH TR PE RIV UA ST IND AC ET RA EA OM C SO IEN L C IVE AM ESP MP HON RR CC OV MP MIT EV EE XCE QU SH VO DI H IND R ON RES AC EQ JU CL T R C S C D F R CO AD CU A INN CO CO S N E C FR

Variables

Fig. 2. Clusters of Key Words in Mission Statements of Non-Co-op Social Enterprises.

Table 5. Cluster Number 1

2

3 4 5

6

Organisation Orientations as Perceived by Adults with Learning Disabilities. Organisation Orientation Variable in Cluster

‘Being aware of the special needs of people’; ‘Making sure everyone is accepted’; and ‘Making sure every person is treated the same as others’ ‘Support/help from management and supervisors’; ‘Support from workmates during work time’; ‘Making sure everyone has a say in what goes on around here’; and ‘Making sure people are praised for what they do’ ‘Going out and meeting others after work’, and ‘Giving people more responsibility as time goes’ ‘Making sure everyone is paid properly for the work they do’ ‘Making sure customers are happy with the things done for them’ and ‘Making sure things are produced on time all the time’ ‘Following the rules of the organissation/workshop’

Orientation Type

Social justice

Supportive/Caring

User autonomy Equity oriented (reward) Productivity

Authority

Of the six orientations obtained through variable clustering, the only orientation scores that vary significantly with organisation status (based on the Kruskal–Wallis Test) are social justice (p-valueo0.05), social support (pvalueo0.01), user autonomy (p-valueo0.001), and productivity orientation

282

ABRAHAM MUKOLO ET AL.

(p-valueo0.001). No significant correlation was found between organisation status and reward and authority orientations (p-value ¼ 0.430 and 0.115, respectively). We found that four organisation orientations vary by organisation status among the eight organisations providing employment to adults with learning disabilities as indicated in Fig. 3. Means rankings depicted in Fig. 3, suggest that, while all four orientations are present in each organisation, the strength of each orientation is highly likely to be influenced by organisation status. Worker co-ops in our study seem to be able to encourage high levels of user/member autonomy without sacrificing social support and social justice orientations and without slackening emphasis on attaining high standards of achievement in both customer service and meeting production targets (i.e., productivity orientation). The difference between the worker co-op and the community association is that the latter is seen as emphasising user-autonomy and social justice but puts modest emphasis on social support and productivity. A detailed case study of one community association (done elsewhere) confirms these observations. For example, this particular association encourages users to be involved in decision-making within the workplace. Secondly, an administration team centrally controls labour for three subsidiary social enterprises of the association and allocates this Ranked Means of Organisation Orientation per Organisation Status Productivity Orientation 110

Organisation Status : 1= Worker Co-op 2= Community Association 3 = Foundation/ Trust 4= Religious Institution

Autonomy Orientation Social Support Orientation Social Justice Orientation

90

Productivity Orientation

70

50

30 0

1

2

3

Organisation Status

Fig. 3.

Levels of Organisation Orientation Reported by Participants.

4

Positive Externalities of Organisation Culture

283

labour by criteria that is not entirely productivity oriented. Therefore, it is possible that increased user-autonomy (as in users taking a lead role in some activities) in this community association is seen by users as a reduction in social support.

4.5. The Spill-Over We assumed that the spill-over between work and non-work domains of life exists if there is significant correlation between workplace social integration (WI) and non-work social integration (nWI). A general linear model in which non-work social integration is determined by workplace social integration, employment type (i.e., open or sheltered employment), independent living status, partner status, years one is employed, hours worked per week, age, sex, disability level, and country of residence of each participant was examined. Both Multi-predictor and linear regression analysis was conducted using SPSS 10.0 for Windows. Firstly, the model was significant (p-valueo0.001 multi-predictor, and p-valueo0.001 linear regression). Both analysis showed that nWI and WI were positively correlated and that the correlation was statistically significant (p-valueo0.001, regression coefficient ¼ 0.368).15

5. DISCUSSION According to this study, levels of participation in decision-making processes and overall social interaction in the workplace seem to be significantly dependent on both organisation characteristics (industry type and legal status) and conditions of work (in particular, employment type). Magnitude of learning disability seems to be of marginal significance in patterns of self-reported social interaction and participation in decision-making processes. However, self-reported social acceptance in the workplace seems to vary more by difference in personal characteristics of working adults with learning disabilities (such as level of disability) than organisation characteristics. Focusing primarily on legal status, our analysis suggests that the main component of workplace social integration that significantly distinguishes worker co-op members from all other participants is the level of involvement in discussing management and control-type issues in the workplace. Evidence presented in Tables 2 and 3 and Fig. 1 indicate that worker co-op

284

ABRAHAM MUKOLO ET AL.

members are the most socially integrated in the workplace by virtue of their direct involvement in decision-making activities/roles at levels that exceed those attained by members of all other social enterprises. Community association members in our sample also score high on workplace social integration by virtue of their direct involvement in decision-making structures and processes albeit to a lesser extent than worker co-op members. Although non-formal contact and undertaking of issues seems more marked in religious institution affiliates and foundations than in community associations, participants from foundations and religious institution affiliates consistently attain lower scores on most indicators of integration in the workplace than community association members. Overall, therefore, people in community associations closely match the magnitude of social integration attained by worker co-op members. Furthermore, though people in religious institutions persistently attain higher integration scores than attained by people in foundations/trust organisations, their scores vary marginally from those of the latter. Therefore, it is possible that community associations have something in common with worker co-ops, primarily their emphasis on democracy and workers rights. Likewise, social enterprises affiliated to religious institutions might have something in common with foundations/trust organisations, primarily their emphasis on care/social support and limited emphasis on democracy and workers rights. The community association also shares some common features with foundation/trust organisations and religious institutions that have little to do with practices in the workers co-op, in particular uncertainty about the achievability and efficacy of workplace democracy. Therefore, a continuum in which the worker co-ops and either religious institutions or foundation/trust organisations occupy opposite ends is implicated by these results. Based on this sample of organisations, working adults with learning disabilities are more likely to be integrated in the decision-making structures and processes of enterprises that emphasise user-involvement in key decision-making, and social justice/equality values than in enterprises that prioritise care-oriented approaches to service provision. This is less likely to be dependent on the characteristics of adults with learning disabilities, their country of residence, and the length of time they have been associated with the enterprises in question. However, while there is some link between organisation status and social integration in non-work domains of life, there is a limited basis for assuming that the legal status of social enterprises is of marked significance in explaining difference in the non-work social integration experiences of adults

Positive Externalities of Organisation Culture

285

with disabilities that took part in this study. Rather, it would appear that differences in independent living status and in degree of learning difficulty (crude as the disability measure we used seems to be) are better explanatory factors for variation in non-work social integration than differences in the legal status of organisations.

6. CONCLUSION According to this study, adults with learning disabilities who work in organisations that prioritise user/member autonomy and emphasise egalitarian values are highly integrated in both work and non-work domains of life. Organisation culture (indicated by organisation status and/or orientation) appears to be significantly associated with both overall social integration and workplace social integration, independent of the personal characteristics of adults with learning disabilities (such as age, sex, degree of disability, independent living status, and partner status), the length of time they are associated with the employer organisation, and difference in country of residence. Therefore, when legal status is the main explanatory variable, there is a high likelihood that organisation cultures in which user-involvement in management and control decision-making is emphasised (e.g., the workers co-op in this study) engender a positive influence on the social interaction experiences of members with learning disabilities in work and non-work domains of life, having regard to differences in demographic factors, employment characteristics, country of residence and level of disability.16 However, while a spill-over between work and non-work domains of life is observed, person-specific factors such as level of disability and independent living status had significant impact on integration in non-work domains of life independent of difference in the work settings from which participants were drawn. Overall, therefore, these results point to the significance of the co-op model of organisation (either co-op legal structure and/or increased user control of the organisation) in the social integration experiences of working adults with learning disabilities. The positive externalities of organisation culture highlighted in this study imply that work organisations can be structured to empower their members (particularly those experiencing social disadvantage) for integration in the wider society. The study also accentuates the importance of workplace democracy (Bernstein, 1976; Nightingale, 1982; Greenberg, 1986; Warhurst, 1998) in enhancing the quality of life of working adults with learning disabilities.

286

ABRAHAM MUKOLO ET AL.

NOTES 1. See discussion on disability inclusion by Northway (1997). 2. The European Council endorsed this view at the March 2000 Summit in Lisbon by declaring that ‘(t)he best way to tackle social exclusion is through a job’. The new EQUAL programme of the European commission is underpinned by this philosophy. While these efforts are beginning to impact positively on the general society’s responses to disability, significant social barriers still exist as evidenced by the lived experiences of some among the 105 adults with learning disabilities, who reported, in confidence, the barriers they have to contend with on a regular basis. 3. Co-ops, in particular workers co-ops, have been identified as one form of selfdirected employment in addition to sole proprietorship (Neufeldt & Albright, 1995). Some studies have noted that most disability employment supports in advanced industrial countries tend to shy away from self-directed employment as a career option for people with disabilities. In an ‘international study of strategies leading to self-directed employment’, Aldred H. Neufeldt and Alison L. Albright (1995) observed that there is more experience of self-directed employment among people with disabilities in low/middle-income countries than in high-income countries. They also noted the limited amount of published literature on self-directed employment of people with disabilities. See also Johann Gudmundsson (1985) justification for the global promotion of co-ops among disabled people by the United Nations. 4. There have been moves and an expectation in the US, for example, that sheltered work programmes would be completely transformed into supported employment programmes (Albin et al., 1994). The changeover has occurred in a few cases but in others sheltered employment has continued to exist alongside supported employment. In Ireland, supported employment tends to be an add-on to traditional supports that include sheltered employment. 5. Commentating on crisis in work organisations of major Western democracies, Donald V. Nightingale (1982, p. 6) observed, ‘A significant component of the increasing problems in the workplace is the contradiction between the values celebrated in the larger society and the values underpinning the workplace.y In the modern organisation, the employee’s freedom is suspended in many important respects, justice is limited, obedience to superiors is demanded, and the workplace is permeated with symbols of authority, deference and subordination’.’’ 6. As experienced in the disability support organisation that has embraced such societal values and principles. 7. A detailed description of the methodology employed to generate the data reported in this paper is available elsewhere (PhD thesis by the first author). Obtaining such rich qualitative data from the learning disabilities sector, that could significantly inform mainstream economics and management studies, was not without its own challenges. There were ethical issues to be tackled (mainly raised by non-disabled professionals and custodians of study participants), and issues of questionnaire design (the content and structure of the SIS questionnaire), administration, reliability of observation/information obtained, and data analysis methods. Our sample is drawn from a population of adults, who traditionally would have limited opportunities to make even basic decisions such as deciding what to eat, wear, who to socialize with, etc.

Positive Externalities of Organisation Culture

287

8. LSD is a multiple comparison test to investigate/reveal groups that are significantly different from each other. In our case, for example, we can examine whether means for worker co-ops are significantly different from those of religious institutions, or compare worker co-op means and foundation/trust organisation means to establish whether they are significantly different or not. Alternatively, inference can be drawn from 95% Confidence Intervals (CI) of means of each group. If two CI intersect it implies that the two groups are not significantly different and vice versa if the two CI do not intersect. 9. Tendency is reflected by correlation coefficients. Influence is inferred from LSD tests. As noted, ‘An Influential value is one which has a serious impact on the regression parameter estimates – that is where the line sits for simple linear regression’ (Woodward, 1999, p. 432). 10. www.fed-vol.com/public/3_fvb/members/mem19.htm. Federation members generally supply information about themselves (i.e., who they are, their main aims and objectives, their mission, and their activities) on the Federation’s website. For example, jobs for this sector are often advertised on this site, so are notices of major upcoming events like conferences, seminars, etc. The website has also been a useful conduit for disseminating information concerning the most recent and far-reaching public sector reforms to affect the learning disabilities sector such as the dissolution of the National Rehabilitation Board, the transfer of parts of the disability support responsibilities from the Department of Health and Children to the Department of Enterprise, Trade and Employment, etc. Information on this site is intended for public consumption. Since the orientations being sought after are not of specific organizations, it was felt that the information on this site was reliable enough to provide a general picture of organization orientations that exist in this sector. 11. For more on using content analysis to distil structures of organisation values see Kabanoff et al. (1995). 12. For example, the number of missing values had to be less than the total number of key words. According to the scores generated, the number of missing values exceeded the number of key words if more than 37 key words were considered in the analysis. 13. Questions in this section of the questionnaire were originally intended to explore organization orientations to ‘elite leadership’ tendencies (Q.20m); democratic systems of decision-making; market orientation (Q.20k); attitudes to justice and equality; and production orientation. These are based on Geert Hofstede’s (1994) typology of organisation orientations, the values structure analysis undertaken by Kabanoff et al. (1995) among Australian organisations and the ‘levels of organisational culture’ concept used by Davies et al. (2000) to distinguish cultures within the health services sector in Britain. 14. The cluster variables method bunches together variables that are ‘close’ to each other when groups are initially unknown. The amalgamation is done in a hierarchical manner with observations closest to each other being joined. We used the Wards linkage (the distance between two clusters being the sum of squared deviations from points of centroids) to calculate inter-cluster distances. Ward’s linkage minimises the within cluster sum of squares (MINITAB Software Package Version 13.0). 15. The significance of the correlation is marginally lost when industry sector and organisation status are factored into the multi-predictor model, the p-value increases from 0.002 to 0.059. Therefore, this spill-over is not significantly impacted by

288

ABRAHAM MUKOLO ET AL.

industry sector and organisation status. However, there are other predictors of the spill-over than difference in employer organisations and industry sector. For example, non-work social integration is marginally correlated with level of disability (p-value ¼ 0.037), sex (p-value ¼ 0.046), and independent living status (p-value ¼ 0.048), when the correlation with workplace social integration is considered as well. Among workplace social integration variables, for example, non-work social integration is only significantly correlated with work–talk involvement (p-valueo0.001; regression coefficient ¼ 0.778) and work-group membership (p-value ¼ 0.04; Spearman’s correlation ¼ 0.208). Combining these two workplace social integration measures one obtains an estimate of magnitude of members’/ users’ participation in decision-making, i.e., the number of issues each user actively discusses with others in the workplace and the number of decision-making bodies each user belongs to in the workplace. Participation in decision-making seems to correlate significantly with non-work social integration (p-valueo0.001; regression coefficient ¼ 0.875) despite difference in workplace settings and personal characteristics of participants. While participation in decision-making varies significantly with organisation status (p-valueo0.001), it also varies with independence in living status (p-value ¼ 0.025). People that live independently tend to be more engaged in decision-making in all organisations than those living under supervised conditions. 16. Without ignoring the strong impact of degree of learning difficulty and independence in living status suggested by our data.

REFERENCES Albin, J. M., Rhodes, L., & Mank, D. (1994). Realigning organizational culture, resources, and community roles: Changeover to community employment. Journal for the Association of Persons with Severe Handicaps, 19(2), 105–115. Arns, P. G., & Linney, J. A. (1993). Work, self, and life satisfaction for people with severe and persistent mental disorders. Psychosocial Rehabilitation Journal, 17(2), 63–79. Aryee, S., Fields, D., & Luk, V. (1999). A cross-cultural test of a model of the work-family interface. Journal of Management, 25(4), 491–511. Bailey, J., Schermerhorn, J., Hunt, J., & Osborn, R. (1986). Managing organisational behavior (2nd ed.). Auckland: Wiley. Bernstein, P. (1976). Workplace democratization: Its internal dynamics. Kent, Ohio: The Comparative Administration Research Institute, The Kent State University Press. Blanc, M. (1998). Social integration and exclusion in France: Some introductory remarks from a social transaction perspective. Housing Studies, 13(6), 781–792. Blau, G. (1999). Testing the longitudinal impact of work variables and performance appraisal satisfaction on subsequent overall job satisfaction. Human Relations, 52(8), 1099–1113. Calvez, M. (1993). Social interactions in the neighborhood: Cultural approach to social integration of individuals with mental retardation. Mental Retardation, 31(6), 418–423. Chadsey-Rusch, J., Linneman, D., & Raylance, B. J. (1997). Beliefs about social integration from the perspective of persons with mental retardation, job coaches, and employees. American Journal of Mental Retardation, 102(1), 1–12.

Positive Externalities of Organisation Culture

289

Chadsey-Rusch, J., Gonzalez, P., Tines, J., & Johnson, R. (1998). Social ecology of the workplace: Contextual variables affecting social interaction of employees with and without mental retardation. American Journal of Mental Retardation, 94(2), 141–151. Chappell, A. L. (2000). Emergence of participatory methodology in learning difficulty research: Understanding the context. British Journal of Learning Disabilities, 28(1), 38–43. Chappell, A. L., Goodley, D., & Lawthom, R. (2001). Making connections: The relevance of the social model of disability for people with learning difficulties. British Journal of Learning Disabilities, 29(2), 45–50. Chatman, J. A., & Jehn, K. A. (1994). Assessing the relationship between industry characteristics and organisational culture: How different can you be? Academy of Management Journal, 37(3), 522–553. Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being different yet feeling similar: The influence of demographic composition and organizational culture on work process and outcomes. Administrative Science Quarterly, 43(December), 749–780. Cohen, A. (1997). Non-work influences on withdrawal cognitions: An empirical examination of an over looked issue. Human Relations, 50(12), 1511–1536. Cohen, A., & Vigoda, E. (1998). The growth of value of good citizenship: An examination of the relationship between civic behavior and involvement in the job. Applied Psychology: An International Review, 47(4), 559–570. Davies, H. T. O., Nutley, S. M., & Russell, M. (2000). Organisational culture and quality of health care (special article). Quality in Health Care, 9, 111–119. http://qhc.bmjjournals.com/cgi/content/full/9/2/111 Ferguson, B., McDonnell, J., & Drew, C. (1993). Type and frequency of social interaction among workers with and without mental retardation. American Journal of Mental Retardation, 97(5), 530–540. Greenberg, E. S. (1986). Workplace democracy: The political effects of participation. Ithaca: Cornell University Press. Gudmundsson, J. (1985). Cooperatives of disabled persons: A guide for promotion and organisation. UN: Committee for the Promotion of Aid to Cooperatives (COPAC) and Centre for Social Development and Humanitarian Affairs of the United Nations Department of International Economic and Social Affairs. Hart, P. M. (1999). Predicting employee life satisfaction: A coherent model of personality, work and non-work experience, and domain satisfaction. Journal of Applied Psychology, 84(4), 564–584. Herzberg, F. (1968). One more time: How do you motivate employees? Harvard Business Review (January–February), 46(57), cited in Bailey, J., Schermerhorn, J., Hunt, J., & Osborn, R. (1986). Managing organisational behavior (2nd ed.). Auckland: Wiley. Heyman, B., Swain, J., Gillman, M., Handyside, E. C., & Newman, W. (1997). Alone in the crowd: How adults with learning difficulties cope with social network problems. Social Science Medicine, 44(1), 41–53. Hofstede, G. (1994). Cultures and organisations – software of the mind: Intercultural cooperation and its importance for survival. London: Harper-Collins Business. Hofstede, G. (1998). Attitudes, values and organisation culture: Disentangling the concepts. Organisation Studies, 19(3), 477–492. Hughes, C., Kim, J. H., & Hwang, B. (1998). Assessing social integration in employment settings: Current knowledge and future directions. American Journal of Mental Retardation, 103(2), 173–185.

290

ABRAHAM MUKOLO ET AL.

Jermier, J. M. (1998). Introduction: Perspectives on organizational control. Administrative Science Quarterly, 43, 232–256. Kabanoff, B., Waldersee, R., & Cohen, M. (1995). Espoused values and organisational change themes. Academy of Management Journal, 38(40), 1075–1104. Kazdin, A. E., & Matson, J. L. (1981). Social validation in mental retardation. Applied Research in Mental Retardation, 2, 39–53. Kilsby, M., & Beyer, S. (1996). Engagement and interaction: A comparison between supported employment and day service provision. Journal of Intellectual Disabilities Research (Aug.), 40(40), 348–357. Knight, P. M., & Westbrook, J. (1999). Comparing employees in traditional job structures vs. telecommuting jobs using Herzbergs’s hygienes & motivators. Engineering Management Journal, 11(1), 15–20. Kroese, B. S. (1997). Cognitive-behaviour therapy for people with learning disabilities: Conceptual and contextual issues. In: B. S. Kroese, D. Dagnan & K. Loumidis (Eds), Cognitive-behaviour therapy for people with learning disabilities (pp. 1–15). London: Routledge. Kroese, B. S., Dagnan, D., & Loumidis, K. (Eds) (1997). Cognitive-behavior therapy for people with learning disabilities. London: Routledge. Leornard, N. H., Beauvais, L. L., & Scholl, R. N. (1999). Work motivation: The incorporation of self-concept based processes. Human Relations, 52(8), 969–998. Lou, L. (1999). Work motivation, job stress and employees’ well-being. Journal of Applied Management Studies, 8(1), 61–72. Mank, D., Cioffi, A., & Yovanoff, P. (1997). Analysis of the typicalness of supported employment jobs, natural supports, and wage and integration outcomes. Mental Retardation, 35(3), 185–197. Montero-Perez, J., Perula-de Torres, L. A., Martinez-de la Iglesia, J., & Jimenez, G. (1991). Comparative study of health related behaviours and habits among school children of two areas of Cordoba. Revista de Sanidad e Higiene Publica, 65(6), 545–555. Neath, J., & Schriner, K. (1998). Power to people with disabilities: Empowerment issues in employment programming. Disability & Society, 13(2), 217–228. Neufeldt, A. H., & Albright, A. L. (1995). An international study of strategies leading to selfdirected employment. International Journal of Practical Approaches to Disability, 19(2), 3–8. Nightingale, D. V. (1982). Workplace democracy: An inquiry into employee participation in Canadian work organisations. Toronto: University of Toronto Press. Northway, R. (1997). Integration and inclusion: Illusion or progress in services for disabled people? Social Policy and Administration (Jun), 31(2), 157–172. Orth-Gomer, K., Rosengen, A., & Wilhelsen, L. (1993). Lack of social support and incidence of coronary heart disease in middle-aged Swedish men. Psychosomatic Medicine, 55(1), 37–43 (Abstract). Putnam, R. D. (Ed.) (2002). Democracies in flux: The evolution of social capital in contemporary society. New York: Oxford University Press Inc. Redaelli, T., Cassinis, A., Cosignani, F., Conti, B., Onofri, M. P., & Dall’Acqua (1992). Interdisciplinary treatment of spina bifida children. Paraplegia, 30(10), 683–689. Rusch, F. R., Wilson, P. G., Hughes, C., & Heal, L. W. (1995). Interaction of persons with severe mental retardation and their non-disabled co-workers in integrated work settings. Behavior Modification, 19(1), 59–77.

Positive Externalities of Organisation Culture

291

Shafer, Kregel, Banks, & Hill (1988). An analysis of employer evaluations of workers with mental retardation. Research in Developmental Disabilities, 9(4), 377–391. Thernlund, G. M., & Samuelsson, M. A. K. (1993). Parental social support and child behaviour problems in different populations and socioeconomic groups: A methodological study. Social Science and Medicine, 36(3), 353–360. van Buskirk, W., & McGrath, D. (1999). Organizational cultures as holding environments: A psychodynamic look at organizational symbolism. Human Relations, 52(8), 805–832. Villeneuve, L., Lebel, P., & Lambert, J. (1992). Stress and social support in relation to psychological distress in community-residing elderly persons in Montreal. Canadian Journal of Public Health, 83(5), 353–356. Warhurst, C. (1998). Recognizing the possible: The organization and control of a socialist labour process. Administrative Science Quarterly, 43, 470–497. Woodward, M. (1999). Epidemiology study design and data analysis: Texts in statistical science. London: Chapman & Hall/CRC.

292

APPENDIX Table A1.

Characteristics of Study Population.

Means, Standard Deviations, and Correlations among Variablesa Variable

S. D.

1

2

3

4

5

6

7

8

9

10

11

12

— 19.2 41.4 23.2 16.2 59% 41% 2.07

3.20

0.06

—

0.00 0.13

0.01 0.12

— 0.03

—

0.07

0.05

0.00

0.17

—

0.04

0.01

0.10

0.04

0.09

—

0.02

0.00

0.10

0.06

0.09

0.10

99.0% 1.0% 74.3% 25.7% 30.5% 59.0% 4.8% 5.7% 70.5% 22.9% 6.7%

—

ABRAHAM MUKOLO ET AL.

1. Age group 1 ¼ 19–25 years 2 ¼ 26–35 years 3 ¼ 36–45 years 4 ¼ 46–65 years 2. Sex 1 ¼ Male 2 ¼ Female 3. Comprehension level 4. Marital status 1 ¼ Single 2 ¼ Married 5. Partner 0 ¼ No 1 ¼ Yes 6. Independent living status 1 ¼ Institutional 2 ¼ Family 3 ¼ Semi-independent 4 ¼ Independent 7. Employment type 1 ¼ Sheltered 2 ¼ Supported open 3 ¼ Open

Mean or Frequency

a

34.10 7.84

7.98 6.01

0.06 0.27 0.16 0.17 0.29 0.06 0.02 0.03 0.05 0.21 0.21 0.17

0.18 0.05 0.25

0.04 0.00 0.12

.18 — 0.16 0.05 0.37 0.14

— 0.22

—

7.6% 13.3% 27.6% 51.4% 0.08

0.10

0.05

0.26 0.16

0.03

0.09

0.16

0.10

0.55

—

0.13

0.17

0.05

0.07

0.04

0.12

0.05

0.03

0.12

0.39

0.16

—

0.13

0.16

0.20

0.05

0.04

0.12

0.10

0.35 0.05

0.06

0.07

0.14

12.4% 87.6% 21.0% 3.8% 21.9% 2.9% 14.3% 13.3% 22.9%

Positive Externalities of Organisation Culture

8. Weekly work hours 9. Years employed 10. Organisation status 1 ¼ Worker co-op 2 ¼ Community association 3 ¼ Foundation/Trust 4 ¼ Religious institution 11. Country 1 ¼ UK 2 ¼ Ireland 12. Industry type 1 ¼ Cafe´/Catering 2 ¼ Print/Print services 3 ¼ Packaging 4 ¼ Bakery/Health food and retail 5 ¼ Horticulture 6 ¼ Wood products 7 ¼ Various 13. Overall life satisfaction 1 ¼ Unhappy 2 ¼ Happy 3 ¼ Very happy

3.9% 17.6% 78.4%

Spearman correlation coefficient.

po0.05; po0.01.

293

294

Table A2.

Workplace Social Integration Variables.

Means, Standard Deviations, and Correlations among Variablesa Variable

3.55

S.D.

2.34

1

— 0.26

2

3

4

5

— 0.18

—

6

—

50.5% 30.5% 12.4% 5.7% 1.0% 6.7% 93.3% 2.65 36.2% 63.8% 8.53

0.77

3.12

0.07

0.09

—

0.15 0.09

0.05 0.02

0.15 0.12

0.90 0.45 0.22 0.38 0.29

—

7

8

9

10

11

12

13

ABRAHAM MUKOLO ET AL.

1. Work–talk involvement 2. Work-group membership 0 Groups 1 Groups 2 Groups 3 Groups 4 Groups 3. Social events attendance 0 ¼ No 1 ¼ Yes 4. Work time interaction 5. Socialisation with workmates 0 ¼ No 1 ¼ Yes 6. Workplace social interaction

Mean or Frequency

a

2.77 2.64 41.9% 58.1% 2.69

0.89

0.21 0.24 0.33

0.12 0.10 0.00

0.00 0.16 0.08

0.04 0.07 0.31 0.09 0.27 0.16

0.23 — 0.27 0.26 — 0.33 0.04 0.22

0.02

0.08

0.12

0.27 0.08

—

0.07

0.08

0.10

0.04

—

0.06

0.12

0.15

0.06

—

0.05

0.38 0.15

0.04

0.69

0.03

0.06

0.17

0.07

0.26 0.08

0.11

0.15

0.25 0.15

0.22 0.16

18.1% 81.9% 20% 80% 11.53 20.67

Spearman correlation coefficient.

po0.05; po0.01.

0.62 0.81

2.46 4.67

—

0.29 0.03 0.22 0.30 0.22 0.37 0.44 0.57 0.47 0.41 0.41 0.51 — 0.76 0.36 0.27 0.39 0.33 0.88 0.35 0.47 0.47 0.29 0.24 0.34 0.73

Positive Externalities of Organisation Culture

7. Confidence to engage 8. Work support satisfaction 9. Socialisation opportunity 0 ¼ No 1 ¼ Yes 10. Work interaction opportunity 11. Work friendship closeness 0 ¼ No 1 ¼ Yes 12. Work friendship satisfaction 0 ¼ No 1 ¼ Yes 13. Workplace social acceptance 14. Workplace social integration

295

296

Table A3.

Non-Work Social Integration Variables.

Means, Standard Deviations, and Correlations among Variablesa Variable

a

mates socialisation

activity involvement social interaction relationship closeness relationship contentment socialisation opportunity social acceptance social integration

Spearman’s correlation coefficient.

po0.05; po0.01.

S.D.

1

2

3

4

5

6

7

— 0.97 0.16 0.26 0.61 0.50 0.82

— 0.20 0.26 0.58 0.51 0.84

— 0.67 0.14 0.59 0.47

— 0.25 0.66 0.53

— 0.83 0.82

— 0.88

— 28.8% 70.5% 4.12 4.19 2.17 2.42 5.38 10.08 14.99

2.0 2.17 0.82 0.70 2.05 2.67 4.17

0.13 0.37 0.19 0.10 0.14 0.21 0.31

ABRAHAM MUKOLO ET AL.

1. Non-work 1 ¼ Yes 0 ¼ No 2. Non-work 3. Non-work 4. Non-work 5. Non-work 6. Non-work 7. Non-work 8. Non-work

Mean or Frequency

THEME IV: PARTICIPATION: PAST, PRESENT, FUTURE

297

This page intentionally left blank

298

GUILD SOCIALISM REVISITED, AND ITS YUGOSLAV COUNTERPART Mario Ferrero ABSTRACT This paper revisits the early-20th-century British blueprint for Guild Socialism and discusses its similarities and differences with labor managed firm (LMF) theory and with the historic Yugoslav system. It finds that the Guild Socialist vision of a corporatist workers’ state based on universal, non-anonymous, multi-party negotiation of incomes, prices, and quantities comes much closer to anticipating the real-world Yugoslav experiment in worker-managed market socialism than the market-syndicalist utopia embodied in the Western economic model of the LMF and economy.

1. INTRODUCTION Guild Socialism was an influential school of thought within the early-20th-century British socialism. It has, since, been all but forgotten and played apparently no part in either the theoretical underpinnings of Yugoslav self-management or the Western economics of the labor-managed

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 299–319 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09010-1

299

300

MARIO FERRERO

firm (henceforth, LMF) and economy.1 This is unfortunate because Guild Socialist theory presciently laid down a set of institutions and working arrangements that bear a striking resemblance to those of former Yugoslavia, especially as it evolved during the 1970s and 1980s. The purpose of this paper is, first, to survey the original writings of Cole, Orage, Hobson, and others and describe the main lines of Guild theory; second, to point out its main differences with LMF theory; and third, to suggest a comparison of the Guild model with the actual Yugoslav system. It has often been noted – among others, by the late Branko Horvat (1986) – that the basic LMF assumptions of a self-managed firm facing competitive markets for all goods and non-labor factors in a laissez-faire environment is a conspicuously inadequate description of the environment in which Yugoslav self-managed firms operated. The LMF model implies a noninterventionist state that concerns itself only with macroeconomic and distributional tasks, and in particular, that does not interfere with a nondiscriminatory supply of capital at market-clearing rental rates. By contrast, in the Yugoslav system, capital subsidization and rationing, state and corporate regulation of entry and exit of firms, and pervasive bargaining and collusion over prices and quantities by an array of producing, consuming, and sociopolitical organizations were in fact the norm. What makes the Guild model interesting in retrospect is precisely that (1) it envisaged a replacement of the all-purpose ‘‘state’’ with a battery of coordinating and governing bodies organized along functional lines, and (2) it envisaged the ‘‘market’’ as an institution for face-to-face contact, bargaining and contracting between producers and users at all levels, notably including investment finance. The tendency to ‘‘producer sovereignty’’ that is inherent in worker control was thus to be balanced by countervailing interest group pressure and resolved by vertical and horizontal coordination through universal bargaining – not quite a market economy as we understand it. Guild Socialism, much like real-world Yugoslavia, was in fact designed as a disguised, cumbersome form of comprehensive, detailed micro-planning of prices and quantities, where the planner’s orders from above were replaced by permanent consultation, negotiation, and ‘‘responsibility for supply’’ – a feature reminiscent of other experiments in market socialism, notably the Hungarian reform. While the main body of this paper is concerned with the writings of a group of socialist thinkers, it is not intended to be an exercise in the history of thought. Disregarding historical contingencies and controversies, it will focus mainly on the works of the most systematic and long-lived member of

Guild Socialism Revisited, and its Yugoslav Counterpart

301

the group, G.D.H. Cole, and endeavor to bring out the main lines of the theoretical blueprint and its evolution with the hindsight of the Yugoslav experience. The paper therefore more properly belongs to the theory of economic systems.

2. THE SYSTEM OF NATIONAL GUILDS AND ITS HISTORICAL OPPONENTS The Guild idea was first put forward by a number of British writers gathered around A.R. Orage’s New Age journal, and received its first full exposition in the book National Guilds, edited by Orage and collecting articles anonymously published on that journal by S.G. Hobson (Orage, 1914). A National Guilds League was founded in 1915 under the leadership of G.D.H. Cole, author of The World of Labour (Cole, 1913), and W. Mellor, and for some time won considerable support in the Trade Unions. After the war, important theoretical contributions were produced, and a practical experiment was tried out by Hobson who started a National Building Guild in 1920. After an initial success, this experiment collapsed and as a consequence the whole movement was discredited. The League was dissolved in 1923 and although the idea lingered on as an inspiring ideal, after 1923 there was no longer an organized Guild movement. As its leader-turned-historian recalls, ‘‘the Guild Socialists found themselves after 1918 squeezed out between the contending factions of Communism and parliamentary social democracy’’ (Cole, 1961, p. 454). Guild Socialism (henceforth, GS) aimed at replacing capitalism with a system that could revive, in modernized form, the spirit of craftsmanship and service to the community of the medieval Gilds (hence the name). It looked for a ‘‘third way’’ between Collectivism or State Socialism, as advocated in particular by the Fabian Society, and Syndicalism.2 The Fabian Society, led by Sidney and Beatrice Webb, advocated policies similar to those advocated by most socialist parties in the Second International and which in time became the platform of the British Labour Party. In their view, the basic flaw of industrial capitalism was the class monopoly of capital ownership, and the way out was public ownership, leading to ‘‘planned production for community consumption’’; on this basis, the Webbs would later become supporters of the Soviet model. The Fabians neither trusted the workers, nor their trade unions, to be willing or able to take over the management of their industries, which would best be left to

302

MARIO FERRERO

officials appointed by the State. The State itself was conceived of primarily as a national organization of consumers rather than a political machinery; a democratic parliament, together with the local governments and the consumer cooperatives, was to be the guardian of the consumers’ interests and ensure public control of the main directions of production and allocations of national income. This is what ‘‘industrial democracy’’ meant for the Webbs. Importantly, the income allocations included shares for collective services, social security, health, and education; in this regard, the Fabians are rightly counted as among the forerunners of the modern Welfare State. Their main difference with the continental socialists and communists of their day was in the means rather than the ends: the collectivist society was to be achieved through gradual reform and peaceful electoral politics rather than class warfare and revolution. On the other hand, the Syndicalist program, which under the leadership of Tom Mann became influential in sections of the Trade Unions, particularly the miners’, was a blend of ideas and practices borrowed from the American Industrial Workers of the World and the French anarchosyndicalists. The Syndicalists distrusted parliamentary politics altogether and relied on direct industrial action; to that purpose they wanted to turn the unions into revolutionary organizations whose goal should be not to protect the workers under the wage system but to abolish it. This required a thorough restructuring of unionism along industry, as opposed to trade or craft, lines, so as to achieve the potential for broad mass mobilization and eventually ‘‘One Big Union’’ (the IWW’s slogan), capable of dealing ‘‘One Big Strike’’ to the capitalist system and paving the way to social revolution. The outcome of this struggle was to be not nationalization of industry, which would merely imply a change of masters over the workers, but takeover from below and worker control on behalf of the whole working class (‘‘the mines to the miners’’). Tellingly, some (though not all) syndicalists and industrial unionists later joined the communists and were comfortable with the notion that ‘‘worker control’’ would be exercised by the workers as a class, through the ‘‘dictatorship of the proletariat’’, which alone could in their eyes justify centralization and discipline enforced from above. By contrast, the Guild Socialists took a ‘‘federalist’’, anti-Marxist stand, and emphasized self-government from below. In their critique, collectivism was bent on nationalization of industry on behalf of the consumer, represented by the State, for purposes of income redistribution, but left the wage system in place and necessarily involved bureaucratic, authoritarian

Guild Socialism Revisited, and its Yugoslav Counterpart

303

management of industry. Syndicalism, on the other hand, aimed at instituting worker control and self-government in industry, but made no provision for the general interests of the community and the consumer. By contrast, the blueprint for GS abolished the wage system, entrusted the monopoly of labor ‘‘by hand and brain’’ in each industry to a National Guild, and provided for consultation and joint decision-making between the Guilds and some institution representing the community as a whole (on which more below). The transition to GS was envisaged as a gradual refusal by workers to sell their labor as a commodity. In the end, this process would make the capitalist system unworkable and would be crowned by nationalization, which, if peaceful and involving compensation to the owners, would turn the former capitalists into useless rentiers. From then on, industry would be under ‘‘social ownership’’, formally vested in the State, but in fact held in trust and controlled by the Guilds. The agents of change were to be the Trade Unions, but for these to become Guilds, i.e. to ready themselves for the task of managing production rather than fighting the employers, three changes were required (Cole, 1913, 1920a). First, the unions should acquire a ‘‘blackleg-proof’’ monopoly of labor in each industry, fighting if necessary for compulsory membership. Second, they should be thoroughly reorganized along industrial lines, as opposed to the craft structure of the day. This implied that the fragmentation of the production chain should be overcome by vertical integration as far as practicable. Third, the unions should extend membership to managers, supervisors, and technicians (‘‘brain workers’’), winning over their loyalty at the expense of the employers. In the new system, the Guilds alone, each including the whole personnel of a given industry, would be in charge of production and appropriate the sales revenues. However, they would not be working for profit: public ownership would permit public control of prices and workers’ incomes, thus protecting consumers and defeating any tendencies for the Guilds to become exploitative monopolies. On this ‘‘public control’’, Guildsmen (as they called themselves) were not of one mind: some (Cole, 1913, 1918) would entrust it to the territorial State as consumers’ representative, others (Hobson, 1920) to a Distributive Guild, leaving the State as protector of producers and consumers alike, still others (Cole, 1920b) to an entirely new communal organization. By whichever means, public control was to enable Guild independence and industrial self-government to be socially beneficial, ushering in industrial democracy for the first time.

304

MARIO FERRERO

3. THE GUILD IN BEING There are four distinguishing marks of the wage system that the Guild is designed to overcome (Cole, 1918, Chap. 6). These are: (1) ‘‘Labor’’ is abstracted from the laborer and made into a commodity. This ends when organized workers perfect their monopoly of labor supply in each industry and are thereby in a position to secure control of capital as well. (2) Wages are paid only when it is profitable for the capitalist to employ labor. This ends when the Guild institutes income security and takes upon itself the charge of ‘‘industrial maintenance’’, i.e. it pays the worker also when he is ill, unemployed, or on short time, and when he retires. (3) In return for his wage, the wageworker surrenders all control over the organization of production. The first principle of Guild organization is to place control of production in the hands of the associated workers, starting at the shop floor level, which implies drawing in the ‘‘brain workers’’ and electing foremen and superiors. (4) In return for his wage, the wageworker surrenders all claims upon the product of his labor. In the Guilds the workers regain a claim on their product through the control of exchange and investment, subject to public supervision (on which more below). Operationally, despite the extensive vertical integration implied by the ‘‘industrial’’ principle of organization, the Guild is to be decentralized to the greatest possible extent, to encourage local initiative and valuable differences. ‘‘The Collectivist Utopia would be a world of public trusts; the Guild Utopia will be a world of producers’ cartels, worked in the interest of the whole community. If the Guild is not to fall into mediocrity, it must preserve the distinctness of works from works, of locality from locality, and of nation from nation. It is the organization of human differences on the basis of human identity’’ (Cole, 1918, p. 254). The basic principle is government from below: ‘‘wherever a body of men has to work under the supervision of a leader or officer, it must have the choice of that officer’’ (ibid., p. 255). Cole then proceeds to lay down detailed arrangements for election and appointment at all levels of the industrial hierarchy. Workshop committees are elected by direct ballot of the workers, and ‘‘Works’’ committees (corresponding to the firm of today) are elected sectionally by ballot of the members of each shop. Then there is the District committee, which consists of (1) works representatives, elected by each Works committee, and (2) craft

Guild Socialism Revisited, and its Yugoslav Counterpart

305

representatives, elected directly by all members of each craft across all works in the district. This committee coordinates production over the district and handles interactions with municipalities and the other Guilds in the district; it is also the main link between the individual works and the Guild as a whole. Next comes the National Guild Executive, which consists of (1) district representatives, elected by general ballot of each district, and (2) craft representatives, elected by general national ballot of each craft. Thus every member of the Guild is to cast two votes for the National Executive, one for his district and one for his craft representative. Finally, there is a National Delegate Meeting with ultimate power of control over the Executive; this is elected by general ballot of the members of each craft in each district. ‘‘The giving to each committeeman of a more restricted but at the same time more alert electorate secures that the individual workers shall not only elect, but also control, their leaders. It converts a paper democracy into a system of true self-government’’ (ibid., p. 262). It will be noticed that this elaborate structure carefully balances out enterprise, territorial, and craft interests, but the larger organs of Guild administration are based largely on the principle of factory representation (Cole, 1920b, p. 49). As to officials, foremen are elected by ballot of all the workers in the shop, the Works Manager by ballot of all the production workers in the works, and the Clerical Manager by ballot of all the clerical workers in the works. Higher executives are chosen by indirect representation: the General Manager of the Works is appointed by the Works committee, the District Secretary by the District committee, and the General Secretary of the Guild by the Executive Committee, subject to ratification by the Delegate Meeting. Experts at all levels are appointed by the respective committees. The Guild’s workers enjoy a high degree of job security. They can be dismissed for serious offenses, given appropriate safeguards (Reckitt & Bechhofer, 1918, pp. 310–311). But in case of redundancies due to slack demand they are to be supported by the Guild, as we have seen, so that it will be in the Guild’s interest to minimize the burden of unemployment benefits (Cole, 1918, p. 168) – that is, as long as labor’s marginal value product is positive. As to the managers, they can ultimately be removed by the workers under them – by whom they were elected (Cole, 1920b, pp. 58–59).

4. INTER-GUILD RELATIONS AND THE MARKET Different Guilds are related to one another by sending representatives to local and regional Guild Councils, and the structure peaks at the national

306

MARIO FERRERO

Industrial Guilds Congress, representing both National Guilds and local Guild Councils – a structure that parallels the Trade Union organization of the day. This Guilds Congress is in effect both the Guild legislature, which lays down general rules and principles, and the ultimate court of appeal that adjudicates differences and disputes between Guilds. It also holds some coercive power to enforce compliance and discipline on individual Guilds: in particular, it has the important task of preserving an open door into the Guilds. Workers are free to move and apply for membership in a particular Guild, subject, however, to the available openings; the Congress will see to it that no Guild is further restricting its labor demand for monopolistic purposes (Cole, 1920b, p. 75). A key role of the Guilds Congress is the fixing of pay rates. The general position of Guild Socialists is that the total national salary fund should be divided among the Guilds in a way strictly proportional to membership, so that per capita remuneration of all workers in the nation should be equal. Recognizing, however, that complete equality of incomes is not a viable proposition for some time to come, each Guild is then left free to allocate its own salary fund among its members as it wishes, possibly allowing for inequalities of skill and occupation. It is the duty of the Guilds Congress to enforce this distribution rule (Reckitt & Bechhofer, 1918, pp. 314–316, 328; Carpenter, 1922, p. 173). In his later work, Cole (1920b, pp. 71–72) comes to the alternative view that salary scales are to be drawn up by each Guild and submitted to the Guilds Congress for review and modification, so as to adjust the claims of various sections of workers. Inequality of pay may then persist as between different Guilds as well. In any case, the canonical question that exercised all brands of socialists, ‘‘who will do the dirty work under Guild Socialism’’, is answered by everyone by allowing not differential pay but shorter hours for unpleasant occupations (Cole, 1920b, pp. 75–76; Reckitt & Bechhofer, 1918, p. 316).3 How is, however, the national salary fund available for distribution to be determined? Each Guild receives revenue from the sale of its output and pools it in the treasury of the Guilds Congress, and applies for funds for depreciation and new investment. The Congress decides on capital allocations to every Guild, agrees with the State on the total amount of taxes to meet expected budget expenditures, and by subtracting taxes and investment funds from total revenues calculates the salary fund or total labor income. Taxation is thus entirely levied at source on the Guilds’ gross income, not on personal income. Moreover, the apportionment of the total tax burden among the Guilds is to be proportional to Guild revenue, or its monetary productivity (which amounts to the same thing, given the membership); and

Guild Socialism Revisited, and its Yugoslav Counterpart

307

since revenue and productivity depend on price, this mechanism is claimed to effectively prevent ‘‘profiteering’’ by the Guild by making monopolistic price increases self-defeating.4 Nevertheless, as there may be goods that are priced above or below cost for social reasons, through excises or subsidies, price fixing is best carried out jointly by the Guilds Congress and the State or some other functional body representing the consumer. Taxes and prices are thus to be determined by the same, joint body. On the other hand, though there is a lack of detail, it is clear that the rate and structure of capital accumulation is a matter for central decision by the Guilds Congress (Cole, 1918, pp. 283–286; Reckitt & Bechhofer, 1918, pp. 316, 328–329). There is no discussion whatever of the entry and exit of firms – a resounding silence to the ears of today’s readers. One is left to speculate that decisions about the setting up of new firms, or the closure of bankrupt ones, would be part of the overall central decisions on investment policy, possibly after consultation with the Guilds concerned. There is free consumer choice in the system, but users and producers are not supposed to meet on a competitive, anonymous market. According to Cole (1918, pp. 272–273, 1920b, pp. 59–60), the ‘‘works’’, or firm, within a Guild would only produce, not sell; exchange is carried out by either the Guild’s District committee or its National committee, as the case may be. This division of functions guarantees self-management in production and national coordination of exchange at the same time. The exchange of intermediate goods is to be arranged directly by the Guilds concerned at various levels, aided by an exchange of representatives on their respective governing boards, like a system of interlocking directorates. This device helps settle prices, quantities, and implicitly, valuations of labor (Orage, 1914, ch. 13; Cole, 1920b, pp. 67–68). The exchange of final goods is to be arranged between Guilds and consumer representatives. The latter are variously specified as either a Distributive Guild, developed out of the Cooperative Societies of the day, whose consumer membership would become universal (Hobson, 1920, pp. 76–78), or the municipalities and the State acting on the consumers’ behalf (Cole, 1918, pp. 272–273), or some nondescript consumer associations to be set up for the purpose (Reckitt & Bechhofer, 1918, p. 329). Clearly, these different specifications imply very different views of the role of political organizations versus voluntary associations. We will see in the next section that in his later work, Cole comes to a comprehensive solution of this key problem. Finally, as noted by a perceptive contemporary commentator on the Economic Journal (Reynard, 1920), the fundamental question of work incentives is usually slighted by Guild Socialists, who reject the very idea of

308

MARIO FERRERO

‘‘payment by results’’ as a disguised survival of the ‘‘profit motive’’ to the detriment of the motive of craftsmanship (e.g. Reckitt & Bechhofer, 1918, pp. 312–314). However, the division of functions proposed by Cole and mentioned above, whereby the factory delivers its output to the Guild’s District committee which takes care of marketing it, implies that the committee pays for this output according to the price list fixed by the National Guild, with due regard to quality. This allows for worker productivity to be rewarded – provided, however, that the central setting of salary funds, discussed above, makes room for it (Cole, 1918, pp. 273–274; also Hobson, 1920, p. 16; skepticism in Reynard, 1920, pp. 325–326).

5. THE GUILDS, THE CONSUMERS, AND THE STATE It has been apparent in the exposition so far that Guild Socialists rely on the principle of function as a guiding principle for industrial and social reorganization, as developed by Tawney (1920) in a related line of thought. In what is perhaps the most mature product of GS theory, Guild Socialism Restated, Cole (1920b) carries this principle to its logical consequences and delivers us the most consistent and thorough account of Guild society. We give here an outline of its main innovations with respect to previous Guild theory. Probably influenced by Hobson, Cole now rejects his previous view (of Fabian legacy) of the State and local governments as representatives of the consumers. Even more, he now denies that government based on territorial representation can be acknowledged as democratic, because ‘‘democracy is only real when it is conceived in terms of function and purpose’’ (ibid., p. 31). It follows that ‘‘there must be, in the Society, as many separately elected groups of representatives as there are distinct essential groups of functions to be performed...for a human being, as an individual, is fundamentally incapable of being represented’’ (ibid., p. 33). The ‘‘omnicompetent State’’ with its ‘‘undifferentiated representation’’ must therefore be destroyed, or ‘‘wither away’’, and be replaced by a system of functional representation. We have already seen Cole’s elaborate provisions for sectional and functional representation in the design of the individual Guild structure and the network of Guilds, culminating in the Guilds Congress. We must now complete the picture on the consumer’s side. Consumers in their functional capacity, or interests, can be distinguished as subjects of personal and collective consumption (local or national utilities). This functional differentiation would be violated if a single body were

Guild Socialism Revisited, and its Yugoslav Counterpart

309

entrusted with representing both interests, or if any such body were political rather than ad hoc. Let then Cooperation represent the consumer in relation to personal consumption, and a Collective Utilities Council represent him in relation to collective consumption. The latter is a nonpolitical body taking over the economic functions of today’s local governments. Furthermore, there are to be Civic Guilds, notably an Education Guild of teachers and a Health Guild of medical personnel, analogous to the Industrial Guilds in everything except that their services are not sold on the market but financed by public funds. Therefore, there must be a Health Council and a Cultural Council representing consumers in their capacity of users of those services. These four consumers’ councils are to be set up at all levels, local, regional, and national. ‘‘This National Collective Utilities Council would be, whereas the State is not and cannot be, a proper national representative of collective consumption’’ (ibid., p. 90). These consumer organizations articulate and specify the consumers’ demands in each sector, and protect the consumer against the producer if need be, by meeting constantly with Guild representatives at all levels. This is how supply and demand are brought into balance. This clears the ground for a picture of the communal organization of Guild Socialist society. There is no more a State that either exercises economic and civic functions or coordinates them. There must be noncoercive self-coordination by the various bodies which require coordination. This self-coordination is to take place within a new institution called the Commune, local, regional, and national, respectively. The Communes at each level include: the Guild Council or Congress representing the industrial Guilds, the Civic Guilds, the Cooperative Council, the Collective Utilities Council, the Cultural Council, and the Health Council. Citizens vote separately for representatives to each of the four consumers’ councils at the local level. The regional councils, comprising the regional Commune, are made up of representatives of the various local councils, subject to recall. The National Commune, which replaces the State of today, is composed of representatives of the National Guilds of all kinds, the National Councils of all kinds, and the Regional Communes themselves. The National Commune has few direct administrative functions left. It arbiters issues of demarcation between functional bodies and adjudicates disputes; it exercises directly only the functions of national defense and non-economic foreign relations. Analogous functions of settling disputes and managing the police force fall to the local and regional Communes. The basic task of the Commune at various levels is the coordination of resource allocation and economic regulation. As explained above, prices are

310

MARIO FERRERO

determined by joint producer and consumer decisions in these bodies. Guild budgets, including both capital requirements and salaries, must first go before the various councils of consumers or citizens and then be evaluated and brought into consistency by the Commune itself. On this basis the comprehensive Commune budget is drawn up; it determines in one shot the rate and structure of investment, the levels of worker incomes, the rate and allocation of Guild taxation, the allocations to the civic services, and the incomes of nonworking individuals. The greater part of this work of financial coordination is in fact not done by the Commune at all, but directly and in consultation by the various functional bodies. Finally, to keep overall monetary balance, the National Commune controls the issue of credit to the Guilds, and the banking system is also communal. A remarkable consequence of such a comprehensive, tight financial planning is that, with labor incomes and capital allocations predetermined by the Commune budget, ‘‘any surplus realized by a Guild in its annual working would pass to the Commune for allocation’’ (ibid., p. 147). This appears to contradict the possibility, previously allowed by Cole, of rewarding productivity at the factory level, as discussed at the end of the previous section. An important implication, stressed by Carpenter (1922, pp. 274–275), is that Cole’s theory neither considers nor permits a party system to exist: the electorate is split up into several functional constituencies, each with its own fixed interests. Politics, and political competition for office, can be no more. The State has indeed ‘‘withered away’’.

6. THE GUILDS WITHIN SOCIALIST PLANNING The final chapter in the story of GS was written once again by G.D.H. Cole, 15 years after Guild Socialism Restated. Now a convert to central planning, but still an ‘‘unrepentant Guild Socialist’’ at heart, Cole wrote a book on economic planning (Cole, 1935) in which he tried to fit his original scheme into the socialist planning framework. He now recognizes that the case against workers’ control is ‘‘far more formidable than I used to admit when I wrote my Guild Socialist books’’; still, it must remain the ultimate end, or ‘‘political democracy will be stultified by the autocracy of the economic system’’ (ibid., p. 329) The argument is basically unchanged, only the road to GS is now expected to be longer and more roundabout. Industrial democracy begins in the workshop, where a Workshop committee must be free to choose how production is to be carried out, while the plan sets what is to be produced.

Guild Socialism Revisited, and its Yugoslav Counterpart

311

From this a Works committee can in time be developed. Then gradually the workers may be entrusted with the election of foremen and managers, and be allowed to send delegates to the regional and national boards in charge of an industry. At first these boards are appointed by the national planning authority, and the workers’ delegates have only advisory functions. At some point, however, the delegates will take over responsibility for choosing the board members, and the industry will become fully self-governing – in fact, a Guild, making proposals and demands on the central planners. This finally transforms the planning authority into a coordinating authority, working on the basis of spontaneous proposals and suggestions coming up from below. The expected sequence of social change is therefore reversed: top-down instead of bottom-up. In the earlier period, the Guilds were envisaged as arising out of the Trade Unions and springing into full life at the moment of nationalization, which was to be brought about from below by industrial action. Now nationalization is seen as coming first as a political measure, and the authorities in control of nationalized industry as created from above as sections of the new planned economy. Hence ‘‘the development of ‘workers’ control’ towards a Guild system can come only through a process of gradual expansion in the powers and functions of workshop bodies created, with Trade Union help, as integral elements of the new industrial order’’ (ibid., p. 336, italics added). As before, no industry can alone have the final say on what it produces or what it pays its workers – only the higher authority has been re-defined. As before, monetary incentives for work must gradually be superseded by a ‘‘responsibility for supply’’ – as we would now say. Unlike before, however, the agents of industrial democracy at the grassroot level are neither required nor expected to be all the workers concerned, but only ‘‘in each factory and workshop a sufficient minority of active and interested men and women to impress their leadership upon the rest’’ (ibid., p. 340). In fact, if not in name, this minority of activists may easily be interpreted to be the local section or cell of the revolutionary workers’ party, which takes upon itself the task of implementing worker control in the workplace, as in the experience of many communist regimes – though Cole would never have stated this much in so many words.

7. COMPARISONS We will best leave it to the reader to make sense of the foregoing intellectual story for himself. It is clear, however, that GS is a world away from Illyria,

312

MARIO FERRERO

and somewhere near former Yugoslavia. We now briefly take up each of these comparisons in turn.

7.1. Guild Socialism and the Theory of Market Syndicalism In his seminal article, Benjamin Ward (1958) rightly chose the label ‘‘market syndicalism’’ for the system he was analyzing, and the economics literature followed in his steps.5 In this economy, the labor-managed firm buys inputs, notably capital, and sells outputs in competitive, anonymous markets where consumer sovereignty prevails. Only the labor market does not exist. Although specific market failures, especially in the entry of new firms and capital accumulation, are likely to recommend state intervention, the state mostly concerns itself with rather conventional tasks of macroeconomic management, anti-trust policy, and income redistribution. There is no room in Illyria for constant face-to-face contact and interaction between organized interests to determine outputs, prices, and incomes. There is no corporate organization over and above the firm. In short, there is no Guild, and hence no need to counterbalance its power by collective or consumer interests. And there is no politics in the LMF model, as in all of neoclassical economics. At the beginning, the Yugoslav experiment was seen as the empirical counterpart to the LMF model (see Ward, 1958, 1967; Vanek, 1970). Until the early 1970s, the persistent inconsistencies between the model and Yugoslav reality were usually explained away as lingering legacies of central planning and control that would in time die out. But then, instead of moving ahead in the direction of further market liberalization and decentralization, Yugoslavia took the opposite course and moved toward a new system of ‘‘negotiated planning’’ (see Section 7.2). This prompted a scholarly reaction and a long sequence of ad hoc amendments to the LMF model in an effort to fill the gaps and fix the cracks in the theoretical construct.6 By the time Yugoslavia collapsed, it was clear that the revisions were inadequate and something fundamental was amiss in the market-syndicalist approach to the subject. By contrast, as we have seen, the market, as we understand it, does not exist under GS. In its stead, there is constant, pervasive, organized interaction, and bargaining between users and producers at all levels. There is industrial democracy in the extreme, but no industrial competition to speak of. The Guild is fully decentralized in its government and operation, but work incentives are muted and ultimately rely on a kind of ‘‘responsibility for

Guild Socialism Revisited, and its Yugoslav Counterpart

313

supply’’ vis-a`-vis a variety of social organizations. Despite self-management, the determination of wages, prices, capital investment, and income distribution is centralized at the top, albeit in a multi-party bargaining framework. The representation of special interest groups is segmented and completely depoliticized, not because Guildsmen gave no thought to politics but because they wanted political government to ‘‘wither away’’. And importantly, as the previous section showed, the Guild system need not spring up from below but can be created from above, through devolution of nationalized industry. This is exactly the way in which, starting from centrally planned socialism, worker-managed market socialism came into being in Yugoslavia. To this we now turn.

7.2. Guild Socialism and Yugoslavia’s Worker-Managed Market Socialism As is well known, the Yugoslav experience with economic organization7 began with a standard Soviet-type, centrally planned system, and started introducing workers’ councils and self-management at firm level in 1950. Unsurprisingly, this proved to be inconsistent with central planning. By the mid-1960s the latter was completely dismantled, markets and prices were considerably liberalized, firms gained greater freedom to distribute profits between capital investment and worker income, and they were encouraged to borrow from, and repay to, banks instead of being granted investment funds from the center. This system, often called ‘‘market self-management’’ (Prasnikar & Svejnar, 1991), was short-lived. The extensive decentralization to the market resulted in increased income inequalities among firms, industries, and regions, and these inequalities provided the main impetus behind a political recentralization of investment finance starting in the early 1970s, which worked through quasi-compulsory savings and transfers between firms, selective soft credit in an inflationary environment, and the bailing out of loss-making firms (the so-called ‘‘socialization of losses’’). This move away from laissez-faire did not, however, involve a switch back to command planning, but gave rise to an entirely new system. The new constitution adopted in 1974 and the Law on Associated Labor enacted in 1976 inaugurated the system of ‘‘integrally planned selfmanagement’’ (Prasnikar & Svejnar, 1991). On the one hand, enterprise decentralization and self-management were carried to unprecedented extremes. Within the large enterprises, each individual production unit was organized into an independent Basic Organization of Associated Labor

314

MARIO FERRERO

(BOAL), with its separate financial accounts, its own assets, its own workers’ council, and its own representation on the workers’ council of the larger enterprise, now called Work Organization of Associated Labor (WOAL). In principle, each BOAL was free to set its own prices, production and employment levels, and to negotiate contractual agreements with other BOALs. On the other hand, in practice, to be able to function, the BOALs were involved in constant consultation and negotiation with one another and with the mother enterprise on everything. Furthermore, local and republican governments, whose power was greatly enhanced under the strongly federalistic provisions of the new constitution, secured a strong say on such crucial matters as price setting, employment, and foreign trade. The ‘‘sociopolitical organizations’’, i.e. the trade unions and the League of Communists, were formally given a role in enterprise decision-making with a view to keeping self-management in line with the ‘‘public interest’’. The multi-party negotiations were formally to take place in institutional settings called ‘‘self-management interest groups’’, set up at all government levels, and to result in ‘‘self-management agreements’’ and ‘‘social compacts’’, to be entered voluntarily by each concerned party but eventually binding on all signatories. ‘‘Planning’’ in the new system was made up of the working out and execution of these myriad agreements. It is unclear how much of all this was just paperwork, but there is little doubt that the transaction costs involved by the operation of the system were substantial. Prout (1985, p. 71) calls the system ‘‘contractual planning’’ to distinguish it from both indicative and command planning, while Ben-Ner and Neuberger (1990) call it ‘‘negotiated planning’’. More sharply, Lydall (1989) traces the dismal economic performance of the 1980s to the paralyzing combination of federalism with socialist self-management, as opposed to competitive self-management. The federalist component was responsible for the quasi-autarkic duplication of production facilities and distribution networks, while the socialist, or social-ownership, component was responsible for the stunting of competition, the prevention of new firm entry and the subsidization of loss-making firms, the political influence on local employment, and the redistribution of investment funds from highincome to low-income firms.8 Lydall (1989, p. 81) concludes that ‘‘each commune, with all its power over local enterprises, is combined with a ‘national’ (republican or provincial) one-party state (y) each is a form of centralized state socialism’’, and the total system is in effect a form of ‘‘feudal socialism’’. It is not difficult to see the logic of this evolution: by pushing BOALs fragmentation and self-management to the extreme, work discipline and

Guild Socialism Revisited, and its Yugoslav Counterpart

315

effort incentives collapsed; by divorcing capital management from financial liability, the soft budget constraint syndrome was perpetuated. Given this, the only way to bring coordination to such an anarchistic system was by quasi-compulsory negotiation among all parties, at all levels, at all times. In this way, the theoretical boundary between Yugoslavia’s worker-managed market socialism and Hungary’s administrative market socialism was in practice blurred. Keren (1992) has noted the similarity between the two experiences of market socialism, which boils down to a politically enforced ‘‘responsibility for supply’’ as the only real force that kept supply, however sluggishly, in line with demand. An important implication of contractual planning of this type is the politicization of economic life at all levels, blurring the boundary between politics and economics. Insiders in responsible positions made this into a theory. For example, Horvat (1969), writing during the market decentralization period, envisages the conversion of the communist party into a ‘‘league of political activists’’, who are supposed to bring the collective interest to bear on all decisions at the grassroot level. Horvat’s stated goal is to go beyond both one-party socialism and multi-party democracy and implement a ‘‘non-party’’ system, in which party and state are both depoliticized and merged with economy and society into one continuous process of collective consultation, negotiation, and agreement. This is claimed to be the Marxian utopia of the ‘‘withering away of the state’’ coming true under worker-managed market socialism. In another place, Horvat (1972) sets forth the goal of doing better than democracy, based on faulty majority rule, and achieving ‘‘polyarchy’’, based on small, homogeneous, participatory decision-making units where in fact unanimity prevails; this is first to be implemented in the organization of the workplace and hence to shape the whole system. Thus, while institutional details differ in many ways, the system toward which Yugoslavia was gravitating in its latest period is remarkably reminiscent of GS. The key ingredients of GS were there, to wit: the abolition of the wage system; the combination of enterprise decentralization and worker control to an extreme degree with extensive vertical integration of firms along ‘‘industrial’’ lines – the Guild, in fact if not in name; the ensuing nonmarket character of the process of firm entry and exit and the suppression of market competition; the elaborate system of graded representation and interlocking boards that gave each workshop a de facto voting right on crucial matters such as income distribution within the firm and investment allocation; the ultimate centralization of basic decisions on prices, capital formation, taxation, and income distribution; the locally decentralized but

316

MARIO FERRERO

‘‘corporatist’’ nature of supply and employment decisions; the contractual, multi-party character of this decision making, both at the central and local levels, involving not only all manner of producers but all kinds of representation of consumer interests and ‘‘sociopolitical’’ bodies; and finally, at least embryonically, the ‘‘withering away’’ of political representation, political competition, political parties, and the political State itself to clear the way for the sovereignty of the producer in the service of the community. In the real history, as we know, it was not the political State but the composite Yugoslav nation that eventually withered away amidst ethnic conflict, and it is not clear that either worker management in general or the BOAL system in particular can be blamed for the undoing of Yugoslavia. Still, the degree of similarity between the old blueprint and the real system is striking enough to give food for thought.9

8. CONCLUSION This paper has surveyed the blueprint for GS after 80 years of neglect. The thrust of the argument has been that the GS vision of a corporatist worker’s state based on universal, non-anonymous, multi-party negotiation of incomes, prices, and quantities comes much closer to anticipating the realworld Yugoslav experiment in worker-managed market socialism than the market-syndicalist utopia embodied in the economic model of the LMF and economy. The central message of this exercise in comparison of a theoretical blueprint with a real-world experiment may simply be stated as follows. The competitive model of the labor-managed economy, or the market-syndicalist utopia, was created by well-meaning Western academic economists. Half a century before them, a group of theorists in close touch with the realities of the British working-class movement thought they knew better. They outlined a blueprint that was as far as anything from competition and decentralized decision-making and envisaged a network of Guilds as the mainstay of the society that was to come after the demise of the wage system. Perhaps not knowingly, the Yugoslav reformers were later to tread on the same path, with a remarkable convergence of features. The earlier, paper-and-pencil experiment in system design thus proved, in retrospect, a much more accurate anticipation of subsequent reality than the academic construct. Perhaps this is as it should be, laying the competitive-syndicalist utopia to rest for good.

Guild Socialism Revisited, and its Yugoslav Counterpart

317

NOTES 1. Vanek’s (1975) reader includes a short selection from Cole (1918). Horvat (1969) briefly mentions Guild Socialism as an historical precedent and has a quote from Cole (1920b), but this is apparently unrelated to the Yugoslav experiment. I have been unable to locate any other citation in the economics literature. 2. The following paragraphs are based on Cole (1956, Chaps. 3 and 4). Cole’s own prominent role in the events does not detract from his fairness of judgment as a historian but adds insight to his account. 3. This is Edward Bellamy’s solution to the incentive problem under equality of incomes (see Cugno & Ferrero, 1984). 4. We know that this is wrong: a revenue tax or an ad valorem tax on a monopoly cuts profits but further reduces output and increases price. 5. Besides Ward (1958, 1967), the classic contributions to the LMF model include Domar (1966), Vanek (1970) and Meade (1972). The subsequent theoretical literature is extensive; Ireland and Law (1982) provide a comprehensive analytic treatment and Bonin and Putterman (1987) a useful survey. A more recent survey, focused on the difficult match between model and reality, is Bonin, Jones and Putterman (1993). 6. The special issue of the Journal of Comparative Economics, 10, 1, edited by Branko Horvat and John Michael Montias, was perhaps a focal point in this revision process. See also the surveys by Bonin and Putterman (1987) and Bonin et al. (1993), and the discussion of the discrepancies between model and reality in Estrin, Moore and Svejnar (1988) and Prasnikar and Svejnar (1991). 7. Useful, concise assessments of the overall Yugoslav experience are Estrin (1991) and Prasnikar and Svejnar (1991). Extended, detailed accounts can be found in Prout (1985), Dyker (1990) and Lydall (1989); I found the latter particularly illuminating. Prasnikar and Svejnar (1988) and Vodopivec (1994) are essential sources for the description and analysis of the firm in the BOAL period, and Uvalic (1992) is a basic reference on investment decisions. Keren and Levhari (1992) is a rare effort at modeling the capital market under socialist self-management, while Keren (1992) is an illuminating comparison of the Yugoslav LMF with other experiments in market socialism. Ben-Ner and Neuberger (1990) is a good review and theoretical discussion of several planned market systems that were tried out in Yugoslavia. Finally, Horvat’s (1969, 1972, 1976) works not only are informed studies but offer a revealing view by a committed insider. 8. The paramount role of social ownership in the non-market allocation of investment is also emphasized by Prasnikar and Svejnar (1991) and Uvalic (1992). In an earlier paper (Ferrero, 1998), I offered a formal model of capital taxation/subsidization as the policy likely to be rationally adopted in a syndicalist society where self-managed firms vote over redistribution schemes, which strongly suggests that laissez faire in matters of investment is a political impossibility in such a society. 9. In the background of this discussion clearly lies the broader theoretical question of the viability of central planning cum worker control in the absence of functioning markets. An armchair exercise in system design such as GS can hardly yield sufficient insights to address such big questions. Nevertheless, it is noteworthy that both the theoretical debate among British socialists and the actual experimentation in

318

MARIO FERRERO

Yugoslavia evolved in the same direction, i.e. away from competitive decentralization and toward negotiated corporatist planning. See especially Ben-Ner and Neuberger (1990) and Keren (1992, 1993) for further discussion of these issues.

ACKNOWLEDGMENTS An earlier draft of this paper was presented at the conference of the International Association for the Economics of Participation (IAFEP) in Halifax (NS), July 8–10, 2004, whose participants provided interesting discussion. I am indebted to a referee and to Panu Kalmi for helpful comments and suggestions.

REFERENCES Ben-Ner, A., & Neuberger, E. (1990). The feasibility of planned market systems: The Yugoslav visible hand and negotiated planning. Journal of Comparative Economics, 14, 768–790. Bonin, J., Jones, D., & Putterman, L. (1993). Theoretical and empirical studies of producer cooperatives: Will the twains ever meet? Journal of Economic Literature, 31, 1290–1320. Bonin, J., & Putterman, L. (1987). Economics of cooperation and the labor-managed economy. New York: Harwood Academic Publishers. Carpenter, N. (1922). Guild Socialism: An historical and critical analysis. New York: Appleton. Cole, G. D. H. (1913). The world of labour. London: Bell & Sons. Cole, G. D. H. (1918). Self-government in industry. London: Bell & Sons. Cole, G. D. H. (1920a). Chaos and order in industry. London: Methuen. Cole, G. D. H. (1920b). Guild Socialism restated. London: Parsons. Cole, G. D. H. (1935). Principles of economic planning. London: Macmillan. Cole, G. D. H. (1956). History of socialist thought, vol. III: The Second International 1889–1914. London: Macmillan. Cole, G. D. H. (1961). History of socialist thought, vol. IV: Communism and social democracy 1914-1931. London: Macmillan. Cugno, F., & Ferrero, M. (1984). Individual incentives by adjusting work hours: Bellamy’s egalitarian economy. Journal of Comparative Economics, 8, 182–206. Domar, E. (1966). The Soviet collective farm as a producer cooperative. American Economic Review, 56, 734–757. Dyker, D. A. (1990). Yugoslavia. Socialism, development and debt. London: Routledge. Estrin, S. (1991). Yugoslavia: The case of self-managing market socialism. Journal of Economic Perspectives, 5, 187–194. Estrin, S., Moore, R. E., & Svejnar, J. (1988). Market imperfections, labor management, and earnings differentials in a developing country: Theory and evidence from Yugoslavia. Quarterly Journal of Economics, 102, 465–478. Ferrero, M. (1998). Public choice in the labor-managed economy. Economic Analysis, 1, 85–100. Hobson, S. G. (1920). National Guilds and the state. London: Bell & Sons. Horvat, B. (1969). An essay on Yugoslav society. White Plains, NY: International Arts and Sciences Press.

Guild Socialism Revisited, and its Yugoslav Counterpart

319

Horvat, B. (1972). An institutional model of a self-managed socialist economy. Reprinted in: Vanek, J. (Ed.), (1975). Self-management: Economic liberation of man. (pp. 127–144), Penguin: Harmondsworth. Horvat, B. (1976). The Yugoslav economic system. White Plains, NY: International Arts and Sciences Press. Horvat, B. (1986). The theory of the worker-managed firm revisited. Journal of Comparative Economics, 10, 9–25. Ireland, N. J., & Law, P. J. (1982). The economics of labour-managed enterprises. London: Croom Helm. Keren, M. (1992). The New Economic System, The New Economic Mechanism, and the Yugoslav LMF: Bureaucratic limits to reform. Economic Systems, 16, 89–111. Keren, M. (1993). On the (im)possibility of market socialism. Eastern Economic Journal, 19, 333–344. Keren, M., & Levhari, D. (1992). Some capital market failures in the socialist labor-managed economy. Journal of Comparative Economics, 16, 655–669. Lydall, H. (1989). Yugoslavia in crisis. Oxford: Clarendon Press. Meade, J. E. (1972). The theory of labour-managed firms and of profit-sharing. Economic Journal, 82, 402–428. Orage, A. R. (Ed.) (1914). National Guilds. An inquiry into the wage system and the way out. London: Bell & Sons. Prasnikar, J., Svejnar, J. (1988). Economic behavior of Yugoslav enterprises. In: D. Jones & J. Svejnar (Eds), Advances in the economic analysis of participatory and labor-managed firms (Vol. 3, pp. 237–311). Greenwich, CT: JAI Press. Prasnikar, J., & Svejnar, J. (1991). Workers’ participation in management vs. social ownership and government policies: Yugoslav lessons for transforming socialist economies. Comparative Economic Studies, 33, 27–45. Prout, C. (1985). Market socialism in Yugoslavia. Oxford: Oxford University Press. Reckitt, M. B., & Bechhofer, C. E. (1918). The meaning of National Guilds. London: Cecil Palmer & Hayward. Reynard, H. (1920). The Guild Socialists. Economic Journal, 30, 321–330. Tawney, R. H. (1920). The acquisitive society. New York: Harcourt, Brace & World. Uvalic, M. (1992). Investment and property rights in Yugoslavia. Cambridge: Cambridge University Press. Vanek, J. (1970). The general theory of labor-managed market economies. Ithaca, NY: Cornell University Press. Vanek, J. (Ed.) (1975). Self-management: Economic liberation of man. Penguin: Harmondsworth. Vodopivec, M. (1994). Appropriability of returns in the Yugoslav firm. Eastern Economic Journal, 20, 337–348. Ward, B. (1958). The firm in Illyria: Market syndicalism. American Economic Review, 48, 566–589. Ward, B. (1967). The socialist economy: A study in organizational alternatives. New York: Random House.

This page intentionally left blank

320

WHITHER SELF-MANAGEMENT? FINDING NEW PATHS TO WORKPLACE DEMOCRACY$ David Ellerman ABSTRACT Today, there is the real possibility that self-management and workplace democracy will follow socialism into the dustbin of history. But the connection of self-management to socialism was misconceived from the beginning. Workplace democracy has its own roots in the historical struggle against slavery and against autocracy. The paper reviews the history of the theory of inalienable rights that applies not only against the self-sale contract and the political contract of subjection but also against of the self-rental or employment contract, today’s contract of subjection for the workplace. The paper concludes with the current debate about corporate governance.

$

Keynote Address for 12th Annual Conference International Association for the Economics of Participation July 8, 2004, Halifax, Canada.

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 321–355 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09011-3

321

322

DAVID ELLERMAN

1. INTRODUCTION: SOCIALISM AND SELF-MANAGEMENT Historically, the idea of workers’ self-management has been entwined or entangled with the idea of ‘‘socialism.’’ Concretely, the connection arose out of the attempt by post-war Yugoslavia to find a third way, a system of selfmanagement distinct from both the state-socialism of the Russian model and the capitalist model in the West. Indeed, this form of work organization had its roots in that bold Yugoslav effort. Many in those days accepted the connection between socialism and the general idea of worker’s selfmanagement or workplace democracy. They thought of democracy in the workplace as a variant of socialism, as a ‘‘self-managed socialism’’ in Branko Horvat’s phrase. But today much has changed. The monstrous construction of Soviet communism has finally collapsed. And, like it or not, all the variants of ‘‘socialism’’ have been drowned in the resulting tidal wave. I would argue that the connection between workplace democracy and socialism was misconceived all along, and that we should welcome rather than regret the sweeping of all the ‘‘socialisms’’ into the dustbin of history. Perhaps there are a few who bonded with the word ‘‘socialism’’ in the pink of youth and now they want to redefine it yet again. But this is not the late 19th century when the definition of ‘‘socialism’’ was still up for grabs. Whether we like it or not, the 20th century has indelibly stamped the word ‘‘socialism’’ with the meaning of government ownership, and it is sentimental nonsense to pretend otherwise. ‘‘Socialists’’ who seek a new world might start by seeking a new word. Today, we face a different challenge – the challenge to find new paths to the underlying principles of workplace democracy so that those principles are not washed away in the deluge along with all the socialisms. Today, there is a real danger that those who favored workers’ self-management will retreat back from democratic first principles to the practice of ‘‘democratic participation’’ in firms still organized on the basis of the employer–employee relationship. How can we begin again to make the case for genuine democracy in the workplace?

2. EFFICIENCY ARGUMENTS One set of arguments, particularly amenable to economists, is about the relative efficiency of more ‘‘democratic’’ or participative firms in contrast to conventional firms. According to the often-implicit premises of these

Whither Self-Management? Finding New Paths to Workplace Democracy

323

arguments, there are no rights violations involved in the conventional employment firms so the choice between the types of firms is just that – a matter of choice. People may choose to set up one type of firm or the other. Proponents of the more democratic firms point out various possibilities for greater efficiency – be it allocative efficiency or X-efficiency. I have never given much credence to efficiency arguments for workplace democracy for two reasons. First, as will become clear, I think the transformation of the workplace is a question of rights, not expediency. Antebellum slavery may well have been an inefficient way to organize work but to analyze slavery as an ‘‘efficiency problem’’ would leave much to be desired. Second, efficiency arguments are not directly about the rights structure of the employment firm so they are always vulnerable to simulation counterarguments. If more bottom-up decision-making proves to be more efficient in a democratic firm, then an ‘‘enlightened’’ employment firm could always try to simulate it by delegating more decision-making to the shop or office floor in a participative scheme. Thus an efficiency argument can never imply that the rights structure should itself be changed.1 Nevertheless, this is as close as many economists – who see only through an efficiency lens – can come to an argument for a ‘‘democratic’’ or at least ‘‘participative’’ firm.

3. HOW CAN THERE BE A RIGHTS VIOLATION IN A TRULY VOLUNTARY EMPLOYMENT RELATION? All would agree that there was an inherent rights violation in slavery. But many will argue that any comparison of the employment contract to slavery is ill-taken. The phrase ‘‘wage slavery’’ is an oxymoron. Consent is a necessary condition for any acceptable relation. Slavery was an involuntary relationship and thus efficiency questions about slavery were always trumped by the rights violation due to the coercive nature of the relationship. They might agree that, leaving aside consent, one can compare the master–slave relation and the employer–employee relation along a scale measuring the amount and duration of the labor bought and sold. Under slavery, the master acquired all of the slave’s services – just as when one buys a machine, one acquires all of the machine’s services (plus the ‘‘scrap’’ at the end of its lifetime). In the employment relation, the employer only acquires the employee’s labor for a certain length of time – just as when one rents a machine, one acquires the machine’s service only for a certain duration.

324

DAVID ELLERMAN

While the master–slave and employment relationships might be compared along a duration scale, the two relationships are fundamentally different in terms of voluntariness. The basic liberal defense of the employment relation is not along the duration scale. The defense is that the employment relationship is voluntary. The emergence of modern liberal society was a movement from the involuntary status of being a slave or serf to a society based on contract – including the employment contract. The bedrock question within the liberal framework is the question of consent or coercion. Since slavery lies on the side of coercion and the employment relation on the side of consent, any comparison between the two is ill-taken – according to this standard view.

4. THE DARK SIDE OF LIBERALISM: CONTRACTARIAN ARGUMENTS FOR SLAVERY AND AUTOCRACY 4.1. Modern Liberal Philosophers We return to the basic question: ‘‘How can there be an inherent rights violation in a fully voluntary contract?’’ This question has an answer, an answer that was hammered out in the anti-slavery and democratic movements. The intellectual roots of the answer can be traced back to the Stoics but the golden thread of the argument descends to modern times from the Reformation and the Enlightenment. But that answer has however been lost to the mainstream of modern liberalism.2 This might be illustrated by considering the two most prominent liberal moral philosophers of our day, John Rawls and Robert Nozick both late of Harvard University. The employment contract is the voluntary contract to rent or hire oneself out to an employer for a certain purpose and time period. Ordinarily, the word ‘‘hire’’ is preferred but I use the synonym ‘‘rent’’ to extract us from the old mental ruts. The words are equivalent. Americans say ‘‘rent a car’’ and the British say ‘‘hire a car’’ but they mean the same thing. As Paul Samuelson puts it: One can even say that wages are the rentals paid for the use of a man’s personal services for a day or a week or a year. This may seem a strange use of terms, but on second thought, one recognizes that every agreement to hire labor is really for some limited period of time. By outright purchase, you might avoid ever renting any kind of land. But in our society, labor is one of the few productive factors that cannot legally be

Whither Self-Management? Finding New Paths to Workplace Democracy

325

bought outright. Labor can only be rented, and the wage rate is really a rental (Samuelson, 1976, p. 569).

The moral philosopher John Rawls (1971) lived his whole life in an economic system based on the renting of human beings. Yet in a lifetime of contemplation and writing on justice , Rawls never hinted that there might be an inherent problem in an economy based on the voluntary renting of human beings. Yes, here and there, a particular relationship might be abused. But in spite of some lip service to the Kantian principle of treating people as endsin-themselves and never simply as means, Rawls saw no problem per se in the relationship of voluntarily renting other human beings. Leaving aside the historically coercive nature of slavery, what about a truly voluntary self-sale contract to sell one’s labor by the lifetime instead of by the hour, week, or month? History has already ruled out such a voluntary slavery contract along with the institution of involuntary slavery. Again, as Paul Samuelson (1976) puts it: Since slavery was abolished, human earning power is forbidden by law to be capitalized. A man is not even free to sell himself; he must rent himself at a wage (p. 52 (emphasis in the original)).

Robert Nozick, not contented to stop at John Rawls’ acceptance of the selfrental contract, argued on strict liberal and indeed libertarian grounds that even the self-sale or voluntary-slavery contract should be accepted. This contract comes in both a collective and individual form. The collective form was historically known as the pact of subjection or pactum subjectionis, wherein a people alienated and transferred their right to govern themselves to a monarch or some other form of a Hobbesian sovereign. Professor Nozick (1974) argued that a free libertarian society should validate that sort of a contract with a ‘‘dominant protective association’’ playing the role of the Hobbesian sovereign (p. 15). And the same reasoning applied to the individual version of the alienation contract. The comparable question about an individual is whether a free system will allow him to sell himself into slavery. I believe that it would (Nozick, 1974, p. 331).

Accordingly, Nozick completely abandoned the notion of inalienable rights from the anti-slavery and democratic movements. But he kept the notion of a ‘‘right’’ that could not be taken away without one’s consent. But that is only a right as opposed to a privilege. Nozick had no notion of an ‘‘inalienable right’’ that may not be taken away even with one’s consent. Even if a system of positive law accepted a voluntary contract to alienate such a right, the anti-slavery and democratic movements argued, as we will

326

DAVID ELLERMAN

see, that such contracts were inherently invalid – and thus that the rights should be recognized as being inalienable. Nozick was not alone in this suggested revision of post-bellum jurisprudence to accept the self-sale contract. Nozick has neoclassical economics as a silent partner in this quest for liberal justice. Allocative efficiency requires full futures markets in all commodities including human labor. Any attempt to truncate self-rental contracts at, say, T years could violate market efficiency since there might be mutually voluntary contracts to buy and sell labor T þ 1 years in the future. Hence, market efficiency requires full future markets in labor – essentially the self-sale contract. Neoclassical textbooks loathe to admit this implication. But the Johns Hopkins University economist Carl Christ made the point quite explicit in no less a forum than Congressional testimony. Now it is time to state the conditions under which private property and free contract will lead to an optimal allocation of resources .... The institution of private property and free contract as we know it is modified to permit individuals to sell or mortgage their persons in return for present and/or future benefits (Christ, 1975, p. 334).

Thus John Rawls finds no justice problem in the self-rental contract and Robert Nozick explicitly and neoclassical economics implicitly accept the self-sale contract. While the 4th of July rhetoric of ‘‘inalienable rights’’ lingers on, the liberal mainstream has essentially lost the inalienable rights theory that descends from the Reformation and Enlightenment – the theory that explains how rights can be violated in a fully voluntary contract.

4.2. The Older History of Self-Sale Contracts Modern liberalism can ignore the idea of rights-violating voluntary contracts since liberalism promulgates a dumbed-down version of the historic debate about slavery as a simplistic morality play of consent versus coercion. The defenders of slavery are pictured as condoning coercion – at least of people with a sufficiently different ethnicity or race. Modern liberalism presents itself as having achieved the superior moral insight that coercion is always wrong – regardless of race or ethnicity. But that is a gross falsification of the actual historical debates. In fact, from ancient times, there have been sophisticated defenses of slavery on contractarian grounds. Roman law, as codified in the Institutes of Justinian, provided three legal means of becoming a slave:

Whither Self-Management? Finding New Paths to Workplace Democracy

327

Slaves either are born or become so. They are born so when their mother is a slave; they become so either by the law of nations, that is, by captivity, or by the civil law, as when a free person, above the age of twenty, suffers himself to be sold, that he may share the price given for him (Institutes Lib. I, Tit. III, 4).

In addition to outright contractual slavery, the other two means were also seen as having aspects of contract. A person born of a slave mother and raised using the master’s food, clothing, and shelter was considered as having agreed to a tacit contract to trade a lifetime of labor for these and future provisions. And Thomas Hobbes, for example, clearly saw a ‘‘covenant’’ in this ancient practice of enslaving prisoners of war. And this dominion is then acquired to the victor when the vanquished, to avoid the present stroke of death, covenants either in express words or by other sufficient signs of the will that, so long as his life and the liberty of his body is allowed him, the victor shall have the use thereof at his pleasure.y It is not, therefore, the victory that gives the right of dominion over the vanquished but his own covenant (Leviathan, II, Chapter 20).

The point is not the factual absurdity of interpreting this as a covenant; the point is the attempt by Hobbes and many others to ground slavery on the liberal basis of explicit or tacit consent. The dumbed-down consentor-coercion liberalism of our day would disagree only on the factual question of what constitutes ‘‘consent.’’ Roman Law thus contemplated three legal means of becoming a slave, and all were based on an implicit or explicit contract. But that is only Roman law; what about the father of modern liberalism, John Locke (1960)? In Locke’s influential Two Treatises of Government (1690), he would not condone a contract which gave the master the power of life or death over the slave. For a Man, not having the Power of his own Life, cannot, by Compact or his own Consent, enslave himself to any one, nor put himself under the Absolute, Arbitrary Power of another, to take away his Life, when he pleases (Second Treatise, y23).

Locke is ruling out a voluntary version of the old Roman slavery where the master could take the life of the slave with impunity. But once the contract was put on a more civilized footing, it would be a rather severe form of the master–servant or employment relationship. For, if once Compact enter between them, and make an agreement for a limited Power on the one side, and Obedience on the other, the State of War and Slavery ceases, as long as the Compact endures.... I confess, we find among the Jews, as well as other Nations, that Men did sell themselves; but, ‘tis plain, this was only to Drudgery, not to Slavery. For, it is evident, the Person sold was not under an Absolute, Arbitrary, Despotical Power (Second Treatise, y24).

328

DAVID ELLERMAN

Moreover, Locke agreed with Hobbes on the practice of enslaving the captives in a ‘‘Just War’’ as a quid pro quo exchange based on the on-going consent of the captive. Indeed having, by his fault, forfeited his own Life, by some Act that deserves Death; he, to whom he has forfeited it, may (when he has him in his Power) delay to take it, and make use of him to his own Service, and he does him no injury by it. For, whenever he finds the hardship of his Slavery out-weigh the value of his Life, ‘tis in his Power, by resisting the Will of his Master, to draw on himself the Death he desires (Second Treatise, y23).

Having thus established that slavery could be based on contract, the liberal defenders of slavery had no need to condone kidnapping. But wherever slavery existed as a settled condition, then – like society itself – the institution could be seen as being based on a tacit contract. For example, Rev. Samuel Seabury gave a sophisticated liberal contractarian defense of antebellum slavery. From all which it appears that, wherever slavery exists as a settled condition or institution of society, the bond which unites master and servant is of a moral nature; founded in right, not in might; .... Let the origin of the relation have been what it may, yet when once it can plead such prescription of time as to have received a fixed and determinate character, it must be assumed to be founded in the consent of the parties, and to be, to all intents and purposes, a compact or covenant, of the same kind with that which lies at the foundation of all human society (1969 (1861), p. 144). Seabury easily anticipated the retort to his classical tacit-contract argument. ‘‘Contract!’’ methinks I hear them exclaim; ‘‘look at the poor fugitive from his master’s service! He bound by contract! A good joke, truly.’’ But ask these same men what binds them to society? Are they slaves to their rulers? O no! They are bound together by the COMPACT on which society is founded. Very good; but did you ever sign this compact? Did your fathers every sign it? ‘‘No; it is a tacit and implied contract’’ (Seabury, 1969, p. 153).

This puts modern contractarian liberals in the uncomfortable position of disagreeing with Seabury only on factual grounds. They are reduced to arguing on empirical grounds that the implied contract for society has ‘‘genuine tacit consent,’’ but that the implied slavery contract does not. It is no surprise that modern liberalism has just avoided this quandary by dumbing down the intellectual history of the slavery debates. The sophisticated liberal contractual arguments go down the memory hole; it is just a question of consent or coercion.

Whither Self-Management? Finding New Paths to Workplace Democracy

329

4.3. The Older History of Contracts of Subjection It was previously noted that there were both individual and collective versions of the contract to alienate the rights of self-governance. The fullblown rump-and-stump version of the individual contract was the self-sale contract previously considered. The collective version was the pact of subjection which alienated and transferred the people’s rights of self-governance to a sovereign who then ruled in the sovereign’s own name – not as a delegate, representative, or trustee of the people. By the contract of subjection, the people became subjects of the sovereign. Here again, modern liberalism has dumbed down the intellectual history of the debate between autocracy and democracy to a simplistic question of coercion or consent. Democracy is presented as ‘‘government by the consent of the governed’’ and non-democratic governments are presented as being based on coercion. But again there was a liberal contractarian defense of non-democratic government from antiquity down to Nozick. We may again start with Roman law in our thumbnail sketch of the contractarian defense of autocracy. The sovereignty of the Roman emperor was usually seen as being founded on a contract of rulership enacted by the Roman people. The Roman jurist Ulpian gave the classic and oft-quoted statement of this view in the Institutes of Justinian (Lib. I, Tit. II, 6): Whatever has pleased the prince has the force of law, since the Roman people by the lex regia enacted concerning his imperium, have yielded up to him all their power and authority (quoted in Corwin, 1955, p. 4; or in Sabine, 1958, p. 171).

The American constitutional scholar, Edward S. Corwin, noted the questions which would arise in the Middle Ages about the nature of this pact. During the Middle Ages the question was much debated whether the lex regia effected an absolute alienation (translatio) of the legislative power to the Emperor, or was a revocable delegation (cessio). The champions of popular sovereignty...took the latter view (Corwin, 1955, p. 4).

It is precisely this question of translatio or concessio – alienation or delegation of the right of government in the contract – that is the key question. The liberal framing of the question as ‘‘consent or coercion’’ misses the real issue. Consent is on both sides of that question; it is a matter of a contract of alienation or a contract of delegation. The alienation version of the contract became a sophisticated tacit contract defense of non-democratic government wherever it existed as a settled condition. And the delegation version of the contract became the foundation for democratic theory.

330

DAVID ELLERMAN

The German legal thinker, Otto Gierke (1966), was quite clear about the alienation-versus-delegation question. This dispute also reaches far back into the Middle Ages. It first took a strictly juristic form in the dispute ... as to the legal nature of the ancient ‘‘translatio imperii’’ from the Roman people to the Princeps. One school explained this as a definitive and irrevocable alienation of power, the other as a mere concession of its use and exercise.... On the one hand from the people’s abdication the most absolute sovereignty of the prince might be deduced,y On the other hand the assumption of a mere ‘‘concessio imperii’’ led to the doctrine of popular sovereignty (pp. 93–94).

The contractarian defense of non-democratic government was based on the translatio interpretation of the tacit social contract. In contrast to theories which would insist more or less emphatically on the usurpatory and illegitimate origin of Temporal Lordship, there was developed a doctrine which taught that the State had a rightful beginning in a Contract of Subjection to which the People was party (Gierke, 1958, pp. 38–39).

In terms of the liberal ‘‘coercion or contract’’ dichotomy, this tradition was grounded foursquare on contract. Indeed that the legal title to all Rulership lies in the voluntary and contractual submission of the Ruled could therefore be propounded as a philosophic axiom (Gierke, 1958, pp. 39–40).

A state of government which had been settled for many years was ex post facto legitimated by the tacit consent of the people. In about 1310, according to Gierke (1958), Engelbert of Volkersdorf is the first to declare in a general way that all regna et principatus originated in a pactum subjectionis which satisfied a natural want and instinct (p. 146).

Skipping over many minor figures, Thomas Hobbes (1588–1679) made the best-known attempt to found non-democratic government on the consent of the governed. Without an overarching power to hold people in awe, life would be a constant war of all against all. To prevent this state of chaos and strife, men should join together and voluntarily transfer the right of selfgovernment to a person or body of persons as the sovereign. This pactum subjectionis would be a covenant of every man with every man, in such manner as if every man should say to every man, I authorize and give up my right of governing myself to this man, or to this assembly of men, on this condition, that you give up your right to him and authorize all his actions in like manner (Hobbes, 1958 (Original 1651), p. 142).

Whither Self-Management? Finding New Paths to Workplace Democracy

331

This contractarian tradition is brought fully up to date in Robert Nozick’s libertarian defense of the contract to alienate one’s right of self-determination to a ‘‘dominant protective association.’’

5. THE COUNTERARGUMENT: INALIENABLE RIGHTS We have seen that the debate from antiquity up to the present about slavery and autocracy was not a simple consent-versus-coercion debate. Instead, there were consent-based arguments for slavery and non-democratic government as being based on certain explicit or implicit contracts. What were the counterarguments against such contracts in the abolitionist and democratic movements? The most basic counterarguments were not about efficiency and were not quibbles about the reality or quality of the consent to explicit or implicit contracts. Instead, arguments developed that there was something inherently invalid in such alienation or translatio contracts, and thus that the rights which these contracts pretended to alienate were in fact inalienable. There is a simple logical structure to the inalienable rights arguments against the alienation contracts, such as the self-sale contract or the pact of subjection. In consenting to such a contract a person was agreeing to, in effect, take on the legal role of a non-person or thing. Yet all the consent in the world would not in fact turn the person into a thing. The person would play an agreed upon role such as obeying the master, and the authorities would ‘‘count’’ that as ‘‘fulfilling’’ the contract to have the role of a nonperson. Then the rights and obligations are assigned as per the contract as it were actually fulfilled. Since the person remained a de facto person with only the contractual role of a thing, the contract was impossible and invalid. A system of positive law that accepted such contracts was only a fraud on an institutional scale. But how are persons and non-persons differentiated? Persons and things can be distinguished on the basis of decision-making and responsibility. For instance, a thing such as a tool can be alienated or transferred from person A to B. Person A can indeed give up making decisions about the use of the tool and person B can take over making those decisions. Similarly, person A does not have the responsibility for the consequences of the employment of the tool by person B. Person B makes the decisions about using the tool and has the de facto responsibility for the results of that use. Thus a contract

332

DAVID ELLERMAN

to sell or rent a tool from A to B can be fulfilled. The decision-making and responsibility for employing the tool can in fact be transferred from A to B. But now replace the tool by person A himself or herself. Suppose that the contract was for person A to sell or rent himself or herself to person B – as if a person was a transferable or alienable instrument or tool. The contract could be perfectly voluntary. For whatever reason and compensation, person A is willing to take on the legal role of a talking instrument (to use Aristotle’s phrase). But the person A cannot in fact transfer decision-making or responsibility over his or her own actions to B. At most, person A can agree to cooperate with B by doing what B says. But that is no alienation or transference of decision-making or responsibility. Person A is still inexorably involved in ratifying B’s decisions and person A inextricably shares the de facto responsibility for the results of A’s and B’s joint activity. Yet a legal system could ‘‘validate’’ such a contract and could ‘‘count’’ obedience to the master or sovereign as ‘‘fulfilling’’ the contract. But such an institutionalized fraud always has one revealing moment where even the most slavishly conforming observers can see the fiction behind the system. That is when the legalized thing would commit a crime. Then the ‘‘thing’’ would be suddenly metamorphosed back into being a person to be held legally responsible for the crime. For instance, an antebellum Alabama court asserted that slaves are rational beings, they are capable of committing crimes; and in reference to acts which are crimes, are regarded as persons. Because they are slaves, they are y incapable of performing civil acts, and, in reference to all such, they are things, not persons (Catterall, 1926, p. 247).

Since there was no legal theory that slaves physically became things in their ‘‘civil acts,’’ the fiction involved in treating the slaves as ‘‘things’’ was clear. And this is a question of the facts about human nature; facts that are unchanged by consent or contract. If the slave had acquired that legal role in a voluntary contract, it would not change the fact that the slave remained a de facto person with the law only ‘‘counting’’ the slave’s non-criminous obedience as ‘‘fulfilling’’ the contract to play the legal role of a nonresponsible entity, a non-person, or a thing. The key insight is the difference in the factual transferability of a thing’s services and our own actions – the person–thing mismatch. I can voluntarily transfer the services of my shovel to another person so that the other person can employ the shovel and be solely de facto responsible for the results. I cannot voluntarily transfer my own actions in like manner. Thus the contract to rent out my shovel is a normal contract that I fulfill by transferring

Whither Self-Management? Finding New Paths to Workplace Democracy

333

the employment of the shovel to its employer. I can certainly voluntarily agree to a contract to be ‘‘employed’’ in like manner by an ‘‘employer’’ on a long- or short-term basis, but I cannot in fact ‘‘transfer’’ my own actions. Where the legal system ‘‘validates’’ such contracts, it must fictitiously ‘‘count’’ my inextricably co-responsible co-operation with the ‘‘employer’’ as fulfilling the employment contract – unless, of course, the employer and employee commit a crime together. The servant in work becomes the partner in crime. All who participate in a crime with a guilty intent are liable to punishment. A master and servant who so participate in a crime are liable criminally, not because they are master and servant, but because they jointly carried out a criminal venture and are both criminous (Batt, 1967, p. 612).

When the ‘‘venture’’ being ‘‘jointly carried out’’ by the employer and employee is not criminous, then the facts about human responsibility are unchanged. But then the fiction takes over. The joint venture or partnership is transformed into the employer’s sole venture. The employee is legally transformed from being a co-responsible partner to being only an input supplier sharing no legal responsibility for either the negative or positive results of the employer’s business.

5.1. Some Intellectual History of the Person–Thing Mismatch Where has this key insight – that a person cannot fit the legal role of a thing – erupted in the history of thought? The Ancients did not see this matter clearly. For Aristotle, slavery was based on ‘‘fact’’; some people were ‘‘talking instruments’’ – marked for slavery ‘‘from the hour of their birth.’’ Treating them as slaves was no more inappropriate for Aristotle than treating a donkey as an animal. The Stoics held the radically different view that no one was a slave by their nature; slavery was an external condition juxtaposed to the internal freedom of the soul. After the Dark Ages, the Stoic doctrine that only the body is enslaved and that the ‘‘inner part cannot be delivered into bondage’’ (Seneca quoted in Davis, 1966, p. 77) re-emerged in the Reformation doctrine of liberty of conscience. Secular authorities who try to compel belief can only secure external conformity. Besides, the blind, wretched folk do not see how utterly hopeless and impossible a thing they are attempting. For no matter how much they fret and fume, they cannot do more

334

DAVID ELLERMAN

than make people obey them by word or deed; the heart they cannot constrain, though they wear themselves out trying. For the proverb is true, ‘‘Thoughts are free.’’ Why then would they constrain people to believe from the heart, when they see that it is impossible? (Luther, 1942 (1523), p. 316).

Martin Luther (1942 (1523)) was explicit about the de facto element; it was ‘‘impossible’’ to ‘‘constrain people to believe from the heart.’’ Furthermore, every man is responsible for his own faith, and he must see it for himself that he believes rightly. As little as another can go to hell or heaven for me, so little can he believe or disbelieve for me; and as little as he can open or shut heaven or hell for me, so little can he drive me to faith or unbelief. Since, then, belief or unbelief is a matter of every one’s conscience, and since this is no lessening of the secular power, the latter should be content and attend to its own affairs and permit men to believe one thing or another, as they are able and willing, and constrain no one by force (p. 316).

Leaving aside some intermediate figures, we might jump over to Francis Hutcheson (1755), the predecessor of Adam Smith in the chair in moral philosophy in Glasgow and one of the leading moral philosophers of the Scottish Enlightenment. The inalienability argument is developed in Hutcheson’s influential A System of Moral Philosophy. Our rights are either alienable, or unalienable. The former are known by these two characters jointly, that the translation of them to others can be made effectually, and that some interest of society, or individuals consistently with it, may frequently require such translations. Thus our right to our goods and labours is naturally alienable. But where either the translation cannot be made with any effect, or where no good in human life requires it, the right is unalienable, and cannot be justly claimed by any other but the person originally possessing it (1755, p. 261).

Hutcheson appeals to the inalienability argument in addition to utility. He contrasts de facto alienable goods where ‘‘the translation of them to others can be made effectually’’ (like the aforementioned shovel) with factually inalienable faculties where ‘‘the translation cannot be made with any effect.’’ Hutcheson goes on to show how the ‘‘right of private judgment’’ or ‘‘liberty of conscience’’ is inalienable. Thus no man can really change his sentiments, judgments, and inward affections, at the pleasure of another; nor can it tend to any good to make him profess what is contrary to his heart. The right of private judgment is therefore unalienable (pp. 261–262).

Hutcheson pinpoints the factual non-transferability of private decisionmaking power. In the case of the criminous employee, we saw how the employee ultimately makes the decisions himself (through ratification and voluntary obedience) in spite of what is commanded by the employer. Short of coercion, an individual’s faculty of judgment cannot in fact be short circuited by a secular or religious authority.

Whither Self-Management? Finding New Paths to Workplace Democracy

335

A like natural right every intelligent being has about his own opinions, speculative or practical, to judge according to the evidence that appears to him. This right appears from the very constitution of the rational mind which can assent or dissent solely according to the evidence presented, and naturally desires knowledge. The same considerations show this right to be unalienable: it cannot be subjected to the will of another: tho’ where there is a previous judgment formed concerning the superior wisdom of another, or his infallibility, the opinion of this other, to a weak mind, may become sufficient evidence (1755, p. 295).

Democratic theory carried over this theory from the inalienability of conscience to a critique of the Hobbesian pactum subjectionis, the contract to alienate and transfer the right of self-determination as if it were a property that could be transferred from a people to a sovereign. Like the mind’s quest for religious truth from which it was derived, self-determination was not a claim to ownership which might be both acquired and surrendered, but an inextricable aspect of the activity of being human (Lynd, 1969, pp. 56–57).

Or as Ernst Cassirer (1963) puts it: There is, at least, one right that cannot be ceded or abandoned: the right to personality y They charged the great logician [Hobbes] with a contradiction in terms. If a man could give up his personality he would cease being a moral being.y There is no pactum subjectionis, no act of submission by which man can give up the state of free agent and enslave himself. For by such an act of renunciation he would give up that very character which constitutes his nature and essence: he would lose his humanity (p. 175).

In the Scottish Enlightenment, the jurist, George Wallace (or Wallis) applied the inalienability argument to slavery. Wallace asserted that: ‘‘Men and their liberty are not in commercio; they are not either saleable or purchasable.’’ He then continues: For these reasons, every one of those unfortunate men, who are pretended to be slaves, has a right to be declared free, for he never lost his liberty; he could not lose it; his prince had no power to dispose of him. Of course, the sale was ipso jure void. This right he carries about with him, and is entitled every where to get it declared. As soon, therefore, as he comes into a country, in which the judges are not forgetful of their own humanity, it is their duty to remember that he is a man, and to declare him to be free (Wallace, 1760, pp. 95–96).

Wallace’s statement illustrates the interplay between de facto and de jure elements, an interplay that is central to understanding the inalienability argument. When he declares that the slave has ‘‘never lost his liberty; he could not lose it,’’ that refers to the slave’s de facto retention of his free will and decision-making capacity (as recognized, for example, in the example of the criminous slave). Yet the law can declare a slave purchase contract as valid, and take a slave’s obedience as fulfilling the contract to be a chattel.

336

DAVID ELLERMAN

Since the slaves remain a de facto human agents in the de jure role of a thing, they are only ‘‘pretended to be slaves’’ by the legal authorities (at least until the slaves commit crimes). The inalienability argument and the legal pretense involved in a voluntary slavery contract has been echoed by the modern Kantian philosopher, George Schrader. Other people by their existence make a demand on us to acknowledge and treat them as persons rather than as things. This is a demand, incidentally, which no man can forfeit by his own volition. No man can, for example, by selling himself as a slave make himself not to be a person (Schrader, 1960, p. 64).

As Wallace noted, there is an element of pretense in the relationship since the slaves remain de facto persons. A man remains a man no matter what his condition in the world. He may not demand in any verbal way that he be treated as a man; in fact, he may even recommend that his humanity be disregarded. But the fact that he continues to exist as a man entails that his claim upon us as a human subject has not been removed.y [N]o man can actually make himself or another to be merely a slave; he can only make him play the role of a slave. It is not difficult to exhibit the deception and bad faith involved in such a relationship (Schrader, 1960, p. 64).

In the American Declaration of Independence, ‘‘Jefferson took his division of rights into alienable and unalienable from Hutcheson, who made the distinction popular and important’’ (Wills, 1979, p. 213). But the theory behind the rhetoric of ‘‘inalienable rights’’ was lost in the transition from the Scottish Enlightenment to the slave-holding society of antebellum America.

5.2. Application to the Employment Contract Today, the inalienability theory has to be recovered from its roots in the critique of the liberal contractarian defense of slavery and autocracy. By dumbing down the question to ‘‘consent or coercion,’’ modern liberalism has lost sight of both the liberal arguments for slavery and autocracy as well as the counterarguments in the form of inalienable rights theory. Once recovered, it is seen that the inalienability arguments apply as well to the individual self-rental contract and the collective pactum subjectionis of the workplace, the individual and collective versions of the employment contract. The mismatch of a person in a non-responsible ‘‘thing’’ role and the non-transferability of decision-making and responsibility apply as well for 8 hours a day as for a lifetime of labor.

Whither Self-Management? Finding New Paths to Workplace Democracy

337

The abolition of the employment relation is a radical conclusion that will be strongly resisted on every front. After the abolition of slavery and the acceptance of political democracy, liberal societies prided themselves on having finally gotten human rights right. Hence, there is strong intellectual resistance to giving any sustained thought to the idea that there might be an inherent rights violation in a liberal economic system based on the voluntary renting of human beings. There is certainly resistance to recovering the hidden history of liberal contractarian arguments for slavery and autocracy – and even to recovering the inalienable rights contra-arguments against those contracts. Today’s employment contract is only the rental version of the self-sale contract and is only the workplace version of the social contract of subjection. Very little sustained thought is necessary to understand the arguments. Take, for example, the approach to the employment contract as the workplace pactum subjectionis. The key to the intellectual history was to understand the distinction between two opposite types of social contract. On the one side was the social contract wherein a people would alienate and transfer their rights of self-determination to a sovereign. The sovereign was not a delegate, representative, or trustee for the people. The sovereign ruled in the sovereign’s own name; the people were subjects. One the other side was the idea of a social contract as a democratic constitution erected to secure the inalienable rights rather than to alienate them. Those who wield political authority over the citizens do so as their delegates, representatives, or trustees; they govern in the name of the people. Now once one understands this fundamental distinction between the alienation and the delegation social contracts, what additional information is needed to make the application to the employment contract? Does anyone think that the employer is the delegate, representative, or trustee of the employees? Does anyone think that the employer manages in the name of the employees?3 Since the answers are so blindingly obvious, the usual response is apparently just to not think about it. Just don’t go there. Then one can fall back on the consent or coercion framework. Democracy is government by the consent of the governed, and the employees give their consent to the employment contract so where is the problem? The other approach to the employer–employee contract is as the contract to hire or rent oneself out – just as one might rent out an instrument or tool. The counterargument was that people cannot in fact transfer the employment of themselves to an employer as they can the employment of a tool. The employer cannot be solely de facto responsible for the results as if the employees were only non-responsible tools. This is again blindingly obvious

338

DAVID ELLERMAN

and fully recognized by the law when the employer and employee commit a crime. Does anyone really think that employees morph into non-responsible instruments when their actions are not criminous? How can one avoid the conclusion that the employees and working employers are jointly de facto responsible for the results of their enterprise? Again, it is better not to think about it. There are many ‘‘stories’’ in conventional economics and legal theory to help one avoid thinking about these issues. As Luther himself emphasized, the mind cannot be forced to go where it does not want to go. Without going through all the ‘‘ducking and diving,’’ perhaps the basic story goes something like this. In the employment contract, the employees know well what is expected of them; in return for their compensation, they are to obey the employer within the scope of the employment contract (which, incidentally, would not include crimes). Similarly, the employer knows that he or she, in return for paying the compensation, is acquiring the right to decide what the employees are to do within the scope of the contract. Everyone knows what everyone else is expected to do. There is no language in the contract about temporarily playing the role of a ‘‘thing’’ or anything like that. It is a perfectly straightforward voluntary contract. When the two parties both fulfill their part (paying compensation and following directions), then the contract is fulfilled and that is the end of the matter.

This is the simple ‘‘obey and get paid’’ story that one should tell oneself to avoid thinking too hard about the actual structure of rights in the employment firm and to avoid thinking about the R-word (responsibility). Before taking note of the rights structure, it is useful to see that the same kind of story can be told about a voluntary master–slave relationship based on selling labor by the lifetime. In the slavery contract, the slaves know well what is expected of them; in return for their consideration (e.g., purchase price and perhaps ongoing food, clothing, and shelter), they are to obey the master within the scope of the slavery contract (which, incidentally, would not include crimes). Similarly, the master knows that he or she, in return for providing the consideration, is acquiring the right to decide what the slaves are to do within the scope of the contract. Everyone knows what everyone else is expected to do. There need be no language in the contract about playing the role of a ‘‘thing’’ or anything like that. It is a perfectly straightforward voluntary contract. When the two parties both fulfill their part (paying consideration and following directions), then the contract is fulfilled and that is the end of the matter.

When the antebellum law talked about slaves having the legal role of ‘‘things,’’ that was an unnecessary extravagance. A slavery contract would need no such language; it is a straightforward quid pro quo, the consideration is given in return for obedience – till death do they part.

Whither Self-Management? Finding New Paths to Workplace Democracy

339

5.3. Structure of Rights in the Employment Firm To see that the employees (or the slaves) have the legal role of nonresponsible entities, one has to apply some analysis to take the mind where the mind does not want to go. In a proprietorship, the proprietor has the legal responsibility (both positive and negative) for the results of the proprietor’s de facto responsible actions. That is, the proprietor is liable for the costs of the used up inputs and the proprietor may claim and sell the output that is produced. Thus the proprietor does not have a non-responsible role. Similarly for the partners in a partnership. In an owner-operated corporation, the corporation is a legal person separate from the owner or owners as individuals. Thus, when the working owner or owners carry out the work of the company, it is technically the company as a separate legal person that has the legal liability for the used up inputs and the legal claim on the produced outputs. But the owners are the legal members of that company, so through their corporate embodiment they have the legal responsibility for the positive and negative results of their de facto responsible actions. In economics, this is sometimes called the role of the ‘‘residual claimant’’ (liable for the input costs and getting the output revenues whose net is the residual). Thus, the owner–operators of a company do not have the legal role of a non-responsible entity or thing. We have seen that the employees are inextricably de facto co-responsible along with the working employers for the results (positive and negative) of their enterprise. But the employees as employees are not legal members of the company. Yet it is the company that has the legal liability for the used up inputs (the employees’ labor simply counting as one of those inputs) and the legal claim on the produced outputs. Since the employees are not legal members of that corporate body, they have no share of the legal responsibility for either the positive or negative fruits of their de facto responsible actions. Thus, it is that the employees take on a legally non-responsible role in the employment contract in spite of there being no language to that effect in the labor contract and in spite of their continuing de facto responsibility. It is important to compare the employees’ role with that of the other factor suppliers who supply to the company actual things or the services of things such as land, machinery, intermediate goods, or loan capital. Those factor suppliers also are not legal members of the company, so they bear none of the legal liability for the costs (their supplied inputs being one of the costs) and have no legal claim on the outputs. Here is where the difference in the factual transferability of persons and things comes into the analysis. The suppliers of things can alienate and transfer their inputs to the employer, so

340

Table 1.

DAVID ELLERMAN

Responsibility for the (Positive and Negative) Results of a Company.

Has legal responsibility Has no legal responsibility

Has de facto Responsibility

Has no de facto Responsibility

Working members of company Employees of company

Absentee ‘‘members’’ (shareholders) of company Suppliers of things to company

those factor suppliers have no de facto responsibility for the employer’s use of the factors. In the following 2 2 table (Table 1), of the four combinations of having or not having legal responsibility and having or not having de facto responsibility, there is only one remaining category to mention, those who have legal responsibility without de facto responsibility for the results of the enterprise. They are the absentee shareholders (persons or institutions) who do not work or otherwise have an active personal role in the enterprise. Yet they like the working shareholders who are also the legal ‘‘members’’ of the company and are thus residual claimants through their corporate embodiment (not as separate persons). All of this analysis of the rights and responsibilities complicates the picture far beyond the simplistic ‘‘obey and get paid’’ story about the employment contract. The employment contract does not have to ‘‘say’’ that the employee takes on a non-responsible role. As long as the legal system accepts the employees’ obedience as fulfilling the contract in return for the wages, then the employer (e.g., the employing corporation) bears all the liabilities for the inputs and thus has the legal claim on the produced outputs. Thus without any such language in the contract, the employees are excluded from any legal responsibility or residual claimancy role (e.g., corporate membership) in spite of their de facto responsibility. Thus persons take on the legal role of non-responsible entities or things in what is conventionally seen as ‘‘a perfectly straightforward voluntary contract.’’ The negative conclusion is that the employment contract should be recognized as being jurisprudentially invalid. Human decision-making and responsibility is in fact not transferable so the contract for the sale of human actions (labor) is inherently invalid. On the positive side, there is the basic juridical principle of responsibility that legal responsibility should be imputed in accordance with de facto responsibility (insofar as possible since accidents may not have a de facto responsible party). In short, the responsibility principle implies that there

Whither Self-Management? Finding New Paths to Workplace Democracy

341

should be no off-diagonal elements in the above table. The people who work in a firm should be the legal members of the firm, and the people who only supply things to the firm should not be members of the firm. All firms should be democratic self-managed firms. 5.4. Other Rights-Violating Voluntary Contracts Sometimes it is easier to understand an argument if one sees parallel applications – in this case, to other types of contracts which would involve inherent rights violations. One of the contracts has been historically realized and the others are hypothetical. A historical example of this sort of institutionalized fiction was the older and now legally invalid coverture marriage contract that ‘‘identified’’ the legal personality of the wife with that of the husband. By marriage, the husband and wife are one person in law: that is, the very being or legal existence of the woman is suspended during the marriage, or at least is incorporated and consolidated into that of the husband; under whose wing, protection, and cover, she performs everything; and is therefore called in our law-French, a femme covert, and is said to be under the protection and influence of her husband, her baron, or lord; and her condition during her marriage is called her coverture (Blackstone, 1959 (1765), p. 83, Section on ‘‘Husband and Wife’’).

The baron–femme relationship established by the coverture marriage contract exemplified the identity fiction in past domestic law. A female was to pass from the cover of her father to the cover of her husband (as in the vestigial ceremony where the bride’s father ‘‘gives away’’ the bride to the groom) – always a ‘‘femme covert’’ instead of the anomalous ‘‘femme sole.’’ The identity fiction for the baron–femme relation was that ‘‘the husband and wife are one person in law’’ with the implicit or explicit rider, ‘‘and that one person is the husband.’’ A wife could own property and make contracts, but only in the name of her husband. Again, obedience counted as ‘‘fulfilling’’ the contract to have the wife’s legal personality subsumed under and identified with that of the husband.4 The coverture marriage contract is now outlawed in favor of the partnership version of the marriage contract but one could imagine a modernized sex-neutral dependency contract. One adult with full capacity would voluntarily agree to become a non-adult ‘‘dependent’’ of another adult, the ‘‘guardian’’ or ‘‘sponsor,’’ in return for whatever consideration. The independent and adult legal personality of the ‘‘dependent’’ would be ‘‘suspended’’ in favor of the sponsor. The dependent could only make contracts and hold property under the name of the sponsor. The relevant identity

342

DAVID ELLERMAN

fiction would be: ‘‘the sponsor and dependent are one person in law – and that one person is the sponsor.’’ The language of the contract could be adjusted so as not to offend modern sensitivities as long as it was understood what the contract means. Obedience by the adult dependent to the sponsor would count as ‘‘fulfilling’’ this contract to be a ‘‘non-adult.’’ This hypothetical modernized sex-neutral version of the coverture contract would be invalid for the same reasons as the original coverture contract, the self-sale contract, or the employment contract. The adult ‘‘dependent’’ remains a de facto adult, the law would accept obedience to the sponsor as ‘‘fulfilling’’ the contract, and then the legal rights would be assigned accordingly as if the ‘‘dependent’’ was actually a non-adult. Another imagined example of an invalid alienation contract would be the consumptive employment contract. Typically, a consumer buys consumption goods, self-manages his or her own consumption activity, and takes the liability for the consumed good as well as appropriates any waste product. The ‘‘employment’’ concept could be applied to consumption by having the consumer instead of paying for the consumer good, to pay an ‘‘employer’’ to ‘‘employ’’ the consumer in the consumption of the good. Then the employer would legally take on the legal liability for the used-up input to the consumption process, would manage the process, and would appropriate any waste product that might be sold, say, to a recycling center. Instead of the employer paying the productive-employee to transfer value-adding labor, the consumptive-employee is paying the employer to accept the transfer of the value-subtracting consumption activity (like a negative form of labor). While some contracts between the elderly and nursing homes might seem to be of this type, this type of consumption employment contract seems little used. In any case, the critique of the production employment contract would apply as well to the consumption version. Responsible human action, net value adding or net value subtracting, is not de facto transferable. In spite of the abundance of legal precedent in the historical contracts such as the self-sale contract, the pactum subjectionis, and the coverture marriage contract, today’s employment contract, and even some hypothetical contracts (the ‘‘dependency contract’’ and the consumptive employment contract), legal theory has yet to focus on this general notion of a voluntary contract for a fully capacitated person to take on a lesser legal role. All these contracts have the same scheme. An adult person with full capacity voluntarily agrees for whatever reason and in return for whatever consideration to accepting a lesser legal role. But they do not in fact alienate their capacity as a person in order to fulfill that lesser legal role. Instead, the law accepts their (non-criminous) obedience to the master as ‘‘fulfilling’’ the contract. Then

Whither Self-Management? Finding New Paths to Workplace Democracy

343

the rights and obligations follow the legal role (e.g., the slave of a master, the subject of a sovereign, the femme covert of her baron, the employee of the employer, and so forth) – as if the person were not in fact a person at all or were not a person of full capacity. The whole scheme amounts to a fiction and fraud on an institutional scale that nonetheless parades upon the historical stage as a contractual institution based on consent. 5.5. New Paths and New Perspectives ‘‘Self-management’’ is associated with ‘‘socialism’’ and ‘‘socialism’’ is associated historically in liberal society with the denial of democracy. Yet we have seen almost the opposite. Liberalism exhibits a learned ignorance of the long history of liberal contractarian defenses of slavery and nondemocratic governments as being based on consent. And liberalism also has ‘‘lost’’ the inalienability theory hammered out in the anti-slavery and democratic movements. Instead, the basic question is posed in liberalism as consent versus coercion. Fundamentally, there are only two ways of co-ordinating the economic activities of millions. One is central direction involving the use of coercion – the technique of the army and of the modern totalitarian state. The other is voluntary co-operation of individuals – the technique of the market place (Friedman, 1962, p. 13).

Since democracy is pictured as being ‘‘government based on the consent of the governed’’ and since the employment firm is also based on consent, there are no basic rights violations in modern liberal society. Contrary to the blinkered vision of liberal apologetics, we have seen that the subtle issues lie all within the domain of consent (little subtlety is required to be against coercion). The ‘‘consent of the governed’’ to a Hobbesian pactum subjectionis is not democracy, and the employment contract is the mini-Hobbesian contract for the workplace. Thus once the question is posed as consent-to-alienation versus consent-to-delegation, then the daunted affinity of ‘‘liberal-capitalism’’ with democracy disappears in its acceptance of the employment firm. A true affinity to democracy would entail not only the abolition of the pactum subjectionis in the political sphere in favor of democratic constitutions but the abolition of the employment contract in favor of all firms and other organizations being organized as workplace democracies. A similar reversal occurs concerning property rights. ‘‘Self-management’’ is associated with ‘‘socialism’’ and ‘‘socialism’’ is historically associated with the denial of private property rights. Yet we have seen almost the opposite. A basic principle in jurisprudence is the responsibility principle

344

DAVID ELLERMAN

that whenever possible legal responsibility should be assigned or imputed according to the de facto responsible party. For instance, in a trial the idea is to make an official decision on the factual question of whether or not the defendant is the de facto responsible party. If so, then legal responsibility is imputed accordingly. But the principle is general. The more positive application of the responsibility principle is the old idea often associated with John Locke that people should appropriate the fruits of their labor. This labor theory of property is both positive and negative since new products are only produced by using up other things as inputs. Hence the question of assigning legal responsibility is two sided, to assign the ownership of the product and the liability for the used-up inputs to the people who, by their de facto responsible actions, produced the outputs by using up the inputs. Hence a private property system based on the basic principle of justice (imputing to people what they are responsible for) would have the legal members of each firm exactly the people who work in the firm. Thus, a system based on justice in private property would entail workplace democracy. In contrast, the current system based on the employment contract ends up not assigning the assets and liabilities created in production to those who created them but rather assigns them to a largely absentee class of shareholders who did not create them. It is also time to move beyond the simplistic liberal morality play of consent-or-coercion and to see the whole history of subjection linked to the consent-based apologetics of the intellectual clerks of the past and present. And it is time to understand the deeper intellectual history of the anti-slavery and democratic movements based on the enduring idea that persons do not fit into the legal role of things – even with consent. This theory, overlooked by modern liberalism, is part of our own history. It bequeaths to our day the call for the abolition of the whole institution of renting human beings. It is time for new paths. New paths indeed. It is precisely the principles of democracy and justice in property that call for the abolition of the employment contract in favor of a private property market economy of democratic firms.

6. CONCLUDING REMARKS: GETTING FROM HERE TO THERE 6.1. The Approach through Employee Ownership I will now turn to the more practical questions of getting from here to there, from our current economic system very much based on the unchallenged

Whither Self-Management? Finding New Paths to Workplace Democracy

345

employment relation to a system of economic democracy where all firms are democratic firms. One strategy is through employee ownership, the firm by firm creation of more-or-less democratic firms by start-ups or buy-outs using legal forms such as worker cooperatives or joint-stock corporations with employee stock ownership plans (ESOPs). This strategy is important for the simple reason that it fosters local experimentation. In spite of all the public rhetoric about political democracy, our society is actually rather anti-democratic in what people do all day long, their work. The very idea of the firm as a small democratic polity seems to be so new and strange that much experimentation is needed to find out how it might work. The downside of the employee-ownership approach is that it ‘‘buys into’’ the fundamental myth that the enterprise is a piece of property with owners. This is a basic misunderstanding of a private property market economy. Land, buildings, machinery, and financial capital all have owners. There are two quite different ways in which the owners of capital can use it. The owners can try to hire in a complementary set of inputs including the ‘‘labor input’’ and then undertake a certain productive opportunity. Or the owners of capital can hire the capital out to others who may undertake production. Who ends up undertaking production is determined by how those contracts are made – by who ends up hiring what or whom. Thus, the undertaking of production – being the firm – is a contractual position, not a pre-existing property right that was part of the ownership of capital. This is rather confusing since the legal parties that make the contracts to undertake production are typically corporations, and corporations do have owners. Hence we hear the loose expressions about the ‘‘ownership of the firm’’ or ‘‘ownership of the enterprise.’’ But there is no ‘‘ownership’’ of a pattern of contracts. Absent the contracts to hire in other factors or even to hire out the corporate assets, the corporation is only an input owner. It is the contracts that determine if a corporation becomes an enterprise undertaking production or a supplier of inputs to other parties who are undertaking productive opportunities. At one point the Studebaker corporation leased one of its factories to Chrysler Corporation. Studebaker ‘‘owned the means of production’’ (the factory) but Chrysler was the firm or enterprise undertaking production using those capital assets. The idea that the ownership of capital includes the ‘‘ownership of the firm’’ using that capital as the fundamental myth of our present economic system. Understanding this myth is the pons asinorum5 in understanding property rights. That is why the name ‘‘capitalism’’ is a misnomer for the

346

DAVID ELLERMAN

employment system. ‘‘Capital’’ does not include the ‘‘ownership of the enterprise.’’ The question of who is the firm (not who ‘‘owns’’ the firm) is determined by the contracts. And the characteristic feature of our present system is the voluntary contract for the renting of the people working in an enterprise, the employment contract. Hence, I have used the phrase ‘‘employment system’’ although older names such as ‘‘wage system’’ or ‘‘wage labor system’’ are also better than the misnomer ‘‘capitalism.’’ Part of the road to workplace democracy is crossing the pons asinorum to see what is actually owned and what is not owned in our current system. Democracy is not at war with private property but with the employment contract to rent human beings. The approach to workplace democracy through employee ownership does not really cross that bridge; it tends to ‘‘buy into’’ the idea of the employees as the new ‘‘owners of the firm.’’ A democratic firm is not a corporation where its class of owners or shareholders is (temporarily) co-extensive with its employees. In an analogous way, manumission (buying a slave’s freedom) is a confused strategy to abolish slavery since it confirms the underlying ownership of the slave. In the employee ownership case, there is even confusion about just what ‘‘ownership’’ is being purchased. Since the other parties to a corporation’s contracts are free to contract with other partners in the future, one cannot ‘‘buy’’ for the future the contractual position that the corporation had in the past. When one buys a corporation, one is actually buying only a set of capital assets organized in a corporate body. One hopes that the ‘‘goodwill’’ of the contractual partners will continue beyond their current contracts but that is not ‘‘purchased’’ along with the corporate shares. Setting aside all the myths and confusions, the heart of a democratic worker ‘‘buy-out’’ is a reversal in the contract between capital and labor where current labor has also become a supplier of some of the capital. In a company reorganized on a democratic basis, a person becomes a member solely on the basis of working in the company, not on the basis of supplying capital. Nevertheless, worker-members may supply capital as part of the original deal or as a membership fee, and they will acquire more capital as part of their labor-based earnings is retained in the firm rather than paid out. This member-supplied capital is recorded in a set of internal capital accounts that bear interest like any other borrowed capital and that is paid out over a period of time. Since the worker’s membership rights (voting and net revenue rights) are based solely on their human participation in the firm (i.e., their de facto responsible activity in the firm), those rights are independent of the capital in their capital account.

Whither Self-Management? Finding New Paths to Workplace Democracy

347

These questions have come up in a poignant way in my own past work on employee buy-outs in the Industrial Cooperative Association and elsewhere. It is safe to say that most or all employees come into a buy-out with full belief in the fundamental myth. They are ‘‘buying the firm’’ so they will then be the ‘‘owners of the firm’’ regardless of whether or not they work in the firm in the future. At what point does the democratic consultant try to persuade the workers to structure the new company so that they are not the new ‘‘owners’’ who may rent future workers. When they think they are ‘‘buying the firm,’’ is it some type of ‘‘bait and switch’’ to then suggest a new company structure where both present and future workers will be members based on their labor and that their membership rights will be independent of the capital in their accounts? These are some of the practical problems in trying to use ‘‘employee ownership’’ – with its attendant intellectual baggage – as a path to workplace democracy. In spite of having spent many years taking this path, I now think that, for both practical and theoretical reasons, other paths should be explored. But before leaving the ‘‘employee-ownership’’ path, I might remark about a related path of ‘‘capital strategies’’ through various union-controlled or public-sector pension funds. Instead of letting management usurp the control rights ‘‘rightfully’’ belonging to the absentee owners of shares, these institutional investors are trying to use those rights to get corporations to do various worthwhile things. This is even being touted as a new form of ‘‘public accountability’’ in something like ‘‘pension fund socialism.’’ In terms of the analogy with slavery, this ‘‘capital strategy’’ is even worse than the manumission strategy of fighting slavery by buying slaves and freeing them. In this case, the idea is, in effect, for good and well-intended people to buy slaves and then to ‘‘do good’’ for them as slaves. If the capital strategy approach was to have any emancipatory potential then the ‘‘labor-oriented’’ institutional investors should use their influence on a company to transform the company into a democratic firm so that, among other things, it would not be ‘‘owned’’ by the absentee suppliers of capital. Since that is rather unlikely to happen, I must be very skeptical that much good will come from the well-meaning absentee suppliers of capital – regardless of the ‘‘labor-oriented’’ label. Since the idea is to convert a firm into a democratic polity whose ‘‘citizens’’ or members are the people being managed in the firm, it is by no means clear that an ownership-based or capital-based strategy is best. Suppose that instead of a residency-based democratic government in a municipality that the city was ruled by the landowners using land-based voting

348

DAVID ELLERMAN

rights. One strategy to change the system would be for all the current residents who were landless and thus could not vote to use one means or another to buy the necessary land and thus get the vote. Then the idea is that they would use their land-based votes to change the system so that all current and future residents could vote regardless of land ownership. It is possible but unlikely for this to work due to means being used contradicting the originally sought-after end. The other strategy where the means contradict the end is a path to workplace democracy through collective bargaining by labor unions. In the late 19th century, there were reform unions who aimed to replace the ‘‘wage system’’ with the ‘‘cooperative commonwealth’’ (see Grob, 1969; Lasch, 1991). But throughout most of the 20th century, the labor movement has withdrawn that challenge and has tried only to get ‘‘more, more, and more’’ within the employment contract. In return, the role of unions in collectively bargaining the employment contract has been enshrined in labor legislation. Thus, trade unions have become part and parcel of the employment system. Any strategy that would expect the unions to work to abolish the employment system must be viewed with some skepticism.

6.2. Other Paths to Workplace Democracy What are the other paths to workplace democracy where the means may not contradict the end? One approach is German-style co-determination. Corporate law is changed so that employees in a company acquire what is typically a minority representation on the supervisory board of a company. Since this representation is independent of share ownership, this approach has the virtue of directly approaching the democratic structure of labor-based voting rights – at least for part of the board representatives. Full membership rights would also include net income rights, and then with the partial retention of net income, the workers would need some structure of internal capital accounts to keep track of their share of retained income. And finally the process could be completed by converting the absentee shares into some form of non-voting preferred shares or variable income bonds. This path to workplace democracy might have some potential in continental Europe. The post-war history in Japan has created another path. With the family ownership of the old zaibatsu groups discredited by the war, the American occupying reformers tried to set an Anglo-American system of stock market capitalism. But instead, the older corporate groups reconstituted themselves

Whither Self-Management? Finding New Paths to Workplace Democracy

349

with sizable cross-holdings of shares. Share cross-ownership within these new keiretsu groups was the method used to neutralize the control of absentee shareholders while keeping the legal structure of the Anglo-American joint stock company. The cross-owners within the group would delegate the control rights (for non-distressed operation) to each company’s management and all the shares were treated as fixed income preferred stock. Thus the membership rights in such firms were in some sense ‘‘internalized’’ to the firm. There was more a transformation in the internal corporate culture than in the legal structure. The managers recognized that they were trustees managing in the name of the company which in human terms meant the people who made up the company. But there was no actual legal structure to enact that accountability. With these caveats, it could be seen as a type of ‘‘employee sovereignty.’’ The fundamental principle underlying the Japanese model of mixed economy is anthropocentricism, or what Keisuke Itami refers to as ‘‘peoplism.’’ Peoplism is given concrete expression in the form of employee sovereignty with the corporation, and an emphasis on the independent, land-owning farmer within agriculture. This principle is clearly different from the ideological foundations of Western capitalism, and it would be incorrect to assume that the Japanese system belongs to the same regime just because it uses market mechanisms extensively and exists side by side with a democratic political system (Sakakibara, 1993, p. 4).

Ronald Dore (1987) has contrasted the company as community model in Japan with the Anglo-American concept as of the company as a piece of property. Koji Matsumoto has emphasized these labor-oriented aspects of the Japanese firm along with comparisons to Yugoslav self-managed firms. Although there is some danger of oversimplification in making such a statement, the most direct description of this situation is that Japanese corporations ‘are controlled by, and exist for, their employees’. Japanese corporations are thus united bodies of corporate employees (Matsumoto, 1991, p. 27).

It is not a coincidence that we have seen these two partial paths toward workplace democracy develop in the two countries where the old system was discredited by defeat in the World War II. Perhaps other paths will develop out of the discrediting of the old system in the post-socialist revolutions of 1989–1990. But it is too soon to tell since the first response has usually been a swing of the pendulum to the other extreme. It is a truism that each country has its unique history. As long as one gets away from the idea that there is One Best Way or blueprint for all private property market economies, then many local experiments will take place. For example, Korea industrialized with the family owned choebals playing a major role. There was no defeat in war to discredit the family owners but

350

DAVID ELLERMAN

there was some discrediting with the East Asian crisis. In any case, diminishing returns is being felt with the generational turnover. Korea is also a relatively small country that would eventually have its major companies under foreign ownership and control unless something is done. They might find their own version of the Japanese path to replace foreign control with insider control and more cooperative labor relations as the family ownership diminishes.

6.3. The Path through the Corporate Governance Debate Where does this leave the Anglo-American economies? I have emphasized that the labor contract is fundamentally different from the other input supply contracts. Human labor is not de facto transferable. Even though the employees have the legal role as the outside suppliers of an input, in fact they are the firm as an organization of people. This leads to a remarkable schizophrenia. There are two firms. There is the firm in the eyes of the law whose members are the shareholders scattered far and wide and who typically trade into the stock simply as an investment without any intent or capacity to play a human role in the firm. This is the firm that has a ‘‘meeting’’ once a year. In contrast to this de jure firm, there is the de facto firm consisting of the people who work in the firm – who have a meeting every working day to actually produce the product and conduct the business of the firm. Table 1 can be rephrased using these notions of the legal (de jure) company and the de facto company. Ordinary consciousness often reflects the de facto company. The employees are often thought of as the members of the organization. Consider the following from a perfectly standard managerial accounting textbook. An organization can be defined as a group of people united together for some common purpose. A bank providing financial services is an organization, as is a university providing educational services, and the General Electric Company producing appliances and other products. An organization consists of people, not physical assets. Thus, a bank building is not an organization; rather, the organization consists of the people who work in the bank and who are bound together for the common purpose of providing financial services to a community (Garrison, 1979, p. 2).

Garrison is talking entirely about the de facto company, not the company as it exists in law. This same redefinition of the firm now has roots in ordinary consciousness. Look at the books on the business shelves in your local bookstore. Find a book that uses some expression like ‘‘members’’ of the

Whither Self-Management? Finding New Paths to Workplace Democracy

351

company. Chances are that the author, like Garrison above, is referring to the employees (including managers) of the firm, not the far-flung shareholders (who are the legal ‘‘members’’). As public stock markets have spread shares far and wide, the idea that the stockholders are in any real sense the ‘‘owners’’ or ‘‘members’’ of a publicly traded company has become a sheer fantasy. There are two groups invested in keeping the fantasy alive. There are economists, lawyers, and assorted ideologues who will gladly adapt themselves to whatever is the ruling orthodoxy and vigorously defend it – until it changes. And there are the managers who have mightily profited from the eclipse of the shareholders. They have every interest in keeping the fantasy of ‘‘shareholders’ capitalism’’ alive as the cover story for the reality of managerial capitalism. John Maynard Keynes (1936) saw that the public stock markets had caused the ‘‘euthanasia of the rentier, of the functionless investor’’ (p. 376) while Adolf Berle and Gardiner Means described the result as the ‘‘separation of ownership and control’’ (Berle & Means, 1932, p. 89). But this delegitimization of the old ‘‘property’’ rationale for membership in the company has so far lead to more de facto managerialism than to a relegitimization of membership based on internal democracy. Here is the entry point for another path to workplace democracy. The separation of ownership and control along with the unaccountability of managers and the resulting abuses has created the ‘‘corporate governance problem.’’ Who is to be the new legitimate members of the company? While a few cherish the hope for the resurrection of the ‘‘responsible private owner’’ in massive institutional investors run by portfolio-managing bureaucrats, others search the horizon for various ‘‘stakeholders’’ who together with the regulatory agencies and law courts might create a ‘‘new accountability.’’ But they are searching for legitimacy in all the wrong places. There already is a class of members who make up the firm, the de facto firm consisting of the people who work in it. And since the staff of a company are the de facto firm, they are the ones who could actually monitor the management of their company and address the corporate governance problem directly. The only cohesive, workable, and effective constituency within view is the corporation’s work force (Flynn, 1973, p. 106).

And there already is a legitimacy norm for questions of governance and democracy. The notion of ‘‘shareholders’ democracy’’ is not only impractical but is incoherent since the shareholders are not themselves under the authority of the managers.

352

DAVID ELLERMAN

The analogy between state and corporation has been congenial to American lawmakers, legislative and judicial. The shareholders were the electorate, the directors the legislature, enacting general policies and committing them to the officers for execution.y Shareholder democracy, so-called, is misconceived because the shareholders are not the governed of the corporation whose consent must be sought (Chayes, 1966, pp. 39–40).

Many parties have their interests affected by a company and better judicial and regulatory means are needed to protect those legitimate outside interests. But those (de facto and de jure) outside parties are not governed or under the authority of company management. Only the de facto firm, the people working in a company are under the authority of the management within the scope of the employment contract. Hence the path to democracy in the workplace is to redefine the de jure firm so that it matched the de facto firm (which would eliminate the off-diagonal elements in Table 2 above). This was well put by the English conservative Lord Eustice Percy 60 years ago. Here is the most urgent challenge to political invention ever offered to the jurist and the statesman. The human association which in fact produces and distributes wealth, the association of workmen, managers, technicians and directors, is not an association recognised by the law. The association which the law does recognise – the association of shareholders, creditors and directors – is incapable of production and is not expected by the law to perform these functions. We have to give law to the real association and withdraw meaningless privilege from the imaginary one (Percy, 1944, p. 38; quoted in Goyder, 1961, p. 57).

Although the corporate governance problem has been with us throughout the 20th century, the path to workplace democracy through the corporate governance question seems even more appropriate today. It is a process of gradually reconstituting the corporation. All the work toward greater participation of workers in the affairs of their company points in this direction. In some respects, the psychological transformation is already there; the people who make up the de facto company are seen as the real members of the company. What remains is finding ways to effect the legal transformation, the withdrawing of privilege from the imaginary company of shareholders and the legal recognition of the actual members of the company. Table 2.

Who’s in and Who’s out of the Two Companies. Inside de facto Company

Inside legal company Outside legal company

Shareholders working in company Employees of company

Outside de facto Company Absentee ‘‘members’’ (shareholders) of company Suppliers of things to company

Whither Self-Management? Finding New Paths to Workplace Democracy

353

NOTES 1. The simulation argument also works the other way. When some economists claim that ‘‘hierarchy’’ is necessary for efficiency and thus a democratic firm is inefficient, they miss the possibility of simulating the hierarchical ‘‘efficiency’’ within a democratic firm. Moreover, as will be seen below, the question between an employment or democratic firm is not hierarchy or not but whether the managerial authority (‘‘hierarchy’’) is based on a contract of alienation or delegation. 2. I use ‘‘liberalism’’ in the European sense as ‘‘classical liberalism,’’ not in the American sense juxtaposed to conservativism. The fundamental tenet of liberalism is a society based on voluntary contract, not coercion (including ‘‘status’’ as a type of coercion). At the enterprise level, the characteristic feature of liberalism is the acceptance of the employment firm based on the employer–employee contract. I use the phrase ‘‘employment firm’’ rather than the misleading phrase ‘‘capitalist firm’’ since the characteristic feature of the conventional firm is the employment contract, not the private ownership of the means of production. More on this anon. 3. ‘‘The manager in industry is not like the Minister in politics: he is not chosen by or responsible to the workers in the industry, but chosen by and responsible to partners and directors or some other autocratic authority. Instead of the manager being the Minister or servant and the men the ultimate masters, the men are the servants and the manager and the external power behind him the master. Thus, while our governmental organisation is democratic in theory, and by the extension of education is continually becoming more so in practice, our industrial organisation is built upon a different basis’’ (Zimmern, 1918, p. 263). 4. Carole Pateman (1970), already well-known for her early work on selfmanagement and participatory democracy , has analyzed this sort of a ‘‘sexual contract’’ in a more general setting where she independently pointed out the connection to the employment contract and the de facto inalienability of labor. ‘‘The contractarian argument is unassailable all the time it is accepted that abilities can ‘acquire’ an external relation to an individual, and can be treated as if they were property. To treat abilities in this manner is also implicitly to accept that the ‘exchange’ between employer and worker is like any other exchange of material property.y The answer to the question of how property in the person can be contracted out is that no such procedure is possible. Labour power, capacities or services, cannot be separated from the person of the worker like pieces of property’’ (Pateman, 1988, pp. 147–150). 5. The ‘‘ass’ bridge’’ was the early theorem in Euclidean geometry curriculum such that if a student could not get across that bridge, then his talents were elsewhere.

REFERENCES Batt, F. (1967). The law of master and servant (5th ed.). London: Pitman. Berle, A., & Means, G. (1932). The modern corporation and private property. New York: MacMillan. Blackstone, W. (1959 (1765)). Ehrlich’s blackstone. In: J. W. Ehrlich (Ed.). New York: Capricorn Books (Original 1765).

354

DAVID ELLERMAN

Cassirer, E. (1963). The myth of the state. New Haven: Yale University Press. Catterall, H. T. (Ed.) (1926). Judicial cases concerning slavery and the Negro. Washington, DC: Carnegie Institute. Chayes, A. (1966). The modern corporation and the rule of law. In: E. S. Mason (Ed.), The corporation in modern society. New York: Atheneum. Christ, C. F. (1975). The competitive market and optimal allocative efficiency. In: J. Elliott & J. Cownie (Eds), Competing philosophies in American political economics (pp. 332–338). Pacific Palisades, CA: Goodyear. Corwin, E. S. (1955). The ‘Higher Law’ background of American constitutional law. Ithaca: Cornell University Press. Davis, D. B. (1966). The problem of slavery in Western culture. Ithaca: Cornell University Press. Dore, R. (1987). Taking Japan seriously. Stanford, CA: Stanford University Press. Flynn, J. J. (1973). Corporate democracy: Nice work if you can get it. In: R. Nader & M. J. Green (Eds), Corporate power in America (pp. 94–110). New York: Grossman Publishers. Friedman, M. (1962). Capitalism and freedom. Chicago: University of Chicago Press. Garrison, R. (1979). Managerial accounting. Dallas: Business Publications Inc. Gierke, O. v. (1958). Political theories of the Middle Age. Boston: Beacon Press. Gierke, O. v. (1966). The development of political theory. New York: Howard Fertig. Goyder, G. (1961). The responsible company. Oxford: Basil Blackwell. Grob, G. (1969). Workers and utopia. Chicago: Quadrangle. Hobbes, T. (1958 (Original 1651)). Leviathan. Indianapolis: Bobbs-Merrill. Hutcheson, F. (1755). A system of moral philosophy. London: Foulis. Keynes, J. M. (1936). The general theory of employment, interest, and money. New York: Harcourt, Brace & World. Lasch, C. (1991). The true and only heaven. New York: Norton. Locke, J. (1960). Two treatises of government. New York: New American Library (Original 1690). Luther, M. (1942 (1523)). Concerning secular authority. In: F.W. Coker (Ed.), Readings in political philosophy (pp. 306–329). New York: Macmillan (Original 1523). Lynd, S. (1969). Intellectual origins of american radicalism. New York: Vintage Books. Matsumoto, K. (1991). The rise of the Japanese corporate system: The inside view of a MITI official. London: Kegan Paul International. Nozick, R. (1974). Anarchy, state, and utopia. New York: Basic Books. Pateman, C. (1970). Participation and democratic theory. Cambridge: Cambridge University Press. Pateman, C. (1988). The sexual contract. Stanford: Stanford University Press. Percy, E. (1944). The unknown state: 16th Riddell memorial lectures. London: Oxford University Press. Rawls, J. (1971). A theory of justice. Cambridge: Harvard University Press. Sabine, G. H. (1958). A history of political theory. New York: Henry Holt and Company. Sakakibara, E. (1993). Beyond capitalism: The Japanese model of market economics. Lanham, MD: University Press of America. Samuelson, P. A. (1976). Economics (10th ed.). New York: McGraw-Hill. Schrader, G. A. (1960). Responsibility and existence. In: C. J. Friedrich (Ed.), Responsibility (pp. 43–70). New York: Liberal Arts Press.

Whither Self-Management? Finding New Paths to Workplace Democracy

355

Seabury, S. (1969). American slavery justified by the law of nature. Miami: Mnemosyne Publishing Company (Orig. 1861) Reprinted. Wallace, G. (1760). A system of the principles of the law of Scotland (Vol. I). Edinburgh. Wills, G. (1979). Inventing America. New York: Vintage Books. Zimmern, A. E. (1918). Nationality & government. London: Chatto & Windus.

This page intentionally left blank

356

THE FUTURE OF BROAD-BASED OPTIONS: WHAT RESEARCH TELLS US Corey Rosen ABSTRACT This paper looks at the research to date on the future of broadly granted stock options (options granted to at least half the full-time employees of a company). In the U.S., granting options broadly became popular in the late 1990s, but has lost some of its appeal in the wake of stock market declines, accounting changes, and increased shareholder concerns about dilution. The data indicate a significant minority of companies will change their plans, but a substantial majority will keep them. The data also indicate changes in accounting rules will not affect stock prices and that broadly granted options are better for corporate performance than narrowly granted options.

1. INTRODUCTION Employee ownership has been much in the news in the last several years. At the height of the dot.com boom, stories abounded of ordinary workers becoming millionaires from stock options after just a few years at a

Participation in the Age of Globalization and Information Advances in the Economic Analysis of Participatory and Labor-Managed Firms, Volume 9, 357–398 Copyright r 2006 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0885-3339/doi:10.1016/S0885-3339(05)09012-5

357

358

COREY ROSEN

company. When the boom collapsed, just as many stories told their tales of woe in seeing what was often just paper wealth evaporate. Soon after the century turned, Enron became enmeshed in financial scandals, destroying its stock and, with it, the retirement savings of many of its employees who were heavily invested in Enron shares in their 401(k) plans. Enron was soon joined by WorldCom, RiteAid, Lucent, and many others companies and, sadly, their employees. Meanwhile, the company most widely identified as employee owned, United Airlines, followed most of the rest of airline industry into severe financial difficulties. Yet at the same time, thousands of companies had converted to 100% employee-owned companies, with ownership broadly distributed among all the workers – and the large majority were doing very well indeed. The attention being paid to employee ownership is perhaps the most compelling testament to just how important this phenomenon has become in the U.S. and, increasingly, around the world. An estimated 39% of working U.S. adults who are employed by companies that have own stock in their employer through options, 401(k) plans, employee stock ownership plans (ESOPs), employee stock purchase plans (ESPPs), or other vehicles, often participating in more than one plan at a time. That is a startling increase over just a few decades ago, when employee ownership was relatively rare. Despite the negative press for employee ownership that has often dominated the headlines, the underlying story is both more encouraging and more prosaic. Overall, employee ownership contributes both to corporate performance and longevity, while employee owners tend to do better in terms of wealth accumulation and even wages than comparable nonemployee owners. The best employee ownership companies – those combining ownership with open-book management and a high-involvement organization of work – contribute most of this differential. For the more conventional employee ownership companies, sharing ownership is usually just another benefit plan, albeit one that usually makes a net contribution to employee financial well-being. One of the most dramatic areas of growth of employee ownership, in the U.S. and abroad, has been the granting of stock options to most or all employees. While especially prevalent in high-technology companies, it is well established in many other companies as well. In 1990, fewer than 1 million employees had options. By 2002, the number had grown to somewhere between 8 and 14 million, depending on whose estimate is used. But changes in accounting rules and greater shareholder concern over dilution could affect that growth and even lead to some retrenchment.

The Future of Broad-Based Options

359

2. NEW WAYS TO COUNT OLD NUMBERS In 2006, companies in the U.S. and abroad will have to show charges to earnings for all of their equity compensation plans. Under the current accounting procedure known as APB 25, companies do not have to show charges to earnings for stock options if the date on which the price and number of the shares offered are known is the same date as when the grant is actually made. These so-called ‘‘fixed options’’ offer employees the right to buy a specific number of shares at a price fixed today for a defined number of years into the future. They do not show up as a charge to earnings, although they do count as outstanding shares for earnings-per-share calculations. If the options are modified (for example, if the price is lowered after the grant is made or the term of the option is extended) or vest only if a performance goal is met, then they are considered ‘‘variable options’’ and must be charged to earnings. Companies can instead use the accounting procedure described in the Statement of Financial Accounting Standards 123 (SFAS 123), which requires them to estimate the expense of stock options using a model that calculates their present value (the theoretical value at which a willing buyer would buy them from a willing seller). As of this writing (June 2005), more than 800 companies have voluntarily chosen to do so. For equity pay other than options, companies must currently show a charge to income. All of this, however, will change when the International Accounting Standards Board (IASB) and the U.S. Financial Accounting Standards Board (FASB) require a new approach to options accounting. The accounting reforms would also require companies to show ESPPs as an expense. These plans allow employees to buy stock at a discount. The most common plan, the so-called Section 423 plan (after the part of the law where the rules are found) allows employees to put aside out of their paychecks up to $25,000 a year in after-tax money to buy stock at up to a 15% discount from (usually) the lower of the price when they make the purchase or the price when they started to put money aside. This period of time, called the offering period, usually lasts between three months and two years. By law, at least all full-time employees with two or more years of service must be able to participate. Currently, companies take no charge to earnings for these plans; under the proposed rules, they would have to show an expense both for the discount and for the expected value of the ‘‘look-back element’’ that allows employees to buy at the lower of the beginning or end of the offering period. Under current rules, all other forms of equity compensation must be expensed, although the proposed new standards will, in many cases, change

360

COREY ROSEN

how that expensing occurs. One result is that companies will face a more level playing field in choosing different forms of equity compensation. This has caused a great deal of hand wringing, controversy, and concern among companies providing stock options. Opponents of accounting reform say it will be the death blow for broad-based equity plans, that it is impossible to accurately value options anyway, that the formulas being proposed are so complex they will add more confusion than clarity, and that current earnings-per-share disclosure rules are adequate. Proponents contend that options are compensation and that investors are being misled by current procedures. Some reformers believe an accounting rule will force companies to reduce outsized options grants to executives. The accounting bodies say that their only concern is accounting clarity; if reforms change how companies use options, that is not within their area of responsibility. In addition to the accounting changes, companies are concerned about new NASDAQ and NYSE rules for shareholder approval of equity compensation plans. These new rules require approval for almost all equity plans except for qualified plans, such as ESOPs and 401(k) plans, as well as ‘‘material modifications’’ of existing plans. Unlike in the past (at least for NYSE companies), brokers holding stock in ‘‘street name’’ (holding shares for investors in their accounts, as is commonly the case) can no longer vote the shares without specific proxies from the investors. This will make shareholder approval somewhat harder to obtain. With many companies running out of approved shares to issue under existing plans, or wanting to institute new or modified plans, this approval process will become especially important. Investors have shown more wariness about the dilution resulting from equity plans and more focus on how these plans are structured. It is important to note at the outset that many people have misinterpreted just what the accounting charges would do. A company would not have a nickel more or less in cash or profits after the accounting changes. Moreover, how a company accounts for its earnings has nothing to do with how it records earnings for tax purposes; each procedure has its own rules. Companies’ tax obligations would thus not change by one penny. All that would change is how a company reports what it has; nothing would change about what it actually does have. Many news stories have, perhaps unintentionally, confused these issues. Despite the lack of real economic impact on cash flow, assets, or other tangible economic considerations, accounting norms do provide shareholders and potential investors with an idea of how profitable a company is.

The Future of Broad-Based Options

361

Changes in accounting procedures that drive reported (even if not actual) profits down raise the specter of shareholders abandoning options companies for companies that use other forms of compensation. Companies that voluntarily expense options currently would have a median reduction of about 13% in their reported earnings. Because early adopters of expensing may include many companies that have fewer outstanding options than some of their peers, this number will probably go up when expensing for all equity pay is required. In response, many companies are looking at how they can reduce their equity compensation plan expenses, whether by changing to another form of equity (as Microsoft did, replacing options with restricted stock), cutting back on who gets equity, reducing the amounts provided, or all of these. On the other hand, a majority (albeit a slim one) of companies plan to stick with what they have. This report addresses three questions: 1. Is the reaction to expensing rational? The assumption is that expensing will drive down earnings per share, which will discourage investors, which will drive down stock prices. But what if investors already incorporate options information into their decisions, as efficient market theories would predict? In that case, companies will be making decisions about changing compensation plans based not on what is best for the company, its employees, and shareholders, but on a chimerical expectation. The several major studies of this issue all strongly concur that expensing will not affect share prices. 2. What are companies planning to do? Several surveys provide information about how companies expect to respond to expensing. Our principal focus here is on how these changes will affect one of the most dramatic developments in equity compensation over the last 15 years: the rise of broad-based options plans and ESPPs. The surveys suggest about 40–45% of the companies will either eliminate awards to nonmanagement employees (for options) or change their plans in such a way that they will have little appeal to these workers (particularly ESPPs). But a majority of the companies plan to retain their broad-based plans. 3. What works best in equity plans, focusing awards on executives or making them broad-based? Decisions on this issue are being based largely on arguments and assumptions rather than on data about what really works. The research on this issue is fairly recent and quite compelling: broadbased ownership improves corporate performance; ownership concentrated at the top either has no effect or is harmful.

362

COREY ROSEN

3. WILL EXPENSING OPTIONS REALLY AFFECT STOCK PRICES? An underlying assumption of those concerned about being forced to expense options is that companies with significant options programs would see their stock prices decline in response. Of course, this is just an assumption. Others have argued that there would not be much effect at all, because this information already exists in the income statement’s footnotes. Markets, they say, already factor options costs into a company’s stock price. A more nuanced argument contends that companies with options programs that do not seem well targeted or well justified in terms of compensation strategies would suffer, but others would not, or that only companies with options programs well outside their industry norms would suffer. Based on the survey and the empirical research reported below, it appears that options expensing will have little impact on stock prices. Expensing could well turn out to be the Y2K problem of equity compensation – much feared, but with little impact. 3.1. The Impact of Expensing on Stock Prices: What Finance Professors Expect To assess these arguments, in January 2003 we at the National Center of Employee Ownership (NCEO) conducted an e-mail survey of 180 finance professors at 30 leading graduate schools of business. We received 37 responses. The professors were asked two questions. The questions and the introductory material appear below: As you know, there have been recent proposals to require companies to expense the present value of option grants to employees as a compensation cost on their income statements. Current rules require companies to show this expense in their footnotes or on their income statements (over 95% choose footnotes). This two-question survey is intended to find out what, if any, impact on stock prices economists expect this change to have. Which of these statements comes closest to representing your views about the short-term impact on stock prices of requiring companies to expense the present value of option grants to employees as a compensation cost on their income statements? (The number picking each answer follows the question.) It will have no significant impact on stock prices because the market already has incorporated this information. (14) It will have a small impact on stock prices. (7)

The Future of Broad-Based Options

363

It will have a substantial impact on stock prices. (2) It will have a small impact on stock prices of companies with options expenses larger than industry norms. (10) It will have a large impact on stock prices of companies with options expenses larger than industry norms. (4)

In other words, only six of the respondents expected that expensing would have a substantial impact on stock prices for any of the companies with options. The remaining 31 respondents expected either no impact or a small impact. We have also found a number of empirical studies so far, none of which suggests that there will be a significant impact. These are summarized below. In general, they show that the market already incorporates expensing data into stock prices. This does not mean that investors do not care about what options expenses are. In fact, the studies generally show that higher expenses are associated with lower stock prices. It does mean that a company that already has high expensing costs in its footnotes is not likely to see a hit to its stock price when it has to move these costs to its income statement once expensing is required. We have presented summaries of the studies we could find through a Web search and through references from experts below. Of course, there is no guarantee that past reactions to accounting changes will play out the same way in the future. All of the studies we could locate that attempted to draw a causal relationship are included; there was no effort to edit out studies because they came to one conclusion or another. We did exclude studies that are associational, that is they indicate that earnings per share would decline or increase by some percentage based on outstanding options, or that companies that have a higher earnings per share impact from expensing have lower or higher share prices over some period of time. These studies do not indicate (nor do they claim to indicate) whether expensing caused a change in stock price; to do that, some kind of before-and-after or other semi-experimental technique (a technique in social science research to try to replicate how a scientific experiment would be done) must be used. The summaries are organized according to how recently the studies were performed.

3.2. Options Expensing Announcement has no Impact on Share Prices Towers Perrin analysts Richard Ericson and Michael Grund (2004) looked at 335 companies that voluntarily chose to expense options between April 2,

364

COREY ROSEN

2001 and August 14, 2003. Share prices were tracked from the day of the company’s earnings declarations and the 150 days before and after. The results were then compared to the S&P 500 and MidCap 400 indices. Adjusted for general market movement, the average price of the announcing companies over the period showed no significant variation from the 900 companies in the companies index. The report concluded that options expensing does not affect share prices and that incentive compensation should be driven by actual costs and benefits, not accounting norms. Like other studies in this section, the research used an even study methodology, in which shareholder reactions to a specific event are analyzed for the companies being studies compared to market index movements during the same time period before and after. 3.3. Expensing Stock Options: What are the Markets Saying Ashish Garg and William T. Wilson (2003) studied whether expensing will affect share price. The study is of particular importance because it was able to hold constant the impact of options expensing on earnings per share in calculating market reactions to expensing announcements. It found that expensing does not cause a change in stock prices. The authors used an event study methodology. Event studies measure stock market reaction to events one or more days before and after the event. The movement is adjusted for movement in stock prices generally as well as by sector or other variables. The study looked at the 140 companies that announced they would expense options between July 8, 2002 and January 16, 2003. The four companies that were already expensing options as of this date were excluded. The event being studied was the issuance of the first press release announcing the change. The authors point out, however, that one problem with this methodology is that other occurrences such as announcements of earnings or dividends, analyst upgrades or downgrades, changes in leadership, or new strategic plans that happen within the same event time frame can confound the results. To deal with that, the authors excluded any company where any other announcements that could have a measurable impact on stock price also occurred. That left 54 companies to evaluate. To make sure that the study included the possibility that news of the decision to expense had not previously leaked to the press, the study looked at the three-day period surrounding the announcement. The results present a very clear picture. The average abnormal share reaction to expensing over the three-day window being studied was only

The Future of Broad-Based Options

365

+0.02% (in other words, stock prices rose 0.02% more than would be expected). The median was –0.52%. Neither result was statistically significant, however. Statistical significance means that researchers can have a high degree of confidence that the results are not random; statistical insignificance thus means that there is a very good chance that any differences found were simply random variation. The results were then broken down by industry, market capitalization, potential reduction in earnings per share attributable to the expensing, and the degree of corporate transparency. None of these subcategories produced statistically significant differences. Only the energy sector had a change of more than 1% ( 1.42%), but this could have been a random result. Other sectors saw positive results, but, again, none that were statistically distinguishable from zero. Market capitalization also yielded no differences. The major criticism of the event studies has been that companies that are voluntarily expensing options are generally those where the impact on earnings per share would be the lowest (the assumption is that these companies thought expensing would give them a favorable public image at little cost to their income statements). But if that argument held up, then companies with higher earnings impacts should show a more negative market reaction. The authors calculated what the earnings per share impacts would be going back to 2001. The average impact was 13.4% lower earnings. But no significant correlation was found between share reaction and the estimated impact on earnings. In fact, companies with a high exposure on earnings actually showed a 1.2% higher abnormal return, although this was not a statistically significant finding. Finally, the authors looked at data from Standard & Poors on how transparent companies were in disclosing financials. Using a 10-point ranking, no systematic differences could be found between more and less transparent companies. In fact, companies with low transparency actually had somewhat below normal returns, and the more transparent companies higher returns, although the findings were not statistically significant. The authors conclude that the data are consistent with the ‘‘semi-strong’’ version of the efficient-market theory. That theory states that the market automatically incorporates any available information on a company into its stock price (the semi-strong version allows for some randomness and other factors, but not on a consistent or significant basis). Expensing, they argue, will likely have little or no impact on stock price movement.

366

COREY ROSEN

3.4. The Watson Wyatt Study, 2003 In a 2003 study, Watson Wyatt looked at 800 companies drawn from the S&P 1,500 that had footnoted options expenses. Looking at expenses in 1999, 2000, and 2001, the study found that the higher the expense, the lower the total shareholder return. Similarly, looking at data from 1999 through 2001, the study found that the larger a company’s increase in options expense over time, the lower shareholder returns were, ranging from an average of –6.7% for the lowest group (where options expenses declined 16%) to –15.1% for the highest (which had a 227% increase). While the report seems to indicate the market already is sensitive to footnote disclosures, it is not possible from these data to know whether the market would be more or less sensitive if expenses showed up in the income statement. Moreover, causality here cannot be ascribed simply to options expensing. It could be that some third factor explains both a change in expense and a change in the market. For instance, companies with higher expenses may be more concentrated in industries that suffered the most in the market downturn, or companies that were relatively weak financially may have had the greatest pressure to increase options expenses in order to attract and retain employees. While the report focused on shareholder reactions to options expensing, the authors of the study also found that ‘‘over the last three years, for the options that are allocated to broad employee groups [this term was not defined by the study], an increase of one percentage point (i.e., from 78% of the options to 79% of the options) is associated with a nearly one percent increase in the firm’s market value. This is equal to nearly $15 million in market capitalization for the median firm in our sample.’’

3.5. The Heritage Foundation Study Norbert Michel and Paul Garwood (2002) conclude that the stock market would react with indifference to mandatory expensing of stock options. The conclusion is based on a study of the market’s reaction to six events from 1993 through 1995 surrounding FASB’s announcement that it would consider requiring stock options expensing and, eventually, its announcement that it would require expensing to appear only in footnotes. The authors used an event study methodology; the event in question was an indication by FASB of what it would do about options expensing. The authors hypothesized that if the market ‘‘cared’’ about expensing, prices for companies with

The Future of Broad-Based Options

367

the highest percentage of options overhang would decline relative to those with the lowest. So when announcements suggested expensing would be required, the prices of companies with a lot of overhang should have gone down, and they should have gone up when the news suggested expensing might not be required. In fact, the authors found no evidence that this happened. Instead, the market appeared indifferent to the announcements.

3.6. The Haidan Li Study The largest and most academic of the articles to date on the impact of expensing on share prices is by Haidan Li (2002). She looked at the threeday periods surrounding both the announcements of 10-K filings (which provide details on corporate earnings) and the three-day period surrounding earnings announcements to the public (these are typically released some time before 10-K filings). The study included data from 1,500 S&P companies. Li found that the market does react negatively (albeit in a relatively minor way) to ‘‘unexpected’’ stock options expenses in the period around 10-K filings, but there is ‘‘no significant association between unexpected stock option expense and unexpected stock returns around earnings announcements.’’ What this means, she says, is that the market already incorporates data from footnote disclosures, indicating that formal expensing would not add much new information. Li created a model for residual income, defined as the sum of the book value of equity and the present value of earnings less a charge for capital. Stock options must be factored into this model, Li argues, by considering existing options as a contribution to capital (because, she assumes, employees are foregoing cash compensation for options), with employees having a claim on future earnings, a situation that must be factored into the residual income calculations. If firms chose not to expense options, she says, ‘‘then the residual income model must be modified through a subtraction of the discounted value of future options.’’ Of course, no accounting proposal would require companies to expense options they are yet to grant, but Li notes that, theoretically, these future options do have an impact on expected earnings. Li’s next step is to create a model that incorporates a Black-Scholes estimate of the impact of options expensing, the tax effect of options, an estimate she makes of future options expenses, analysts’ expectations of future earnings, share price, and book value (this greatly oversimplifies an extremely complex and mathematical presentation). Using a sample of 1,500

368

COREY ROSEN

companies from the S&P 500, the S&P 400 MidCap, and the S&P 600 Small Cap companies, Li collected data on employee options from 1997 through 2000. The final sample included 1,113 companies that had all the necessary data. The model indicates a significant negative correlation between future employee stock options and predicted future residual earnings. But do shareholders already incorporate this information into stock prices? To answer that, Li conducted a three-day (before-and-after) event study in which she analyzed whether changes in options expenses correlated with unexpected shareholder returns around the filing of companies’ 10-K reports. She did the same analysis for earnings announcements. Changes in options expenses were significantly related to stock prices after the 10-K filings, but the magnitude of the explained difference is small, only 1.2% (a study can find a small difference as a result of a variable that is still statistically significant – not likely to occur at random – if the sample size is large). By contrast, predicted options expenses were not related to stock prices in the pre-10-K filing announcements. In other words, Li concludes, the information in the 10-K filings (the current footnote disclosures) is sufficient for shareholders to determine company valuations.1

4. WHAT WILL COMPANIES DO ABOUT BROAD-BASED EQUITY AWARDS IN RESPONSE TO EXPENSING? Employee ownership has grown dramatically in the last three decades in the United States. The 2002 General Social Survey from the National Opinion Research Center indicates that about 25 million employees currently own stock in their employer, a number that amounts to about 40% of the employees who work for companies that have stock (a summary of these data is provided in the appendix). Of these, we estimate that about 10 million own stock through stock options, while 12–15 million buy stock through ESPPs. Because these groups overlap extensively, the total may be in the range of 15–18 million people owning stock through these plans. (Many of these employees may also own stock in their employer through other plans as well.) As the following data suggest, however, these numbers could be reduced, perhaps significantly, as companies react (and, we will argue, overreact) to accounting reforms. A number of recent surveys provide information on how companies say they will react to requirements that options and other equity awards be

The Future of Broad-Based Options

369

shown as an expense on their income statements. While it is well-known that the new rules will affect conventional stock option plans, less known is that it will also affect ESPPs. FASB and IASB both seem to be moving toward requiring companies to expense any discounts offered to employees in these plans. The surveys consistently show that just fewer than half of the companies responding plan to cut back the eligibility and/or the size of grants for employees below management levels. On the other hand, very few companies have any intention of reducing eligibility or the size of awards at the senior executive level, and relatively few at the management level. In addition, a third to 45% of the companies plan either to eliminate ESPPs or change their structure so as to provide lower discounts, eliminate their ‘‘look-back’’ features, or shorten their offering periods. The ironic result is that expensing could further concentrate ownership in the hands of a small number of highly paid people, arguably precisely the opposite consequence that at least some options reformers had in mind. If companies do what they are telling these researchers they intend to do, it will make a mockery of corporate reform, as senior executives simply find new ways to use reform to argue for their advantage. These surveys must be viewed with some caution. First, many companies make employees eligible for an equity compensation plan but do not actually provide equity to many of the eligible employees because they do not meet some performance or other criteria. Only actual experience will indicate how many employees will get how much less equity in the post-expensing era. It could be, for instance, that some companies say they will maintain current eligibility rules, but then raise the standards for actually getting an award. Second, there is arguably more concern about expensing than is merited, as argued above. If that turns out to be the case, companies may change their minds about restructuring their plans if they find that the market does not react strongly to expensing. After all, it is never easy to take benefits away, especially when company leaders are not sharing in the sacrifice. Finally, and probably most importantly, the job market could recover, creating the same kinds of labor market conditions that prevailed a few years ago that had companies competing for labor with ownership stakes. It could be very difficult for a company to eliminate its options or its ESPP (or cut back substantially on its benefits) when a competitor continues to offer these plans. The surveys indicate that more than half of the companies plan to keep their programs intact, a number we expect to end up somewhat higher as the specter of expensing becomes less threatening. That will make staying in the relatively Scrooge-like minority tough to do. In any event, the Bureau of Labor Statistics has predicted that in the next decade there will be

370

COREY ROSEN

a shortage of 10 million skilled workers. Yet with continuing global competition, it will be hard for companies to compete for labor based solely on cash compensation without raising their costs above what they can afford to pay. It is important to recognize that companies choosing to cut back on broad ownership plans will achieve relatively little in the way of reducing outstanding options. Nonmanagement ownership accounts for only about 30% of all the options in companies with broad-based plans. ESPPs would account for a small fraction of that. Simply reducing the outsized grants to senior management in these companies would make it possible to retain broad-based plans at meaningful levels of awards. (These data come from the NCEO, 2000.) Below, we summarize the results from major surveys.

4.1. The State of Stock Options 2003 In the study of Sibson Consulting and WorldatWork (2004), data from 363 public companies responding to an e-mail survey were used to determine what changes companies made in their stock option plans in 2003 compared to 2002. Unlike the studies that follow, this study looked not at intentions for the future, but actual current changes. Table 1 indicates that companies made few changes in eligibility for options, except at the non-exempt employee level, where eligibility declined from 37% in 2002 to 27% in 2003. Note that eligibility is not the same as actually receiving a grant. For non-exempt employees, other data indicate that eligibility may overstate the percentage of who will get a grant by a factor of about two, but this will vary widely from company to company.

Table 1.

Eligibility by Employee Level. % Eligible

Top executives Non-executive managers Professionals Sales staff Non-exempt employees Source: Sibson Consulting and WorldatWork (2003).

2002

2003

98 86–97 63–65 67 37

99 88 64 50 27

The Future of Broad-Based Options

371

The value of the options granted (as opposed to options already held) in 2003 was more likely to decline than increase for non-exempt employees. In 29% of the cases, the decrease was greater than 20%, while in 11% it was 5–19%. In 27% of the cases, it stayed the same, while it went up 5–19% in 15% of the cases and increased by more than 20% in 5% of the cases (for unexplained reasons, the numbers do not add to 100%). A similar pattern was reported for exempt employees below the executive level, except that the grant size remained the same about 10% more of the time. Executives, predictably, did better, with 47% staying the same, and with fewer showing significant decreases. Just over half the companies reported that concerns about the impact of dilution and the anticipated implementation of new accounting rules were major factors in driving their changes. Interestingly, while 42% of the companies said the value of options in motivating employees was a major factor in options changes in 2002, only 31% did in 2003. This may reflect other Sibson data showing that, contrary to popular expectation, employee attitudes toward stock options actually showed very little change from 2000 to 2003. Companies may have learned that options have a fairly steady effect on employees even against the backdrop of fluctuating markets.

4.2. 2003 Technology Stock Compensation Survey: Looking beyond Options Deloitte and Touche (2003) surveyed public and private companies about their plans for equity compensation programs post-expensing. They received 196 responses (the number surveyed was not indicated), 175 of which were from technology companies. The results generally report only on the technology companies. Among publicly traded companies, ‘‘69% said they offer options to 90% or more of their employees. Nearly all the private companies – 96% – said this was their practice.’’ Note the wording here indicates companies actually offer options widely, not just make employees eligible for them. In the small sample of non-technology companies, only 21% offered options to more than 90% of their employees, a number that is very consistent with prior larger surveys by various organizations on the prevalence of options. Most of the respondents expressed concerns that their options programs were threatened both by expensing (their top concern) and shareholder reluctance to grant additional shares for option programs. Respondents were planning a variety of responses to these challenges, but it is noteworthy that 45% of the public companies and 70% of the private

372

COREY ROSEN

Table 2.

How Companies will Change Their Stock Option Plans.

Type of Change

Limit eligibility to mgrs. and execs. Stop granting options Limit grants to select employees below management Reduce size of grants across the board Reduce size of grants below exec. level Reduce size of grants below exec./mgr. level Reduction only at mgmt/exec. levels Base eligibility on increased performance standards Reduce new hire grants

Public Companies (%)

Private Companies (%)

11.6 4.4 25.3 33.3 9.4 9.9 3.3 17.1 14.3

7.5 2.4 18.0 11.4 3.6 5.1 1.5 17.4 11.4

Source: Deloitte & Touche (2003).

companies have no intention of changing their programs. Table 2 indicates the principal kinds of changes that are contemplated (the numbers are for the entire sample; Deloitte reported this as just the percentages of companies planning changes). Percentages may add to more than 100 as companies could choose multiple responses. Table 2 indicates that 41% of the public companies will either stop granting options altogether or restrict options to a limited number of nonmanagement employees. By contrast, only 28% of private companies plan that approach. Reducing grant size is also popular, with 46% of the public companies planning that change, but only 22% of private companies planning to do so. Almost none of the companies, however, will reduce executive grant size. Despite all the discussion about tying grants to performance, only 17% of the companies in both samples planned this approach. Similar levels of change may be in store for ESPPs. Among respondents, 63% said they would make some changes (Deloitte & Touche, 2003) (only public companies were surveyed). The most troubling finding for ESPP participants is that the look-back feature, the most attractive aspect of an ESPP, would be eliminated in 49% of the responding companies. Its elimination would most likely substantially reduce participation, leading some companies to drop the plans altogether. 4.3. Responding to Mandatory Option Expensing: Human Resources and Investor Solutions Mellon Financial (2003) ‘‘flash survey’’ was sent to 1,600 companies, 167 of which responded. Respondents had to be in a public company and currently

The Future of Broad-Based Options

373

provide stock options or plan an IPO in the next year. Forty-eight percent of the respondents were in high-tech, 33% in general industry, and 19% in life sciences. Similarly to the Deloitte and Touche (2003) survey, 67% of the high-technology companies and 77% of life sciences companies provide options to more than two-thirds of non-exempt employees, compared with 24% of general industry respondents. Seventy-two percent of all companies offer ESPPs. Companies also may have other employee ownership plans, with 15% of general industry companies having ESOPs and 69% matching 401(k) deferrals with company stock. Only 12% of the respondents have no additional ownership opportunities beyond options. One-third of all companies responding would eliminate options for nonexempt employees if expensing were required, compared with 31% of hightechnology companies, 46% of general industry companies, and 29% of life sciences companies. While these numbers seem somewhat less significant than the Deloitte responses, remaining eligible and actually getting an award are not the same thing. Forty-seven percent of the respondents indicated that the percentage of nonexempt employees who would actually get options would be substantially lower; the rest said it would be only slightly lower or the same. Taken together, this indicates that almost half the respondents will be eliminating or substantially reducing options for non-exempt employees. Where options program will be significantly decreased, 54% of the companies said that nothing would be done to make up for the change; one-third said there would be other long-term incentives. Not surprisingly, top executives suffer no such pain. Only 1% of the companies said they would stop granting options at this level, and only 3% said the number of options granted would change significantly. Other executives do only slightly worse. Seven percent of the companies with ESPPs said they would eliminate their programs in response to expensing, 48% said they would cut back the discount to 5%, and 45% said they would keep the 15% discount. The survey did not ask about retaining the look-back period or changing the offering period. The responses indicating that a 5% discount would be used reflect an assumption by companies that they will not have to expense ESPPs if they have this limited discount. Under SFAS 123 (the current optional approach in the United States for fully expensing options) that would be allowed. This is probably not an accurate assumption, however, as indications are that any discount will show up as an expense. In 2004, the flash survey was again conducted, using the same basic approach. This time, there were 108 responses. Of these, 44% were technology companies, 37% in general industry, and 19% were life sciences companies.

374

COREY ROSEN

Forty-nine percent of the respondents reported that they made options or similar awards available to non-exempt employees on an annual basis, 52% made at least half their exempt non-managerial employees eligible on an annual basis, and almost all the companies made all their top executives eligible. Almost all (87%) the life sciences companies made non-exempt employees eligible, compared to 33% for general industry, and 31% for high-technology companies. Sixty-three percent of the companies have ESPPs, 9% have ESOPs (but 20% of general industry companies), and 51% 401(k) plans with company stock. Only 13% have no other equity vehicle in addition to their options or similar plan. All but a few of the ESPPs have a look-back period except in the general industry sector, where 64% do. Ninety-four percent of the companies provide a 15% discount. In response to proposed accounting rule changes, 45% of all companies as well as technology companies, 67% of general industry companies, and 38% of life sciences would eliminate option eligibility for non-exempt employees. While non-managerial exempt employees would mostly still remain eligible (although 33% of general industry companies would eliminate option grants for these employees), 40% of the companies say the amounts these employees would get would be much lower. Not surprisingly, executives will suffer no such slings and arrows, and even where options grants go down, companies report something else will take their place for people at executive levels, but not much at other levels (61% of nonexempt and 58% of exempt non-managerial employees would not have anything added to make up for the loss of options). Companies generally say they will keep their ESPPs, but two-thirds of the companies say they will either cut back the discount, the look-back period, or both. About half the companies in all sectors would eliminate the lookback period, while between 10% and 33% (depending on the industry) would cut it back. Half also say they would cut back on the discount.

4.4. Future of Equity: 2003 Update The Mercer Human Resource Consulting (2003) study had 205 respondents, 92% of which were public. The respondents represented a broad mix of industries, but most were very large, with 53% having more than 10,000 employees. The survey touched only briefly on the issue of principal concern here, namely how changes in expensing rules will affect the distribution of stock options and other equity awards among the work force.

The Future of Broad-Based Options

375

Sixty-five percent of the companies said they had already made some changes to their equity compensation programs in 2003 in response to the expensing debate and other issues. Of the companies that did make changes:

63% reduced the number of options granted or intended to be granted. 46% reduced the number of individuals who receive options. 57% introduced new equity vehicles. 8% reduced options and substituted at least some cash.

Note that this means that 30% of the companies said they had already reduced the number of employees who will receive or have received options. The data did not indicate just what the nature of these reductions were, however. Of the 76 companies that said they introduced new equity programs, 63% said they provided restricted stock with service-based vesting and 14% restricted stock with performance vesting. Other approaches had only spotty use. The new equity vehicles affected employees broadly in only 10% of the cases; everywhere else they focused primarily on executives and upper management. 4.5. Stock Options Today: Down but Not Out A 2003 survey of 165 of Deloitte & Touche Consulting’s (2004) largest clients indicates that 46% of the companies plan to reduce options in the coming year and 29% have already done so. Concerns about dilution were a primary concern for 33% of the companies planning on reductions and a secondary concern for 30%. Accounting changes were also a major issue, with 32% of the companies planning changes citing this as a primary concern and 19% citing them as a secondary concern. In fact, 25% of the companies planning changes say that they will run out of shares approved by shareholders for equity grants in less than one year, while 14% say they will run out in one year, 28% in two years, and 33% in three years or less. Forty-five percent of the companies said that options did not have the value they had thought as employee incentives. In a finding that is consistent with other surveys, 36% of the companies say they will limit eligibility to management and/or executives (it was not clear, however, how many of these companies currently extend options much beyond management). Another 25% are making across-the-board reductions, while 13% said they will simply eliminate options. Of companies that had extended options below the

376

COREY ROSEN

management level, 51% who granted options to exempt employees will no longer do so, and 54% who granted them to nonexempt employees will stop.

4.6. Implications Of course, it is impossible to predict how the market will respond with any precision, but at this point, it seems safer to bet that the impact of expensing will be nominal. If this is the case, companies should certainly not make wholesale changes in their plans just to respond to a nonexistent problem. Instead, they should design their equity compensation so that it serves their real economic and human resource needs, not their accounting convention needs. So what would that look like? Companies that want options plans to work should de-emphasize the outsized grants being made to executives and emphasize instead broad ownership programs. Ironically, some corporate leaders are saying (or perhaps threatening) that if expensing is required, they will do just the opposite – keep grants for top people and cut out broad grants. These data suggest that would be a serious mistake for companies and their shareholders. Opponents of accounting reform will blame these developments on FASB, IASB, and those in Congress and elsewhere who have pushed for expensing options. But this argument is not entirely convincing, especially when companies complain that reforms will make it impossible for them to compete because they cannot offer equity across the board. If broad-based equity is so important, then accounting changes should not force companies to make drastic alterations in how they pay people. Accounting reforms do not compel companies to focus their equity awards on limited numbers of people. They may provide an excuse for scaling back to companies that never were committed to broad ownership in the first place, but went along because of competitive pressures. And they may provide a rationalization that executives who want to focus ownership on themselves can use to justify their excesses. But the research in this report strongly argues that these rationalizations are flimsy and self-serving. As the next section shows, ownership spread broadly works much better for everyone involved. That does not mean it will work for every company in every case; there will still be scandals, disappointments, and plans that provide only token benefits. But there will be spectacular successes as well. In the end, an economy that treats its workers more the way Southwest Airlines, Cisco, Starbucks, and many others do is a far better economy than one where ownership is reserved for the already privileged.

The Future of Broad-Based Options

377

5. EMPLOYEE OWNERSHIP AND CORPORATE PERFORMANCE: DOES IT MATTER WHO GETS EQUITY? As companies face the prospect of expensing options and other equity plans as well as tougher shareholder approval rules, clearly many, if not most, will want to think through who gets how much ownership more carefully. Ultimately, corporate boards and compensation committees need to ask which strategies work best for long-term corporate performance and shareholder returns. Fortunately, there is now considerable research on how the distribution of equity affects both these measures. Studies of broad-based equity plans and corporate performance indicate a consistently positive relationship; studies of equity compensation geared primarily for executives come to mixed and uncertain conclusions. The most comprehensive summary of the issue of broad-based ownership and corporate performance comes from Joseph Blasi, Douglas Kruse, and Aaron Bernstein (2003). The authors analyzed research on what they call ‘‘partnership capitalism,’’ namely sharing equity with most or all employees in one fashion or another (ESOPs, stock options, ESPPs, and other approaches) as well as employee involvement and profit-sharing programs. Most of this research is on ESOPs, because this is the best-established method for sharing ownership widely. There have been recent studies on options as well, however. It is worth quoting the authors’ conclusions at length: In the past twenty-five years, researchers have done more than seventy empirical studies of these forms of risk sharing [ed.’s note: by ‘‘these forms,’’ they mean partnership capitalism]. Taken together, the studies provide compelling evidence for the net gain that the partnership approach can produce for a company’s shareholdersy. The results surprised even us, not because they were so positive, but because they were so extensive and so uniformy. The most striking conclusion: Every major study found that investors came out ahead if their company adopted key elements of partnership capitalismy. We added up all the conclusions and averaged them into a single finding for each of the four elements. Roughly speaking, we found that the partnership approach improves a company’s productivity level by about 4 percentage points, compared to firms that don’t adopt such practices. Total shareholder returns increase by some two percentage points relative to other firms. Profit levels – as measured by return on assets, return on equity, and profit margins – jump by about 14%y The gains in profits and returns came after the dilution borne by outside shareholders has been factored in.

Below, we break out the most important studies on this topic, looking primarily at those that show a before-and-after effect, rather than just a

378

COREY ROSEN

correlation. Correlational studies (studies that show that companies with certain ownership characteristics are better performers than those without them) are limited by the fact that it is not possible to know whether greater ownership is a cause or effect of better performance. As numerous as studies of broad-based ownership plans are, far more numerous are studies looking at executive pay, perhaps because funding is more readily available. One of the most thorough and recent analyses of these studies provides a useful summation. John Core, Wayne Guay, and David Larcker (2003) state: There is presently no theoretical or empirical consensus on how stock options and managerial equity ownership affect firm performancey. A limitation of this research is that the causal direction of the relation between equity incentives and performance in uncleary. Rather than higher equity incentives producing better future firm performance, it may be the case that firms expecting better future performance grant more equity.

In the material that follows, we summarize key research on these topics. Because our expertise is focused on broad-based plans, we present more detail on this research, relying primarily on summaries of executive ownership research by leading scholars in the field.

5.1. Broad-Based Ownership and Corporate Performance Most of the research in this field applies to ESOP companies. As of yet, only a few major studies have been done on the impact of broad-based stock option plans on corporate performance. In ESOPs, the findings also apply primarily to closely held companies, although there are some data on public companies. By contrast, the options studies look primarily at public companies. There are currently no studies on the effects of ownership through 401(k) plans or ESPPs on corporate performance.

5.2. ESOPs and Private Companies In the largest and most significant study to date of ESOPs in closely held companies, Kruse and Blasi (2000) found that ESOPs increase sales, employment, and sales/employee by about 2.3–2.4% per year over what would have been expected absent an ESOP. ESOP companies are also somewhat more likely than their competitors to still be in business several years later. This is despite (or perhaps because of) the fact that ESOP companies are

The Future of Broad-Based Options

379

substantially more likely than comparable companies to offer other retirement benefit plans along with their ESOPs. Kruse and Blasi obtained files from Dun and Bradstreet on companies that had adopted ESOPs between 1988 and 1994. They then matched these companies to non-ESOP companies that were comparable in size, industry, and region. They then looked for which of these companies had sales and employment data available for a period of three years before the plan’s start and three years after. The sales and employment growth data were then compared for each year for each paired company. They also checked the companies’ filings with the Department of Labor to determine which of the companies had other retirement-oriented benefit plans. Finally, they looked to see what percentage of the companies remained in business in the 1995 through 1997 period. The process yielded 343 ESOP companies and 343 pairs for the overall sample. However, missing data meant that employment data were available for only 254 ESOP companies and 234 pairs, sales data for 138 ESOP companies and 77 pairs, and sales/employee data for 115 ESOP companies and 65 pairs (some pair companies could be used for more than one ESOP company). The results showed that ESOP companies perform better in the postESOP period than their pre-ESOP performance would have predicted. The table below shows the difference in the pre- to post-ESOP period for ESOP companies’ sales growth, employment growth, and growth in sales per employee: Difference in post- to pre-ESOP performance, ESOP versus comparable non-ESOP companies Annual sales growth +2.4% Annual employment growth +2.3% Annual growth in sales per employee +2.3% It might be assumed that sales per employee would not go up by 2.3% per year since sales and employment growth differences were about the same, but, the researchers explain, the differing compositions of the samples for the measures makes such a simple comparison misleading. The relative growth numbers might seem small at first glance, but projected out over 10 years, an ESOP company with these differentials would be a third larger than its paired non-ESOP match. The first study to show a specific causal linkage between employee ownership and corporate performance was by Michael Quarrey and Corey

380

COREY ROSEN

Rosen (1987). The study looked at the performance of employee ownership companies for five years before and after they set up their ESOPs. It indexed out market effects by looking at how well employee ownership companies did relative to competitors in the pre- and post-ESOP periods, then subtracted the difference. For example, if a company were growing 3% per year faster than its competitors in the pre-ESOP period, and 6% per year faster in the post-ESOP period, there would be a +3.0% difference attributable to the ESOP, other things being equal. The study found that ESOP companies had sales growth rates 3.4% per year higher and employment growth rates 3.8% per year higher in the postESOP period than would have been expected based on pre-ESOP performance. When the companies were divided into three groups based on how participatively managed they were, however, only the most participative companies showed gains. These companies grew 8–11% per year faster than they would have been expected to, while the middle group did about the same, and the bottom group showed a decline in performance. Participation alone, however, is not enough to improve performance. A large number of studies show that the impact of participation absent ownership is short-lived or ambiguous. Ownership seems to provide the cultural glue to keep participation going. These two studies remain the most significant analyses to date of the impact of ESOPs and corporate performance in closely held companies, but other studies, particularly by Gorm Winther (1995) of firms in New York and Washington, reached very similar conclusions. It is important to note that differences in compensation do not explain the improved performance of ESOP companies; again, research by Kruse and Blasi (2000) finds that ESOP companies are much more likely than comparable companies to offer other retirement plans, and a comprehensive study in Washington state found that ESOP companies pay 5–12% higher wages (Kardas, Keogh, & Schiff, 1998).

5.3. Research on Public Companies and ESOPs The data for public companies are much more ambiguous. A study by Hamid Mehran (1999) found that ESOPs in 382 publicly traded companies increased their returns on assets (ROA) 2.7% over what would otherwise have been expected. The study looked at each company’s financial returns for two years prior to the plan’s implementation and four years after. Each company was compared to industry norm ROA figures for both periods. Mehran also found that for the 303 ESOP companies surviving the entire

The Future of Broad-Based Options

381

four-year post-ESOP study period, ROA was 14% higher than the comparison group scores, while for the 382 companies as a group, ROA was 6.9% higher for the four-year period. More than 60% of the companies experienced increases in their stock prices in the two-day period following public announcement of the ESOP, with the average increase for all companies at 1.6%. This suggests that the stock market now reacts positively to ESOPs, a change from the pattern in the 1980s when ESOP announcements were often seen as an indicator that a company was trying to prevent a hostile takeover. In 1992, Kruse, Blasi, and Michael Conte created an ‘‘Employee Ownership Index’’ (EOI) that tracked the average percentage increase in the stock prices of all publicly traded companies with public records of 10% or more employee ownership and more than $50 million in market value. The EOI was subsequently maintained for some time by American Capital Strategies, an investment bank based in Bethesda, Maryland, and its returns were published quarterly in the NCEO’s Journal of Employee Ownership Law and Finance. The EOI grew 193% from 1992 through 1997, while the Dow was up 145% and the S&P 500, 140%. The authors did not attribute any causal relationship between the ESOPs and these numbers, however. Other studies look at before-and-after results, with mixed conclusions. Donald Collat (1995) found that public companies that did not set up their ESOPs in response to takeover threats saw their operating margins improve 2.1% per year compared with their pre-ESOP performance. The study looked at companies for three years before and after the ESOP, indexing for market effects. Takeover-threat ESOPs, however, saw a decline of 3.3%. Mary Ducy, Zahid Iqbal, and Aige Akhigbe (1997) found that ESOP companies showed a decline in operating cash flow of 0.2–2.1% in post-ESOP performance, also using a three years before, three years after measure, and again indexing for market effects. While these are the most thorough of several studies on public company ESOPs, others come to similarly mixed conclusions. Finally, Blair, Kruse, and Blasi (2002) found that companies that are publicly traded and at least 20% or more owned by an ESOP are more organizationally stable than comparable non-ESOP companies. Looking at companies between 1983 and 1996, the study found that 74.1% of the ESOP companies remained as independent operations while only 37.8% of the comparison companies did (these figures changed to 59.3% and 51.1% for the period 1983 through 1997, however). None of the ESOP companies went bankrupt, but 25% of the comparison companies did. These mixed results are probably explained by three factors. First, a 1997 NCEO study found that public companies generally seem to view employee

382

COREY ROSEN

ownership solely as a benefit plan, not part of an explicit organizational culture, as many closely held companies do. Second, ESOPs in public companies tend to own much smaller percentages of company stock than ESOPs in closely held companies. Some studies have indicated this as a factor in how effective ESOPs are. Finally, in many cases, public company ESOPs simply replaced existing 401(k) plans to which the company contributed its stock. Now the company used an ESOP to make this contribution instead. Hence, the ‘‘before’’ was really not much different from the ‘‘after,’’ so not much could be expected to change.

5.4. Broadly Granted Stock Options and Stock Prices Eric Hager (2003) analyzed how shareholders reacted to announcements of broad-based options grants. Announcements were excluded if they did not indicate employees would receive stock options, but were included if employees only or employees in addition to managers, executives, directors, and/or consultants received options. Hager looked at companies in Canada and the U.S. For companies to be included in the study, trading data for the period from 120 days prior to the announcement to one day after it had to be sufficiently available. Announcements had to involve actual grants, rather than just stock option plans. Hager points out that the announcement of a plan can be misleading because a company may make a certain number of employees eligible for a plan but actually only make grants to a much smaller number. So Hager looked at actual grant practices. A total of 91 grant announcements were suitable for inclusion in the event study for Canada and 54 in the U.S. The U.S. grants were from 1993–2002 and Canada from 1995 through 2002. Hager used a standard event-study methodology to extract abnormal shareholder returns, that is, returns greater or less than what would have been predicted in the day following the announcement based on how a model accounting for other companies in the industry performed. For Canadian companies, Hager found that returns were up 2.13% over what would have been expected for broad grants, and 2.33% greater when more than 1% of the equity was granted. The results were not affected by looking at grants only after September 1, 2000, when the markets began to fall sharply. For U.S. companies, Hager found that returns were up 1.78% over what would have been expected for broad grants when more than 1% of the equity was granted, but there was no significant relationship when less was granted. Again, the results were not significantly affected by looking at grants

The Future of Broad-Based Options

383

only after September 1, 2000. The findings for the broad-based grants were statistically significant (that is, not likely to be a result of random variation).

5.5. broadly granted stock options and corporate performance To date, there have only been only two large-scale assessments of the impact of broad-based stock option plans on corporate performance. The most important, and largest, was a study by Kruse, Blasi, Kroumova, and Sesil (2001), using data provided by the NCEO. The study sample was drawn from the 1998 NCEO Current Practices in Stock Option Plan Design study. That study sent surveys to 1,360 companies that were identified as possibly having broad-based option plans, which we defined as plans in which more than 50% of full-time employees would actually receive options. We received 141 responses. For the Rutgers study’s purposes, 105 companies provided usable data. The authors used a before-and-after approach to the data to reduce or eliminate sampling bias issues. Results were based primarily on the 91% of the sample companies that were publicly traded. Data were gathered on productivity, ROAs, Tobin’s Q (a complex financial measure of ROAs that produced similar results to the ROAs measure and is not reported here), and total shareholder return. These were then compared to all non-broad-based stock option companies in their industries of similar size (the full sample group) and to paired comparisons of matched non-broad-based stock-option companies (the paired sample). Because few companies had discrete plan start dates early enough to perform a comprehensive before-and-after analysis, the researchers, as a substitute, analyzed companies in the period 1985–1987 and 1995–1997, reasoning that few, if any, of the companies had option plans in the earlier period and most had them in the later period. Comparisons were made with non-stock-option companies for the two periods and the difference subtracted. In effect, the earlier period results provided a baseline to measure the performance in the later period. If a stock-option company had productivity 3% greater than its peers in the earlier period and 6% greater in the later period, then it could be argued that the plan improved relative performance on this measure by 3%. The study found that productivity rates did improve with the institution of a plan. The difference between productivity scores for the overall sample from the pre-plan period (1985–1987) to the post-plan period (1995–1997) was 14.8% when the comparison group was all non-option companies and

384

COREY ROSEN

16.8% when looking just at paired comparisons. Sampling error can be strongly rejected as an explanation for these results. Return on assets showed a similar pattern. Here the stock-option companies showed an improvement of 2.5% on ROA relative to the full sample in the post-plan period compared with the pre-plan period. When just paired comparisons are used, the improvement was 2.05%. Again, sampling error is very unlikely to have caused these results. Total shareholder return, however, showed no statistically significant difference in the relative performance during the two periods, meaning any measured change could simply reflect random sampling error. The researchers thus believe that the any value consequences of dilution caused by broad-based options seems counterbalanced by increased productivity. Looking simply at how the companies did in the period 1992–1997, without trying to adjust for market effects, a similar pattern emerged. Productivity growth was 1% per year greater and ROAs 5.8% greater, but shareholder return was not statistically distinguishable. Christopher Ittner, Richard Lambert, and David Larcker (2003) looked at options and corporate performance, using data from 159 ‘‘new economy’’ companies providing detailed responses on their stock plans for an iQuantic/Buck (now simply Buck) Consulting survey. They found that the performance effects of option programs depended on how the options were distributed. The study found that broad-based equity grants were the norm in this sector. Companies with greater growth opportunities, larger market capitalization, better past stock market performance, and lower leverage all used options and other equity grants aggressively at all employment levels. The researchers then looked at whether stock returns subsequent to option grants improved or declined. To measure this, they developed a statistical model in which ‘‘return’’ equaled the ‘‘continuously compounded stock price return in the 12-month period following grant.’’ They also created estimation models for burn rate (here defined as the rate at which new equity is issued to employees) and overhang (the percentage of total shares outstanding divided into the currently issued and unexercised equity awards plus shares that have been authorized for issuance for equity awards). The results were controlled by a variety of factors, including several industry indicators to factor out returns unrelated to the grants themselves. They also controlled for company size and book-to-market ratios (a measure of balance sheet versus stock market values), R&D expenditures, and advertising sales. The data looked at returns in 2000 relative to 1999. The results showed that deviations from the norm for overhang and burn rate were unrelated to stock price performance, suggesting that variations in

The Future of Broad-Based Options

385

how much equity is granted do not affect abnormal stock price movements (movements away from what would otherwise have been expected). On the other hand, who gets equity does make a significant difference. Larger than usual grants to executives (CEOs, vice presidents, and directors) did not significantly affect stock prices. Grants to managers, ‘‘individual contributors’’ (critical non-management employees), technical employees, and exempt non-technical employees, by contrast, resulted in significantly greater than expected stock price growth. The researchers explained that the model suggests that, for instance, ‘‘for a 20% increase in the ratio of equity to salary for similar new economy companies, there was a 5.1% increase in annual returns if the grant was to technical employees; for non-technical employees, the return was 2.7%.’’ Looking only at non-exempt employees (hourly employees not exempt from the Wages and Hours Act), the study found a small negative relationship between stock price and grants of more than the benchmark amounts to these lower-level employees.2 These results are somewhat suspect, however, in that this group of companies had very, very few such employees (the iQuantic/Buck Web site indicated that only about 2% of the work force fell into this category), so these results cannot easily be compared to companies where non-exempt employees are a large percentage of the total work force. A Starbucks or Bank of America, for instance, presents a very different picture in terms of the importance of equity grants to nonmanagement employees than does a small software company. It should be kept in mind that this study looked only at changes in stock prices over the 12 months after equity grants. Stock price movements over any relatively short period, even when controlled for industry factors, can include considerable randomness and, in any event, are often not closely related to actual company performance. Concluding that any group of employees received equity awards in year one, causing them to behave differently enough so that company performance improved enough in year two to change market responses as measured by stock prices is obviously somewhat dangerous, as the authors, of course, understand. Nonetheless, this carefully done study does lend further support to the notion that the typical assumptions that equity awards to executives are what ‘‘really matter’’ does not appear to be correct, while broader grants to those who actually do the day-to-day work do matter. Sesil and Kroumova (2004) use NCEO lists of companies with broadbased options in 1998 and 2000 to evaluate the impact of these plans on productivity. The study was just completed as this brief was issued and is now in submission for publication. The study looked at two issues:

386

COREY ROSEN

(1) would stock options be less effective in times of declining share prices than rising prices? and (2) does the effectiveness of broad-based stock options depend on the size of the company (specifically, do employees react more positively in smaller companies where their individual efforts have a more visible effect)? To analyze this issue, Sesil and Kroumova studied companies providing broad-based stock options using NCEO lists. Companies were included only if they provided stock options to 50% or more of their non-management employees and were in business from 1995 through 2002. From this, two datasets were created, one of 463 companies for 1995–1997, a period of rising stock prices, and one of 367 companies for 2000–2002, a period of falling prices. Companies were then classified as small (under 500 employees), medium (500–5,000), or large (over 5,000). Using data from Compustat, the researchers matched the broad options companies with comparable companies in their 2-digit SIC codes. Because the researchers could not know when the plans were established, they used a technique called a random effects estimator to correct for any bias these omitted data might introduce. Productivity was measured as a standard Cobb–Douglas function, a measure of productivity that, as used here, looks at how labor and capital combine to produce output. The researchers adjusted the formula to account for changes in employment and capital over time. In the 1995–1997 period, they found that companies with broad-based options had productivity levels 20–33% higher than comparable firms. The smallest companies and largest companies registered at about a 20% differential, medium sized at 33%. In 2000–2002, medium and large-sized companies retained these differentials; the small company differential declined by a little more than 1%. The results indicate that the declining stock market did not undermine the impact of broad options. Moreover, contrary to popular perception that the incentive effects of options should be lower in larger companies (because individual employee efforts seem to matter less), company size does not seem to be consistently related to performance. 5.6. Studies on Executive Equity Compensation and Corporate Performance Ittner et al. (2003) provides a good lead-in for this section on research on executive equity compensation and corporate performance. Studies on this topic are legion and support can be found for almost any position. As noted at the outset of this section on research, however, these studies typically suffer from some unavoidable methodological problems that make it

The Future of Broad-Based Options

387

exceedingly difficult to disentangle causes and effects. If company executives get more equity compensation in companies where stock prices are higher than industry norms (or subsequently increase more), does that mean that this higher equity pay caused the result? That might be the case, but consider alternative explanations: Given that executives often effectively determine or highly influence how boards decide to pay them, might executives who believe their stock price is going to outperform the market not want more stock pay? Because these executives have information not yet available to the investing public, they are in a good position to assess the role stock should play in their pay. If a company’s stock is performing well, or is expected to perform well, but its cash position is less robust, might not compensation committees favor more stock in executive pay? In other words, might better stock performance cause rather than result from executive compensation practices? If executives are paid more in stock, might they not be tempted to take actions that push up short-term stock prices even though longer-term corporate performance might suffer? While clearly this has happened in dubious if not illegal ways in some companies, the actions need not be so sinister. An executive might postpone needed R&D expenditures, for instance, or reduce staffing, both of which would pump up short-term earnings but make it difficult to respond effectively to future challenges. In less well-designed research, studies simply look at whether executive equity pay is higher in companies with relatively stronger stock prices. This begs the question of whether executives caused the higher prices or, alternatively, stock was just a more readily available corporate currency. These studies, therefore, have value only if they show the converse. If there is no relationship, there is not much reason to argue that executive pay causes improved stock prices (nor is there reason to say it hurts them). More careful studies look at subsequent stock price movements, controlled by industry and other factors (as Ittner et al., 2003 did), but there too the results are much more persuasive in arguing casualty where there is no relationship than when they find a positive or negative relationship. With these limitations in mind, we can look at some of the major studies in this field. This is not a comprehensive list, but it is also not selective. We ran a web search or recent articles, examined footnotes in those articles for articles we missed, and looked at the extensive references in Kruse et al. (2003) and included any studies since 1998 that appeared methodologically rigorous. We looked particularly at studies of studies, as they provide useful summary information. We found that we ended up agreeing with what these summary

388

COREY ROSEN

studies all report: there is no consistent relationship between executive equity compensation and improved returns to shareholders or corporate performance. In fact, the weight of evidence makes it very difficult to argue that paying executives more than industry norms in equity pay improves performance, and that these norms themselves may be entirely unjustified. 5.7. What the Studies Find Perhaps the most useful summary study is by Robert Grams (2003). He notes that there are three main criticisms of current stock option grant design: (1) Executives are encouraged to assume excessive levels of risk because while stock options provide rewards for success, they do not offer a downside risk should an executive’s approach cause failure. (2) A company’s stock price is influenced by a number of factors; this weakens the connection between executive performance and their gains from equity-based pay. (3) Large stock holdings with short holding periods encourage focus on short-term stock price, rather than on long-term corporate health and wealth creation. Grams surveyed the literature on the relationship between executive ownership and company performance, reviewing 229 studies. He found the correlations were too weak to draw substantial conclusions. A more complex view, he argues, would conclude that when insider ownership (including by directors) exceeds 20–30%, management entrenchment can occur, making change and adaptability less likely. Below 20–30%, increased management ownership may be effective, but above that there is either no relationship or a negative one. Grams notes, however, that the research often suffers from serious flaws, including counting all vested stock options and those that are scheduled to vest within 60 days (the standard required for proxy reporting). Grams contends that when the data are looked at in terms of actual ownership versus contingent awards (equity awards yet to vest), a clearer relationship emerges.3 Grams says three main conclusions can be drawn from empirical research on publicly traded companies that provide stock options to those at relatively small or large compensation levels: (1) In large companies, giving an unusually large portion of all option grants to executives compared to other employees does not generate improved corporate performance.

The Future of Broad-Based Options

389

(2) Research on the causal relationship between option award sizes and company performance is mixed and can be inconsistent. (3) Although not definitive, research on CEO pay suggests that option grants might improve company performance at lower grant levels, but cause harm at above-average grant levels. Grams argues that options have proven to be an inefficient way to reward executives and that more direct ownership should be encouraged instead, such as through stock purchase programs. This argument is taken up by Jack Dolmat–Connell (2003), who argues that the level of actual executive stock ownership drives company performance. The study looked at 106 of the largest companies in 10 diverse industries as well as paired comparisons of 17 industry performance leaders and performance laggards (measured by stock price performance). It then analyzed whether there were significant differences in terms of beneficial ownership and stock options held. The companies with the highest beneficial ownership by executives ranked, on average, in the 94th percentile in terms of the following year’s stock price growth; companies in the 75th percentile of beneficial ownership ranked in the 63rd percentile in terms of stock growth. By contrast, companies ranked at the bottom or in the 25th percentile in terms of beneficial ownership averaged about the 23rd percentile in terms of the next year’s stock price growth. When stock options holdings relative to actual ownership were compared, companies where the ratios of options to ownership were highest did the worst, and vice versa. The same result held in paired comparisons of 17 industry leaders versus 17 industry laggards. In each case, the leaders had more beneficial ownership by executives. These findings are statistically impressive, but Dolmat-Connell goes too far in arguing a causal relationship. Top executives have access to privileged information, and where they expect stock prices to go up in the short term, they would be expected to hold on to shares and buy shares. By contrast, executives holding a lot of options would be reticent to exercise and hold them if they were skeptical about short-term stock price movements. Moreover, as Grams points out, studies that look at short-term stock price movements necessarily are subject to a lot of random variation because short-term (one to three years) stock prices fluctuate for reasons that are only poorly related to corporate performance. Presumably, though, if ownership matters, it matters because it motivates executives to improve performance, which improves stock prices. The randomness of stock price movements in the following year makes this argument hard to support with

390

COREY ROSEN

research. A more sinister explanation, of course, would be that executives who own more stock actually manipulate corporate earnings to jack up stock prices. While this clearly has happened in recent years, we would not argue that this is a widespread practice. Another comprehensive and current analysis is by Core et al. (2003). They note that ‘‘despite considerable prior research, the performance consequences of equity ownership remain open to question.’’ They find some studies indicating options and other equity compensation seem related to better stock performance and some that do not, but indicate that none of these studies really can prove a causal connection. Much of the research they cite focuses on whether options are an efficient way to pay executives given that theory and empirical data indicate that executives value their options grants at less, and often considerably less, than their Black–Scholes values. What this means is that the estimated cost to the company of an equity award (Black–Scholes) is greater than the value the executive perceives, so the executive has to get more equity to arrive at an equivalent value to what would have been delivered by cash. Other research, however, suggests that equity compensation may be more efficient for some companies depending on their cash positions, their risk profiles, the executives’ risk profiles, and other issues.4 Blasi and Kruse (2004) analyzed compensation for the top five executives and corporate performance in the 1,500 largest U.S. publicly traded companies. The study found that executive compensation, most of which has been in the form of options, increased in years when the stock prices of their companies went up. But when Blasi and Kruse examined whether increases in total compensation (again, primarily in options) were related to subsequent increases in stock prices or total shareholder return over the next three or five years, they found a slightly negative relationship (declines of less than 1% per year). They found the same result when they looked at the ultimate profit made from options exercises as opposed to the Black–Scholes value at grant. The study examined what are technically called ‘‘marginal’’ increases. In this context, that does not mean ‘‘small,’’ but rather is a way percentage increases in compensation result in corresponding increases (or decreases) in company performance. This study is one of the most comprehensive analyses of whether increases in executive compensation or options awards do what they are supposed to do – provide incentives for top management to improve returns to shareholders. Whether these increases in compensation actually motivate executives to behave any differently is unclear, but it is very clear that they do not result in the expected gains for shareholders.

The Future of Broad-Based Options

391

Finally, a Watson Wyatt study (2004) found that employees undervalue stock options by 30–50% relative to the Black–Scholes value. The study was based on a survey of 650 high-income individuals. It found that employees who were better able to identify their company’s stock price worked for larger companies, and worked for companies with better stock price performance all placing smaller discounts on options, while employees getting larger grants and with higher incomes discount options more. Restricted stock is only discounted by a mean of 18%, by contrast. Unlike options, restricted stock actually gives employees shares of stock. The stock does not actually belong to employees, however, until certain requirements are met, such as working for a defined number of years or meeting a performance target. Because employees can benefit from restricted stock even if the value of shares declines, fewer shares of restricted stock are needed to provide the same amount of economic value. It should be noted that the Black–Scholes formula is widely considered to overstate the value of options somewhat, although not by 30–50%. Also of interest is that higher-income employees and those getting larger grants discount options more, meaning that very large grants to very high-income employees (such as top executives) are the least likely to get a proportionate bang for their buck. So companies have to keep granting more and more options to higher-income employees to provide an incentive, while grants to nonmanagement employees can produce a much better return on investment.

6. CONCLUSION It is important to view any study of the relationship between compensation strategies and corporate performance with some caution. People’s behavior in organizations is highly overdetermined. Think of all the things that motivate you and your colleagues at work – pay, benefits, relationships, personal issues outside work, individual work ethics, the content of what you do, work organization, relationships with superiors, what you ate that day, and on and on. Identifying any set of factors across a company and saying that is what makes the difference is no easy task. But this just tells us what motivates people. Attempts to then link motivation to individual performance, and then to link individual to corporate performance, face a similarly daunting list of complications. For instance, does working harder really improve corporate performance much if work structures prevent people from sharing ideas and information on how to do things better?

392

COREY ROSEN

Ideal studies look at large numbers of companies, identify measurable changes that have been instituted in discrete ways and specific times (rather than creeping changes, such as gradual increases in benefits), index out industry effects, and compare before-and-after performance. This is practical in the case of employee ownership, because these plans usually have specific starting points. It is harder with executive compensation because it is rare to find executives who started off with no ownership, then received significant amounts into their tenure. Nonetheless, the data are strikingly consistent. Broad-based ownership seems to improve corporate performance most of the time. Narrowly focused ownership has at best an uncertain impact and, in most analyses, a neutral or negative one. It is thus disheartening that many corporations now report that they will move away from what has been proven to work for shareholders toward something that appears not to work. But in a free market, how can this be? If it is more rational for a company to share ownership more widely, won’t the market make that happen? To a considerable extent, it has, as witnessed by the tremendous growth of broadbased ownership over the last 30 years. As the General Social Survey data indicate, it appears that about 40% of employees who work for companies that have stock actually own stock in their employers. That is an astonishing number. If the data reported earlier indicating about 40% of companies with broad-based option plans and about a third of companies with ESPPs will drop them as valid, this would still only reduce the percentage of employees holding stock in their employer by perhaps 5%, from about 40% to about 35%. We believe the actual behavior will be less dramatic. Still, enormous amounts of seemingly irrational equity grants are going to top executives. The explanation is simple. While it may not be rational for the company they head to do this, it is entirely rational for them. Given the short tenure of many top executives, the fact that narrowly distributing ownership may not be good for stock prices in the long run is hardly their rational concern as individual wealth maximizers. As long as they and the boards who go along with them can control their compensation, there is not much reason to expect greed to morph into responsibility. That kind of change will depend on shareholders taking much more activist views on ownership distribution than they currently do. It is not enough for institutional investors to rage, as they now often do, about ‘‘too much dilution.’’ They need to rage too about who gets how much as well. Whatever happens to the use of broad-based options, the future for broad-based employee ownership looks very positive. There are several reasons for this. First, as the economy increasingly emphasizes innovation,

393

The Future of Broad-Based Options

creativity, and information over the simple rote repetition of production routines, employees are being asked to contribute a great deal more of their intellectual capital. This is true at all kinds of companies at all levels, whether it is the stock clerk with a better idea for counting inventory or the programmer with better software for recording it. As employees have to make more decisions about more things, companies naturally want, and need, them to think and act like owners – people who understand what needs to be done to make a profit and have an incentive to act on it. Second, in the U.S. and in other countries, there are significant tax incentives for employee ownership in various forms (ESOPs in the U.S., for instance, have substantial benefits). Policymakers in many countries have decided that broad-based ownership is a free market solution to providing greater economic justice without compromising economic incentives. As author Thomas Friedman (2005) has argued in his new book, The World Is Flat, employee ownership is a very effective means to deal with the inequalities and other challenges globalization poses to ordinary employees. Third, as noted, employee ownership can provide a meaningful competitive advantage. While this advantage must compete against the more narrow short-term interest of other corporate players, in the long run, it is bound to become more important.

NOTES 1. A number of other studies looked at how the impact of expensing affects stock prices, but in an associational way rather than a way that could indicate whether mandatory expensing per se would lead to a different result in stock prices than already obtained through footnote disclosures. Studies of this type are described in Li’s paper and include Aboody (2002), Aboody, Barth, and Kasznik (2002), Rees and Stott (1999), and Bell, Landsman, Miller, and Yeh (2002). 2. The Wages and Hours Act divides employees into hourly (non-exempt) and salaried (exempt) employees and defines who must be included in each category. ‘‘Non-exempt’’ employees are thus hourly employees and are subject to the requirements of the Act. Generally, the Act sets rules for overtime pay and other compensation issues for hourly employees, but it does not cover employee ownership plans or many other benefit arrangements. 3. He cites a study by Ofek and Yermack (2000) showing that executive options rarely lead into increases in actual ownership, except perhaps at low levels of initial ownership. 4. For further details on studies on this issue, go to Encycogov.com. Its table ‘‘Managerial Ownership and Performance – Empirical Studies’’ groups dozens of studies on the issue according to their principal hypotheses. The site identifies many more studies documenting a zero or negative relationship between managerial ownership and company performance, rather than a positive relationship.

394

COREY ROSEN

5. The individual data come from several questions included in the General Social Survey (GSS), a 2002 random sampling of working adults performed by the National Opinion Research Center (NORC) of the University of Chicago. The organizational data come from NORC’s National Organizations Survey (NOS), performed in 2003. Joseph Blasi and Douglas Kruse of Rutgers and Richard Freeman of Harvard, all affiliated with the Shared Capitalism Project of the National Bureau of Economic Research, organized the questions and did subsequent analysis. Further information on the data is to be found on the NCEO’s web site, http://www.nceo.org.

REFERENCES Aboody, D. (2002). Market valuation of employee stock options. Journal of Accounting and Economics, 22, 357–391. Aboody, D., Barth, M., & Kasznik, M. (2002). SFAS 123 stock-based compensation expense and equity market values. Working paper, University of California at Los Angeles and Stanford University. Bell, T., Landsman, B., Miller, B., & Yeh, S. (2002). The valuation implications of employee stock option accounting for profitable computer software firms. The Accounting Review, 77(4), 971–996. Blair, M., Kruse, D., & Blasi, J. (2002). Is employee ownership an unstable form or a stabilizing force. Employee Ownership and Corporate Performance, National Center for Employee Ownership (Winter). Blasi, J., & Kruse, D. (2004). Corporate governance, executive compensation, and strategic human resource management from 1992–2002. Rutgers University, Paper in submission. Blasi, J., Kruse, D., & Bernstein, A. (2003). In the company of owners. New York: Basic Books. Collat, D. (1995). Public company ESOPs and corporate performance. National Center for Employee Ownership. Core, J., Guay, W., & Larcker, D. (2003). Executive equity compensation and incentives: A survey. FRBNY Economic Policy Review, (April), 27–50. Deloitte & Touche. (2003). Technology stock compensation survey: Looking beyond options. Deloitte & Touche. Deloitte & Touche. (2004). Stock options: Down but not out. Deloitte & Touche. Dolmat-Connell, J. (2003). Ownership, not options, drives performance: Rethinking and restructuring executive compensation. WorldatWork Journal, Q4, 19–27. Ducy, M., Iqbal, Z., & Akhigbe, A. (1997). Employee stock ownership plans and cash flow performance of publicly traded firms. American Business Review, 15(2). Ericson, R., & Grund, M. (2004). Options expensing announcement has no impact on share prices. Towers Perrin, March 31. Friedman, T. (2005). The world is flat: A brief history of the twenty-first century. New York: Allen Lane. Garg, A., & Wilson, T. (2003). Expensing stock options: What are the markets saying. The Journal of Employee Ownership Law and Finance, 15(4), 3–21. Grams, R. (2003). Consequences of executive stock options and ownership. WorldatWork Journal, Q3.

The Future of Broad-Based Options

395

Hager, E. (2003). Do employee stock option grant announcements affect shareholder value? Journal of Employee Ownership Law and Finance, National Center for Employee Ownership, (Summer), 12–18. Ittner, C. D., Lambert, R. A., & Larcker, D. F. (2003). The structure and performance of equity grants to employees of new economy firms. Journal of Accounting and Economics, 34(1–3), 89–127. Kardas, P., Keogh, J., & Schiff, A. (1998). Wealth and income consequences of employee ownership: A comparative study from Washington state. National Center for Employee Ownership. Kruse & Blasi (2000). Largest study yet shows ESOPs improve corporate performance. National center for Employee Ownership. www.nceo.org Kruse, D., & Blasi, J. (2003). Corporate governance, executive compensation, and strategic human resource management from 1992–2002. Paper in submission. Kruse, D., Blasi, J., Kroumova, M., & Sesil, J. (2001). Stock options, corporate performance, and organizational change. National Center for Employee Ownership. Li, H. (2002). Employee stock options, residual income valuation and stock price reaction to SFAS 123, draft document. ERN Labor Journals, (December). Mehran, H. (1999). Unleashing the ownership dynamic – Creating connections through employee ownership – A research summary. Hewitt Associates. Mellon Financial. (2003). Responding to mandatory option expensing: Human resources and investor solutions. Published on website of Mellon Financial. Mercer Human Resource Consulting. (2003). Future of equity – 2003 update. Mercer Human Resource Consulting. Michel, N., & Garwood, P. (2002). Expensing employee stock options: Lifting the fog. Center for Data Analysis of the Heritage Foundation report 02–06. Ofek, E., & Yermack, D. (2000). Taking stock: Equity-based compensation and the evolution of managerial ownership. Journal of Finance, 55(June), 1367–1384. Quarrey, M., & Rosen, C. (1987). Employee ownership and corporate performance. Harvard Business Review, (September/October), 126–131. Rees, L., & Stott D. (1999). The value-relevance of stock-based employee compensation disclosures. Working paper, Texas A&M University. Sesil, J., & Kroumova, M. (2004). Broad-based stock options before and after the market meltdown. Paper in submission. Sibson Consulting, & WorldatWork. (2004). The state of stock options 2003. WorldatWork. Watson Wyatt Consulting. (2004). How do employees value stock options: Results from a survey. Watson Wyatt Consulting. Winther, G. (1995). Employee ownership: A comparative analysis of growth performance. Aalborg: Aalborg University Press.

APPENDIX. THE INCIDENCE OF EMPLOYEE OWNERSHIP IN THE U.S. The General Social Survey Data on Employee Ownership in the U.S. collected in 2002 indicate that 23.3% of all employees working for for-profit companies report owning stock in their companies, while 14.4% hold stock

396

COREY ROSEN

options. The two categories are exclusive of one another, but other data indicate that almost all employees holding options also own company stock in other ways. That would mean that approximately 25 million Americans own employer stock through ESOPs, options, stock purchase plans, 401(k) plans, and other plans; while 14.6 million hold stock options (and, usually, other stock). In 2002, 4.8% of employees received option grants, but because many programs are not annual, the number holding options is much larger. Looked at on the basis of companies, 37.8% of for-profit companies provide employee ownership plans for 50% or more of their employees, and 16.3% grant options to a similarly broad group.5 The data provide the most representative analysis of the penetration of employee ownership to date. Previously, the NCEO had estimated that 25 million Americans own stock through one kind of employee ownership plan or another, essentially the same as the results here. We also estimated that about 10 million employees held stock options in 2002, less than the 14.6 million estimated in this survey. The survey may somewhat overstate the number of options holders, however, because some employees may have misinterpreted the question on options to include participation in an employee stock purchase plan (although a prior question made a distinction between stock options and direct purchases of stock). Because of the differing estimating techniques and inevitable ambiguities of data collection and interpretation in this area, however, variations in results are to be expected. These numbers, as well as any other estimates, should always be understood with the caveat that they cannot be precise. The data would likely be different if compiled in the coming year or two, because many companies say they will cut back on broad-based options, as well as ESPPs, in response to expected changes in accounting rules. We estimate that the number of employees holding company stock could drop by 3–4 million people as a result, but that competitive labor market pressures will likely make these short-term changes. In addition to these overall findings, the survey looked at how much employees actually own through their plans. Here, respondents were asked to indicate how much their ownership stakes would be worth, assuming they would be fully vested, if they sold it today. The results are presented in Table A1. The mean and median differ so much because a small minority of employees hold very large stakes. Nonetheless, the median values are impressive, especially considering that many of these plans are fairly recent, and employees will build up considerable additional value over time. Those who argue that employee ownership usually provides only a trivial financial stake

397

The Future of Broad-Based Options

Table A1.

Size of Ownership Stake.

Mean in dollars Median in dollars Mean as % of annual pay Median as % of annual pay

$84,409 $10,000 99.6 21.2

Source: General Social Survey on employee ownership.

Table A2.

Variations in Stock Ownership.

Have Company Stocka (%)

Have Stock Options (%)

Industry Computer services Comm./utilities Finance/insurance Manufacturing Others

58.3 55.3 39.8 31.5 16.8

56.5 42.6 27.1 20.0 7.4

Earnings o15,000 $15,000–$30,000 $30,000–$50,000 $50,000–$75,000 $75,000+

5.5 18.0 28.4 36.6 50.7

4.0 9.7 14.9 24.3 41.3

Company size o50 Employees 50–499 Employees 500+ Employees

12.0 25.1 38.7

6.0 18.0 23.4

Note: The complete results are available on the NCEO’s Web site, www.nceo.org. Source: General Social Survey on employee ownership. a Does not include stock options.

clearly are off the mark; these numbers are not too far from the median 401(k) holdings (the mean was about $57,000 in 2002; recent median data were not available, but were $11,300 in 1996 and would probably be about $17,000 today). Most participants in ESOPs, ESPPs, and stock option plans also are in 401(k) or other, usually (but not always) diversified retirement plans. Publicly traded companies are much more likely to offer employee stock programs than are closely held companies (16.4% compared to 4.2% of companies), partly because it is easier for them to do so and partly because

398

COREY ROSEN

many closely held companies are partnerships, LLCs, or sole proprietorships and do not have stock. Similarly, larger companies are more likely to have stock programs. There is a gradual increase in the prevalence of employer stock as company size increases in all closely held companies, with 46.8% of employees in companies with more than 2,000 workers holding stock. Public companies also show this pattern, but the numbers level out at around 21% of employees in plans for companies with more than 500 employees. As expected, employees in the technology sector are the most likely to own stock, followed by finance/insurance. Surprisingly, union members are about as likely to hold stock as the overall employee population. Tenure does not seem much of a factor after two years, but, predictably, people making higher salaries are more likely to be included. Table A2 provides more details.