Social Capital: Reaching Out, Reaching in

Social Capital Social Capital Reaching Out, Reaching In Edited by Viva Ona Bartkus University of Notre Dame, USA an...

Author: Viva Ona Bartkus | James H. Davis

71 downloads 6889 Views 3MB 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

Social Capital

Social Capital Reaching Out, Reaching In

Edited by

Viva Ona Bartkus University of Notre Dame, USA and

James H. Davis University of Notre Dame, USA

Edward Elgar Cheltenham, UK • Northampton, MA, USA

© Viva Ona Bartkus and James H. Davis 2009 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2008939736

ISBN 978 1 84720 072 3 Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall

Contents List of ﬁgures List of tables List of contributors Acknowledgements

vii viii ix xv

Introduction: the yet undiscovered value of social capital Viva Ona Bartkus and James H. Davis PART I

OVERVIEW

1 What is social capital? Elinor Ostrom 2 Network duality of social capital Ronald S. Burt 3 On the costs of conceptualizing social ties as social capital Robert M. Fishman PART II

1

17 39 66

SOCIAL CAPITAL: REACHING OUT

4 How’s the job? Are trust and social capital neglected workplace investments? John F. Helliwell, Haifang Huang and Robert D. Putnam 5 Social capital eﬀects on student outcomes Maureen T. Hallinan 6 Communities, schools and voter turnout: a case study in social norms David E. Campbell 7 Experimental approaches to the diﬀusion of norms David W. Nickerson 8 Capitalizing on connections: social capital and strategic management Janine Nahapiet

v

87 145

160 186

205

vi

PART III

Contents

SOCIAL CAPITAL: REACHING IN

9 Social capital creation: collective identities and collective action Roderick M. Kramer 10 Connecting to brokers: strategies for acquiring social capital Daniel J. Brass 11 Trust, distrust and building social capital Roy J. Lewicki and Chad T. Brinsﬁeld 12 Understanding social capital: in whom do we trust? Darryl Stickel, Roger C. Mayer and Sim B. Sitkin 13 Organizational trust and social capital James H. Davis and Viva Ona Bartkus

239 260 275 304 319

Conclusion: frontiers of social capital research Viva Ona Bartkus and James H. Davis

339

Index

357

Figures 2.1 2.2 2.3 2.4 2.5 2.6 4.1 6.1 6.2 6.3 10.1 11.1 12.1 13.1

Manager network illustrating opportunities for brokerage and closure Performance and brokerage Closure and banker reputation stability Closure slows network decay, especially in new relationships Discussion network in a virtual organization Network duality failure modes Life satisfaction at different levels of trust in management, from 2003 Canadian ESC Political competition and voter turnout in the 2004 presidential election Political competition and motivations for voting Impacts on the probability of voting in 1980 Brokerage requires connecting to brokers Integrating trust and distrust: alternative social realities Stickel matrix of forms of social control Model of organizational trust and social capital

vii

41 44 48 51 54 61 91 168 171 179 266 281 313 323

Tables I.1 1.1 1.2 4.1 4.2 4.3 4.4 4A.1 5.1 5.2 5.3

5.4

7.1 7.2 7.3 11.1 11.2 13.1 13.2 13.3 C.1 C.2

Definitions of social capital Relationships between governance structure and physical condition of irrigation systems Relationships between governance structure and cropping intensity of irrigation systems Comparing well-being effects of income and workplace trust Expanded models on the well-being effects of income and workplace trust Formation of workplace trust Compensating differentials Descriptive statistics Descriptive statistics for social network variables Descriptive statistics for variables in regression analyses Regression of 2004 reading and mathematics achievement on social networks and background factors for Catholic schools in the Chicago School Study Effect of social network factor in regressions of non-academic outcomes on social networks and background factors for Catholic schools in the Chicago School Study OLS regression results modeling the peer effects of roommates and suitemates Effect of peer-to-peer mobilization Contagion effect of voting within the household Integrating trust and distrust: alternative social realities Contributors to enhanced or reduced social capital Chronbach alphas and correlations Regression – dependent variable: organizational trust in team Regression – dependent variable: organizational trust in program Benefits (costs) of social capital Benefits (costs) of social capital within management arena

viii

3 29 29 94 96 116 131 136 152 155

156

157 192 196 199 283 288 329 332 333 340 341

Contributors Viva Ona Bartkus is an Associate Professor of Management at the University of Notre Dame. Dr Bartkus graduated summa cum laude from Yale University with Masters and Bachelors degrees in economics, and then completed her Doctorate and Masters in international relations at Oxford University, while on a Rhodes Scholarship. Her book investigating nationalism, self-determination and why groups attempt secession, The Dynamic of Secession, was published by Cambridge University Press (1999). Her current research and teaching interests concentrate on two distinct areas: the social capital of communities that enables collaboration and the leadership approaches most eﬀective in solving complex business problems. Her book, Getting It Right, with Ed Conlon, published by Jossey Bass (2008), argues that successful leadership depends on making values-based problem-solving a habit of mind. Prior to joining the faculty at Notre Dame, Dr Bartkus spent ten years serving healthcare, industrial, retail and high-tech clients to overcome their strategic, operational and organizational challenges while at the global management consulting ﬁrm of McKinsey & Company, the last four years as a partner of the ﬁrm. Daniel J. Brass is J. Henning Hilliard Professor of Innovation Management at the Gatton College of Business and Economics at the University of Kentucky. He is associate editor of Administrative Science Quarterly and Director of LINKS: The International Center for the Study of Social Networks in Business. He is a pioneer in the study of social networks in organizations with more than 1000 citations to his work. Chad T. Brinsﬁeld is a doctoral candidate in Labor and Human Resource Management at the Fisher College of Business, the Ohio State University, where he also received his MBA. His extensive management and entrepreneurial experience in the ﬁelds of marketing, operations and logistics informs his research interests including employee silence, trust within organizations and organizational commitment. Ronald S. Burt is the Hobart W. Williams Professor of Sociology and Strategy at the University of Chicago Graduate School of Business. Dr Burt’s research, consulting and teaching concern the social structure of competitive advantage (for example, Brokerage and Closure, 2005, Oxford University ix

x

Contributors

Press). Applications focus on manager networks (how people of diverse backgrounds create social capital and its eﬀect on their careers) and the network structure of market proﬁts (how the structure of producers, suppliers and consumer relations deﬁnes competitive advantage among producers). Past positions held by Dr Burt include the Shell Professor of Human Resources at the Institute European d’Administration d’Aﬀaires (INSEAD), 1998–2001; Director of Raytheon’s Leadership Institution, 2000–03; and Vice-President of Strategic Learning, Raytheon Company, 2001–03, with the broad task of establishing a Leadership Institute to enhance the strategic thinking of middle and senior managers in a collaborative network across the company. David E. Campbell is the John Cardinal O’Hara, CSC Associate Professor of Political Science at the University of Notre Dame. He is the author of Why We Vote: How Schools and Communities Shape Our Civic Life (Princeton University Press), editor of A Matter of Faith: Religion in the 2004 Presidential Election and co-author of Democracy at Risk (all with the Brookings Institution Press). He has also published numerous articles in academic journals. James H. Davis is the John F. O’Shaughnessy Professor of Family Enterprises and the Ray and Milan Siegfried Director of Entrepreneurial Studies at the University of Notre Dame. He received his PhD in corporate strategy from the University of Iowa, his MBA from Idaho State University and his MEd and BA from Brigham Young University. His research has appeared in many publications including the Strategic Management Journal, the Academy of Management Review, Journal of Applied Psychology and International Journal of Value-Based Management. His publication on ‘trust’ has been recognized as the best research for the second decade of the Academy of Management Journal. Robert M. Fishman is Professor of Sociology and Kellogg Institute Fellow at the University of Notre Dame. His most recent books include Democracy’s Voices, winner of the 2005 Honorable Mention for Best Book within the Political Sociology Section of the American Sociology Association, and The Year of the Euro (co-edited with Anthony Messina). He is currently engaged in a collaborative cross-national project on factors shaping the evolution of priestly vocations and in a macro-level comparison of democratic practice in Portugal and Spain focusing on the enduring consequences of the fundamentally diﬀerent paths to democracy followed by those neighboring countries in the 1970s. Maureen T. Hallinan is the White Professor of Sociology and Director of the Center for Research on Educational Opportunity, Institute for

Contributors

xi

Educational Initiatives at the University of Notre Dame. She received a BA from Marymount College, Tarrytown, NY, an MS from the University of Notre Dame and a PhD in Sociology and in Education from the University of Chicago. Professor Hallinan’s research is primarily in the sociology of education. She was the principal investigator of the study Comparative Analysis of Best Practices in Public and Private Schools, a ﬁve-year project funded by the US Department of Education to determine the best practices adopted by educators to promote student learning and social development. Professor Hallinan has been the recipient of research grants from the National Science Foundation, the National Institute of Child Health and Human Services, the US Department of Education and the Spencer Foundation. Professor Hallinan has published over 100 articles in professional journals, is the author or editor of eight books and has several chapters in edited volumes. She edited The Handbook of the Sociology of Education, 2000 (New York, Plenum Press) and most recently School Sector and Student Outcomes, 2006 (University of Notre Dame Press). She was editor of Sociology of Education from 1981 to 1986. Professor Hallinan was President of the American Sociological Association in 1995–96 and President of the Sociological Research Association in 2000. She was elected to the National Academy of Education in 1999 and was Vice-President from 2000 to 2004, and is the recipient of the Willard Waller Award from the American Sociological Association (2004). John F. Helliwell is Arthur J.E. Child Foundation Fellow of the Canadian Institute of Advanced Research and Co-director (with George Akerlof) of its program on Social Interactions, Identity and Well-Being. He is also Professor Emeritus of Economics at the University of British Columbia, a member of the National Statistics Council, Fellow of the Royal Society of Canada and Oﬃcer of the Order of Canada. His books include Globalization and Well-being (UBC Press, 2002). Recent articles include ‘Well-being, social capital and public policy: what’s new?’ (Economic Journal, March 2006), and ‘Well-being and social capital: does suicide pose a puzzle?’ (Social Indicators Research, 2007). Haifang Huang is an Assistant Professor in Economics at the University of Alberta. His ﬁelds of research include economics of well-being and social capital. His recent papers include ‘How’s your government? International evidence linking good government and well-being’ (with John Helliwell, forthcoming in British Journal of Political Science) and ‘How’s the job? Well-being and social capital in the workplace’ (working paper with John Helliwell).

xii

Contributors

Roderick M. Kramer is the William R. Kimball Professor of Organizational Behavior at the Stanford Business School. He is the author or co-author of more than 100 scholarly articles. His work has appeared in leading academic journals, as well as popular journals. He is the co-author of numerous books, including Negotiation in Social Contexts, Trust in Organizations, Power and Inﬂuence in Organizations and Trust and Distrust Within Organizations. He has been on the editorial boards of numerous leading academic journals including Administrative Science Quarterly and Organization Science. Roy J. Lewicki is the Dean’s Distinguished Professor and Professor of Management and Human Resources at the Max M. Fisher College of Business, Ohio State University. Professor Lewicki received his BA in psychology from Dartmouth College and his PhD in social psychology from Columbia University. Professor Lewicki maintains research and teaching interests in the ﬁelds of negotiation and dispute resolution, trust development, managerial leadership, organizational justice and ethical decisionmaking. He is the author of numerous research articles in these ﬁelds, as well as an extensive repertoire of teaching materials. He is an author or editor of 31 books, including The Master Negotiator (Jossey Bass/Wiley, 2007), Negotiation (McGraw-Hill/Irwin, 5th edn 2006) and Making Sense of Intractable Environmental Conﬂicts: Frames and Cases (Island Press, 2003). He was selected as a Fellow of the Academy of Management in 2003, and in 2005 he received the 16 000-member Academy of Management’s Distinguished Educator Award. Roger C. Mayer is a Professor of Management at the University of Akron. He previously served on the faculties of Notre Dame, Purdue and Baylor. Dr Mayer received a PhD in organizational behavior and human resource management from the Krannert Graduate School of Management at Purdue University. Mayer’s research is focused on trust, employee decision-making, attitudes and productivity, and has been published in such leading journals as Academy of Management Journal, Academy of Management Review, Journal of Applied Psychology, Journal of Organizational Behavior and Strategic Management Journal. He is a member of the editorial review board for Academy of Management Journal. Janine Nahapiet is Associate Fellow of Said Business School and Templeton College, University of Oxford. She has served on the faculties of Templeton College, Oxford University, London Business School and Imperial College, London and was until recently Lead Research Fellow of the ESRC/EPSRC Advanced Institute of Management Research. She has

Contributors

xiii

a BA (Hons) in Sociology and Psychology from the University of Sheﬃeld, a Graduate Diploma and PhD from the University of London and an MA by special resolution from the University of Oxford. She has published widely on aspects of strategy and organization and her co-authored paper ‘Social capital, intellectual capital and the organizational advantage’ was adjudged the Academy of Management Review Best Paper in 1998. She served on the advisory committee of the ESRC Research Program on the Evolution of Business Knowledge and on the Regulatory Decisions Committee of the UK Financial Services Authority. Her research focuses on the links between social capital, innovation and knowledge; creating value through networks; the theory and practice of cooperation; strategy and organization in professional service ﬁrms; and organization theory and design, especially innovative forms of organizing. She recently moved to a portfolio career, enabling her to focus her academic research and work with executives on the challenge of building organizational capabilities that meet the needs of the twenty-ﬁrst century. David W. Nickerson is an Assistant Professor of Political Science at the University of Notre Dame. His research uses ﬁeld experiments to study voter mobilization and how behaviors and norms diﬀuse through social networks. He has worked extensively with partisan and non-partisan groups to evaluate and hone outreach techniques. Examples of his academic work can be found in the American Journal of Political Science, the Journal of Politics and Political Analysis. Elinor Ostrom is Arthur F. Bentley Professor of Political Science, Codirector of the Workshop in Political Theory and Policy Analysis, Indiana University, Bloomington, and Founding Director, Center for the Study of Institutional Diversity, Arizona State University. She is a member of the American Academy of Arts and Sciences, the National Academy of Sciences and the American Philosophical Society, and a recipient of the Frank E. Seidman Prize in Political Economy, the Johan Skytte Prize in Political Science, the Atlas Economic Research Foundation’s Lifetime Achievement Award, the John J. Carty Award for the Advancement of Science and the Cozzarelli Prize from the Proceedings of the National Academy of Sciences. Her books include Governing the Commons (1990), Rules, Games, and Common-Pool Resources (1994, with Roy Gardner and James Walker), Local Commons and Global Interdependence: Heterogeneity and Cooperation in Two Domains (1995, with Robert Keohane), Trust and Reciprocity: Interdisciplinary Lessons from Experimental Research (2003, with James Walker), The Commons in the New Millennium: Challenges and Adaptations (2003, with Nives Dolsak), The Samaritan’s Dilemma: The Political Economy of Development Aid (2005, with Clark Gibson, Krister Andersson and Sujai

xiv

Contributors

Shivakumar), Understanding Institutional Diversity (2005) and Understanding Knowledge as a Commons: From Theory to Practice (2007, with Charlotte Hess). Robert D. Putnam is the Peter and Isabel Malkin Professor of Public Policy at Harvard, Visiting Professor at the University of Manchester and Director of the Harvard–Manchester program on Social Change. He is a member of the National Academy of Sciences, a Fellow of the British Academy and past President of the American Political Science Association. In 2006, Putnam received the Skytte Prize, one of the world’s highest accolades for a political scientist. His two landmark books Making Democracy Work (1993) and Bowling Alone (2000) underlie the surge of interest in social capital. He is currently working on three major empirical projects: (1) the changing role of religion in contemporary America; (2) the eﬀects of workplace practices on family and community life; and (3) practical strategies for civic renewal in the United States in the context of immigration and social and ethnic diversity. Sim B. Sitkin is a Professor of Management and Faculty Director of the Center on Leadership and Ethics at Duke University’s Fuqua School of Business. His research concerns the eﬀects of organizational control and leadership on risk-taking and accountability, sense-making and trust, learning, and change and innovation. His publications have appeared in numerous scholarly and managerial journals and books. He recently completed serving on the Board of Governors of the Academy of Management, as senior editor of Organization Science and as associate editor of the Journal of Organizational Behavior. Darryl Stickel completed his PhD at the Fuqua School of Business at Duke University. Upon completion of his studies, Dr Stickel went to work for the global management consulting ﬁrm of McKinsey & Company. His doctoral thesis focused on building trust in hostile environments. The combination of a deep theoretical understanding of trust and the applied problem-solving skills learned at McKinsey allowed Dr Stickel to found a small consulting agency, Trust Unlimited, which works with a wide variety of organizations on trust-related problems.

Acknowledgements No book dedicated to investigating social capital would be possible without broad collaboration. Ours is no diﬀerent. A simple reference on this page, unfortunately, will not be suﬃcient tribute to the many colleagues that provided their encouragement. Dean Carolyn Woo, Dean of the Mendoza College of Business at Notre Dame, provided ongoing advice and guidance. The Kellogg, Kroc and Nanovic Institutes at Notre Dame generously supported both this scholarly inquiry into the dynamics of social capital and its interdisciplinary research conferences. Our work beneﬁted from the strong support of our publishers, and the patient editing of Sarah Smith and Laura Gerber. Finally, we are grateful to the Bartkus and Davis families. This book would not have been possible without their unwavering belief in our aspirations.

xv

Introduction: the yet undiscovered value of social capital Viva Ona Bartkus and James H. Davis At heart, social capital is a simple concept – that is, relationships matter. Every day we build and maintain relationships with family, friends and colleagues, relationships which we enjoy in their own right, but which also serve to help us through tough times or enable us to achieve our goals. We also intuitively understand that excessive solidarity within a community can become a liability and lead to discrimination and prejudice. Despite its simplicity, social capital has taken the social sciences by storm. However, it has also remained controversial, even with regard to its basic deﬁnition – a rare achievement after two decades of vibrant debate. The question arises – why do scholars bother when resolution eludes the ﬁeld? In other words, why does social capital generate so much interest? At their core, social capital theories help explain behavior. Economic, political and sociological models of human behavior are often unable to explain signiﬁcant diﬀerences in human performance. By examining social capital, scholars hope to overcome the limits of these paradigms. Social capital scholarship contributes crucial theoretical insights by expanding the inquiry into the causes of behavior to include a wealth of new additional factors – such as trust, networks and norms. In eﬀect, social capital scholarship provides a distinct lens that can explain human motivation and actions, especially collaboration, and outcomes. Moreover, this research explores the paradox of cooperative behavior without sacriﬁcing the microassumptions that are the foundation for explanations of macro-phenomena. Despite its complexity and conceptual diﬃculties, the persistent interest in social capital thus arises from the fact that it helps us make sense of why people do what they do. The overarching objectives of this Introduction – and of the entire book – are to highlight the potential, undiscovered value of social capital and to revitalize an interdisciplinary dialogue that builds our understanding of that value. An overview of the manifold ways to deﬁne social capital sets the stage for an argument regarding its major theoretical contributions and a discussion of the inevitable controversies that follow. The scholarship in 1

2

Social capital

the rest of the chapters contributes ground-breaking thinking on many of the controversies – whether around social capital’s conceptualization, impact, creation mechanisms or measurement. After highlighting the most provocative insights from each chapter, the Introduction closes with a few hopes for the book and social capital research in general.

HOW DO SCHOLARS DEFINE SOCIAL CAPITAL? Progress in accumulating knowledge requires parsimonious theories supported by quantiﬁable measurement and assessment. This aspiration, however, can sometimes prove diﬃcult to achieve in the social sciences. A theoretically meaningful deﬁnition of social capital is essential if we are to build successive waves of research on a sound foundation. But before adopting such a deﬁnition, scholars must be convinced that the beneﬁts associated with using it for future research outweigh its inherent limitations or costs. Table I.1 represents a broad snapshot of current deﬁnitions for social capital. At an intuitive level, social capital represents the resources that arise from relationships and that can accrue to either the individual or the collective. These assets can help individuals reach their own goals or assist members of the collective in working toward the common good. Even in this most basic sense, multiple challenges to the deﬁnition of social capital emerge around the type of construct, level of analysis (individual vs collective) and separation of sources from eﬀects. Most researchers treat social capital as a construct with two factors: relationship structure – for example, network size, conﬁguration and density; and relationship content – for example, norms, institutions and values. This is because neither structure nor content can provide meaningful insights on their own. Investigation of both aspects is critical to understanding social capital’s potential impact. Unlike many other concepts in the social sciences, social capital faces issues around the appropriate level of analysis. As illustrated in the following example, the collective’s social capital can beneﬁt the individual, who can use his relationships to learn enhanced skills, gain new information or access opportunities. Alternatively, in the context that Burt (2005) describes as ‘closure’ and Putnam (2000) as ‘bonding social capital’, the beneﬁts may accrue to an in-group, a subgroup of the collective, by increasing the overall trust in the group through faster information flows and better surveillance and sanctioning of unacceptable behavior. Finally, the broader collective may beneﬁt. Burt describes the situation of ‘brokerage’ and Putnam ‘bridging social capital’ in which the broader collective accesses a greater variety

3

Woolcock (1998), p. 153 Ostrom and Ahn (2003), p. xiv Adler and Kwon (2002), p. 23

Nahapiet and Ghoshal (1998), p. 243

Robert Putnam (1993), p. 169 Burt (1992), p. 9

James Coleman (1994), p. 302

Glenn Loury (1992), p. 100 Pierre Bourdieu (with Wacquant, 1992), p. 19

Lyda Judson Hanifan (1920), p. 22

Alexis de Tocqueville (1945 [1840])

‘Self-interest is the original motive to the establishment of justice. [As society develops], a sympathy (or moral sense) with public interest is the source of the moral appropriation, which attends that virtue.’ ‘Americans of all ages, all conditions, and all dispositions constantly form associations . . . Wherever at the head of some new undertaking you see the government in France, or a man of rank in England, in the United States you would be sure to ﬁnd an association.’ ‘. . . that in life which tends to make these tangible substances count for most in the daily lives of a people; namely, goodwill, fellowship, sympathy, and social intercourse among the individuals and families that make a social unit.’ ‘. . . naturally occurring social relationship among persons which promote or assist the acquisition of skills and traits valued in the marketplace.’ ‘Social Capital is the sum of resources, actual or virtual, that accrue to an individual or a group by virtue of possessing a durable network of more or less institutionalized relationships of mutual acquaintance and recognition.’ ‘Social Capital is deﬁned by its function. It is not a single entity, but a variety of diﬀerent entities having two characteristics in common: they consist of some aspect of social structure, and they facilitate certain actions of individuals who are within the structure.’ ‘. . . features of social organization, such as trust, norms, and networks, that can improve the eﬃciency of society by facilitating coordinated action.’ ‘. . . friends, colleagues, and more general contacts through whom you receive opportunities to use your ﬁnancial and human capital.’ ‘. . . the sum of the actual and potential resources embedded within, available through, and derived from the network of relationships possessed by an individual or social unit. Social Capital thus comprises both the network and the assets that may be mobilized through that network.’ ‘. . . the information, trust, and norms of reciprocity inherent in one’s social network.’ ‘. . . social capital as an attribute of individuals and their relationships that enhances their ability to solve collective action problems.’ ‘Social capital is the goodwill available to individuals or groups. Its source lies in the structure and content of the actor’s social relations. Its eﬀorts flow from the information, influence and solidarity it makes available to the actor.’

Deﬁnitions of social capital

David Hume (1751, in McNally (1994), p. 169)

Table I.1

4

Social capital

of information flows, opinions and opportunities. Social capital research will forever face this complexity. But part of its inherent value lies in its ability to connect these diﬀerent levels in a meaningful way. Scholars’ deﬁnitions tend to emphasize either social capital’s sources or its eﬀects. With regards to its sources, a lively discussion persists, for instance, regarding whether relationships should be treated as instrumentally focused actions (means toward ends) or actions taken for their own merit or value (ends in themselves). Both motivations can create the relationships and social structures on which social capital depends. For our purposes, we deﬁne social capital using both Putnam’s functional approach and Nahapiet and Ghoshal’s descriptive one. For Putnam, social capital represents ‘the features of social organization, such as trust, norms, and networks, that can improve the eﬃciency of society by facilitating coordinated action’ (Putnam et al.,1993, p. 169). For Nahapiet and Ghoshal, social capital is ‘the sum of actual and potential resources embedded within, available through, and derived from the network of relationships possessed by an individual or social unit’ (1998, p. 243). Both deﬁnitions are critical to understanding the concept as broadly as possible and highlighting its contributions to the social sciences.

HOW DOES SOCIAL CAPITAL CONTRIBUTE TO EXPLAINING HUMAN BEHAVIOR? As argued earlier, social capital theories expand the factors – such as trust, networks and norms – that we can use to explain troublesome human behaviors such as cooperation. These new elements place it in strong contrast to neoclassical economics, a discipline that applies the rational choice model of human decision-making. This paradigm originally modeled motivation on an assumption of atomistic, selﬁsh and rational human beings. Although economists have already unraveled some of its assumptions, the model still treats cooperation – except under severely prescribed conditions – as a seemingly irrational endeavor. Yet, as one colleague pointed out, even economists cooperate. Social capital’s addition of new factors has systematically chiseled away at some of the neoclassical economic assumptions about cooperation. We begin our exploration of social capital’s contributions with grounding conditions for cooperation: that it must rest on some underlying trust in behavior, where trust means a belief about another’s intrinsic motivation, indeed about their trustworthiness. Two aspects of social capital – the networks and norms in which individuals are embedded – reinforce cooperative behavior by changing the

Introduction

5

pay-oﬀs for certain actions. They also challenge the neoclassical assumption of atomistic humans. Not many of us have lived the life of Robinson Crusoe, alone on his island for 17 years before encountering his new friend and companion, Friday. Community networks, norms and ongoing relationships directly aﬀect how individuals make decisions about cooperation. Networks particularly can reinforce trustworthy behavior. Burt (2000) argues that dense horizontal relationships, which he calls closure networks, transmit information, encourage surveillance and enforce sanctions on deviant behavior. However, one could be 100 percent selﬁsh and 100 percent rational, as neoclassical homo economicus may be, and still cooperate in this context of strong incentives for appropriate behavior associated with closed social networks. Norms provide further reasons for members of a community to interact in a trustworthy manner. These rules prescribe acceptable behavior and provide the mechanism to punish or reward unacceptable behavior. James Coleman (1990) argues that when a norm or rule exists, members of a group willingly transfer some authority over their behavior to their fellow members. Putnam (1993) argues that the degree to which a community adopts the norm of reciprocity indicates the proportion of its members who act in a trustworthy manner. Once again, the selﬁsh and rational homo economicus would act in a trustworthy manner when faced with strong, well-enforced norms. Signiﬁcant ongoing research has provided empirical evidence to moderate neoclassical assumptions of universal selﬁshness. Some individuals have been shown to be as concerned with another’s welfare as their own. The University of Michigan World Values Survey, which asks: ‘Do you believe that most people can be trusted or that you cannot be too careful in dealing with people?’ consistently shows wide variation in responses. This information supports the idea that diﬀerent people, communities and societies display diﬀerent levels of trustworthiness. Social capital theories thus seem to moderate further rational choice assumptions on human behavior. The incentives set up by networks of relationships and norms of reciprocity would encourage members of a collective to act in a trustworthy manner.

WHAT ARE THE CONTROVERSIES SURROUNDING SOCIAL CAPITAL? Perhaps it is not surprising that social capital still faces vehement criticisms. Social capital research has introduced concepts such as trustworthiness, trust, networks and norms into the ﬁeld of economics, and into the broader

6

Social capital

social sciences. These concepts are diﬃcult to deﬁne and even more diﬃcult to quantify, thus going against the tradition of the social sciences, and in particular economics, which have long prided themselves on their rigor, analysis and quantiﬁable measurement. It is time we turn our attention to these criticisms. The major direct criticisms of social capital center on whether it is capital at all; that it lacks the required building blocks to be considered capital. These issues include the inability to invest in social capital at time t and gain the beneﬁts in t+1, the inability to alienate or, in other words, ‘sell’ social capital, and the problem of its measurement. Scholars continue to debate the appropriate conceptualization of social capital and to disagree on its impact and the mechanisms that can create it. They also continue to debate the intricate relationship between social capital and trust. The scholarship collected here provides new and pioneering insights into each of these fundamental questions. In Part I, Elinor Ostrom, Ronald Burt and Robert Fishman provide overviews of social capital research. They assess its current contributions and challenges in order to provide a ‘State of the Union’ perspective. In so doing, each advocates a starkly diﬀerent point of view about social capital’s conceptualization. Their debate is between an expansionist view and a minimalist one (Ostrom and Ahn, 2003). In Chapter 1, Ostrom argues for a more expansive view of social capital that emphasizes the importance not only of networks, but also the institutions or norms that they foster. Carefully placing social capital in the context of all human-made capital, she contends that scholars must study the diverse forms of social capital if they want to understand its underlying design principles and use these to create tailored forms of social capital that can address speciﬁc collective action problems. By contrast, Ron Burt advocates a more minimalist approach. He contends that limiting the inquiry to the structure and characteristics of networks is the way that scholars can begin to understand the mechanisms by which social capital aﬀects outcomes. Burt’s closely argued theory (2005) contends that social capital beneﬁts lie in managing between brokerage and closure relationships. Building new connections among individuals and groups where none exist, brokerage concerns the advantages associated with greater exposure to diﬀerent ideas, practices and opinions. Closure is about the beneﬁts of eﬃciency – the productivity associated with all members of a community knowing one another. Robert Fishman takes an even more focused approach by arguing that the substantial costs of conceptualizing social ties and associated norms as social capital far outweigh the beneﬁts. The complexity of disparate sources, eﬀects and levels of analysis prevents such an approach. Rather, he suggests that scholars should analyze types of connections and their

Introduction

7

associated types of eﬀects. Although few scholars think of social capital as undiﬀerentiated stock or infer its existence through overarching eﬀects, Fishman’s caution is a wise one.

HOW DOES SOCIAL CAPITAL INTERACT WITH OTHER DISCIPLINES? Social capital as a concept both ‘reaches out’and ‘reaches in’. Part II investigates how social capital ‘reaches out’ to other disciplines to help understand their dynamics better. John Helliwell, Haifang Huang, Robert Putnam, Maureen Hallinan, David Campbell, David Nickerson and Janine Nahapiet investigate how social capital impacts upon workplace productivity, economic development, education, governance and competitive advantage. Part III, in contrast, traces how other disciplines such as psychology, network analysis, and normative and trust research can ‘reach in’ and can help elucidate the building blocks on which social capital rests. Roderick Kramer, Daniel Brass, Roy Lewicki, Chad Brinsﬁeld, Darryl Stickel, Roger Mayer, Sim Sitkin, James Davis and Viva Bartkus investigate social capital’s formation by studying its structural, relational and cognitive antecedents. In Part II of this volume, authors re-examine some of the long-standing beliefs that underpin economics, sociology, political science and business through the lens provided by a social capital perspective. In Chapter 4, John Helliwell, Haifang Huang and Robert Putnam employ a social capital perspective to question orthodoxies in the labor market. They argue that wage diﬀerentials may be based more on workplace trust than on the arduousness and hazard of the job – an explanation ﬁrst advocated by Adam Smith. The authors demonstrate that a climate of trust in the workplace – that is, social capital – enhances the subjective well-being of employees. Indeed, their research indicates that moving one point on a ten-point scale of workplace trust has the same eﬀect on life satisfaction as a 40 percent change in income. While social capital in the workplace is diﬃcult to build and relatively easy to destroy, the size of this compensating diﬀerential raises the question of why companies are not doing more to increase trust. Perhaps some of the yet undiscovered value of social capital lies in the possibility that building workplace trust can help managers and employees make tremendous gains in productivity and satisfaction. In some of the earliest scholarship regarding social capital, James Coleman (1988) argued that the superior academic results of Catholic school students were generated by the intergenerational closure and strong bonds among the Catholic community surrounding the school. After being

8

Social capital

normalized for socio-economic class, race, age and other factors, this superior academic achievement became known as the ‘Catholic school advantage’. In the ﬁrst ever replication of Coleman’s massive study, Maureen Hallinan retraces Coleman’s steps by conducting a longitudinal study of Catholic and public (that is, state-funded) school children in Chicago. Her ﬁndings in Chapter 5 question Coleman’s results in terms of academic achievement, but indicate that strong social capital networks do impact upon non-academic factors such as liking school, feelings of safety, trust in the neighborhood, participation in clubs and helping others in the community. In a comparable assault on orthodoxy – this time in political science – David Campbell proposes an explanation for voting behavior. He does not use the reigning paradigm of homo politicus – that people vote to defend their interests when elections are close; rather, he poses the theory of homo civicus – that people vote because they believe it is their civic duty. In Chapter 6, Campbell argues that voting is a civic-motivated activity and that social norms facilitating civic-mindedness are transmitted through the social closure network fostered in school. As a consequence, students demonstrate these learned norms – such as voting – for years to come. Although Hallinan and Campbell come from diﬀerent disciplines, they broadly agree on two major insights: ﬁrst, that social capital requires both networks and norms; and second, that schools are causal agents that can lead to more public engagement, collective identiﬁcation and, therefore, the future creation of social capital. They also concur that only those networks in which parents and educators enforce strong norms of either scholarship or civic-mindedness can be expected to increase academic achievement and civic-oriented behavior. Most social capital theory rests on the assertion that networks of dense relationships create and reinforce norms. However, a ‘chicken vs egg’ issue exists when examining communities: what portion of norm strength comes from self-selection into a network versus transmission of norms through the network? We are drawn to those with whom we share similarities of background, experience or values. Yet we also take on the values of those with whom we interact. In Chapter 7, David Nickerson uses a series of controlled ﬁeld experiments to address these critical questions. In Chapter 8, Janine Nahapiet extends her previous scholarship regarding the ways in which a ﬁrm’s social capital creates intellectual capital, and ultimately competitive advantage. She contends that a series of developments at the close of the twentieth century – including increasing connectedness due to technology change and deregulation, a shift from tangible to intangible assets and in particular the increased importance of human capital, knowledge management and relationships – has resulted in shifting

Introduction

9

organizational boundaries and even entirely new ways of organizing economic activity. Consequently, driving competitive advantage through innovation depends even more centrally on the appropriable resources latent in the relationships that social capital represents.

HOW CAN WE CREATE AND MAINTAIN SOCIAL CAPITAL? If many scholars of multiple disciplines argue for its tremendous impact, as they do in Part II, why then is there not more social capital? In fact, why do leading experts such as Putnam (2000) argue that – at least in the US – social capital is low and falling? Part III explores the actual mechanisms – cognitive, structural and relational – by which we foster and maintain social capital. Each author develops insights into the puzzling question of how we can create social capital. In Chapter 9, Roderick Kramer argues that cognition – to identify with others in a group – is a critical initial step to building social capital. He advances an integrated framework that shows how individuals’ social identities influence their willingness to collaborate, to help each other and to contribute to the collective’s reservoir of social capital in general. Their identiﬁcation with the entire collective – rather than personal or subgroup identiﬁcation – facilitates this behavior. He then describes in detail how such psychological transformations of identity can occur. By contrast, in Chapter 10 Daniel Brass argues that the strategy behind creating social capital should be more structural – focused on the structure and type of one’s relationships – rather than cognitive. He analyzes the structural conditions under which it may be to the individual’s advantage to connect to brokers, who usually stand to gain more from the relationship than the individual. However, by creating relationships through strategic actions, individuals can beneﬁt from the resulting social structure and create social capital. Trust among members of a collective is critical to any discussions of cooperation, and a collective’s resulting social capital. To address the many diﬃculties with trust’s conceptualization, Mayer et al. (1995) develop a parsimonious model of trust – as a phenomenon between two people – and its causes and outcomes. They argue that the antecedents to trust are the trusting party’s perception of the other party’s ability to do what they say they can do, of their integrity, and of their benevolence toward the trusting party. Yet the question remains: how to scale dyadic trust to the level of the entire collective, especially in circumstances where one neither knows all other members nor can monitor and perhaps sanction their actions?

10

Social capital

In Chapter 11, Roy Lewicki and Chad Brinsﬁeld briefly proﬁle recent theoretical and empirical research regarding trust, more speciﬁcally characterizing the types of trust and distinguishing between trust and distrust. Because trust and distrust are diﬀerent from each other, Lewicki and Brinsﬁeld argue that both can coexist within relationships. They then draw the implications of building trust and managing distrust for creating and sustaining social capital within a community. Darryl Stickel, Roger Mayer and Sim Sitkin investigate the speciﬁc decision as to why an individual would invest in social capital. In Chapter 12, the authors argue that individuals must trust in the collective prior to making such investments, and thus they explore in detail how individuals make the decision to trust others. They then examine the social changes which may be leading both to declines in trust speciﬁcally and to increases in the perception of the risks associated with such social investments more broadly, as a way to begin to understand the general decline in social capital, at least in the United States. Regardless of the challenges trust adds to the social capital equation, its importance is without doubt. Kenneth Arrow called our attention to it by pointing out (1972, p. 357): ‘Virtually every commercial transaction has within itself an element of trust, certainly any transaction conducted over a period of time.’ Lewicki, Brinsﬁeld, Stickel, Mayer and Sitkin add important insights into the complicated relationship between trust and social capital.

HOW DO WE MEASURE SOCIAL CAPITAL? Among the many factors critical to the widespread adoption of any new concept is its measurability. Comprehensive deﬁnitions of social capital are extremely problematic in this regard. They are multidimensional with varying levels and units of analysis. In addition, their forms change over time. Scholars have therefore measured social capital both through direct measurement and by capturing the positive or negative externalities or consequences that social capital causes. At the macro level, Putnam (2000) and the other scholars have used composite indices to attempt direct measurement. For Putnam, social capital is the aggregate of associations in a society and their membership. By contrast, Partha Dasgupta and Ismail Serageldin (2000) argue that the fact that no ‘natural system of units’ of social capital exists ‘does not matter in principle, because in any given context, you can measure their worth by the extent to which mutual beneﬁts can be realized’. At the macro level, Dasgupta argues that social capital can be measured from its generated

Introduction

11

externalities, more speciﬁcally in total factor productivity. Proxies for social capital using trust and civic engagement have also been attempted. At the micro level, World Bank scholars attempted to create an index of social capital at the family and community levels based on the density and characteristics of formal and informal networks. Entering into this varied landscape of proposals for social capital measurement, James Davis and Viva Bartkus conduct empirical tests of their theoretical model of social capital. The authors argue that organizational trust creates the environment that community members need if they are to engage in activities together. This atmosphere mitigates the risks that individuals take in collaborating. They further argue that a community’s level of organizational trust, and thus, social capital, depends directly on the strength of its networks, the depth and breadth of its norms, and the abilities of its members and the group itself. This approach enables them to use social capital’s antecedents to predict its presence rather than inferring its existence from a composite index of factors or an assessment of its associated positive or negative externalities. By combining the theoretical model with its empirical test on a sample of students and study teams in a business college, Davis and Bartkus indicate ways that social capital can be better understood and measured. In closing, the time has come to pause for theoretical reflection and clear empirical testing of the integrity of the social capital concept. The fact that social capital eﬀects have occurred in areas as disparate as democracy and governance, economic development, education and labor economics suggests that further insights will likely arise by building bridges and conversations among researchers across the social sciences. It is with these goals in mind, as well as a willingness and commitment to learn each others’ languages, that leading scholars on social capital converged on the University of Notre Dame on 1–3 April 2006 to foster just such a dialogue. They were determined to overcome the increasing degree of specialization within academic disciplines that can preclude a natural and free-flowing dialogue. This book is a product of those discussions. It showcases new innovative research, but also enables leading scholars to continue their dialogue on the themes outlined here. Our goal is to generate further debate regarding social capital; if we achieve this, then we will consider the book a success.

12

Social capital

REFERENCES Adler, P.S. and S.W. Kwon (2002), ‘Social capital: prospects for a new concept’, Academy of Management Review, 27 (1), 17–40. Arrow, Kenneth (1972), ‘Gifts and exchanges’, Philosophy and Public Aﬀairs, 1, (Summer) 343–62. Arrow, Kenneth (2000), ‘Observations on social capital’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 3–5. Becker, Gary (1993), Human Capital, 3rd edn, Chicago, IL: University of Chicago Press. Bourdieu, P. (1985), ‘The forms of capital’, in J.G. Richardson (ed.), Handbook of Theory and Research for the Sociology of Education, New York: Greenwood, pp. 241–58. Bourdieu, P. and L.J.D. Wacquant (1992), An Invitation To Reflexive Sociology, Chicago, IL: University of Chicago Press. Burt, Ronald S. (1992), ‘The social structure of competition,’ in Structural Holes: The Social Structure of Competition, Cambridge, MA and London, UK: Harvard University Press, pp. 8–30. Burt, Ronald S. (1997), ‘The contingent value of social capital’, Administrative Science Quarterly, 42, 339–65. Burt, Ronald S. (2000), ‘Network structure’, Research in Organizational Behavior, 22, 345–423. Burt, Ronald S. (2002), ‘Bridge decay’, Social Networks, 333–63. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Capital, Oxford: Oxford University Press, pp. 333–63. Burt, Ronald S. and Marc Knez (1996), ‘Trust and third-party gossip’, in R.M. Kramer and T.R. Tyler (eds), Trust in Organizations: Frontiers of Theory and Research, Thousand Oaks, CA: Sage, pp. 255–92. Coleman, James (1987), ‘Norms as social capital’, in Gerard Radnitzky and Peter Bernholz (eds), Economic Imperialism: The Economic Approach Applied Outside the Field of Economics, New York: Paragon House Publishers, pp. 133–55. Coleman, James (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94 (Supplement: Organizations and Institutions: Sociological and Economic Approaches to the Analysis of Social Structure), S95–S120. Coleman, James (1990), Equality and Achievement in Education, Boulder, Co: Westview. Coleman, James (1994), Foundations of Social Theory, Cambridge, MA: Belknap Press. Coleman, James S. (2000), ‘Analytical foundations: social capital in the creation of human capital’, in P. Dasgupta and I. Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 13–39. Collier, Paul and Jan Willem Gunning (1999), ‘Explaining African economic performance’, Journal of Economic Literature, 37 (1), 64–111. Dasgupta, Partha (2000a), ‘Social capital and economic performance analytics’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 1–31. Dasgupta, Partha (2000b), ‘Economic progress and the idea of social capital’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 325–424.

Introduction

13

Dasgupta, Partha and Ismail Serageldin (eds) (2000), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank. Davis, J.H., D. Schoorman, R.C. Mayer and T.H. Hoon (2000), ‘The trusted general manager and business unit performance: empirical evidence of a competitive advantage’, Strategic Management Journal, 21 (5), 563–76. de Tocqueville, Alexis [1840] (1945), ‘Of the use which the Americans make of public associations in civil life’, ‘Of the relation between public association and the newspapers’, ‘Relation of civil to political association’ and ‘How the Americans combat individualism by the principle of self-interest rightly understood’, in Phillips Bradley (ed.), Democracy in America, Volume 2. Hanifan, Lyda Judgson (1916), ‘The rural school community center’, Annals of the American Academy of Political and Social Science, 67, 130–38. Hanifan, Lyda Judgson (1920), The Community Center, Boston, MA: Silver Burdett & Company. Helliwell, John H. and Robert Putnam (1995), ‘Economic growth and social capital in Italy’, reprinted in P. Dasgupta and I. Serageldin (eds) (2000), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 253–68. Knack, Stephen and Philip Keefer (1997), ‘Does social capital have an economic payoﬀ? A cross-country investigation’, Quarterly Journal of Economics, 112 (4), 1251–88. Loury, Glenn (1992), ‘The economics of discrimination: getting to the core of the problem’, Harvard Journal for African-American Public Policy, 1, 91–110. Mayer, Roger C. and J.H. Davis (1999), ‘The eﬀect of the performance appraisal system on trust for management: a ﬁeld quasi-experiment’, Journal of Applied Psychology, 84, 123–36. Mayer, Roger C., J.H. Davis and F.D. Schoorman (1995), ‘An integrative model of organizational trust’, Academy of Management Review, 20 (3), 709–34. McNally, David (1994), Political Economy and the Rise of Capitalism: A Reinterpretation, Berkeley, CA: University of California Press. Nahapiet, Janine and Sumantra Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23 (2), 242–66. Narayan, D. (1999), ‘Bonds and bridges: social capital and poverty’, World Bank papers. Narayan, Deepa and Lant Pritchett (2000), ‘Social capital: evidence and implications’, in P. Dasgupta and I. Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 269–95. Ostrom, Elinor (1990), Governing the Commons: The Evolution of Institutions for Collective Action, Cambridge: Cambridge University Press. Ostrom, Elinor (1992), Crafting Institutions for Self-Governing Irrigation Systems, San Francisco, CA: ICS Press, pp. 19–39. Ostrom, Elinor (1998), ‘A behavioral approach to the rational choice theory of collective action’, American Political Science Review, 92 (1), 1–22. Ostrom, Elinor (2000), ‘Social capital: a fad or a fundamental concept?’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 172–214. Ostrom, Elinor and T.K. Ahn (2003), Foundations of Social Capital, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Ostrom, Elinor and James Walker (2002), Trust and Reciprocity: Interdisciplinary Lessons for Experimental Research, New York: Russell Sage Foundation.

14

Social capital

Putnam, Robert D. (1993), ‘The prosperous community: social capital and public life’, American Prospect, 13, 35–42. Putnam, Robert D. (1995a), ‘Bowling alone: America’s declining social capital’, Journal of Democracy, 6 (1), 65–78. Putnam, Robert D. (1995b), ‘Tuning in, tuning out: the strange disappearance of social capital in America’, PS: Political Science and Politics, 38 (4), 664–83. Putnam, Robert D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Putnam, Robert D. (ed.) (2002), Democracies in Flux: The Evolution of Social Capital in Contemporary Society, New York: Oxford University Press. Putnam, Robert D., Robert Leonardi and Raﬀaella Y. Nanetti (1993), ‘Social capital and institutional success’, in Making Democracy Work: Civic Traditions in Modern Italy, Princeton, NJ: Princeton University Press. Woolcock, Michael (1998), ‘Social capital and economic development: toward a theoretical synthesis and policy framework’, Theory and Society, 27 (2), 151–208. Woolcock, Michael (2001), ‘The place of social capital in understanding social and economic outcomes’, Canadian Journal of Policy Research/Revue canadienne de recherche sur les politiques. Government of Canada’s Policy Research Secretariat, 2 (1), 1–8. Woolcock, Michael and Deepa Narayan (2000), ‘Social capital: implications for development theory, research, and policy’, World Bank Research Observer, 15 (2), 225–49.

PART I

Overview

1.

What is social capital? Elinor Ostrom

In this chapter, I will argue that the concept of social capital is extremely important as well as useful for all of the social sciences. After reviewing the current situation, I will lay out a careful set of deﬁnitions of types and forms of human-made capital and clarify the individual and shared characteristics of these concepts. Next, I will illustrate the importance of social capital by comparing the performance of irrigation systems in Nepal with high levels of physical capital and low social capital with those that have low levels of physical capital but high social capital. Finally, I will conclude with reflections on some of the broad lessons learned about social capital. Social capital research has grown steadily since James S. Coleman’s systematic analysis in 1988 (see Bourdieu, 1986; Loury, 1977 for earlier uses of the concept). In the early 1990s, researchers started employing the concept as a major building block in their scholarship (Burt, 1992; Ostrom, 1992). In 1993, Robert Putnam et al.’s celebrated book, Making Democracy Work, catapulted social capital research into a widespread and lively phase of development (see Knack and Keefer, 1997; Dasgupta, 2003; Fukuyama, 2000; Lake and Huckfeldt, 1998; Ostrom and Ahn, 2003b). Despite this attention, however, it remains unclear whether social capital will be useful for the social sciences over the long term. Academic research can be aﬄicted by fads and fashions, just like other ﬁelds of endeavor. Answering this question requires a clear, precise deﬁnition as well as well-designed, robust research protocols. A multitude of deﬁnitions currently exist. While some are vague or narrow, many are based on reasonable intuition and social theory. Almost all reflect two basic assumptions: social capital is a resource that is available to members of a social network, and social structure is often the type of capital that all members of a group can access to promote their interests. As Putnam (1993) aptly put it, social capital is never ‘the private property of any of the persons who beneﬁt from it’ (p. 170). While a single deﬁnition is not required for social capital to be meaningful over the long term, whatever ones are adopted need to be clear, precise and robust. A quick review of some of the commonly accepted deﬁnitions can provide some perspective on the challenge we face in achieving these goals. 17

18

Overview

The abundant criticism of social capital research – and its subsequent rejection by some scholars – illustrates the importance of our task. Arrow (1999, p. 4) is a good example of such criticism. He argued that the concept of social capital must meet the three criteria associated with capital in general: extension in time, deliberate sacriﬁce in the present for future beneﬁts, and alienability. Most deﬁnitions of social capital do not meet the second and third criteria, and Arrow therefore rejected the validity of social capital on these grounds. After a comprehensive review of the current deﬁnitions, I will present the deﬁnition of social capital that will be used throughout the rest of this chapter.

CURRENT DEFINITIONS OF SOCIAL CAPITAL Many of the past and current deﬁnitions center on social networks. In 1965, Jacobs deﬁned a community’s social capital in terms of the networks of ‘strong, cross-cutting personal relationships that develop over time’ and that create ‘a basis for trust, cooperation, and collective action’ among community members (cited in Nahapiet and Ghoshal, 1998, p. 243). Another deﬁnition (Bourdieu and Wacquant, 1992) treats social capital as an aggregate of actual or potential resources linked to possession of a stable network of relationships. Putnam (2000) uses the term to refer to the characteristics of a social organization, such as trust, norms and networks, that facilitate actions that can increase societal eﬃcacy. As he says: The core idea of social capital theory is that networks have value . . . social contacts aﬀect the productivity of individuals and groups. Whereas physical capital refers to physical objects and human capital refers to properties of individuals, social capital refers to connections among individuals – social networks and the norms of reciprocity and trustworthiness that arise from them. (p. 19)

Further, Lin (2001) discussed social capital as an investment in social relations with expected returns. Coleman’s (1990) deﬁnition, in contrast, emphasized the achievement side of social capital and was functional in nature. He stated that social capital is a resource easily accessible by the collective that makes possible ‘the achievement of certain ends that would not be attainable in its absence’ (p. 302). It is the product of social structure that encompasses more elements than simply network conﬁgurations. Other, broader deﬁnitions also exist. Burt (2000) treats social capital as a metaphor for the fact that wellconnected people do better in life. Portes (1998) sees social capital as the positive consequences of social relations.

What is social capital?

19

Outstanding Issues with Current Research Questions of social capital’s future longevity and usefulness as a concept arise from how the concept has been used and abused in the literature (Solow, 1999; Arrow, 1999; Fine, 2001; Durlauf, 2002). Solow (1999) noted that, despite the importance and relevance of the underlying ideas, most current social capital research is plagued by ‘vague ideas’ and ‘casual empiricism’. He and other critics claim that current research allows almost any aspect of social structure to be deﬁned as potentially yielding social capital. In response to these issues, many researchers focus on social networks rather than social structure. This narrow approach has allowed scientists not only to predict the eﬀects of social capital on network members but also to identify the mechanisms that transmit these eﬀects. For this reason, Lin (2001), Burt (2000) and others suggest that social capital theory should apply only to social networks. Moreover, economists also have substantial problems with social capital. The ﬁrst issue relates to their inability to specify additive units of social capital that can be used to explain increases in production. A basic production function shows how a flow of products is created by a combination of land, labor and capital. If we add more of one factor to ﬁxed amounts of the other two, we will generate more of the ﬁnal product – although at a diminishing marginal rate of return. Their second struggle relates to the familiar and vexing dilemma that relates to all collective action. How do you explain people’s willingness to invest time and eﬀort to build social capital when all of the members of a network can reap beneﬁts irrespective of their individual contribution to that capital’s production and maintenance? Even if all individuals in the network recognize the value of the capital, they can still ‘free ride’ on the eﬀorts of others (Hardin, 1968; Olson, 1965). If social capital is to be a valuable concept, then its theory must be able to identify the conditions under which members are willing to help create and sustain the reservoir of social capital (Ahn and Ostrom, forthcoming). Current thinking states that two elements influence the form such behaviors or commitments take: (1) the nature of the network or social connections between the members; and (2) the particular form of the pertinent behaviors (for example, Brief and Motowidlo, 1986; Uzzi, 1997). Researchers categorize the types of social capital into two types of embeddedness: structural and relational (Nahapiet and Ghoshal, 1998). First, structural embeddedness demonstrates the ‘impersonal conﬁguration of linkages among people or units’ (Nahapiet and Ghoshal, 1998, p. 244). An example is the density of network ties linking interdependent actors within

20

Overview

a formal system (Burt, 1992, 1997, 2002). Second, relational embeddedness reflects the ‘kinds of personal relationships people have developed over time with each other through a history of interaction’ (Nahapiet and Ghoshal, 1998, p. 244). The intimate and repeated exchanges observed within fraternal organizations and business networks exemplify this latter form (for example, Boissevain, 1974; McEvily and Zaheer, 2005; Uzzi, 1997). Behaviors will vary greatly based on situations and the type of network involved. For instance, the more seasoned and experienced members in an informal network of amateur mushroom collectors must be willing to create and sustain an ongoing, voluntary social structure that educates novice members in safe mushroom detection and preparation (Fine and Holyﬁeld, 1996). In a network of taxi drivers working at night in dangerous areas, social capital depends upon members’ willingness to maintain an informal communication system alerting each other to problem areas (drunks, criminal types, and so on) as well as roadside risks (Gambetta and Hamill, 2005). In a more formal setting, such as a society or nation-state, individuals must be willing to contribute to the creation of what are often more formal networks that produce and help sustain various forms of public goods (for example, Bonacich and Schneider, 1992; Messick and Brewer, 1983; Dasgupta and Serageldin, 1999; Murnighan et al., 1994; Pew Research Center, 1996; Putnam, 1993, 2000; Olson, 1965). The time has come when we need to revisit social capital’s foundations – not just limit our deﬁnitions to avoid some of the issues mentioned above. What is the core meaning of human-made capital, of which social capital is one part? To make social capital a lasting concept, we must arrive at a theoretically meaningful and accepted deﬁnition for it. The rest of this chapter deals with this issue: ﬁrst, by deﬁning human-made capital; then by discussing some of the similarities and diﬀerences between social capital and other forms of human-made capital; and ﬁnally, by illustrating some of social capital’s speciﬁc characteristics through a case study.

WHAT IS THE MEANING OF HUMAN-MADE CAPITAL? The ﬁrst step in building a useful deﬁnition of social capital is to distinguish between ‘natural capital’ and ‘human-made capital’, which includes social capital. Capital, in its most basic sense, is a set of assets capable of generating future beneﬁts for at least some individuals (Lachmann, 1978). On the one hand, natural capital is all of the resources of the earth that humans did not build. The oceans, the atmosphere and biodiversity are all forms of natural capital that humans use – and frequently overuse.

What is social capital?

21

Human-made capital, on the other hand, is created by humans. Humans build it by spending time and eﬀort to transform something into something else that they hope will increase beneﬁts later on. Humanmade capital can be divided into three types: physical capital, human capital and social capital. Physical capital includes investments in physical infrastructure and tools – for example, roads, factories, computers – that the builder hopes will increase the productivity of future activities. Human capital represents individual or group skills and knowledge that can potentially increase the flow of future beneﬁts. Social capital is a set of relationships among members of a group and the values that they share that enable them to solve collective problems in the present and future. Types of Human-Made Capital Physical capital Of the three types of human-made capital, physical is the easiest type to understand and measure. Builders usually create it by deferring consumption and investing in physical infrastructure and tools that will enhance future production. This capital is often assessed by its monetary value, the simplest option being that one assigns the monetary value that the owner would receive if the capital were sold. However, putting a value on physical capital can involve the same type of problems that occur when valuing other forms of capital. Whether any form of human-made capital can be bought or sold, and at what value, depends on more than its type. Public highways are a very important form of physical capital, but assigning a monetary value to a highway system demands many calculations beyond the highway’s initial cost. Human capital When scholars ﬁrst introduced the concept of human capital (Schultz, 1961) it took time for this concept and ways of measuring it to be accepted. Now, human capital is recognized as a major factor in economic productivity (Becker, 1964; Hardin, 1999). In regard to human capital, an individual defers current consumption and pleasurable activities to acquire better skills and knowledge that can potentially increase future beneﬁts. When young people go to college, they invest valuable time and a great deal of their own or their parents’ money. They also defer income. In another example, an athlete builds human capital by eating healthy foods and working out, and so develops endurance, muscular strength and good health. A musician spends time and eﬀort practicing to enhance their skills on an instrument.

22

Overview

Social capital Social capital is a set of relationships and shared values created and used by multiple individuals to solve collective problems in the present and future. It reflects how small to large groups interact culturally, normatively, structurally and institutionally. It also describes the eﬀects these interactions have on individual incentives and behavior and the resulting economic, political and other possible changes (Pasotti and Rothstein, 2002). Social capital is relevant whenever individuals’ joint endeavors are critical to achieve a collective goal. However, it may not only create opportunities but also constrain those who are involved in collective problem-solving (Birner and Wittmer, 2003). A friendship network is a good example of social capital. While one may build this network just to gain a capital asset, real friends spend substantial time together. They gain a reputation for being friendly and reliable, for reciprocating favors, and for being ‘good buddies’ – someone who will help a friend when they would prefer to be doing something else. All of these behaviors are driven by the obligation that comes with being a friend. Building an eﬀective friendship network requires deferral of time and eﬀort from individual activities; even if not done consciously, this activity counts as creating social capital. An individual may then draw upon the resources inherent in their friendship network to gain assistance in diﬃcult times or to gain access to new opportunities. Social capital can grow with appropriate use – unlike physical capital, which wears out and requires continuous maintenance. On the one hand, families who help one another strengthen their network as they use it. On the other hand, if families begin to take advantage of past trust and not reciprocate favors, social capital may disintegrate very rapidly. Once destroyed or weakened, social capital can be far more diﬃcult to rebuild than physical capital. Shared Characteristics of Human-Made Capital While we have highlighted a few of the diﬀerences between the various types of human-made capital, they share many more common traits. All types of human-made capital are either capital assets or flow assets, also known as consumption goods. A capital asset is used to generate other assets in the future. A flow asset, on the other hand, is consumed or used up in the short run. Most of us have consumption goods in our refrigerators or pantries. We will either eat these at a future meal or combine them with other assets to generate a new product (for example, bread, a pasta dish, salad). When wheat is used to make bread, both the wheat and the bread are flow assets because they are consumed in the short run. On the

What is social capital?

23

other hand, when a farmer builds or buys a new hoe, he is creating a capital asset. The hoe will not be consumed immediately but will be used in weeding and other tasks that will increase the probability of more horticulture products in the future. However, no type of capital is guaranteed to produce future beneﬁts. Capital may remain ‘dead’ or ‘unused’ after great eﬀorts have been expended to build it. When a factory shuts down without a new owner transforming the building and equipment into another use and hiring workers to make new products, the factory becomes ‘dead capital’. Knowledge and further resources – at least time and eﬀort – must be expended to make this capital asset produce beneﬁts. Capital may also have negative – rather than positive – eﬀects. It can constrain future events or even be used for purposes that produce harm rather than beneﬁts. A new highway, for example, enables people to travel to new locations but also constrains the type of vehicle that can be used, the speed of traveling and where the highway can be crossed. A weapons factory is a tragic example of physical capital that will protect some individuals while harming others. Education can make scientiﬁc breakthroughs or write computer viruses. Social networks can create a safe and friendly neighborhood or violent gangs. This ‘dark side’ of social capital can also involve police trusting each other not to report excessive force (Langbein and Jorstad, 2002), corporations or nations colluding with one another to create cartels (Hoﬀman and Libecap, 1995), or members of the Maﬁa undertaking illegal economic activities (Gambetta, 1988; see Ostrom, 1999). While we talk about human-made capital as something made by humans, such capital can also be created as a by-product of other activities. A path through the forest provides an excellent example. Deer start a path; humans then discover and follow the deer path. After enough humans have used it, the path may become a way for humans to get to work, gather products or undertake other tasks more rapidly. The path is now a form of shared capital for those who need to get through the forest. It may have been built without self-conscious eﬀort. But to keep the path working over time, users will need to invest eﬀort in maintenance such as removing fallen trees, weeding and ﬁlling in holes. Other examples abound. A farmer who makes his own farm implements learns skills (human capital) that he can use again to make more farm implements (physical capital). A member of a swimming team creates human capital by building her muscular strength but also gains social capital by learning about her team members, including whom she can trust to engage fairly. With physical capital, the forms all have diﬀerent characteristics. Common forms include: buildings, roads, factories, schools,

24

Overview

irrigation systems and portable tools. A multitude of ‘sub-forms’ can also exist within a common form. The number of diﬀerent hammers designed for very special use as tools, for example, is extremely large. It is reported that over 500 diﬀerent hammers were in active production toward the latter part of the nineteenth century in Birmingham, England, alone (Petroski, 2006, p. 2). Human capital has many diﬀerent forms of individual knowledge and skill. Social capital also has multiple forms, including belonging to a club, a workplace or an alumni association, or having a network of friends. However capital is created, it is important to remember that it always involves multiple forms. As a result, the same capital asset that is productive in one setting may be counterproductive (or worse) in another setting where it is not well ﬁtted for the problem that needs to be solved. Consequently, functional deﬁnitions of capital in general, and social capital in particular, are not as useful in the social sciences (Durlauf, 2002).

HOW DO YOU MEASURE HUMAN-MADE CAPITAL? Measuring the value of capital can be challenging. Frequently, multiple variables can aﬀect the ultimate estimate of value. While the method of assessment matters, it is not as important as the speciﬁc form of capital being assessed, the particular characteristics of the local situation, or the proﬁciency of the human agent involved. Using monetary values is not only challenging but relies heavily on assumptions that may, or may not, be valid. Because it seems ‘easier’, scholars sometimes tend to use monetary values (for example, for some physical and human capital like a factory or college degree). However, doing so requires substantial knowledge about the date of acquisition, the initial investment costs, the speciﬁc sector, the amount of maintenance invested over time, and the future demands for or returns on this particular type of capital. Estimating the industrial capital or overall educational achievement of a nation requires even further assumptions and is even more diﬃcult. The meaningfulness of the calculated value will depend heavily on the way the question is asked, the detailed information available, the accounting formulae employed and the way the resulting answer is used. The value of all human-made capital depends on the situation in which it exists or is employed. Thus, the speciﬁc processes or problems being considered – not the method of assessment – will determine the value. For instance, a hammer designed for one kind of nail and one kind of surface may fail to perform well when used for a diﬀerent nail or surface. The value of the capital also depends on the skill or proﬁciency of the human agency.

What is social capital?

25

Without a skilled carpenter, a hammer may inadvertently destroy a woodworking project rather than help build it. In a rigorous study of 60 villages in Rajasthan, India, Krishna (2001) illustrates the importance of tailoring the measurement approach to the situation when assessing social capital’s signiﬁcance. He evaluated social capital’s role in improving the performance of local economic development initiatives and, more speciﬁcally, with respect to four watershed development and poverty reduction projects in the region. To do so, he conducted almost 2000 individual interviews in the 60 villages. To develop his measures of social capital, Krishna carefully selected indicators that ﬁt the local setting. He rejected an indicator often used in the literature – the density of formal organizations. Putnam (1993, 2000) used this to measure social capital in Italy as well as the United States. Krishna points out: Density of formal organizations is a particularly inappropriate indicator for Rajasthan villages. Hardly any formal organizations have been set up voluntarily by villages in Rajasthan, and nearly every formal organization in these villages is a companion to a state agency and executor of its program. (2001, p. 6)

Krishna draws instead on a scale that focuses primarily on informal networks (Krishna and Uphoﬀ, 1999). To create this scale, he questioned a sample of respondents about: their membership in labor-sharing groups; how they dealt with crop disease and natural disasters; the trust they would feel if they owned a farm with another family; the solidarity between their village leaders and regular residents; and the levels of reciprocity that were demonstrated in child rearing. Krishna also measured various forms of human agency. He particularly focused on the presence or absence of leaders who might be able to employ social capital to achieve higher performance in development programs. Krishna ultimately also created an index of economic development performance that includes overall poverty reduction as the performance metric. To assess social capital’s role in achieving reduced poverty levels, Krishna conducted three basic regression analyses. In the ﬁrst, he examined how various types of human agency were associated with his index of development performance. None of the agency variables turned out to be signiﬁcant, and the overall ﬁt was relatively low. Literacy rates, however, were signiﬁcant. In the second model, he assessed the impact of social capital on his performance measure and found some added improvement. In his third model, he created an interaction term by multiplying the index of social capital with a measure of new leadership in each village. Krishna found that this interaction term was highly signiﬁcant in its impact on the

26

Overview

economic development performance index for a village, but that the measure of social capital by itself dropped somewhat in importance. As Krishna concluded: Social capital and the capacity of new leaders both matter for development performance, and they matter in interaction with each other . . . The higher the value of social capital, the greater the eﬀect made by diﬀerences in agency strength. Conversely, the greater the agency variable, the more the diﬀerence in performance on account of social capital. (2001, p. 10)

Krishna’s study illustrates the importance of tailoring assessments of social capital to diﬀerent cultural, social and economic settings and how human agency is essential to productive capital.

ARE INSTITUTIONS SOCIAL CAPITAL? When many scholars limit their focus in social capital research to the role of network structure, one could reasonably ask how institutions might contribute to social capital. In this chapter, we treat institutions as a form of social capital rather than just an outcome of social capital. Much scholarship supports the observation that institutions play a signiﬁcant role in fostering cooperation (North, 1990; Oakerson, 1993; Evans, 1996; Rothstein, 1998, 2005). My own work has shown how diverse forms of institutions may enhance shared norms of trustworthiness, trust and reciprocity (see Ostrom, 1998; Ostrom and Ahn, 2003a). Part of what determines a researcher’s point of view is whether they use a static or dynamic view of the world and social capital. In a static view, one variable cannot be both a cause and an eﬀect. In a dynamic view, however, investments in capital in one time period can generate capital that can be used to build more capital in the next time period. A simple example involves physical capital: by building a factory to produce delivery trucks, a company uses physical capital to create more physical capital (for example, more delivery trucks). When one adopts a more dynamic view of social processes, one form of capital is frequently used to build more of the same form of capital, or of diﬀerent forms of capital. The debate about whether or not institutions are a type of social capital depends on both the researchers’ view of the world and the way they ask their questions about social capital. For Putnam and his colleagues (1993), social capital is the independent variable that aﬀects the dependent variable of institutions. Putnam uses the social capital that evolved in northern Italy as the main causal variable that made formal governance institutions work

What is social capital?

27

more eﬀectively in northern vs southern Italy. In his analysis, institutions are a dependent variable whose performance depends on the forms of social capital that have been created. Even Putnam, however, refers to the diﬀerent institutional history of the two regions to explain the growth of social networks in one region versus another (see also Sabetti, 2002). Other scholars treat institutions as one type of social capital. Douglass North (1990) deﬁnes institutions as the ‘rules of the game’ that individuals use to organize their activities within and across organizations. The rules are both formal and informal. ‘Rules-in-form’ (formal) are written documents from a legislature, court, administrative agency, for-proﬁt ﬁrm or not-for-proﬁt organization; they deﬁne most of the institutional structure of the entity. ‘Rules-in-use’ (informal) may evolve with less self-conscious investment, but are critical as they represent the norms underpinning the behavior a collective may deem acceptable. Such rules (Sproule-Jones, 1993) are commonly understood and enforced prescriptions about what may, must or must not be done (see also Crawford and Ostrom, 2005 [1995]). Within a dynamic paradigm of social capital, institutions aﬀect and are aﬀected by the trustworthiness of members of the group and other characteristics of the relevant networks (Pasotti and Rothstein, 2002). Evidence indicates that institutions are a potential way to increase individuals’ trust in others and to leverage existing networks to create better institutions. In other words, institutions at time t may be used to revise, update or build new trust, reciprocity and even new institutions at time t + 1. Of course, if institutions do not help individuals solve the problems they face, they may actually impede the development of new forms of social capital. It is important for researchers to remember that rules that work well in one setting may generate entirely diﬀerent incentives and outcomes in other situations (Ostrom, 1997).

AN ILLUSTRATION OF THE IMPORTANCE OF INSTITUTIONS AS A TYPE OF SOCIAL CAPITAL Research comparing the performance of two types of Nepali irrigation systems – one with high levels of physical capital and low social capital, the other with low levels of physical capital and social capital – helps illustrate the importance of institutions as a form of social capital. Farmers in Asia use their evolved knowledge (human capital) of water delivery to engineer complex irrigation systems with dams, tunnels and water diversion structures of varying size and complexity (physical capital). These systems succeed by using agreed-upon rules (social capital) for allocating water and

28

Overview

sharing responsibilities for providing the needed labor, materials and money to build the systems in the ﬁrst place and maintain them over time (Shivakoti and Ostrom, 2001). Over the years, farmers built paddy rice irrigation systems. Because regional princes governed Nepal until 1848, the central government was not involved in planning, building or maintaining these systems. Even when the ruling Rana family consolidated power in the mid-nineteenth century, irrigation received very little national attention. In the mid-1950s, however, Nepal established the Department of Irrigation (DOI) and it created a series of Five-Year Plans. Since then, the Asian Development Bank, the World Bank, CARE, the International Labor Organization and other donors have invested very large sums to design and construct large-scale, agency-managed irrigation systems (AMIS) in some Nepalese regions. No one really knows how many farmer-managed irrigation systems (FMIS) currently exist. The best estimate in 1995 was around 20 000, covering 75 percent of the total irrigated land (APROSC and JMA, 1995). The existence of multiple, diverse systems provides an excellent opportunity to compare the performance of irrigation systems organized by the farmers themselves with that of systems designed by engineers working for a donor or a national government. Since the early 1990s, colleagues associated with the Asian Institute of Technology in Bangkok, Thailand; Kathmandu University in Nepal; and Indiana University have jointly studied this situation (Benjamin et al., 1994; Lam et al., 1994; Shivakoti et al., 2005; Regmi, 2006).1 We developed the Nepal Irrigation Institutions and Systems (NIIS) database, which currently has information about 231 irrigation systems located in 29 out of the 75 districts in Nepal2 (Joshi et al., 2000). To the surprise of many of our colleagues, our consistent ﬁnding is that on average FMIS outperform AMIS on multiple dimensions. Most AMIS are funded by international donors who hire highly educated engineers to design sophisticated physical capital. One would expect that these systems – with strong metal headworks, automatic gates and concrete-lined ﬁeld canals – would easily outperform the FMIS’s low-tech systems where crude canal headworks serve unlined irrigation canals. Let me provide a very brief overview of our ﬁndings.3 Three evaluations of the physical condition of the irrigation system, and ultimately the amount of water provided to irrigated ﬁelds, reveal that a larger proportion of FMIS maintains the system in excellent or moderately good condition as contrasted to AMIS (see Table 1.1). The condition measured is not the basic construction of the system but whether the system – however constructed – is being maintained (see Lam, 1998 for a full deﬁnition of how these concepts were deﬁned). The better physical condition of their canals enables FMIS to capture increased levels of cropping intensity (the number

29

What is social capital?

Table 1.1 Relationships between governance structure and physical condition of irrigation systems Physical condition of irrigation systems Overall condition Technical efficiency Economic efficiency

Types of governance structure FMIS (%)

AMIS (%)

18.2 67.4 14.4 28.9 62.8 8.3 33.2 63.5 3.3

8.4 45.8 45.8 12.5 50.0 37.5 12.5 52.1 35.4

Excellent [37] Moderately good [144] Poor [48] Highly efficient [58] Moderately efficient [137] Inefficient [33] Highly efficient [66] Moderately efficient [140] Inefficient [23]

Chisquare value

Sig.

23.02

.00

27.30

.00

45.35

.00

Note: Number of irrigation systems is in brackets. Source: Joshi et al. (2000), p. 78.

Table 1.2 Relationships between governance structure and cropping intensity of irrigation systems Cropping intensity

Intensity at head end Intensity at tail end

Types of governance structure

High [142] Low [72] High [123] Low [87]

FMIS (%)

AMIS (%)

70.2 29.8 65.1 34.9

52.2 47.8 34.1 65.9

Chisquare value

Sig.

5.27

.02

13.74

.00

Note: Number of irrigation systems is in brackets. Source: Joshi et al. (2000), p. 80.

of crops grown during a year) at both the head and tail end of the canal, as shown in Table 1.2. As a result, the farmers’ investment in keeping their systems in good physical condition pays oﬀ in signiﬁcantly higher agricultural productivity. For those who have made visual inspections of these irrigation systems, the consistent ﬁndings of these ongoing studies are surprising. FMIS are

30

Overview

usually built with local raw materials such as stones, trees, mud and bricks, while AMIS use modern engineering tools. The fact that the human and social capital of Nepali farmers signiﬁcantly outperforms the sophisticated physical capital and considerable ﬁnancial backing and human capital of AMIS projects is a major feat. We need to explore the organization of FMIS vs AMIS further to begin to unravel the reasons for these results. About two-thirds of both FMIS and AMIS have formal written rules that include provisions for imposing ﬁnes on farmers for not contributing resources to operate and manage the systems (Joshi et al., 2000, p. 75). Nonetheless, the speciﬁc rules the farmers in FMIS use to govern their systems on a day-to-day basis – allocation, monitoring, punishment – vary substantially from one system to another. Because the ‘oﬃcial’ guard on FMIS is a rotating position among the farmers, monitoring and maintenance are primarily performed by those who helped craft the speciﬁc rules of their system. The farmers have a strong interest in seeing their system perform well and ensuring that others are not getting a free ride or taking more water than their oﬃcial share. In contrast, international development agencies usually design the rules in AMIS with little regard to the situation in the local community. The written rules of AMIS are quite similar on all systems. Yet, international agencies have not invested in educating local farmers or even DOI oﬃcials on the speciﬁcs of these rules. In eight out of ten AMIS, an oﬃcial guard is hired, while only six out of ten FMIS rely on an oﬃcial guard who is actually a farmer temporarily fulﬁlling that position (Joshi et al., 2000, p. 75). The presence of the guard, however, does not increase the likelihood that ﬁnes will be imposed on those farmers who have not contributed resources to operate the irrigation system. Fines are imposed on 75 percent of the FMIS when farmers break a rule; they are imposed on only 38 percent of the AMIS (Joshi et al., 2000, p. 76). Further research reveals more qualitative ﬁndings that farmers also follow their system’s rules to a greater extent and tend to achieve a higher level of mutual trust on FMIS than on the AMIS (ibid.). The study of irrigation systems in Nepal is only one of the empirical studies on the impact of social capital in the last 20 years. We have repeatedly found higher performance when those most aﬀected by a process have a strong voice in crafting the rules related to that process and are engaged in monitoring conformance with the rules (Gibson et al., 2000; Varughese and Ostrom, 2001; Hayes and Ostrom, 2005; Gibson et al., 2005a; Poteete and Ostrom, 2004; Dietz et al., 2003). For a more comprehensive overview of this research on institutions as a form of social capital, see McGinnis (1999a, 1999b, 2000). In summary, what matters is not just whether or not an institution is present but whether its form is well ﬁtted to both the ecological and social

What is social capital?

31

conditions facing participants and the nature of the collective-action problem that they face (Anderies et al., 2004).

LESSONS FROM STUDYING INSTITUTIONS AS A TYPE OF SOCIAL CAPITAL By now, I hope I have convinced the reader that social capital is a useful concept for the social sciences, one that can be carefully deﬁned and even more carefully measured. I also hope that its rich diversity of forms is clear and meaningful. To make the forms relevant, however, they must be tailored to the problems individuals are trying to solve. Members of the group must also know how to employ these forms easily if they are to succeed. Unfortunately, many well-meaning scholars and organizations devoted to helping individuals in developing countries do not understand the importance of social capital and of ﬁtting institutional rules to diverse settings (Gibson et al., 2005b; Andersson, 2004; Shivakumar, 2005). In irrigation, for example, the international development funding is rarely allocated appropriately between building major new projects and further developing systems that build on the existing ones created by local participants (Araral, 2005). This misallocation of funds toward major new projects has resulted in lower than expected returns on signiﬁcant international investments. The World Bank alone contributed over $10 billion in loans for irrigation projects between 1983 and 1999 (Pitman, 2002, p. 12; see also Yudelman, 1985). International donors were contributing about $2 billion per year during the 1990s (Winpenny, 1994). Hugh Turral (1995, p. 1) captured the judgment of many analysts by concluding that: ‘irrigation schemes have often under-performed in economic terms, and ﬁeld research has highlighted substantial shortcomings in management (operation and maintenance), equity, cost-recovery and agricultural productivity’. Some critics, like William Easterly (2001, 2006), assert that most of the funding spent by international aid agencies since the 1960s has, tragically, not achieved promised results (see also Gibson et al., 2005b). The confluence of many complex factors caused these wasted investments. What a tragic, missed opportunity to improve human well-being in the developing world. That being said, research indicates that one of the causes was the hubris of experts who relied on simple models of ‘the best’ engineering plans and idealized sets of rules, if they paid any attention to institutions at all. The investment made by the United States Agency for International Development (USAID)-funded Rapti Integrated Rural Development Project in the Chiregad system in the Rapti Valley of Nepal

32

Overview

is an excellent example of this (Hilton, 1990, 2001). The new system was built in a location that had been irrigated by ﬁve farmer-built and managed irrigation systems that had evolved diverse rules ﬁtted to their own locations. The engineers building the new system did not even take the existence of these systems into account except to leave parts of them standing to serve as distribution canals.4 After construction, the Nepal Department of Irrigation started what they called a ‘participatory process’ to involve the farmers in managing the system since they had previously ignored them during the construction process. The DOI formed a Water User Committee and appointed the chair of the local village panchayat to head this committee even though he did not own any land in the service area, was not involved in the original irrigation system and had no useful previous experience. This chair then appointed the rest of the committee without including any of the water managers of the previous systems, who were all trying to distribute water once it reached (if it did) their established distribution canals. When Hilton interviewed the oﬃcial members of the Water User Committee, they indicated that they rarely met. They could not provide any information about the system characteristics or how it operated. She also found that the area receiving irrigation water after the Chiregad system was built was smaller than the area served earlier by the ﬁve independent systems, that maintenance levels were lower, and that other farmermanaged systems in the Rapti Valley were achieving higher agricultural productivity than the farmers served by the new system. If this were just one story about a single project, we could write it oﬀ as simply an anomaly. Unfortunately, such stories occur regularly throughout the developing world and are happening more and more frequently as government oﬃcials exert more authority over local resources. As Peter Evans (2004, pp. 31–2) articulates: ‘Currently, the dominant method of trying to build institutions that will promote development is to impose uniform institutional blueprints on the countries of the global South – a process that I call “institutional monocropping”.’ Many of these imposed rules-in-form are never used in practice. Powerful individuals within a community are frequently given further advantages (for example, control over the resources), leading to major problems in economic development. The poor and the powerless are the ones who pay the most. Moreover, Lant Pritchett and Michael Woolcock (2004) point out the problem of imposing uniform solutions, especially when the dominant topdown bureaucratic paradigm is the solution used by development agencies. They argue that imposing uniform social capital is counterproductive. Their graphic descriptions of the systematic failure of development agencies in improving services to rural areas, including irrigation, are

What is social capital?

33

telling. Donor activity often amounts to sending ‘experts’ who copy institutions in ‘Denmark’ to launch institutions in ‘Djibouti’. Institutions designed by people living in Denmark to work in Denmark are likely to be successful. When imposed on others, however, the likelihood of success is very low. To sum up, social capital – particularly in its form as institution or norms – needs to be tailored both to the unique circumstance and to the type of collective-action problem that it is meant to solve. As scholars, we need to encourage the study of diverse forms of social capital to get at underlying design principles (Ostrom, 1990). Only in this way can participants draw on knowledge of their own ecological and social situation and use this to design speciﬁc forms of social capital that will prove to be most helpful in their local context.

NOTES 1. Earlier we worked closely with the Irrigation Management Systems Study Group at the Institute of Agriculture and Animal Science, Tribhuvan University, but the current Nepal Irrigation Institutions and Systems database is housed at Kathmandu University. 2. Until recently, there has been considerable rebel activity in Nepal that has disrupted many activities, especially in the countryside, and created many tragedies for Nepali farmers. The ﬁndings discussed in this chapter are based on data, most of which were collected in earlier peaceful times. 3. Readers who wish to dig deeper are encouraged to read Lam (1998), Joshi et al. (2000), Shivakoti and Ostrom (2001, and the extensive references cited therein) and Shivakoti et al. (2005). 4. Walter Coward (1980) reported similar investments in physical capital while ignoring the extensive investments of farmers in both physical and social capital (see also Coward, 2006).

REFERENCES Ahn, T.K. and Elinor Ostrom (forthcoming), ‘Social capital and collective action’, in D. Castiglione, J.W. van Deth and Guglielmo Wolleb (eds), Handbook of Social Capital, Oxford: Oxford University Press. Anderies, J. Marty, Marco Janssen and Elinor Ostrom (2004), ‘A framework to analyze the robustness of social-ecological systems from an institutional perspective’, Ecology and Society, 9 (1), 18. Andersson, Krister (2004), ‘Who talks with whom? The role of repeated interactions in decentralized forest governance’, World Development, 32 (2), 233–49. APROSC and JMA (1995), Nepal Agriculture Perspective Plan (Final Report) – Main Document, Kathmandu: Agricultural Projects Service Center and John Miller Associates. Araral, Eduardo (2005), ‘Bureaucratic incentives, path dependence, and foreign aid: an empirical institutional analysis of irrigation in the Philippines’, Policy Sciences, 38 (2–3), 131–57.

34

Overview

Arrow, Kenneth (1999), ‘Observations on social capital’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 3–5. Becker, Gary S. (1964), Human Capital, New York: Columbia University Press. Benjamin, Paul, Wai Fung Lam, Elinor Ostrom and Ganesh Shivakoti (1994), Institutions, Incentives, and Irrigation in Nepal: Decentralization Finance and Management Project Report, Burlington, VT: Associates in Rural Development. Birner, Regina and Heidi Wittmer (2003), ‘Using social capital to create political capital: how do local communities gain political influence? A theoretical approach and empirical evidence from Thailand’, in Nives Dolsak and Elinor Ostrom (eds), The Commons in the New Millennium: Challenges and Adaptations, Cambridge, MA: MIT Press, pp. 291–334. Boissevain, J. (1974), Friends of Friends: Networks, Manipulators, and Coalitions, Oxford: Blackwell. Bonacich, Phillip and Sherry Schneider (1992), ‘Communication networks and collective action’, in W.B.G. Liebrand, D.M. Messick and H.A.M. Wilke (eds), Social Dilemmas: Theoretical Issues and Research Findings, New York: Pergamon Press, pp. 225–45. Bourdieu, Pierre (1986), ‘Forms of capital’, in John G. Richardson (ed.), Handbook of Theory and Research for the Sociology of Education, New York: Greenwood Press, pp. 241–58. Bourdieu, Pierre and L.J.D. Wacquant (1992), An Invitation to Reflexive Sociology, Chicago, IL: University of Chicago Press. Brief, Arthur P. and Stephan J. Motowidlo (1986), ‘Prosocial organizational behaviors’, Academy of Management Review, 11 (4), 710–25. Burt, Ronald S. (1992), Structural Holes: The Social Structure of Competition, Cambridge, MA: Harvard University Press. Burt, Ronald S. (1997), ‘The contingent value of social capital’, Administrative Science Quarterly, 42, 339–65. Burt, Ronald S. (2000), ‘Network structure’, Research in Organizational Behavior, 22, 345–423. Burt, Ronald S. (2002), ‘Bridge decay’, Social Networks, 24, 333–63. Coleman, James (1990), Equality and Achievement in Education, Boulder, CO: Westview Press. Coward, E. Walter (1980), Irrigation and Agricultural Development in Asia: Perspectives from Social Sciences, Ithaca, NY: Cornell University Press. Coward, E. Walter (2006), ‘Property landscapes in motion’, Keynote Address at the International Association for the Study of Common Property, 11th Biennial Global Conference, Bali, Indonesia, June. Crawford, Sue E.S. and Elinor Ostrom (2005), ‘The grammar of rules’, in Elinor Ostrom (ed.), Understanding Institutional Diversity, Princeton, NJ: Princeton University Press, pp. 137–74; originally published 1995 in American Political Science Review, 89 (3), 582–600. Dasgupta, Partha (2003), ‘Social capital and economic performance: analytics’, in Elinor Ostrom and T.K. Ahn (eds), Foundations of Social Capital, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 309–39. Dasgupta, Partha and Ismail Serageldin (eds) (1999), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank. Dietz, Thomas, Elinor Ostrom and Paul Stern (2003), ‘The struggle to govern the commons’, Science, 302 (5652), 1907–12.

What is social capital?

35

Durlauf, Steven N. (2002), ‘Bowling alone: a review essay’, Journal of Economic Behavior and Organization, 47 (3), 259–74. Easterly, William (2001), The Elusive Quest for Growth: Economists’ Adventures and Misadventures in the Tropics, Cambridge: Cambridge University Press. Easterly, William (2006), The White Man’s Burden: Why the West’s Eﬀorts to Aid the Rest Have Done So Much Ill and So Little Good, New York: Penguin Press. Evans, Peter (1996), ‘Government action, social capital and development: reviewing the evidence on synergy’, World Development, 24 (6), 1119–32. Evans, Peter (2004), ‘Development as institutional change: the pitfalls of monocropping and the potentials of deliberation’, Studies in Comparative International Development, 39 (4), 30–52. Fine, Ben (2001), Social Capital versus Social Theory: Political Economy and Social Science at the Turn of the Millennium, London: Routledge. Fine, Gary A. and Lori Holyﬁeld (1996), ‘Secrecy, trust and dangerous leisure: generating group cohesion in voluntary organizations’, Social Psychology Quarterly, 59 (1), 22–38. Fukuyama, Francis (2000), ‘Social capital and civil society’, IMF Working Paper WP/00/74, Washington, DC: International Monetary Fund. Gambetta, Diego (1988), ‘Maﬁa: the price of distrust’, in Diego Gambetta (ed.), Trust: Making and Breaking Cooperative Relations, London: Basil Blackwell, pp. 158–75. Gambetta, Diego and Heather Hamill (2005), Streetwise: How Taxi Drivers Establish Customers’ Trustworthiness, New York: Russell Sage Foundation. Gibson, Clark, Margaret McKean and Elinor Ostrom (eds) (2000), People and Forests: Communities, Institutions, and Governance, Cambridge, MA: MIT Press. Gibson, Clark, John Williams and Elinor Ostrom (2005a), ‘Local enforcement and better forests’, World Development, 33 (2), 273–84. Gibson, Clark, Krister Andersson, Elinor Ostrom and Sujai Shivakumar (2005b), The Samaritan’s Dilemma: The Political Economy of Development Aid, Oxford: Oxford University Press. Hardin, Garrett (1968), ‘The tragedy of the commons’, Science, 162, 1243–8. Hardin, Russell (1999), ‘Social capital’, in James Alt, Margaret Levi and Elinor Ostrom (eds), Competition and Cooperation, New York: Russell Sage Foundation, pp. 170–89. Hayes, Tanya and Elinor Ostrom (2005), ‘Conserving the world’s forests: are protected areas the only way?’, Indiana Law Review, 38 (3), 595–617. Hilton, Rita (1990), Cost Recovery and Local Resource Mobilization – An Examination of Incentives in Irrigation Systems in Nepal: Decentralization Finance and Management Project Report, Burlington, VT: Associates in Rural Development. Hilton, Rita (2001), ‘Institutional incentives for resource mobilization in farmermanaged and agency-managed irrigation’, in Ganesh P. Shivakoti and Elinor Ostrom (eds), Improving Irrigation Governance and Management in Nepal, Oakland, CA: ICS Press, pp. 150–76. Hoﬀman, Elizabeth and Gary Libecap (1995), ‘The failure of governmentsponsored cartels and development of federal farm policy’, Economic Inquiry, 33, 365–82. Joshi, N.N., Elinor Ostrom, Ganesh Shivakoti and Wai Fung Lam (2000), ‘Institutional opportunities and constraints in the performance of farmer-

36

Overview

managed irrigation systems in Nepal’, Asia-Paciﬁc Journal of Rural Development, 10 (2), 67–92. Knack, S. and P. Keefer (1997), ‘Does social capital have an economic payoﬀ? A cross-country investigation’, Quarterly Journal of Economics, 112, 1251–88. Krishna, Anirudh (2001), ‘Moving from the stock of social capital to the flow of beneﬁts: the role of agency’, World Development, 29 (6), 925–43. Krishna, Anirudh and Norman Uphoﬀ (1999), ‘Mapping and measuring social capital: a conceptual and empirical study of collective action for conserving and developing watershed in Rajasthan, India’, Social Capital Initiative Working Paper 13, Washington, DC: World Bank. Lachmann, Ludwig M. (1978), Capital and Its Structure, Kansas City, MO: Sheed Andrews and McMeel. Lake, Ronald La Due and Robert Huckfeldt (1998), ‘Social capital, social networks, and political participation’, Political Psychology, 19 (3), 567–84. Lam, Wai Fung (1998), Governing Irrigation Systems in Nepal: Institutions, Infrastructure, and Collective Action, Oakland, CA: ICS Press. Lam, Wai Fung, Myungsuk Lee and Elinor Ostrom (1994), ‘An institutional analysis approach: ﬁndings from the NIIS on irrigation performance’, in Jennifer Sowerwine, Ganesh Shivakoti, Ujjwal Pradhan, Ashutosh Shukla and Elinor Ostrom (eds), From Farmers’ Fields to Data Fields and Back: A Synthesis of Participatory Information Systems for Irrigation and Other Resources, Colombo, Sri Lanka: International Irrigation Management Institute, and Rampur, Nepal: IAAS, pp. 69–93. Langbein, Laura and Connie M. Jorstad (2002), ‘Productivity in the workplace: cops, culture, communication, cooperation, and collusion’, working paper, Washington, DC: American University, School of Public Aﬀairs. Lin, Nan (2001), Social Capital: A Theory of Social Structure and Action, Cambridge: Cambridge University Press. Loury, Glenn (1977), ‘A dynamic theory of racial income diﬀerences’, in P.A. Wallace and A. LeMund (eds), Women, Minorities, and Employment Discrimination, Lexington, MA: Lexington Books, pp. 153–88. McEvily, Bill and Akbar Zaheer (2005), ‘Does trust still matter? Research on the role of trust in inter-organizational exchange’, in R. Bachman and A. Zaheer (eds), Handbook of Trust Research, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 280–302. McGinnis, Michael (2000), Polycentric Games and Institutions: Readings from the Workshop in Political Theory and Policy Analysis, Ann Arbor, MI: University of Michigan Press. Messick, David M. and Marilyn B. Brewer (1983), ‘Solving social dilemmas: a review’, in L. Wheeler and P. Shaver (eds), Annual Review of Personality and Social Psychology, Beverly Hills, CA: Sage, pp. 11–44. Murnighan, J.K., J.W. Kim and A.R. Metzger (1994), ‘The volunteer dilemma’, Administrative Science Quarterly, 38, 515–38. Nahapiet, Janine and Sumantra Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23 (2), 242–66. North, Douglass C. (1990), Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press. Oakerson, Ronald (1993), ‘Reciprocity: a bottom-up view of political development’, in Vincent Ostrom, David Feeny and Hartmut Picht (eds), Rethinking

What is social capital?

37

Institutional Analysis and Development: Issues, Alternatives, and Choices, San Francisco, CA: ICS Press, pp. 141–58. Olson, Mancur (1965), The Logic of Collective Action: Public Goods and the Theory of Groups, Cambridge, MA: Harvard University Press. Ostrom, Elinor (1990), Governing the Commons: The Evolution of Institutions for Collective Action, New York: Cambridge University Press. Ostrom, Elinor (1992), Crafting Institutions for Self-Governing Irrigation Systems, San Francisco, CA: ICS Press. Ostrom, Elinor (1998), ‘A behavioral approach to the rational choice theory of collective action’, American Political Science Review, 92 (1), 1–22. Ostrom, Elinor (1999), ‘Social capital: a fad or a fundamental concept?’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 172–214. Ostrom, Elinor and T.K. Ahn (2003a), ‘Introduction’, in Elinor Ostrom and T.K. Ahn (eds), Foundations of Social Capital, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. xi–xxxix. Ostrom, Elinor and T.K. Ahn (eds) (2003b), Foundations of Social Capital, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Ostrom, Vincent (1997), The Meaning of Democracy and the Vulnerability of Democracies, Ann Arbor, MI: University of Michigan Press. Pasotti, Eleonara and Bo Rothstein (2002), ‘In the market for ideas: a quest for a uniﬁed conceptual approach in political science’, paper delivered at the Annual Meeting of the American Political Science Association, Boston, MA, 29 August –1 September. Petroski, Henry (2006), Success through Failure: The Paradox of Design, Princeton, NJ: Princeton University Press. Pew Research Center for the People and the Press (1996), Trust and Citizen Engagement in Metropolitan Philadelphia: A Case Study, Washington, DC: Pew. Pitman, George (2002), Bridging Troubled Waters: Assessing the WB Water Resources Strategy, Washington, DC: WP-Operations Evaluation Department. Portes, A. (1998), ‘Social capital: its origins and applications in modern sociology’, Annual Review of Sociology, 24, 1–24. Poteete, Amy and Elinor Ostrom (2004), ‘Heterogeneity, group size, and collective action: the role of institutions in forest management’, Development and Change, 35 (3), 435–61. Pritchett, Lant and Michael Woolcock (2004), ‘Solutions when the solution is the problem: arraying the disarray in development’, World Development, 32 (2), 191–212. Putnam, Robert (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Putnam, Robert (with Robert Leonardi and Raﬀaella Nanetti) (1993), Making Democracy Work, Princeton, NJ: Princeton University Press. Regmi, Ashok (2006), ‘The role of group heterogeneity in collective action: a look at the interrelationship between irrigation and forests. Case studies from Chitwan, Nepal’, PhD dissertation, Indiana University. Rothstein, Bo (1998), ‘State building and capitalism: the rise of the Swedish bureaucracy’, Scandinavian Political Studies, 21 (2), 287–306. Rothstein, Bo (2005), Social Traps and the Problem of Trust, Cambridge: Cambridge University Press.

38

Overview

Sabetti, Filippo (2002), The Search for Good Government: Understanding the Paradox of Italian Democracy, Montreal: McGill-Queens University Press. Schultz, Theodore W. (1961), ‘Investments in human capital’, American Economic Review, 51 (1), 1–17. Shivakoti, Ganesh and Elinor Ostrom (2001), Improving Irrigation Governance and Management in Nepal, Oakland, CA: ICS Press. Shivakoti, Ganesh, Douglas Vermillion, Wai-Fung Lam, Elinor Ostrom, Ujjwal Pradhan and Robert Yoder (eds) (2005), Asian Irrigation in Transition: Responding to Challenges, New Delhi, India: Sage. Shivakumar, Sujai (2005), The Constitution of Development: Crafting Capabilities for Self-Governance, New York: Palgrave Macmillan. Solow, Robert M. (1999), ‘Notes on social capital and economic performance’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 6–10. Sproule-Jones, Mark (1993), Governments at Work: Canadian Parliamentary Federalism and Its Public Policy Eﬀects, Toronto: University of Toronto Press. Turral, Hugh (1995), ‘Recent trends in irrigation management: changing directions for the public sector’, London: Overseas Development Institute, no. 5, September. Uzzi, B. (1997), ‘Social structure and competition in interﬁrm networks: the paradox of embeddedness’, Administrative Science Quarterly, 42, 35–67. Varughese, George and Elinor Ostrom (2001), ‘The contested role of heterogeneity in collective action: some evidence from community forestry in Nepal’, World Development, 29 (5), 747–65. Winpenny, J.T. (1994), Managing Water as an Economic Resource, New York: Routledge. Yudelman, Montague (1985), ‘The World Bank and agricultural development: an insider’s view’, World Resources Paper no. 1, Washington, DC: World Resources Institution.

2.

Network duality of social capital Ronald S. Burt1

This chapter is about balance between brokerage and closure, the two primary mechanisms by which social networks constitute social capital. Brokerage involves building connections across groups to increase exposure to diverse opinion and practice. Brokerage is associated with growth and innovation. Closure involves strengthening connections within a group to focus the group on a limited set of opinions and practice. Closure is associated with trust and alignment, ultimately enhancing eﬃciency. The balance between brokerage and closure is usually analyzed in terms of where to invoke the mechanisms: maximum advantage occurs when a closed network secures alignment within a team and team members have brokerage networks beyond the team (see Burt, 2005, pp. 126–66, for review and illustrative evidence). The balance has also been discussed, though never to my knowledge with beneﬁt of network data, in terms of when to invoke the mechanisms: brokerage and closure are in perpetual cycle as a network duality, mending and undoing one another; brokerage followed by closure, followed by brokerage, and so on. It is misadventure to break the cycle. Business examples abound. General Electric chief executive oﬃcer (CEO) Jack Welsh emphasized ‘integrated diversity’ in the 1980s – a unity believed to work only ‘when the elements of that diversity, GE’s thirteen business units, were strong in their own right’ (Slater, 1999, p. 97). Rhone-Poulenc CEO JeanRené Fourtou emphasized the importance of preserving ‘le vide’ (literally, vacuum or empty space, or in network terms, structural holes): ‘Le vide has a huge function in organizations.’ ‘Shock comes when diﬀerent things meet. It’s the interface that’s interesting’; Stewart (1996, p. 165). Gernot Grabher (1993) concluded that economic development in the German Ruhr was hurt by rigid specialization enforced in dense interorganizational networks, the same networks that enhanced the region’s prior growth. AnnaLee Saxenian (1994) concluded that the San Francisco technology community in Silicon Valley had a competitive advantage over Boston’s Route 128 because flexible specialization endured in Silicon Valley. Clayton Christensen (1997) drew attention to the frequent corporate failures that result when an industry leader is too focused on improving what made the 39

40

Overview

company great rather than anticipating what will make the next industry leader great. John Hagel and John Seely Brown (2005) caution against pushing eﬃciency so far as to eliminate the ‘productive friction’ that creates value. These are a sampling of many such discussions. My purpose in this chapter is to make more explicit the network mechanisms underlying such discussions. I begin with a quick statement of the two network mechanisms responsible for social capital, oﬀering illustrative evidence, then describe the mechanisms as they came together in a speciﬁc management initiative that illustrates a common network-duality failure mode, here discussed as ‘premature consensus’. The moral is that bridging our diﬀerences today creates a risk of decimating future growth. Being aware of the risk is a ﬁrst-line defense encouraging balance between brokerage and closure. More generally, the risk is less with more flexible embedding networks such as competing coalitions or safe-harbor common areas.

STRUCTURAL HOLES Figure 2.1 is a sociogram of the network around a manager. Dots represent people. Lines represent relationships. The manager has eight direct contacts. Those eight contacts deﬁne the manager’s immediate network. Of course, the network does not ‘belong’ to the manager. It is co-owned with contacts. A more accurate label for Figure 2.1 would be ‘the interface between manager and social structure’, but the label is clumsy. The structure of relations among the contacts deﬁnes the immediate network around the manager, which is typically what is meant when people talk about a person’s network. Beyond the immediate network in Figure 2.1 are a host of indirect contacts, friends of friends at various distances. Hollow dots represent people with whom the manager has indirect contact through his or her eight direct contacts. Dashed lines represent connections with and among the indirect contacts. Inherent in the network is a level of social capital, a competitive advantage the manager enjoys as a result of the network. Two facts from empirical research are the foundation for network models of social capital. Firstly, people cluster into groups and tribes deﬁned by the organizations with which they aﬃliate, the projects in which they are involved, the oﬃces where they work, the places they meet and shared interests they discover. Secondly, communication is more frequent within groups than it is between groups, such that people in the same group come to have similar views of the history that led to today, similar views of proper opinion and practice, and similar views of how to go forward into the future. People tire of

41

Figure 2.1

Manager network illustrating opportunities for brokerage and closure

42

Overview

repeating arguments and stories explaining why they believe and behave the way they do. They make up shorthand phrases to reference whole paragraphs of text with which colleagues are familiar. Jargon flourishes. Not only jargon, but a system of phrasing, opinions, symbols and behaviors deﬁning what it means to be a member of the group. What was once explicit knowledge interpretable by anyone becomes tacit knowledge meaningful only to insiders. With time, new combinations and nuances emerge. The tacit knowledge becomes more complex, making it more diﬃcult to move to other groups. Much of what we know is not readily understood beyond the colleagues around us. Explicit knowledge converted into local, tacit knowledge makes information sticky, such that holes tear open in the flow of information between groups. Holes in the social structure of communication, or more simply ‘structural holes’, are missing relationships that inhibit information flow. A hole ‘is a buﬀer, like an insulator in an electric circuit’ (Burt, 1992, p. 18). Structural holes are a source of eﬃciency at the same time as they are a source of growth. As a source of eﬃciency, structural holes are boundary markers in the division of labor. By not having to attend to the interpretations of people beyond the boundary around my specialty, I can focus on deepening my knowledge of what I already know pretty well. Without structural holes, we would be overwhelmed with the diversity of knowledge out there – and I expect that we would quickly establish structural holes to re-establish a sense of control over our lives. Structural holes are simultaneously a source of growth from the hardy souls among us who reach out to broker connections across the holes to create new combinations of existing opinion and practice.

BROKERAGE AND GROWTH As a network form of social capital, brokerage is about the advantage of exposure to variation in opinion and practice provided by building connections across structural holes, an advantage associated with performance in the form of innovation and growth (see Burt, 2005, Chapters 1–2, for review). Brokerage is measured in terms of the opportunities a network provides to coordinate across structural holes. Where everyone knows everyone else, there are no structural holes to broker. The more disconnected the manager’s contacts, the more likely his or her network is to span structural holes in the surrounding organization and market. In the earlier example in Figure 2.1, the ﬁve contacts to the east have no connections with one another and their contacts do not connect. Therefore, the eastern part of the network is rich in structural holes.

Network duality of social capital

43

People who connect across structural holes – call them network brokers, connectors or entrepreneurs – have a vision advantage in detecting and developing lucrative opportunities. People who have no contact with one another often employ more limited problem-solving and practices in their work. Because network brokers are more exposed to the diversity of these diverse opinions and practices, brokers have a vision advantage in selecting alternative ways to go, synthesizing new ways to go, and detecting likely supporters and opponents to implementing a proposed way to go. Thus, people with strong relations to otherwise disconnected groups have a competitive advantage in detecting and developing rewarding opportunities. Figure 2.2 contains illustrative evidence. People are sorted across the horizontal axes of the graphs by network constraint, a concentration measure of the extent to which a person’s network time and energy are consumed by a single group (for example, Burt, 1992, Chapter 2). Ideographs at the extremes of the horizontal axes illustrate network structures deﬁning high and low constraint. To the right, everyone you know knows one another. Your network is like a straitjacket locking you into one way of thinking. To the left, you are freed from the constraint of any one person or group by having connections with multiple groups, as indicated by the structural holes between your contacts. Network constraint approaches its minimum value of zero. Figure 2.2A shows the association between brokerage and performance. The data come from eight study populations: salesmen, the supply chain managers in Figure 2.2B, investment bankers, human resource oﬃcers, engineers, operations people, and organizations in America as well as organizations in France and Singapore (Burt, 2005, pp. 34–46). The vertical axis is a residual z-score measure of performance relative to peers. A performance indicator is regressed over background factors such as job rank, kind of work, geographic location, experience, and so on, where performance includes compensation, annual evaluations and promotions. Some people do better than expected. Some do worse. That residual performance relative to how people ‘like you’ perform is the vertical axis in Figure 2.2A. A score of zero indicates a level of performance typical for someone with your background. For the investment bankers, for example, it is the bonus compensation typical for someone with your job rank and your years in the company. For engineers in one of the study populations, it is the age at which people like you were promoted to your current job rank where ‘like you’ means the same job rank, education, gender, race, functional area, region of the country, and so on. A score of one on the vertical axis in Figure 2.2A indicates someone one standard deviation ahead of their peers. The graph shows a strong, negative correlation – more network constraint, weaker performance. People who have networks

44

Overview 2.5

A. Social capital of brokerage is evident from compensation, evaluations, and promotion rates better than peers

Z-score relative performance

(compensation, evaluation, promotion)

2.0 1.5

55%

28%

1.0 0.5

17%

0.0 –0.5

5%

10%

–1.0 –1.5 –2.0 Z = 2.78 – 0.82 In (C) r = –0.53

–2.5 5

15

25

35

45

55

75

85

95

B. Which can be traced back to a vision advantage in detecting and developing good ideas

1.5

Average Z-score idea value

65

1.0

33%

2%

0.5

64% Z = 3.35 – 0.84 In (C) r = –0.80

0.0

–0.5

–1.0

85%

5

15

25

35

45

55

65

75

85

Network constraint (C) many

structural holes

few

95

C. And is much of the predicted variance: network constraint (white), job rank (grey), and other factors (black). The first pie is investment banker compensation and analyst election to the All-America Research Team. Second pie is supply chain and HR manager compensation. Third pie is early promotion to senior job rank in a large electronics firm.

Note: Circles are average scores on the vertical axis (Z) for a five-point interval of network constraint (C) within each study population. Dashed line goes through mean values of Z for intervals of C. Bold line is performance predicted by the natural logarithm of C. Other details are in the text.

Figure 2.2

Performance and brokerage

that span structural holes (to the left in the graph) perform above their peers. People with connected contacts (to the right in the graph) perform below their peers. The advantage manifest as brokers enjoying higher compensation, more positive evaluations, and more likely promotion, can be traced to a vision

Network duality of social capital

45

advantage illustrated in Figure 2.2B. The graph shows an association between brokerage and good ideas. Supply chain managers in a large American electronics ﬁrm were asked to describe their best idea for improving the value of the supply chain organization to the company. Their descriptions were judged by two senior vice-presidents in the supply chain, from which I computed standardized scores measuring idea ‘value’ for the vertical axis of Figure 2.2B (Burt, 2005, pp. 91–2). The strong, negative correlation in Figure 2.2B shows more positive evaluations of the ideas proposed by managers connected to otherwise disconnected groups in the organization. The two graphs in Figure 2.2 show a statistically signiﬁcant association between brokerage and performance. The association is also substantively signiﬁcant. Pie charts in Figure 2.2C show how performance associations with brokerage compare with associations with other predictors. The ﬁrst pie shows that brokerage accounts for a little more than half of the explained variance in investment banker compensation and analyst recognition in the All-America Research Team (Burt, 2007). Job rank contributes a ﬁfth of the variance. Demographic and geographic factors account for the remaining explained variance. The second pie chart shows a small network eﬀect in bureaucratic organizations. Within two functional organizations – supply chain in an electronics company (Burt, 2005) and human resources in a ﬁnancial organization – brokerage contributes a statistically signiﬁcant, but substantively small, 10 percent of predictable compensation diﬀerences between managers. Demographic and geographic factors make about the same contribution (9 percent). What really matters is job rank. In a bureaucracy, compensation is a function of job rank. The key to high compensation is high job rank. Compensation diﬀerences between the supply chain and HR managers are largely determined by job rank (81 percent). But getting to a senior job rank is largely a function of network brokerage. The third pie chart in Figure 2.2C shows that brokerage contributes two-thirds of the explained variance in early promotion to senior job rank in a large electronics company (Burt, 1992). Thus, compensation remains a result of brokerage, but the eﬀect is indirect through job rank in a bureaucratic organization – network brokers are more likely to get promoted to senior rank early, where they enjoy higher compensation. Figure 2.2 illustrates an empirical result that has become familiar over the last two decades: people whose networks span structural holes are at higher risk of detecting and developing good ideas, because of which they enjoy higher compensation than peers, more-positive evaluations and faster promotions.

46

Overview

CLOSURE AND STABILITY Closure is about the beneﬁt of protection from diverse opinion and practice, protection provided by building connections that do not span structural holes. This is a form of social capital associated with eﬃciency gains in performance (see Burt, 2005, Chapters 3–4, for review). Closure is measured by the extent to which everyone in a network is connected to everyone else, through a central person in the network, or through direct connections between people in the network. Back in Figure 2.1, the manager and his or her three contacts to the west are densely connected, in part directly and in part through several friends of friends. The closed network is detrimental with respect to the vision advantage of brokerage, but can be an advantage with respect to coordinating work. Each bridge relation that coordinates across a structural hole increases closure, so it is useful to know how the emergent, more-closed network is tied to value. Reputation is the mechanism by which closure has its eﬀect. As connections close the network around a manager, people are more informed about one another and calibrate with respect to one another. Reputations emerge to distinguish the peripheral from the best among us. Some people are eminent, respected members of a network; others are peripheral, barely considered legitimate members. To preserve reputation among colleagues well informed about one another’s behavior, people are careful to behave well (which lowers the risk of trusting colleagues within the network) and people work to keep up with colleagues (which lowers cost within the network by increasing the quality and quantity of work and decreasing the need for a supervisor to monitor individual behavior). For reputation to have its salutary eﬀects, there has to be a credible threat that a person’s reputation will persist to aﬀect future relationships. From an individual’s work in one project group, word gets around deﬁning their reputation, which precedes them into their next project group. If positive reputation quickly dissolves, reputation loses its attraction as an incentive to align with colleagues because yesterday’s good behavior is too soon forgotten. If negative reputation quickly dissolves, reputation loses its coercive power because yesterday’s poor behavior is too soon forgotten. ‘Too soon’ is relative. It could be a day, a month, a year. Relative stability is the key. Reputation has to persist longer than the productive relations it facilitates and the hurtful relations it protects against. Stability cannot be taken for granted. Network closure varies from low to high, so closure-induced stability must vary. How does stability change with closure? How closed must a network be before there is a credible cost for losing reputation?

Network duality of social capital

47

Figure 2.3 contains illustrative evidence on investment banker reputations in a ﬁnancial organization during the 1990s (Burt, 2005, Chapter 4). Banker reputation is measured here as it is measured in the organization: by the average evaluation a banker receives from colleagues. Each year, people in the bonus pool are asked to name colleagues with whom they worked closely during the year, and describe how it was to work with each colleague (4 outstanding, 3 good, 2 average, 1 poor; these are my synonyms for the words actually used). A banker’s average rating is then used to guide promotion and bonus decisions. Despite massive change in relationships from one year to the next, reputations persist. Three of four relationships cited this year are not cited next year, but a banker’s reputation this year is correlated 0.6 with his or her reputation next year. Intrigued by stable reputations in chaotic networks, I raised the issue over drinks with one of the organization’s senior people. He looked puzzled, then patiently explained to me that ‘of course’ employee reputations are stable. They are the company’s market index of employee quality. A good employee this year is a good employee next year, regardless of the colleagues with whom the employee works. Reputations are expected to go up and down a little depending on personalities and business opportunities, but good employees continue to be good employees, and weak employees are weeded out. In other words, the division head had a human-capital explanation for reputation stability. Able people receive good evaluations. Weak people receive poor evaluations. Reputation is correlated over time because human capital continues over time, certainly between adjacent years. I had a social-capital explanation. Colleague evaluations are based on limited personal experience mixed with the experiences of colleagues with whom work is discussed. The more connected the colleagues making evaluations, the more likely their evaluations are in part formed by stories they have shared about the object evaluated. The human-capital and social-capital explanations can be tested against each other. If individual ability is the reason for reputation stability over time, then stability should be independent of connections between colleagues. An able employee should receive good evaluations whether the colleagues who made the evaluations work together (that is, are more connected) or work in separate parts of the organization (that is, are less connected). If reputation stability is deﬁned by colleagues sharing stories about the employee, then stability should be higher when colleagues are more connected because they are more likely to have shared stories about the employee. Evidence in Figure 2.3 supports the social-capital explanation: reputation stability increases dramatically with network closure.

48

Overview

R(t+1) 4 3

0.8

2 1

1

2

3

4

0.7

Rt (13-person subsample)

Correlation between banker’s reputation this year and next

0.6 0.5 0.4 0.3 0.2

R(t+1) 4

0.1

3 2 1

e.g. sociogram bottom right

0.0 1

2

3

Rt

4

0 banker

2

4

6

8

Average number of mutual contacts linking banker this year with colleagues

10 or more banker

Note: Dots are average correlations at each level of closure. Bold regression line through the hollow dots describes stability in positive reputations (8.1 routine t-test). Thin regression line through solid dots describes stability in negative reputations (6.1 routine t-test). Dashed line goes through mean corrrelations across all of the bankers at each level of closure.

Figure 2.3

Closure and banker reputation stability

Reputation stability is measured on the vertical axis by correlation between reputations in adjacent years within a subsample of the six bankers with less-closed networks and the six bankers with more-closed networks (Burt, 2005, p. 209n). Bankers at the top of the vertical axis have reputations this year very similar to their reputations next year. Closure is measured on the horizontal axis by the extent to which an employee is evaluated by connected colleagues. For each colleague citing an employee in a particular year, the number of mutual contacts is the number of people citing the employee that year and connected to the colleague by an evaluation. An employee’s score on the horizontal axis in Figure 2.3 is the employee’s average number of mutual contacts with evaluating

Network duality of social capital

49

colleagues. For this illustration, I rounded scores to the nearest of the 11 integer categories on the horizontal axis (see Burt, 2007, Table 2, for regression results with continuous measures and controls). Lines in the graph show reputation stability increasing with closure. Where colleagues have no contact with one another, banker reputation this year has no correlation with reputation next year (0.09 correlation). Do the same work with interconnected colleagues, and reputation this year is a good predictor of reputation next year (0.73 correlation for ten or more mutual colleagues). And the closure eﬀect is separate from work quality: the bold and thin solid lines in Figure 2.3 show that the stability of positive and negative reputations increases similarly with closure. Consider two hypothetical bankers who work with ten colleagues this year. One works with colleagues segregated in the organization so they do not cite one another in the annual peer evaluations (illustrated by the sociogram at the bottom left in Figure 2.3). That banker would be over the ‘0’ on the horizontal axis in Figure 2.3. The second banker works with ﬁve colleagues who work together in one division and another ﬁve colleagues who work together in a second division (sociogram to the bottom right in Figure 2.3). The second banker would be over the ‘4’ on the horizontal axis. Even when both bankers do good work, it is the second banker’s work that will be remembered. The bold solid line in Figure 2.3 shows that a banker doing good work for colleagues not connected with each other can expect to be forgotten. The exact correlation expected between the banker’s reputation this year and next year is given by the level of the bold solid line over the ‘0’ on the horizontal axis. The correlation is indistinguishable from random noise. Bankers work with so many new contacts each year that their work is quickly forgotten – unless the colleagues with whom they work talk to each other. For the second banker, the one who worked with two groups of connected colleagues, reputation has an expected correlation of 0.57 over time. What carries a banker’s reputation into the future is gossiping colleagues. An implication is that you do not own your reputation. Rather, the people who own your reputation are the people in whose conversations it is built, and the goal of those conversations is not accuracy so much as bonding between the speakers (Burt, 2005, Chapter 4). You are merely grist for the gossip-mill through which colleagues strengthen their relationships with each other. Coleman (1988, p. S107) had it right when he opined that: ‘Reputation cannot arise in an open structure.’ Closure’s stabilizing eﬀect on reputation can be traced back to a stabilizing eﬀect on the individual relationships in which reputation is deﬁned. Closure creates an endogenous force for the status quo that secures and expands the boundary around a network, protecting new relations from

50

Overview

decay until they are self-sustaining. Figure 2.4 shows how this works by mapping decay against age for colleague relations between the bankers (see Burt, 2005, pp. 197–208, for details). Decay, on the vertical axis, is the probability that a relationship cited this year is not cited next year. Age, on the horizontal axis, is the years for which a relationship has been continuously observed. The lines show closure slowing decay. Bridge relations – that is, relations that span structural holes – almost all decay during their ﬁrst year. Ninety-one percent of bridge relations decay during their initial 13 months. Bridge relations have to survive on their own merit. There are no mutual attachments to keep unproductive relations in place. In contrast, when you and I have mutual colleagues, we keep bumping into one another even if we would prefer otherwise. Figure 2.4 shows that relations formed within a closed network – that is, relations new this year between bankers who have many mutual colleagues – decay relatively slowly. The bold decay line in Figure 2.4 peaks at 0.47 probability of decay in relations that have lasted 22 months.

A VIRTUAL ORGANIZATION I now wish to apply these ideas to a practical problem facing the leadership of a West Coast high-tech manufacturing organization at the turn of the twenty-ﬁrst century (hereafter, ‘the ﬁrm’). The ﬁrm was composed of four business divisions, each of which was doing well in its established product markets. However, the product markets were changing rapidly. Internet considerations were intruding everywhere. Small companies seemed to be chipping away at the ﬁrm’s markets with new products rapidly developed in response to emerging opportunities. Leadership was advised by an expensive external consultancy that the ﬁrm was missing lucrative opportunities that lay between the markets on which the four businesses were focused. For example, the consumer market was evolving to combine voiceover data in electronic data transmission along with video signal. This evolution is apparent today in digital broadcasting within companies and the home, and to handheld devices. However, at the time, the technology had alternative ways it could develop and customer channels were only vaguely coming into focus. Some leaders in the ﬁrm believed the consultant-deﬁned opportunities were less real than imagined. Moreover, the four businesses were doing well with their established products. Earnings would be less certain with less familiar products. Shifting a business division to go after the opportunity would be risky. Senior leadership faced the generic issue facing all organizations of size at the turn of the twenty-ﬁrst century: how to

51

Figure 2.4

Probability that relation decays before next year

0.0

0.2

0.4

0.6

0.8

1.0

0

1

2

3

(years)

8

9

Relationships embedded in network of one or more mutual colleagues

4 5 6 7 Relationship duration

Bridge relationships

Closure slows network decay, especially in new relationships

(relation cited this year is not cited next year)

10

One of the 25% most embedded relationships

52

Overview

harvest the eﬃciencies and growth of coordination across the enterprise made possible by technological advances without giving up productivity within the existing businesses. Rather than modify the ﬁrm’s structure, the leadership team decided to form a virtual organization, a cavalry unit that cut across the existing four businesses. The goal of the virtual organization was to identify and develop market opportunities that lay between the existing businesses. Prospects could be a new customer looking for a product that the ﬁrm was already shipping, an emerging market for which the ﬁrm could develop a lucrative product quickly from what was already going on in the ﬁrm, or an emerging market where the ﬁrm would have an advantage in creating new products from combinations of its existing technologies. A senior person from each business was given the job of recruiting people to the virtual organization. The sociogram at the top of Figure 2.5 describes the virtual organization as a network of discussion relations. These data were gathered about a year after the virtual organization had been launched. Each dot is a person. Dots with an X over them indicate the four people who were to recruit people to the virtual organization. The boundaries of the virtual organization were identiﬁed with snowball sampling. Each X-dot leader was asked to name the people on whom he or she most depended for collaboration in the virtual organization. The people named were asked who they most depended on, and so on. Shape and color indicate business division. For example, black squares indicate people drawn from business D. Product details are not needed for this illustration. I refer to the businesses as A, B, C and D. There was progress in the ﬁrst year. Most obviously, groups of people emerged. Moreover, the X-dot leaders did not try to own the virtual organization. The four X-dots in Figure 2.5A are not central in the discussion around them. For example, using ‘number of discussion partners’ as a centrality metric, the X-dot leader to the southeast of Figure 2.5A named two discussion partners. The two discussion partners are much more central, one has seven discussion partners and the other has 11. Similarly, the other X-dot leaders are connected to discussion partners more central than themselves. Visible progress notwithstanding, the ﬁrst year was unsatisfactory. The most obvious issue was that the virtual organization looked too much like the formal organization. The problem could have been driven by the way people were recruited; leaders had mobilized people in their own division and those people turned to colleagues they already knew well. There was little evidence of people making new contacts. Discussion partners at the top of Figure 2.5 had known one another for eight years on average, which

Network duality of social capital

53

was well before the launch of the virtual organization in the previous year. With respect to recruiting within one’s own division, notice that the shape and color of each X-dot leader always matches the shape and color of his or her discussion partners. Note also the two structural holes in the virtual organization that correspond to boundaries between businesses in the formal organization. People from business A (white circles) are concentrated in a cluster to the northwest of Figure 2.5A. Two people from the business are network brokers into the adjacent cluster. On the other side of the virtual organization, to the southeast in Figure 2.5A, there is a cluster of people drawn from business D (black squares). They are connected into the adjacent cluster by one network broker in business C (white square). The evidence of myopia – people focused on familiar colleagues within their own division – was troubling: eﬃciencies and opportunities across business units were being missed. Stories had come back from the ﬁeld about the groups stepping on one another’s toes, and the toes of the established businesses, in presenting customers with multiple, contradictory images of the company. It was disconcerting to see people respond to the new initiative by turning to the same people they had turned to in the past. People did not seem to ‘get’ the virtual-organization strategy. As an intervention to facilitate the virtual organization, the network analysis in Figure 2.5A was used to identify key people to send to workshops on managing informal organizations. Always in the background was the fact well known among managers that company leadership was unhappy with progress made in the ﬁrst year. After another year, the virtual organization looked very diﬀerent, as you can see in Figure 2.5B. More people were involved in the virtual organization, but what is striking is the coordination across businesses. People became more connected, and more connected with new acquaintances. Path distance is the shortest number of links required to connect two people in a network: discussion partners are one link distant, friends of friends are two links distant, and so on. After the ﬁrst year in operation, people in the virtual organization were separated by four and a half links. At the end of the second year, people on average were separated by three and a half links. People were working more often with new contacts. In the ﬁrst year, people cited discussion partners they had known a long time (eight years on average). At the end of the second year, that number had been cut in half: discussion partners had known each other for four years on average. Cutting the average in half means that people made a lot of new contacts.

54

Figure 2.5

Discussion network in a virtual organization

55

Figure 2.5

(cont.)

56

Overview

EVALUATION Did the changes in Figure 2.5 strengthen the virtual organization? The short answer is yes; yes with respect to brokerage, closure and senior opinion. With respect to brokerage, connections spanned the structural holes that previously Balkanized the virtual organization, and people were connecting with new acquaintances. After two years in operation, a sense of growth and optimism developed as people worked on resolving diﬀerences in opinion and business practice across the company. With respect to closure, increased connections across the virtual organization created the sense that people more often recognize one another as participants in the virtual organization, a special group apart from usual businesses. Positive relations developed in the course of working through previous diﬀerences in opinion and practice. Reputations developed for the people active in the virtual organization. Maintaining reputation within the virtual organization became its own incentive to work harder to make the organization a success. Senior opinion was quite positive about the change displayed in Figure 2.5. Attention shifted to more pressing issues. New People More than New Network Further analysis – out of the limelight of senior management politics – revealed a dark side to the change. As an external consultant to the project, I was initially struck by the level of turnover between the two years. I described the contrast between Figures 2.5A and 2.5B in terms of people connecting across the businesses. In fact, the network did not change so much as the people changed, and the new people established a new network. Of 88 people active in the ﬁrst year of the virtual organization, 37 continued through the second year – less than half. The other 51 returned to their regular jobs in the businesses. That means 67 new people entered in the second year (104 total in Figure 2.5B minus 37 continuing). The turnover seemed high: 58 percent of last year’s people left, 64 percent of this year’s people are new. High turnover need not indicate instability. If the people most connected during the ﬁrst year continued to be the people most connected through the second year, then the virtual organization would stay on course despite massive turnover among people marginally involved in the organization. However, turnover ran right to the center of the virtual organization. It was so high that leadership in the ﬁrst year could not predict leadership in the second year. Among the people with above-average numbers of discussion partners in either year, their number of discussion partners in the ﬁrst year had no correlation with discussion partners in the second year (r = 0.01).

Network duality of social capital

57

Some of the people central in the ﬁrst year were marginal in the second. Some of the new entrants were among the most central at the end of the second year. This is not a quality issue. People active in the ﬁrst year were no more or less able than the people active in the second year. I suspected when I ﬁrst noticed the turnover that able people might be avoiding the political bother of the virtual organization and less-able people were ﬁnding refuge there, or new hires were being assigned to the virtual organization as a temporary assignment until a permanent position was deﬁned. However, leavers, stayers and entrants were similar kinds of people (51, 37 and 67 people respectively): They were the same age on average (1.82 F(2,152), P ~ 0.57). There are no statistically signiﬁcant diﬀerences in their years with the ﬁrm (2.30 F(2,152), P ~ 0.10), nor in their relative numbers joining the ﬁrm within the previous year (1.91 chi-square, 2 d.f., P ~ 0.39). Job evaluations were no diﬀerent across the three categories (1.92 chi-square, 4 d.f., P ~ 0.75, for annual evaluations distinguishing poor, good and excellent work). In short, the people leaving, staying or entering were comparably experienced and able. Nevertheless, it was clear that what seemed to be change in connections between people was more precisely change in the people connected. Premature Consensus on Good Ideas A second, more serious concern emerged after the second year. People active in the virtual organization seemed to come to premature consensus on good ideas. Beyond my impressions from listening to conversations between people active in the virtual organization, I have two indicators of premature consensus: one indicator of consensus, the other of consensus premature. This evidence is not conclusive, but it is consistent with the eﬀects of network closure illustrated in Figures 2.3 and 2.4. An indicator of consensus is the link that emerged, in the third year, between the virtual organization and budgets for interdivisional new products. The virtual organization had no budget. In contrast, ﬁnancial support could be solicited from the research and development (R&D) budget to facilitate the development of speciﬁc new products that involved more than one of the businesses. Coincident with launching the virtual organization, seven of several new-product proposals were funded at various levels depending on need and promise. Seven small teams of people worked on each of the seven new products. Company support for each team was visible when team budgets increased or decreased from one year to the next. The teams varied in their connection into the virtual organization. Team members were sometimes mentioned as discussion partners by the people

58

Overview

active in the virtual organization. Some were mentioned often. Many were never mentioned. I computed an ‘integration’ index for the seven teams by counting the number of times that team members were named as discussion partners by people in the virtual organization, named in year one (Figure 2.5A) then a second index of how often team members were named in year two (Figure 2.5B). For the ﬁrst year of the virtual organization, there was no association between new-product team integration into the virtual organization and the extent to which the team budget increased or decreased going into the second year. The diﬀerent market groups had diﬀerent perspectives on what would work in reaching the market, and the new-product teams were variably connected with people in the four market groups. The virtual organization came together in the second year as illustrated in Figure 2.5B, and during that year new-product team connections into the virtual organization became strongly associated with budget change. Three new-product teams contained individuals often mentioned in year two as discussion partners by people central in the virtual organization. Budgets for these ‘strongly connected’ teams increased from year two to year three by an average of 72 percent. Three teams were sometimes mentioned for discussion in the virtual organization. The budgets for these ‘somewhat connected’ teams increased by an average of 18 percent. One new-product team was composed of people never mentioned as discussion partners in the virtual organization. The budget for the one ‘social isolate’ team was decreased by 70 percent. I have few data here. There are only seven teams. But the pattern is clear: the more your team is connected into the virtual organization, the better your funding. My second indicator – anecdotal evidence of consensus premature – concerns successes attributed to the virtual organization. I checked with a company executive a year after Figure 2.5 to learn what had come of the experiment. The virtual organization was still operating. It was credited with two successes: a new customer for an existing company product and a new product that combined two existing products from one of the divisions. These are successes, but neither was a new product across divisions. In fact, as the executive explained: ‘The businesses were very involved in winning the two contracts since the work was going to be done in the businesses [virtual organization had no employees], so labor and facilities had to be secured. If you were to ask the businesses, they would probably claim credit for the wins.’ Was consensus in the virtual organization a little too early? The virtual organization made worthy contributions, but the contributions were closely related to activity already ongoing in the established busi-

Network duality of social capital

59

nesses. People central in the virtual organization were under time pressure because the operation had been running for a year without result. There is no certain answer to the question, but similarity between the virtual organization’s contributions and activities already ongoing in the established businesses raises a question: How far beyond the consensus reached in the virtual organization was a mix of study-ﬁrm technologies that would constitute a genuinely new product? Moral of the Story Consensus is a good and necessary thing, but coming to it prematurely can freeze out superior courses of action. Tentative trials with alternative courses of action is essential to lower risk in business exploration. The problem I see in the described virtual organization is too much focus on ﬁnding common ground. It is naive to think that subject-matter experts are also expert in exchanging ideas across previously segregated subjects. Brokerage experience matters (Burt and Ronchi, 2007). Employees in the study ﬁrm were accustomed to life in a corporate hierarchy. In the virtual organization they established consensus within a hierarchy. So much attention was given to building bridges across divisions and market groups that too little was paid to preserving the diﬀerences that were a competitive advantage in the target markets. The people involved were able. But people with diﬀerent backgrounds, tackling a shared task, often get excited about ﬁnding common ground. They can lose sight of the fact that preserving their unique strengths is what makes it productive for them to get together. First-year participants in the virtual organization were connected to the established businesses. Second-year participants were connected to one another. In the excitement of connecting across the separate market groups, and so across the ﬁrm, people in the virtual organization came to consensus about how to go after markets not already targeted by company businesses. As experts in new-market business, their consensual opinion informed the allocation of funds to new-product teams, illustrated by team budget increases closely associated with connections to people in the virtual organization. Consensus in the study-ﬁrm virtual organization narrowed the variety of new products pursued and overlapped to a great extent with ongoing activity in the established businesses. The moral here is succinctly embodied in a business mantra I ﬁrst heard in IDEO, America’s premier design ﬁrm: ‘Fail often to succeed sooner’ (Kelly, 2001, p. 232). We learn by trying alternatives. To do that, you need alternatives and the ability to pursue them.

60

Overview

THE MORE GENERAL POINT Every act of brokerage is implicitly an act of closure. Every step taken to coordinate across a structural hole to try something new increases closure, locking you into a course of action. The competitive advantage by which social networks constitute social capital is a balancing act between brokerage and closure. Social capital is an intersection of two functionally distinct networks: a ‘diﬀerentiating’ network in which people are distinguished by skills or resources, and an ‘embedding’ network in which people with complementary skills or resources are brought together to better pursue their interests. Brokerage is about positioning bridge relations in the embedding network to span structural holes in the diﬀerentiating network. Closure is about reinforcing bridges in the embedding network to harvest bridge value. The task for business and civic leaders is to strike a balance between brokerage and closure. They must decide whether the time is right to tighten connections within the group to obtain the trust and eﬃciency beneﬁts of a closed network, or build bridges beyond the group to obtain the innovation and growth beneﬁts of a brokerage network. Imbalance deﬁnes the four failure modes in Figure 2.6. The columns distinguish situations in which people are thinking about building social capital from situations in which they have taken successful action, either by establishing a beneﬁcial bridge between two previously disconnected groups (brokerage) or by closing the network around a community such that members feel a sense of identity and reputation within the community (closure). Across the ﬁrst row in the ﬁgure, too much brokerage will erode coordination into a chaos of inconsequential personal opinion. Before action is taken, this failure mode is apparent in a lack of eﬀective action despite people trying. Eﬀort is moving in too many diﬀerent directions, or discussion is in so many directions that it stymies eﬀective collaborative eﬀort in any direction (for example, Burt, 2005, pp. 240–44). After a successful bridge has been developed, this failure mode is apparent in people squandering the success of the bridge by again moving in too many diﬀerent directions. Across the second row in the ﬁgure, too much closure locks the organization into rigid adherence to past practice, group-think, and people passively waiting for orders (see Burt, 2005, Chapter 4, for review). Before action is taken, this failure mode is apparent in people having an exaggerated sense of themselves (hubris, which thrives on the lack of data typical in a closed network) and searching through a narrow range of alternatives. Elements of this failure mode were illustrated by the virtual organization in Figure 2.5. After a community is successful, the too-much-closure failure mode is apparent in people resisting ideas inconsistent with their previous

61

Figure 2.6 Network duality failure modes

Too Much Closure (gossip-induced hubris, group-think, rigidity and stereotyping)

Too Much Brokerage (chaos of new ideas and agency problems)

Premature Consensus (narrow search, risk local maximum)

No Consensus (no effective action, dissipated resources)

Before Action

Dogmatic Consensus (rigid boundaries, hard edges to structural holes)

No Consensus (ineffective action, dissipated success)

After Action

62

Overview

success. A hardened social shell forms around the community, protecting them from ideas not invented here. I focused on the second row of Figure 2.6 in this chapter because it is the more likely failure mode for the many management initiatives intended to strengthen coordination across an enterprise. Structural holes allow diﬀerences among us to develop, which are the foundation for tomorrow’s new ideas. Hagel and Brown (2005, p. 87) discuss the importance of having company processes by which people exchange opinion and practice as expected in the Figure 2.5 virtual organization: ‘Productive friction occurs when people with diverse and appropriate specializations creatively resolve diﬃcult business issues. But to gain its full beneﬁt, companies must also establish processes . . . to help them reflect on the practices emerging from these collaborations, recognize patterns, and increase awareness of high-impact solutions.’ There are two management strategies for ﬁnding the balance between brokerage and closure. One is to shift artfully between brokerage and closure, implementing suﬃcient brokerage to prevent premature consensus from closure followed by suﬃcient closure to ensure that something concrete results from the good ideas provided by brokerage. This strategy is well suited to smaller groups such as work teams and project groups in which leaders can keep a ﬁnger on the emotional pulse of the group so as to avoid the failure modes in Figure 2.6. In large organizations, shifting between policies of brokerage or closure can be confusing to employees and unwieldy to implement. The failure modes in Figure 2.6 are likely. A second strategy, especially appropriate to larger groups, is to take away the monopoly rights of the embedding network in favor of rights more flexibly tied to the underlying diﬀerentiation network (Starbuck, 1996; Mayer-Schönberger, 2007). The virtual organization in the study ﬁrm is an example of a monopoly embedding network. Employees were encouraged to propose new products where they saw value, but getting senior attention could be diﬃcult. The virtual organization was the company-approved vehicle for proposing new products across divisions. Traditional examples of monopoly embedding networks are churches, corporations, guilds, marriage or government more generally. These are embedding organizations in the sense that they provide rules and obligations for the conduct of interpersonal relations. They are also jealous organizations. They do not typically encourage aﬃliations with competitors. Your priest would be surprised to meet your rabbi. Your employer would be unhappy to discover your job working evenings and weekends with another employer. Government presumes control over all activity on its sovereign soil. Flexible embedding networks are emerging rapidly in response to coordination issues created by the global connections possible with the wireless

Network duality of social capital

63

Internet. The general idea is to harvest the trust and collaboration beneﬁts of closure’s reputation mechanism without incurring the rigidity of monopolistic closed networks. For example, multiple embeddings can compete as coalitions. The virtual organization in the study ﬁrm consisted initially of three groups (Figure 2.5A). Each group competed as a coalition recruiting participants. There were three network brokers who were active in two groups, but most employees were active in a single group. Instead of creating closure across the groups (Figure 2.5B), senior leadership could have preserved the separate groups but focused on ensuring that they learned from one another’s trials and that each was aware of progress made in the others to maintain competitive reputation pressure between the groups. The ﬁrm would thus function like a standards-setting body that deﬁnes rules for routine exchange between the competing coalitions. GSM played such a role in the mobile phone market (originally Groupe Spéciale Mobile, now Global System for Mobile Communication). Launched in 1982 for communication across Europe and taking oﬀ in 1992, GSM has been adopted as an interoperability standard by most of the world. Focused on domestic communication, the US did not adopt, choosing instead to live with a variety of competing standards not compatible with GSM. The GSM standard meant that ﬁrms such as Ericsson, Nokia and Samsung could invest in new features expecting returns from a large GSM market across countries. The result: mobile phones outside the US are richer in exciting capabilities. Apple adopted the GSM standard for its feature-rich iPhone, which limits initial US distribution in favor of distribution in the larger global market. A more extreme approach to flexible embedding is for senior leaders to create a physical or virtual ‘common’ space in which employees sketch ideas in search of collaborators elsewhere in the ﬁrm. This approach is extreme in that it is meant to harvest the value of interpersonal collaboration with minimal intrusion from a governing organization (a social network version of Saltzer et al.’s, 1984, end-to-end argument for computer network design). The corporation functions here as a security force to protect open communication while employees play in the common space. This imagery can be seen in safe-harbor regulations that limit liability on the condition that action is taken in good faith. For example, the Federal Communications Commission facilitates open communication by deﬁning a safe-harbor period from 22:00 to 6:00 for the broadcast of material indecent for children. The Digital Millennium Copyright Act facilitates open communication by deﬁning safe-harbor provisions that limit Internet service provider liability for copyright infringement by users. The imagery of an intellectual commons or safe harbor also can be seen in corporate policies that authorize employees to spend a percentage of their work-week developing ideas independent of their assigned tasks. Able

64

Overview

employees do not require authorization to be imaginative, but can require a company policy to counterbalance a supervisor impatient for the employee to complete an assigned task. The above are progressive solutions. Monopoly solutions are more common, like the virtual organization in the manufacturing ﬁrm in Figure 2.5. In this, eﬀorts to build social capital put me in mind of early tobacco farmers in Virginia. Tobacco plants drain soil nutrients such that it is diﬃcult to grow crops on the same land for many years. Farmers, chasing the proﬁts of increasing English demand for tobacco, worked a plot of land until it was destroyed, then moved to a new plot and destroyed the new land (for example, Kulikoﬀ, 1988, pp. 47–54). Similarly, eﬀorts to harvest social capital by bridging structural holes can eliminate the diversity that made the bridges initially valuable. Like Virginian tobacco farmers, managers focus on the short-term beneﬁt of building bridges across the structural holes between groups. The long-term cost of farmers draining the soil, or managers eliminating diﬀerences in perspective and practice, is borne by the sovereign organization and future residents. Senior people would be wise to remember Jean-René Fourtou on managing le vide, the structural holes, in organizations and markets – some instances of le vide should be preserved as catalysts for future advantage.

NOTE 1. I am grateful to the University of Chicago Graduate School of Business for ﬁnancial support of work on this chapter. Portions of this chapter were presented at the 2006 Academy of Management meetings and the ‘Interdisciplinary Conference on Social Capital’ organized by Viva Bartkus at the Mendoza College of Business at the University of Notre Dame. Second-year network data on the virtual organization described here were gathered by a colleague who preferred to remain anonymous. I appreciate him sharing the data with which I compare the organization before and after the network intervention. The text was improved in response to comments by Viva Bartkus, Philip Evans, Don Ronchi and Toby Stuart.

REFERENCES Burt, Ronald S. (1992), Structural Holes: The Social Structure of Competition, Boston, MA: Harvard University Press. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Capital, New York: Oxford University Press. Burt, Ronald S. and Don Ronchi (2007), ‘Teaching executives to see social capital: results from a ﬁeld experiment’, Social Science Research, 36 (September), 1156–83. Christensen, Clayton M. (1997), The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail, Boston, MA: Harvard Business School Press.

Network duality of social capital

65

Coleman, James S. (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94 (Supplement), 95–120. Grabher, Gernot (1993), ‘The weakness of strong ties: the lock-in of regional development in the Ruhr area’, in Gernot Grabher (ed.), The Embedded Firm: On the Socioeconomics of Industrial Networks, London: Routledge, pp. 255–77. Hagel III, John and John Seely Brown (2005), ‘Productive friction’, Harvard Business Review, February, 82–91. Kelly, Tom (2001), The Art of Innovation: Lessons in Creativity from IDEO, America’s Leading Design Firm, New York: Doubleday. Kulikoﬀ, Allan (1988), Tobacco and Slaves: The Development of Southern Cultures in the Chesapeake, 1680–1800, Chapel Hill, NC: University of North Carolina Press. Mayer-Schönberger, Viktor (2007), ‘Useful void: the art of forgetting in the age of ubiquitous computing’, Harvard University, Kennedy School of Government Working Paper RWP07-022. Saltzer, Jerome H., David P. Reed and David D. Clark (1984), ‘End-to-end arguments in system design’, ACM Transactions in Computer Systems, 2, 277–88. Saxenian, AnnaLee (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Boston, MA: Harvard University Press. Slater, Robert (1999), Jack Welch and the GE Way, New York: McGraw-Hill. Starbuck, William H. (1996), ‘Unlearning ineﬀective or obsolete technologies’, International Journal of Technology Management, 11, 725–37. Stewart, Thomas A. (1996), ‘The great conundrum – you vs. the team’, Fortune, November, 165–6.

3.

On the costs of conceptualizing social ties as social capital Robert M. Fishman1

I oﬀer in this chapter a friendly, yet emphatic, critique of a term – and concept – that serves to animate and draw attention to much work that I admire and from which a great deal can be learned. In the pages that follow I elaborate a series of rather substantial and interrelated costs of conceptualizing social ties as ‘social capital’ (hereinafter, SC), costs that – I contend – impinge upon our ability to understand empirical reality. This critique is intended to be friendly, for I thoroughly share the commitment of SC theorists to identify and account for positive outcomes that can be attributed to social ties and tie-supporting norms. Indeed, in the closing section of this chapter, I oﬀer a theoretical claim intended to enhance our ability to locate just such positive eﬀects of social connections, but I propose to do so in a way that underscores diﬀerences, rather than commonalities, in the causal impact of social relations and norms. I argue that despite the admirable quality of much work done within the SC framework, this school’s conceptualization and labeling of useful social ties, and related norms, SC detracts from the collective explanatory eﬀorts of social scientists and introduces unnecessary confusion into scholarly discourse. I acknowledge that the term draws attention to important ﬁndings but argue that such attention comes at a price and that it detracts from social scientists’ pursuit of their most fundamental goals: explanation and conceptual understanding. The critique developed in this chapter represents an elaboration of an argument I ﬁrst oﬀered in Democracy’s Voices,2 a book that argues for the ability of boundary-crossing social ties and conversations to reshape public life in a way that enhances democratic quality. In that book I examine – and reject – the possibility that the SC approach provides a more useful framework than the analysis of particular types of ties for exploring the impact of social connections and conversations. In this essay I extend that earlier argument by reviewing evidence from recent scholarship, taking up the loose metaphor-only version of SC scholarship, and oﬀering a positive case 66

Costs of conceptualizing social ties as social capital

67

for the beneﬁts of conceptualizing useful social ties through at least two analytical prisms – rather than just one. I ask readers to accept just one fundamental methodological and theoretical assumption: that the mere formulation and elaboration of a concept does not by itself establish whether it usefully reflects – or promotes the study of – underlying empirical reality. Indeed, the scholarly eﬀort to determine whether concepts are useful or counterproductive stands as an important task for social scientiﬁc progress. This assumption, and the related claim (which readers may or may not fully accept) that concepts in and of themselves are not real things, but rather intellectual constructs that must be judged precisely on the basis of their usefulness for answering meaningful questions about reality, represent a constitutive element of the Weberian approach to social science to which I subscribe. Although I have avoided the use of quotation marks around the term SC in all but a few places, I do ask readers to leave open the question of whether the term is a fruitful or confusing way of understanding the empirical realities that we all seek to capture, and in that sense I invite readers mentally to place quotation marks around the term. I argue that unlocking the enormous (but multifaceted) causal force of social connections – as modern sociology, much contemporary political science, scholars in other ﬁelds and theoretical precursors such as de Toqueville have sought to do – requires us to diﬀerentiate among types of connections and types of eﬀects. There is a strong basis in scholarly works, both classic and contemporary, for distinguishing among types of ties – and their eﬀects. Indeed, empirical studies are replete with instances of precisely such diﬀerences, the most widely recognized modern classic on this point being the pioneering scholarship of Mark Granovetter on the dissimilar labor market impact of strong and weak ties.3 The broader importance of searching for the dissimilar eﬀects of social relations is nicely captured by Barry Wellman’s compelling formulation: ‘Which types of ties and networks give what kinds of social support?’4 Yet despite abundant evidence that diﬀering forms of social connection do generate quite dissimilar consequences, the concept I critique may encourage scholars – and readers of their work – to conceive of social connections and tie-supporting norms as constituting one undiﬀerentiated ‘stock’ or amount of SC that can be measured by a single metric and that exerts one overarching eﬀect characterized by its underlying similarity in disparate arenas or settings. One purely terminological issue merits brief attention: the use of the term SC typically rests on a series of distinctions that for some, such as Elinor Ostrom in her stimulating formulation in this volume, begin with ‘physical capital’, which is taken to represent the most common sense in which the term ‘capital’ is employed. However, it is not clear that the

68

Overview

conventional understanding of capital can be fully reduced to ‘physical capital’. Capital in its standard sense embraces some items – such as ﬁnancial instruments – that have a value not primarily reflective of their qualities as physical things. However, such items, along with physical resources, do have an economic exchange value, thus providing capital in its basic sense with an important property, fungibility,5 which allows us to measure meaningfully the total amount of capital possessed by an individual or collective entity. For these reasons, I prefer to understand capital in its most conventional sense as economic, rather than simply physical, capital. Scholars have formulated a series of other ‘capitals’ – cultural, social, political and more recently spiritual – as theoretical tools thought to promote scientiﬁc explanation, but whether these conceptual adaptations of capital really do facilitate our ability to identify and understand causal dynamics in spheres other than the economic market requires serious attention, which is precisely the point of this discussion. This chapter’s critique is not equally applicable to all currents within the SC approach. I ﬁrst examine both the allure and the signiﬁcant pitfalls of the most extreme, or ‘more-than-metaphor’, version of SC theory which takes social connections and related norms to be strongly similar to economic capital. This perspective assumes that individuals and social aggregates – be they ﬁrms, local communities, countries or some other collective entity – possess a total ‘stock’ or amount of SC that can be meaningfully measured and which exerts a generalized positive eﬀect on the pursuit of myriad objectives. Some, but not all, scholars operating within this framework assume that SC, like economic capital, should be understood as an investment that is made by actors instrumentally pursuing the positive beneﬁts that can be predicted to follow from the growth of their total stock of SC. If valid, this approach would vastly facilitate the study and understanding of social ties and trust, for it oﬀers the promise of providing one uniﬁed metric and conceptual lens for studying a wide array of eﬀects that scholars have studied separately. Yet by the same token, if not valid, this approach presents scholars with major impediments to the pursuit of their explanatory objectives. Secondly I examine the appeal and the much more limited costs of adopting a looser ‘metaphor-only’ understanding of SC. This perspective takes social ties and trust to be similar enough to economic capital to render the term SC suggestively useful, yet dissimilar enough to require a conceptual apparatus and methodology fundamentally unlike those appropriate for the study of economic capital. This approach does pursue some useful distinctions among types of ties and eﬀects. If we could easily limit the use of the term and concept to this loose metaphor-only sense in which one argues that the causal force of social connections or norms is somewhat

Costs of conceptualizing social ties as social capital?

69

(but not precisely or measurably) akin to economic capital, many of the objections presented in this essay would vanish. But such a restriction is unlikely to prevail, even though some scholars clearly do limit their own use of the term to this diﬀuse sense. By its very nature, the metaphor of capital, when used to conceptualize the causal force of social connections, encourages those who employ it to pursue a strong – more than mere metaphor – theory.6 Moreover, I argue that even the loose metaphor-only sense of SC, despite its surface appeal, can generate unnecessary confusion and may actually distract our focus from some important explanatory challenges. Thirdly, I present a theoretical rationale and empirical support for examining the causal impact of social connections through more than one conceptual prism, and I argue that the study of social ties and their causal impact proves most fruitful precisely when such analytical distinctions are pursued. Finally I place this critique within the theoretical perspective that animates it, namely the Weberian approach to social science.7

THE MORE-THAN-METAPHOR SOCIAL CAPITAL APPROACH AND ITS PITFALLS If the eﬀects of social connections, and norms supportive of such connections, can be usefully understood as constituting one more or less cohesive and measurable underlying phenomenon that proves relevant for analyzing disparate human endeavors, the theory and terminology of social capital are thoroughly validated. If this condition held, it would be preferable to view speciﬁc types of ties and associated norms as mere manifestations – and components – of a deeper underlying reality, namely the total pool of SC held by an individual or collective entity. Much scholarship makes essentially this assumption, and treats speciﬁc types of social relations and/or trust as mere data points in the operationalization of what is taken to be a meaningful and more general phenomenon conceptualized as total SC and represented in one summary measure. Network ties, associational membership, attitudes of trust and much more are used to calculate a generalized measure of SC which, in the extreme, may then be viewed by some scholars as more real than any of its components taken on their own. This approach – which essentially forfeits the opportunity to examine the speciﬁc causal impact of each type of tie, and instead calculates a summary measure of total SC – poses numerous problems. These problems include methodological and theoretical drawbacks as well as the overwhelming weight of a great deal of empirical evidence on the wide range of variation in the causal dynamics manifested by diﬀering social ties and norms. Perhaps the most fundamental diﬃculty concerns the viability of

70

Overview

constructing one summary measure of total SC, as practitioners of the more-than-metaphor approach frequently attempt. If a summary measure is to be calculated we require some uniﬁed metric for assigning weights to all those social relations and norms held to contribute to SC. In the absence of such a uniﬁed metric it is diﬃcult to see how we can meaningfully assign actors or collective entities with a total score, an overall stock, of SC. In addressing this point we encounter a fundamental contrast between social relations and economic capital. The economic resources of a ﬁrm or individual can be assigned monetary values allowing one to construct an overall measure of value. Granted, even economic capital proves diﬃcult to measure in a fully satisfactory way: some of the resources possessed by economic actors are more readily exchangeable than others, and measures of a ﬁrm’s worth – such as its capitalization and its book value – may diverge substantially from one another. Yet despite all the diﬃculties involved in assessing a ﬁrm’s or an individual’s total stock of capital, it remains the case that monetary values and market exchanges do provide us with a plausible basis for constructing summary measures of wealth and value. There is no plausibly equivalent metric for assessing the total ‘social wealth’ of an individual or a collective actor. One cannot exchange family ties for professional connections, or the reverse, and trust cannot be bought, sold or mechanically exchanged. Yet is there another available metric for assigning weights to particular social connections or norms, in the eﬀort to construct a summary measure of SC? Is there any alternative to relying on exchange value, an approach which appears to be thoroughly implausible given that in the modern world social relations are not exchangeable on the market? Two alternatives appear worthy of some consideration, but they lead us in rather diﬀerent directions. We could assess the contribution of speciﬁc social connections to an actor’s presumed total stock of SC by attempting to ascertain the exchange value of information or resources obtained through the tie in question. Alternatively we could assess the opportunity cost of building and sustaining the same ties. The time and eﬀort involved could conceivably be captured through focused research instruments. However, this dual strategy does not prove to be compelling, for the two approaches would often yield quite diﬀerent scores, thus placing in question the validity of both. The value of resources which can be said to flow through ties is only imperfectly (and often indirectly) related to the energy and intentions involved in constructing those ties; many positive rewards that flow to those with (certain types of) social ties have been shown to be second-order consequences or by-products of social relations established for other reasons. The seeming impossibility of generating a meaningful summary measure of ‘total SC’ raises serious doubts about the conceptual commitment to

Costs of conceptualizing social ties as social capital?

71

viewing social relations and norms as constituting one undiﬀerentiated overall amount. Yet this is not the most weighty objection to the view that individuals and collective entities hold an overall stock of SC. Perhaps the greatest problem posed by this approach concerns its failure to capture the multiple causal logics manifested in the diverging consequences of diﬀerent types of social ties. Given that the components used to calculate an overall measure of SC generate quite diﬀerent causal outcomes – and often fail to covary with one another – as much research shows, it follows that in this instance the explanatory power of the ‘whole’, which is to say any given summary measure of SC, is actually lesser than that of the sum of its parts. As Stephen Smith and Jessica Kulnych note in their critical essay on SC: ‘the terminology of social capital oversimpliﬁes the character of such relationships and actually obscures the vast diﬀerences in their eﬀects’.8 Trying to capture the causal impact of particular types of social relations through a conceptual lens that focuses in undiﬀerentiated fashion on the sum total of all social relations is highly counterproductive. This is why some SC theorists do introduce distinctions within the large set of phenomena they label with this term, and when they do so the explanatory beneﬁts of such distinctions become clear. For example, in Chapter 4 of this book, John Helliwell and his collaborators Haifang Huang and Robert Putnam ﬁnd that domain-speciﬁc measures of trust oﬀer stronger predictions than a generalized trust measure in the search for determinants of workplace well-being.9 Moreover, the domain-speciﬁc variables diﬀer in their explanatory power, with workplace trust holding the greatest predictive power. The number of types of social connections and related norms (such as trust) that have been presented by scholars as instances of SC – and that have been included in general measures of total SC – is quite large; the list of component indicators varies quite substantially from one researcher to another. In light of this vast diversity it should come as no surprise that SC researchers have often encountered substantial dissimilarities across time and cases in the level obtained by the alleged components of total SC and in the directionality of their change. For example, trust may decline while associational membership holds steady, and other disparities in the evolution of phenomena thought to form part of SC have also been found.10 Furthermore, some seemingly straightforward indicators, such as associational membership, may mask causally relevant distinctions among changing forms of group membership in contemporary societies.11 The degree of interaction among members of associations is highly variable. If all SC indicators exerted exactly the same causal impact – or if the explanatory power of each such indicator could be assigned a given weight that

72

Overview

remained constant for all those causal processes we wish to understand – diﬀerences in the pattern and direction of change in SC components would not be of major scientiﬁc concern. Yet that is not the case. The diﬀering causal consequences of diverse types of ties and norms render highly questionable any summary measure that rests on components that vary at least somewhat independently of one another. The more-than-metaphor understanding of SC also encourages some scholars to conceptualize all useful ties and norms as the fruit of instrumentally oriented investment-like behavior thought to have established and sustained those ties and norms. This analytical prism seems to hold considerable logic if one adopts the SC framework as one’s approach for studying beneﬁcial eﬀects of social relations. After all, if social connections and trust are simply one form of capital, why not assume that they rest on the same sort of calculations that underpin the growth of economic capital? Yet even one of the founding theorists of the SC school, James Coleman, argued explicitly against the assumption that the usefulness of social relations need necessarily reflect prior eﬀorts and calculations speciﬁcally designed to produce such beneﬁcial eﬀects. The evidence against this assumption is too strong to sustain such a claim, at least in its most extreme form. Indeed, many useful consequences of social relations have been shown to be by-products of connections initially established for other reasons, thus rendering the instrumental-investment prism insuﬃcient (but not always inaccurate) as a guide to understanding fully the force of social relations. The decision to label social relations as a form of capital and to discuss social dynamics in the language of economics tends to privilege an analytical prism that is helpful for understanding some – but not all – useful ties and norms. Enthusiasts of the more-than-metaphor usage of SC might well respond that the pitfalls outlined here are all avoidable. Surely it is true that many SC theorists are aware of at least some distinctions to be made in the eﬀects that flow from ties. Moreover, numerous SC scholars explicitly avoid the restrictive instrumentalism-only view of social relations. In his monumentally important study of the decline in civic engagement and sociability in the United States, the distinguished SC scholar Robert Putnam does diﬀerentiate between ‘bridging’ and ‘bonding’ SC, thus oﬀering a case in point of the openness of some within this school to the study of diﬀerences among types of ties.12 Nonetheless, the more-than-metaphor approach, by its very nature, encourages scholars to calculate and discuss summary measures of total SC, as Putnam himself does after introducing that distinction. Some SC scholars may exclude certain types of social linkages from their summary measures – and conceptualizations – as in the case of Putnam’s convincing exclusion of vertical clientelistic ties in his study of

Costs of conceptualizing social ties as social capital?

73

the civic foundations of eﬀective democracy.13 Yet in a sense, speciﬁc exclusions serve only to underscore the broader case against a summary measure composed of dissimilar components. To a somewhat lesser degree, the strong version of SC theory may also encourage scholars to embrace fully the exclusively instrumentalist view of social relations that lies closest to the analytical prism appropriate for understanding economic investments. Thus in various ways, scholarly eﬀorts and attention are diverted from the necessary search for explanations resting on distinctions among social connections, such as the diﬀerence between ties that are constructed and valued for instrumental reasons centered on the external consequences of the social relations in question and other ties that participants establish and take to be meaningful because of their intrinsic qualities.

THE METAPHOR-ONLY UNDERSTANDING OF SOCIAL CAPITAL Many if not all the objections I have presented thus far would appear to diminish in magnitude or even disappear entirely if references to SC could be understood to constitute nothing more than a suggestive but quite inexact metaphor. This is the approach taken in practice by those SC scholars who diﬀerentiate among types of social relations in their analysis while also eschewing any eﬀort to calculate a total pool of SC. Ronald Burt in his sociological tour de force, Brokerage and Closure,14 is quite explicit in his adherence to such a metaphor-only understanding. Indeed he notes that ‘the concept begins as a metaphor about advantage’.15 Instead of attempting to theorize an underlying – and measurable – causal unity encompassing all instances of social relational advantage, he focuses on two structurally distinct forms of such advantage that he labels ‘brokerage’ and ‘closure’. On the basis of a thorough review of empirical evidence and a wide-ranging theoretical synthesis, Burt shows how two very diﬀerent forms of network-based advantage – leading to diﬀerent types of beneﬁts – rest on: (1) facilitating a connection between otherwise unrelated social worlds (that is, brokerage); and (2) strengthening social relations within relatively closed worlds (that is, closure). This theoretical framework proves highly successful in identifying and making sense out of important forms of advantage rooted in structured social relations. Moreover it avoids the (counterproductive) attempt to calculate an overall stock of socially rooted advantage possessed by individuals or collective actors. Does this metaphor-only approach succeed in avoiding the pitfalls of the more-than-metaphor usage? Certainly, to some degree the answer to that

74

Overview

question must be aﬃrmative, but that does not mean that this looser understanding of SC is without its own costs. Burt articulates his SC framework around precisely the sort of distinction among types of ties and eﬀects that this chapter advocates – and which would be masked by limiting analysis to one undiﬀerentiated conception of all useful social connections. But by placing the discussion of brokerage and closure on the terminological terrain of SC his analysis may inadvertently create conditions that inhibit the collective pursuit of scientiﬁc clarity and accumulation; the metaphor of capital, based as it is on a scientiﬁc concept, encourages many scholars to attempt a more-than-metaphor formulation. As Burt himself acknowledges: ‘Clear-thinking observers can be frustrated with the vagaries of social capital left as a metaphor. Social capital is the Wild West of academic work. There are no skill or intellectual barriers to entry.’16 Many scholars will ﬁnd it inherently frustrating to limit the understanding of SC to that of a loose metaphor based on a hard concept (that is, economic capital) rather than pursuing a theoretical elaboration far more precise than any loose metaphor can be. If the usefulness of social connections is to be discussed through a metaphor provided by the allusion to economic capital, why not attempt to formulate a theorization of SC as conceptually parsimonious and as amenable to measurement as economic capital? Thus the metaphor-only approach, compelling as it may seem, leaves the scholarly ﬁeld with a massive unresolved tension between rather broadly shared aspirations for a parsimonious and undiﬀerentiated conceptualization (and measurable operationalization) of all network-based advantage, and the reality that such aspirations cannot be persuasively met. The result is both unnecessary terminological confusion and a fair bit of work (often by superb scholars) that searches for a fountain of wisdom – in the form of a cohesive understanding of all beneﬁcial social connections – which cannot be found, and which undervalues the search for the many sources of more delimited social knowledge remaining to be discovered. Needless to say, this state of aﬀairs is counterproductive for the scientiﬁc enterprise. The Wild West, if we adopt Burt’s imagery, hardly seems the most promising model for scientiﬁc discourse and advancement. Burt himself appears ambivalent about the conceptual and terminological anarchy that ensues, remarking that ‘The variety is as interesting and exciting as it is corrosive to cumulative work.’17 The metaphor-only understanding cannot easily command and channel most of the scholarly discourse that takes place on a terminological terrain which encourages a parsimonious overarching understanding of all beneﬁcial social relations, that is to say, a more-thanmetaphor usage of SC. Yet even if, against all odds, the metaphor-only approach came to be the only understanding of SC in scientiﬁc usage, it would still entail some costs

Costs of conceptualizing social ties as social capital?

75

for scholarly investigation, albeit far smaller ones than those posed by the more expansive understanding of the term. I argue below that a fully adequate understanding of the enormous array of consequences that flow from diverse patterns of social relations requires us to adopt multiple analytical prisms. I contend that at least one analytical lens quite helpful for examining some social connections – and their consequences – highlights processes that are so far removed from the workings of economic capital as to render them almost inconceivable from the standpoint of analysis shaped by the metaphor of capital. The metaphor of capital does succeed in underscoring, and drawing attention to, the magnitude of the ‘pay-oﬀ’ to be obtained by taking seriously the causal consequences of social ties and norms. But that metaphor encourages scholars to look for parallels of social life to economic capital far greater than what the evidence can sustain, and in the process it distracts much attention from important questions – and crucial distinctions – concerning exactly how and where social relations generate important consequences. Moreover, despite the exemplary scholarly quality of much work carried on under the SC banner, and the undeniable signiﬁcance of many ﬁndings conveyed in that work, unsolvable disagreements over how to conceptualize and operationalize SC lead to unnecessary confusion. Even at its best the SC approach to the study of social relations and norms carries meaningful costs for the scientiﬁc enterprise.

PRISMS FOR THE STUDY OF SOCIAL CONNECTIONS – AND THEIR CONSEQUENCES This chapter’s argument heavily emphasizes diﬀerences among types of ties and eﬀects, a perspective on social relations that I now amplify. I argue that social connections are understood by their participants – and should be studied by social scientists – through more than one analytical prism. A lengthy and distinguished tradition in the study of social connections seeks to understand them by focusing on their structural, or relational, conﬁguration. From this network-based standpoint, the most central feature of ties is to be found in their placement within the overall pattern of social relations that prevails in a population. Burt’s distinction between brokerage and closure is a prime example of such an approach.18 With a foundation in this structural study of network connections, many analysts move on to examine the implications of one or another pattern of connection for the flow of resources and information. Without undervaluing this approach, I argue for the parallel importance of another gestalt for understanding social connections, one that helps to

76

Overview

illuminate some – but not all – social connections that carry signiﬁcant causal consequences. Most crucially, I insist that the location of social connections in the overall network structure of a population does not presuppose the content of the social interactions in question or the meaning ascribed to them by the actors involved. Whereas some ties are most usefully understood through a prism that emphasizes their placement in extensive network structures (and the role of the ties in promoting social processes such as resource or information flows extending well beyond direct dyadic – that is, two-way – connections), other ties are more usefully understood through a prism that emphasizes the intrinsic importance and meaning that participants place on their direct dyadic interactions. The theoretical distinction I emphasize here counterposes those ties that are understood by their participants fundamentally as means to obtain an end or ends external to the direct dyadic social interaction with others that are valued intrinsically by their participants. The contrast between instrumentally focused action and action rooted in a sense of its intrinsic merit is a fundamental element of the Weberian approach to sociology but its relevance for the study of social ties and their impact has been underdeveloped. With a foundation in my own empirical ﬁndings in Democracy’s Voices and recent work by other scholars, I argue that culturally or subculturally rooted practices and meanings surrounding social ties help to determine whether given ties are best understood through a lens focused on their external network-based instrumentality or one emphasizing their intrinsic interactive worth. Moreover, as we shall see, those ties that are embedded in actor understandings emphasizing the external network-based instrumentalities oﬀered by dyadic linkages – which is to say, the indirect access to distant resources and power opened up by such linkages – may hold a weaker ability to generate certain positive outcomes than ties rooted in very diﬀerent actor understandings. To specify further, there may be elements of instrumentality present in ties that are valued intrinsically, but in such cases the goals pursued by participants are internal to the interaction in question and not (as in the opposing case) external resources to which the participants hope to gain access by cultivating the tie. Two simple illustrations help to make the point: a student may value interaction with a teacher because she or he enjoys learning (and assumes it likely to be of use at some point), thus placing intrinsic importance on the student–teacher relation, or because the student believes that the teacher stands as an influential gatekeeper controlling access to external opportunities and resources such as future employment, even if he or she does not enjoy or care about learning. Similarly, an employee in a medium-sized company may value interaction with his or her immediate

Costs of conceptualizing social ties as social capital?

77

supervisor as enjoyable conversation and also as an opportunity to learn useful lessons from the supervisor or, alternatively, may value that interaction only as a conduit to resources that the employee hopes to obtain through the intermediary (network-activating) role of the supervisor. Of course, in reality these possibilities are not mutually exclusive. My objective here is to introduce opposing ideal types even though much empirical reality may be located somewhere between these two theoretical poles. To reiterate the main point: the distinction I draw focuses not on the presence or absence of instrumental motives but instead on the way in which participants value and understand their direct social interactions, as a means to obtain access to resources made available through the network linkages of their dyadic partners or as intrinsically valuable (and/or useful) interactive experiences. Crucially, I also argue that the causal impact of ties may vary as a function of this distinction, and that for certain ends, ties that are intrinsically valued as interaction may prove more powerful in their ability to reshape outcomes than ties constructed and understood through an external-instrumentalist prism that focuses on access to goods outside the direct interaction. Moreover, this distinction can be drawn, and may prove causally decisive, for social ties that seem more or less indistinguishable from one another from a purely structuralist perspective that emphasizes only the pattern of network linkages. From this chapter’s perspective, the meaning that those involved place on ties proves decisive. Thus I emphasize the conceptual and explanatory signiﬁcance of a distinction among types of ties that the SC approach, even in its metaphor-only version, would be unlikely to take note of. After all, capital is to be understood through the development of future wealth that it facilitates, and not through the intrinsic value that some may place on their riches. Ties that involve interactions which are valued intrinsically, and not as a means to obtain external resources (embedded in networks to which the ties provide access), are best understood through an analytic prism far removed from the capital metaphor, but for certain ends they prove more powerful than those ties best understood through a network-based prism focusing on structurally determined access to resources external to the direct dyadic ties. The extent to which ties shape at least some meaningful outcomes is not a function of the degree to which the ties ﬁt the metaphor of capital – even in the loosest sense. There is much evidence in favor of the proposition that social ties can and should be understood not only from the perspective of their placement in network structures but also from the standpoint of their actual content, the nature of what goes on in the interactions in question, and the meaning placed on those relations by participants. The important ﬁndings of Bearman and Parigi show that the conversational themes articulated in

78

Overview

social relations and thought to be important by contemporary Americans are enormously varied, encompassing many topics – often far removed from instrumental matters or strategies – and that the simple existence of network ties does not fully determine whether those ties will take the form of conversations which their participants think of as important.19 Evidence culled from several centuries of Japanese history by Eiko Ikegami in her important recent book, Bonds of Civility,20 analyzes the emergence of wide-ranging cultural associations and ties that were valued by their participants on the basis of the intrinsic content of the aesthetically oriented activity taking place. She shows that these ties, and the network revolution they represented, ushered in new forms of expression and social connectedness with carry-over consequences for the political system. Aesthetically oriented ties facilitated social contacts among individuals from disparate positions in social hierarchies, and they did so without centering these interactions on status-oriented strategies and goals external to the cultural content of the ties. In both of these cases, the signiﬁcance that participants place on interactions and the content of the interactions cannot be mechanically read from a purely relational characterization of the ties. Ties may (or may not) be valued, instead, on the basis of the direct dyadic interactions that characterize them. My own work in Democracy’s Voices makes a case for the explanatory signiﬁcance of a distinction based precisely on the extent to which ties are valued either for their intrinsic dynamics or as a means to obtain access to external power-based resources. The data and analysis presented in that book show that boundary-crossing ties between intellectuals and workers hold the ability to reshape democratic discourse in important ways, but that this capacity is contingent on just the sort of distinction outlined above. The empirical ﬁndings of my survey analysis and qualitative ﬁeldwork establish that, in at least one politically deﬁned subculture, boundary-crossing ties hold an extraordinary ability to reshape the discursive horizons of local communities – and as a result, the broader quality of public life in democracy – but that structurally similar ties in another politically deﬁned subculture thoroughly lack such an ability to reshape public discourse. On the basis of those empirical ﬁndings about the causal strength of social ties in one subculture – contrasted with their relative impotence in another subculture – I develop the broader theoretical claim that ‘ties-asconversation’ hold the ability to reshape the discursive practice of their participants, with fundamental carry-over signiﬁcance for public life, whereas ‘ties-as-brokerage’ lack that ability. The ties took the form of ‘conversation’ where subculturally based understandings encouraged participants to view interactions between workers and intellectuals as intrinsically valuable. In that setting, listening, as well as speaking, came to be a highly valued

Costs of conceptualizing social ties as social capital?

79

component of interactions. Such conversational ties vastly increased the likelihood that local leaders and their communities would articulate their interests through ‘globalizing discursive horizons’ of interest to a national audience instead of through narrow defensive localism. Where boundarycrossing ties took the form of ‘brokerage’ facilitating access to otherwise unavailable external resources, or sources of power, they lacked such an ability. In both instances the ties in question connected dissimilar social worlds, but whether that connection can best be understood as brokerage or conversation proved to vary by subculture. Thus I argue that for certain important concerns, such as shaping the quality of democratic public life, ties-as-conversation, which can only be fully understood through an analytic prism emphasizing the intrinsic value of direct interaction, hold far greater causal weight than ties-as-brokerage, which are to be understood through an analytic prism emphasizing the access they oﬀer to otherwise unattainable network-based sources of power or resources. This distinction builds on a great deal of excellent conceptual work by social scientists21 and on careful empirical investigation, but the eﬀort to diﬀerentiate between ties best understood through one analytic prism and those most readily understood through a quite dissimilar prism does not reflect currently predominant approaches and it runs directly counter to the spirit of work, shaped by the metaphor of capital, which searches for underlying unity in the causal logics underpinning the impact of ties and norms. This chapter argues that just such distinctions are required if we are to uncover fully the explanatory signiﬁcance of social connections. All ties can be placed analytically in network structures (shaping access to more or less distant resources) and all ties do involve at least minimal conversational interaction, but the extent to which any given dyadic social tie is primarily understood by its participants – and can be usefully read by scholars – through one of these prisms or the other is highly variable from case to case, and this variation proves to have important causal consequences. This is not the place to examine the dynamics shaping actor understandings of their ties, which is to say determining whether they best ﬁt one prism or the other. However, it is possible to suggest the following. Culturally or subculturally based understandings of ties appear to be crucial in shaping the meanings and practices that actors adopt, but the magnitude of the incentives available to those who do make use of network ties primarily to increase their access to external resources is also surely relevant. The framework I propose searches for scientiﬁcally relevant distinctions among types of social connections, and underscores the considerable causal force – for at least some purposes – of those ties that are valued as intrinsically worthwhile interactions and that are not primarily seen by their

80

Overview

participants as brokerage-like avenues to influence and power-based resources. A good deal of important recent work at the intersection of political and cultural sociology is consonant with this claim. The aesthetic networks richly analyzed by Ikegami are a powerful example of social relations built on the basis of the intrinsic meaning placed on interactions, and yet these interactions were capable of generating wide-ranging second-order consequences for the larger social and political order.22 Paul Lichterman’s compelling analysis of the varying ability of church groups to bridge social divisions in the contemporary United States identiﬁes as a source of special power to do so the extent to which such groups treat their interactions reflexively, examining their own conduct and experiences self-critically.23 Again, the intrinsic importance placed on interactions themselves proves to be a source of strength in stitching together boundary-crossing social connections of societal signiﬁcance. Gianpaolo Baiocchi’s skillful examination of the broad socio-political signiﬁcance of participatory budgeting in Porto Alegre, Brazil notes the limitations of community activists who perceived participatory institutions only as avenues to secure particular resources, and the strength of (frequently poor) activists who developed within those institutions a discourse of the common good emphasizing general civic purposes.24 In this instance, diﬀerent groups of activists articulated their socio-political engagement around relatively dissimilar prisms emphasizing either localistic access to external resources or overarching civic commitments that valued the intrinsic worth of participation. In all of these recent studies, distinctions similar to those emphasized by this chapter prove highly useful for scholarly analysis. Social engagements focused on the intrinsic value of interactions, connections and discussion hold considerable power, but they are not best understood through the metaphor of capital.

CONCLUSION Nothing argued above is intended to place exclusive explanatory and theoretical weight on interactions conceived primarily as intrinsically meaningful, or to undervalue the importance of those network-based understandings of social relations that focus instead on access to more or less distant resources and on the advantages provided (or denied) by one’s relational location. A fully adequate understanding of social relations and their causal impact requires us to employ both prisms and to examine their relative explanatory relevance in diﬀering settings and for predicting diﬀerent desired outcomes. But this claim is not intended to represent an all-encompassing approach to the study of social ties; it is oﬀered here as one example of the sort of

Costs of conceptualizing social ties as social capital?

81

distinction that scholars can usefully develop and deploy in the search for an understanding of empirical reality. The push for excessive theoretical parsimony – or, worse still, the tendency to reify concepts – can only get in the way of an adequate understanding of empirical reality, as the Weberian perspective insists. The Weberian approach to social science, as I understand it, abhors essentialisms of all sorts – be they economic, cultural, social relational or otherwise – and instead builds its explanations on the embrace of a series of seeming tensions or intellectual oppositions.25 This approach treats distinctions between seemingly contradictory causal logics and social processes as potentially fruitful in the search for understanding of an empirical world more complex than any one theoretical claim can fully capture on its own. The perspective that animates this chapter values those concepts proven to be useful, but in the tradition of Weber understands even those concepts to be inherently limited. This perspective pursues the explanatory usefulness of embracing conceptual tension and of recognizing the inherent complexity of the empirical world. The promise of conceptual and terminological order seemingly provided by the SC school exerts a powerful attraction on many of the very best minds in the social sciences, but any embrace of the metaphor of capital for understanding social relations comes at a certain cost for it discourages scholars from delineating and exploring distinctions crucial for an adequate understanding of processes and outcomes in the world around us.

NOTES 1. 2. 3. 4. 5.

6. 7. 8. 9. 10. 11. 12.

For very useful comments on an earlier draft of this chapter I wish to thank Gianpaolo Baiocchi, Viva Bartkus, Erika Summers Eﬄer, Jaume Lopez, Julia Lopez, Paul Lichterman and Jaime Ros. Fishman (2004). See Granovetter (1974). Wellman (1992). Money may be taken as the metric by which economic capital can be measured and exchanged but there is scholarly controversy over the extent to which money is fully fungible. See the important analysis by Viviana Zelizer (1994), The Social Meaning of Money. I am indebted to Thomas Miley for insisting on this point in discussions after a seminar in which I presented an earlier version of this argument at the Universitat Pompeu Fabra (Barcelona). For an interpretation and defense of the Weberian methodological perspective, see Fishman (2007). Smith and Kulnych (2002), on this point see p. 173. Also see Helliwell and Huang (2005). See Hall (1999) and Paxton (1999). Skocpol and Fiorina (1999) and Wuthnow (1998). Putnam (2001).

82 13. 14. 15. 16. 17. 18. 19. 20. 21.

22. 23. 24. 25.

Overview Putnam (1993). Burt (2005). Ibid, p. 4. Ibid, p. 5. Ibid. For pioneering work in this tradition see White et al. (1976). Bearman and Parigi (2004). Ikegami (2005). A great deal of work has sought to develop and put to use the concept of brokerage. For important theoretical statements, see the previously cited work of Burt, as well as Gould and Fernandez (1989). On conversations and interactions themselves as the basis of network relations, and for a careful review of relevant literature, see the theoretically rich analysis of Ann Mische (2003). On conversations from an SC perspective see Nahapiet (2005). For important theorizations of the interface between culture and social relations see Emirbayer and Goodwin (1994) and Lizardo (2006). See Ikegami (2005). Lichterman (2005). Baiocchi (2005). I develop this understanding of the Weberian perspective in ‘On being a Weberian’ (Fishman, 2007).

REFERENCES Baiocchi, Gianpaolo (2005), Militants and Citizens: The Politics of Participatory Democracy in Porto Alegre, Stanford, CA: Stanford University Press. Bearman, Peter and Paolo Parigi (2004), ‘Cloning headless frogs and other important matters: conversation topics and network structure’, Social Forces, 83(2), 535–57. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Closure, Oxford: Oxford University Press. Emirbayer, Mustafa and Jeﬀ Goodwin (1994), ‘Network analysis, culture and the problem of agency’, American Journal of Sociology, 99(6), 1411–54. Fishman, Robert M. (2004), Democracy’s Voices: Social Ties and the Quality of Public Life in Spain, Ithaca: Cornell University Press. Fishman, Robert M. (2007), ‘On being a Weberian (after Spain’s March 11–14): notes on the continuing relevance of Weber’s methodological perspective’, in Laurence McFalls (ed.), Max Weber’s ‘Objectivity’ Reconsidered, Toronto: University of Toronto Press. Gould, Roger V. and Roberto M. Fernandez (1989), ‘Structures of mediation: a formal approach to brokerage in transaction networks’, Sociological Methodology, 19, 89–126. Granovetter, Mark (1974), Getting a Job, Cambridge: Harvard University Press. Hall, Peter A. (1999), ‘Social capital in Britain’, British Journal of Political Science, 29, 417–61. Helliwell, John and Haifang Huang (2005), ‘How’s the job? Well-being and social capital in the workplace’, National Bureau of Economic Research, Working Paper 11759, p. 18. Available at www.nber.org/papers/w11759. Helliwell, John, Haifang Huang and Robert D. Putnam (2008), ‘How’s the job? Are trust and social capital neglected workplace investments?’, Chapter 4, in this volume.

Costs of conceptualizing social ties as social capital?

83

Ikegami, Eiko (2005), Bonds of Civility: Aesthetic Networks and the Political Origins of Japanese Culture, Cambridge: Cambridge University Press. Lichterman, Paul (2005), Elusive Togetherness: Church Groups Trying to Bridge America’s Divisions, Princeton: Princeton University Press. Lizardo, Omar (2006), ‘How cultural tastes shape personal networks’, American Sociological Review, 71(5), 778–807. Mische, Ann (2003), ‘Cross-talk in movements: reconceiving the culture–network Link’, in Mario Diani and Doug McAdam (eds), Social Movements and Networks: Relational Approaches to Collective Action, Oxford: Oxford University Press, pp. 258–80. Nahapiet, Janine (2005), ‘Knowledge, learning and social relationships: the role of conversation’, paper presented at The Microfoundations of Organizational Capabilities and Knowledge Processes, Copenhagen Business School, 1–2 December. Ostrom, Elinor (2008), ‘What is social capital?’, Chapter 1, this volume. Paxton, Pamela (1999), ‘Is social capital declining in the United States? A multiple indicator assessment’, American Journal of Sociology, 105(1), 88–127. Putnam, Robert (1993), Making Democracy Work: Civic Traditions in Modern Italy, Princeton: Princeton University Press. Putnam, Robert (2001), Bowling Alone, New York: Touchstone Books. Skocpol, Theda and Morris Fiorina (eds) (1998), Civic Engagement in American Democracy, Washington, DC: Brookings Institution Press, New York: Russell Sage Foundation. Smith, Stephen Samuel and Jessica Kulnych (2002), ‘It may be social, but why is it capital? The social construction of social capital and the politics of language’, Politics and Society, 30(1), 149–86. Wellman, Barry (1992), ‘Which types of ties and networks give what kinds of social support?’, in Edward Lawler, Barry Markovsky and Cecilia Ridgeway (eds), Advances in Group Processes, Greenwich, CT: JAI Press, pp. 207–35. White, Harrison C., Scott A. Boorman and Ronald L. Breiger (1976), ‘Social structure from multiple networks. I. Blockmodels of roles and positions’, American Journal of Sociology, 81, 730–80. Wuthnow, Robert (1998), Loose Connections: Joining Together in America’s Fragmented Communities, Cambridge: Harvard University Press. Zelizer, Viviana (1994), The Social Meaning of Money, New York: Basic Books.

PART II

Social capital: reaching out

4.

How’s the job? Are trust and social capital neglected workplace investments? John F. Helliwell, Haifang Huang and Robert D. Putnam1

This chapter investigates trust and social capital as potentially beneficial yet neglected investments in the workplace. We define social capital as networks together with shared norms, values and understandings that facilitate cooperation within or among groups. We use subjective measures of life satisfaction to value the climate of trust in the workplace, set in the context of life as a whole. Using data from two separate Canadian surveys, and one large US survey, we find that the climate of trust in the workplace is strongly related to subjective well-being, even after allowing for individual personality differences. For example, results from one of the surveys suggest that moving one point on a ten-point scale of workplace trust affects life satisfaction about the same amount as a 40 per cent change in income. The size of these effects, often referred to as compensating differentials, suggests unrecognized opportunities for managers and employees to increase satisfaction and productivity by building better workplaces. Throughout the chapter, we examine the determinants of workplace trust and explore differences among subgroups of workers. We also consider why such large unrecognized opportunities for improving the quality of workplace social capital remain, and suggest some plausible investment strategies for capturing this potential.

INTRODUCTION The chapter began with the recent finding that workplace trust and other measures of the quality of life on the job have strikingly high incomeequivalent values (Helliwell and Huang, 2005). These values are so great that they suggest the existence of many unexploited opportunities for investment in the workplace. This chapter examines these possibilities in more 87

88

Social capital: reaching out

detail. First, we discuss our methodology to attach values to trust and social capital in the workplace. Then we consider the latest evidence of the values of non-financial aspects of work, with special attention to trust on the job, often taken as a key indicator of the quality of workplace social capital. We use the Organisation for Economic Co-operation and Development (OECD) definition of social capital as ‘networks together with shared norms, values and understandings that facilitate co-operation within or among groups’ (OECD, 2001, p. 41). Because there are many established survey measures of trust in general and within specific domains, many previous attempts to establish the implications of social capital have used trust as an independent variable, particularly where economic growth rates are the variable to be explained.2 We also give trust a central role, partly because it has such well-measured and clearly defined linkages to subjective well-being, but also because it has strong empirical and conceptual links to ‘trustworthiness’ which in turn deserves to rank foremost among norms that facilitate social capital, and thereby cooperation within or among groups.3

METHODOLOGY Most previous attempts to value non-financial aspects of jobs have focused on job hazards, and have used wage equations with safety as the independent variable and wages or incomes as the dependent one. The most obvious estimation problem is that posed by unmeasured differences in employee ability and training. With the usual assumption that safety is a normal good, more able or better-trained workers are in a position to choose jobs that produce more income and more safety. Such formulae often follow an approach similar to the one below. In this equation, where yi is the earnings level for worker I and Xi is a vector of job characteristics applicable to worker i’s job, the compensating differentials – the prevailing market valuations of job attributes – are estimated by the coefficient vector . The Zi are measured characteristics of worker i, and the Zui are unmeasured characteristics of the worker, the job, or the local marketplace. The i are the assumed error terms. A typical estimation form is: ln(yi) Xi + Zi + Zui + i

(4.1)

However, multiple econometric and other difficulties exist with this approach. Cross-sectional data often underestimate the compensating differentials – the estimation of .4 This occurs because the unobservable

How’s the job?

89

earning potential becomes part of the error term, thus making the error term (i ) correlated with both the dependent wage variable (yi) and the job characteristics (Zi). Intuitively, this happens because people with greater ability will generally use their market attractiveness to acquire jobs that have both high wages and good working conditions. This creates a positive association between more enjoyable jobs and higher wages, thereby masking the trade-off between pay and work environment that workers face. These econometric difficulties suggest it may be more promising to use subjective well-being as a direct measure of utility. In this case, compensating differentials would be the ratio of the well-being effects of job characteristics over the well-being effect of income. In the equations below, and

are measures of marginal utility and the compensating differentials () reflect the ratio of , the job characteristics coefficient, over , the coefficient on the log of income. Our specific approach has three steps: first, to create a linear general utility function U(y, X) at the equilibrium point; second, to estimate the marginal contributions of job characteristics and income to life satisfaction; and third, to calculate the compensating differentials directly from these estimated coefficients, as ratios of the job characteristics coefficients to the coefficient on the (log of) income. To the extent that jobs are actually available with the characteristics in question, these ratios should reflect the prevailing market valuations of job attributes.5 In its general form, the proposed strategy has LS stand for life satisfaction, y(y) is the functional form of income, x(X) is the functional form of job attributes, Zi are all other controls. The compensating differentials from equation (4.1) are then derived using equation (4.3). The transformations with () are needed in some cases to have uniform coefficients and , which denote marginal utilities, across heterogeneous agents. These functional forms accommodate a concave utility. The proposed strategy is as follows: LSi y(yi) + x(Xi) + Zi +i /

(4.2) (4.3)

We have built multiple elements into the equations so that they can cover a variety of situations. First, by measuring income in its log form instead of its level – with y(y) log(y) – we are able to estimate compensating differentials as proportions of income. This approach reflects both standard economic assumptions and the many empirical results suggesting that less affluent agents derive greater utility from extra income. Second, for job attributes, we initially adopted a simplistic view that their per unit

90

Social capital: reaching out

contribution to workers is the same regardless of income or level of X, so that x(Xi) Xi. Third, we use Zi to control for many observed heterogeneities across agents, including, in some tests, personality differences. We also search for, and find, evidence that marginal valuations of income and other job characteristics might differ among groups of workers; namely, that those found in low-quality working environments might, on average, care less about the non-financial aspects of their jobs.

EVIDENCE ON THE VALUE OF WORKPLACE TRUST AND SOCIAL CAPITAL Our results are based on two recent Canadian surveys and one US survey. The two Canadian surveys are the 2003 second-wave SSHRC-supported Equality, Security, and Community Survey (ESC hereafter), and the 2003 General Social Survey-17: Social Engagement Survey (GSS hereafter). The US survey is the 2000/01 Social Capital Community Benchmark Survey (US Benchmark hereafter). Our sample includes all those respondents who are currently employed and on whom we have observations on the other key variables. The sample size is about 2500, 10 500 and 16 500 for the three surveys, respectively. The main results will be based on the Canadian ESC survey because its second wave was specifically designed to assess some aspects of workplace relations, while also measuring their importance for life satisfaction. This survey also examined trust in management which had the largest effect on life satisfaction. The other two surveys measured trust among colleagues, instead of trust in management, and covered fewer other aspects of the job. None of the surveys asks about both trust among co-workers and trust in management. As a result, we cannot address directly some of the interesting differences among bonding, bridging and linking social capital (OECD, 2001, p. 42) in the workplace. We can utilize informal triangulation to suggest hypotheses for future research. For example, Canadian ESC data show that respondents who are union members are far less likely to rate their workplaces high on the extent to which management can be trusted. US Benchmark and Canadian GSS data show that union members and non-union employees rate trust in colleagues similarly. These findings suggest that the type of bonding social capital unions provide is more likely to be associated with trust among colleagues than between labour and management. These results also suggest that trust of management is likely to be high when high levels of bridging and linking social capital exist. By contrast, linking social capital may be lower in union shops. This may be

91

How’s the job? 8.6 8.4 8.2 8 7.8 7.6 7.4 7.2 7 6.8

5 or below: 27%

6: 12%

7: 18%

8: 20%

9: 12%

10: 11%

Figure 4.1 Life satisfaction at different levels of trust in management, from 2003 Canadian ESC because workers in jobs marked by low-quality relations between themselves and management are more likely to approve unionization, and because the bonding social capital fostered in a union environment may encourage solidarity against management. We start by presenting our ESC data on life satisfaction and trust in management. Figure 4.1 divides the ESC respondents according to how highly, on a scale of 1 to 10, they would rate the level of trust that workers have in management at their workplace. Respondents’ ratings of the workplace trust environment are mapped against their average life satisfaction (see Figure 4.1). The respondents are drawn randomly from the whole working population, and hence include workers and managers. In principle, these two groups could both answer the question in the same way, because it specifically addresses not the respondent’s own level of trust in management, but the general climate of trust between workers and management. The bars in the chart show by their height the average level of life satisfaction recorded, again on a scale of 1 to 10, by those giving each of the possible answers to the workplace trust question. It is noteworthy that the average level of workplace trust is quite high, and that there are very few who rate trust at 5 or below on the ten-point scale. The number is small enough that we gathered all these respondents together for the purposes of Figure 4.1. There is a very strong positive relation between assessments of workplace trust and average measures of life satisfaction. An increase in workplace trust equal to one standard deviation (2.3 on the ten-point scale, as shown

92

Social capital: reaching out

in the Appendix) produces an increase of 0.33 points in life satisfaction (see Table 4.1). Thus an increase of one point in the ten-point trust scale is associated with an increase of life satisfaction of 0.17 points. How big is this effect? One way to measure it is to divide respondents by income class, and see how much life satisfaction rises with income, with and without accounting for all of the other variables. The ESC results reported in Helliwell and Putnam (2005, Figure 17.2d) suggest that moving from the bottom ( $20 000 annual household income) to the top ( $100 000) is matched by an increase of less than 0.3 points on the ten-point life satisfaction scale, roughly of the same magnitude as the benefit from an increase in workplace trust of one standard deviation, or equivalently a jump in ranking that covers about one-third of the sample, starting from the middle. For our current research we are mainly using personal income in logarithmic form so that we can interpret the estimated compensating differential as proportional changes of income. In Table 4.1, we include estimates from the ESC2, the GSS and the US Benchmark Survey. These surveys use different scales for their workplace trust variables. For comparability across surveys, we use the standardized score instead of the survey response to the questions.6 All three surveys show that workplace trust is an important factor in explaining interpersonal differences in happiness or life satisfaction. The ESC results suggest that if typical employees move from one job to another, and their rating of trust rises by 2.3 points, then their life satisfaction will also rise as long as their new income is more than two-fifths as large as before. In the original ten-point scale, a one-point increase in trust in management is associated with an approximate 40 per cent change in income. The results in the other two larger surveys asking about workplace trust among co-workers are directionally similar (see Table 4.1). Although the type of trust being asked about is conceptually different, employees attach just as much or more value to trust than in the ESC survey.7 People are often sceptical of these results. If workplace trust is so valuable to workers, and is probably also good for productivity, why are there any workplaces that exhibit low levels of trust? This is a good question, and one that we shall try to address. But first, to deal with some of these issues, Helliwell and Huang (2005) employed a large number of tests to ensure that these values are not an artefact of some special features of the particular survey, of the nature of life satisfaction data, of individual personality differences, or of correlations with some critical omitted variables affecting life satisfaction. The study presented consistent results from three Canadian surveys: the 2003 Ethnic Diversity Survey (EDS), the GSS and the ESC. The GSS contains questions designed to capture features of the respondent’s personality, many of which might make them more inclined

How’s the job?

93

to report high life satisfaction and also to see the bright side of their work. One such personality indicator in the GSS is what Statistics Canada labels the ‘mastery scale’. This index is based on a principal component analysis of the extent of agreement with the following statements: ‘I have little control over the things that happen to me’; ‘There is really no way I can solve some of the problems I have’; ‘There is little I can do to change many of the important things in my life’; ‘I often feel helpless in dealing with the problems of life’; ‘Sometimes I feel that I’m being pushed around in life’; ‘What happens to me in the future depends mainly on me’; ‘I can do just about anything I really set my mind to do’. Although these individual measures of personality were, as expected, strong determinants of reported life satisfaction, they did not materially affect the key coefficients of interest, and as a result, the compensating differentials were robust across personality differences. In Table 4.1, we add the personality measure to the last column. The results show that even though the mastery scale enters strongly and significantly, the estimated compensating differential, which is measured by the ratio of coefficients, drops only marginally from 1.3 to 1.2.8 In Table 4.2 we turn to more fully specified models for life satisfaction in the ESC and GSS surveys for Canada and happiness in the US Benchmark survey. Workplace trust continues to exhibit significant well-being effects in regressions that add basic demographic information and numerous controls. While the coefficients drop somewhat when measures of general trust and trust in other domains are included, this is to be expected. Personality and other unmeasured differences among respondents might lead them to be optimistic or dyspeptic in their assessments of trust. Table 4.2 – which is made up of three sub-tables, one for each survey – depicts the results using life satisfaction in the ESC and GSS surveys and happiness in the US Benchmark survey. The first column in each sub-table is always for the largest possible sample of working population, which is our focus for now. The findings are striking. Workplace trust is the single most important trust variable in the Canadian ESC and GSS results and it is rivalled only by trust in police in the US Benchmark survey results. In addition, all of the measures of domain trust are highly significant statistically and contribute substantial fractions of the variation in well-being. Table 4.4 shows the compensating differentials, which are expressed as the percentage change of income that has an equal effect on life satisfaction as when workplace trust moves one-third of a standard deviation. This covers about 10 per cent of the sample when starting from the mean. For the ESC and the US Benchmark survey, the estimates are both about 32 per cent. The estimate is higher for the GSS, reaching 60 per cent of household income.

94

0.311 [0.038]**

0.263 [0.048]**

0.313 [0.048]**

0.335 [0.038]**

Survey linear regression

Regression method

Job: Trust in management, standardized score Log of personal income (see note 2) Trust in co-workers, standardized score Log of household income (see note)

Life satisfaction, 1–10 point scale

Canadian ESC2 Working population (see note 1)

0.306 [0.018]**

0.256 [0.019]**

0.246 [0.032]**

0.31 [0.019]**

0.3 0.272 [0.033]** [0.032]**

Survey linear regression

Life satisfaction, 1–10 point scale

Canadian GSS Working population

Comparing well-being effects of income and workplace trust

Dependent variable

Sample

Table 4.1

0.125 [0.006]**

0.157 [0.008]**

Survey linear regression

Happiness: 1–4 point scale

0.133 [0.008]**

0.112 [0.005]**

US Benchmark Working population (see note 1)

95

4.677 7.972 [0.510]** [0.016]** 2522 10676 0.05 0.05

4.643 4.972 [0.367]** [0.355]** 9346 9346 0.01 0.06

3.332 [0.007]** 16519

0.05

4.678 [0.355]** 9039

0.11

2.238 [0.130]**

0.03

1.637 [0.083]** 16519

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. Self-employed is excluded in the two Canadian surveys. US Benchmark survey does not provide information on self-employment status. Therefore the sample may include self-employed. 2. The Benchmark survey does not provide information on personal income; household income is used. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Observations R-squared

7.952 5.204 [0.033]** [0.507]** 2522 2522 0.04 0.01

(cont.)

Mastery scale net of income effect(GSSonly) Constant

Table 4.1

0.07

1.902 [0.080]** 16519

96

Job: Has variety of tasks Job: Have enough time Job: Free of conflicting demands Job: Makes own decision Dummy, union member Self-perceived health status, scaled 1 to 5

Log of personal income Job: Trust in management, standardized score Job: Requires skill

0.281 [0.029]**

0.194 [0.039]** 0.189 [0.025]**

0.218 [0.072]** 0.142 [0.037]**

0.275 [0.053]**

0.235 [0.109]* 0.393 [0.105]** 0.199 [0.079]* 0.206 [0.067]** 0.029 [0.106] 0.138 [0.049]** 0.28 [0.029]**

Union members

0.154 [0.041]** 0.165 [0.026]**

Whole sample

Immigrant

0.284 [0.034]**

0.18 [0.045]** 0.244 [0.034]**

0.179 [0.068]**

0.118 [0.065] 0.27 [0.061]**

Life satisfaction; 1–10 point scale

The rest

Canadian ESC2, year 2003 Working population

0.319 [0.032]**

0.216 [0.049]** 0.171 [0.027]**

The rest

0.107 [0.098]

0.126 [0.074] 0.369 [0.083]**

Visible minority

0.308 [0.030]**

0.211 [0.046]** 0.173 [0.026]**

The rest

Expanded models on the well-being effects of income and workplace trust, ESC2 survey ordered probit

Dependent variable

Sample

Table 4.2(a)

97

Marital status: Married Marital status: As married Marital status: Divorced Marital status: Separated

Age Group: 65 up

Age Group: 55–64

Age Group: 45–54

Age Group: 35–44

Age Group: 25–34

Male

0.132 [0.039]** 0.231 [0.084]** 0.203 [0.089]* 0.189 [0.093]* 0.064 [0.108] 0.198 [0.274] 0.365 [0.064]** 0.357 [0.086]** 0.202 [0.099]* 0.24 [0.120]*

Whole sample

0.128 [0.039]** 0.239 [0.084]** 0.212 [0.090]* 0.195 [0.093]* 0.068 [0.109] 0.161 [0.267] 0.373 [0.064]** 0.373 [0.084]** 0.196 [0.099]* 0.238 [0.115]*

(cont.)

Dependent variable

Sample

Table 4.2(a)

0.091 [0.073] 0.41 [0.209] 0.411 [0.201]* 0.322 [0.209] 0.047 [0.238] 0.339 [0.537] 0.401 [0.108]** 0.633 [0.136]** 0.287 [0.159] 0.177 [0.210]

Union members

Immigrant

0.136 [0.050]** 0.201 [0.097]* 0.177 [0.107] 0.222 [0.111]* 0.02 [0.128] 0.11 [0.308] 0.371 [0.079]** 0.232 [0.108]* 0.124 [0.123] 0.278 [0.129]*

0.03 [0.096] 0.379 [0.179]* 0.127 [0.176] 0.102 [0.192] 0.395 [0.218] 0.466 [0.395] 0.183 [0.117] 0.328 [0.185] 0.359 [0.190] 0.6 [0.245]*

Life satisfaction; 1–10 point scale

The rest

Canadian ESC2, year 2003 Working population

0.159 [0.047]** 0.215 [0.100]* 0.248 [0.105]* 0.236 [0.103]* 0.083 [0.129] 0.01 [0.347] 0.448 [0.074]** 0.398 [0.094]** 0.147 [0.120] 0.117 [0.127]

The rest

0.003 [0.131] 0.464 [0.177]** 0.29 [0.193] 0.27 [0.215] 0.087 [0.299] 0.909 [0.472] 0.027 [0.174] 0.435 [0.294] 0.697 [0.274]* 0.575 [0.502]

Visible minority

0.151 [0.043]** 0.216 [0.096]* 0.205 [0.099]* 0.187 [0.102] 0.041 [0.120] 0.258 [0.287] 0.438 [0.068]** 0.384 [0.088]** 0.125 [0.106] 0.206 [0.114]

The rest

98

Contacts with neighbours Number of membership or extent of activity

Education: With university degree Contacts with family member outside household Contacts with friends

Marital status: Widowed Education: High school Education: Between

0.206 [0.185] 0.09 [0.096] 0.099 [0.083] 0.127 [0.090] 0.153 [0.071]* 0.373 [0.085]** 0.04 [0.072] 0.008 [0.013]

0.431 [0.085]** 0.091 [0.073] 0.009 [0.013]

Whole sample

0.182 [0.188] 0.124 [0.094] 0.118 [0.082] 0.164 [0.088] 0.158 [0.071]*

(cont.)

Dependent Variable

Sample

Table 4.2(a)

0.249 [0.160] 0.158 [0.136] 0.027 [0.023]

0.137 [0.307] 0.159 [0.195] 0.02 [0.167] 0.025 [0.176] 0.141 [0.151]

Union members

Immigrant

0.49 [0.114]** 0.054 [0.085] 0.006 [0.016]

0.216 [0.257] 0.111 [0.111] 0.178 [0.106] 0.236 [0.105]* 0.149 [0.091] 0.364 [0.189] 0.154 [0.144] 0.024 [0.029]

0.551 [0.294] 0.269 [0.231] 0.235 [0.220] 0.343 [0.220] 0.093 [0.147]

Life satisfaction; 1-10 point scale

The rest

Canadian ESC2, year 2003 Working population

0.469 [0.099]** 0.064 [0.083] 0.002 [0.015]

0.091 [0.251] 0.116 [0.104] 0.117 [0.091] 0.12 [0.099] 0.139 [0.089]

The rest

0.071 [0.270] 0.447 [0.186]* 0.026 [0.036]

0.802 [0.637] 0.023 [0.360] 0.193 [0.336] 0.203 [0.327] 0.117 [0.178]

Visible minority

0.467 [0.093]** 0.034 [0.078] 0.011 [0.014]

0.318 [0.207] 0.127 [0.098] 0.155 [0.086] 0.188 [0.093]* 0.167 [0.082]*

The rest

99

Cut3: Constant

Cut2: Constant

Importance of religion Frequency of attending religious services Cut1: Constant

Trust in police

Trust in neighbours

Trust in general

0.121 [0.046]** 0.223 [0.062]** 0.162 [0.099] 0.126 [0.088] 0.022 [0.100] 1.448 [0.375]** 1.616 [0.373]** 1.762 [0.376]**

1.18 [0.374]** 1.348 [0.370]** 1.492 [0.373]**

Whole sample

0.117 [0.047]* 0.233 [0.063]** 0.162 [0.098] 0.153 [0.087] 0.029 [0.099]

(cont.)

Dependent Variable

Sample

Table 4.2(a)

1.185 [0.724] 1.376 [0.734] 1.511 [0.731]*

0.143 [0.087] 0.069 [0.115] 0.134 [0.161] 0.241 [0.149] 0.156 [0.156]

Union members

Immigrant

1.069 [0.441]* 1.227 [0.433]** 1.378 [0.431]**

0.11 [0.056] 0.298 [0.077]** 0.184 [0.133] 0.12 [0.115] 0.03 [0.124]

0.132 [0.058]* 0.223 [0.071]** 0.131 [0.115] 0.156 [0.099] 0.033 [0.108] 1.505 [0.460]** 1.657 [0.455]** 1.818 [0.467]**

0.121 [0.690] 0.086 [0.689] 0.195 [0.691]

The rest

0.058 [0.110] 0.271 [0.151] 0.183 [0.180] 0.178 [0.205] 0.03 [0.224]

Life satisfaction; 1-10 point scale

The rest

Canadian ESC2, year 2003 Working population

0.002 [0.878] 0.224 [0.864] 0.267 [0.873]

0.08 [0.128] 0.252 [0.187] 0.185 [0.212] 0.401 [0.225] 0.117 [0.258]

Visible minority

1.343 [0.432]** 1.495 [0.431]** 1.679 [0.440]**

0.113 [0.052]* 0.228 [0.071]** 0.108 [0.109] 0.127 [0.092] 0.007 [0.107]

The rest

100

Immigrant

The rest

0.416 2.117 [0.696] [0.464]** 1.104 2.51 [0.697] [0.466]** 1.419 2.864 [0.698]* [0.472]** 2.072 3.613 [0.700]** [0.471]** 2.857 4.577 [0.711]** [0.473]** 3.405 5.2 [0.712]** [0.475]** 539 1979

Life satisfaction; 1-10 point scale

The rest

1.789 1.657 [0.723]* [0.418]** 2.17 2.17 [0.734]** [0.421]** 2.493 2.518 [0.739]** [0.427]** 3.288 3.206 [0.743]** [0.427]** 4.216 4.132 [0.744]** [0.431]** 4.831 4.738 [0.746]** [0.433]** 856 1662

Union members

Canadian ESC2, year 2003 Working population The rest

0.417 1.99 [0.871] [0.432]** 1.039 2.432 [0.868] [0.438]** 1.449 2.761 [0.873] [0.445]** 2.027 3.512 [0.877]* [0.443]** 2.994 4.434 [0.896]** [0.444]** 3.521 5.053 [0.904]** [0.445]** 323 2195

Visible minority

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. Self-employed is excluded in the two Canadian surveys. US Benchmark survey does not provide information on self-employment status. Therefore the sample may include self-employed. 2. The Benchmark survey does not provide information on personal income; household income is used. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Observations

Cut9: Constant

Cut8: Constant

Cut7: Constant

Cut6: Constant

Cut5: Constant

Cut4: Constant

Whole sample

1.769 2.042 [0.366]** [0.372]** 2.24 2.52 [0.369]** [0.377]** 2.578 2.863 [0.376]** [0.383]** 3.294 3.588 [0.375]** [0.382]** 4.212 4.517 [0.378]** [0.384]** 4.815 5.128 [0.377]** [0.385]** 2518 2516

(cont.)

Dependent Variable

Sample

Table 4.2(a)

101

Log of household income

Age group: 45–54

Age group: 35–44

Age group: 25–34

Log of household income Trust in co-workers, standardized score Self-perceived health status, scaled 1 to 5 Male

Dependent Variable

Sample

0.099 [0.029]** 0.179 [0.017]** 0.386 [0.016]** 0.081 [0.026]** 0.219 [0.058]** 0.381 [0.060]** 0.413 [0.065]** 0.099 [0.029]**

0.171 [0.061]** 0.194 [0.031]** 0.407 [0.026]** 0.064 [0.044] 0.302 [0.116]** 0.503 [0.117]** 0.478 [0.120]** 0.171 [0.061]**

Whole sample

Immigrant

0.22 [0.064]** 0.222 [0.041]** 0.461 [0.039]** 0.089 [0.066] 0.579 [0.183]** 0.685 [0.189]** 0.718 [0.194]** 0.22 [0.064]**

Life satisfaction; 1-10 point scale

The rest

0.079 [0.032]* 0.17 [0.021]** 0.38 [0.021]** 0.083 [0.033]* 0.194 [0.067]** 0.327 [0.069]** 0.395 [0.077]** 0.079 [0.032]*

Union members

The rest

0.051 [0.031] 0.166 [0.019]** 0.371 [0.018]** 0.081 [0.030]** 0.19 [0.062]** 0.364 [0.063]** 0.406 [0.067]** 0.051 [0.031]

Canadian ESC2, year 2003 Working population

0.179 [0.055]** 0.198 [0.038]** 0.406 [0.036]** 0.096 [0.056] 0.26 [0.123]* 0.369 [0.128]** 0.423 [0.135]** 0.179 [0.055]**

Visible minority

0.05 [0.033] 0.173 [0.019]** 0.383 [0.018]** 0.074 [0.030]* 0.213 [0.065]** 0.389 [0.066]** 0.428 [0.071]** 0.05 [0.033]

The rest

Table 4.2(b) Expanded models on the well-being effects of income and workplace trust, Canadian GSS-17 survey ordered probit

102

Marital status: Married

Age group: 65 up

Age group: 55–64

Age group: 45–54

Age group: 35–44

Age group: 25–34

0.194 [0.031]** 0.407 [0.026]** 0.064 [0.044] 0.302 [0.116]** 0.503 [0.117]** 0.478 [0.120]** 0.39 [0.129]** 0.402 [0.296] 0.211 [0.064]**

0.17 [0.021]** 0.38 [0.021]** 0.083 [0.033]* 0.194 [0.067]** 0.327 [0.069]** 0.395 [0.077]** 0.283 [0.084]** 0.141 [0.232] 0.277 [0.047]**

0.222 [0.041]** 0.461 [0.039]** 0.089 [0.066] 0.579 [0.183]** 0.685 [0.189]** 0.718 [0.194]** 0.437 [0.198]* 0.393 [0.279] 0.076 [0.093]

0.179 [0.017]** 0.386 [0.016]** 0.081 [0.026]** 0.219 [0.058]** 0.381 [0.060]** 0.413 [0.065]** 0.31 [0.069]** 0.017 [0.185] 0.259 [0.038]**

Immigrant

Trust in co-workers, standardized score Self-perceived health status, scaled 1 to 5 male

Union The rest members (see note) Life satisfaction; 1–10 point scale

Whole sample

0.166 [0.019]** 0.371 [0.018]** 0.081 [0.030]** 0.19 [0.062]** 0.364 [0.063]** 0.406 [0.067]** 0.363 [0.077]** 0.024 [0.220] 0.323 [0.042]**

The rest

Canadian GSS-17, year 2003 working population

Dependent variable

Sample

Table 4.2(b) (cont.)

0.198 [0.038]** 0.406 [0.036]** 0.096 [0.056] 0.26 [0.123]* 0.369 [0.128]** 0.423 [0.135]** 0.217 [0.156] 0.037 [0.412] 0.203 [0.074]**

Visible Minority

0.173 [0.019]** 0.383 [0.018]** 0.074 [0.030]* 0.213 [0.065]** 0.389 [0.066]** 0.428 [0.071]** 0.354 [0.076]** 0.061 [0.205] 0.303 [0.044]**

The rest

103

Education: With university degree Contacts with family member outside household Contacts with friends 0.124 [0.088]

0.221 [0.056]**

0.272 [0.069]**

0.124 [0.124]

0.152 [0.147] 0.365 [0.158]* 0.088 [0.133] 0.532 [0.396] 0.057 [0.160] 0.019 [0.146] 0.158 [0.150] 0.408 [0.090]**

0.166 [0.077]* 0.223 [0.103]* 0.166 [0.088] 0.324 [0.174] 0.348 [0.128]** 0.252 [0.110]* 0.313 [0.115]** 0.187 [0.077]*

0.237 [0.044]** 0.28 [0.066]** 0.105 [0.056] 0.227 [0.124] 0.283 [0.061]** 0.262 [0.054]** 0.378 [0.060]** 0.232 [0.045]**

0.266 [0.055]** 0.319 [0.088]** 0.071 [0.074] 0.157 [0.171] 0.251 [0.068]** 0.251 [0.062]** 0.417 [0.070]** 0.26 [0.055]**

Immigrant

Marital status: As Married Marital status: Divorced Marital status: Separated Marital status: Widowed Education: High school Education: Between

Union The rest members (see note) Life satisfaction; 1–10 point scale

Whole sample

0.238 [0.061]**

0.262 [0.047]** 0.275 [0.074]** 0.142 [0.062]* 0.17 [0.130] 0.333 [0.065]** 0.305 [0.057]** 0.386 [0.064]** 0.163 [0.052]**

The rest

Canadian GSS-17, year 2003 working population

Dependent variable

Sample

Table 4.2(b) (cont.)

0.302 [0.121]*

0.201 [0.106] 0.393 [0.133]** 0.074 [0.124] 0.365 [0.297] 0.16 [0.128] 0.165 [0.124] 0.335 [0.131]* 0.313 [0.086]**

Visible Minority

0.191 [0.062]**

0.258 [0.051]** 0.254 [0.077]** 0.145 [0.063]* 0.187 [0.139] 0.316 [0.068]** 0.285 [0.059]** 0.372 [0.066]** 0.193 [0.052]**

The rest

104

Importance of religion Frequency of attending religious services Cut1: Constant

Confidence in police

Trust in neighbours

Contacts with neighbours Number of membership or extent of activity Trust in general

Dependent variable

Sample

Table 4.2(b) (cont.)

0.199 [0.686]

0.041 [0.052] 0.073 [0.123] 0.338 [0.111]** 0.282 [0.088]** 0.103 [0.088]

0.076 [0.031]* 0.243 [0.069]** 0.342 [0.065]** 0.189 [0.051]** -0.066 [0.056]

0.448 [0.330]

0.115 [0.077] 0.012 [0.068]

0.6 [0.365]

0.093 [0.039]* 0.325 [0.083]** 0.336 [0.076]** 0.136 [0.063]* 0.041 [0.069]

0.049 [0.059] 0.03 [0.040]

Union The rest members (see note)

0.067 [0.047] 0.021 [0.033]

Whole sample

1.073 [0.723]

0.157 [0.086] 0.212 [0.160] 0.401 [0.139]** 0.162 [0.143] 0.085 [0.144]

0.087 [0.115] 0.014 [0.079]

Life satisfaction; 1–10 point scale

Immigrant

1.127 [0.359]**

0.054 [0.033] 0.256 [0.078]** 0.354 [0.073]** 0.206 [0.056]** 0.05 [0.060]

0.055 [0.052] 0.01 [0.037]

The rest

Canadian GSS-17, year 2003 working population

0.425 [0.615]

-0.045 [0.071] 0.077 [0.139] 0.101 [0.126] 0.063 [0.114] 0.018 [0.126]

0.109 [0.100] 0.025 [0.070]

Visible Minority

0.998 [0.375]**

-0.08 [0.034]* 0.302 [0.082]** 0.435 [0.074]** 0.233 [0.057]** 0.08 [0.061]

0.041 [0.054] 0.02 [0.038]

The rest

105

Observations

Cut9: Constant

Cut8: Constant

Cut7: Constant

Cut6: Constant

Cut5: Constant

Cut4: Constant

Cut3: Constant

Cut2: Constant

Dependent variable

Sample

Table 4.2(b) (cont.)

0.128 [0.317] 0.219 [0.309] 0.461 [0.311] 1.105 [0.312]** 1.543 [0.312]** 2.357 [0.314]** 3.376 [0.315]** 4.144 [0.317]** 8794

Whole sample

0.543 [0.660] 0.812 [0.647] 1.078 [0.651] 1.66 [0.648]* 2.131 [0.648]** 3.013 [0.649]** 4.055 [0.649]** 4.858 [0.652]** 3239

0.294 [0.352] 0.096 [0.347] 0.327 [0.347] 1.002 [0.348]** 1.426 [0.348]** 2.208 [0.350]** 3.218 [0.352]** 3.971 [0.354]** 5555

Union The rest members (see note)

1.18 [0.727] 1.639 [0.718]* 1.901 [0.718]** 2.59 [0.720]** 2.977 [0.719]** 3.757 [0.724]** 4.829 [0.730]** 5.574 [0.734]** 1343

Life satisfaction; 1–10 point scale

Immigrant

0.689 [0.336]* 0.378 [0.329] 0.14 [0.330] 0.498 [0.330] 0.956 [0.331]** 1.787 [0.332]** 2.799 [0.333]** 3.574 [0.335]** 7451

The rest

Canadian GSS-17, year 2003 working population

0.75 [0.589] 1.139 [0.591] 1.37 [0.596]* 1.998 [0.596]** 2.403 [0.594]** 3.167 [0.599]** 4.119 [0.603]** 4.832 [0.607]** 1729

Visible Minority

0.673 [0.361] 0.343 [0.351] 0.091 [0.351] 0.568 [0.352] 1.022 [0.353]** 1.859 [0.354]** 2.899 [0.355]** 3.685 [0.356]** 7065

The rest

106

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. In GSS, the union members are identified from memberships in union or professional associations. In order to filter out members of professional associations, both self-employed and those with high income (>$90 000, or top 5% earners) are excluded from the definition of union members. 2. The GSS does not provide very detailed information on ethnicity. The visible minority that is identifiable from the survey is limited to Asians and Aboriginals. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Table 4.2(b) (cont.)

107 0.355 [0.013]** 0.116 [0.027]** 0.01 [0.046] 0.035 [0.045] 0.041 [0.057]

0.352 [0.031]** 0.147 [0.069]* 0.215 [0.140] 0.288 [0.152] 0.283 [0.140]*

Age group: 45–54

Age group: 35–44

Age group: 25–34

0.103 [0.022]** 0.093 [0.012]**

0.045 [0.064] 0.1 [0.032]**

0.096 [0.022]** 0.093 [0.011]** 0.027 [0.029] 0.354 [0.012]** 0.118 [0.025]** 0.029 [0.043] 0.059 [0.038] 0.063 [0.050]

The rest

Log of household income Trust in co-workers, standardized score Participation in labour organizations Self-perceived health status, scaled 1 to 5 male

Union members Happiness; 1–4 point scale

Whole sample

0.088 [0.023]** 0.124 [0.013]** 0.022 [0.035] 0.367 [0.015]** 0.111 [0.030]** 0.007 [0.053] 0.056 [0.047] 0.062 [0.054]

Non-Hispanic white

US Benchmark survey, year 2000/01 Working population

Dependent Variable

Sample

Table 4.2(c) Expanded models on the well-being effects of income and workplace trust, US Benchmark survey, survey ordered probit

0.113 [0.043] 0.051 [0.020]* 0.145 [0.051]** 0.324 [0.023]** 0.111 [0.042]* 0.07 [0.054] 0.039 [0.063] 0.031 [0.110]

The rest

108

Marital status: Married Marital status: As married Marital status: Divorced Marital status: Separated Marital status: Widowed Education: High school Education: Between

Age group: 65 up

Age group: 55–64

Dependent variable

Sample

Table 4.2(c) (cont.)

0.024 [0.045] 0.068 [0.081] 0.264 [0.037]** 0.128 [0.049]* 0.043 [0.039] 0.226 [0.066]** 0.057 [0.072] 0.115 [0.048]* 0.118 [0.052]*

Whole sample

0.295 [0.184] 0.222 [0.274] 0.331 [0.108]** 0.132 [0.158] 0.018 [0.144] 0.279 [0.210] 0.105 [0.191] 0.001 [0.156] 0.093 [0.174]

Union members

0.067 [0.050] 0.095 [0.093] 0.255 [0.034]** 0.128 [0.047]** 0.043 [0.041] 0.224 [0.071]** 0.046 [0.077] 0.132 [0.051]* 0.123 [0.057]*

Happiness; 1–4 point scale

The rest

0.05 [0.058] 0.076 [0.106] 0.304 [0.042]** 0.203 [0.051]** 0.037 [0.046] 0.296 [0.101]** 0.123 [0.081] 0.017 [0.093] 0.028 [0.101]

Non-Hispanic White

US Benchmark survey, year 2000/01 Working population

0.001 [0.082] 0.212 [0.245] 0.205 [0.058]** 0.02 [0.088] 0.014 [0.088] 0.154 [0.081] 0.126 [0.136] 0.184 [0.058]** 0.168 [0.069]*

The rest

109

Trust in police

Trust in neighbours

Contacts with neighbours Number of membership or extent of activity Trust in general

Education: With university degree Contacts with family member outside household Contacts with friends

Dependent Variable

Sample

Table 4.2(c) (cont.)

0.041 [0.183] 0.214 [0.084]* 0.12 [0.031]** 0.321 [0.118]** 0.148 [0.183] 0.203 [0.055]** 0.339 [0.145]* 0.37 [0.135]**

0.111 [0.010]** 0.152 [0.048]** 0.301 [0.064]**

0.175 [0.022]** 0.25 [0.048]** 0.348 [0.040]**

Union members

0.036 [0.054] 0.161 [0.032]**

Whole sample

0.169 [0.023]** 0.238 [0.049]** 0.344 [0.043]**

0.109 [0.010]** 0.131 [0.050]* 0.329 [0.061]**

0.045 [0.055] 0.152 [0.035]**

Happiness; 1–4 point scale

The rest

0.173 [0.030]** 0.252 [0.050]** 0.445 [0.056]**

0.139 [0.012]** 0.156 [0.059]* 0.276 [0.076]**

0.07 [0.095] 0.141 [0.038]**

Non-Hispanic white

US Benchmark survey, year 2000/01 Working population

0.187 [0.048]** 0.295 [0.086]** 0.24 [0.064]**

0.058 [0.021]** 0.156 [0.058]** 0.301 [0.092]**

0.126 [0.082] 0.203 [0.064]**

The rest

110

0.529 [0.221]* 1.583 [0.230]** 3.933 [0.226]** 13293

0.2 [0.053]** 0.167 [0.034]**

Whole sample

0.298 [0.691] 0.857 [0.707] 3.206 [0.707]** 1718

0.029 [0.186] 0.112 [0.111]

Union members

0.632 [0.229]** 1.673 [0.235]** 4.027 [0.230]** 11575

0.231 [0.051]** 0.174 [0.041]**

Happiness; 1–4 point scale

The rest

0.505 [0.258] 1.566 [0.258]** 4.005 [0.259]** 9962

0.217 [0.056]** 0.109 [0.037]**

Non-Hispanic white

US Benchmark survey, year 2000/01 Working population

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. Self-employed is excluded in the two Canadian surveys. US benchmark survey does not provide information on self-employment status. Therefore the sample may include self-employed. 2. The Benchmark survey does not provide information on personal income; household income is used. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Observations

Cut3: Constant

Cut2: Constant

Frequency of attending religious services Cut1: Constant

Importance of religion

Dependent Variable

Sample

Table 4.2(c) (cont.)

0.619 [0.463] 1.678 [0.466]** 3.882 [0.473]** 3331

0.166 [0.102] 0.305 [0.061]**

The rest

How’s the job?

111

The effects of other variables – such as age, marital status, health – on well-being are similar to those found by Helliwell and Putnam (2005). Age shows a U-shaped pattern where well-being is higher for the young and the old than for those in the middle. Married persons have higher well-being than divorced ones. People in better physical health have higher well-being. Respondents with more social capital also have higher well-being; social capital is measured in this case by the extent and intensity of involvement in a variety of family, neighbourhood and community networks. Those with stronger religious beliefs tend to report greater life satisfaction in both countries, supplemented in the US Benchmark sample by positive happiness effects from more frequent involvement in church activities. Income’s effects are significant enough in both surveys to calculate compensating differentials with a reasonable degree of statistical significance – especially in the ESC data. However, more education appears to have no well-being bonus, once its indirect effects flowing through health, income and social capital are separately accounted for. In Table 4.2 we also split our samples in several ways to see if our general findings can be replicated in subsamples of the working population. The first split is between employees who are members of unions and those who are not. This split, which is analysed in more detail in a companion paper (Helliwell and Huang, 2006), has much more salience in Canada than in the United States. As shown by the sample and subsample properties reported in the Appendix, in both countries union members have higher incomes and are better educated and older. They have much lower trust in management (in the ESC, the only survey to ask this question), but do not seem to differ from non-union workers in their degree of trust in their colleagues. Union members are if anything slightly happier, on average, especially in the ESC sample, where the difference is revealed by a significant positive coefficient on the union variable in the first column of Table 4.2. The union share is far higher in the Canadian sample, reflecting the aggregate differences in unionization in the two countries.9 The happiness equations for the Benchmark sample split by union status are in all respects indistinguishable, showing that the general specification is equally applicable to union and non-union members.10 The ESC results show more significant differences in coefficients between the union and non-union samples. The most important of these, for our consideration of workplace trust, is that union members systematically attach more importance to income and less to the climate of workplace trust than do non-members. This may reflect to some extent the results of job matching, since the ESC evidence suggests that the union respondents are in general more highly paid, and have lower levels of trust in management than do the non-union respondents. Thus those who care relatively

112

Social capital: reaching out

more about income than about the climate of workplace trust are more likely than are other workers to either seek or remain in a union environment. That is one interpretation of the pattern of coefficient differences between the union and non-union workers samples by the ESC. The differences are substantial, as the compensating differential for workplace trust is more than twice as large for non-union as for union respondents. It is essentially only for these two key coefficients that the life satisfaction results differ significantly between the union and non-union subsamples of the ESC. Similar patterns are revealed by the GSS results. Union members assign greater weight on income, and as a result have lower compensating differential for trust in colleagues. We turn next to samples split by immigration and ethnic status. The Benchmark survey does not include a split based on immigration status, so the only split is based on ethnicity, with non-Hispanic whites in one group and the rest of the population in the other. The ESC sample is split two different ways: immigrant and Canadian-born, and visible minorities (nonwhites) and the rest of the sample. In the ESC sample, as shown in the Appendix, both the immigrant and the visible minority employee subsamples have lower life satisfaction but equal job satisfaction, and greater trust in management,11 than the rest of the population. The immigrant and visible minority subsamples also have substantially more respondents with university degrees. The ESC sample provides some evidence that both the immigrant and visible minority respondents value income less highly, and the climate of workplace trust more highly, than do other respondents.12 Most other coefficients are similar in the two groups. However, the same pattern is not present in the results from either the US Benchmark or the Canadian GSS. In the US study, non-Hispanic whites assign much greater weight to workplace trust than does the rest of the population. The latter attach greater weight to income instead. Part of the variation may come from differences in income gaps between Canada and the US. While the income gap exists in both countries, it is much wider in the United States. In the GSS, the income coefficients are small and weak for Canadian-born and those who are not part of the visible minority. However, the imprecise income coefficients produce large standard errors on the compensating differentials. In sum, while there are some striking differences between the various subsamples, we consistently find workplace trust as one of the most important variables in explaining well-being, across groups of populations, across surveys and across countries. Given the importance of workplace trust for so many workers, and its effect on productivity, the question remains: why do any workplaces exhibit low levels of trust? We address this question next.

How’s the job?

113

WHAT EXPLAINS DIFFERENCES IN WORKPLACE TRUST? Because workplace trust has such large well-being effects, it is important to discover how such trust is determined, and how it might be built. To do so in a satisfactory manner would require workplace-based surveys to supplement our current surveys, which provide little by way of workplace vs job characteristics. Nonetheless, we gleaned what we could from our current data (see Table 4.3). The strongest variable explaining differences in trust in management is the respondent’s membership in a union. When we split the sample between union and non-union members, we find that unionized workplaces, or at least unionized employees, have lower trust in management. This may reflect the fact that the organizing climate for unions may be better in workplaces where management is not trusted. The adversarial nature of bargaining may also encourage and sustain bonding social capital among union members and discourage trust of management. Interestingly, the mistrust of management does not reflect general mistrust; in fact, union members generally have slightly higher social trust than non-members (see Appendix). Trust equations (Table 4.3) also support this finding by showing that non-union employees are the only ones with a positive linkage between general trust and trust in management. On the ESC survey, respondents with higher incomes systematically rate trust in management lower. This is true both in simple regression (t 7.7) as well as the more fully specified equations shown in Table 4.3. Not surprisingly, trust in management is rated higher by those who do not expect to lose their jobs in the next 12 months (t 7.6 in a simple regression). Those who have changed jobs more frequently in the past 12 months rate trust in their current jobs slightly higher than do others, perhaps reflecting the role that the expected trust climate had on their job choice. Diversity in the workplace – as measured by education, age, religion, ethnicity, language, income or social class – has no effect in the Canadian climate of trust in management. Only a few differences appear when the sample is split by immigration or visible minority status. The most striking one is that the number of job moves in the last 12 months correlates negatively with trust in management; suggesting that immigrants and visible minorities may have less opportunity to move to workplaces characterized by higher trust in management. However, because these respondents rate management’s trustworthiness more highly than other respondents (see Appendix), they may have less reason to search for and move to higher-trust workplaces. Overall different patterns emerge for the effects of job characteristics on workplace trust and life satisfaction. Several characteristics of the

114

Social capital: reaching out

respondent’s job, which may well be shared with other jobs in their workplace, correlate significantly with the climate of workplace trust. Jobs with a variety of tasks and with adequate time available to perform them have positive effects. Positions that are free of conflicting demands and involve a higher degree of decision-making produce even stronger increases in trust. Such jobs may have more managerial tasks and a more favourable view of management (see Table 4.3). This engagement may be more important than closeness to management, because the decision-making effect is just as strong for union members as for non-members. Interestingly, this link does not appear in the visible minority and immigration groups. Other individual variables with significant linkages to the assessments of workplace trust include health, and frequency of church attendance. Skill content is not significant for workplace trust. This contrasts with life satisfaction, which correlates positively with jobs that require skills but not with those requiring decision-making (see Table 4.2). The latter difference may be because the strains of decision-making make the job more engaging, but increase the costs on the home front (Helliwell and Huang, 2005). Only workplace-specific data will permit us to separate the effects of the respondent’s own income from those of a workplace with higher paid jobs. The GSS and the US Benchmark ask respondents to evaluate trust in their co-workers. Unlike the findings from the ESC, which measured trust in management, union members and non-members show similar levels of trust in co-workers. In the ESC, the cross-group difference was half a standard deviation (or less). The US Benchmark survey also asked about the frequency of socialization with co-workers outside of work. Not surprisingly, this variable is positively and significantly correlated with trust in colleagues. The positive effect of general trust is also not surprising, as workplace trust is a key source of evidence for each respondent’s overall assessment of the extent to which others can be trusted.

WHY MIGHT UNDERINVESTMENT BE OCCURRING, AND WHAT MIGHT BE DONE ABOUT IT? The question remains: if this is true, then why have companies not created more workplaces with high levels of trust? If this is to reflect a fully informed equilibrium, then it must be correspondingly expensive to change the climate of trust in management within a workplace, and for employees to find and move to more trustworthy workplaces. Alternatively, perhaps excessive emphasis on pay as the principal benefit of employment, and the

How’s the job?

115

lack of information about the importance of workplace trust and social capital, may also have deflected what might otherwise have been a systematic search for ways to improve life in the workplace. Pay, which is more easily measured and compared than social relations, is often seen as the most important way of rewarding employees. This misperception is probably exacerbated by the content of university courses in economics (Rabin, 2002), management training and the general structure of management, which concentrate on conventionally measured economic outcomes. Human relations departments, who might make the case for social capital and workplace trust, are frequently on the decision-making periphery, expected to find new employees to replace those who leave, and to hold birthday and retirement parties for those who remain. There are experimentally well-established features of decision-making that might also explain why firms and employees pay inadequate attention to workplace trust and social capital. In their quasi-experimental study of the well-being effects of lottery outcomes for undergraduate houses, Dunn et al. (2003) found, as expected, that students would focus too much on the obvious (but less important) physical features of the different houses, and not enough on subtle (but more important) differences in their social environments. This was the case even though students, when asked to rank the relative importance of social and physical features of their environments, were aware that the social environment was more important. In a similar vein, Frey and Stutzer (2004), in their survey of the related literature, conclude that: ‘Goods and activities characterized by stronger intrinsic attributes (such as spending time with family and friends and pursuing hobbies) are undervalued compared to those characterized by stronger extrinsic attributes (such as most consumer goods).’ Two other factors – the silo-like separation of management functions and the emphasis on delivering short-term profits – can reduce the capacity or will to find and repair situations of low trust. We have ourselves encountered situations where even experienced senior executives have been uneasy about learning too much about the climate of trust within their organizations. This is an understandable risk-avoidance strategy, but is quite likely to be counterproductive. Enduring trust – and often higher productivity – is best built on the mutual willingness to collaborate in weaving the fabric of mutual trust before it starts to crumble. Recent measures designed to reduce the perceived risks of corporate misbehaviour or malfeasance may unintentionally lower trust and exacerbate some of the situations described above (what might be described as Sarbanes–Oxley effects). These measures are often introduced to reduce the risks of malfeasance or at least control the potential legal liabilities. Nevertheless, they may have inadvertently damaged the climate of trust by

116

: Religion

Job: Has variety of tasks Job: Have enough time Job: Free of conflicting demands Job: Makes own decision Commonality with co-workers : Social class or income bracket : Language group

Log of personal income Job: Requires skill

0.272 [0.094]** 0.33 [0.204] 0.018 [0.197] 0.122 [0.144] 0.529 [0.150]** 0.611 [0.147]**

0.034 [0.057] 0.063 [0.044] 0.082 [0.045]

0.025 [0.028] 0.006 [0.024] 0.047 [0.025]

Union members

0.183 [0.043]** 0.203 [0.110] 0.288 [0.101]** 0.327 [0.093]** 0.59 [0.081]** 0.526 [0.092]**

Whole sample

Immigrant

The rest

0.033 [0.032] 0.007 [0.030] 0.036 [0.029]

0.134 [0.052]* 0.139 [0.132] 0.356 [0.119]** 0.485 [0.115]** 0.649 [0.092]** 0.482 [0.120]**

0.09 [0.058] 0.025 [0.046] 0.018 [0.045]

0.116 [0.083] 0.05 [0.212] 0.614 [0.190]** 0.303 [0.168] 0.319 [0.176] 0.215 [0.188]

0.008 [0.032] 0.025 [0.032] 0.068 [0.028]*

0.235 [0.057]** 0.239 [0.129] 0.27 [0.127]* 0.333 [0.109]** 0.629 [0.086]** 0.68 [0.108]**

Workplace trust, 1–10 point scale

The rest

Canadian ESC2; Working population

Formation of workplace trust, ESC2 sample survey ordered probit

Dependent Variable

Sample

Table 4.3(a)

0.015 [0.076] 0.068 [0.061] 0.047 [0.071]

0.198 [0.088]* 0.03 [0.301] 0.691 [0.260]** 0.076 [0.237] 0.449 [0.244] 0.456 [0.269]

Visible minority

0.028 [0.033] 0.013 [0.027] 0.068 [0.026]*

0.204 [0.050]** 0.227 [0.114]* 0.224 [0.112]* 0.346 [0.105]** 0.598 [0.089]** 0.577 [0.099]**

The rest

117

(cont.)

Age group: 45–54

Age group: 35–44

Age group: 25–34

Dummy, union member Sense of job security, 1–4 point Self-perceived health status, scaled 1 to 5 male

: Level of education

: Age

: Ethnicity

Dependent Variable

Sample

Table 4.3(a)

0.019 [0.026] 0.022 [0.027] 0.028 [0.027] 0.54 [0.055]** 0.172 [0.025]** 0.112 [0.030]** 0.109 [0.047]* 0.053 [0.102] 0.029 [0.103] 0.1 [0.106]

Whole sample

0.221 [0.047]** 0.057 [0.057] 0.117 [0.088] 0.223 [0.226] 0.128 [0.235] 0.137 [0.227]

0.006 [0.042] 0.066 [0.045] 0 [0.051]

Union members

Immigrant

The rest

0.14 [0.031]** 0.145 [0.036]** 0.098 [0.059] 0.128 [0.109] 0.055 [0.115] 0.206 [0.122]

0.015 [0.032] 0 [0.033] 0.031 [0.032]

0.054 [0.055] 0.072 [0.057] 0.05 [0.060] 0.62 [0.125]** 0.142 [0.056]* 0.03 [0.067] 0.162 [0.115] 0.071 [0.217] 0.058 [0.232] 0.002 [0.237] 0.056 [0.030] 0.019 [0.031] 0.034 [0.030] 0.512 [0.059]** 0.2 [0.030]** 0.127 [0.033]** 0.095 [0.054] 0.096 [0.117] 0.123 [0.117] 0.199 [0.119]

Workplace trust, 1–10 point scale

The rest

Canadian ESC2; Working population

0.11 [0.080] 0.057 [0.092] 0.046 [0.087] 0.426 [0.173]* 0.187 [0.081]* 0.143 [0.083] 0.05 [0.134] 0.148 [0.234] 0.102 [0.266] 0.022 [0.268]

Visible minority

0.025 [0.027] 0.013 [0.030] 0.04 [0.028] 0.551 [0.056]** 0.171 [0.028]** 0.125 [0.032]** 0.121 [0.052]* 0.061 [0.119] 0.062 [0.117] 0.153 [0.121]

The rest

118

Education: With university degree

Marital status: Married Marital status: As married Marital status: Divorced Marital status: Separated Marital status: Widowed Education: High school Education: Between

Age group: 65 up

0.058 [0.255] 0.086 [0.710] 0.033 [0.119] 0.024 [0.168] 0.01 [0.165] 0.04 [0.182] 0.29 [0.392] 0.371 [0.166]* 0.386 [0.153]* 0.174 [0.160]

The rest

Immigrant

The rest

0.23 [0.142] 0.112 [0.405] 0.058 [0.080] 0.042 [0.121] 0.289 [0.144]* 0.048 [0.128] 0.272 [0.260] 0.164 [0.138] 0.268 [0.131]* 0.236 [0.135]

0.294 [0.152] 0.294 [0.420] 0.012 [0.070] 0.026 [0.108] 0.187 [0.128] 0.038 [0.110] 0.213 [0.204] 0.243 [0.122]* 0.324 [0.116]** 0.139 [0.125]

0.045 [0.271] 0.524 [0.583] 0.226 [0.150] 0.358 [0.197] 0.078 [0.235] 0.135 [0.236] 1.354 [0.491]** 0.149 [0.250] 0.348 [0.236] 0.496 [0.239]*

0.19 [0.124] 0.268 [0.355] 0.025 [0.068] 0.02 [0.099] 0.156 [0.107] 0.031 [0.097] 0.013 [0.208] 0.229 [0.114]* 0.324 [0.106]** 0.214 [0.111]

Union members

Age group: 55–64

Whole sample

Canadian ESC2; Working population

Workplace trust, 1–10 point scale

(cont.)

Dependent Variable

Sample

Table 4.3(a)

0.314 [0.407] 1.876 [0.651]** 0.035 [0.238] 0.465 [0.393] 0.007 [0.338] 0.064 [0.384] 0.417 [1.048] 0.193 [0.357] 0.276 [0.334] 0.598 [0.372]

Visible minority

0.285 [0.138]* 0.493 [0.364] 0.026 [0.069] 0.012 [0.105] 0.172 [0.111] 0.062 [0.101] 0.009 [0.211] 0.196 [0.116] 0.309 [0.105]** 0.129 [0.112]

The rest

119

(cont.)

Importance of religion Frequency of attending religious services

Trust in police

Trust in neighbours

Contacts with neighbours Number of membership or extent of activity Trust in general

Contacts with family member outside household Contacts with friends

Dependent Variable

Sample

Table 4.3(a)

0.297 [0.167] 0.139 [0.136] 0.003 [0.027] 0.148 [0.112] 0.127 [0.130] 0.248 [0.169] -0.214 [0.180] 0.319 [0.188]

0.159 [0.054]** 0.079 [0.084] 0.123 [0.094] 0.161 [0.095] 0.316 [0.097]**

0.003 [0.132]

Union members

0.147 [0.105] 0.044 [0.076] 0.004 [0.015]

0.158 [0.072]*

Whole sample

Immigrant

The rest

0.163 [0.058]** 0.056 [0.100] 0.094 [0.116] 0.178 [0.122] 0.345 [0.115]**

0.073 [0.121] 0.005 [0.099] 0.011 [0.020]

0.21 [0.093]*

0.113 [0.118] 0.187 [0.158] 0.359 [0.201] 0.061 [0.194] 0.248 [0.193]

0.113 [0.188] 0.07 [0.174] 0.041 [0.034]

0.214 [0.147]

0.191 [0.061]** 0.004 [0.102] 0.112 [0.105] 0.202 [0.116] 0.314 [0.115]**

0.119 [0.123] 0.053 [0.087] 0.021 [0.018]

0.177 [0.086]*

Workplace trust, 1–10 point scale

The rest

Canadian ESC2; Working population

0.037 [0.156] 0.342 [0.207] 0.675 [0.254]** 0.081 [0.268] 0.104 [0.258]

0.137 [0.274] 0.246 [0.246] 0.059 [0.043]

0.029 [0.209]

Visible minority

0.176 [0.059]** 0.063 [0.091] 0.052 [0.102] 0.181 [0.103] 0.36 [0.105]**

0.152 [0.114] 0.108 [0.082] 0.017 [0.017]

0.173 [0.082]*

The rest

120

(cont.)

Cut7: Constant

Cut6: Constant

Cut5: Constant

Cut4: Constant

Cut3: Constant

Cut2: Constant

Number of employers in past 12 months (equals 3 if greater than 3) Year since immigration Cut1: Constant

Dependent Variable

Sample

Table 4.3(a)

1.347 [0.457]** 1.04 [0.453]* 0.678 [0.449] 0.353 [0.450] 0.07 [0.453] 0.469 [0.452] 0.991 [0.452]*

0.1 [0.051]

Whole sample

2.046 [1.124] 1.651 [1.118] 1.279 [1.113] 0.935 [1.114] 0.496 [1.117] 0.114 [1.112] 0.458 [1.113]

0.291 [0.112]**

Union members

Immigrant

The rest

0.643 [0.521] 0.412 [0.516] 0.049 [0.508] 0.268 [0.512] 0.69 [0.518] 1.111 [0.517]* 1.615 [0.517]**

0.031 [0.056] 0.007 [0.004] 2.087 [0.907]* 1.897 [0.912]* 1.602 [0.909] 1.254 [0.907] 0.717 [0.918] 0.206 [0.908] 0.299 [0.911]

0.223 [0.092]*

1.148 [0.612] 0.808 [0.605] 0.423 [0.602] 0.093 [0.604] 0.316 [0.607] 0.699 [0.608] 1.241 [0.610]*

0.21 [0.065]**

Workplace trust, 1–10 point scale

The rest

Canadian ESC2; Working population

2.482 [0.989]* 1.981 [0.970]* 1.676 [0.963] 1.274 [0.944] 0.815 [0.949] 0.249 [0.933] 0.278 [0.935]

0.163 [0.116]

Visible minority

0.202 [0.544]* 0.907 [0.539] 0.53 [0.535] 0.206 [0.535] 0.221 [0.536] 0.603 [0.537] 1.132 [0.536]*

0.148 [0.061]*

The rest

121

(cont.)

1.641 [0.453]** 2.204 [0.451]** 2150

Whole sample

1.086 [1.112] 1.731 [1.097] 735

Union members

Immigrant

The rest

2.282 [0.519]** 2.829 [0.518]** 1415

1.011 [0.910] 1.573 [0.903] 458

1.891 [0.609]** 2.469 [0.608]** 1692

Workplace trust, 1–10 point scale

The rest

Canadian ESC2; Working population

0.972 [0.934] 1.469 [0.915] 280

Visible minority

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. ESC2 sample has been filtered out the self-employed. US benchmark survey does not provide information on self-employment status. Therefore the sample include self-employed. 2. The Benchmark survey does not provide information on personal income, therefore household income is used. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Observations

Cut9: Constant

Cut8: Constant

Dependent Variable

Sample

Table 4.3(a)

1.785 [0.536]** 2.367 [0.534]** 1870

The rest

122

Marital status:

Age group: 65 up

Age group: 55–64

Age group: 45–54

Age group: 35–44

Age group: 25–34

Log of household income Union status, to be added Self-perceived health status, scaled 1 to 5 male

0.159 [0.029]** 0.008 [0.029] 0.092 [0.016]** 0.087 [0.029]** 0.124 [0.060]* 0.114 [0.059] 0.025 [0.063] 0.151 [0.073]* 0.503 [0.213]* 0.058

0.178 [0.034]**

0.116 [0.020]** 0.1 [0.038]** 0.141 [0.072]* 0.115 [0.071] 0.075 [0.075] 0.184 [0.089]* 0.635 [0.273]* 0.016

0.087 [0.056]

0.047 [0.027] 0.051 [0.045] 0.01 [0.115] 0.015 [0.115] 0.142 [0.119] 0.172 [0.134] 0.291 [0.314] 0.136

Whole Union The rest sample members (see note)

0.075 [0.060] 0.08 [0.070] 0.117 [0.037]** 0.07 [0.067] 0.305 [0.208] 0.204 [0.212] 0.267 [0.209] 0.094 [0.228] 0.191 [0.509] 0.069

Trust in colleagues, 1–5 point scale

Immigrant

0.201 [0.033]** 0.022 [0.032] 0.081 [0.018]** 0.083 [0.031]** 0.113 [0.064] 0.112 [0.062] 0.012 [0.066] 0.206 [0.078]** 0.522 [0.236]* 0.123

The rest

Canadian GSS-17; Working population

Formation of workplace trust, Canadian GSS-17 survey ordered probit

Dependent Variable

Sample

Table 4.3(b)

0.1 [0.055] 0.054 [0.065] 0.139 [0.036]** 0.105 [0.063] 0.179 [0.139] 0.175 [0.141] 0.235 [0.144] 0.363 [0.170]* 0.161 [0.301] 0.128

Visible Minority

0.202 [0.034]** 0.01 [0.033] 0.077 [0.018]** 0.076 [0.032]* 0.219 [0.067]** 0.193 [0.064]** 0.11 [0.069] 0.077 [0.078] 0.576 [0.241]* 0.133

The rest

123

Education: With university degree Contacts with family member outside household Contacts with friends

[0.065]* 0.056 [0.078] 0.204 [0.110] 0.002 [0.104] 0.134 [0.204] 0.001 [0.119] 0.041 [0.100] 0.145 [0.104] 0.037 [0.081] 0.253 [0.093]**

[0.039] 0.048 [0.047] 0.109 [0.067] 0.04 [0.068] 0.103 [0.129] 0.046 [0.061] 0.011 [0.052] 0.102 [0.057] 0.048 [0.048]

0.233 [0.057]**

0.22 [0.071]**

[0.048] 0.038 [0.059] 0.065 [0.085] 0.059 [0.090] 0.264 [0.167] 0.062 [0.068] 0.032 [0.060] 0.048 [0.068] 0.086 [0.060]

The rest

Immigrant

0.177 [0.122]

[0.104] 0.087 [0.146] 0.099 [0.163] 0.165 [0.213] 0.83 [0.438] 0.049 [0.171] 0.015 [0.135] 0.103 [0.142] 0.043 [0.112]

Union member (see note)

Married Marital status: As married Marital status: Divorced Marital status: Separated Marital status: Widowed Education: High school Education: Between

Whole sample

0.233 [0.063]**

[0.042]** 0.06 [0.050] 0.118 [0.074] 0.017 [0.067] 0 [0.132] 0.084 [0.064] 0.044 [0.056] 0.18 [0.062]** 0.002 [0.054]

The rest

Canadian GSS-17; Working population

Trust in colleagues, 1–5 point scale

(cont.)

Dependent variable

Sample

Table 4.3(b)

0.227 [0.119]

[0.083] 0.113 [0.112] 0.137 [0.148] 0.286 [0.177] 0.582 [0.303] 0.101 [0.151] 0.063 [0.127] 0.023 [0.134] 0.04 [0.102]

Visible Minority

0.219 [0.066]**

[0.044]** 0.039 [0.052] 0.104 [0.076] 0.032 [0.071] 0.025 [0.139] 0.033 [0.064] 0.03 [0.056] 0.167 [0.062]** 0.066 [0.054]

The rest

124

Importance of religion Frequency of attending religious services Socializing with co-workers

Confidence in police

Trust in neighbours

Contacts with neighbours Number of membership or extent of activity Trust in general 0.472 [0.053]** 2.19 [0.127]** 0.408 [0.114]** 0.262 [0.084]** 0.306 [0.089]** 0.087 [0.020]**

0.069 [0.012]**

0.362 [0.083]** 0.087 [0.067]

0.449 [0.032]** 2.086 [0.079]** 0.384 [0.066]** 0.083 [0.050] 0.093 [0.055]

0.246 [0.053]** 0.007 [0.038]

0.067 [0.015]**

0.437 [0.039]** 2.034 [0.098]** 0.377 [0.082]** 0 [0.064] 0.016 [0.073]

0.198 [0.066]** 0.052 [0.044]

Whole Union The rest sample members (see note)

0.014 [0.027]

0.357 [0.080]** 2.392 [0.215]** 0.205 [0.166] 0.169 [0.130] 0.255 [0.145]

0.1 [0.141] 0.075 [0.095]

Trust in colleagues, 1–5 point scale

Immigrant

0.084 [0.013]**

0.477 [0.035]** 2.004 [0.082]** 0.429 [0.071]** 0.074 [0.055] 0.096 [0.060]

0.279 [0.056]** 0.033 [0.040]

The rest

Canadian GSS-17; Working population

Formation of workplace trust, Canadian GSS-17 survey ordered probit

Dependent Variable

Sample

Table 4.3(b)

0.058 [0.026]*

0.404 [0.069]** 1.971 [0.174]** 0.458 [0.142]** 0.258 [0.119]* 0.347 [0.132]**

0.119 [0.117] 0.077 [0.080]

Visible Minority

0.076 [0.013]**

0.46 [0.036]** 2.121 [0.087]** 0.365 [0.075]** 0.042 [0.055] 0.027 [0.060]

0.292 [0.056]** 0.019 [0.043]

The rest

125

1.77 [0.310]** 0.949 [0.306]** 0.2 [0.306] 1.503 [0.307]** 8773

1.149 [0.600] 0.296 [0.596] 0.92 [0.599] 2.328 [0.600]** 3236

1.884 [0.361]** 1.07 [0.359]** 0.056 [0.358] 1.313 [0.358]** 5537

0.94 [0.623] 0.056 [0.623] 1.106 [0.619] 2.419 [0.622]** 1339

Trust in colleagues, 1–5 point scale

Immigrant

2.258 [0.357]** 1.451 [0.354]** 0.293 [0.354] 1.02 [0.355]** 7434

The rest

0.845 [0.590] 0.023 [0.592] 1.095 [0.587] 2.374 [0.586]** 1725

Visible Minority

2.352 [0.366]** 1.525 [0.364]** 0.357 [0.364] 0.962 [0.365]** 7048

The rest

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. In GSS, the union members are identified from memberships in union or professional associations. In order to filter out members of professional associations, both self-employed and those with high income (>$90 000, or top 5% earners) are excluded from the definition of union members. 2. The GSS does not provide very detailed information on ethnicity. The visible minority that is identifiable from the survey is limited to Asians and Aboriginals. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Observations

Cut4: Constant

Cut3: Constant

Cut2: Constant

Cut1: Constant

Whole Union The rest sample members (see note)

Canadian GSS-17; Working population

Formation of workplace trust, Canadian GSS-17 survey ordered probit

Dependent Variable

Sample

Table 4.3(b)

126

Age group: 65 up

Age group: 55–64

Age group: 45–54

Age group: 35–44

Age group: 25–34

Log of household income Frequency; socialized with co-workers outside of work Participating in labour organizations Self-perceived health status, scaled 1 to 5 male

Dependent variable

Sample

0.027 [0.058] 0.005 [0.002]*

0.062 [0.037] 0.169 [0.068]* 0.116 [0.107] 0.086 [0.104] 0.044 [0.120] 0.133 [0.148] 0.434 [0.235]

0.161 [0.036]** 0.074 [0.013]** 0.01 [0.020] 0.059 [0.041] 0.041 [0.043] 0.057 [0.052] 0.134 [0.065]* 0.323 [0.098]**

Union members

0.035 [0.018] 0.006 [0.001]**

Whole sample

Non-Hispanic White

0.075 [0.013]** 0.017 [0.023] 0.051 [0.041] 0.034 [0.043] 0.058 [0.054] 0.131 [0.067] 0.314 [0.099]**

0.037 [0.019] 0.006 [0.001]** 0.216 [0.039]** 0.098 [0.017]** 0.02 [0.024] 0.061 [0.053] 0.046 [0.062] 0.044 [0.064] 0.069 [0.081] 0.253 [0.116]*

0.091 [0.022]** 0.006 [0.001]**

Trust in colleagues, 1–4 point scale

The rest

US Benchmark; working population

Table 4.3(c) Formation of workplace trust, US Benchmark community survey survey ordered probit

0.038 [0.073] 0.021 [0.019] 0.011 [0.045] 0.048 [0.069] 0.051 [0.071] 0.007 [0.078] 0.244 [0.093]* 0.397 [0.199]

0.016 [0.038] 0.006 [0.001]**

The rest

127

Education: With university degree Contacts with family member outside household

0.046 [0.079] 0.061 [0.131] 0.004 [0.131] 0.128 [0.169] 0.273 [0.227] 0.07 [0.161] 0.268 [0.153] 0.395 [0.183]* 0.019 [0.066]

Non-Hispanic White

0.076 [0.054] 0.079 [0.059] 0.028 [0.058] 0.029 [0.090] 0.04 [0.128] 0.04 [0.088] 0.038 [0.089] 0.198 [0.085]* 0.012 [0.036]

0.075 [0.047] 0.115 [0.043]* 0.066 [0.045] 0.064 [0.051] 0.014 [0.107] 0.11 [0.051]* 0.196 [0.060]** 0.339 [0.054]** 0.014 [0.032]

0.07 [0.043] 0.107 [0.046]* 0.057 [0.047] 0.073 [0.046] 0.048 [0.102] 0.099 [0.052] 0.201 [0.058]** 0.348 [0.062]** 0.012 [0.029]

The rest

Marital status: Married Marital status: As married Marital status: Divorced Marital status: Separated Marital status: Widowed Education: High school Education: Between

Union members Trust in colleagues, 1–4 point scale

Whole sample

US Benchmark; working population

Dependent Variable

Sample

Table 4.3(c) (cont.)

0.011 [0.057] 0.158 [0.062]* 0.075 [0.089] 0.082 [0.089] 0.032 [0.136] 0.083 [0.077] 0.24 [0.079]** 0.31 [0.088]** 0.019 [0.067]

The rest

128

Non-Hispanic White

Importance of religion Frequency of attending religious services Cut1: Constant

Trust in police

Trust in neighbours

Contacts with neighbours Number of membership or extent of activity Trust in general

0.143 [0.615]

0.535 [0.058]** 0.895 [0.161]** 0.9 [0.115]** 0.162 [0.152] 0.079 [0.096]

0.541 [0.021]** 0.892 [0.047]** 0.863 [0.051]** 0.038 [0.065] 0.022 [0.030]

0.115 [0.210]

0.068 [0.027]* 0.068 [0.101] 0.275 [0.134]*

0.117 [0.011]** 0.012 [0.038] 0.093 [0.069]

0.126 [0.225]

0.542 [0.025]** 0.893 [0.053]** 0.86 [0.058]** 0.072 [0.067] 0.017 [0.030]

0.941 [0.239]**

0.546 [0.029]** 0.758 [0.057]** 0.86 [0.056]** 0.025 [0.069] 0.028 [0.033]

0.092 [0.013]** 0.089 [0.056] 0.257 [0.074]**

0.126 [0.013]** 0.006 [0.042] 0.064 [0.077]

The rest

Contacts with friends

Union members Trust in colleagues, 1–4 point scale

Whole sample

US Benchmark; working population

Dependent Variable

Sample

Table 4.3(c) (cont.)

0.209 [0.385]

0.469 [0.054]** 0.9 [0.089]** 0.711 [0.072]** 0.086 [0.128] 0.003 [0.068]

0.13 [0.031]** 0.078 [0.053] 0.058 [0.116]

The rest

129

0.801 [0.212]** 2.044 [0.211]** 13242

Whole sample

0.985 [0.604] 2.237 [0.601]** 1708

Union members

Non-Hispanic White

0.806 [0.224]** 2.049 [0.223]** 11534

0.152 [0.234] 1.109 [0.234]** 9928

Trust in colleagues, 1–4 point scale

The rest

US Benchmark; working population

Notes: Standard errors in brackets. * significant at 5%; ** significant at 1%. 1. ESC2 sample has been filtered out the self-employed. US benchmark survey does not provide information on self-employment status therefore the sample includes self-employed. 2. The Benchmark survey does not provide information on personal income, therefore household income is used. 3. Please refer to Appendix Table 4A.1 for descriptive statistics by samples.

Observations

Cut3: Constant

Cut2: Constant

Dependent Variable

Sample

Table 4.3(c) (cont.)

1.27 [0.391]** 2.524 [0.388]** 3314

The rest

130

Social capital: reaching out

increasing reporting burdens while simultaneously creating the perception that employees and managers are not to be trusted. Lack of trust breeds suspicion, but actions designed to assuage suspicion may have the corollary effect of lessening the extent to which employees and managers feel themselves to be collaboratively engaged. A final element of human psychology that might help to explain chronic underinvestment in workplace trust and social capital is present-biased preferences (O’Donoghue and Rabin, 1999), which represent an excessive concentration on the short term. Within economics, this is part of the explanation for the coexistence of gambling and insurance, and the widespread incidence of addictive behaviour and of high-interest credit-card debt. In the case of trust and social capital, this is complicated by the asymmetry whereby it takes long-sustained investment to build trust but only a few moments of careless inattention, or a longer period of benign neglect, to destroy it. So what can be done to maintain and improve workplace trust? Our results do not speak directly to the design of strategies to build and maintain social capital in the workplace, but the empirical and well-being literature suggests a few general approaches. All of these invite and need a stronger experimental base before they are implemented. The first and simplest strategy might be to assess, or reassess, each major change in administrative procedures, especially those designed to tighten administrative controls and increase reporting. The important thing is to ensure that they are done in such a way as to increase rather than decrease trust. This follows the ‘do no harm’ principle, and needs to be implemented before, rather than after, such changes are made, as it is easier to damage trust than to restore it. A second strategy might be to encourage and document experimental approaches and initiatives in decision-making and reporting structures within workplaces that may yield positive results. The lack of solid research on the creation and maintenance of social capital within the workplace needs to be filled. For an experimental strategy to work, it must be recognized by all that many or even most of the trials will fail. That is the whole point of an experimental strategy – only by trying many alternatives will the right answers come to light. Third, strategies to build social capital and engender trust are almost sure to be more effective if they are driven by innovation and energy from the bottom rather than the top of the organization. Why? Life satisfaction is based on matching feelings of engagement and efficacy. For employees to feel both engaged and efficacious, they must be convinced that they have had an instrumental role in design and an effective role in delivering what happens in the workplace. This sharing of both design and execution could build social capital and trust of a high order.

131

Whole sample

Whole sample*

Union members

Canadian ESC2, year 2003 Working population The rest

Immigrant

Compensating differentials, derived from estimates of Table 4.2

Whole sample

Union members

The rest

Canadian GSS-17, year 2003 Working population

Coefficient of the log of household income 0.099 0.171 0.079 Coefficient of the standardized score of trust in management 0.179 0.194 0.17 Ratio of coefficients 1.81 1.13 2.15

Panel-b

0.051 0.166 3.25

0.222 1.01

The rest

0.22

Immigrant

Coefficient of the log of personal income 0.19 0.154 0.22 0.18 0.12 Coefficient of the standardized score of trust in management 0.19 0.165 0.14 0.24 0.27 Ratio of coefficients 0.97 1.07 0.65 1.36 2.29 Approximated percentage income change associated with one third of a standard deviation, such a movement covers roughly 10% of the sample** 32% 36% 22% 45% Statistically Standard error*** 7% 9% 8% 11% insignificant

Panel-a

Table 4.4

2.93

0.79

1.11

0.198

0.179

Identifiable minority

3.46

0.173

0.05

The rest

Statistically insignificant

0.37

0.17

26% 6%

0.13

Visible minority

0.22

The rest

27% 6%

0.82

0.17

0.21

The rest

132

(cont.)

Whole sample

Union members

The rest

US Benchmark survey, year 2000/01 Working population White nonHispanic

The rest

37% 13%

Statistically insignificant

Notes: * this particular regression includes other job characteristics as the right hand side variables. ** this is simply one third of the ratio of coefficients. *** standard error is calculated from the delta method from the variance co-variance matrix of the estimated coefficients.

Coefficient of the Log of household income 0.10 0.05 0.10 0.09 0.11 Coefficient of the standardized score of trust in management 0.09 0.10 0.09 0.12 0.05 Ratio of coefficients 0.97 2.22 0.90 1.41 0.45 Percentage income change associated with one third of a standard deviation, such a movement covers roughly 10% of the sample** 32% Statistically 30% 47% Statistically Standard error*** 9% insignificant 8% 13% insignificant

Panel-c

Percentage income change associated with one third of a standard deviation, such a movement covers roughly 10% of the sample** 60% 38% 72% 34% Statistically Standard error*** 18% 14% 30% 12% insignificant

Table 4.4

How’s the job?

133

NOTES 1.

We are grateful for invaluable access to General Social Survey (GSS) and Equality, Security and Community Survey (ESC) data provided through the University of British Columbia (UBC) Research Data Centre supported by Statistics Canada, the Social Sciences and Humanities Research Council, and UBC. This is part of the Social Interactions, Identity and Well-Being research programme of the Canadian Institute for Advanced Research, and the research is also supported by grants from the Social Sciences and Humanities Research Council of Canada. 2. See, for example, Knack and Keefer (1997), Helliwell (1996). 3. See Putnam (2000) and Halpern (2005). 4. This is demonstrated more explicitly by Hwang et al. (1992) and Helliwell and Huang (2005). 5. A very similar approach has been used by Frey and Stutzer (2004) to value commuting time in Switzerland, and by Van Praag and Baarsma (2005) to estimate compensating differentials for aircraft noise in the neighbourhoods surrounding Amsterdam Airport. 6. The original survey response from ESC2 has a ten-point scale. The GSS has a five-point scale. The US Benchmark has a four-point scale. After standardization, all these trust variables have zero mean and a standard deviation of one. The ratio of the coefficient of the standardized trust measures to that of income is the estimate of the income equivalent (or compensating differential) for a movement of one standard deviation. Starting from the mean, such a movement covers about 35 per cent of the sample. 7. There are some other non-comparabilities among the three surveys, as the ESC2 and the GSS measure well-being using a ten-point assessment of life satisfaction while the Benchmark Survey employs a four-point assessment of happiness. Helliwell and Putnam (2005) show that happiness and life satisfaction questions do produce slightly different response patterns, when asked of the same subjects, although the signs and sizes of most of the estimated coefficients are similar. Using ratios of standardized coefficients deals reasonably well with the issue posed by using different scales for the answers to the trust questions in these surveys. The ESC sample excludes the self-employed, an adjustment that was not possible for the Benchmark sample. Furthermore, in the ESC2, we use the preferred measure of personal income, but in the US Benchmark, the only income measure is the household income. In the GSS, we use household income for consistency with our earlier papers. 8. We adjusted the mastery scale to remove its correlation with income. We had observed that the effects of income on life satisfaction are much smaller when the unadjusted mastery scale is included in the equation. This may be because those with better coping personalities are more able to find and hold higher-paying positions. It may also be because those who have higher incomes, from any source, may feel better placed to deal with whatever comes their way. The coefficient on non-financial job satisfaction, on the other hand, is unchanged by addition of the mastery scale. Combined with the lower estimated effects of income, this means that making explicit allowance for individual personality differences raises rather than lowers the size of the resulting estimates of compensating differentials. We wish our estimates of these differentials to err if anything on the conservative side (because they are likely to be thought surprisingly large), so we base our results on a measure of mastery scale adjusted to remove its correlation with income. This restores the income coefficient to what it was without the inclusion of the mastery scale. This also makes it easier to compare the GSS results with those from the other surveys, since the ESC and the EDS do not have personality variables. 9. This is so despite the fact that the Benchmark question refers to involvement with a labour organization, while the ESC data are derived from a more specific question about union membership. 10. The US Benchmark results differ slightly in that the non-union workers show an (insignificantly) greater average income effect than do the union workers.

134 11.

12.

Social capital: reaching out But they have lower general social trust. Soroka et al. (2007) show that this difference in social trust can be explained mainly by differences in the climate of social trust in the countries from which the immigrants came, confirming the earlier US results of Rice and Feldman (1997) and Putnam (2000). In the Benchmark survey, however, trust of co-workers is more important for nonHispanic whites than for other employees. This is likely to reflect that trust between management and workers is different in important ways from trust among co-workers.

REFERENCES Dunn, Elizabeth W., Timothy D. Wilson and Daniel T. Gilbert (2003), ‘Location, location, location: the misprediction of satisfaction in housing lotteries’, Personality and Social Psychology Bulletin, 29 (11), 1421–32. Frey, Bruno S. and Alois Stutzer (2004), ‘Economic consequences of mispredicting utility’, Working Paper 218, University of Zurich: Institute for Empirical Research in Economics. Halpern, David (2005), Social Capital, Cambridge: Polity Press. Helliwell, John F. (1996), ‘Economic growth and social capital in Asia’, NBER Working Paper No. 11759, Cambridge: National Bureau of Economic Research. Helliwell, John F. (2003), ‘How’s life? Combining individual and national variables to explain subjective well-being’, Economic Modelling, 20, 331–60. Helliwell, John F. and Haifang Huang (2005), ‘How’s the job? Well-being and social capital in the workplace’, NBER Working Paper No. 11759, Cambridge: National Bureau of Economic Research. Helliwell, John F. and Haifang Huang (2006), ‘Well-being and the workplace: gender and labour relations’, paper prepared for the Mackenzie King Festschrift for Paul Weiler, Cambridge, MA, 3–4 November. Helliwell, John F. and Robert D. Putnam (2005), ‘The social context of well-being’, in Felicia A. Huppert, Nick Baylis and Barry Keverne (eds), The Science of WellBeing, Oxford: Oxford University Press, pp. 435–59. Hwang, Hae-shin, W. Robert Reed and Carleton Hubbard (1992), ‘Compensating wage differentials and unobserved productivity’, Journal of Political Economy 100 (4), 835–58. Knack, S. and P. Keefer (1997), ‘Does social capital have an economic payoff ? A cross-county investigation’, Quarterly Journal of Economics, 112 (4), 1251–88. O’Donoghue, Ted and Matthew Rabin (1999), ‘Doing it now or later’, American Economic Review, 89 (1), 103–24. OECD (2001), The Well-Being of Nations: The Role of Human and Social Capital, Paris: Organisation for Economic Co-operation and Development, Centre for Educational Research and Innovation. Van Praag, B.M.S. and B.E. Baarsma (2005), ‘Using happiness surveys to value intangibles: the case of aircraft noise’, Economic Journal, 115 (January), 224–46. Putnam, Robert D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Rice, Tom W. and Jan L. Feldman (1997), ‘Civic culture and democracy from Europe to America’, Journal of Politics, 59 (4), 1143–72.

How’s the job?

135

Smith, A. (1850), An Inquiry into the Nature and Causes of the Wealth of Nations, 4th edn, edited by J.R. McCulloch, Edinburgh: Adam and Charles Black; 1st edition 1786. Soroka, Stuart, John F. Helliwell and Richard Johnston (2007), ‘Measuring and modelling interpersonal trust’, in Fiona Kay and Richard Johnston (eds), Social Capital, Diversity and the Welfare State, Vancouver: UBC Press, pp. 95–132.

136

7.95 7.26 10.45 40.49 0.48 0.52 0.34 6.73 0.65 0.65

2523

2523

2523

2484 2523 2523 2523

2523

2523

2523

Life satisfaction, 1–10 point scale Job satisfaction, 1–10 point scale Log of personal income age male married Education: With university degree Trust in management, 1–10 point scale Trust in general, 0–1 scale Confidence that neighbour will return the wallet

Mean

Obs

0.34

0.47

2.28

10.57 0.50 0.50 0.47

0.70

1.88

1.63

Std dev

Working population excluding self-employed

Variable

Sample: Canadian ESC

Table 4A.1 Descriptive statistics: ESC

APPENDIX

858

858

858

847 858 858 858

858

858

858

Obs

0.66

0.67

5.95

42.70 0.44 0.55 0.37

10.59

6.95

8.05

Mean

0.33

0.46

2.30

9.90 0.50 0.50 0.48

0.55

1.94

1.58

Std. Dev.

Working population, union members

1665

1665

1665

1637 1665 1665 1665

1665

1665

1665

Obs

0.65

0.64

7.14

39.35 0.51 0.51 0.32

10.38

7.42

7.90

Mean

0.35

0.48

2.16

10.73 0.50 0.50 0.47

0.76

1.83

1.65

Std dev

Working population, non-union members

137

(cont.)

0.83 0.78 0.86 0.65 0.51 0.77 3.25 0.34 0.21 0.13

2523

2522

2523

2523

2522

2523

2487

2523

2523

2523

Confidence that police will return the wallet Job: Requires skill, 0–1 scale Job: Has variety of tasks, 0–1 scale Job: Have enough time, 0–1 scale Job: Free of conflicting demands, 0–1 scale Job: Makes own decision, 0–1 scale Sense of job, security 1–4 point scale Dummy: Union member Dummy: Immigrant Dummy: Visible minority

Mean

Obs

0.33

0.41

0.47

0.86

0.29

0.33

0.32

0.23

0.27

0.28

Std. Dev.

Working population excluding self-employed

Variable

Sample: Canadian ESC

Table 4A.1

858

858

858

849

858

858

858

858

857

858

Obs

0.09

0.18

3.30

0.74

0.47

0.59

0.87

0.81

0.83

Mean

0.29

0.38

0.86

0.29

0.33

0.34

0.21

0.25

0.28

Std dev

Working population, union members

1665

1665

1638

1665

1664

1665

1665

1665

1665

Obs

0.15

0.23

3.22

0.78

0.53

0.69

0.85

0.76

0.83

Mean

0.35

0.42

0.87

0.28

0.33

0.31

0.24

0.27

0.28

Std dev

Working population, non-union members

138

(cont.)

1.81 1.87 0.80

10.87 0.50 0.49 0.50 2.16

0.49 0.35

0.31

7.74

7.24

10.40

41.60 0.55 0.59 0.44

6.85

0.58

0.60

0.79

Life satisfaction, 540 1–10 point scale Job satisfaction, 540 1–10 point scale Log of personal 540 income/household income age 521 male 540 married 540 Education: With 540 university degree Trust in 540 management, 1–10 point scale Trust in general, 540 0–1 scale Confidence that 540 neighbour will return the wallet, 0–1 scale Confidence that 540 police will return the wallet, 0–1 scale

Std dev

Mean

Working population, immigrant

Obs

Variable

Sample: Canadian ESC

Table 4A.1

1983

1983

1983

1983

1963 1983 1983 1983

1983

1983

1983

Obs

0.84

0.67

0.67

6.70

40.20 0.46 0.50 0.31

10.46

7.27

8.01

Mean

0.27

0.34

0.47

2.31

10.48 0.50 0.50 0.46

0.68

1.88

1.57

Std dev

Working population, Canadian born

324

324

324

324

312 324 324 324

324

324

324

Obs

0.79

0.57

0.54

6.86

38.92 0.53 0.60 0.46

10.31

7.30

7.56

Mean

0.31

0.36

0.49

2.09

10.82 0.50 0.49 0.50

0.83

1.77

1.87

Std dev

Working population, visible minority

2199

2199

2199

2199

2172 2199 2199 2199

2199

2199

2199

Obs

0.84

0.66

0.67

6.72

40.72 0.48 0.51 0.32

10.47

7.26

8.01

Mean

0.27

0.34

0.47

2.31

10.52 0.50 0.50 0.47

0.68

1.90

1.58

Std dev

Working population, the rest

139

(cont.)

2.99

0.28

0.42

540

540

0.53

540

529

0.66

540

0.73

0.81

540

540

0.76

540

Job: Requires skill, 0–1 scale Job: Has variety of tasks, 0–1 scale Job: Have enough time, 0–1 scale Job: Free of conflicting demands, 0–1 scale Job: Makes own decision 0–1 scale Sense of job security, 1–4 scale Dummy: Union member Dummy: Immigrant Dummy: Visible minority

Mean

Obs

0.49

0.45

1.00

0.31

0.33

0.33

0.27

0.29

Std dev

Working population, immigrant

Variable

Sample: Canadian ESC

Table 4A.1

1983

1983

1958

1983

1982

1983

1983

1982

Obs

0.05

0.36

3.31

0.78

0.51

0.65

0.87

0.78

Mean

0.22

0.48

0.81

0.28

0.33

0.32

0.22

0.26

Std dev

Working population, Canadian born

0.24 0.70

324

2.99

0.71

0.53

0.67

0.79

0.73

Mean

324

319

324

324

324

324

324

Obs

0.46

0.43

0.99

0.31

0.31

0.32

0.28

0.29

Std dev

Working population, visible minority

2199

2199

2168

2199

2198

2199

2199

2198

Obs

0.14

0.35

3.28

0.77

0.51

0.65

0.87

0.79

Mean

0.35

0.48

0.84

0.28

0.33

0.32

0.22

0.26

Std dev

Working population, the rest

140

(cont.)

Happiness: 1–4 point scale Log of household income age male married Education: With university degree Trust in co-workers, 1–4 scale general trust, 0–1 scale Trust in neighbours, 0–1 scale Confidence in police, 0–1 scale Dummy: Union member non-Hispanic white

Variable

Sample: US Benchmark

Table 4A.1

3.33 10.79 39.75 0.46 0.52 0.40 3.40 0.55 0.75 0.76 0.13 0.75

16519

16519 16519 16519 16519

16519

16519

16220

16305

16519

16519

Mean

16519

Obs

0.43

0.34

0.28

0.28

0.48

0.79

12.08 0.50 0.50 0.49

0.68

0.58

Std dev

Working population NOT excluding self-employed

2136

2101

2111

2136

2136

2136 2136 2136 2136

2136

2136

Obs

0.73

0.74

0.76

0.56

3.34

41.58 0.54 0.55 0.41

10.89

3.34

Mean

0.44

0.30

0.27

0.48

0.83

11.34 0.50 0.50 0.49

0.58

0.58

Std dev

Working population, union members

14383

14204

14109

14383

14383

14383 14383 14383 14383

14383

14383

Obs

0.75

0.77

0.75

0.55

3.41

39.47 0.45 0.51 0.39

10.78

3.33

Mean

0.43

0.28

0.28

0.48

0.78

12.16 0.50 0.50 0.49

0.69

0.58

Std dev

Working population, non-union members

141

(cont.)

3.37 10.87 40.89 0.46 0.55 0.43 3.54 0.62 0.81 0.81 0.13

12399

12399

12399 12399 12399 12399

12399

12399

12206

12250

12399

Happiness: 1–4 point scale Log of household income age male married Education: With university degree Trust in coworkers, 1–4 scale general trust, 0–1 scale Trust in neighbours, 0–1 scale Confidence in police, 0–1 scale Dummy: Union member

Mean

Obs

Working population, non-Hispanic white

Variable

Sample: US Benchmark

Table 4A.1

0.33

0.26

0.25

0.47

0.70

12.10 0.50 0.50 0.50

0.64

0.56

Std dev

4120

4055

4014

4120

4120

4120 4120 4120 4120

4120

4120

Obs

0.14

0.63

0.59

0.35

2.99

36.32 0.45 0.42 0.29

10.58

3.21

Mean

Working population, all else

0.34

0.32

0.31

0.46

0.91

11.34 0.50 0.49 0.45

0.75

0.61

Std dev

142

(cont.)

1.42 0.97 0.38 11.43 0.50 0.50 0.55 0.66 0.14 0.91 0.44 0.30 0.26 0.32 0.41 0.49 0.26 0.24 0.39 0.36 0.47

7.95 3.82 0.18 39.51 0.52 0.55 11.05 10.50 0.25 3.91 0.26 0.50 0.67 0.52 0.42 0.57 0.67 0.73 0.45 0.31 0.32

Life satisfaction Trust in co-workers Immigrant status Age % married % male Log of household income Log of personal income Mastery Scale Health Status University degrees Frequency of visiting relatives Friendship Contact with neighbours Membership General trust Trust in neighbours Confidence in police Importance of religion Frequency of attending religious services % union

Std dev

Mean

Whole sample

Variable

Sample: Canadian GSS

Table 4A.1

1.00

7.96 3.89 0.15 40.80 0.54 0.54 11.15 10.64 0.26 3.96 0.36 0.51 0.67 0.53 0.58 0.63 0.69 0.74 0.46 0.32

Mean

0.00

1.37 0.91 0.36 10.73 0.50 0.50 0.47 0.49 0.14 0.89 0.48 0.29 0.25 0.31 0.34 0.48 0.24 0.22 0.39 0.37

Std dev

Union members*

Working population

0.00

7.95 3.79 0.20 38.92 0.52 0.55 10.99 10.41 0.25 3.89 0.21 0.49 0.67 0.52 0.35 0.54 0.66 0.73 0.45 0.30

Mean

0.00

1.45 0.99 0.40 11.70 0.50 0.50 0.58 0.73 0.14 0.92 0.41 0.31 0.26 0.32 0.42 0.50 0.26 0.24 0.40 0.36

Std dev

The rest

143

(cont.)

1.53 1.02 0.00 11.04 0.47 0.50 0.59 0.70 0.15 0.94 0.49 0.33 0.27 0.29 0.41 0.49 0.26 0.24 0.43 0.41 0.44

7.77 3.67 1.00 41.98 0.67 0.56 11.02 10.51 0.22 3.83 0.40 0.41 0.63 0.49 0.39 0.57 0.64 0.73 0.51 0.39 0.26

Life satisfaction Trust in co-workers Immigrant status Age % married % male Log of household income Log of personal income Mastery Scale Health Status University degrees Frequency of visiting relatives Friendship Contact with neighbours Membership General trust Trust in neighbours Confidence in police Importance of religion Frequency of attending religious services % union

Std dev

Mean

Immigrants

0.33

7.99 3.85 0.00 38.97 0.49 0.54 11.05 10.49 0.26 3.93 0.23 0.52 0.68 0.53 0.43 0.56 0.67 0.73 0.44 0.29

Mean

Working population

Variable Label

Sample: Canadian GSS

Table 4A.1

0.47

1.40 0.96 0.00 11.45 0.50 0.50 0.54 0.66 0.14 0.90 0.42 0.29 0.25 0.32 0.41 0.49 0.25 0.23 0.38 0.35

Std dev

Canadian born

144

(cont.)

1.52 1.02 0.50 11.15 0.50 0.50 0.60 0.68 0.15 0.94 0.47 0.33 0.27 0.31 0.41 0.50 0.27 0.25 0.42 0.38 0.44

7.85 3.68 0.50 38.46 0.55 0.56 10.96 10.41 0.23 3.86 0.32 0.46 0.65 0.49 0.40 0.52 0.62 0.71 0.46 0.33 0.26

Identifiable visible minority Mean Std dev

0.33

7.98 3.86 0.09 39.82 0.52 0.54 11.07 10.52 0.26 3.93 0.24 0.51 0.67 0.53 0.43 0.58 0.68 0.74 0.45 0.30

Mean

Working population The rest

0.47

1.39 0.95 0.28 11.50 0.50 0.50 0.53 0.66 0.14 0.90 0.43 0.29 0.25 0.32 0.41 0.49 0.25 0.23 0.39 0.36

Std dev

Note: Union member is defined as those who answered positively to the question ‘In the past 12 months, were you a member or participant in a union or professional association?’, excluding self-employed and those with personal income above ca $90 000 (top 5% of the paid workers who are union members or professional organization).

Life satisfaction Trust in co-workers Immigrant status Age % married % male Log of household income Log of personal income Mastery scale Health status University degrees Frequency of visiting relatives Friendship Contact with neighbours Membership General trust Trust in neighbours Confidence in police Importance of religion Frequency of attending religious services % union

Variable

Sample: Canadian GSS

Table 4A.1

5.

Social capital eﬀects on student outcomes Maureen T. Hallinan1

The primary goal of this chapter is to test James Coleman’s hypothesis that social capital, in the form of intergenerational social closure, increases student achievement. To test this hypothesis, I conducted both descriptive and inferential analyses on a longitudinal set of data in Catholic schools in Chicago. Surprisingly, the analyses did not show that networks of intergenerational social closure increase student academic achievement, either overall or indirectly. Instead, they indicated that networks of intergenerational closure increased liking for school and other non-academic variables. Networks with social closure have a statistically signiﬁcant, positive eﬀect on feeling able to count on neighborhood adults for safety, trusting people in the neighborhood, participating in school clubs or organizations, believing it important to help others in the community, attending religious services, and feeling safe in school. They have a statistically signiﬁcant, negative eﬀect on the number of times a student is sent to the principal’s oﬃce. A multitude of deﬁnitions exist for social capital, ranging from tight deﬁnitions used within very speciﬁc situations to broad-based concepts that can apply to a wide variety of social contexts. Given the focus of this chapter, I will be using Coleman’s deﬁnition (1990): social capital is a product of social structure that produces beneﬁts. Coleman’s initial fascination with the dynamics of structure and change in social systems led him to social capital and empirical analyses of schools as social systems. In order to gain theoretical speciﬁcity, he focused on social networks within schools. His work reinforced that of Lin (2001), Burt (2001) and others who suggest that the application of social capital theory should be restricted to social networks. This approach enables researchers to predict the eﬀects of social capital on network members more precisely and to identify the mechanisms that transmit these eﬀects more conﬁdently. Burt (2001) deﬁned two important characteristics of social networks: social closure and structural holes. A network with social closure is a discrete, densely connected group without links to other networks. Structural 145

146

Social capital: reaching out

holes exist between two networks when a small number of ties connect them. Networks with social closure are sources of social capital because dense ties enable members to beneﬁt from easy access to information held by other members. Members of such networks also tend to develop norms that govern the behavior of group members and establish sanctions for violations of those norms. Members usually comply with these norms either because they agree with them or because they are unwilling to risk sanctions. This conformity makes behavior more predictable and provides a basis for trust. In this case, people trust each other to act according to the prescribed norms and to help enforce them. Networks with structural holes provide a diﬀerent kind of social capital – one based on exchanges of information with others outside the network. Members of densely connected groups tend to be self-involved and share information and resources only among themselves. They are unlikely to initiate communication with members of other groups. However, if a few connections exist between a densely connected group and another group, the links may provide an opportunity to pass information between groups. Such linkages can provide new information which may increase innovation, creativity and opportunities within the original network. In Coleman’s research (1988, 1987), he related social capital – and speciﬁcally networks with intergenerational social closure – to academic achievement. In such networks, a student’s parents are friends with the parents of the student’s friends. Coleman claimed that when parents belong to a network with intergenerational social closure, the network becomes a rich source of social capital. Students beneﬁt when parents share information about educational matters. If one parent in the group can identify the outstanding teachers in a child’s school, another understands how class assignments are made and a third knows excellent tutors – and they share this knowledge – then all students in the group may beneﬁt from increased educational opportunities. Coleman (1987) also hypothesized that networks with intergenerational social closure create social capital through promoting norms of acceptable behavior, in this case striving toward academic excellence. He claimed that norms evolve in networks where parents have high academic standards and goals for their children. Through social interactions, these parents reinforce each other’s commitment to their children’s education. They also help enforce behavioral norms that promote their children’s learning. A network of parents who are committed to their children’s education can make it easier to encourage children to study because their friends are likely to be studying as well. If parents agree on the importance of completing homework, studying, taking challenging courses and avoiding behaviors that would negatively aﬀect class work, they are apt to require these behaviors

How’s the job?

147

of their children, and encourage their children’s friends to engage in similar ones. Finally, Coleman (1987) claimed that members of networks with intergenerational social closure create social capital by developing trusting relationships with each other. Trust is the basis for expectations and obligations among members. A person who trusts other members is willing to act in the best interests of those members, convinced that they will reciprocate. Networks with intergenerational social closure are particularly suited to developing such trusting relationships because parents and students are connected by close friendship ties. Coleman relied on these ideas to explain diﬀerences across school sectors in student achievement. In analyses of the High School and Beyond survey, Coleman (Coleman et al., 1981) found that students in Catholic schools achieve higher standardized test scores and higher high school graduation rates than those in public (that is, state-funded) schools, controlling for background factors. Coleman’s ﬁnding became known as the ‘Catholic school advantage’. Coleman argued that three particular characteristics of Catholic schools create more social capital to support student academic achievement than public schools do. These characteristics – size, religious commitment and organizational structure – foster parent–student networks with intergenerational social closure with close ties among parents, teachers and administrators. The smaller size of Catholic schools makes it easier for Catholic school parents to get to know each other and their children’s friends. The fact that Catholic schools are typically associated with a parish or church also strengthens social ties. Finally, Catholic schools typically have a rigorous academic curriculum. Coleman’s social capital explanation for the Catholic school advantage has not been tested rigorously because the network data needed to do so are not yet available in large school surveys. However, Morgan and Sorensen (1999) attempted to use the National Education Longitudinal Study (NELS) data to conduct an initial test of Coleman’s argument that intergenerational social closure contributes to the Catholic school advantage. Their ﬁndings only partially supported Coleman’s; unfortunately, because one of their assumptions is dubious, their conclusions must be questioned. Morgan and Sorensen (1999) distinguished between two kinds of schools: norm-enforcing schools and horizon-expanding schools. Normenforcing schools have networks with intergenerational social closure and a school community with close ties among students, parents, teachers and administrators. Horizon-expanding schools have social networks with bridges to persons outside the school. However, I should note that because

148

Social capital: reaching out

the NELS survey does not include a measure of horizon-expanding schools, Morgan and Sorensen made the assumption that schools with little social closure are thus, by their deﬁnition, horizon-expanding. Because this assumption is highly questionable, their results must be viewed with caution. Morgan and Sorensen (1999) agreed with Coleman (1987), Bryk et al. (1993) and others that schools with intergenerational social closure and communal organization create social capital that can positively influence the academic behavior of students. But, they also claimed that horizonexpanding schools create a diﬀerent type of social capital that could also increase student achievement. They argued that parents in horizonexpanding schools conduct their social interactions with persons outside the school community. These other social links can also provide incremental information about educational opportunities which parents might not otherwise have access to for their children. Morgan and Sorensen (1999) then examined whether norm-enforcing or horizon-expanding schools more eﬀectively promote student achievement. Relying on the 1988, 1990 and 1992 waves of the NELS survey, they estimated social capital eﬀects on gains in Mathematics achievement in public and Catholic schools. Their ﬁndings show that networks with intergenerational social closure have a positive eﬀect on student achievement in Catholic schools, but a negative eﬀect in public schools. They interpret this result as showing that horizon-expanding schools are more eﬀective in promoting student learning than norm-enforcing schools. However, there are other possible explanations for these ﬁndings than Morgan and Sorensen’s interpretations. In this case, for example, some public schools may also have intergenerational social closure that enforces norms other than those of striving for academic excellence. Thus, we must be cautious in interpreting Morgan and Sorensen’s results. My approach to testing Coleman’s hypothesis is diﬀerent and as analytical as possible. Because of real-world data constraints, I had to limit myself to descriptive and inferential analyses. To make the sample as representative as possible, I used data from a large, longitudinal study of sixth through twelfth grade students in Chicago’s public and Catholic Archdiocesan schools. Because only the Catholic data contained questions about networks, I did not analyze the public (state-funded) school data. Before starting, I determined what analyses would be possible given available data. A comprehensive, analytically rigorous approach would be almost impossible as it would require speciﬁcity in the data that was not available. I would have to be able to map the social ties of nearly all the parents and students in each school. Because various norms may exist in diﬀerent networks, I would have to know all the norms in all the networks

Social capital effects on student outcomes

149

as well as what student behaviors they influence. I would have to determine whether schools and intergenerational social networks vary in the strength of the norms governing academic achievement. As if all this were not enough of a challenge, it would be essential to know whether the information parents receive through their links with other networks is relevant to their children’s schooling. If so, one might be seeing the eﬀect of networks with structural holes rather than intergenerational social closure in student academic achievement outcomes. I also had to determine whether the school is the correct unit of analysis. Parents and students in a school community rarely belong to a single, large social network. Rather, several parent and student networks usually exist within the context of a particular school. Some of these various networks may be densely connected, others loosely. Some may exhibit intergenerational social closure, others consist solely of parents or of students. Some may have links to individuals or groups outside the school while others may not. Based on the diﬀerences between networks, the correct unit of analysis may be smaller than the entire school. Given these realities, I kept the analyses descriptive, inferential and extremely focused. The primary independent variable is a social capital factor measuring network density among students, their friends and their friends’ parents. The dependent variable, student Reading or Mathematics achievement, is measured as change in scale score from seventh to eighth grade on the Terra Nova test in English and Mathematics. In the inferential analysis, I used student responses to survey questions about their attitudes and behaviors to examine network eﬀects on non-academic outcomes. The rest of this chapter outlines my approach, identiﬁes the data sample, describes the analyses and discusses the conclusions. Where appropriate, tables have been included; more detailed analytical information is at the end of the chapter.

APPROACH: SAMPLE The data came from the Chicago School Study, a large, longitudinal examination of best practices in public and Catholic elementary and secondary schools. Students in the sixth through eleventh grades in Chicago public schools and in the sixth through twelfth grades in Catholic schools in the Chicago Archdiocese completed surveys. The Chicago Archdiocese includes the city of Chicago and surrounding suburbs. The analyses focus only on the Catholic schools because the public school survey did not include network questions.

150

Social capital: reaching out

The Archdiocese of Chicago administers about 260 elementary and 43 high schools; speciﬁc school years vary slightly due to school openings, closings and consolidations. Slightly more than half of these schools are in the city while the remainder are in surrounding suburbs. Principals, teachers and students in 75 percent of the Catholic schools who agreed to participate in the study received the survey questionnaire. Approximately 80 percent of the elementary school students and 60 percent of the high school students in participating schools chose to respond. The descriptive and inferential analyses are based on responses from 4366 students in the Catholic elementary schools. This number represents approximately half of the Chicago Catholic schools’ eighth-grade population. The analyses only dealt with students who completed the Chicago School Survey and who had Terra Nova test scores in both the seventh and eighth grades. This standardized test is mandatory for all seventh-graders in the Catholic schools, and over 99 percent of the seventh-grade students take the test every year. A special administration of the test was given to eighth-graders in 2004, and approximately 85 percent of all eighth-grade students participated. Of the eighth-grade students who returned a survey in 2004, 4859 students had valid test scores in both Reading and Mathematics as seventh-graders in 2003 and as eighth-graders in 2004. After dropping cases due to missing data on other variables, the sample for these analyses is 4366 students.

DESCRIPTIVE ANALYSES Variables The descriptive analysis relates student responses on network questions to school characteristics and student background. School variables include locale, measured as urban or suburban, and grade level, measured as sixth through eighth grade or ninth through twelfth grade. Student background variables are gender and race or ethnicity, measured as White, African American, Hispanic and Other (Asian, Native American, biracial, other). These variables are also used in the inferential analysis. The independent variable is a social capital factor measuring network density among students, their friends and their friends’ parents. The factor is based on student responses to three items: ● ● ●

‘How many of your best friends do your parents know?’ ‘How many of your best friends’ parents do you know?’ ‘How many of your best friends’ parents do your parents know?’

Social capital effects on student outcomes

151

Student responses were none, some, most and all, coded 1–4. A single factor solution explains 73 percent of the variance (eigenvalue = 2.2) with all three items loading at 0.82 to 0.88. Findings The data show that elementary school students and females have denser networks of intergenerational social closure than high school students and males. In terms of race or ethnicity, white students in elementary and high school have the densest social networks, while only slight diﬀerences are found across minority groups. For locale, suburban schools have denser networks than urban schools at both elementary and high school levels. Table 5.1 presents the means and standard deviations of the network variables by elementary and high school levels and locale, as well as by gender, race and ethnicity. These results hold across each of the network questions. While the statistics are not surprising, they do raise questions about students’ social networks. We need to conduct further research to determine why and how student and school characteristics influence the formation of networks of intergenerational social closure.

INFERENTIAL ANALYSES Variables The inferential analyses control for several additional student background variables including gender, race or ethnicity, socio-economic status, family structure and age. Family structure is a categorical variable measured by only one adult in the household. Mother’s education is a categorical variable, measured as: some high school, high school graduate, some college, college graduate and post-college education, coded 1–5. Student age is measured in months, centered at zero for students who were the appropriate age in January 2004. School variables include locale, coded as urban or suburban, and proportion of students on free or reduced-cost lunch. Reading and Mathematics achievements in seventh grade are also included as controls. The dependent variable, student Reading or Mathematics achievement, is measured as change in scale score from seventh to eighth grade on the Terra Nova test in English and Mathematics. Scale scores are more readily interpretable than grade equivalent, national percentile or other raw score transformations when measuring growth in achievement (Kolen and

152

Social capital: reaching out

Table 5.1 Descriptive statistics for social network variables by grade, locale, gender and race/ethnicity

6th–8th grade Urban Suburban 9th–12th grade Urban Suburban 6th–8th grade Male Female 9th–12th grade Male Female 6th–8th grade White Black Hispanic Other (Asian, Native American, biracial, other) 9th–12th grade White Black Hispanic Other (Asian, Native American, biracial, other)

How many of your best friends do your parents know?

How many of your best friends’ parents do you know?

mean

s.d.

mean

s.d.

mean

s.d.

3.3 3.5

0.9 0.8

3.2 3.4

0.8 0.7

2.9 3.2

0.9 0.8

2.9 3.1

0.9 0.9

2.9 3.1

0.9 0.8

2.5 2.6

0.9 0.9

3.3 3.5

0.9 0.8

3.2 3.4

0.8 0.7

3.0 3.2

0.9 0.8

2.9 3.1

0.9 0.9

2.9 3.1

0.9 0.8

2.5 2.6

0.9 0.9

3.5 3.1 3.2 3.3

0.7 1.0 0.9 0.9

3.4 3.1 3.0 3.2

0.7 0.9 0.9 0.8

3.2 2.8 2.7 3.0

0.8 1 0.9 0.9

3.1 2.9 2.8 2.8

0.8 1.0 1.0 1.0

3.1 3.0 2.8 2.8

0.8 1.0 0.9 0.9

2.6 2.6 2.3 2.4

0.8 1 0.9 0.9

6th–8th, N = 19,346 9th–12th, N = 15,056 Note:

6th–8th, N = 19,550 9th–12th, N = 15,147

How many of your best friends’ parents do your parents know?

6th–8th, N = 19,470 9th–12th, N = 15,112

Each social network variable is coded (1) None, (2) Some, (3) Most and (4) All.

Social capital effects on student outcomes

153

Brennan, 2004). Seventh-grade achievement tests were administered in March 2003, and eighth-grade tests were given a year later in March 2004. The most important independent variable is one used in the descriptive analysis: a social capital factor measuring network density among students, their friends and their friends’ parents. The factor is based on student responses to three items outlined earlier, including: ‘How many of your best friends do your parents know?’, ‘How many of your best friends’ parents do you know?’, ‘How many of your best friends’ parents do your parents know?’ Student responses were once again none, some, most and all, coded 1–4. A single factor solution explains 73 percent of the variance (eigenvalue 2.2) with all three items loading at 0.82 to 0.88. Analyses of network eﬀects on non-academic outcomes are based on student responses to survey questions about their attitudes and behaviors. For one of these dependent variables, designated ‘like school’, I combined student responses to four items on the survey: ● ● ● ●

‘I usually look forward to school.’ ‘I wish I didn’t have to go to school.’ ‘I’m bored in school.’ ‘I’m glad to get back to school after summer vacation.’

Answers to these questions were coded 1–4 for strongly agree to strongly disagree. They yielded one factor with an eigenvalue greater than 1 (eigenvalue = 2.3). The factor explains 59 percent of the variance with four items loading in the appropriate direction with absolute values of 0.70 to 0.81. Another seven dependent variables were based on student responses to the following seven questions: ●

● ● ●

● ●

‘I can count on adults in my neighborhood to see that children are safe and do not get into trouble’ (strongly disagree to strongly agree, coded 1–4). ‘People in this neighborhood can be trusted’ (strongly disagree to strongly agree, coded 1–4). ‘How many times were you sent to the oﬃce this year?’ (student selfreport). ‘How often do you participate in school clubs or organizations?’ (never, once in a while, once a week, almost every day, every day, coded 1–5). ‘It is important to help others in my community’ (strongly disagree to strongly agree, coded 1–4). ‘How safe do you feel in school?’ (not safe, somewhat safe, mostly safe, very safe, coded 1–4).

154 ●

Social capital: reaching out

‘Outside of school time, how often do you attend religious services at your church?’ (never, a few times a year, once a month, 2–3 times a month, once a week, more than once a week, daily, several times a day, coded 1–8).

Model Ordinary least squares (OLS) regression models were used to analyze the eﬀects of social networks on student achievement in Reading and Mathematics. Two of the control variables in the model, mother’s education and proportion of student body on free or reduced-cost lunch, appear to be non-linear. To take this into account, a squared term was added for each of these variables and orthogonalized to the linear term. OLS regression models were also used to determine the eﬀects of social networks on various non-academic outcomes. Findings The means show growth in Reading and Mathematics achievement from seventh to eighth grade. Mother’s education has a mean of 3.53, indicating that the mothers of students in the study averaged some college education. The social network and like school variables are factors with a mean of zero and range of 1 to 1. The means and standard deviations of the variables in the multivariate inferential analyses are found in Table 5.2. The means of the non-academic variables provide information about students’ attitudes and behavior in and outside of school. On average, the students in the sample felt that they could count on and trust neighborhood adults. They were sent to the oﬃce for disciplinary reasons two to three times during the school year. They participated in school clubs slightly more than once a week, agreed that it is important to help others in their community, and attended religious services three to four times a month. Students, on average, felt safe or very safe in school. The results of the OLS regression of test scores on the social network factor showed that the eﬀect of the variable of intergenerational social closure on both Reading and Mathematics achievement is close to zero and fails to attain statistical signiﬁcance. The regression controlled for seventh-grade Reading and Mathematics test scores, student background and school characteristics. Table 5.3 presents the total eﬀect of these networks on student achievement. Although the analyses show no total eﬀect, it is possible – though unlikely – that social networks influence achievement but the diﬀerent influences sum to zero total eﬀect. In other words, there may be a combination of positive (or negative) direct eﬀects

Social capital effects on student outcomes

155

Table 5.2 Descriptive statistics for variables in regression analyses, N = 4366

2004 Reading achievement 2004 Mathematics achievement 2003 Reading achievement 2003 Mathematics achievement Female Black Latin Other race Age in months One adult in household Mother’s education Mother’s education squared Missing mother’s education Internet access at home More than 50 books at home Prop. in school on free/reduced lunch Free lunch prop., squared Social network factor Like school factor Count on neighborhood adults Trust people in neighborhood Times sent to office Participate in school clubs/organizations Important to help others in community How often attend religious services Feel safe in school

mean

s.d.

701.25 709.70 691.81 695.15 0.53 0.07 0.15 0.12 0.43 0.09 3.53 13.66 0.11 0.92 0.81 0.18 0.02 0.01 0.22 2.70 2.74 2.63 3.22 3.18 4.22 3.57

32.13 37.32 30.21 35.45 0.50 0.26 0.36 0.32 3.74 0.28 1.10 7.53 0.32 0.27 0.39 0.27 0.09 0.99 0.96 0.79 0.75 2.16 1.16 0.62 1.44 0.64

countered by negative (or positive) indirect eﬀects. To examine this possibility, I ran models that assessed the indirect eﬀects of social networks on achievement, as mediated by each of the eight non-academic student outcomes. The estimates of the indirect eﬀects showed that the non-academic variables do not transmit the eﬀects of intergenerational social networks on academic achievement. The eﬀects themselves were trivial with a magnitude less than 0.01 (results not shown). That is, a change of one standard deviation on the social network factor, working through any of the nonacademic variables, leads to a change of less than one-hundredth of a standard deviation in academic achievement.

156

Social capital: reaching out

Table 5.3 Regression of 2004 reading and mathematics achievement on social networks and background factors for Catholic schools in the Chicago School Study, N = 4366 Reading achievement

2003 Reading achievement 2003 Mathematics achievement Female Black Latin Other race Age in months One adult in household Mother’s education Mother’s education squared Missing mother’s education Internet access at home More than 50 books at home Prop. in school on free/reduced lunch Free lunch prop., squared Social network factor Constant

b

sig

0.53 0.23 9.87 0.93 0.85 1.03 0.22 0.18 1.24 0.15 1.83 1.86 3.50 3.04 13.92 0.16 163.52

*** *** ***

*

*** ** ***

Mathematics achievement b 0.18 0.72 4.40 8.35 1.68 1.65 0.21 2.86 0.53 0.11 0.72 2.05 2.60 0.02 21.67 0.27 87.10

sig *** *** *** ***

* *

** *** ***

Note: * p 0 .05, ** p 0 .01, *** p 0 .001.

This analysis does not support Coleman’s proposition that social capital as exempliﬁed through intergenerational social closure promotes student academic achievement; however, a question remains as to whether networks with social closure increase student achievement under other conditions. In this analysis, little is known about the content of the interactions among network members, particularly students’ parents. If the appropriate conditions are in place, networks with social closure might aﬀect achievement; in other words, if members discuss their children’s education, endorse strong norms regarding student academic performance and trust other network members. These conditions need to be taken into account in future tests of Coleman’s hypothesis. While academic achievement was not aﬀected, another question centered on whether social capital in networks with social closure might aﬀect other aspects of a student’s development. Table 5.4 presents eight OLS regressions showing the eﬀects of networks with intergenerational social

157

Social capital effects on student outcomes

Table 5.4 Effect of social network factor in regressions of non-academic outcomes on social networks and background factors for Catholic schools in the Chicago School Study, N 4366 Dependent variables

Effect of social network factor b

Like school factor Count on neighborhood adults Trust people in neighborhood Times sent to office Participate in school clubs/organizations Important to help others in community How often attend religious services Feel safe in school

0.16 0.13 0.11 0.22 0.27 0.10 0.19 0.15

sig *** *** *** *** *** *** *** ***

Notes: * p 0.05, ** p 0.01, *** p 0.001. Effects of other variables in the model not shown.

closure on a broad array of student non-academic outcomes. The results are similar for all the student outcomes studied. Networks with social closure have a statistically signiﬁcant, positive eﬀect on liking school, feeling able to count on neighborhood adults for safety, trusting people in the neighborhood, participating in school clubs or organizations, believing it important to help others in the community, attending religious services and feeling safe in school. Social networks have a statistically signiﬁcant, negative eﬀect on the number of times a student is sent to the oﬃce. This dramatic evidence of influence on non-academic outcomes has important implications. The ﬁnding that networks with social closure increase religious and community participation and feelings of safety and responsibility addresses some of the issues raised in Putnam’s widely known research on social engagement. Putnam (2000) claims that contemporary Americans now belong to fewer groups, join fewer social organizations and participate less in political and social activities than in previous generations. The research reported here suggests that networks with social closure may lead to greater human engagement in the future. Indeed, a small number of close links to friends or neighbors through a school environment can teach students, and even may help catalyze their parents, to join larger organizations, assume greater responsibility for friends and neighbors, and commit themselves more deeply to political and civic activities.

158

Social capital: reaching out

CONCLUSION Like Coleman and others, this chapter argues that one source of social capital is networks with social closure. Such networks generate social capital by facilitating the transmission of information from one network member to another. Members typically share information on various topics, but such information serves as social capital with respect to student achievement only when the information is relevant to a student’s education. When members of social networks discuss their children’s schooling and hold and enforce strong academic norms, they may be able to increase academic achievement. If the parents’ primary goal is to improve their children’s academic outcomes, they would be well advised to join a social network where educational matters are discussed and norms governing pro-academic behavior are promoted. Trust also acts as social capital. Members of networks with social closure develop trusting relationships with each other through their interactions and shared interests. They share expectations and obligations. These norms may govern children’s academic, social, political, religious, moral or ethical behavior. The network members rely on other network members to support them in working toward the established norms. My research showed that network density was closely tied to a number of non-academic outcomes, which may suggest that many of the networks examined shared information and norms about these issues rather than academic achievement. Whatever the norms, social capital among network members with intergenerational social closure beneﬁts students through the encouragement and support of the parents within the network. Future research needs to determine what type of information is being shared among parents and what norms are being established, and then examine student outcomes in that light. This will require greater conceptual depth and analytical rigor. This type of rigor is necessary to develop a robust theory of social capital that will not fade away as a vague intuitive notion but rather will emerge as a solid, living concept that can withstand intellectual challenge and provide heuristic value.

NOTE 1. The author thanks the US Department of Education, grant number R215K010011, for supporting this project. She is also grateful to the administrators, teachers and students whose participation made this research possible. Without my colleagues, none of this would have happened: Warren Kubitschek supervised the data collection and contributed methodological and statistical expertise; Catherine Liu and Deb Kraklow assisted with the literature review; and Cheryl Pauley contributed editorial expertise. They all have my gratitude.

Social capital effects on student outcomes

159

REFERENCES Bryk, Anthony, Valerie E. Lee and Peter Holland (1993), Catholic Schools and the Common Good, Cambridge, MA: Harvard University Press. Burt, Ronald S. (2001), ‘Structural holes versus network closure as social capital’, in N. Lin, K. Cook and R.S. Burt (eds), Social Capital: Theory and Research, New York: Walter de Gruyter, pp. 31–56. Coleman, James S. (1987), ‘The relations between school and social structure’, in M.T. Hallinan (ed.), The Social Organization of Schools: New Conceptualizations of the Learning Process, New York: Plenum Press, pp. 177–204. Coleman, James S. (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94, 95–120. Coleman, James S. (1990), Foundations of Social Theory, Cambridge, MA: Harvard University Press. Coleman, James S., Thomas Hoﬀer and Sally Kilgore (1981), Public and Private Schools: An Analysis of High School and Beyond: A National Longitudinal Study for the 1980’s, Washington, DC: National Center for Education Statistics. Kolen, Michael J. and Robert L. Brennan (2004), Test Equating, Scaling, and Linking: Methods and Practices, 2nd edn, New York: Springer. Lin, Nan (2001), ‘Building a network theory of social capital’, in N. Lin, K. Cook and R.S. Burt (eds), Social Capital: Theory and Research, New York: Walter de Gruyter, pp. 3–29. Morgan, Stephen L. and Aage B. Sorensen (1999), ‘Parental networks, social closure, and mathematics learning: a test of Coleman’s social capital explanation of school eﬀects’, American Sociological Review, 64, 661–81. Putnam, Robert D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. US Department of Education (2002), CD-ROM: NELS: 88/2000 Public Use Data Files and Electronic Codebook – Base Year through Fourth Follow-up, Washington, DC: National Center for Education Statistics.

6.

Communities, schools and voter turnout: a case study in social norms David E. Campbell

This chapter advances the provocative claim that it is not necessary to have a close election in order to observe high voter turnout. Rather, in contrast to conventional wisdom among election observers, we have reason to believe that turnout can increase as electoral competition decreases. This claim is predicated upon the existence of two types of voters: homo politicus, who votes in highly competitive situations because she believes her vote can make a diﬀerence; and homo civicus, who votes because he believes that voting is a civic duty. Close examination of voter turnout in both highly and non-competitive elections (as measured by political heterogeneity and homogeneity) reveals that turnout can be high in both situations. In the latter case, civic norms that reinforce voter turnout are essential. The chapter then explores the question of where these norms are generated. While communities are responsible for many norms, some norms are formed earlier in life – thanks to families, friends and schools. A longitudinal study of adolescents reveals that the presence of a vibrant civic climate in high school has a strong correlation with voting as an adult. The logic – or rather, illogic – of collective action is an enduring puzzle in social science (Olson, 1965). Why do people engage in cooperative activity when they could simply free ride on the eﬀorts of others? Why contribute to charity? Why donate blood? Why vote? Indeed, why vote? The question is clearly signiﬁcant; whoever votes determines who governs. While an important subject in any era, America’s current level of low voter turnout heightens this issue’s urgency. Explaining why we vote also has implications beyond the political realm, as understanding voting can help us understand collective action more generally.1 Leaders interested in generating collective action and social capital within their organizations can potentially learn much from such a study.

160

Communities, schools and voter turnout

161

HOMO POLITICUS The question of why we vote has occupied a small army of political scientists and economists for decades. It has been particularly puzzling for those scholars who consider the voter to be a member of the species homo economicus – that is, those who subscribe to the rational choice school of human behavior. In a long-noted paradox, the inﬁnitesimal chance that any individual voter could cast a deciding vote means that a cost–beneﬁt analysis of voter turnout should always result in greater costs than beneﬁts (Downs, 1957; Tullock, 1967). Voting is thus perhaps the quintessential dilemma of collective action. The costs to voting must be paid by each voter: it takes time, energy and information. The beneﬁts, however, are diﬀuse and, in the language of economics, ‘non-excludable’. In other words, public policies enacted by those voted into oﬃce aﬀect everyone, whether they voted or not. By any rational consideration no one should ever vote. However, while in most US elections voter turnout is low, it is far from zero. As irrational as it might be, lots of people vote. It is clear that viewing people as homo economicus will not take us very far in understanding why we vote. Instead, we must view people as being Janus-faced. Depending on their context, people can be either homo politicus or homo civicus. ‘Politicus’ refers to the fact that voters turn out when political conflict is ignited – when they are motivated by the desire to defend or advance their interests in the political arena. ‘Civicus’ refers to a motivation for turning out that is based on a sense of civic obligation and therefore immune from the heat of political combat. Expressed in operational terms, homo politicus appears when contests are competitive and elections are close. An extensive literature largely, but by no means entirely, conﬁrms the hypothesis: that voter turnout rises with electoral competition. In summing up this literature, one scholar has written: ‘The close relationship of electoral competition and voter turnout at the individual level and at the system level is a frequently veriﬁed proposition, bordering upon the status of a law’ (Gray, 1976). Researchers have observed the competition–turnout connection in a variety of political contexts, including states in presidential elections,2 counties in presidential elections (Hoﬀstetter, 1973), congressional districts,3 states in gubernatorial elections,4 and state legislative districts (Caldeira and Patterson, 1982). In addition, a few studies have found that the closeness of the national election can spur political participation at other levels.5 However, most research indicates that the signiﬁcant relationship is between localized electoral competition and voter turnout, even in elections for national or statewide oﬃces.

162

Social capital: reaching out

The experience of the 2000 presidential contest, which was as close as they come, is instructive. To many observers it seemed that turnout must have gone up – a lot. Many newspapers ran stories, sometimes with distinctly purple prose, on how the tight margin led to record turnout. The day after the election the Atlanta Constitution wrote: ‘Georgians packed the polls in heavy numbers Tuesday – ﬁred out of their ballot box apathy of the past eight years by a tight presidential contest’ (Stepp, 2000). The Independent of London was equally colorful: ‘Energised by a nail-bitingly close race, voters across the United States flocked to the polls in strikingly large numbers yesterday’ (Gumbel, 2000). In less colorful language, the Los Angeles Times reported record voter participation in counties around LA, attributing it to ‘the possibility that, in a race so tight, every vote might actually count’ (Fiore, 2000). However, the real increase in voter turnout compared to 1996 was small. According to Michael McDonald and Samuel Popkin,6 turnout in 2000 was 55.6 percent, three points higher than in 1996. Three percentage points is an increase, but a turnout of 55.6 percent hardly suggests voters flocking to packed polls. Should we be surprised by this paltry increase? Perhaps not. Although the bulk of the research on electoral competition and turnout argues that close elections boost turnout, there are a few dissenting voices. A crosscurrent in the literature suggests that perhaps, as John Ferejohn and Morris Fiorina (1975) memorably put it, closeness only counts in horseshoes and dancing – not in voter turnout. Here and there various authors present evidence that questions the conventional wisdom. Ferejohn and Fiorina ﬁnd little evidence that the perceived closeness of an election aﬀects turnout. Similarly, James Robinson and William Standing (1960) report that in Indiana there has not historically been a correlation between the closeness of an election and voter turnout. Carroll Foster (1984) also reports that when turnout is measured at the state level there is a ‘weak, unstable or nonexistent’ relationship. The same inconsistency exists in congressional elections. Margaret Conway (1981) ﬁnds mixed results when testing for a relationship between electoral margin and turnout in congressional elections. In fact, turnout has actually been higher in less-competitive districts in some years. While Gary Cox and Michael Munger (1989) ﬁnd a closeness–turnout relationship in congressional elections, they stress that it is quite modest. Based on this array of studies it would appear that the ‘law’ linking competitive elections and turnout was made to be broken. Fast forward to 2004. There was every reason to expect a close vote and therefore a high level of voter turnout. This time, turnout was relatively high compared to the low standard of contemporary presidential

Communities, schools and voter turnout

163

elections: 60 percent, according to Michael McDonald. This ﬁnding raises an interesting question: given that the elections of 2000 and 2004 were equally close, why was turnout so much higher in 2004 than 2000? The answer lies in the fact that the literature actually proposes two complementary explanations for why competitive elections enhance voter turnout. Explanation 1 is that voters feel they are more likely to aﬀect the outcome of an election when it is perceived to be close.7 Explanation 2 is simply that political elites, particularly parties, strategically concentrate their resources in electorally competitive places.8 This type of party mobilization contributes to voter turnout.9 Both of these processes clearly work in tandem; as a result, both explanations may be at play in any one election. Political professionals mobilize voters where the electoral margin is narrow, but their appeals are only eﬀective if they can successfully convince voters that in a close contest every vote matters. So, how was 2004 diﬀerent from 2000? First, voter mobilization eﬀorts in 2004 had greater scale and sophistication. The eﬀorts at getting out the vote in 2004, by supporters of both George W. Bush and John Kerry, dwarfed what was done in 2000. This fact, coupled with the diﬀerence in turnout, provides strong evidence that explanations 1 and 2 both apply to the 2004 election. High turnout results from an intensive get-out-the-vote eﬀort, which in turn is most likely to be eﬀective when and where an election is extremely close. Voter mobilization, spurred by a close election, appears to be suﬃcient for high turnout. However, the evidence from recent elections leaves a question about whether an election has to be close to spur high voter turnout. I wish to advance the provocative claim that it is not necessary to have a close election in order to observe high voter turnout. Rather, in contrast to conventional wisdom among election observers, we have reason to believe that turnout can increase as electoral competition decreases. This hypothesis is based on the idea of a norm of civic duty that motivates people; it postulates the existence of homo civicus – people who vote because they believe it is their civic duty. And, as explained below, that sense of civic duty is likely to be stronger in places where elections are not competitive.

HOMO CIVICUS We have seen that a long-standing literature in political science concludes that where elections are competitive, voters behave as homo politicus. However, the social capital literature gives us reason to think that where

164

Social capital: reaching out

elections are won overwhelmingly by one party, voters are better described as homo civicus – voting out of adherence to a social norm that voting is a duty. To understand why the social capital literature leads us to expect noncompetitive elections to produce high voter turnout, we must ﬁrst deﬁne our terms. A multitude of deﬁnitions of social capital exist; I employ one similar to that adopted by Robert Putnam (1993, 2000) and James Coleman (1990): social capital has two components – social networks and social norms. The research presented in this chapter further reinforces Elinor Ostrom’s argument in Chapter 1 of this book of the importance of institutions or norms as a form of social capital. For Ostrom, social capital is a set of relationships and shared values (that is, norms) created and used by multiple individuals to solve collective problems in the present and future. What do I mean by a social norm? Like any social science term, its deﬁnition is contested. For some, a norm is simply a behavioral regularity, something people generally do. The deﬁnition I employ, however, speciﬁes that a norm is rooted in a sense of obligation. My use of the term concurs with two other deﬁnitions, one more formal, the other less so. Sociologist Thomas Voss deﬁnes a norm as: ‘a regularity such that members of [a population] expect that nonconformity will (with positive probability) be punished with (negative) sanctions’ (2001, p. 109). Legal scholar Richard McAdams deﬁnes a social norm as a ‘decentralized behavioral standard that individuals feel obligated to follow’ (1997, p. 381).10 In a memorable turn of phrase, some authors refer to a norm’s ‘oughtness’ – it is something members of a community feel they ought to do, even if they do not always do it (Hechter and Opp, 2001). The seminal work of Coleman contains the theoretical tools to understand how civic norms are propagated. A norm exists, Coleman suggests, when individuals willingly transfer some authority over their behavior to fellow members of their community. ‘[A] norm concerning a speciﬁc action exists when the socially deﬁned right to control the action is held not by the actor but by others’ (1990, p. 243). A norm is enforced within a community through the use of social sanctions. The term ‘sanctions’ may seem vaguely sinister, but it refers to nothing more than the signals people send to one another in everyday social interactions. These sanctions are typically subtle – a disapproving look, a raised eyebrow, the whispered label of ‘shirker’ – but are nonetheless eﬀective (Knack, 1992). They are also largely a matter of anticipated reactions; I anticipate how my neighbors will perceive my behavior, and act to meet their assumed approval (or avoid their disapproval). Putnam uses the example of a neighborhood in which there is a norm to rake one’s leaves to illustrate this type of social sanction:

Communities, schools and voter turnout

165

The norm of keeping lawns leaf-free is powerful in my neighborhood . . . and it constrains my decision as to whether to spend Saturday afternoon watching TV. This norm is not actually taught in local schools, but neighbors mention it when newcomers move in, and they reinforce it in frequent autumnal chats, as well as by obsessive raking of their own yards. Non-rakers risk being shunned at neighborhood events, and non-raking is rare. (1993, p. 171)

This process of norm enforcement through social sanctions only occurs when three conditions hold. First, the prescribed behavior is generally recognized as something that individuals ought to do (or, as the case may be, not do). In Putnam’s example, leaf-raking is widely accepted as normative within his neighborhood. Analogously, voting is deeply ingrained as normative within the collective American psyche, and thus clearly meets this condition. The second and third conditions apply not to the norm, but to the people potentially subject to it. They must recognize the legitimacy of other members within their community to enforce the norm, and they must have social relationships with those people for sanctions to be eﬀective. Again returning to Putnam’s example, the leaf-raking norm only works because residents of his neighborhood care what others think of them, and interact with them in ways that allow sanctions for non-raking to be applied. In Chapter 9 of this book, Kramer talks about these conditions in terms of social identity. If people are operating at a personal level, they will focus on their own interests and norms and sanctions will be less eﬀective. However, if they identify either with an ‘in-group’ – for example, their block or church – or the larger collective – for example, their city or denomination – sanctions and rewards will be much more eﬀective. Members who identify with these groups will also be much more aware of and sensitive to the norms of these groups. Homogeneity within a community facilitates the formation of social identity and the fulﬁllment of the conditions mentioned above. First, people with common backgrounds and beliefs are more likely to arrive at consensus on normative behavior. Second, commonality among members of a community enhances the legitimacy with which they view one another’s opinions. Third, people generally want to associate with others who are ‘like them’. It is this last criterion that leads us to expect that voter turnout will be high in places where elections are not competitive. Jurisdictions that are not electorally competitive are often politically homogeneous. That is, most people are surrounded by others who are ‘like them’ and who probably share similar political views. Burt (2001) deﬁned such a situation as one of social closure within a community. A community with social closure is a discrete, densely connected group. Its sources of social capital lie in that

166

Social capital: reaching out

dense ties enable members to beneﬁt from easy access to information held by others in the community. Such communities develop norms that govern behavior, and because conformity makes behavior more predictable, it provides the basis for trust among community participants. Moreover, several studies have found that social capital, as deﬁned as social closure and however measured, is higher in communities that are more homogeneous (however that is assessed).11 The existing literature typically measures community homogeneity with demographic variables such as race, ethnicity and income. This is done because demographic characteristics are thought to reflect individuals’ values, interests or preferences (terms I use interchangeably). These measures capture what it is about people that leads them to think that others are like them. To this list I add partisanship, which is arguably an even better proxy for preferences than race, ethnicity or income. Consider the following thought experiment. Imagine that you were asked to predict whether two randomly selected people had similar preferences on a particular issue. You can only know one thing about them, and you do not know the issue. What is the one thing that you would want to know? I might suggest that more than their ethnic group, income level or even their race, you would do well to ﬁnd out their partisanship (that is, whether they identify as a Republican or a Democrat), as this is the single best predictor of preferences across a wide range of issues. Partisanship, I contend, is an even better predictor than race, which is undeniably a salient factor in American politics.12 While race shapes much of American politics, a measure of racial heterogeneity will not necessarily reflect diﬀerences in preferences, opinions and values. This is because White voters are not simply the inverse of African-American voters. While African-Americans consistently favor liberal policies, Whites are not uniformly conservative (even on issues related to race).13 Even if the reader accepts that partisanship is an important predictor of issue preferences, the theory still rests on people choosing to associate with others of similar political views. I concede that for many Americans the partisanship of their friends and acquaintances is not something that occupies much of their attention. Americans are not likely to ask one another about their partisan preferences in casual conversation. This might suggest, therefore, that the partisan composition of a community is not terribly relevant for social norms. I contend that in the contemporary United States partisanship is actually a good proxy for one’s cultural outlook, which in turn corresponds to a bundle of issue preferences. You may not know whether your neighbors are Republicans, but you are likely to know something about their lifestyle, like whether both husband and wife work outside the home or whether they

Communities, schools and voter turnout

167

attend church regularly. People who have similar lifestyles are, in turn, likely to form networks of reciprocity with one another. And matters of lifestyle, or what some might call cultural values, are increasingly related to partisanship in America (Layman, 2001). A careful reading of the literature on community homogeneity leads us through the following chain of logic: If: (a) voting is a civically motivated activity and; (b) social norms facilitating civically motivated collective action are stronger in networks characterized by homogenous preferences and; (c) partisanship is our most consistent measure of (or proxy for) preferences then: (d) voter turnout should rise with partisan, or political, homogeneity. And if places with homogeneous political preferences also tend to have non-competitive elections – an almost self-evident claim – it follows that voter turnout should rise in places where elections are won by wide margins. It all ﬁts together in a tidy, logical package. The catch, of course, is the long line of research that argues exactly the opposite: that turnout rises where elections are competitive. We are thus left with competing expectations about the relationship between political homogeneity and turnout. On the one hand, the literature on competitive elections, bolstered by the conventional wisdom among lay observers of the political scene, suggests that political competition should ignite turnout. But on the other hand, a close reading of the social capital literature suggests that turnout should be high in communities where elections are not competitive and are thus politically homogeneous.

HOMO POLITICUS VS HOMO CIVICUS So which is it? Does political competition lead to higher or lower turnout? The answer is both. Figure 6.1 displays what I mean. US counties are arrayed on the x-axis according to their electoral competitiveness. Average voter turnout is shown on the y-axis. We see a curvilinear relationship

168

0

30

60

90

1

2

3

4 5 6 7 Deciles of competitiveness

8

9

10

Figure 6.1

Political competition and voter turnout in the 2004 presidential election (counties)

Note: Competitiveness is measured as the share of the vote taken by the winning candidate in each county. Voter turnout is measured as the number of voters divided by the voting-age population. Alaska and Hawaii do not have counties, and so these results are only for the 48 contiguous states.

Voter turnout (%)

Communities, schools and voter turnout

169

where voter turnout is highest in communities with very low competitiveness (and probably high homogeneity) or where presidential elections are most competitive. Figure 6.1 suggests that voters sometimes behave as homo politicus and other times as homo civicus. Admittedly, though, this evidence is far from conclusive. It could be that places where elections are either highly – or not at all – competitive have other characteristics that spur high levels of turnout. However, more in-depth analysis reveals that this curvilinear relationship is robust across a variety of situations. It holds up when statistical models account for potentially confounding factors that the voluminous literature on turnout has shown to predict voter participation. These include mean education level, household income, level of urbanization, population density, residential mobility and the region of the country in which the county is located. The relationship is also present when controlling for ethnic, racial and income heterogeneity, as well as political competition at the state level. Furthermore, it is robust when modeling individual-level (as compared to county-level) data. Finally, the same relationship holds up in other presidential election years, not just 2004. We are thus able to reject the alternative explanation that the voter turnout in the two diﬀerent situations is simply due to other characteristics of those counties or the people who reside in them. But it is one thing to demonstrate what does not explain a statistical relationship, and quite another to show what does. Is there any aﬃrmative evidence that political homogeneity leads people to adhere to a civic norm? The most straightforward way to determine why people vote is to ask them. If my explanation for high turnout in non-competitive communities is correct, voters there should report turning out because of civic obligation more frequently than in highly competitive elections. The opposite would also be true: voters in places where elections are highly competitive should indicate a more instrumental rationale – for example, influencing the election – for casting a ballot. Fortunately, it is possible to test these two hypotheses with data from the Citizen Participation Study. This extensive analysis of Americans’ civic and political engagement asked people to report their motivation for participating in various forms of collective action. It asked voters about their motivations for turning out; the question succinctly captures the essence of civically motivated participation as I have deﬁned it.Voters in the most recent local or national election were asked whether they voted because it was their ‘duty as a citizen’, indicating whether this was a ‘not very, somewhat, or very important’ reason. Conversely, voters were also asked to indicate the importance of the ‘chance to influence government policy’ as a motivation, which reflects a vote motivated by political concerns.

170

Social capital: reaching out

While these two questions are ideally worded, the way in which they were asked was not optimal for our purposes. Theory predicts that duty trumps policy in homogeneous communities, and policy trumps duty in heterogeneous ones with competitive elections. The best test of voters’ motivations would require them to make a trade-oﬀ between the two motivations. However, the Citizen Participation Study asked voters about each motivation separately. As a result, respondents could rate each one as equally important. Not surprisingly, very high proportions of voters reported that both duty and policy motivated their decision to vote. Sixty-one percent indicated that influencing government policy was a very important reason why they voted, with another 28 percent saying that it was somewhat important. The distribution of responses is even more skewed for the question about duty. Seventy-seven percent of voters reported that civic duty was a very important reason for casting a ballot, while only 5 percent reported that it was not very important. That such a high percentage of respondents report voting because of these two motivations makes them diﬃcult to model, as it is hard to explain variation when there is little variation to explain. A more reﬁned measure of voters’ motivations would allow for a better test of the dual motivations theory. Nonetheless, if we are able to discern a relationship with such a skewed dependent variable, this should bolster our conﬁdence that we have found evidence to support the theory. The hypotheses are straightforward. First, voters should cite civic motivation more frequently than political motivation in communities that are politically homogeneous, with civic motivations declining as heterogeneity rises. Concomitantly, a political motivation should be the least common response in homogeneous communities, but become more common as heterogeneity rises. Each of the two self-reported motivations is regressed on communitylevel political heterogeneity, controlling for an array of potentially confounding factors.14 Political heterogeneity ranges from the least competitive (high homogeneity) to the most competitive (high heterogeneity).15 The results displayed in Figure 6.2 conﬁrm our hypotheses. The probability that a voter reports duty as a very important reason for voting drops nine points as you move from the least to the most competitive county. Conversely, the likelihood that people would cast a politically motivated vote rose signiﬁcantly as political competition rose – an increase of 13 points. In other words, we see homo civicus in the least politically competitive communities, and homo politicus where competition is highest. All the control variables are held constant at their mean values. Why do people vote? Sometimes they do so because they believe their political interests are at stake. But just as often they turn out because of the social capital – as reflected by the civic norms – within their community.

171

Minimum

Maximum Political competition

Political competition and motivations for voting

Results calculated from model described in the text.

50

55

60

65

70

75

80

85

90

Figure 6.2

Note:

Percentage

Influence policy

Civic duty

Motivation

172

Social capital: reaching out

THE EFFECT OF CLASSROOM INSTRUCTION The discussion thus far has been about community norms and how powerful they can be in prompting – or prohibiting – behavior. But are the norms that encourage – or discourage – voting just the ones reflected in a person’s current community? Research has shown that the most eﬀective norms are internalized. The internalization of a norm means: ‘that an individual comes to have an internal sanctioning system which provides punishment when he carries out an action proscribed by the norm or fails to carry out an action prescribed by the norm’ (Coleman, 1990, p. 293). We might say that a norm has been internalized when you act in accordance with it even when no one else is looking. ‘Socialization’ is the process by which one internalizes a norm. As young people undergo socialization, they are imprinted with norms that have the potential to guide their behavior throughout their lives. Given that voting is a norm in most of American society, it is likely that voter turnout is shaped by where people were when they were young as well as where they are now. In my book Why We Vote: How Schools and Communities Shape Our Civic Life (Campbell, 2006), I describe in detail how adolescents who attend schools that foster a sense of civic responsibility develop into adults who vote. Interestingly, little evidence exists that they learn this sense of civic duty from formal classroom instruction. Political scientists have long looked to schools as an important source of preparation for engagement in the political process (Merriam, 1931; Merriam, 1934). However, in this ﬁrst wave of empirical research, little evidence was uncovered that civics instruction had any bearing on civic attitudes and/or behavior in either the short or long term. In summarizing this literature, Richard Niemi and Jane Junn remark: ‘the accepted wisdom in the political science profession is that civics classes have little or no eﬀect on the vast majority of students’ (1998, p. 16). Scholars have oﬀered numerous reasons for the absence of such a link between civics courses and civic attitudes and behavior. In many cases, civics instruction is not conﬁned to a single course, or to school at all, in a way that is unique among academic subjects. Unlike politics, you are not going to learn much algebra or chemistry from reading the newspaper or watching television. Students, however, can absorb a lot of political information from their environment, making classroom instruction redundant. Furthermore, even within schools, politically relevant material is covered in the course of studying subjects other than what was once called civics (and is more commonly known today as social studies). Pamela Conover and Donald Searing have found that this is especially true in literature classes (2000). Finally, Paul Beck and M. Kent Jennings (1982) suggest that because civics instruction is

Communities, schools and voter turnout

173

more or less constant across diﬀerent schools, it is unlikely that there are diﬀerential eﬀects on political knowledge or eﬃcacy. As political scientists’ attention began to turn away from adolescent socialization to other topics, there was little new research on the impact of classroom civics instruction. And so the conventional wisdom remained that classroom civics instruction had little or no impact on civic outcomes. Recently, scholarly attention has again turned to civic education, with research that questions the long-standing conventional wisdom that civics instruction has very little impact. These classes, it turns out, appear to have an eﬀect on civic education after all. However, this new evidence only underscores the dramatically lower expectations within the research community for the impact of civics instruction. Niemi and Junn (1998) have analyzed the 1988 National Assessment of Educational Progress (NAEP) civics exam and found that classroom instruction in civics (usually social studies courses) correlates with a modestly higher score on the test, even when controlling for the students’ demographic characteristics known to aﬀect academic performance. However, this new evidence also underscores the dramatically diminished expectations scientists have with regard to the impact of civics instruction.

THE EFFECT OF THE SCHOOL EXPERIENCE IN ADOLESCENCE But what about the eﬀect of school experiences beyond the formal curriculum? Given the central importance of school in the life of the typical adolescent, and the historic mandate for schools to prepare their students for a lifetime of active and engaged citizenship, could it really be that school experiences have such a limited eﬀect on students’ civic preparation? While it is perhaps natural to look ﬁrst to a school’s formal curriculum, this does not exhaust the ways in which schools can prepare their students for a lifetime of active involvement in their communities. Schools provide a civic education in many, often more subtle ways than formal classes. Past research has clearly shown that institutions – of which a school is one – can enhance shared norms of trustworthiness, trust and reciprocity (Ostrom, 1998; Ahn and Ostrom, 2003). A school also provides its students with experience as members of that community, preparing them for membership in other, larger communities. In the words of a recent report on civic education in America’s schools: Schools are communities in which young people learn to interact, argue, and work together with others, an important condition for future citizenship.

174

Social capital: reaching out

Schools have the capacity to bring together a heterogeneous population of young people – with diﬀerent backgrounds, perspectives, and vocational ambitions – to instruct them in common lessons and values. They can also bring young people into signiﬁcant relationships with adult role models. (Carnegie Corporation and CIRCLE, 2003)

In other words, schools are communities and, consequently, can provide a venue in which young people learn to contribute to the good of a community. Hallinan’s recent research presented in Chapter 5 of this book on the eﬀects of intergenerational social closure on student achievement supports this concept. While network density did not relate to academic achievement in her study, it strongly correlated to a host of non-academic variables. It increased liking school, feeling safe at school and trusting people in the neighborhood. It also had a signiﬁcant positive eﬀect on participation in school clubs or organizations, attending religious services and helping others in the community. Further, it reduced the number of times a student is sent to the principal’s oﬃce. As is clear from this and other research, schools are places where young people can absorb the norms that deﬁne good citizenship. Or, at least, that is a working hypothesis. Empirical evidence puts this proposition to the test. Studying how the collective norms held within the high school you attend as an adolescent might aﬀect your electoral engagement as an adult requires data that simultaneously meet two criteria. First, the study must be contextual. We require measures, not only of individuals’ attitudes, behavior and characteristics, but also of the characteristics of the social environment within their school. Second, the data must be longitudinal. Adolescents interviewed in high school must be tracked into adulthood, with follow-up interviews to determine their level of civic and political engagement. The combination of both elements in a single data set requires a considerable investment of resources (not to mention a fair degree of patience, as the study must unfold over a number of years). The Youth Parent Socialization Study (YPSS) ﬁts the bill. The YPSS began in 1965, when interviews were conducted with a representative sample of high school seniors and their parents. In addition to the 1965 wave of the study, youth and parents were then re-interviewed in 1973 and 1982, thus fulﬁlling the longitudinal criterion. Importantly, in addition the contextual data were provided by additional data from samples from each student’s high school. The data, gathered from roughly 125 seniors in each school, complement the interviews with adolescent panel members and their parents in 1965. It is therefore possible to construct school-level measures for every question asked of the individual students, meeting the need for contextual data. For example, we can determine whether an individual student intends to attend college. Moreover, from the aggregation of the

Communities, schools and voter turnout

175

entire school sample, we can also estimate the percentage of students within that school who plan on going to college. The ability to aggregate responses from the school population provides a strategy to overcome the fundamental challenge of studying the impact of social norms on any behavior, namely ﬁnding an indicator of the norms within a community. The YPSS makes it possible to gauge the civic norms within a high school with a question that asked students: ‘What three things about a person are most important in showing that he is a good citizen?’ Students were then shown a list of choices, all designed to capture diﬀerent dimensions of citizenship, including religious involvement, adherence to the law and a sense of privatism (or minding one’s own business). For the purpose at hand, one response relates directly to voter participation: ‘He votes in elections.’ The beauty of this measure is its simplicity. Because the question relies on the students’ initiative to identify voting as a component of good citizenship, it largely avoids the problem of having most everyone reflexively aﬃrm that voting is a normative expectation, which tends to happen when respondents are simply asked whether they think voting is important. As a result, there is a considerable amount of variation in the percentage of students in each high school who endorse voting as a component of good citizenship. The number ranges from 46 to 85 percent across the 77 schools from which students were sampled (with a mean of 70 percent).16 The percentage of students in each school who endorse voting as a component of good citizenship is the operational deﬁnition of the school’s ‘civic climate’. The ﬁrst hypothesis to be tested is that young people in a school climate that fosters engagement as a civic duty should be more likely to envision themselves as active participants upon reaching adulthood. In other words, are adolescents surrounded by peers who identify voting as a norm more likely to see themselves as active citizens? Anticipated participation is measured straightforwardly, as respondents in the high school sample were asked: ‘Looking ahead to the time when you are on your own, what about actual participation in public aﬀairs and politics? How active do you think you will be in these matters?’ In response, they had three choices: not very active, somewhat active and very active. A question that asks about young people’s anticipated participation is an interesting window into what they see as normative, and is just as much a guide to their present state of mind as a reliable prediction of their future behavior. It is likely that, to an adolescent, a question which essentially says, ‘Will you be civically engaged in the future?’ is really asking, ‘Do you think people should be engaged?’ Or, perhaps more accurately, ‘Do you think people like you should be engaged?’

176

Social capital: reaching out

In essence, the model tests whether adolescents surrounded by peers who identify voting as a norm are more likely to envision themselves as active citizens. In order to be certain that we were not observing the eﬀects of the individual’s own sense of civic duty or of other confounding factors, we controlled for a large number of school and individual variables. At the school level, these include the mean education level of the students’ parents, which serves to account as a proxy for the general socio-economic status of the school’s population. The model further accounts for the average length of time that students have lived in their community, as it is likely that civic norms are more likely to be inculcated in places where people have deep roots (Coleman, 1988). The model also controls for the degree of political diversity within the school population. At the individual level, controls include parents’ education as well as whether the student plans on attending college. Just as at the aggregate level, the model accounts for each individual’s residential stability. Owing to long-standing gender diﬀerences in levels of engagement the model also controls for gender. Perhaps most critically, the degree to which each individual endorses voting as a sign of good citizenship is also included in the equation. In other words, the model isolates the impact of being surrounded by others who hold that voting is a norm by controlling for the individual’s own sense of civic duty. As expected, adolescents who report that good citizens exercise their right to vote also report that they anticipate being publicly engaged upon reaching adulthood. However, there is also an eﬀect of being immersed in a civic duty-rich environment over and above the individual’s own sense of civic obligation. In terms of its magnitude, it is on a par with educational expectations and parents’ education, and far exceeds the impact of gender. We have seen that being surrounded by peers who share the belief that voting is a social norm leads adolescents to anticipate that they will participate in public aﬀairs when they reach adulthood. To the extent that social norms have an impact on behavior in the here and now, this is to be expected – especially if adolescents’ reports of what they will do in the future reflect their current perception of that activity’s normative value. As discussed above, adolescents internalize social norms. That is, the adoption of norms in adolescence aﬀects not only the present, but also the future. If a norm encouraging voting is internalized in adolescence, we should also see that it aﬀects public engagement in adulthood. While these results suggest a link between norms in one’s youth and public engagement as an adult, the evidence is hardly deﬁnitive. For one, we do not know whether adolescents’ stated expectation that they will be politically engaged at some unspeciﬁed point in the future actually translates into engagement in adulthood. Consider the model of anticipated

Communities, schools and voter turnout

177

participation, therefore, to lead to the critical question: determining whether the anticipation of engagement in adolescence corresponds to actual behavior in adulthood. Speciﬁcally, does the civic climate of one’s high school aﬀect the likelihood of turning out to vote in the years following high school?

THE LINK BETWEEN ADOLESCENT ENGAGEMENT AND ADULT VOTING Adolescence is an especially signiﬁcant period for the internalization of social norms because it aﬀects not just the present but also the future. If a norm encouraging public engagement, namely voting, is internalized then, we should see the civic climate of a high school reach into adulthood to aﬀect public engagement. Our inquiry centers on the second wave of follow-up interviews, conducted in 1982, which asked the study’s participants to report on whether they turned out in the 1980 presidential election. This was 15 years after the panel members’ graduation from high school. At the individual level, this model controls for education level, marital status, gender and length of time in one’s present community, all factors past research has shown have a bearing on voter turnout (Wolﬁnger and Rosenstone, 1980; Rosenstone and Hansen, 1993). In addition to these factors in the present, it is also important to pay attention to the past. Since the home is a critical factor aﬀecting an individual’s civic development, it is important to control for the engagement level of students’ parents. That is, students with participatory parents might be inclined to report that such engagement is an indication of good citizenship, and so three measures of parental engagement have been included. First is an index of the parents’ level of political activism. Another accounts for whether, in 1964, parents reported taking an ‘active part’ in ‘local or community aﬀairs’ or ‘anything of that kind’. The third records whether the parent who participated in the study voted in the most recent (1964) presidential election. Finally, the model also controls for whether the individual explicitly endorsed voting as an essential component of good citizenship. In addition to individual-level control variables, the model also includes characteristics of the respondent’s high school. School-level controls include the mean level of parents’ education, residential mobility, the aggregate level of anticipated participation and political diversity within the school environment. Of course, for the purpose of testing the impact of civic norms, the most signiﬁcant school-level measure is its civic climate, namely the percentage of students who view voting as an obligation of

178

Social capital: reaching out

citizenship. Finally, the model also accounts for the political composition of the community in which respondents were living when they were interviewed as adults, as a rough indicator of the extent to which voter turnout is driven by civic or political motivations, in the present. The dependent variable is whether the respondent voted in the 1980 presidential election or not. Does the civic climate of one’s high school aﬀect the likelihood of turning out to vote in the years following high school? Simply put, it does. The impact of civic climate on voter turnout is signiﬁcant, accounting for an increase of ten percentage points when the civic climate goes from its minimum level to its maximum one and other variables are held at their means. It has a stronger impact than a shift in political homogeneity (seven points). The eﬀect is also comparable to other well-established variables such as having a college education (13 points), being married (11 points) and parental political activism (12 points).17 Figure 6.3 depicts these results.

CONCLUSIONS We have seen how the civic climate of one’s high school has an impact on voter turnout at least 15 years following high school. However, the pithiness of this statement perhaps obscures its full signiﬁcance. First, we have seen that it is the civic climate of the school community that has a robust impact on participation in adulthood, not the degree to which an adolescent’s peers see themselves as active in public aﬀairs. What matters is that an adolescent’s environment is populated with a high percentage of peers who express that voting is an indicator of good citizenship – homo civicus. A high school context characterized by adolescents who are politically engaged, and anticipate being so in the future, does not have a similar eﬀect. Second, we have seen that it is the norms within the adolescent’s community, deﬁned in this case as the high school, that matter. In fact, once we account for environmental factors, an individual’s own belief that voting is a civic duty does not have an impact on voting as an adult. Third, strong civic norms within a school community reach into the future. Where you were then aﬀects what you do now. We are thus reminded that individual-level factors are only part of the story explaining participation, and political behavior, more generally. While most social scientists recognize the hazards of inferring individuallevel behavior from aggregate data, it can also be the case that inferring individual-level behavior without information about individuals’ social

179

0

5

Impacts on the probability of voting in 1980

Results calculated from model described in the text.

Figure 6.3

Note:

Married

Parents’ campaign activism (1965)

Education

Political homogeneity (present community)

High school civic climate (1965)

10 Percentage points

15

20

180

Social capital: reaching out

contexts is equally fallacious (Huckfeldt and Sprague, 1993). Individuals do not act, nor are they acted upon, in isolation. Rather, norms are inculcated within communities, such as the family, the neighborhood and – as we have seen – the school.

BEYOND HOMOGENEITY Because of the earlier focus on the link between homogeneity and civic norms, I want to emphasize what I am not suggesting. A quick glance at the foregoing discussion might suggest that homogeneity is the solution to our civic ills – conservatives should be shunted oﬀ to their communities and schools, and liberals to theirs. However, a closer reading of both the empirical results and the supporting theoretical apparatus clariﬁes why doing this will not pull us out of our participatory tailspin. Even though I have used measures of communities’ partisan heterogeneity as proxies for the presence of civic norms, this does not mean that the strength of norms is inevitably a function of a community’s political composition. Nor is the strength of the norms necessarily determined by religious, ethnic, racial, income or any other type of demographic heterogeneity. Norms are strong where people consider themselves to have something ‘in common’ with others. The perception of what we have in common is entirely malleable. Today, partisan preferences unite us (and, consequently, also divide us). Tomorrow, other factors could – and hopefully will – overshadow partisanship as the glue bonding people together. A sensible conclusion to draw from these results is that any school-based reform with the objective of enhancing voter turnout among the rising generation should focus on ways to foster civic norms. A promising course of action to foster civic norms, therefore, is to identify ways to foster a sense of commonality among a school’s students. At this point, I am not suggesting that the means for doing so are clear. Rather, I submit that the suggestive evidence presented here points us in the direction for future research. We need to learn how schools can foster cohesiveness, linked to a sense of civic obligation among their students. It would be premature to advocate a silver bullet as the way to strengthen the civic commitment of today’s youth, as we know little about the process of civic development among adolescents. While these data do not enable us to determine precisely what should be done, they do suggest what we should be looking for: ways to thicken civic norms.

Communities, schools and voter turnout

181

BEYOND VOTING These ﬁndings about the importance of norms are not limited to voter turnout or to schools. Voting is only one form of collective action, and schools are only one type of institution in which norms are fostered. Other institutions, including ﬁrms, can also inculcate norms encouraging cooperative behavior. And there is evidence that, as with schools, the ﬁrms in which people spend their early, and thus formative, years imprint behavioral norms for the duration of their careers (Higgins, 2005). As scholars of business increasingly turn to the study of social capital, they are well advised to take heed that social capital not only entails the extent of individuals’ social networks, but also the communal normative culture in which those individuals are embedded.

NOTES 1.

This chapter sketches out my explanation for why people vote; greater depth can be found in my book Why We Vote: How Schools and Communities Shape Our Civic Life (Campbell, 2006). 2. Patterson and Caldeira (1984), Caldeira et al. (1990), Cebula and Murphy (1980), Nagler and Leighley (1992), Milbrath (1971), Cox (1988), Cox and Munger (1989), Filer and Kenny (1980), Gray (1976), Kau and Rubin (1976), Chambers and Davis (1978), Settle and Abrams (1976), Kim et al. (1975), Leighley and Nagler (1992), Shachar and Nalebuﬀ (1999). 3. Crain et al. (1987), Caldeira et al. (1985), Silberman and Durden (1975). 4. Patterson and Caldeira (1983), Barzel and Silberberg (1973). 5. Filer et al. (1993), Rosenstone and Hansen (1993). 6. These turnout ﬁgures are derived from the estimates of Michael McDonald and Samuel Popkin, who have done political science a valuable service by pointing out that accurate turnout ﬁgures require an accurate count of the population eligible to vote. While the verdict is still out on the utility of their adjusted ﬁgures (Freeman, 2001), as their calculations of the voting eligible population are estimates at best, at the very least we can say that they are an improvement over using the voting-age population, which includes people who are not eligible to vote (ex-felons in some states, non-citizens in all states). See McDonald and Popkin (2001) and http://elections.gmu.edu/voter_turnout.htm. 7. Barzel and Silberberg (1973), Crain et al. (1987), Filer and Kenny (1980). 8. Chambers and Davis (1978), Copeland (1983), Cox and Munger (1989), Key (1949), Patterson and Caldeira (1983), Settle and Abrams (1976). 9. Gosnell (1927), Green and Gerber (2004), Kramer (1970), Rosenstone and Hansen (1993), Wielhouwer and Lockerbie (1994). 10. Cited in Schudsen (2001). 11. Alesina and LaFerrara (2000), Costa and Kahn (1993), Glaeser et al. (2000), Knack and Keefer (1997), Vigdor (2001). 12. Carmines and Stimson (1989), Kinder and Sanders (1996). 13. For more evidence on this point, see Chapter 2 of Why We Vote (Campbell, 2006). 14. At the individual level, the control variables include education level, family income, age, age squared, marital status, race, gender, years lived in the community and political interest. At the community level, the controls include the percentage with a college degree, mean household income, the degree of urbanization, population density, mean

182

15.

16. 17.

Social capital: reaching out residential mobility and mean commuting time. See Why We Vote (Campbell, 2006), Chapter 3, for the parameter estimates and more explanation for the choice of these controls. The level of political heterogeneity within a county is calculated as the mean level of competitiveness of presidential elections from 1980 to 1988 (since the survey was done in 1990). Ordered logistic regression is the estimator, since the dependent variable consists of three ordered categories. Ninety-seven schools were attended by members of the YPSS student panel, meaning that contextual data were collected for about 80 percent of the panel members’ schools. In other words, I have estimated the probability of voting by multiplying the coeﬃcients in the model by reasonable values for the variables they represent. In this case, I have opted to vary civic climate from its minimum to maximum value. Selecting a more restricted range would reduce its absolute impact on the likelihood of voting, but not its impact relative to other variables.

REFERENCES Ahn, T.K. and Elinor Ostrom (2003), Foundations of Social Capital, Cheltenham, UK: Edward Elgar, pp. xi–xxxix. Alesina, Alberto and Eliana La Ferrara (2000), ‘Participation in heterogeneous communities’, Quarterly Journal of Economics, August, 847–904. Barzel, Yoram and Eugene Silberberg (1973), ‘Is the act of voting rational?’, Public Choice, 16, 51–8. Beck, Paul Allen and M. Kent Jennings (1982), ‘Pathways to participation’, American Political Science Review, 76 (1), 94–108. Caldeira, Gregory and Samuel C. Patterson (1982), ‘Contextual influences on participation in US state legislative elections’, Legislative Studies Quarterly, 7 (3), 359–81. Caldeira, Gregory A., Samuel C. Patterson and Gregor A. Markko (1985), ‘The mobilization of voters in congressional elections’, Journal of Politics, 47 (2), 490–509. Caldeira, Gregory A., Aage R. Clausen and Samuel C. Patterson (1990), ‘Partisan mobilization and electoral participation’, Electoral Studies, 9 (3), 191–204. Campbell, David E. (2006), Why We Vote: How Schools and Communities Shape our Civic Life, Princeton, NJ: Princeton University Press. Carmines, Edward G. and James A. Stimson (1989), Issue Evolution: Race and the Transformation of American Politics, Princeton, NJ: Princeton University Press. Carnegie Corporation and CIRCLE (2003), ‘The civic mission of schools’, available at www.civicmissionofschools.org. Cebula, Richard J. and Dennis R. Murphy (1980), ‘The electoral college and voter participation rates: an exploratory note’, Public Choice, 35, 185–90. Chambers, William N. and Philip C. Davis (1978), ‘Party competition and mass participation: the case of the democratizing party system, 1824–1852’, in Joel H. Silbey, Allan G. Bogue and William H. Flanigan (eds), The History of American Electoral Behavior, Princeton, NJ: Princeton University Press, pp. 174–97. Coleman, James S. (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94 (Supplement), S95–S120. Coleman, James S. (1990), Foundations of Social Theory, Cambridge, MA: Harvard University Press.

Communities, schools and voter turnout

183

Conover, Pamela Johnston and Donald D. Searing (2000), ‘A political socialization perspective’, in Lorraine M. McDonnell, P. Michael Timpane and Roger Benjamin (eds), Rediscovering the Democratic Purposes of Education, Lawrence, KS: University Press of Kansas, pp. 91–124. Conway, M. Margaret (1981), ‘Political participation in midterm congressional elections: attitudinal and social characteristics during the 1970s’, American Politics Quarterly, 9 (2), 221–44. Copeland, Gary W. (1983), ‘Activating voters in congressional elections’, Political Behavior, 5 (4), 391–401. Cox, Gary W. (1988), ‘Closeness and turnout: a methodological note’, Journal of Politics, 50 (3), 768–75. Cox, Gary W. and Michael C. Munger (1989), ‘Closeness, expenditures, and turnout in the 1982 house elections’, American Journal of Political Science, 83 (1), 217–31. Crain, W. Mark, Donald R. Leavens and Lynn Abbott (1987), ‘Voting and not voting at the same time’, Public Choice, 53, 221–9. Downs, Anthony (1957), An Economic Theory of Democracy, New York: Harper & Brothers. Ferejohn, John A. and Morris P. Fiorina (1975), ‘Closeness counts only in horseshoes and dancing’, American Political Science Review, 69 (3), 920–25. Filer, John E. and Lawrence W. Kenny (1980), ‘Voter turnout and the beneﬁts of voting’, Public Choice, 35, 575–85. Filer, John E., Lawrence W. Kenny and Rebecca B. Morton (1993), ‘Redistribution, income, and voting’, American Journal of Political Science, 37 (1), 63–87. Fiore, Faye (2000), ‘Decision 2000’, Los Angeles Times, 8 November, p. 20. Foster, Carroll B. (1984), ‘The performance of rational voter models in recent presidential elections’, American Political Science Review, 78 (3), 678–90. Freeman, Richard B. (2001), ‘What, me vote?’ paper for the Russell Sage Foundation Project on the Social Dimensions of Inequality. Glaeser, Edward L., David I. Laibson, Jose A. Scheinkman and Andrei Shleifer (2000), ‘Measuring trust’, Quarterly Journal of Economics, 115 (3), 811–46. Gosnell, Harold F. (1927), Getting Out the Vote: An Experiment in the Stimulation of Voting, Chicago, IL: University of Chicago Press. Gray, Virginia (1976), ‘A note on competition and turnout in the American states’, Journal of Politics, 38 (1), 153–8. Green, Donald P. and Alan S. Gerber (2004), Get Out the Vote! How to Increase Voter Turnout, Washington, DC: Brookings Institution Press. Gumbel, Andrew (2000), ‘Apathy is the loser as crucial race bucks trend of falling turnout’, Independent, 8 November, p. 7. Hechter, Michael and Karl-Dieter Opp (2001), ‘Introduction’, in Michael Hechter and Karl-Dieter Opp (eds), Social Norms, New York: Russell Sage Foundation, pp. 3–24. Higgins, Monica (2005), Career Imprints: Creating Leaders Across an Industry, San Francisco, CA: Jossey-Bass. Hoﬀstetter, C. Richard (1973), ‘Inter-party competition and electoral turnout: the case of Indiana’, American Journal of Political Science, 17 (2), 351–66. Huckfeldt, Robert and John Sprague (1993), ‘Citizens, contexts, and politics’, in Ada W. Finifter (ed.), Political Science: The State of the Discipline II, Washington, DC: American Political Science Association. Kau, James B. and Paul H. Rubin (1976), ‘The electoral college and the rational vote’, Public Choice, 27, 101–7.

184

Social capital: reaching out

Key, V.O. (1949), Southern Politics in State and Nation, New York: Random House. Kim, Jae-On, John R. Petrocik and Stephen N. Enokson (1975), ‘Voter turnout among the American states: systemic and individual components’, American Political Science Review, 69 (1), 107–23. Kinder, Donald R. and Lynn M. Sanders (1996), Divided by Color: Racial Politics and Democratic Ideals, Chicago, IL: University of Chicago Press. Knack, Stephen (1992), ‘Civic norms, social sanctions, and voter turnout’, Rationality and Society, 4 (2), 133–56. Knack, Stephen and Philip Keefer (1997), ‘Does social capital have an economic payoﬀ? A cross-country investigation’, Quarterly Journal of Economics, 62, 1251–88. Kramer, Gerald H. (1970), ‘The eﬀects of precinct-level canvassing on voter behavior’, Public Opinion Quarterly, 34 (4), 560–72. Layman, Geoﬀrey (2001), The Great Divide: Religious and Cultural Conflict in American Party Politics, New York: Columbia University Press. Leighley, Jan E. and Jonathan Nagler (1992), ‘Individual and systemic influences on turnout: who votes?’, Journal of Politics, 54 (3), 718–40. McAdams, Richard H. (1997), ‘The origin, development, and regulation of norms’, Michigan Law Review, 96 (November), 338–433. McDonald, Michael P. and Samuel L. Popkin (2001), ‘The myth of the vanishing voter’, American Political Science Review, 95 (4), 963–74. Merriam, Charles E. (1931), The Making of Citizens: A Comparative Study of Methods of Civic Training, Chicago, IL: University of Chicago Press. Merriam, Charles (1934), Civic Education in the United States, New York: Scribner’s Sons. Milbrath, Lester W. (1971), ‘Individuals and government’, in Herbert Jacob and Kenneth N. Vines (eds), Politics in the American States: A Comparative Analysis, Boston, MA: Little, Brown & Company. Nagler, Jonathan and Jan Leighley (1992), ‘Presidential campaign expenditures: evidence on allocations and eﬀects’, Public Choice, 73, 319–33. Niemi, Richard G. and Jane Junn (1998), Civic Education: What Makes Students Learn, New Haven, CT: Yale University Press. Olson, Mancur (1965), The Logic of Collective Action: Public Goods and the Theory of Groups, Cambridge, MA: Harvard University Press. Ostrom, Elinor (1998), ‘A behavioral approach to the rational choice theory of collective action’, American Political Science Review, 92 (1), 1–22. Patterson, Samuel C. and Gregory A. Caldeira (1983), ‘Getting out the vote: participation in gubernatorial elections’, American Political Science Review, 77 (3), 675–89. Patterson, Samuel C. and Gregory A. Caldeira (1984), ‘The etiology of party competition’, American Political Science Review, 78 (3), 691–707. Putnam, Robert D. (1993), Making Democracy Work: Civic Traditions in Modern Italy, Princeton, NJ: Princeton University Press. Putnam, Robert D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Robinson, James A. and William H. Standing (1960), ‘Some correlates of voter participation: the case of Indiana’, Journal of Politics, 22 (1), 96–111. Rosenstone, Steven J. and John Mark Hansen (1993), Mobilization, Participation, and Democracy in America, New York: Macmillan. Schudsen, Michael (2001), ‘The objectivity norm in American journalism’, Journalism, 2 (2), 149–70.

Communities, schools and voter turnout

185

Settle, Rusell F. and Buron A. Abrams (1976), ‘The determinants of voter participation: a more general model’, Public Choice, 27, 81–9. Shachar, Ron and Barry Nalebuﬀ (1999), ‘Follow the leader: theory and evidence on political participation’, American Economic Review, 89 (3), 525–47. Silberman, Jonathan and Garey Durden (1975), ‘The rational behavior theory of voter participation: the evidence from congressional elections’, Public Choice, 23, 101–8. Stepp, Diane R. (2000), ‘Georgians scurry up and wait to cast their ballots’, Atlanta Constitution, 8 November, p. 1A. Tullock, Gordon (1967), Toward a Mathematics of Politics, Ann Arbor, MI: University of Michigan Press. Vigdor, Jacob (2001), ‘Community composition and collective action: analyzing initial mail response to the 2000 Census’, Duke University, unpublished. Voss, Thomas (2001), ‘Game-theoretical perspectives on the emergence of social norms’, in Michael Hechter and Karl-Dieter Opp (eds), Social Norms, New York: Russell Sage Foundation, pp. 105–36. Wielhouwer, Peter W. and Brad Lockerbie (1994), ‘Party contacting and political participation, 1952–90’, American Journal of Political Science, 38 (1), 211–29. Wolﬁnger, Raymond E. and Steven J. Rosenstone (1980), Who Votes?’, New Haven, CT: Yale University Press.

7.

Experimental approaches to the diﬀusion of norms David W. Nickerson

Many theories of social capital contend that dense ties within networks develop norms that govern the behavior of group members. Dense networks also establish surveillance and sanctioning systems for violations of those norms. Compliance with those norms provides the basis for trust in the community as behavior becomes more predictable. Coleman (1988), Burt (2005) and many other scholars argue that networks with closure are sources of social capital precisely because densely connected groups transmit and reinforce group norms. Thus, social networks are perceived as critical in the diﬀusion and reinforcement of important civic and other norms. Unfortunately, selection into social networks makes it extremely diﬃcult to isolate the unique role played by the network in the formation of norms. Speciﬁcally, how does one disentangle the influence of the network over time from that of the traits that influenced the person to join the network in the ﬁrst place? Randomized ﬁeld experiments oﬀer a method for measuring the role of these two processes. Researchers can pursue three diﬀerent experimental strategies to measure norm creation in the ﬁeld: create a randomized social network; randomly manipulate interactions within an existing network; or trace the eﬀect of an exogenous shock through the network. This chapter briefly explains the logic motivating experiments and provides an example of each type of experiment to study norm formation within social networks.1 Friends, family members, co-workers and neighbors influence one anothers’ actions and attitudes. Neighbors tend to vote in blocs2 and children often follow their parents’ lead in political behaviors and beliefs.3 These phenomena lead some to believe that interpersonal influence is obvious and ubiquitous. This very ubiquity, however, makes accurate detection of interpersonal influence a complicated endeavor. For instance, neighborhoods often exhibit a remarkable homogeneity with regards to aesthetics, social status and even norms. Social pressures may lead to conformity regarding lawn care standards, but it is also possible that the family chose to move into the neighborhood because they felt 186

Experimental approaches to the diffusion of norms

187

comfortable with the existing norms. Casual discussions about politics can create agreement on political matters, but people living in close proximity often possess similar incomes and, thus, share similar material interests that shape political views. Nearly every example of norm formation within social networks possesses an equally plausible counter-explanation involving selection. Randomized experiments oﬀer a method of isolating the eﬀects of interpersonal influence.

THE BENEFITS OF RANDOMIZATION For our purposes, an experiment is any random application of a factor on a pool of subjects. Classic experiments randomly divide subjects into a treatment group, which receives an intervention, and a control group, which does not.4 Because the division is random, the treatment group and control group should be comparable. In other words, the likelihood that a subject receives the intervention is not aﬀected by their age, income, education, gregariousness, curiosity or any other trait. An experiment has been correctly designed when a subject’s assignment to treatment or control provides no information about their demographic or other characteristics. The chief attraction of experiments is the ease with which they can isolate unique influences of the manipulated treatment. The empirical leverage aﬀorded by experiments can be illustrated by thinking of one person’s behavior as a function of another person’s behavior and outside factors. Equation (7.1) models the eﬀect of person 1’s behavior on person 2 where A2 represents the attitudes or actions of person 2, c is a constant, A1 the attitudes or actions of person 1, X are the variables that can be measured and controlled for, g(H) are the known causes of both people’s attitudes, and is the idiosyncratic causes of person 2’s behavior: A2 c A1 X g(H)

(7.1)

Because H is unmeasured and correlates with A1, E[ | A1] 0; as a result, estimates of may be biased using most observational techniques. But suppose that an experimenter devised a study where subjects were randomly divided into a treatment group that was exposed to the opinions of their neighbors (A1 1), and a control group that was never informed of how neighbors felt (A1 0). We could then apply equation (7.1) to both the treatment group, denoted by subscript T, and the control group, denoted by subscript C. Equation (7.2) presents the diﬀerence between the treatment and control groups:

188

Social capital: reaching out

A2C cC A1C XC g(GC ) C A2T cT A1T XT g(GT ) T A2T A2C cT cC (A1T A1C ) (XT XC ) g(GT ) g(GC ) T C

(7.2)

Because the treatment and control groups are randomly assigned, the manipulated factor A1 should be the only independent variable that diﬀers between the two groups. With suﬃciently large sample sizes the observed, unobserved and contingent factors balance out. This is expressed mathematically as E(cT cC ) 0, E(XT XC ) 0, E[g(GT ) g(GC )] 0 and E(T C ) 0. Thus, the treatment is dichotomous, equation (7.2) can be reduced to: A2T A2C (A1T A1C )

(7.3)

In other words, to estimate the eﬀect of one person’s opinions on another person’s opinions, , the experimenter simply needs to subtract the mean of the control group from the mean of the treatment group when a randomized experiment is conducted.5 The experiment’s design and careful implementation thereby allow the simple subtraction of two averages to yield an unbiased estimate of a particular form of interpersonal influence.6 Laboratory versus Field Experiments Laboratory experiments are a commonly employed tool in social science. They have examined the eﬀect of discussing news with others (for example, Iyengar and Kinder, 1987), the role of personal information in trust games (for example, Wilson and Musick, 1997), exposure to discordant ideas on political tolerance, and the spread of rumors (for example, Allport and Postman, 1940). Unfortunately, such studies remove subjects from the social networks the experimenter is trying to assess. Because the setting is artiﬁcial, ﬁndings may or may not translate to the original networks. While such studies provide interesting insights into interactions in anonymous settings (of which there are many), they appear to be severely limited in their ability to deliver ﬁndings about the original social networks. Ultimately, the extent to which laboratory ﬁndings translate to actual neighborhoods needs to be tested. If the goal is to describe interactions in settings in which subjects exist on a day-to-day basis, ﬁeld experiments are essential to conﬁrm and extend results from the lab.

Experimental approaches to the diffusion of norms

189

There is no obvious method of making organic social networks amenable to experimental manipulation. In many instances, the task will prove impossible. However, in principle, randomized ﬁeld experiments can study interpersonal and social influence in three ways: (1) create a randomized social network; (2) control the flow of communication within an existing social network; or (3) introduce exogenous shocks into the network and trace the eﬀect of the shock.7 Examples of each one of these tactics are presented in order.

CREATING AND MEASURING A RANDOMIZED NETWORK Observational studies of neighborhoods such as Oliver (2001) and Huckfeldt and Sprague (1995) may report biased estimates of influence and participation because subjects self-selected into the communities and neighborhoods where they live. But what if it were possible to randomly assign residents within a city to neighborhoods? Random assignment would solve the confounding influence of selection, and the models Huckfeldt and Sprague analyze would yield unbiased estimates of the influence of political context. Obviously, such an experiment would be unfeasible and unethical. However, randomized assignment may be possible within specialized communities. There are many settings in which persons are assigned to groups: recruits to camps and barracks in the military; poor families to housing developments, apartment buildings and speciﬁc apartments (see Katz et al., 2001); prisoners to prisons, cells and blocks; and college students to dorms and rooms (see Sacerdote, 2001; Winston and Zimmerman, 2003; Stinebrickner and Stinebrickner, 2001; Hoel et al., 2004; McEwan and Soderberg, 2004; Marmaros and Sacerdote, 2002; Duncan et al., 2003). The cases of college students and prisoners are particularly interesting because the governing authority assigns not only rooms but also room- or cellmates. After assignments are made, people explore the social topography and form relationships with those around them.8 Usually assignments within these communities use a conscious decisionmaking process. Many colleges ask incoming students to answer surveys of varying lengths to determine the compatibility of roommates and create demographic balance within dorms. But in the instances where assignments are made randomly (not just haphazardly), an experiment has been conducted and all a researcher has to do is collect the data. Because all the persons of a certain type have been randomly distributed, there is no risk that civically engaged persons opted into the same apartment building or

190

Social capital: reaching out

that all violent criminals are in a particular cellblock. If interpersonal influence eﬀects are strong enough, pockets of norms, behaviors and opinions will form beyond what is attributable to random chance. If the context eﬀects are weak, then any pockets of homogeneity will probably be the result of sampling variance (that is, a group of Republicans placed together by random chance) and estimates of influence will not reject the null hypothesis of individuals being unaﬀected by their roommates.9 It is important to note what is measured when assignments are controlled within a social network. Such experiments measure context eﬀects and not individual interactions. While neighbors (and perhaps roommates) are externally imposed, the subject remains free to select the people with whom they interact the most – their friends. A regression analysis considering the influence of the norms in an apartment building will yield accurate estimates of the eﬀect of these norms on the behavior and beliefs of the individual subject. In contrast, a survey comparing the attitudes of friends will be biased because of self-selection and unobserved heterogeneity. The randomization of one facet of social interaction does not purge the entire process of bias.10 An experiment that used randomized social networks to study the spread of norms was conducted during the 2002–03 school year at a college that randomly assigned new students to housing (see Nickerson, 2005, Chapter 3). The school is typical of residential campuses across the USA. It is a somewhat selective, mid-sized private school in a suburban East Coast setting. Unlike students at Dartmouth, Reed and Williams College, the students at this college were not the academic elite. Racially, the student body mirrors broader society; however, with tuition, room and board costing more than $30 000, the students come from wealthier than average families. Students were randomly assigned to rooms within suites of dormitories. Rooms generally contained two students and suites consisted of four rooms sharing a living space and bathroom. Given the relatively intimate accommodations, one could reasonably expect behavioral contagion among both roommates and suitemates. At the beginning of the school year, there should be no correlation between the behaviors and beliefs of randomly selected roommate pairs, as every student has an equal chance of being placed with a given roommate of the same gender.11 The predisposition of roommates (and suitemates) constitutes the treatment to which a student was exposed. Some students had apathetic roommates with no real political opinions, while others were presented with strident Democrats or Republicans. Non-random clusters of opinions and behaviors at the end of the school year (that is, roommates possess correlated behaviors and beliefs) would provide evidence of influence. To measure the influence, one can model student i’s end-of-year attitudes (A2i), which are a function of her attitudes in the beginning of the year (A1i)

Experimental approaches to the diffusion of norms

191

and the starting opinions of her roommates (A1Ri) and suitemates (A1Si). To account for the single-sex nature of rooms and suites, a dummy variable for gender is also included. Equation (7.4) summarizes the regression equation to be used where c is the constant and i represents the unobserved, idiosyncratic causes of a student’s opinion at the end of the year:12 A2i c 1A1i 1A1R 2A1S 2Female i i

i

(7.4)

Data for the study came from a 28-question, two-wave panel survey administered during the second day of freshman orientation and the week prior to ﬁnal exams in the spring. Roughly half of the entering class took both the ﬁrst and second wave of the survey. However, 45 percent of the students did not complete the full year of study and not all respondents who had roommates and suitemates completed the survey. As a result, the sample of students answering both surveys with pre-survey information for roommates and suitemates is relatively small (N = 230). Fortunately, beginning-of-the-year attitudes explain most of the variance in end-ofyear attitudes, so the study has suﬃcient statistical power to detect influence.13 Questions probed a range of political beliefs ranging from broad ideology to sentiments regarding speciﬁc groups or attitudes about particular programs. Table 7.1 presents the results for those questions exhibiting influence.14 The bottom line of Table 7.1 is that roommates and suitemates can shift a person’s opinions, but the shift is neither radical and nor does it necessarily breed like-mindedness. Imagine two students who begin the year indiﬀerent to homosexuals (that is, rating of 0). One student is placed with a roommate who has an unfavorable view of homosexuals (that is, 2), the other student with a roommate with a favorable view of homosexuals (that is, 2). By the end of the year, our results predict that one student will have a slightly unfavorable view of homosexuals (that is, 0.5), the other a slightly favorable view (that is, 0.5). Even more striking is the fact that a student’s attitude at the beginning of the year is by far the best predictor of the student’s opinion at the end of the school year.15 The coeﬃcients are much larger than the peer eﬀect coeﬃcients, with most in the range of 0.35 to 0.56 and statistically signiﬁcant. For the homosexual question, the pre-test coeﬃcient is over twice as high as the roommate coeﬃcient and the coeﬃcient for the pot smokers question was over three times larger than the suitemate one. Some behaviors do seem to shift somewhat based on roommate and suitemate influence. Few would be surprised that students whose roommates enjoy discussing politics also report discussing politics (though they are no more interested in politics at the end of the year). Similarly, students

192

Discuss politics

0.10 (0.26) 0.25*** (0.03) 0.12*** (0.04) 0.13** (0.06) 0.12 (0.16) 230 0.23

Number of days a week you read a newspaper 0.33 (0.47) 0.34*** (0.04) 0.18*** (0.05) 0.15* (0.08) 0.33 (0.20) 230 0.26

Number of days a week watch television 0.59*** (0.13) 0.37*** (0.06) 0.15** (0.06) 0.20** (0.10) 0.30* (0.17) 202 0.20

Drinking age

Notes: 1. The dependent variable is the post-test response of the student. 2. Unless otherwise stated, questions were on a 5-point agree/disagree scale. 3. Favor/unfavorable ratings are on a 7-point scale. 4. Implies * p 0.1; ** implies p 0.05; *** implies p 0.01.

Constant

0.06 (0.14) Pretest 0.57*** Answer (0.07) Pretest 0.18** Roommate (0.07) Pretest 0.21* Suitemate (0.12) Female 0.24 (0.18) N 204 Adj-R-sq 0.30

Variable

0.23 (0.14) 0.70*** (0.05) 0.19*** (0.06) 0.24** (0.10) 0.05 (0.20) 197 0.51

Favorable / unfavorable pot smokers

0.09 (0.14) 0.66*** (0.06) 0.14** (0.07) 0.16 (0.12) 0.02 (0.18) 198 0.38

Legalize marijuana

Table 7.1 OLS regression results modeling the peer effects of roommates and suitemates

0.20 (0.12) 0.56*** (0.05) 0.24*** (0.05) 0.16 (0.10) 0.46** (0.19) 204 0.51

Favorable / unfavorable homosexuals

0.53*** (0.12) 0.36*** (0.04) 0.10** (0.05) 0.14 (0.09) 0.38*** (0.12) 202 0.36

Homosexual teachers

Experimental approaches to the diffusion of norms

193

whose suitemates regularly read the newspaper or watch television are more likely to consume such readily available media. However, even opinion questions on the legal drinking age and marijuana may capture behaviors. Roommates who procure illicit substances may alter attitudes simply by lowering barriers to usage. Attitudes concerning homosexuality are the only exception to this pattern. A possible explanation is that, freed from direct parental oversight, college represents the ﬁrst opportunity for young people to encounter homosexuality. The intimate living conditions with non-family members may increase the relevance of the topic. The combination of unfamiliarity and importance probably creates an environment where interpersonal influence is strong. The results of the study are not surprising, but they are instructive. By far the best predictors of a student’s attitudes at the end of the year are his or her attitudes at the beginning of the year. Moreover, no peer eﬀects whatsoever were detected on bedrock political attitudes such as partisan identiﬁcation, taxes and welfare. The views of roommates and suitemates matter, but only on behavioral, unfamiliar or highly salient topics, and the eﬀects are seen only on the margins.

RANDOMIZED STRUCTURED INTERACTIONS WITHIN AN EXISTING NETWORK If answers about common residential life are desired, one must study behavior and interactions within the existing social network. The second experimental design seeks to do precisely this by controlling and measuring interaction between subjects in a current network. The literature on interpersonal influence within a network posits two kinds of mechanisms for convergence in attitudes and actions. In the ﬁrst, friends and neighbors explicitly transfer information and ideas. This ﬁts the classic Norman Rockwell image of neighbors talking over the fence or meeting at the corner store. Politics, for instance, may be touched upon only briefly, but communication of ideas around politics is explicit. The second mechanism involves a diﬀusion of norms and expectations among peers. Friends and neighbors do not talk about politics as such, but subtle cues of approbation and disapproval are sensed by individuals. Eventually, residents internalize community norms without resorting to formal processes. The two mechanisms, both of which involve learning and conditioning, are mutually reinforcing. In Chapter 6 of this book, David Campbell uses observation data to examine the eﬀect of both mechanisms in creating and reinforcing norms on voter turnout.

194

Social capital: reaching out

When considering these two approaches, the ﬁrst is easier to tackle experimentally because its communication is direct and explicit.16 The researcher simply needs to force members of the network to comply with a protocol that dictates the content of conversations. This can be done by recruiting members to participate in the experiment. The model is similar to how Huckfeldt and Sprague (1995) ask subjects to provide the names of people with whom they discuss politics. Huckfeldt and Sprague (1995) asked subjects questions about past interactions with the friends listed (and then interviewed the friends in the same fashion). In the type of experiment described here, the volunteers are asked to initiate communication with a select set of friends. The list of friends is randomly divided into the various treatment groups and the volunteer is then instructed to carry out the protocol. The type of treatment can vary by topic (for example, donations to candidates, voting, volunteerism, trimming the lawn or simply holding an opinion) or by mode of delivery (for example, face-to-face conversations, phone calls or e-mail). The important thing is that the researcher – not the volunteer – determines who is contacted and how. This ensures the treatment and control groups do not diﬀer in a subtle way.17 An experiment controlling the flow of information through social networks was conducted during the 2002 Congressional election. Its goal was to determine the extent to which explicit appeals from friends and coworkers could increase voter turnout. Past ﬁeld experiments had demonstrated that a brief face-to-face conversation with a stranger could increase vote turnout by 8–10 percentage points (Gerber and Green, 2000; Green et al., 2003). Given reputation eﬀects and increases in social pressure, the hypothesis was that encouragement to vote from friends should be even more eﬀective. Laboratory experiments have found this to be the case across a wide range of settings (for example, Davis and Rusbult, 2001; Mackie et al., 1990; Brock, 1965) and political organizations have acted accordingly (for example, the voter recruitment program of the American Federation of Labor and Congress of Industrial Organization (ALF–CIO); MoveOn.org’s friendship networks; the Republican Party’s peer-to-peer mobilization in 2004). There is good reason to believe that organizations can harness social networks to change norms of voter turnout. An alternative hypothesis is that friends who are willing to volunteer time to encourage voting may operate in social circles where few people can be mobilized. The majority of the volunteer’s friends will already vote, so it will be impossible to motivate them to do more. Another potential issue is that serial abstainers residing in social networks where voting is the norm probably have a reason for not voting and will resist strongly. They may be ineligible or simply detest politics.

Experimental approaches to the diffusion of norms

195

An experiment was designed to test these hypotheses; volunteers were asked to list friends, neighbors and co-workers they thought were unlikely to vote in the Congressional election. These individuals were then randomly assigned to be either contacted by the volunteer about voting or left alone.18 After the election, the volunteers reported the names of the people they spoke with and oﬃcial voter ﬁles were checked to determine who had voted. Analyses were run to determine whether subjects assigned to the treatment or control group were more likely to vote. Volunteers were recruited from two sources. The ﬁrst source was a nonproﬁt organization named Vote for America that mobilized voters via peerto-peer networks. The organization had been active during the 2000 Presidential election in Rhode Island. It was eager to evaluate its model in an area where its volunteers were experienced, well trained and received robust organizational oversight and support. Thirty-one volunteers were recruited, who listed 481 friends to be mobilized. One-fourth of each volunteer’s list was assigned to the control group and the remaining threefourths were targeted for mobilization. The second sample of volunteers was recruited online through e-mail appeals. Ultimately, 65 volunteers successfully completed the protocol and lived in regions where voter databases were readily accessible. These 65 volunteers listed a total of 374 friends and neighbors; these were evenly divided into treatment and control conditions. Given the lack of training and support for Internet volunteers, one might expect the volunteers recruited over the Internet to be less eﬀective than Vote for America participants, so the two experiments are analyzed separately. Table 7.2 reports the results for both experiments. Neither experiment demonstrated any increase in voter turnout after accounting for slight imbalances between treatment and control groups. In fact, individuals who were approached by a friend regarding the election were on average 4.6 percentage points less likely to vote in the Internet experiment and 3.2 percentage points less likely to vote in the Vote for America experiment. While a demobilization eﬀect from peer-to-peer contact is extremely hard to believe, the experimental results strongly suggest that peer-to-peer mobilization is less eﬀective than traditional mobilization from a stranger. Given the consensus in the literature on the strength of personal ties, this null ﬁnding is surprising. Two explanations immediately suggest themselves. The ﬁrst is that selection into homogeneous social networks outweighs the importance of reputation and persuasion. Many people declined to participate in the experiment because ‘Everyone I know votes’. For such individuals, contacting members of their social circles will do little to increase turnout because everyone is planning to vote already.

196

Social capital: reaching out

Table 7.2 Effect of peer-to-peer mobilization Internet volunteers Model 1 Discussion () Prior voter history Member of major party Captain Dummies Constant N Adj-Rsq

0.048 (0.064)

Yes 0.412** (0.149) 374 0.13

Model 2 0.022 (0.051) Yes

Yes

Model 3 0.046 (0.051) Yes 0.190** (0.045) Yes

0.268* 0.219 (0.121) (0.118) 374 374 0.43 0.47

Vote for America (Rhode Island) Model 1

Model 2

0.030 (0.048)

0.038 0.032 (0.035) (0.033) Yes Yes

Yes

Yes

Model 3

0.275** (0.033) Yes

0.224 0.078 0.062 (0.205) (0.152) (0.143) 481 481 481 0.06 0.48 0.55

Notes: 1. Dependent variable is voter turnout in the 2002 general election. 2. Coefficients are derived from two-stage least squares using assignment as an instrument for contact. 3. Numbers in parentheses represent standard errors. 4. * implies p 0.05 and ** implies p 0.01.

Participating volunteers were instructed to target people whom they thought might not vote. It might be possible that individuals who operate in social circles where voting is the norm but who regularly abstain cannot be mobilized. A second explanation is that friends and neighbors actually are less eﬀective at mobilizing voters. Perhaps developed trust between two individuals has narrow bounds and requests outside of these boundaries have little or no influence. The knock on a door from a friend to ask for civic participation may have seemed artiﬁcial. Why bring up voting after many years? The appeal may have even been perceived as insulting. If this explanation is true, then it is possible that in this case norms are transferred not through direct conversations but through subtext and social expectations. While the study is obviously small, it is suggestive of how interpersonal influence may need to be indirect in some situations.

Experimental approaches to the diffusion of norms

197

EXOGENOUS SHOCKS TO A NETWORK It is not always possible to manipulate structured interactions within a network. An alternative strategy allows the researcher to focus on the influence of an external shock as it diﬀuses through the network. The exogenous shock design is versatile and can address a range of topics. The diﬀusion of information, the contagion of behaviors and the passage of opinions through influence are all designed and analyzed the same way. It can also deal with almost any topic and whether the subjects should be informed, persuaded, indoctrinated or something else. However, attention to detail is paramount to avoid selection bias. The key step is to establish baseline rates for the behavior or belief in question. In order to measure change, the researcher needs to know what the subject’s starting point is and then what the diﬀerence is.19 There are two primary methods for doing this. The more traditional one involves taking a measure of the network prior to the introduction of the exogenous shock (for example, a survey). The results obtained after the intervention are then compared to this baseline. While surveys are not randomized experiments, they can oﬀer a scientiﬁc mode of analysis in this setting. This is because the researcher is actively providing a treatment rather than passively measuring the eﬀect of past interactions. Other researchers can replicate the pattern of diﬀusion measured by the researcher at diﬀerent times and places, on diﬀerent topics. This model of analysis is akin to a physicist dropping a rock in a pool and attempting to model the waves.20 The second model uses a placebo as the reference group. Two parallel exogenous interventions are made and they each serve as the other’s point of reference. Both treatment and placebo need to meet two stringent requirements. First, the two interventions must not lead to the same outcome. If the treatment of interest is expected to improve political knowledge, then the placebo should be unrelated – even indirectly – to politics. In addition, the two treatments must travel through the network at roughly the same rate. A placebo that does not travel far from the point of its introduction would be a poor comparison for a treatment that spreads through a network quickly. Selecting two treatments that meet both requirements can be diﬃcult. The extent to which a design meets these two criteria is an empirical matter and can be diﬃcult to gauge prior to conducting the study.21 However, thinking through the realities can often give experimenters a sense of major diﬀerences. For instance, encouraging families to recycle paper is not a very good placebo for a treatment encouraging families to recycle aluminum because the two behaviors are likely linked. A rumor concerning

198

Social capital: reaching out

police brutality is likely to spread through a neighborhood much more quickly than a piece of information on the mayor’s electoral platform. An experiment using the exogenous shock design was conducted during the 2002 Congressional Primaries in Denver and Minneapolis. The social network considered was very small – households with two registered voters – and the behavior traced was voter turnout. Various two-voter households in Denver and Minneapolis were randomly selected to receive face-to-face encouragement on either of two subjects: recycling or voting.22 The mobilization eﬀect of face-to-face canvassing is well established (see Gerber and Green, 2000), so it makes an excellent exogenous shock. The hypothesis is that people who answer the door and receive the Get Out The Vote (GOTV) message will turn out at higher rates than people who answer the door and receive the recycling message. But what happens to the person in the household who did not answer the door? How much of the boost in turnout from canvassing is passed on to that other person? The design is notable in a few respects. First, nothing in the collection of data relies upon subject responses. The paid canvasser recorded who answered the door (if anyone), and the county clerk recorded who voted. Any measurement error is therefore rare random errors in transcription. Second, the placebo-controlled design oﬀers a perfectly comparable reference group. Persons who answered the door are compared only with other persons who answered the door. Similarly, influence is examined only among persons whose cohabitants answered the door. The only diﬀerence between the groups is the treatment provided.23 Finally, two-voter households are convenient to work with for the experiment, but this analysis would be impossible with non-experimental data. Researchers often ignore selection with regards to city or neighborhood, but there is no avoiding the selection of spouses and roommates. The selection of cohabitants is not an issue in the slightest since the randomization of the treatment occurs only within the set of people who have selected each other. The experiment considers a unique and small social network, but it measures voter contagion very precisely. When pooled together, the two experiments showed a 9.8 percentage point boost in turnout from being exposed to the GOTV message. In Denver, the campaign contacted 33 percent of the households in both the GOTV and recycling conditions for a total of 563 households receiving either the treatment or the placebo. Among the people who answered the door, those receiving the voting appeal were 8.6 percentage points more likely to vote than people answering the door and hearing the recycling message. In Minneapolis, 45 percent (394) of the target households were successfully contacted. The direct treatment eﬀect was a robust 10.9 percentage points.

Experimental approaches to the diffusion of norms

199

Table 7.3 Contagion effect of voting within the household

Contagion ()

Model 1

Model 2

Model 3

0.614** (0.233)

0.564* (0.220) 0.165** (0.040) 0.041 (0.034) 0.134* (0.053) 0.128** (0.045) 0.001 (0.030)

0.587** (0.217) 0.155** (0.037) 0.041 (0.035) 0.119* (0.049) 0.120** (0.042) 0.014 (0.028) 0.002 (0.004) 0.000 (0.000) 0.020 (0.027) 0.019 (0.046) 0.087 (0.101) 955 0.55

Voted 2001 Voted 2000 General Voted 2000 Primary Voted 1999 Voted 1998 General Age Age-squared Female Denver Constant N Adjusted R sq

0.059 (0.055) 0.072 (0.054) 955 0.45

0.044 (0.046) 0.113** (0.033) 955 0.56

Notes: 1. Dependent variable = voter turnout in 2002 Primary of person in household who did not answer the door. 2. Endogenous variable = voter turnout in 2002 Primary of person in household who did answer the door. 3. Instrument = assignment to treatment or placebo condition. 4. Voter history is for person who did not answer the door (same as the dependent variable). 5. Numbers in parentheses represent standard error s. 6. * implies p 0.05 and ** implies p 0.01.

Surprisingly, over 60 percent of this boost in turnout was passed on to the other person in the household. Persons not answering the door in Denver households exposed to the voting appeal were 5.5 percentage points more likely to vote and in Minneapolis the boost in turnout was 6.4 percentage points over households assigned to the recycling message. Pooled together, we ﬁnd registered voters in the household were 6 percentage

200

Social capital: reaching out

points more likely to vote even though they were not directly exposed to the voting message. Table 7.3 uses two-stage least squares to pool the results and control for past voter history and demographic information.24 The magnitude of the eﬀect is substantial and indicates that voting is a highly contagious behavior. A person who may have been 20 percent likely to vote in the Primary becomes 80 percent likely to turn out should their housemate vote. A model where the spouse’s or housemate’s turnout is the only variable explains over 45 percent of the variance in a person’s decision to vote.25 These experiments suggest that the behavior of spouses and housemates has an extremely large impact on the decision to vote.

CONCLUSION Similarities between pairs of people can be the result of a number of factors: interpersonal influence, similar pre-existing dispositions, structural incentives, a selection process that weeds out dissimilar individuals, exposure to identical exogenous factors, or social linkages. Unfortunately, ‘The observational data that would be necessary to disentangle each of these component parts fully and directly are unavailable, and very nearly unimaginable’ (Huckfeldt and Sprague, 1995, p. 164). Because of this, scholars who use observational data to try and estimate the eﬀect of interpersonal influence, , are often required to make unveriﬁable assumptions. In contrast, randomized ﬁeld experiments require minimal assumptions about the process that generated the data or the relationship between the variables. Part of the reason that experiments have not been used more in social science is that the questions asked are necessarily focused and therefore ‘small’. The required control over the factor of interest limits the number of questions a researcher can answer in a given study and increases the diﬃculty in conducting research. This means studies on social capital, networks and influence will need to be tightly focused, often dealing with one of the narrower slices or deﬁnition of social capital. The virtue of this approach is that experiments allow small questions to be answered accurately. The researcher is forced to use the incremental notion of intellectual progress found in the natural sciences. The three experiments described here illustrate how one might apply experiments to social networks. The ﬁndings are interesting and informative, but readers anticipating the ﬁnal word on interpersonal or influence will be sorely disappointed. That said, the three experiments contained in this chapter paint a picture of a subtle and slow process of norm diﬀusion. While voter turnout is contagious within a household, explicit appeals from members of a social network to vote are wholly ineﬀective. For some reason, the eﬀort to

Experimental approaches to the diffusion of norms

201

harness the power of the social network to encourage the norm of voting backﬁred. Similar limits were observed with regards to opinion contagion among ﬁrst-year college students. Contagion was observed only on a select few topics and behaviors, while individuals remained unmoved on the vast majority of questions. Moreover, the attitudinal contagion was highly constrained by pre-existing beliefs. These results run counter to many observational studies where readers are led to believe that we live in a world where citizens are constantly buﬀeted by the attitudes and actions of neighbors. Every person is a party switcher who just happened to move into a particular neighborhood in a particular suburb. The truth is that the diﬀusion and formation of community norms is a slow process. A detailed and incremental experimental approach can unravel the bigger questions of just how social capital is formed under conditions of social closure in a community. It is my hope that the modest examples included in this chapter will spur other scholars to pursue more ambitious experiments in the ﬁeld.

NOTES 1. 2.

Each experiment is presented at length in Nickerson (2005). V.O. Key (1949) oﬀers one of the earliest and most famous examples of context eﬀects in voting. 3. Newcomb and Svehla established this using modern survey techniques in 1937 (Newcomb and Svehla, 1937), but the observation dates back at least as far as Plato. Jennings and Niemi (1974) provide a thorough analysis of this topic. 4. Application of the treatment can be split as many ways as the researcher desires. Medical researchers often experiment with diﬀerent doses of the same drug. Lotteries like Powerball could be thought of as income experiments with the range of winnings constituting the diﬀerent treatment groups (and the losers serving as the control group). Nor does the treatment need to involve a single factor. Factorial design can easily accommodate interactions between multiple treatments. For instance, suppose one wanted to know whether citizens are mobilized by leaflets, a knock on the door, or a combination of both. To test this, registered voters in a neighborhood could be randomly assigned to receive only a leaflet, only a knock on the door, both the leaflet and the knock on the door, or no contact at all from the campaign. The voter turnout rates between the four groups could then be compared to determine which technology is more cost eﬀective in increasing turnout. This type of protocol could be extended to any number of treatments. Sample size is the only limitation upon the use of the technique. 5. Experimental analysis is not devoid of assumptions. However, the analysis above reduces the number of assumptions required in the analysis. No assumptions were required about the process that generated the data, the distributions of the variables or the relationship between the variables. Other assumptions were still necessary. For instance, equation (7.2) could not be simpliﬁed without E(CT CC) 0 and E(T C) 0. That is, it is presupposed that the treatment and control groups have the same intercepts and residual errors on average. However, these two assumptions rest upon the Law of Large Numbers and executing the protocol successfully. It should also be pointed out that observational studies make the same assumptions, so experimental interventions are not

202

6.

7.

8.

9. 10.

11. 12.

13. 14. 15. 16. 17.

18.

Social capital: reaching out adding to the presuppositions of the model. The uncertainty associated with the experimental estimate of is due to sampling error, not model misspeciﬁcation. The simplicity of the unbiased estimator of in equation (7.3) does not imply that more complicated econometric estimators cannot be used. In fact, modeling assumptions are better satisﬁed by experimental rather than observational data. Where observational data require covariates, multiple stages and modeling assumptions to claim E [/AI] 0, the condition holds by construction of the variable AI in experiments. Adding covariates to the analysis reduces variance and thereby shrinks standard errors, but covariates are unnecessary to achieve unbiased estimates of because omitted variable bias is not a concern. These three protocols are not mutually exclusive; in principle, a researcher could design an experiment where communication about a controlled external event takes place within a randomly constructed social network. However, given the diﬃculty in executing each experiment on its own, combining the diﬀerent types may be overkill. This set-up diﬀers from lab experiments in that the subjects are living within the social context as opposed to operating in it for brief periods of time. The interactions are not artiﬁcial by deﬁnition because the constructed framework really is the social network in which the subject resides. To be more precise, the null hypothesis will not be rejected 95 percent of the time. It is perfectly acceptable for the randomization to occur within strata. There may be important policy concerns that trump the researcher’s desire for a purely random sample: the severity of the crime and the predicted risk of future violence help to determine the assignment of prisoners to wards; rooms (and sometimes floors) of dorms are gender segregated; and the military provides diﬀerent training and assignment to specialized troops such as intelligence or communications. While such stratiﬁcation needs to be accounted for and limits the conclusions to be drawn, it in no way invalidates the experiments as a whole. The researcher simply needs to examine the context eﬀect within the strata created rather than in the system as a whole. The net eﬀect is conducting several concurrent experiments on slightly diﬀerent populations. It is possible that men and women diﬀer with regards to a behavior or a belief on average. As a result, it is important to control for gender when studying roommate pairs. Following the lead of Sacerdote (2001), when a student has two roommates, the views of both roommates will be averaged together to calculate A1Ri. Similarly, A1si represents the average views of person i’s suitemates. Averaging loses information but in no way biases the analysis. No meaningful correlation was found between pre-existing attitudes among either roommates or suitemates. The Q-statistic testing the homogeneity of treatment eﬀect across questions is ﬁrmly rejected. This ﬁnding also validates the survey instrument, suggesting that the responses are measures of genuine opinion and not an artifact of taking the survey itself. Experiments testing the gradual diﬀusion of norms are better oﬀ examining constructed social networks. One problem that may arise is treatment contagion where members of the control group interact with members of the treatment group and are exposed to some portion of the treatment. The problem of contagion can be bypassed in two ways. First, the researcher can simply select a set of subjects who are unlikely to infect one another. The distance need not be spatial; it could be along any dimension, such as social or temporal. The second solution is to vary randomly the density of the treatment (that is, the subjects in one set receive a high dosage of the treatment while subjects in another set receive a low dosage). If the treatment is contagious, then the estimated treatment eﬀects will be smaller for populations with a high percentage of subjects assigned to the treatment group than for populations with a low percentage of subjects provided the treatment. While this will not eliminate contamination of the control group by the treatment group, it does provide a means of detecting the source of error. Randomization tests indicated that the treatment and control were extremely similar with regards to observable characteristics. The only noticeable diﬀerence is that the

Experimental approaches to the diffusion of norms

19. 20.

21.

22. 23. 24. 25.

203

control group for the Internet volunteer sample was slightly less likely to be registered than the treatment group. However, this diﬀerence did not cross traditional thresholds of statistical signiﬁcance and there is no reason to assume the randomization failed in any sense. This point may sound pedantic and simplistic, but it is often ignored. A fabulous example of this protocol was provided by Dodd (1952). One in ﬁve people in a town of 950 people were told a six-word slogan for a coﬀee company and were told that anyone able to repeat the slogan in two days would receive a free pound of coﬀee. The next day a plane dropped 30 000 leaflets over the town. The leaflets stated that one in ﬁve people in town knew a six-word slogan and anyone able to repeat the slogan would receive a free pound of coﬀee. The day after the massive literature drop, researchers knocked on every door in the town to see who knew the slogan (and to hand out the coﬀee). Dodd conducted similar experiments in other cities and using diﬀerent modes of mass communication. There is no reason why the two protocol designs cannot be combined. The panel framework has the advantage of mapping the network prior to the intervention so the system is more contained and easier to trace. The placebo design has the advantage of controlling for all temporal eﬀects since both interventions take place at the same time. The strengths of each design cover the weaknesses of the other, so the dual design is optimal. The added complexity makes the execution of the protocol more diﬃcult, but that is the only downside. Households were also assigned to a control group that received no visit from the campaign. The purpose of the control group was simply to verify that the mobilization eﬀect detected was genuine. The similarity of the treatment and placebo groups on observed characteristics is striking. Every randomization check supports this conclusion. A bivariate probit model accounting for the dichotomous nature of voter turnout yields substantively identical results. The predictive accuracy is all the more astounding when the dichotomous nature of voting is taken into account.

REFERENCES Allport, Gordon W. and Leo J. Postman (1940), The Psychology of Rumor, New York: Henry Holt. Brock, Timothy C. (1965), ‘Communicator–recipient similarity and decision change’, Journal of Personality and Experimental Social Psychology, 1, 650–54. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Capital, New York: Oxford University Press. Coleman, James (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94 (supplement), 95–120. Davis, Jody L. and Caryl E. Rusbult (2001), ‘Attitude alignment in close relationships’, Journal of Personality and Social Psychology, 81, 65–84. Dodd, Stuart C. (1952), ‘Testing message diﬀusion from person to person’, Public Opinion Quarterly, 16, 247–62. Duncan, Greg J, Johanne Boisjoly, Dan M. Levy, Michael Kremer and Jacque Eccles (2003), ‘Empathy or antipathy? The consequences of racially and socially diverse peers on attitudes and behaviors’, working paper, Harvard University. Gerber, Alan S. and Donald P. Green (2000), ‘The eﬀects of canvassing, telephone calls, and direct mail on voter turnout: a ﬁeld experiment’, American Political Science Review, 94, 653–63.

204

Social capital: reaching out

Green, Donald P., Alan S. Gerber and David W. Nickerson (2003), ‘Getting out the vote in local elections: results from six door-to-door canvassing experiments’, Journal of Politics, 65, 1083–96. Hoel, Jessica, Jeﬀrey Parker and Jon Rivenburg (2004), ‘Peer eﬀects: do ﬁrst-year classmates, roommates, and dormmates aﬀect students’ academic success?’, paper presented at the Higher Education Data Sharing Consortium Winter Conference, 14 January 2005, retrieved from http://www.reed.edu/~hoelj/ 0HEDS.pdf. Huckfeldt, Robert and John Sprague (1995), Citizens, Politics, and Social Communication: Information and Influence in an Election Campaign, New York: Cambridge University Press. Iyengar, Shanto and Donald R. Kinder (1987), News that Matters, Chicago, IL: University of Chicago Press. Jennings, M. Kent and Richard G. Niemi (1974), The Political Character of Adolescence: The Influence of Families and Schools, Princeton, NJ: Princeton University Press. Katz, Lawrence F., Jeﬀrey R. Kling and Jeﬀrey B. Liebman (2001), ‘Moving to opportunity in Boston: early results of a randomized housing mobility study’, Quarterly Journal of Economics, 116, 607–54. Key, Vladimer Orlando (1949), Southern Politics in State and Nation, New York: Alfred A. Knopf. Mackie, Diane M., Leila T. Worth and Arlene G. Asuncion (1990), ‘Processing of persuasive in-group messages’, Journal of Personality and Social Psychology, 58, 812–22. Marmaros, David and Bruce Sacerdote (2002), ‘Peer and social networks in job search’, European Economic Review, 46, 870–79. McEwan, Patrick J. and Kristin A. Soderberg (2004), ‘Peer eﬀects at a single-sex college’, unpublished manuscript, Wellesley College. Newcomb, Theodore M. and G. Svehla (1937), ‘Intra-family relationships in attitude’, Sociometry, 1, 180–205. Nickerson, David W. (2005), ‘Measuring interpersonal influence’, doctoral dissertation, Yale University. Oliver, J. Eric (2001), Democracy in Suburbia, Princeton, NJ: Princeton University Press. Sacerdote, Bruce I. (2001), ‘Peer eﬀects with random assignment’, Quarterly Journal of Economics, 116, 681–704. Sacerdote, Bruce (2003), ‘Peer eﬀects with random assignment: results from Dartmouth roommates’, Quarterly Journal of Economics, 166 (2), 681–704. Stinebrickner, Todd R. and Ralph Stinebrickner (2001), ‘Peer eﬀects among students from disadvantaged backgrounds’, working paper, University of Western Ontario. Wilson, John and Marc Musick (1997), ‘Who cares? Toward an integrated theory of volunteer work’, American Sociological Review, 62, 694–713. Winston, Gordon C. and David J. Zimmerman (2003), ‘Peer eﬀects in higher education’, Working Paper no. 9501, Cambridge, MA: National Bureau of Economic Research, http://www.nber.org/papers/w9501.

8.

Capitalizing on connections: social capital and strategic management Janine Nahapiet1

Social connections and relationships are a major influence on organizational success. They always have been. However, they are increasingly prominent in current thinking about strategy and organization. Indeed, scholars now argue that we should view the ﬁrm as a portfolio of relationships rather than a portfolio of either businesses or capabilities. This view has profound implications for strategic management. Why are we seeing this renewed emphasis on social connections? What theories help us understand and take action if strategy is viewed from a relational perspective? Fundamental changes at the end of the twentieth century radically challenged conventional wisdom regarding competitive success. Two issues emerged: the crucial importance of the knowledge economy and the idea of collaborative advantage. Both implied the need for a new model of strategy – one built around value creation through social relationships. This chapter’s core argument is that social capital provides a distinctive perspective that can address directly these important challenges facing strategic management. Social capital theory focuses on the value of social connections; scholars argue that ﬁrms well endowed with social capital create competitive advantage through better access to opportunities, options and resources through their relationships. A growing body of evidence supports this view. The distinctive contribution of social capital lies in elucidating the themes that are emerging as central to strategic management. Recent scholarship shows how social capital helps to explain the mechanisms involved in creating and exploiting collaborative advantage and intellectual capital, both particularly critical to innovation and the knowledge economy. Social capital also lends valuable insights into strategy under uncertainty. Finally, ﬁrms need to understand how to organize and manage networks and relationships, if they are to capitalize on their connections.

205

206

Social capital: reaching out

STRATEGIC MANAGEMENT IN THE TWENTYFIRST CENTURY: CHANGE AND CHALLENGE The strategist’s core questions concern the purposes, direction, choices, changes, governance and performance of organizations in multiple contexts: industry, market, social, economic and political (Pettigrew et al., 2002). As these contexts change so too do the solutions sought to these questions. A series of developments at the close of the twentieth century changed our social institutions in signiﬁcant and unmistakable ways. As a result, both scholars and practitioners are reconsidering their core assumptions regarding economic organization, competitive success, even what language, concepts and approaches are central to eﬀective management (DiMaggio, 2001; Walsh et al., 2006; Nahapiet, 2008b). Prahalad and Hamel (1994) provide a comprehensive overview of the major forces that drove change in the competitive environments of the US, Europe and Japan in the 1990s. These drivers range from deregulation to environmental concerns to changing customer expectations to technological discontinuities. One must now include subsequent developments such as post-9/11 political uncertainties, changing demographics and turbulence in ﬁnancial markets. And thus, the challenges for strategy development in these highly uncertain times are great and made increasingly complicated by multiple new trends. To begin with, that information technology, combined with deregulation and globalization, is creating much stronger connections across increasingly permeable boundaries leads scholars to think more about the networked society (Castells, 2000). The major diﬃculties that ﬁnancial institutions face as they try to manage the consequences of the US subprime mortgage crisis are just a single example – although a vivid one. The assumptions that underlie strategy development are also shifting. Companies now view intangible assets as the foundation for success. Although early thinking in business strategy focused on physical assets, researchers in the 1980s started emphasizing the role of less tangible resources in fostering competitive advantage, especially knowledge and capabilities. Attention then turned from economies of scale and scope to ideas such as intelligent enterprise (Quinn, 1992), core competence (Prahalad and Hamel, 1994) and dynamic capabilities (Teece et al., 1997) – perspectives brought together in the knowledge-based view of the ﬁrm (Eisenhardt and Santos, 2002). Further, organizations are emphasizing value creation more. Both scholars and practitioners argue that the key to sustained superior performance lies more in the ability to create value than in any mechanism to appropriate it (Ghoshal and Moran, 1996; Moran and Ghoshal, 1999). Accordingly, innovation becomes an increasingly critical element of strategy.

Capitalizing on connections

207

In pursuit of both eﬃciency and innovation, strategies are increasingly built around assets accessed, not owned, by the ﬁrm. With the major expansion of outsourcing, the growing importance of alliances, including across borders, and the emergence of open-sourced and networked innovation, has come the recognition that ownership is not necessary to control or beneﬁt from assets. This represents a radical departure for scholars and executives. Venktraman and Subramaniam (2002) suggest that strategic theory has now reached a third era in which the ﬁrm is no longer seen as a portfolio of businesses or as a portfolio of capabilities, but rather as a portfolio of relationships. Once again, this change puts connectedness, interdependence and collaborative advantage centre stage. All of these changes raise fundamental questions for strategy. For example, if value creation increasingly occurs between organizations, what is now the appropriate unit of analysis for understanding and developing strategy? How are we to understand the relationship between strategy and organization in a world in which intangible assets underpin success? What provides the sources of strategic advantage in conditions of uncertainty? Finally, if the ability to identify and manage eﬀective relationships represents the critical driver of strategy and competitive advantage, what theories help illuminate and guide the development of eﬀective relationships, both within and outside organizations? To answer these questions, scholars have developed a range of relational theories, that is, theories that focus on the connections and patterns of exchange between actors, the dynamics of cooperation and competition, and the routines for creating and leveraging relational assets. Three constructs in particular have dominated this work: embeddedness (Granovetter, 1985), networks (Kilduﬀ and Tsai, 2003) and communities (Heckscher and Adler, 2006). This chapter argues that social capital oﬀers a distinctive fourth perspective that both draws together important aspects of the other perspectives and oﬀers deeper insights into the dynamics of value creation through social connections.

SOCIAL CAPITAL AND STRATEGIC MANAGEMENT: SETTING THE AGENDA The core proposition of social capital theory is that social ties constitute a valuable resource for the conduct of social aﬀairs, enabling individuals and social groupings to achieve outcomes that they could not otherwise achieve, or could only do so at extra cost (Coleman, 1988; Burt, 1992; Putnam, 1993a, 1995). From a theoretic perspective, social capital is an especially powerful lens for addressing current strategy questions. As a relational

208

Social capital: reaching out

theory, it takes as its prime unit of analysis the connections between actors. Further, it can be applied across diﬀerent levels of analysis, from an individual person to a group, organization, community, region or even nation, thereby enabling study of the impact and interrelationships of these diﬀerent levels. It therefore provides a valuable way to characterize an organization’s complete set of relationships, including those that cross institutional boundaries. Social capital theory can also be applied to a wide range of organizations – public, private and voluntary, large and small, formal and informal, established and emergent – and this helps advance understanding of strategy across a variety of settings. No universally accepted deﬁnition of social capital yet exists. Moreover, as Farr (2004) notes, writers sometimes discuss the core idea of social capital without necessarily using the term itself; de Tocqueville is a prominent example. Still others use the term without the construct meaning we now attach to the term. Given this reality, this chapter follows Schuller et al. (2000) and includes work that addresses key aspects of the concept even though it may not use the term. The deﬁnition of social capital that informs this chapter and has been influential in strategic management studies is that of Nahapiet and Ghoshal (1998). We deﬁne social capital as: ‘the sum of the actual and potential resources embedded within, available through and derived from the network of relationships possessed by an individual or social unit. Social capital comprises both the network and the assets that may be mobilized through that network’ (1998, p. 243). Three attributes of social capital enable its distinctive contribution to the management ﬁeld: its resource-based perspective, its ability to combine multiple dimensions of relationships and its focus on performance outcomes (Nahapiet, 2008a). The ﬁrst attribute is based on social capital’s view of connections both as a resource and as providing access to resources – particularly information, influence, social credentials and solidarity. As Woolcock and Narayan (2000a) observe, intuitively, one’s family, friends and associates constitute an important asset, one that can be enjoyed or called on in a crisis, and/or leveraged for material gain. They argue that what is true for individuals also holds for groups and communities to which we can add organizations. Treating social relations as assets is directly relevant to the resourcebased view of the ﬁrm adopted by strategic management research since the 1990s. According to Barney (1991) a ﬁrm’s competitive advantage lies in its unique conﬁguration of resources – physical, human and organizational – that are valuable, rare, durable, imperfectly imitable and non-tradable (Barney, 1991; Dierickx and Cool, 1989). A resource is non-imitable if it possesses tacitness, causal ambiguity, time compression diseconomies and

Capitalizing on connections

209

interconnectedness, as well as path-dependence and social complexity (Reed and De Fillippi, 1990). Social capital possesses these characteristics as well – hence its potential as a strategic resource. Social capital also draws together multiple dimensions of relationships often kept separate in other studies. Nahapiet and Ghoshal (1998) propose that three dimensions of social capital – the structural, the relational and the cognitive – are all important in understanding social relations. Our analysis is part of a growing body of work that seeks to understand the interplay among the diﬀerent aspects of social capital. These approaches are broader than those of either network scholars that focus primarily on the structure of ties and their consequence (Baker, 1990; Burt, 1992, 2005) or those researchers who concentrate on the relational dimensions, especially trust and norms (for example Putnam, 1993a, 1995; Fukuyama, 1995, 1999). Because social capital research emphasizes the performance outcomes of social connections, it ﬁts naturally with strategic management. For example, in their research on biotechnology start-ups, Maurer and Ebers (2006) demonstrate not only the importance of all three dimensions of social capital, but also how the ability to reconﬁgure these dimensions over time aﬀects start-up performance. Beyond these three features, research on social capital addresses directly many of the central strategic questions of our era. First, social capital theory is widely regarded as providing an important theory of cooperation and thus oﬀering valuable insights into collaborative advantage. The theory also advances our understanding of strategic knowledge processes both within the ﬁrm and more generally. Though not yet prominent, I believe social capital perspectives can also inform our understanding of strategymaking under uncertainty by providing insights into the ways in which social networks and relationships provide both flexibility and strategic options. Finally, as research turns increasingly to the processes that build, sustain and leverage social capital, its insights shed light on an area of growing interest – the links between strategizing and organizing.

SOCIAL CAPITAL AND COLLABORATIVE ADVANTAGE An enduring, central theme in strategic management is the search for sources of sustainable competitive advantage. Clearly ﬁrms have always had important connections with other organizations, especially customers and suppliers. Previously, the dominant assumption guiding strategy has been that ﬁrms should maximize their individual advantage through

210

Social capital: reaching out

minimizing their dependence on external parties, particularly for strategically important activities and resources (Porter, 1980, 1998). That assumption is now changing as ﬁrms recognize that many of the skills and resources essential to their future success lie outside the ﬁrm’s boundaries and thus outside their direct control. Yet the idea of strategy as ‘a control agenda’ remains strong (Galaskiewicz and Zaheer, 1999). The challenge now is how to achieve control in the context of extended enterprise and networked relationships. Many years ago, Mary Parker Follett distinguished between ‘power-over’ and ‘power-with’; she proposed that managers should adopt the idea of power as co-active and jointly developed, resulting in what she termed ‘interactive influence’ and ‘integrated control’ (Follett, 1918, 1998). It is this broad idea of joint endeavour, mutual influence and collective action between organizations that underpins current interest in what authors have variously called the ‘collaborative’ or ‘cooperative’ advantage. Eﬀective collaboration rests on the ﬁrm’s ability to identify, make, sustain and beneﬁt from connections and relationships productively and eﬃciently and to create joint value through these connections. Collaborative advantage derives from its ability to do this especially well both internally and externally and in ways that other ﬁrms cannot eﬃciently replicate (Nahapiet et al., 2005). Because social capital research focuses on patterns of association, exchange and cooperation, it oﬀers immediate insights into this area. For instance, the World Bank views social capital as: ‘the norms and social relations embedded in social structures that enable people to coordinate action to achieve desired goals’. Putnam (1995) oﬀers a similar deﬁnition. For him, social capital refers to: ‘features of social organizations such as networks, norms and social trust that facilitate coordination and cooperation for mutual beneﬁt’. Social capital research contributes to our understanding of collaborative advantage in three important ways: it demonstrates empirically the strategic signiﬁcance of social connections, it explains various factors that produce high or low levels of social capital, and it elucidates the core constructs underpinning both social connections and collaborative advantage: embeddedness, reciprocity, appropriability, latency and convertibility. In a recent review of research on the links between social capital and interorganizational relationships, I show that social capital plays an important part in the formation, evolution and performance of these relationships (Nahapiet, 2008a). For example, the evidence indicates that ﬁrms are more likely to form relationships with those organizations with which they already have a connection (Walker et al., 1997; Ahuja, 2000a). This ﬁnding reflects the increased trust and openness that frequently comes with experience but it may also reflect the greater opportunities that connections may

Capitalizing on connections

211

produce. Uzzi’s research on the New York City apparel industry has shown that arm’s-length relationships and embedded ties are not only distinct forms of interorganizational exchange, but that the former can produce competitive advantages that are diﬃcult for the latter to emulate (Uzzi, 1996, 1997). Embedded ties set expectations for trust and reciprocity which then facilitate pooled resources and cooperation. As a result, embedded ties increase the economic eﬀectiveness and competitiveness of ﬁrms along several crucial dimensions – organizational learning, risk-sharing and speed to market. Over recent years, several organizations have provided instructive examples of the strategic advantages that can come through collaboration, including Benetton, Toyota, IDEO and more recently Li & Fung. In addition, research shows that relation-speciﬁc assets can aﬀect not only the performance of individual ﬁrms but also the evolution of entire industries over time. For example, Powell et al. (1996) found that the locus of innovation in the biotechnology industry is the network – not the individual ﬁrm. Those ﬁrms unable to position themselves in these learning networks are at a competitive disadvantage. Gomes-Casseres (1996) similarly emphasizes the increasing importance of competition between constellations and networks – rather than individual ﬁrms – in several industries (for example, biotechnology, microprocessors and aircraft). Further, research provides important insights into the factors likely to result in relatively high or low levels of social capital. Nahapiet and Ghoshal (1998) identify four factors conducive to social capital development: time, interdependence, interaction and closure. Adler and Kwon (2002) explore the importance of wider features of the social structure and their implications for three proximate sources of social capital – opportunity, motivation and ability. Other scholars have identiﬁed additional influences that range from the role of institutional factors aﬀecting collaboration in biomedicine (Swan et al., 2007) to the role of personal attributes (Oh et al., 1999). But perhaps the most important way that social capital research helps explain collaborative advantage is by identifying the ﬁve critical constructs that underpin cooperation, collaboration and collective action. These constructs explain why and how the connections between social actors facilitate collective action. Two of these illuminate aspects of exchange: embeddedness and reciprocity. The other three deal with the potentiality of connections: appropriability, latency and convertibility. In practice, they can be viewed as variables – with much research focused on how each of them operates. For example, when is reciprocity the dominant norm governing exchange? In what contexts are social ties appropriable and by whom?

212

Social capital: reaching out

Embeddedness is a fundamental concept of social capital research and highlights the importance of social dimensions in business exchange. Polanyi (1944) articulated its central idea: that economic activity is best understood in the context of ongoing patterns of social relations that influence behaviour and outcomes. Granovetter (1985) distinguishes between two important aspects of these social relations – the concrete personal relations between speciﬁc people and the networks of such relations – and analysed their consequences on the creation of trust in economic life. Uzzi (1996) suggests that the structure and quality of social ties among ﬁrms shape economic action by creating both unique opportunities and access to those opportunities (see also Ahuja, 2000b). Hagedoorn (2006) built on these ideas by showing how diﬀerent types and levels of embeddedness aﬀect the formation of interﬁrm partnerships. Embeddedness highlights the social dimension of business, while social exchange theory provides its core theoretical foundations. Bourdieu (1985) observes that social capital resides in relationships and that relationships are created through exchange. While there are many variants of exchange theory, Biggart and Delbridge (2004) identify two systems of exchange in which the particular nature of the parties’ relation to each other is taken into account when parties conduct economic exchange. They describe these as ‘associative’ and ‘communal’ systems and note speciﬁcally that it is in these systems that social capital can become important as a way of facilitating exchange. If economic exchange is embedded in social relations, then a central question is: what are the norms that govern exchange? Reciprocity is the dominant exchange principle in much social capital research – the expectation that exchange will be mutual. Reciprocity is closely linked in the literature to trust. Coleman observes that: ‘[I]f A does something for B and trusts B to reciprocate in the future, this establishes an expectation in A and an obligation on the part of B. This obligation can be conceived as a credit slip held for A for performance by B . . . These credit slips constitute a large body of credit that A can call in if necessary’ (1988, p. S102). Biggart and Delbridge (2004) suggest that reciprocity is critical to both associative and communal systems. In associative systems such as strategic alliances and vertical and horizontal networks, reciprocity provides the foundation for mutual expectations. In communal systems of exchange – in which actors share identity in a community or have some basis for a shared bond – group norms such as reciprocity are even more important. Here, social capital scholars have drawn particular attention to the importance of social obligations that come with membership of a group but also the ‘credentials’ or collectively owned capital available to members of that group (Bourdieu, 1985, p. 249). Examples in the business world include those organizations

Capitalizing on connections

213

identiﬁed as ‘clans’, such as some professional service ﬁrms and their alumni associations. However, as Biggart and Delbridge note, while these two systems are analytically distinct, in practice they are often combined, providing powerful foundations for cooperative and collaborative activity. Putnam (1993) describes an even stronger form of reciprocity in social networks and communities – generalized reciprocity. This is the principle that operates when a person does something of value for you without expecting anything immediately in return and perhaps without even knowing you, conﬁdent that down the road you or someone else will return the favour. He argues that a society that relies on generalized reciprocity is more eﬃcient, more trustful and more open to collaboration. Baker (2000) sees reciprocity as a powerful principle underpinning social exchange across a wide range of contexts, from international relations to the help oﬀered by individuals to each other within an organization. This emphasis on reciprocity links to the idea of social capital as ‘goodwill’ and it helps move exchange from a world of primarily short-term, individualistic and instrumental actions to a more social and enduring basis for exchange relations. It thus reflects a model of social relations very diﬀerent from the selfinterest theories that some observers have ascribed as foundational to work on social capital (Wasserman et al., 2006). Several commentators view appropriability as the central principle for social capital theory (Adler and Kwon, 2002). Widely regarded as a recurring feature of social life, appropriability expresses the idea that social connections of one type often can be transferred from one situation to another, that is, they can be used for diﬀerent purposes and in diﬀerent networks. Putnam (1993a) similarly notes that social capital typically consists in ties, norms and trust transferable from one social setting to another. Examples include the transfer of trust from family and religious aﬃliations to work situations (Fukuyama, 1995), the development of personal relationships into business exchanges (Coleman, 1990; Uzzi, 1996, 1997) and the leveraging of individual contacts for organizational purposes (Bouty, 2000). In practice, appropriability may be more nuanced and problematic. For example, in her research on informal resource exchanges between research and development (R&D) researchers across organizational boundaries, Bouty (2000) distinguishes between potential exchange situations and possible ones – the latter reflecting scientists’ judgements about the strategic and social context in which exchange relations are embedded. Thus: ‘the better a partner is known and the less direct a competitor he or she is, the less diﬀerence exists between possible and available resources – between what the scientist can do and is ready to do’ (Bouty, 2000, p. 55). Others have also focused on the links between individual and organizational appropriability – with research showing that in some contexts what is

214

Social capital: reaching out

appropriated may be used to maximize personal rather than organizational interests (Newell et al., 2004). Coleman notes that social capital in one context may be useless or even harmful in others, yet the signiﬁcance of appropriability is that it suggests the relatively boundless opportunities that may be available through social ties, and hence their potential value as a strategic resource. This idea of ‘potentiality’ leads to another important though infrequently discussed feature of social capital – the idea that much of the value in social relationships is latent. Although people and organizations have connections of many diﬀerent types, the degree to which they can and will be activated and for what purpose is often unknown. The Chambers Dictionary deﬁnes ‘latent’ as hidden, not visible, dormant, undeveloped but capable of development. This implies that many of the possibilities oﬀered by connections go unrecognized at a particular time and in a speciﬁc context – their activation often requires the re-examination of these connections in a new or diﬀerent context. In their study of biotechnology start-ups referred to earlier, Maurer and Ebers (2006) demonstrate how social capital represents latent value in connections which may be activated later and used to create value. Maurer and Ebers found that small initial diﬀerences in the social connections of founders had important implications for their business orientation and performance much later, aﬀecting particularly their willingness to engage with important, non-scientiﬁc external constituencies. Speciﬁcally, some crucial actors in the more successful ﬁrms had exposure to and interest in diﬀerent social circles and their networks extended beyond scientiﬁc communities. For example, one came from a family of entrepreneurs; another had an aﬃnity for dealing with ﬁnanciers. Maurer and Ebers (2006) comment that: ‘although these informants themselves did not regard these earlier contacts as important for their ﬁrms or as part of their social capital at the time of founding’, these connections appear subsequently to have motivated entrepreneurs to specialize in those external constituencies with which they felt most familiar and which turned out to be important for their business. Thus in subtle ways, social capital comprises a latent set of connections to be activated or built upon as and when situations occur, develop or are created or recognized. The ﬁfth core insight that social capital theory brings to an understanding of collaborative advantage is that of convertibility. Bourdieu (1985) states that social capital can be converted, in certain conditions, into other forms of capital, especially economic capital. Schuller et al. (2000) suggest that Bourdieu regards social capital as a multiplier for economic and cultural capital. Ample empirical evidence, both anecdotal and researchbased, supports this assertion. For example, Baker (2000) cites research

Capitalizing on connections

215

showing that 75 per cent of start-ups and new businesses ﬁnd and secure ﬁnancing through the ‘informal investing grapevine’ – the social networks of capital-seekers and investors. Indeed, these networks can be viewed as informal capital markets. More recently, in their research on high-growth ventures before an initial public oﬀering, Florin et al. (2003) found a positive relationship between a venture’s social capital and its ability to accumulate both human and ﬁnancial capital. This provides the venture with a durable source of competitive advantage. Similarly, Hochberg et al. (2007) demonstrate the importance of several social network features for venture capital performance. Echoing earlier points, they emphasize that networking’s role is about more than just making connections; it is about building partnerships based on mutual beneﬁt and reciprocity.

A KNOWLEDGE ECONOMY IS A RELATIONAL ECONOMY In a series of reflections on the near future, Peter Drucker (2001) argued that the next society will be a knowledge society – ‘knowledge will be its key resource and knowledge workers will be the dominant group in the workforce’. In many ways, this is already evident not only for more advanced economies but also for ﬁrms and organizations. A recent report on global innovation suggests that one of the main drivers of productivity disparities is diﬀerences in knowledge – both in terms of individual skills and competencies and in terms of the knowledge-based capabilities of ﬁrms (Advanced Institute of Management Research, 2008). These ﬁndings have profound implications for all engaged in theory, policy and practice. In particular, they highlight not only the importance of human and intellectual capital for today’s economy, but also the signiﬁcance of social capital since a knowledge economy is a relational economy. In seeking a theoretical explanation for the knowledge-based view of the ﬁrm, we have argued that social capital is critical for the creation of intellectual capital (Nahapiet and Ghoshal, 1998). First, our deﬁnition of intellectual capital as ‘the knowledge and knowing capability of a social collectivity’ reflects our view that knowledge is fundamentally a social phenomenon. Knowledge is data whose meaning is socially created and sustained – in groups, communities, professions and organizations. Thus, a rich understanding of knowledge and knowing of necessity draws us into analysing its social origins and embeddedness, whether in scholarly communities, communities of practice or business organizations (Brown and Duguid, 2001). This is particularly the case for collective knowledge, which scholars suggest is the most strategically signiﬁcant type of knowledge for ﬁrms.

216

Social capital: reaching out

The second reason why social capital matters is its role in innovation – that is, the creation of new intellectual capital is the foundation for value creation in ﬁrms and results from combination and exchange of knowledge (Nahapiet and Ghoshal, 1998). It is now widely accepted that eﬀective knowledge production requires that diﬀerent kinds of expertise and information be combined, and that where these belong to diﬀerent people or groups, social exchange must occur. It follows that, since social capital signiﬁcantly influences both exchange and combination of knowledge, it is a major influence on the creation of intellectual capital. In addition, we have suggested that the relationship between social and intellectual capital is not one-way. Intellectual capital can be important for social capital formation; for example, research shows that sharing certain forms of knowledge enhances a ﬁrm’s social capital. It may be more appropriate to view their development as co-evolutionary. Since we ﬁrst put forward this argument, subsequent developments have only served to strengthen our conviction that a knowledge economy is a relational economy. Changes within companies – such as the move to horizontal organization and to projects rather than jobs – mean that increasingly the ﬁrm is viewed as a distributed knowledge system in which the ability to draw on knowledge from all its parts is critical for success (Doz et al., 2001). This is leading many executives to reconsider their organizational design as they seek to develop more eﬀective relationships between people and organizational units and to make cooperation the norm for their enterprise (Nahapiet et al., 2005). The more rapidly knowledge develops and the more diverse its sources, the more ﬁrms are turning to relational contracting and collaboration with other ﬁrms since it is no longer possible for them to be world leaders in all ﬁelds that may be relevant to their business. As I have argued above, their ability to access, leverage and create new knowledge will depend critically on their relationships with others. Finally, as Drucker suggests, human capital has become the key resource in a knowledge economy. The ability of organizations to attract and develop the best talent – either by employment or engagement in its projects through open-sourcing and networked innovation – is signiﬁcantly influenced by their ability to build and sustain eﬀective connections and relationships with talented people, individually and collectively. Thus social capital is growing in prominence as a factor shaping the human capital available to organizations and thus as an important influence on yet another element of strategic advantage. In addition to the above points, a substantial body of research explores the links between social capital and knowledge processes; however, a comprehensive analysis of this ﬁeld is beyond the scope of this chapter.2 What I seek to do here is two things: ﬁrst, to highlight key research themes

Capitalizing on connections

217

and ﬁndings; and second, to identify important issues that remain underexplored. Where appropriate, I will discuss further insights into these relationships in the subsequent sections.

ALL THREE DIMENSIONS OF SOCIAL CAPITAL AFFECT KNOWLEDGE PROCESSES One of the enduring questions for social capital scholars is: how should we view social capital? In particular, scholars have long argued that social capital is not a one-dimensional concept. Responding to this concern, we argued for the importance of considering multiple facets of social capital – independently and together. Subsequent empirical research has supported this view. Arguably, the strongest stream of social capital research focuses on structural social capital, examining the ways in which diﬀerent patterns and conﬁgurations of ties influence information and knowledge processes. Two themes dominate this area: the importance of network centrality and the eﬀects of diﬀerent ties. Studies consistently show that actors located in central network positions are better positioned to exploit internal capabilities and obtain preferential access to external resources. Indeed, in their discussion of the ﬁrm as a portfolio of relationships, Venktraman and Subramaniam (2002) suggest that centrality is the pivotal construct for understanding economies of expertise. Moreover, the research on the diﬀerent types of ties examines bridging, bonding, linking and especially the relative value of brokerage and closure. In addressing this theme, Burt (2005, and Chapter 2 in this volume) shows that brokerage is associated with growth and innovation and achieves its impact through creating a vision advantage. Closure is associated with trust and alignment and achieves its impact primarily through reputation eﬀects. Much work has also been done on the implications for knowledge of the relational dimension of social capital. Trust and trustworthiness remain dominant concerns, with research providing mounting evidence for both the direct eﬀects that trust has on important organizational phenomena, such as information sharing, and its enabling eﬀects, where trust creates or enhances the conditions that promote organizational outcomes such as cooperation (McEvily et al., 2003). For example, Larson (1992) shows the importance of trust and reciprocity as integrating and controlling devices that facilitate learning among successful entrepreneurial networks. Several aspects of relational quality feature strongly in research, particularly following Szulanski’s study (1996) demonstrating that arduous relationships form a barrier to eﬀective knowledge transfer. Quality of

218

Social capital: reaching out

relationships between individuals – particularly the amount of time spent face to face – is especially crucial where the task or knowledge is complex (Hansen, 1999) or where the knowledge requires signiﬁcant local adaptation (Hargadon, 1998). Similarly, where the task requires innovation, a range of relational processes and interaction strategies can be crucial (Cuervo-Cazurra and Un, 2004). However, not all studies support the importance of relationship quality. Yli-Renko et al. (2001), for example, do not ﬁnd the expected relationship between knowledge acquisition and the quality of relationships between young technology-based ﬁrms and their customers. Finally, the importance of identity as a feature of relational social capital has come increasingly to the fore, particularly in research on networks and communities of practice (Brown and Duguid, 2001). The connections between social identity and social capital are explored in detail by Kramer in Chapter 9 in this volume. Although generating less interest than the structural and relational dimension, there is a growing body of work that supports the importance of cognitive social capital in knowledge processes. The central idea here is that communities develop unique social and cognitive repertoires that guide their interpretations of the world and influence their interactions with others, based on whether or not they share a common interpretive frame (Boland and Tenkasi, 1995; Brown and Duguid, 2001). In a study of the UK health sector, Ferlie et al. (2005) found that social and cognitive factors accounted for the diﬀerential spread of evidence-based medical innovations in multi-professional organizations. Innovations spread eﬀectively when professionals had a foundation of shared identity and values that encouraged and enabled adoption of new treatments. By contrast, innovations failed to spread when both social and cognitive or epistemological boundaries existed between professional groups. The former frequently involved disputes concerning roles and social boundaries between doctors and nurses; the latter, diﬀerent assumptions held by various professional groups about what constitutes appropriate evidence. Recent research is now examining not only the three dimensions of social capital separately, but also their interrelationships. For example, within organizations, structural social capital is a particularly potent influence on the transfer of explicit, relatively well-understood information and knowledge, whereas cognitive and relational social capital are more influential for complex, uncertain or tacit knowledge (Hansen, 1999). McEvily and Marcus (2005) observe a similar pattern in interorganizational relations. In their exploration of social capital and knowledge transfer, Inkpen and Tsang (2005) show how the three dimensions aﬀect knowledge flow in three common network types: intra-corporate networks, strategic alliances and industrial districts. Their analysis describes not only the interrelationships

Capitalizing on connections

219

between the three social capital dimensions, but also the interplay between individual and organizational social capital in diﬀerent institutional settings. Speciﬁcally, they hypothesize that organizational social capital takes priority over individual social capital in explaining knowledge transfer in strategic alliances, whereas the reverse is true in industrial districts. For intra-corporate networks, they suggest that the social capital that impacts upon knowledge transfer begins at an organizational level and is subsequently enhanced by social capital developments at the individual level.

THE LINKS BETWEEN SOCIAL CAPITAL AND KNOWLEDGE PROCESSES ARE OFTEN COMPLEX One of the strong messages from recent research is that the relationships between social and intellectual capital are frequently contingent. What is eﬀective in one context may be less so in another. In his research on managerial performance, Moran (2005) shows how structural embeddedness plays a stronger role in explaining more routine, execution-oriented tasks, whereas relational embeddedness, that is, closeness and relational trust, plays a stronger role in explaining new, innovation-oriented tasks. Even the eﬀects of network centrality may be contingent. For example, Stam and Elfring (2008) found that a combination of high network centrality and extensive bridging ties strengthened the link between entrepreneurial orientation and new venture performance. However, network centrality actually weakened that relationship where ﬁrms had few bridging ties. Time may be an important factor too. For example, Soda et al. (2004) have shown that in the Italian TV production industry, current structural holes rather than past ones enhance current network performance; the opposite is true for closure. While both structural holes and closure appear important, they matter at diﬀerent points in time. It is important to remember that social capital can have detrimental as well as positive implications for knowledge, learning and ultimately performance. In a research ﬁnding reminiscent of Powell and Smith-Doerr’s (2003) observation that the ties that bind may become the ties that blind, Maurer and Ebers (2006) found that relational lock-in may lead to cognitive lock-in. In other words, executives who identiﬁed themselves primarily as scientists and whose network membership comprised largely other scientists found themselves locked into these relationships, which made it diﬃcult for them to create new connections and thereby access a wider set of resources necessary to help develop their ventures. Even more importantly, some networks can prevent innovation and can be anti-competitive.

220

Social capital: reaching out

SOCIAL CAPITAL IS A CRITICAL SUCCESS FACTOR IN INNOVATION All the research reviewed in this chapter points to social capital’s influence across the entire range of knowledge processes – acquisition, exchange, transfer and exploitation. However, remember the critical importance of innovation for value creation – at the level of the ﬁrm, the network and the economy. It is here that the impact of social capital is of greatest strategic signiﬁcance. A full review of the ﬁeld of innovation falls outside the scope of this chapter but several key ﬁndings are highlighted below. Collaborative advantage and innovation are inextricably linked – arguments that support the former also apply to the latter. Increasing evidence shows that scientiﬁc and technological breakthroughs require contributions of many actors working in networks. As innovation processes become more complex and products more modular, knowledge tends to be scattered among a large number of organizations (Baldwin and Clark, 2000). Firms must therefore collaborate to harness this knowledge. Research shows that the relevant set of organizations can include key customers (YliRenko et al., 2001), suppliers, alliance partners, and knowledgeable individuals and groups. Formal and informal relations, as well as individual and organizational social capital, signiﬁcantly influence the eﬀectiveness of ﬁrms’ networks. For example, managerial networking across contexts helps determine whether promising practices are adopted. However, research is only now starting to track these complex relationship dynamics. An organization’s ability to build eﬀective internal cooperation and coordination helps determine success in managing external connections and relationships as well. In my own research, I have consistently found that organizations appear to ﬁnd it harder to build eﬀective cooperation and collaboration across internal organizational boundaries than they do, say, with key customers – a pattern observed by others too (Gulati, 2007). Whether internal or external, ﬁrms’ ability to organize and manage key relationships strategically is critical for innovation success and will be explored in more detail later in this chapter.

FUTURE DIRECTIONS: TOWARDS A PRACTICE PERSPECTIVE? So far I have argued that the social capital perspective demonstrates that a knowledge economy is a relational economy and that social connections help produce assets such as information, knowledge and learning.

Capitalizing on connections

221

However, notwithstanding its major contributions thus far, this approach is still too narrow, I believe, to understand fully this critical interrelationship. In particular, the dominant viewpoint of the links between knowledge and social relations is built on a signiﬁcant but still limited view of both knowledge and the social processes and practices in which it is created and embedded. Speciﬁcally, the issue arises from the prevailing model of knowledge – one Cook and Brown (1999) describe as an epistemology of possession, that is, one which views knowledge as something that people possess. Further, the prevailing model of relationships views them as conduits or pipes through which information and knowledge flow (Boland and Tenkasi, 1995). This may reflect the dominance of a network perspective in much social capital research. When taken together, the resources ‘available through’ aspect of our deﬁnition of social capital dominates research in this area. By focusing more on knowledge transfer and knowledge exchange rather than knowledge and knowing in social interaction, extant social capital studies largely ignore important insights from research in adjacent domains, particularly the epistemology of practice which views knowing as action and focuses on the coordinated activities of individuals and groups doing their ‘real work’. As a consequence, I believe our understanding is impoverished. One classic study that represents the practice approach is Orr’s (1998) anthropological research on technical representatives who service and repair photocopiers at customers’ sites. Orr shows how technicians create both a shared understanding, that is, their collective knowledge, and their shared identity as they do their work – in real time. This view of social resources led us to include the phrase ‘embedded within’ as part of our deﬁnition of social capital and to argue that social and intellectual capital frequently co-evolve. Yet this view is still under-represented in social capital research; the ﬁeld would beneﬁt from more exploration in this genre. Two important and highly relevant research streams can help scholars make progress in this area. First, the practice perspective has become a clear line of research for knowledge studies (Osterlund and Carlile, 2005). This work emphasizes the salience of a community rather than a network, most obviously but not exclusively in studies of communities of practice. Although not using the term ‘social capital’, these studies oﬀer valuable insights into the social processes involved in the creation and application of knowledge. As Orlikowski observes, the practice perspective ‘suggests that knowing is not a static embedded capability or stable disposition of actors, but rather an ongoing social accomplishment, constituted and reconstituted as actors engage in the world of practice’ (2002, p. 249). Her work provides a diﬀerent and important lens for understanding the

222

Social capital: reaching out

practicalities of innovation, which she describes as ‘knowing how to innovate’. Second, one of the fastest-growing research areas in strategic management is ‘strategy as practice’. These studies focus on the doing of strategy: who does it, what they do, how they do it, what they use and what implications this has for shaping strategy (Jarzabkowski and Spee, 2008). The interest in strategy praxis – the flow of activity in which strategy is accomplished – could help refocus research on the practices that connect social and intellectual capital. For example, challenging the influential positions of Putnam (1993a) and Saxenian (1994), Cohen and Fields (2000) have recently shown that it is cooperative activity and collective engagement that build trust in Silicon Valley rather than community norms of trust leading to cooperative activity. A practice perspective on social capital would advance research in important ways. By embracing community as well as network, it would renew focus on relational and cognitive dimensions to rebalance the current dominance of structural perspectives. By moving beyond exchange to studying interaction, it could open the black box of activities that are so essential for innovation. In so doing it would provide a rich perspective on the social dimensions of the knowledge economy that would complement the prevailing one today.

SOCIAL CAPITAL AND STRATEGIC MANAGEMENT IN CONDITIONS OF UNCERTAINTY The challenges associated with managing change and uncertainty are enduring themes in strategic management and receive ever more attention in both practitioner and scholarly publications. As Brown and Eisenhardt commented in 1998: ‘[W]e bring up the issue of change because change is the striking feature of contemporary business. Change is ubiquitous – in every industry, in every geography, in every ﬁrm’ (Brown and Eisenhardt, 1998, p. 3). In 2008, the pace of change and perceived levels of uncertainty are still greater. The question now is: how can organizations develop strategy in these conditions? Although the vast literature on strategic change oﬀers a plethora of answers and approaches – dynamic capabilities, structured chaos, strategic agility, managing on the edge, scenario planning, strategic options – a relatively small set of recurring themes emerge. For example, in their recent study of strategic agility, Doz and Kosonen (2008) ﬁnd that winning at what they call ‘fast strategy’ hinges on a few deciding diﬀerences. They identify three in particular: strategic sensitivity, that is, the early awareness of

Capitalizing on connections

223

incipient trends and converging forces; the ability to take key decisions and make real commitments fast; and resource fluidity, that is, the ability to mobilize and deploy flexibly key resources. To these, others add the ability to create new opportunities and to keep options open. While I do not claim that these themes are exhaustive, they are widely acknowledged as important. From my earlier arguments, it is also clear that social capital oﬀers important insights into these critical activities. Some insights are not new to those engaged in the change agenda, but bringing them together within a social capital perspective deepens the understanding of the critical importance of social connections for strategy in the context of uncertainty. The argument here is that both social networks and social communities shape the flow and integrity of knowledge which then enables ﬁrms to detect and interpret early signs of change – strategic sensitivity. The information beneﬁts of social capital, in terms of access, timing and referrals, are well known (Burt, 2005). Further, in opening their discussion of strategic sensitivity, Doz and Kosonen (2008) observe: ‘[I]nsight is not gained in a vacuum . . . Insight results from connection, from a rich network of creative interactions . . . So maximising the channels and integrity of knowledge exchange with the outside world is a precondition for sensitivity’ (Doz and Kosonen, 2008, p. 55). They describe four elements of IBM’s approach: being connected to the world; being connected internally; relying on business development experiments and corporate venturing; and fostering an open strategic dialogue. These are strikingly similar to the practices reported in management innovation studies (Birkinshaw et al., 2007b) and that I observed in Skandia’s extensive intellectual capital development programme. Skandia fostered experiments, enquiry and dialogue alongside what I called ‘the pursuit of permeability’ – a wide-ranging set of practices for bringing the outside world into Skandia and taking the inside out (Nahapiet, 2008a). What is particularly evident is the value of partners in exploring, probing, experimenting and learning. As Hagel and Brown (2005) suggest in their studies of how organizations beneﬁt from managing on the edge, by connecting with other specialized institutions, ﬁrms create the opportunity for leveraged capability-building – getting better faster by working with others. Relationship quality is central to building collective commitments rapidly among the top management team – Doz and Kosonen’s second theme. Their discussion echoes the arguments made earlier for collaboration’s importance, for the shift to power-with from power-over, and for the model of collaborative community described by Adler and Heckscher (2006). Unlike strategic sensitivity, which is signiﬁcantly shaped by network diversity, structural holes, and strong and weak ties, collective commitments draw more heavily on relational and cognitive social capital and the

224

Social capital: reaching out

concept of community. Trust, shared values and common identity are pivotal in creating a context that enables leveraging individual diﬀerences in highly interdependent tasks and decision-making. These are central concerns for social capital scholars. Despite this, to date, social capital research on board and senior management members has typically focused on their external relations and networks. An urgent priority now is to refocus research on the internal structure and dynamics of senior executives’ relationships, identifying the conditions, activities and practices that shape the development of trust, reciprocity and collaboration. As with my earlier comments on social capital as practice, I believe there is already much research in adjacent ﬁelds, speciﬁcally on boards, senior management teams and strategic practices such as strategy workshops and scenariobuilding, that could accelerate this work. Agile resourcing is the third element of fast strategy. For Doz and Kosonen, the dominant concern is resource fluidity – the ability to attract and reconﬁgure resources quickly and flexibility. Bouty (2000) argues that: ‘the concept of organization should be knowledge based and should not be deﬁned in terms of resources as property but in terms of the ability to raise key resources and knowledge’. Most of this chapter has already discussed the many ways in which social capital shapes actors’ ability to access and leverage key resources; this topic will also be addressed later in the section on organizing for social capital. There does, however, seem to be a fourth element crucial for strategic management in conditions of uncertainty – the ability to create and maintain strategic opportunities, that is, to make investments that build the opportunity to respond and beneﬁt from future contingent events. This capability draws together elements of both strategic sensitivity and agile resourcing. Scenarios and real options theory are the two dominant strategy perspectives that address the creation of future opportunities. Given organizational theory’s reluctance to embrace the idea that organizations can proactively exploit risk rather than just absorb it, Kogut and Kulatilaka (2001) suggest that capabilities can be viewed as real options that permit the ﬁrm to make the best response to future market opportunities. Building from this perspective, I suggest that social capital brings distinctive insights to this area. To begin with, opportunities have long been recognized as one of the core beneﬁts of social capital. Uzzi, for example, describes the advantages for the textile network in terms of the flexibility to explore new relationships and the opportunities they bring (Uzzi, 1996, 1997). Ahuja (2000a) similarly shows the importance of both inducements and opportunities arising from interﬁrm linkages. However, the opportunities available to organizations come not only from their existing ties. Delbridge and

Capitalizing on connections

225

Mariotti (2003) argue that in understanding the beneﬁts of social networks we need to go beyond the strong–weak tie dichotomy to understand the impact of other facets of ties on innovation in dynamic environments. In particular, they draw attention to potential ties, that is, embryonic relationships that have the opportunity for further development; and latent ties, that is, those relationships that are currently in a state of inactivity. Their empirical work in the motor sport industry shows how ﬁrms adapt, augment and renew their knowledge base not only through current network relationships but also through exploring new opportunities by deliberately establishing potential ties while suspending others that are redundant, that is, allowing them to become latent. These latent ties remain available for reactivation later if circumstances change. Second, social capital’s insights into the appropriability of connections also imply that the range of resources available to organizations may be wider and greater than it appears. An organization’s connections are also open to mobilization as needs arise. However, much more work is needed to understand the fungibility of connections in diﬀerent settings. A third feature of social capital – reputation – can be important in creating opportunities in uncertain times. Social capital in the form of reputation can mitigate risk and open up business opportunities that might otherwise not exist. In our deﬁnition of social capital we include the resources ‘derived from the network of social relationships’ to highlight the beneﬁcial advantages that status can convey, since actors’ reputations are constructed in part from the identities of their associates. This is especially important where the quality of potential partners is diﬃcult to establish or unknown – as in the early stages of company start-ups, entrepreneurial ventures and areas of new technology (Stuart et al., 1999). In such contexts, where there is little direct experience and track record for people and organizations seeking support, the reputations of their associates and exchange partners provide proxy measures of quality and potential success. This reflects the idea of trust transferability in which initial trust impressions are based in a source other than the trustee (McEvily et al., 2003). The ways in which organizations build and sustain their networks and relationships may provide important resources that determine future opportunities and business options. However, internal and external initiatives that build social capital are not costless. They may be regarded as ineﬃcient, frequently generating low returns on investment since one cannot know in advance which connections will prove to be the most useful and at what time. It is important to judge these investments in terms of what North (1990) describes as adaptive rather than allocative eﬃciency. Adaptive eﬃciency is concerned with encouraging decentralized decision processes, entrepreneurship, innovation and creative activity of all sorts.

226

Social capital: reaching out

North’s observation that ‘we are far from understanding how to achieve adaptively eﬃcient economies’ (1990, p. 81) remains true today. It should become a priority to extend research on the role of social capital in building strategic options.

ORGANIZING FOR SOCIAL CAPITAL: CURRENT ISSUES The core proposition of this chapter is that ﬁrms perform better because, through their connections and relationships, they are better able to access and beneﬁt from strategic opportunities and resources that impact upon performance. The question follows naturally: if social capital is so valuable as a resource, how do organizations go about capitalizing on connections? How do ﬁrms organize for social capital? This issue lies right at the centre of one of the largest debates in social capital research: whether social capital is a by-product of other activities or a resource that can be deliberately developed. Coleman (1988) adopts the former position: social capital is a by-product of activities engaged in for other purposes. Indeed, this reinforces his notion of appropriability. By contrast, Bourdieu (1985) takes the latter position: social capital can and should be deliberately developed. He describes the network of relationships as: ‘the product of investment strategies, individual or collective consciously aimed at establishing or reproducing relationships that are directly usable in the short or long term’. Strategic management research adopts both positions. In their discussion of strategy as guided evolution, Lovas and Ghoshal (2000) seem more oriented to a view of social capital as by-product: ‘to the extent top management can influence where and how employees use their time and energy, they can also influence what human and social capital is created and maintained’ (2000, p. 891). By contrast, Maurer and Ebers (2006) show clearly that executives can use their approaches to relationship management to reconﬁgure their connections deliberately, and thus their social capital. This ability distinguished between high and low performers in biotechnology ﬁrms. In practice, both emergence and design seem important in the evolution of social capital. Even Coleman acknowledges that some forms of social capital directly result from investment; he identiﬁes intentional organization, particularly business organization, as the most prominent of these. In this speciﬁc context, he argues that: ‘like other forms of capital, social capital requires investment in the designing of the structure of obligations and expectations, responsibility and authority, and norms (or rules) and

Capitalizing on connections

227

sanctions that will bring about an eﬀectively functioning organization’ (Coleman, 1990, p. 313). Beyond those already discussed, two additional themes stand out in current research on the organizational aspects of social capital. First, organizing for social capital involves both leveraging existing connections and building productive new ones; and second, it involves understanding the importance of brokers and their activities. Current discussions of strategic organization emphasize the idea of organizational ambidexterity – the ability to balance two diametrically opposed organizational qualities: alignment and adaptability (Gibson and Birkinshaw, 2004). Alignment maximizes the present, leverages existing ideas and markets, exploits current assets and manages evolutionary change. Adaptability anticipates change, explores possible futures and manages revolutionary transformation. Although the term ‘ambidexterity’ normally applies to strategic organization, it may also be useful in considering the dual challenges of organizing for social capital, since networks and relationships also need to be ambidextrous. Organizations must see, understand and manage their important existing ties and connections – active and latent. Networks and communities do matter. Yet social capital scholars consistently show that both individuals and organizations are relatively poor at doing just this. They rarely understand the dynamics and signiﬁcance of their social networks. However, the relational perspective is and has been taken seriously in some functional areas, especially marketing. For example, marketing scholars have focused their attention on the value of loyalty, the importance of account management, and the role of interactions and networks (Ford, 1990). Further research shows that the creation of alliance function can positively impact upon the success rate and value created through alliances (Kale et al., 2001). In professional services, sometimes entire teams are recruited because of their internal relationships and their external ones with major clients. But it would seem that at least in the Anglo-Saxon world, we are still at a relatively early stage in aligning organizational design with strategies to sustain and leverage social capital in existing networks and relationships. Running strategy as a portfolio of relationships is still a considerable challenge. Curiously, more research may be happening on building exploratory networks to foster innovation. Whether this represents greater research interest or greater strategic activity remains an open question. We have already touched on the importance of social capital for adaptive eﬃciency and option creation – yet how do ﬁrms organize for this? How do they build networks for unknown worlds? How do they design for serendipity? Birkinshaw et al.’s (2007) recent research examines how organizations

228

Social capital: reaching out

develop networks in environments characterized by discontinuous change. They describe a three-stage process – ﬁnding, forming and performing. Their approaches to network-building reflect diﬀerent challenges arising from two network features: how diﬃcult is it to ﬁnd potential partners? and how easy is it to form a relationship with them? Their study provides rich insights both into the ways in which ﬁrms build connections, turning networks into performing partners, and into the enormous latent value of networks. The role of actors, especially brokers, represents a second key theme in social capital research. The position of broker is the organizational manifestation of the importance of network centrality. Organizational studies have long recognized the power of brokers; brokers have been viewed either as lynchpins (Likert, 1961, 1967) active in making connections between diﬀerent individuals and groups, or as gatekeepers (Allen, 1977) drawing together valuable information from distributed networks but largely keeping their various ties apart. Social capital research adopts both perspectives on brokers. Burt (2005), for example, has repeatedly shown that brokers consistently do better; their opportunities to beneﬁt arise from spanning structural holes. His approach emphasizes Simmel’s tertius gaudens and the beneﬁts that accrue to brokerage through strategic separation. Recently, though, Obstfeld (2005) has argued for an alternative brokering orientation – Simmel’s tertius iungens – which he describes as a behavioural orientation toward connecting people in one’s social network by either introducing disconnected individuals or facilitating new coordination between connected individuals. His research on networks and innovation in automotive manufacturing demonstrates that a tertius iungens orientation, dense social networks and diverse social knowledge predict involvement in innovation. His work represents a growing interest in what brokers do, how they bring diﬀerent actors together, and the importance of social knowledge in innovation, that is, knowledge of the social and political dynamics that surround innovation eﬀorts. Dhanaraj and Parkhe (2006), who are also interested in how brokers operate to create value, focus on hub ﬁrms rather than individuals in brokerage positions. Their analysis, motivated by the lack of research on network processes, focuses on network orchestration – which they deﬁne as the set of deliberate, purposeful actions taken by hub ﬁrms as they seek to create value (expand the pie) and extract value (gain a larger slice of the pie) from the network. They identify three critical orchestration processes: managing knowledge mobility, managing innovation appropriability and managing network stability. These processes include many features designed to build relational social capital, such as identity, trust and reputation, as well as contexts that encourage reciprocity. Various publications about Li &

Capitalizing on connections

229

Fung and Toyota provide complementary insights into how these companies have orchestrated extensive networks of relationships to create value.

CAPITALIZING ON CONNECTIONS: SUMMARY AND CONCLUSIONS My aim has been to make the case that social capital provides a distinctive perspective on strategic management and addresses directly several of the major current questions in the ﬁeld. It advances our understanding of the emerging model of strategy as a portfolio of relationships and of collaboration as a vital foundation of strategic advantage. Its importance lies in the links between social and intellectual capital and the signiﬁcance of relationships in a knowledge economy. Of growing interest is social capital’s potential as a resource in times of uncertainty, which leads to the imperative of more careful investigation of the organizational processes that foster and leverage this critical relational asset. Several points of tension within the social capital ﬁeld must be acknowledged. They include the relative importance of structural versus relational approaches, and the balance between network and community in understanding social capital. Most social capital research deals with the positive eﬀects of social capital, perhaps underemphasizing its negative impact and ‘dark side’. This is true for my own chapter. However, an aspect which may be even more important is that most social capital research referenced here and in the literature is located in a Western and largely Anglo-Saxon tradition. It is thus framed from an individualistic rather than a collectivist perspective. This Western frame of reference has important consequences for how we understand the dynamics and consequences of social relations. Initial tests of Western models of social capital in other cultures already suggest important limitations on the conclusions drawn thus far. In a recent study of structural hole theory in diﬀerent cultural contexts, Ziao and Tsui (2007) ﬁnd that the typical collectivist culture of China dampens the eﬀects of structural holes. In organizations that foster a high commitment culture – one that emphasizes mutual investment between people – the control beneﬁts of structural holes conflict with the dominant spirit of cooperation, and their information beneﬁts do not materialize. They conclude that brokers that operate in accordance with structural hole theory neither ﬁt nor succeed within China’s collectivist values. Issues raised by such cross-cultural research require urgent further exploration by both scholars and practitioners. Such cross-cultural research has, I believe, two further important implications. First, paradoxically, it reminds

230

Social capital: reaching out

us that our knowledge is socially embedded – in a particular culture at a particular time. This is true of our research. The cultural assumptions about relationships and the many ways these may impact upon relational capital should come to the fore in future research. Second, it also reminds us that the relational assumptions underpinning collaborative advantage may be more diﬃcult to achieve in highly individualistic settings. That there are alternative ways of managing brokering is evident from Obstfeld’s (2005) research, but the model of collaborative community may be too much for many people and ﬁrms. In his review of the near future, Drucker (2001) argues that the ‘multinationals of 2025 are likely to be held together and controlled by strategy. There will still be ownership of course. But alliances, joint ventures, minority stakes, know-how agreements, and contracts will increasingly be the building blocks of a confederation. This kind of organization will need a new kind of top management’ (Drucker, 2001, p. 5). I argue here that it will also need a new kind of strategy – one in which a profound understanding of networks, communities, relationships and social capital is central.

NOTES 1. I am grateful to Mendoza College of Business at the University of Notre Dame for funding the interdisciplinary conference on social capital at which some of the ideas developed in this chapter were ﬁrst presented. I acknowledge the unfailing encouragement and valuable insights of Viva Bartkus and Peter Smith Ring in our ongoing conversations about the potential of social capital research. 2. For a good review of the links between networks and innovation, see Pittaway et al. (2004) and Bessant et al. (2005).

REFERENCES Adler, Paul and Charles Heckscher (2006), ‘Towards collaborative community’, in Charles Heckscher and Paul Adler (eds), The Firm as a Collaborative Community: Reconstructing Trust in the Knowledge Economy, Oxford: Oxford University Press, pp. 11–105. Adler, P.S. and S.W. Kwon (2002), ‘Social capital: prospects for a new concept’, Academy of Management Review, 27 (1), 17–40. Advanced Institute of Management Research (2008), Competing on Knowledge: The UK’s Global Innovation Challenge, London: Advanced Institute of Management Research. Ahuja, Gautam (2000a), ‘The duality of collaboration: inducements and opportunities in the formation of interﬁrm linkages’, Strategic Management Journal, 21 (3), 317–43. Ahuja, Gautam (2000b), ‘Collaborative networks, structural holes, and innovation: a longitudinal study’, Administrative Science Quarterly, 45 (3), 425–55.

Capitalizing on connections

231

Allen, Thomas J. (1977), Managing the Flow of Technology, Cambridge, MA: MIT Press. Baker, Wayne E. (1990), ‘Market networks and corporate behavior’, American Journal of Sociology, 96, 589–625. Baker, Wayne E. (2000), Achieving Success Through Social Capital: Tapping the Hidden Resources in Your Personal and Business Networks, San Francisco, CA: Jossey-Bass. Baldwin, C.Y. and K.B. Clark (2000), Design Rules: The Power of Modularity, Cambridge, MA: MIT Press. Barney, J. (1991), ‘Firm resources and sustained competitive advantage’, Journal of Management Studies, 17 (1), 99–120. Bessant, John, Robert Phelps and Richard Adams (2005), External Knowledge: A Review of the Literature Addressing the Role of External Knowledge and Expertise at the Key Stages of Business Growth and Development, London: Advanced Institute of Management Research. Biggart, Nicole and Rick Delbridge (2004), ‘Systems of exchange’, Academy of Management Review, 29 (1), 28–49. Birkinshaw, Julian, John Bessant and Rick Delbridge (2007a), ‘Finding, forming, and performing: creating networks for discontinuous innovation’, California Management Review, 49 (3), 67–84. Birkinshaw, Julian, Stuart Crainer and Michael Mol (2007b), ‘Special report on management innovation’, Business Strategy Review, 18 (1), 45–73. Boland, Richard and Ramkrishnan Tenkasi (1995), ‘Perspective making and perspective taking in communities of knowing’, Organization Science, 6 (4), 350–72. Bourdieu, Pierre (1985), ‘The forms of capital’, in J.G. Richardson (ed.), Handwork of Theory and Research for the Sociology of Education, New York: Greenwood, pp. 241–58. Bouty, Isabelle (2000), ‘Interpersonal and interaction influences on informal resource exchanges between R&D researchers across organizational boundaries’, Academy of Management Journal, 43 (1), 50–65. Brown, John Seely and Paul Duguid (2001), ‘Knowledge and organization: a socialpractice perspective’, Organization Science, 12 (2), 198–213. Brown, Shona and Kathleen Eisenhardt (1998), Competing On The Edge: Strategy As Structured Chaos, Cambridge, MA: Harvard Business School Press. Burt, Ronald S. (1992), ‘The social structure of competition’, in Structural Holes: The Social Structure of Competition, Cambridge, MA: Harvard University Press. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Capital, Oxford: Oxford University Press. Castells, Manuel (1996), The Rise of the Network Society, Oxford: Blackwell Publishing. Cohen, Stephen S. and Gary Fields (2000), Social Capital and Capital Gains: The Anatomy of An Entrepreneurial Region, Palo Alto, CA: Stanford University Press. Coleman, James (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94 (Supplement, Organizations and Institutions: Sociological and Economic Approaches to the Analysis of Social Structure), S95–S120. Coleman, James (1990), Foundation of Social Theory, Cambridge, MA: Harvard University Press.

232

Social capital: reaching out

Cook, Scott and John Seely Brown (1999), ‘Bridging epistemologies: the generative dance between organizational knowledge and organizational knowing’, Organization Science, 10 (4), 381–400. Cuervo-Cazurra, Alvaro and C.A. Un (2004), ‘Strategies for knowledge creation in ﬁrms’, British Journal of Management, 15 (1), 27–41. Delbridge, Rick and Francesca Mariotti (2003), Creating Knowledge and Overcoming Redundancy in Networks: Beyond the Strong/Weak Tie Dichotomy, London: Advanced Institute of Management Research. Dhanaraj, C. and A. Parkhe (2006), ‘Orchestrating innovation networks’, Academy of Management Review, 31 (3), 659–69. Dierickx, I. and K. Cool (1989), ‘Asset stock accumulation and sustainability of competitive advantage’, Management Science, 35 (12), 1504–11. DiMaggio, Paul (2001), The Twenty First Century Firm, Princeton, NJ: Princeton University Press. Doz, Yves and Mikko Kosonen (2008), Fast Strategy: How Strategic Agility will Help You Stay Ahead of the Game, Philadelphia, PA: Wharton School Publishing. Doz, Yves L., José Santos and Peter Williamson (2001), From Global to MetaNational: How Companies Win in The Knowledge Economy, Cambridge, MA: Harvard Business School Press. Drucker, Peter (2001), ‘The next society’, The Economist, 3 November. Eisenhardt, Kathleen M. and Filipe M. Santos (2002), ‘Knowledge-based view: a new theory of strategy’, in Andrew Pettigrew, Howard Thomas and Richard Whittington (eds), Handbook of Strategy and Management, London: Sage Publications, pp. 139–64. Farr, J. (2004), ‘Social capital: a conceptual history’, Political Theory, 32 (6), 6–33. Ferlie, Ewan, Louise Fitzgerald, Martin Woo and Chris Hawkins (2005), ‘The nonspread of innovations: the mediating role of professionals’, Academy of Management Journal, 48 (1), 117–34. Florin, J., M. Lubakin and W. Schulze (2003), ‘A social capital model of high growth ventures’, Academy of Management Journal, 46 (3), 374–84. Follet, Mary Parker (1918; 1998), The New State: Group Organization: The Solution of Popular Government, University Park, PA: Penn State Press. Ford, David (ed.) (1990), Understanding Business Markets: Interaction, Relationships, and Networks, London: Academic Press. Fukuyama, Francis (1995), ‘Social capital and the global economy’, Foreign Aﬀairs, 74 (5), 89–103. Fukuyama, Francis (1999), Social Capital and Civil Society, International Monetary Fund Working Paper, WP/00/74, 1–18. Galaskiewicz, J. and A. Zaheer (1999), Research in the Sociology of Organizations, Greenwich, CT: JAI Press. Ghoshal, Sumatra and Peter Moran (1996), ‘Bad for practice: a critique of the transaction cost theory’, Academy of Management Review, 21, 31–48. Gibson, Christina and Julian Birkinshaw (2004), ‘The antecedents, consequences and mediating role of ambidexterity’, Academy of Management Journal, 47 (2), 209–26. Gomes-Casseres, Benjamin (1996), The Alliance Revolution: The New Shape of Business Rivalry, Cambridge, MA: Harvard University Press. Granovetter, M. (1985), ‘Economic action and social structure: the problem of embeddedness’, American Journal of Sociology, 91 (5), 481–510.

Capitalizing on connections

233

Gulati, Ranjay (2007), Managing Network Resources: Alliances, Aﬃliations, and Other Relational Assets, Oxford: Oxford University Press. Hagel, John and John Seely Brown (2005), The Only Sustainable Edge: Why Business Strategy Depends on Productive Friction and Dynamic Specialization, Cambridge, MA: Harvard Business School Press. Hansen, M.T. (1999), ‘The search-transfer problem: the role of weak ties in sharing knowledge across organization subunits’, Administrative Science Quarterly, 44, 82–111. Hargadon, Andrew (1998), ‘Firms as knowledge broker’: lessons in pursuing continuous innovation’, California Management Review, 40 (3), 209–27. Hargadon, J. (2006), ‘Understanding the cross-level embeddedness of interﬁrm partnership formation’, Academy of Management Review, 31 (3), 670–80. Heckscher, Charles and Paul Adler (eds) (2006), The Firm as a Collaborative Community: Reconstructing Trust in the Knowledge Economy, Oxford: Oxford University Press. Hochberg, Y.V., A. Ljundquist and Y. Lu (2007), ‘Whom you know matters: venture capital networks and investment performance’, Journal of Finance, 112 (1), 251–301. Inkpen, A. and E. Tsang (2005), ‘Social capital, networks, and knowledge transfers’, Academy of Management Review, 30 (1), 146–65. Jarzabkowski, Paula and A.P. Spee (2008), ‘Strategy as practice: a review and future directions for the ﬁeld’, forthcoming in International Journal of Management Reviews. Kale, Prashant, Jeﬀrey Dyer and Harbir Singh (2001), ‘Value creation and success in strategic alliances: alliancing skills and the role of the alliance function’, European Management Journal, 18 (5), 463–71. Kilduﬀ, Martin and Wenpin Tsai (2003), Social Networks and Organizations, London: Sage Publications. Kogut, B. and B. Kulatilaka (2001), ‘Capabilities as real options’, Organization Science, 12 (6), 744–58. Larson, Andrea (1992), ‘Network dyads in entrepreneurial settings: a study of the governance of exchange relationships’, Administrative Science Quarterly, 37 (1), 76–104. Likert, R. (1961), New Patterns of Management, New York: McGraw-Hill. Likert, R. (1967), The Human Organization, New York: McGraw-Hill. Lovas, Bjorn and Sumantra Ghoshal (2000), ‘Strategy as guided evolution’, Strategic Management Journal, 21 (9), 875–96. Maurer, Indre and Mark Ebers (2006), ‘Dynamics of social capital and their performance implications: lessons from biotechnology start-ups’, Administrative Science Quarterly, 51 (2), 282–92. McEvily, B., V. Perrone and A. Zaheer (2003), ‘Trust as organizing principle’, Organizational Science, 14 (1), 91–103. McEvily, Bill and Alfred Marcus (2005), ‘Embedded ties and the acquisition of competitive capabilities’, Strategic Management Journal, 26 (11), 1033–55. Moran, Peter (2005), ‘Structural versus relational embeddedness: social capital and managerial performance’, Strategic Management Journal, 26 (12), 1129–51. Moran, Peter and Sumantra Ghoshal (1999), ‘Markets, ﬁrms, and the process of economic development’, Academy of Management Review, 24 (3), 390–412.

234

Social capital: reaching out

Nahapiet, Janine (2002), ‘Toward a framework for the institutionalization of renewal’, paper presented at the Annual Meeting of the Academy of Management Meeting, Denver. Nahapiet, Janine (2008a), ‘Social capital and interorganizational relations’, in Steve Cropper, Mark Ebbers, Chris Huxham and Peter Smith Ring (eds), Oxford Handbook of Interorganizational Relations, Oxford: Oxford University Press. Nahapiet, Janine (2008b), ‘There and back again? Organization studies 1965–2006’, in Sue Dopson, Michael Earl and Peter Snow (eds), Mapping the Management Journey: Practice, Theory, and Context, Oxford: Oxford University Press, pp. 580–606. Nahapiet, Janine and Sumatra Ghoshal (1998), ‘Social capital, intellectual capital and the organizational advantage’, Academy of Management Review, 23 (2), 242–66. Nahapiet, Janine, Lynda Gratton and Hector Rocha (2005), ‘Knowledge and relationships: when cooperation is the norm’, European Management Review, 2, 3–14. Newell, S., C. Tanley and J. Huang (2004), ‘Social capital and knowledge integration in an ERB project team: the importance of bridging and bonding’, British Journal of Management, 15, S43–S58. North, Douglass C. (1990), Institutions, Institutional Change, and Economic Performance, Cambridge: Cambridge University Press. Obstfeld, David (2005), ‘Social networks, the Tertius Iungens orientation, and involvement in innovation’, Administrative Science Quarterly, 50 (1), 100–130. Oh, Hongseok, Martin Kilduﬀ and Daniel Brass (1999), ‘Communal social capitals, linking social capital, and economic outcomes’, paper presented at the Annual Meeting of the Academy of Management, Chicago, IL. Orlikowski, Wanda J. (2002), ‘Knowing in practice: enacting a collective capability in distributed organizing’, Organization Science, 13 (3), 249–73. Orr, Julian E. (1998), ‘Images of work’, Science, Technology, and Human Values, 23 (4), 439–55. Osterlund, Carsten and Paul Carlile (2005), ‘Relations in practice: sorting through practice theories on knowledge sharing in complex organizations’, Information Society, 21, 91–107. Pettigrew, Andrew, Howard Thomas and Richard Whittington (eds) (2002), Handbook of Strategy and Management, London: Sage Publications. Pittaway, Luke, Maxine Robertson, Kamal Munir, David Denyer and Andy Neely (2004), ‘Networking and innovation: a systematic review of the evidence’, International Journal of Management Reviews, 5–6 (3–4), 137–68. Polanyi, Karl (1944), The Great Transformation, Boston, MA: Beacon Press. Porter, Michael (1980), Competitive Strategy: Techniques for Analyzing Industries and Competitors, New York: Free Press. Powell, Walter and Laurel Smith-Doerr (2003), ‘Networks and economic life’, in Neil Smelser and Richard Swedberg (eds), The Handbook of Economic Sociology, New York: Russell Sage Foundation and Princeton University Press, pp. 368–402. Powell, Walter, Kenneth Koput and Laurel Smith-Doerr (1996), ‘Interorganizational collaboration and the locus of innovation: networks of learning in biotechnology’, Administrative Science Quarterly, 41 (1), 116–45. Prahalad, C.K. and Gary Hamel (1990), ‘The core competence of the corporation’, Harvard Business Review, 68 (3), 79–91.

Capitalizing on connections

235

Prahalad, C.K. and Gary Hamel (1994), ‘Strategy as a ﬁeld of study: why search for a new paradigm?’, Strategic Management Journal, 15 (Special Issue), 5–16. Putnam, Robert (1993a), Making Democracy Work: Civic Traditions in Modern Italy, Princeton, NJ: Princeton University Press. Putnam, Robert (1993b), ‘The prosperous community: social capital and public life’, American Prospect, 13 (Spring), 35–42. Putnam, Robert (1995), ‘Bowling alone: America’s declining social capital’, Journal of Democracy, 6 (1), 65–78. Quinn, James B. (1992), Intelligent Enterprise: A Knowledge and Service Based Paradigm for Industry, New York: Free Press. Reed, Richard and Robert DeFillippi (1990), ‘Causal ambiguity, barriers to imitation, and sustainable competitive advantage’, Academy of Management Review, 15 (1), 88–102. Saxenian, A. (1994), Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Cambridge, MA: Harvard University Press. Schuller, Tom, Stephen Baron and John Field (2000), ‘Social capital: a review and critique’, in Stephen Baron, John Field and Tom Schuller (eds), Social Capital: Critical Perspectives, Oxford: Oxford University Press, pp. 1–38. Soda, Guiseppi, Alessandro Usai and Akbar Zaheer (2004), ‘Network memory: the influence of past and current networks on performances’, Academy of Management Journal, 47, 893–906. Stam, W. and T. Elfring (2008), ‘Entrepreneurial orientation and new venture performance: the moderating role of intra- and extra-industry social capital’, Academy of Management Journal, 51 (1), 97–111. Stuart, Toby E., Ha Hoang and Ralph C. Hybels (1999), ‘Interorganizational endorsements and the performance of entrepreneurial ventures’, Administrative Science Quarterly, 44 (2), 315–49. Swan, J., M. Robertson, A. Goussevskaia, S. Newell and M. Bresnen (2007), ‘The role of institutional diﬀerences in biomedical innovation processes: a comparison of the UK and US’, International Journal of Health Technology and Management, 8 (3–4), 333–53. Szulanski, Gabriel (1996), ‘Exploring internal stickiness: impediments to the transfer of best practices within the ﬁrm’, Strategic Management Journal, 17, 27–43. Teece, D., G. Pisano and A. Shuen (1997), ‘Dynamic capabilities and strategic fit’, Strategic Management Journal, 18 (7), 509–33. Uzzi, B. (1996), ‘The sources and consequences of embeddedness for the economic performance of organizations: the network eﬀect’, American Sociological Review, 61, 674–98. Uzzi, B. (1997), ‘Social structure and competition in interﬁrm networks: the paradox of embeddedness’, Administrative Science Quarterly, 42, 35–67. Venktraman, N. and Mohan Subramaniam (2002), ‘Theorizing the future of strategy: questions for shaping strategy research in the knowledge economy’, in Andrew Pettigrew, Howard Thomas and Richard Whittington (eds), Handbook of Strategy and Management, London: Sage Publications, pp. 461–74. Walker, Gordon, Bruce Kogut and Weijan Shan (1997), ‘Social capital, structural holes, and the formation of an industry network’, Organization Science, 8 (2), 109–25. Walsh, J.P., A.D. Meyer and C.B. Schoonhoven (2006), ‘A future for organization theory: living in and living with changing organizations’, Organization Science, 17 (5), 657–71.

236

Social capital: reaching out

Wasserman, Stanley, Katherine Faust and Noshir S. Contractor (2006), ‘Testing multitheoretical, multilevel hypotheses about organizational networks: an analytic framework and empirical example’, Academy of Management Review, 31 (3), 681–703. Woolcock, Michael and Deepa Narayan (2000a), ‘Social capital: implications for development theory, research, and policy’, World Bank Research Observer, 15 (2), 225–49. Woolcock, M. and Deepa Narayan (2000b), ‘Social capital: implications for development theory research and policy revisited’, The Search for Empowerment Social Capital as Idea and Practice at the World Bank, Bloomﬁeld, CT: Kumarian Press. Yli-Renko, Helena, Erkko Antio and Harry Sapienza (2001), ‘Social capital, knowledge acquisition, and knowledge exploitation in young technology-based firms’, Strategic Management Journal, 22 (6–7), 587–613. Ziao, Zhining and Anne Tsui (2007), ‘When brokers may not work: the cultural contingency of social capital in Chinese high-tech ﬁrms’, Administrative Science Quarterly, 52 (1), 1–13.

PART III

Social capital: reaching in

9.

Social capital creation: collective identities and collective action Roderick M. Kramer

For a proper understanding of collective action, the existence of a collective identity should be recognized. (Gupta et al., 1997, p. 301)

The quality of community life depends, in no small measure, on the willingness of its members to cooperate with each other. Such cooperation can assume many forms, including sharing vital information, conserving scarce resources and contributing to public goods. When members of a community are reluctant to cooperate in such a fashion, multiple problems arise. In extreme instances, such failures of cooperation and information-sharing can have even more catastrophic and tragic results, as the 9/11 Commission Report Investigating Terrorist Attacks in the United States concluded. The importance of social capital to organizational performance is well established. However, without strong cooperation between individuals, social capital would not exist. This chapter asserts that identiﬁcation with the collective facilitates social capital creation. It explores the theoretical and empirical underpinnings for this idea and puts forth an integrated framework that shows how individuals’ social identities – and the connection of those identities with the collective – influence their willingness to contribute, or not contribute, to the reservoir of social capital available to that collective. It also draws on conceptual insights and empirical ﬁndings from several closely related streams of social psychological theory and research, including social identity theory, the Common Ingroup Identity model, self-categorization theory, and related theories pertaining to the social nature of the self.1 As described in Chapter 1 of this book, ‘What is social capital?’ by Elinor Ostrom, social capital is a form of capital that, along with physical and human capital, is created by humans. Its basic deﬁnition reflects its core elements: a set of relationships and shared values created and used by multiple individuals to solve collective problems in the present and future.2 Social capital also reflects how small to large groups interact culturally, structurally and institutionally. It describes the eﬀects these interactions 239

240

Social capital: reaching in

have on individual incentives and behavior and the resulting economic and political change (Pasotti and Rothstein, 2002). For a more detailed discussion, please consult Chapter 1. Given the signiﬁcant beneﬁts ascribed to it, it is hardly surprising that social scientists have spent considerable time identifying the antecedents of social capital. A more thorough understanding of these factors is critical if groups, institutions and societies are truly to harvest social capital’s beneﬁts (Putnam, 1993b). Considerable progress has already been made along this investigative path.3 However, one important area needs further theoretical exploration: social capital’s psychological underpinnings. These are the cognitive, motivational and aﬀective factors that influence individuals’ willingness to contribute to social capital. Because social capital is derived from the actions of individuals within networks, scholars need to know more about the conditions under which individuals will invest in the creation and maintenance of social capital. One of the most important elements in this investigation is a full appreciation of the vexing dilemma faced by individuals investing in social capital: because each beneﬁts from the existence of social capital whether or not he or she contributes to it, there are often few perceived incentives for personal investment. Individuals can easily place the responsibility on someone else as it was not ‘their job’. Even if all of the individuals in a collective recognize the value of a social network, the decision to contribute to its development and maintenance is problematic: after all, individuals can, if so inclined, ‘free ride’ on the largesse of others.4 Quickly the situation degenerates into one where no one invests in and maintains the social capital of the community. The combination of this dilemma with the importance of understanding what drives people to create social capital makes a study of its psychological underpinnings imperative. The rest of this chapter focuses on a potential driver of these choices – a form of psychological embeddedness that focuses on individuals’ social identities – and how this factor helps create social capital. When experts have considered this basic dilemma, they have often approached the problem in terms of identifying those forms of social and structural embeddedness that bind actors together to create social capital.5 Nahapiet and Ghoshal (1998) identify two important forms of such embeddedness. The ﬁrst form – structural – is the ‘impersonal conﬁguration of linkages among people or units’ (p. 244). An example is the density of network ties linking interdependent actors within a formal system (Burt, 1992, 2005). To illustrate further, James Coleman (1988) and Maureen Hallinan in Chapter 5 of this book describe this type of structural social

Social capital creation

241

capital in the dense and reinforcing relationships among parents, teachers, priests and school administrators, which Coleman calls ‘intergenerational social closure’, and which both scholars investigate to understand improved academic outcomes for Catholic school students. The second form is relational, because it reflects the ‘kinds of personal relationships people have developed over time with each other through a history of interaction’ (Nahapiet and Ghoshal, 1998, p. 244). The intimate and repeated exchanges observed within fraternal organizations and business networks exemplify this latter form.6 In Chapter 1 of this book, Ostrom illustrates the importance of relational social capital when she argues that the relative success of farmer-managed irrigation systems in Nepal is due to the evolved set of institutions or rules-in-use the farmers themselves developed over time to govern their irrigation systems. Beyond these two well-established forms of social capital, I will argue that a third form of embeddedness exists that binds actors together to create social capital. The third form is inherently more psychological. This psychological embeddedness derives from a sense of common or shared social identity among individuals.7

PSYCHOLOGICAL TIES THAT BOND AND BRIDGE: THE ROLE OF SOCIAL IDENTITIES IN SOCIAL CAPITAL CREATION Research on social identity8 and the social self9 has demonstrated the variety of identities one individual might have and the importance of these various roles for social judgment and behavior. The identities help individuals make sense of who they are, their connections to other people around them, and how they should act in social situations.10 As a result, individuals can navigate the social order more eﬀectively.11 While identities make social navigation ‘simpler’, understanding how a particular identity functions in a given context can be extremely complex. Why? Because people simultaneously possess multiple identities (Ashforth and Johnson, 2001). An individual working for the Federal Bureau of Investigation (FBI) may think of herself as a ﬁeld agent – a member of a small, elite team investigating terrorist activities at a regional level. In this situation, the main institutional contrast for her may be between the ﬁeld team and home oﬃce agents. This same agent may think of herself as a general FBI agent; in this case, the main contrast for her is between the FBI and other governmental agencies such as the US Marshals. Finally, she may think of herself as a member of a newly created Homeland Security network, making salient the ‘super-ordinate’ or collective-level identity that

242

Social capital: reaching in

includes both the FBI and US Marshals. Any one of these potentially competing organizational identities might be more meaningful at diﬀerent points. Each changes how the agent perceives herself, how she deﬁnes her obligations and entitlements, and how she construes her relationships with other professionals with whom she works. This chapter asserts that three distinct psychological identities are relevant for understanding an individual’s willingness to contribute to the creation of social capital. These are: (1) the individuals’ personal identities within the institution; (2) their subgroup or in-group identities; and (3) their super-ordinate or collective identities.12 These three types of identities often coexist. Personal identities correspond to people’s self-representations – in other words, how they think of themselves as unique, separate persons or workers. According to Brewer and Gardner (1996), the personal level of social identity reflects the ‘diﬀerentiated, individuated self-concept most characteristic of studies of the self in Western psychology’ (p. 84). For example, a policy analyst in a think-tank might think of herself as distinctive or unique because she is the only person who possesses a particular set of essential econometric skills. In contrast, individuals’ subgroup or in-group identities reflect the primary groups to which an individual belongs within the institution. The same policy analyst might think of herself as a member of the department studying economic change in developing countries. Finally, individuals’ collective identities correspond to the largest relevant organizational aggregate. In the present example, our policy analyst enjoys a common or shared identity with all of the individuals who work at that think-tank. Research suggests that one’s behavior is driven by the interactions between the type of identity that is most important at the moment, and the speciﬁc situational factors. The salience (or importance) of a given identity varies across situations.13 A multitude of situational and contextual factors often determine whether one’s personal, subgroup or collective identities are most relevant at the moment. The psychological activation of these diﬀerent self-representations produces distinct eﬀects on individuals’ perceptions of themselves and of others with whom they interact, on their motivation to act in social situations, and on their evaluative orientation toward their actions (that is, how they feel about them). When personal identities are salient, individuals tend to construe the choices available to them in relatively individualistic terms. They are likely to act in comparatively self-interested ways. In contrast, when in-group identities are foremost, individuals are likely to perceive their choices in terms of impact on the in-group as a unit. At this level of identiﬁcation, intergroup-level comparisons start to take priority because individuals think about how well their in-group is doing relative to other out-groups

Social capital creation

243

around them. Finally, when collective identity is the most relevant, individuals are likely to see themselves and their behavior in terms of impact on the collective to which they belong. Attention to personal- or in-grouplevel interests and outcomes tends to decrease. Empirical support for these propositions is substantial. Initial experimental evidence has shown that the activation of individuals’ collective identities enhanced their willingness to engage in collective behavior. Numerous laboratory studies since the 1990s have revealed that simply increasing the salience of a shared social identity – whether in-group or collective – can, by itself, enhance individuals’ willingness to create and maintain a public good across a variety of situations.14 Other studies have shown that collective identiﬁcation, when measured as an individual diﬀerence variable, also predicts such behavior.15 Field studies of investment in collective outcomes further bolster this conclusion. For example, Gupta et al. (1997) found that collective identiﬁcation with an organization was related to individuals’ levels of commitment to, and participation in, that organization’s mission. Kelly and Kelly (1994) found that one of the strongest predictors of individuals’ participation in collective action on behalf of a group was their identiﬁcation with the group. A more recent study16 (Tyler and Blader, 2000) revealed that identiﬁcation with a community or organization influences individuals’ attitudes toward their own contribution to collective resources. To summarize, numerous empirical studies show that collective identiﬁcation facilitates individuals’ willingness to invest in collective-level outcomes across contexts that range from small, experimental groups to large organizations and social groups. The aggregate results point to a single conclusion: social contextual cues which make individuals’ collective identities salient make it more likely that individuals will engage in behaviors that create and maintain social capital. These results have been observed across multiple studies that involved a wide range of dilemmas and employed a variety of diﬀerent experimental manipulations and measures of collective identity. The eﬀect emerges even when the mechanisms for making an individual’s collective identity relevant are fairly minimal. For example, either a brief group discussion 17 or ad hoc categorization into minimal groups 18 can produce the eﬀect. Explicating the Links Between Collective Identiﬁcation and Social Capital Creation Although we have discussed ample empirical evidence of the relationship between social identiﬁcation and social capital, the chapter has yet to explore why an individual’s particular self-identity might change behaviors

244

Social capital: reaching in

so dramatically. How does the process of making individuals’ collective identities salient influence their self-perceptions and/or perceptions of the collective to which they belong? Why should identiﬁcation with a particular organization or community increase individuals’ willingness to act on its behalf or to further the interests of its members? An analysis of the cognitive, motivational, aﬀective and behavioral consequences of social identiﬁcation can provide a single answer to all these questions. But to ﬁnd that answer, we need to examine the diﬀerent ways individuals think about the types of social capital available to them before we can understand how various types of identiﬁcation influence individuals to create diﬀerent forms of social capital. Depending on the type of social capital, individuals can view the costs and beneﬁts of creating social capital very diﬀerently. Actors can look at the social capital they create in at least three distinct ways: personal social capital, bonding social capital (at the ingroup level) and bridging social capital (at the collective level). They echo the three types of identity. Investments in social capital often generate individual-level or ‘personal social capital’. At this individual level, the direct beneﬁts decision-makers derive from their investments in a social network tend to be what they care about most (see Burt, 1992, 2005 for particularly detailed accounts at this level of analysis). As one Stanford University Master’s of Business Administration (MBA) student answered when asked to explain why she spent so much time building an extensive network during business school: ‘For the rest of my life, I will be able to draw on these investments.’ A second level is ‘bonding social capital’, which Putnam (2000) deﬁned at the in-group level. As Putnam argued: ‘bonding social capital is good for undergirding speciﬁc reciprocity and mobilizing solidarity’ within important, proximate subgroups to which an individual belongs (p. 22). As mentioned earlier, the intergenerational networks of teachers, parents and school administrators within a Catholic parish who systematically encourage students to strive for academic excellence represent a particularly vivid example of bonding social capital. The third level, which Putnam characterized as ‘bridging social capital’, is the form of social capital that is more inclusive of people across social distinctions. Because it is deﬁned at the collective level, Putnam argued, bridging social capital has the potential to ‘generate broader identities and reciprocity, whereas bonding social capital bolsters our narrower selves’ (p. 23). One could think of the broad network of professionals in Silicon Valley who are involved in technology start-up investing as an example of bridging social capital. Silicon Valley venture capitalists, bankers, lawyers and entrepreneurs consciously build relationships across normal industry boundaries in order to access opportunities and information quickly.

Social capital creation

245

Individuals can justify their actions as legitimate choices at any one of these three levels. For example, in a situation where personal identity is foremost, individuals can feel they are appropriately advancing and protecting their own self-interests by engaging in actions that increase the reservoir of personal social capital available to them. Thus, the MBA student described above can feel she is fairly reaping the beneﬁts of her earlier investments when she gains something from her Stanford network. When in-group identity is primary, individuals may construe their actions in terms of increasing the store of bonding social capital available to their in-group. Thus, the MBA student may feel her network-building activities beneﬁt other Stanford students within the network by allowing all of them to compete better against Harvard MBAs. Finally, in situations where a more inclusive, collective-level identity is relevant, individuals may see their actions in terms of their impact on the reservoir of bridging social capital available to everyone within the salient collective category. In the case of the MBA student, she may feel she is building a network from which all business students interested in a particular domain (for example, venture capital) may beneﬁt, irrespective of where they happened to get their MBA. Note there is no single, dominant or ‘correct’ choice for this dilemma: individuals’ view of what constitutes an appropriate or legitimate choice corresponds to the type of identity that is salient at that moment.

FROM ‘I’ TO ‘US’ TO ‘WE’: A TRANSFORMATIONAL ANALYSIS OF CHOICE BEHAVIOR IN COLLECTIVE CONTEXTS I have covered the importance of identity and the diﬀerent ways that people view social capital based on the identity that is relevant at that moment. Given my earlier conclusion about the importance of identity, the next question to answer is why activating diﬀerent levels of psychological identity influences social actors’ decision-making and behaviors. The concept of psychological transformations is a ﬁrst step on this path. Kelley (1979) developed this concept to explain why participants in experimental games often responded to the same dilemma in dramatically diﬀerent ways. In Prisoner’s Dilemma games he observed that some individuals consistently chose cooperative responses while others behaved more competitively. This occurred despite the fact that the frequent outcome for a cooperative response could be negative. To account for these diﬀerences, Kelley argued that individual players often downplay the formal structure of the games and psychologically ‘transform’ them into more personal and often highly idiosyncratic representations. Individuals’

246

Social capital: reaching in

social values and attitudes would be particularly important in influencing their choices so that, for some, cooperative choices are ‘good’ and ‘rational’, whereas for others competitive choices might seem ‘rational’ and fully justiﬁable. Although Kelley focused primarily on individual diﬀerence variables that influenced transformational tendencies, subsequent research has elaborated on the role social identities play in the transformation process.19 This work has identiﬁed at least three important and distinct categories of psychological transformation, which I characterize as cognitive, motivational and hedonic. Any one of these transformations can change the level of salient social identity as well as how the individual views outcomes that have uneven beneﬁts for diﬀerent parties (for example, themselves vs other individuals, their group vs others). Cognitive Transformations Research has identiﬁed two important and interrelated cognitive transformations. These are called ‘self-categorization eﬀects’ and ‘socialcategorization eﬀects’ (Turner, 1987). Both types of eﬀects are driven by the particular social identity that individuals adopt. Of particular psychological importance is the fact that, as individuals identify more with the ingroup and collective levels, there is a corresponding ‘shift toward the perception of [the] self as an interchangeable exemplar of some social category and away from the perception of self as a unique person’ (Turner, 1987, p. 253). That individuals’ self-categorization (that is, perception of themselves in social situations) is a result of the level of salient social identity is well documented. When personal identities are foremost, people tend to categorize themselves in terms of their unique traits and dispositions. The self is seen as a distinct entity and provides the basis for self-construal and selfevaluation. When individuals categorize themselves at the in-group level, they think in terms of the characteristics they share with other members and also how they are diﬀerent from people who are not members of their in-group (so-called ‘out-group’ members). They compare actions or inactions against what a ‘good’ (that is, prototypic) member does or does not do. Finally, when individuals categorize themselves with collective-level social identities, they think in terms of the characteristics they have in common with the collective as a whole. The level of active social identity also influences how individuals socially categorize other individuals with whom they interact. When personal-level identities are dominant, individuals tend to perceive others as distinct individuals and make contrasts between ‘me’ and ‘you’ when evaluating

Social capital creation

247

outcomes. When identiﬁcation shifts to the in-group level, individuals categorize others based on their stereotypes regarding their own in-group and the various out-groups with which they are competing for resources. They contrast ‘us’ with ‘them’. Finally, when identiﬁcation rises to the collective level, individuals tend to perceive other people in terms of the entire collective. As a result, others are evaluated in terms of the prototype or stereotype of the collective. Motivational Transformations Various levels of social identity aﬀect individuals’ social motivational orientation.20 Social motivational orientation is the psychological weight or value that individuals assign to their own outcomes versus those for others in situations where the outcomes are interdependent.21 When personal-level identities are foremost, individuals’ motivational orientation tends to be around their own self-interests: for example, the personal social capital they accrue. At the in-group level, motivational orientation focuses on advancing and protecting the in-group’s welfare or interests by increasing the reservoir of bonding social capital available to the group. Finally, when identity moves to the collective level, motivational orientation is dominated by how individual actions aﬀect the collective welfare. Thus, the impact of actions or inactions on the reservoir of bridging social capital would be most relevant. In discussing why social identity influences individuals’ views of an outcome’s value, Brewer (1979) theorized that one consequence of moving from the personal to the in-group and collective levels of identity is a ‘reduced diﬀerentiation between one’s own and others’ outcomes’ which allows for ‘increasing the weight given to collective outcomes in individual decision-making’ (p. 322). Along similar lines, Mael and Ashforth (1992, p. 103) suggested that increasing the relevance of collective group boundaries helps foster a perceived oneness between individuals’ own fate and the collective outcome – to perceive a common fate; in short, taking us, metaphorically speaking, from ‘I’ to ‘us’ to ‘we’. The increase in attraction among group members that accompanies social identiﬁcation with the group (Hogg and Abrams, 1988) could be another motivation for individuals to engage in collective behavior. Because this attraction is based speciﬁcally on awareness of shared membership, this form of social attraction can be distinguished from interpersonal attraction, ‘which is inter-individual attraction based upon idiosyncratic preferences and ﬁrmly rooted in close personal relationships’ (Hogg and Abrams, 1988, p. 107). As a consequence, individuals are less likely to draw sharp distinctions between their own outcomes and the outcomes obtained by others in the group. Several studies support this line of reasoning.22

248

Social capital: reaching in

Research suggests one other reason why increasing individuals’ collective identity might enhance their motivation to engage in collective behavior. A large body of social psychological research has shown that people’s behavior is often influenced by self-presentational concerns (Leary, 1995). As Leary noted, self-presentational concerns reflect many motives, including the desire to maintain positive personal and social identities and to maintain positive emotions about oneself. Research in this area aﬃrms a simple but important point: in many social situations, people care greatly about what others think of them. In elaborating on the self-presentational aspects of choice in social dilemma situations, Kelley (1979) argued that individuals’ behavior in such situations is important not only in its eﬀect on the material outcomes they obtain (for example, the monetary pay-oﬀs in an experimental game), but also because of a variety of less tangible but no less important psychological and social outcomes. In particular, Kelley noted, behavior in such situations aﬀords individuals an opportunity to show others with whom they are interacting important dispositions and cherished identities; for instance, that they are loyal, trustworthy and cooperative group members. From this perspective, displays of collective behavior provide a way for individuals to aﬃrm the value they associate with membership in the collective (see Lind and Tyler, 1988). They are, in other words, one important mechanism for a sort of ‘collective self-aﬃrmation’ – a group-level motive not unlike individuallevel forms of self-aﬃrmation observed in other contexts (see Steele, 1999). Thus, engaging in collective behavior enables group members to communicate to others the symbolic importance they attach to their shared identity. As March (1994) noted in this regard, social identities help individuals ‘establish and celebrate their ties with others and their place in a social order of relationships that they honor’ (p. 63). Evidence in support of such self-presentational motives comes from several studies. Kramer et al. (1993), for example, examined selfpresentational concerns in a two-person bargaining situation. They found that when a shared social identity was made important to negotiators, the negotiators’concerns about appearing fair and cooperative increased. In contrast, when individual-level or diﬀerentiating identities were reinforced, individuals tended to be more concerned about their own performance (how well they personally negotiated, had outperformed the other party, and so on). Hedonic or Aﬀective Transformations Hedonic or aﬀective transformations are the third link between collective identity and the willingness and likelihood of engaging in social capitalcreating behaviors. One’s level of social identity influences the particular

Social capital creation

249

reference points used to evaluate the hedonic beneﬁts of actions (for example, the net gains or losses in pleasure or satisfaction associated with choices). In the past few years, there has been a great deal of interest in the relationship between hedonic reference points and individuals’ assessments of their subjective well-being.23 This research suggests that diﬀerent reference points influence people’s perceptions of the hedonic consequences associated with their choices (Elster and Loewenstein, 1992). These points aﬀect whether people feel good or bad about the outcomes they obtain, as well as the level of satisfaction or regret they associate with their actions (Gilovich et al., 1995). Related research on hedonic self-regulation suggests that people are sometimes able to anticipate or forecast the hedonic states associated with diﬀerent choices.24 Elster and Loewenstein (1992) framed this competence in terms of the notion of ‘hedonic anticipation’. ‘Like memory’, they argued, ‘anticipated experiences aﬀect current utility through the consumption [eﬀect] . . . [enabling us] in eﬀect to consume events before they occur’ (p. 225). This human capacity to enjoy anticipated events, they went on to posit, ‘acts as multipliers of experience, causing individuals to experience the hedonic impact of events repeatedly before they occur’ (p. 225). This research implies that individuals’ level of identity can strongly aﬀect their hedonic reference points and anticipation of an event – and thus choices. When identiﬁcation moves from the personal to the in-group to the collective level, there are corresponding changes in the hedonic reference points individuals use to evaluate anticipated aﬀective states and outcomes (literally, how good they feel about themselves and those outcomes). When identiﬁcation is at the personal level, the reference point for evaluating good versus bad outcomes tends to be the self: how much (hedonically) do I, as an individual, gain or lose by this action? When in-group-level identities are relevant, hedonic outcomes are deﬁned at the in-group level: how have I helped or hurt my group and therefore how do I feel? Finally, when collective identity is salient, the hedonic reference point is anchored around collective gains or losses (that is, how my actions or inactions have helped or hurt the collective). If individuals perceive collective behavior as a way of aﬃrming positive relationships with other members, they are also likely to see these aﬃrmations as intrinsically pleasurable and therefore hedonically selfrewarding or reinforcing. As Simon (1991) observed: ‘Identiﬁcation with the “we”, which may be a family, a company, a city, a nation, or the local baseball team, allows individuals to experience satisfactions (to gain utility) from successes of the unit thus selected’ (p. 36, emphasis added). Consistent with Simon’s argument, there is some evidence that

250

Social capital: reaching in

decision-makers’ satisfaction with outcomes is influenced, at least in part, by the level of social identity that is foremost when they evaluate those outcomes.25 This same logic implies that, when collective identities are relevant to individuals, they are more likely to anticipate that socially unacceptable choices will lead to negative hedonic outcomes, including feelings of guilt, shame or fear (especially of being labeled disloyal or of social sanction or ostracism). For example, the self-perception that one is free-riding on the eﬀorts of others could be potentially guilt-inducing. A fuller exploration of the hedonic consequences of choice might clarify some of the intense aﬀective reactions people sometimes experience in choice dilemmas, with respect to both their own behavior and, even more dramatically, the behavior of others. As numerous studies have shown, individuals ﬁnd the prospect of getting the so-called ‘sucker’s pay-off’ (that is, cooperating with others when others do not reciprocate) highly aversive. Non-trivial levels of psychological distress such as anger, irritation and even mild depression about human nature are sometimes observed in experimental games involving students who do not know each other, have not seen each other (these games are often played over linked computer terminals), and that involve very small monetary amounts (a few dollars at most). These reactions make little rational or logical sense. However, from the standpoint of social identity, these intense subjective reactions are entirely comprehensible. Social identities can engender a form of what Rotter (1980) termed ‘moralistic trust’ – trust viewed as a duty or obligation for individuals to engage in trustworthy actions. Rotter noted that trust behavior sometimes reflects an individual’s ‘belief in the moral rightness of trust [rather than] an expectancy of risk in trusting others’ (p. 4). In other words, people sometimes engage in trust-related actions not simply or only because they expect to beneﬁt from such actions, but rather because they think they ought to engage in such behaviors. Although Rotter conceptualized moralistic trust as an individual diﬀerence variable, a social identity-based logic emphasizes the notion that such trust is predicated upon individuals’ beliefs regarding what it means to be a good member of a group, including being virtuous, honorable, loyal, and making contributions. As March (1994) aptly suggested in this regard, individuals ‘can violate a logic of consequences and be considered [merely] stupid or naive, but if they violate the moral obligations of identity, they will be condemned as lacking in elementary virtue’ (p. 65). Consistent with this argument, there is some evidence that moralistic trust is stronger in individuals with higher levels of group identiﬁcation and leads to more intense reactions to perceived trust violations, including aﬀective reactions and retaliatory aggression (Kramer, 1999).26

Social capital creation

251

CONCLUSIONS AND IMPLICATIONS My primary aim in this chapter has been to explore some of the psychological underpinnings of social capital creation. Speciﬁcally, the chapter presents a framework that articulates cognitive, motivational and aﬀective processes that foster individuals’ willingness to contribute to the creation and maintenance of social capital. Recent theory and research have converged on the view that social capital resides in the social connections or networks linking interdependent actors and thus constitutes a valuable collective resource. Social capital, however, does not emerge as a bolt from the blue or simply fall from the sky. It must be created, nurtured and maintained. A social identity-based framework contributes to social capital theory in that it helps explain why and how it is created, nurtured and maintained. An identity-based analysis revealed that creating a collective culture could increase the level of social capital associated with a group or network by activating and validating individuals’ thinking and acting in terms of their collective identities. Several lines of empirical evidence support this contention. First, evidence from social psychological research suggests that heightened perceptions of ‘common fate’ contribute to the emergence of collective identities. As Brewer and Miller (1996) noted in reviewing this evidence: ‘perceptions of shared interests or common fate do seem to be crucial to the emergence of collective identity’ (p. 43). In support of this argument, they discussed Gurin and Townsend’s (1986) ﬁnding that a ‘sense of common fate . . . proved to be the most important predictor of collective orientation’ (p. 43). Research has also demonstrated that recategorizing groups in terms of collective-level boundaries can help override or attenuate the impact of competitive, intergroup orientations.27 In the parlance of the social capital theorist, this recategorization helps pave the way for a transition from bonding to bridging social capital. Other recent research suggests that even relatively subtle linguistic cues may play a powerful role in evoking collective identities and enhancing collective behavior. The use of collectivizing language, such as ‘we’ instead of ‘I’, may foster collective orientations and behaviors.28 Another approach is to draw attention to and/or increase the relevance of collective identities through structural network arrangements. As March (1994) observed in this regard, social structures and arrangements often function much like ‘stage managers’, providing ‘prompts that evoke particular identities in particular situations’ (p. 72). This identity-based logic of judgment and choice contributes to our understanding of the relationship between trust and social capital. Social capital theorists have frequently argued that trust constitutes one of the

252

Social capital: reaching in

important consequences of social capital.29 As Putnam (1993a, p. 169) crisply put it, social networks ‘allow trust to become transitive and spread’. It has been less obvious or clear, however, just what is the nature of trust that is involved. The concerns about trust that arise in collective contexts are quite diﬀerent from those that arise in other social settings; they can also be more problematic. The grounds on which trust in others is predicated in collective contexts are less clearly articulated than the grounds for trust associated with more intimate settings, such as interpersonal relationships or small groups. Trust can also be harder to develop; because of the greater size, social heterogeneity and structural complexity of large collectives, individuals do not have the opportunity to engage in the sort of incremental and repeated exchanges that have been shown to facilitate the development of interpersonal or dyadic trust (Lindskold, 1978; Rotter, 1980). Similarly, many of the informal social mechanisms that foster trust development within small, homogeneous groups obviously lose their eﬃcacy in larger collectives (Olson, 1965). Consequently, models of interpersonal trust development (Rotter, 1971) and trust development in small groups (Golembiewski and McConkie, 1975) are often not applicable. As Putnam (1993a) aptly concluded, therefore: ‘In larger, more complex settings, a more impersonal or indirect form of trust is required’ (p. 171). Identity-based trust can be a candidate for this type of trust. When predicated explicitly upon awareness of shared membership in a network, such ‘depersonalized trust’ (Brewer, 1981) may be particularly eﬀective. Professors within Stanford University’s Business School share a collective identity. When this collective identity is most salient, my colleagues and I can build a depersonalized trust even though we do not all know one another and nor have we shared the repeated exchanges normally needed to foster dyadic trust. Because it represents a generalized expectancy predicated on a positive in-group stereotype, identity-based trust can function as a substitute or proxy for other processes on which generalized social trust is usually contingent (Yamagishi and Yamagishi, 1994). Because this type of trust is conferred on the basis of shared membership in a collective, individuals may perceive less need to establish individual trust before engaging in exchanges or transactions with other members. Furthermore, trust of this sort is less likely to depend on the expectation that any speciﬁc member of the network will act in a trustworthy fashion. Paraphrasing Uzzi (1997), individuals can ‘roll over’ their experiences from one member to another. In this respect, such depersonalized trust operates as a form of what Messick and Kramer (2001) characterized as presumptive or ‘shallow’ trust. This level of trust enables individuals to engage more readily in the sort of spontaneous and extended actions that generate social capital. Another factor that may help sustain this depersonalized trust is that

Social capital creation

253

individuals may manifest a ‘leniency bias’ to other members of the collective (Brewer, 1996): when in doubt, they will give other members the beneﬁt of the doubt. Filling in the many remaining gaps in social capital theory continues to be an important enterprise. For a comprehensive theory of social capital, the challenge lies in its micro foundations. At least some of the details must include the psychological mechanisms that contribute to social capital creation. Furthermore, I wholeheartedly agree with Ostrom’s call in Chapter 1 for more research on the diverse forms of social capital and then linking speciﬁc forms of social capital with speciﬁc consequences. One further unresolved question centers on the causal relationships between collective identiﬁcation and the creation of social capital. This chapter, which has focused on the link between collective identiﬁcation and social capital creation, has treated collective identiﬁcation as an independent variable. However, a strong prima facie case can be made for the argument that the creation of social capital enhances collective identiﬁcation in return. Paraphrasing Putnam (1993a, p. 171), just as trust begets cooperation and cooperation begets trust, so I would argue that collective identiﬁcation begets social capital and social capital begets collective identiﬁcation. Bridging constructs are one potential way to ﬁll some of these gaps – speciﬁcally those around the linkages among macro- and micro-level antecedents and the consequences of social capital. Cappeli and Sherer (1991) proposed that the social sciences needed conceptual models that connect or bridge micro (psychological) and macro (sociological and political) accounts of human behavior. A social identity-based model of judgment and choice provides a platform for such a bridge, allowing us to understand better how individuals resolve the inherent tensions between investing in individual versus bonding versus bridging forms of social capital. It can also help address an important issue identiﬁed by Putnam (2000): ‘how the positive consequences of social capital – mutual support, cooperation, trust, institutional eﬀectiveness – can be maximized and the negative manifestations – sectarianism, ethnocentrism, corruption – minimized’ (p. 22). Indeed, a fully developed contingency theory linking levels of salient social identity to social capital-creating or undermining behaviors would help us understand what might happen in various real-world collective contexts. In closing, I would emphasize that an elaboration of the social psychological foundations of social capital does more than simply add to an already long and impressive list of variables related to its emergence. A richer and fuller account of the micro-foundations of social capital would enhance our understanding of the distinctive behaviors associated with its

254

Social capital: reaching in

creation in diﬀerent kinds of networks and the particular consequences those behaviors generate. Because networks, and their intra-group dynamics, are not all alike, a social psychological framework might also contribute to a more comprehensive account of social capital.

NOTES 1. 2.

3. 4. 5. 6. 7.

8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18. 19. 20.

Sedikides and Brewer (2001), Tyler et al. (1999). Considerable attention has been aﬀorded to the issue of how best to deﬁne social capital (see, for example, Burt, 2005; Cook, 2005; Nan, 2002). In this chapter, I follow Putnam’s (2000) lead, positing that social capital refers speciﬁcally to ‘connections among individuals – social networks and the norms of reciprocity and trustworthiness that arise from them’ (p. 19, emphasis added). From the standpoint of this deﬁnition, cooperation and other forms of collective behavior constitute important consequences of social capital. Burt (2005), Cook (2005), Putnam (2000). Hardin (1968), Latane (1986), Olson (1965). Granovetter (1985), Nahapiet and Ghoshal (1998). For example, Boissevain (1974), McEvily and Zaheer (2004), Yamagishi and Sato (1986), Uzzi (1997). There are obviously many other bases one might invoke in trying to explain individuals’ willingness to invest in the creation of social networks, including altruistic motives and even self-interested motives. Because of space limitations, I will not discuss these various motives, or the evidence in favor of them, further. For example, Abrams and Hogg (1990), Brewer and Kramer (1986), Hogg and Abrams (1988), Kramer and Brewer (1984). For example, Brewer (1991), Brewer and Gardner (1996), Simon (1999) Tyler and Blader (2000). Brewer (1991), Deaux et al. (1999). Deaux et al. (1999), Simon (1999). As one of the reviewers helpfully pointed out, Brewer and her colleagues (for example, Brewer and Gardner, 1996; Sedikides and Brewer, 2001) have emphasized the interpersonal or relational self as an intermediate level of self-construal between the individual and collective selves. The notion of a relational self makes considerable sense in the context of many forms of interpersonal interdependence, such as dyadic forms of mixedmotive conflict. In the context of institutions and other work-related hierarchies, however, I would argue that subgroup identiﬁcations play a more prevalent and important role in helping individuals locate themselves socially, especially in terms of deﬁning their group-level loyalties, obligations and sense of entitlements (see, for example, Brewer and Kramer, 1986). For example, Cota and Dion (1986), McGuire and Padawer-Singer (1976). For example, Brewer and Kramer (1986), Brewer and Schneider (1990), Dawes et al. (1988), De Cremer and Van Vugt (1995), Gaertner et al. (1989), Kramer and Brewer (1984, 1986). Deaux et al. (1999), Kramer and Goldman (1995), Piliavin (1990). For example, Tyler and Blader (2000), Tyler and Degoey (1996). Caporael et al. (1989) for a review. Brewer and Kramer (1986), Brewer and Schneider (1990), Kramer and Brewer (1984), Kramer et al. (1993). See for example, De Cremer and Van Vugt (1999), Kramer and Brewer (1986), Kramer and Goldman (1995), Sellers et al. (1998). Although early research treated these social motivational orientations largely as individual diﬀerence variables (on the assumption that they reflected fairly stable values and

Social capital creation

21. 22. 23. 24. 25.

26.

27. 28. 29.

255

attitudes), more recent studies have emphasized the extent to which such orientations are responsive to social contextual cues that, in eﬀect, activate or make more salient diﬀerent social motivational orientations (Brewer and Kramer, 1986; De Cremer and Vugt, 1999; Kramer and Goldman, 1985). These studies demonstrated, for example, that the level of social identity made salient to decision-makers influences whether their behavior reflects individualistic versus cooperative orientations. Kelley (1979), McClintock and Liebrand (1988). Hogg and Hardie (1991). Kahneman et al. (1999). For example, Ainslie and Haslam (1992), Gilbert (2002), Schelling (1984, 1992). Kramer et al. (1993), Thompson et al. (1995). Research on the well-known ‘basking in reflected glory’ phenomenon (Cialdini et al., 1976) provides further suggestive evidence of this link between positive social identiﬁcations (for example, being a member of a school with a winning football team) and positive hedonic states (feeling personal pride or deriving pleasure). A further, rather interesting possibility suggested by this thread of reasoning is that individuals from whom collective-level identities are salient may sometimes be willing to engage in compensatory actions to make up for the comparatively selﬁsh or noncooperative actions of others in the collective (Brewer and Kramer, 1984). For example, Gaertner et al. (1989), Kramer and Brewer (1984). Brewer and Gardner (1996), Kramer et al. (1993). For example, Coleman (1990), Fukuyama (1995), Granovetter (1985), Putnam (1993a, 1993b), Uzzi (1997).

REFERENCES Abrams, D. and M.A. Hogg (1990), Social Identity Theory, New York: SpringerVerlag. Ainslie, G. and N. Haslam (1992), ‘Self-control’, in G. Loewenstein and J. Elster (eds), Choice Over Time, New York: Russell Sage. Ashforth, B.E. and S.A. Johnson (2001), ‘Which hat to wear? The relative salience of multiple identities in organizational contexts’, in M.A. Hogg and D.J. Terry (eds), Social Identity Processes in Organizational Contexts, Philadelphia, PA: Psychology Press, pp. 31–48. Boissevain, J. (1974), Friends of Friends, Oxford: Basil Blackwell. Brewer, M.B. (1979), ‘In-group bias in the minimal intergroup situation: a cognitive-motivational analysis’, Psychological Bulletin, 86, 307–24. Brewer, M.B. (1991), ‘The social self: on being the same and diﬀerent at the same time’, Personality and Social Psychology Bulletin, 17, 475–82. Brewer, M.B. (1996), ‘In-group favoritism: the subtle side of intergroup discrimination’, in D.M. Messick and A. Tenbrunsel (eds), Codes of Conduct: Behavioral Research and Business Ethics, New York: Russell Sage, pp. 101–17. Brewer, M.B. and W. Gardner (1996), ‘Who is this “we”? Levels of collective identity and self representations’, Journal of Personality and Social Psychology, 71, 83–93. Brewer, M.B. and R.M. Kramer (1986), ‘Choice behavior in social dilemmas: eﬀects of social identity, group size, and decision framing’, Journal of Personality and Social Psychology, 50, 543–9. Brewer, M.B. and N. Miller (1996), Intergroup Relations, Paciﬁc Grove, CA: Brooks-Cole Publishing Co.

256

Social capital: reaching in

Brewer, M.B. and S. Schneider (1990), ‘Social identity and social dilemmas: a double-edged sword’, in D. Abrams and M.A. Hogg (eds), Social Identity Theory: Constructive and Critical Advances, New York: Springer-Verlag, pp. 169–8. Burt, R.S. (1992), Structural Holes, New York: Cambridge University Press. Burt, R.S. (2005), Brokerage and Closure: An Introduction to Social Capital, New York: Oxford University Press. Cappelli, P. and P.D. Sherer (1991), ‘The missing role of context in OB: the need for a meso-level approach’, in L.L. Cummings and B.M. Staw (eds), Research in Organizational Behavior, Vol. 13, Greenwich, CT: JAI Press, pp. 55–110. Cialdini, R., R. Borden, A. Thorne, M. Walkder, S. Freman and L. Sloan (1976), ‘Basking in reflected glory: three (football) ﬁeld studies’, Journal of Personality and Social Psychology, 34, 463–76. Coleman, J. (1990), Foundations of Social Theory, Cambridge, MA: Harvard University Press. Cook, K.S. (2005), ‘Networks, norms, and trust: the social psychology of social capital’, Social Psychology Quarterly, 68, 4–14. Cota, A.A. and K.L. Dion (1986), ‘Salience of gender and sex composition in ad hoc groups: a test of distinctiveness theory’, Journal of Personality and Social Psychology, 50, 770–76. De Cremer and Van Gugt (1999), ‘Social identiﬁcation eﬀects in social dilemmas: a transformation of motives’, European Journal of Social Psychology, 29, 871–93. Deaux, K., A. Reid, K. Mizrahi and D. Cotting (1999), ‘Connecting the person to the social: the functions of social identiﬁcation’, in T.R. Tyler, R.M. Kramer and O.P. John (eds), The Psychology of the Social Self, New York: Routledge, pp. 91–113. Elster, J. and G. Loewenstein (1992), ‘Utility from memory and anticipation’, in G. Loewenstein and J. Elster (eds), Choice over Time, New York: Russell Sage, pp. 213–34. Fukuyama, F. (1995), Trust: The Social Virtues and the Creation of Prosperity, New York: Free Press. Gaertner, S.L., J. Mann, A. Murrell and J. Dovidio (1989), ‘Reducing intergroup bias: the beneﬁts of recategorization’, Journal of Personality and Social Psychology, 57, 239–49. Gilbert, D.T. (2002), ‘Decisions and revisions: the aﬀective forecasting of changeable outcomes’, Journal of Personality and Social Psychology, 82, 503–14. Gilovich, T and V.U. Medvec (1995), ‘The experience of regret: what, when, and why’, Psychological Review, 102, 379–95. Golembiewski, R.T. and M. McConkie (1975), ‘The centrality of interpersonal trust in group processes’, in C.L. Cooper (ed.), Theories of Group Processes, London: John Wiley & Sons, pp. 131–85. Granovetter, M. (1985), ‘Economic action and social structure: the problem of embeddedness’, American Journal of Sociology, 91, 481–510. Gupta, D.K., C.R. Hofstetter and T.F. Buss (1997), ‘Group utility in the micro moti-vation of collective action: the case of membership in the AARP’, Journal of Economic Behavior and Organization, 32, 301–20. Gurnin, P. and A.Townsend (1986), ‘Properties of gender identity at their implications for gender consciousness’, British Journal of Psychology, 25,139–48. Hardin, G. (1968), ‘The tragedy of the commons’, Science, 162, 1243–8.

Social capital creation

257

Hogg, M.A. and D. Abrams (1988), Social Identiﬁcations: A Social Psychology of Intergroup Relations and Group Processes, London: Routledge. Hogg, M.A. and E.A. Hardie (1991), ‘Self-attraction, personal attraction and selfcategorization: a ﬁeld study’, Personality and Social Psychology Bulletin, 17, 175–80. Kahneman, D., D. Diener and N. Schwarz (1999), Well-being: The Foundations of Hedonic Psychology, New York: Russell Sage Foundation. Kelley, H.H. (1979), Personal Relationships, Hillsdale, NJ: Lawrence Erlbaum. Kelly, C. and J.E. Kelly (1994), ‘Who gets involved in collective action? Social psychological determinants of individual participation in trade unions’, Human Relations, 47, 63–87. Kramer, R.M. (1999), ‘Aﬀective reactions to trust violations: a social identity perspective’, paper presented at the annual Academy of Management Meetings. Kramer, R.M. and M.B. Brewer (1984), ‘Eﬀects of group identity on resource use in a simulated commons dilemma’, Journal of Personality and Social Psychology, 46, 1044–57. Kramer, R.M. and M.B. Brewer (1986), ‘Social group identity and the emergence of cooperation in resource conservation dilemmas’, in H.A.M. Wilke, D.M. Messick and C.G. Rutte (eds), Experimental Social Dilemma, Frankfurt: Verlag Peter Lang, pp. 357–89. Kramer, R.M. and L. Goldman (1995), ‘Helping the group or helping yourself ? Social motives and group identity in social dilemmas’, in D. Schroeder (ed.), Social Dilemmas, New York: Praeger. Kramer, R.M., P. Pommerenke and E. Newton (1993), ‘The social context of negotiation: eﬀects of social identity on negotiator judgment and decision making’, Journal of Conflict Resolution, 37, 633–54. Latane, B. (1986), ‘Responsibility and eﬀort in organizations’, in P.S. Goodman (ed.), Designing Eﬀective Work Groups, San Francisco, CA: Jossey-Bass, pp. 277–304. Leary, M.R. (1995), ‘Self-Presentation: Impression Management and Interpersonal Behaviour, Boulder, CO: Westview. Lind, E.A. and T.R. Tyler (1988), ‘The Social Psychology of Procedural Justice, New York: Plenum. Lindskold, S. (1978), ‘Trust development, the GRIT proposal, and the eﬀects of conciliatory acts on conflict and cooperation’, Psychological Bulletin, 85, 772–93. Mael, F. and B.E. Ashforth (1992), ‘Alumni and their alma mater: a partial test of the reformulated model of organizational identiﬁcation’, Journal of Organizational Behavior, 13, 103–23. March, J.G. (1994), A Primer on Decision Making, New York: Free Press. March, J.G. and J.P. Olsen (1989), Rediscovering Institutions: The Organizational Basis of Politics, New York: Free Press. McClintock, C.G. and W.B.G. Liebrand (1988), ‘Role of interdependence structure, individual value orientation, and another’s strategy in social decision making: a transformational analysis’, Journal of Personality and Social Psychology, 55, 396–409. McEvily, B. and A. Zaheer (2004), ‘Architects of trust: the role of network facilitators in geographical clusters’, in R.M. Kramer and K.S. Cook (eds), Trust and Distrust in Organizations, New York: Russell Sage Foundation, pp. 189–214. McGuire, W.J. and A. Padawer-Singer (1976), ‘Trait salience in the spontaneous self-concept’, Journal of Personality and Social Psychology, 33, 743–54.

258

Social capital: reaching in

Messick, D.M. and R.M. Kramer (2001), ‘Trust as a form of shallow morality’, in K.S. Cook (ed.), Trust in Society, New York: Russell Sage Foundation, pp. 89–118. Nahapiet, J. and S. Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23, 242–66. Nan, Lin (2002), Social Capital: A Theory of Social Structure and Action, Cambridge: Cambridge University Press. Olson, M. (1965), The Logic of Collective Action, New Haven, CT: Yale University Press. Pasotti, E. and B. Rothstein (2002), ‘In the market for ideas: a quest for a uniﬁed conceptual approach in political science’, paper presented at the annual meeting of the American Political Science Association, 29 August–1 September 2002, Boston, MA. Piliavin, J. (1990), ‘Why do they give the gift of life? A review of research on blood donors since 1977’, Transfusion, 30, 444–59. Putnam, R. (1993a), Making Democracy Work, Princeton, NJ: Princeton University Press. Putnam, R.D. (1993b), ‘The prosperous community: social capital and public life’, America, 4, 1–13. Putnam, R.D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Rotter, J.B. (1971), ‘Generalized expectancies for interpersonal trust’, American Psychologist, 26, 443–52. Rotter, J.B. (1980), ‘Interpersonal trust, trustworthiness, and gullibility’, American Psychologist, 35, 1–7. Schelling, T.C. (1984), ‘The intimate contest for self-command’, in T.C. Schelling (ed.), Choice and Consequence, Cambridge, MA: Harvard University Press. Schelling, T.C. (1992), ‘Self-command: a new discipline’, in G. Loewenstein and J. Elster (eds), Choice over Time, New York: Russell Sage, pp. 167–76. Sedikides, C. and M.B. Brewer (2001), Individual Self, Relational Self, and Collective Self, Philadelphia, PA: Psychology Press. Sellers, R.M., M.A. Smith, N. Shelton, S. Rowley and T.M. Chavous (1998), ‘Multidimensional model of racial identity: a reconceptualization of African American racial identity’, Personality and Social Psychology Review, 2, 18–39. Simon, B. (1999), ‘A place in the world: self and social categorization’, in T.R. Tyler, R.M. Kramer and O.P. John (eds), The Psychology of the Social Self, Mahwah, NJ: Erlbaum, pp. 47–69. Simon, H.A. (1991), ‘Organizations and markets’, Journal of Economic Perspectives, 5, 34–8. Steele, C. (1999), ‘The psychology of self-aﬃrmation: sustaining the integrity of the self’, in R.F. Baumeister (ed.), The Self in Social Psychology, New York: Psychology Press, pp. 372–90. Thompson, L., K. Valley and R.M. Kramer (1995), ‘The bittersweet feeling of success: an examination of social perception in negotiation’, Journal of Experimental Social Psychology, 31, 467–92. Turner, J.C. (1987), Rediscovering the Social Group: A Self-Categorization Theory, Oxford: Basil Blackwell. Tyler, T.R. and S.L. Blader (2000), Cooperation in Groups: Procedural Justice, Social Identity, and Behavioral Engagement, Philadelphia, PA: Psychology Press.

Social capital creation

259

Tyler, T.R. and P. Degoey (1996), ‘Collective restraint in social dilemmas: procedural justice and social identiﬁcation eﬀects on support for authorities’, Journal of Personality and Social Psychology, 69, 482–97. Tyler, T., R.M. Kramer and O. John (1999), The Psychology of the Social Self, Englewood Cliﬀs, NJ: Erlbaum. Uzzi, B. (1997), ‘Social structure and competition in interﬁrm networks: the paradox of embeddedness’, Administrative Science Quarterly, 42, 35–67. Yamagishi, T. and K. Sato (1986), ‘Motivational bases of the public goods problem’, Journal of Personality and Social Psychology, 50, 67–73. Yamagishi, T. and K. Yamagishi (1994), ‘Trust and commitment in the United States and Japan’, Motivation and Emotion, 18, 129–66.

10.

Connecting to brokers: strategies for acquiring social capital Daniel J. Brass

It is not wise to be the ﬁrst or the second when it is the third who beneﬁts. This is not a temporal sequence; rather, I am referring to advantages of the tertius gaudens – the third who beneﬁts from the disunion of the other two (Simmel, 1950). Research in organizations has shown that it is wise to be the broker – the third who connects two disconnected actors (for example, Brass et al., 2004; Burt, 2005; Fernandez-Mateo, 2007). The lack of connection between the ﬁrst and second is often referred to as a ‘structural hole’. As Burt (2005) has argued, the broker is in the position to access and control the information flow between the two disconnected people or groups and acquire social capital. At the individual level of analysis, the advantages of structural holes translate into power (Brass, 1984), better performance (Mehra et al., 2001), promotions (Brass, 1984; Burt, 1992), career success (Seibert et al., 2001) and creativity (Brass, 1995; Burt, 2004). Studies at the interorganizational level of analysis also suggest that social capital accrues to the broker organization. Brokerage has been related to ﬁrm survival (Koput and Powell, 2003; Oh et al., 2006), innovation (Stuart and Podolny, 1999; McEvily and Zaheer, 1999), market share (Rowley and Baum, 2004) and performance (Provan and Milward, 1995). Why then would anyone want to be the ﬁrst or the second? This chapter examines the question of why we use brokers – people who connect two other people who are themselves not connected – if the broker is the one who primarily beneﬁts from the exchange. The broker wins because he can access and control the information flow between the two other actors. Risk is created by the possibility that the broker may engage in opportunistic behavior, playing you oﬀ against the other party. Connecting to brokers carries with it this vulnerability. Yet, there are many everyday examples of connecting to brokers. We seek out real-estate agents to buy or sell our homes; start-up businesses seek venture capitalists; we go to banks for loans, lawyers for advice, and Consumer Reports for information. All retail outlets broker between customers and wholesalers, yet 260

Connecting to brokers

261

we frequent them. And millions of people have connected to perhaps the most successful broker of recent times: eBay. In seeking to determine why we use brokers, this chapter explores the ramiﬁcations of two strategies for developing social capital: the closure strategy, which focuses on connections with friends of friends; and the broker strategy, where people connect to diverse, disconnected others. With whom should we connect? What is the best strategy for acquiring social capital? By examining these two approaches, as well as their pros and cons, we may draw insights as to how to build social capital as an individual, an organization and a society.

SOCIAL CAPITAL We must ﬁrst deﬁne social capital – a task that has received much attention and produced multiple deﬁnitions. This chapter adopts a multilevel deﬁnition: ‘The sum of the actual and potential resources embedded within, available through, and derived from the network of relationships possessed by an individual or social unit’ (Nahapiet and Ghoshal, 1998, p. 243). This deﬁnition encompasses the individual approach with its focus on accessing and controlling resources exchanged through relationships with others (as illustrated by the above references to the advantage of the broker position). It also includes the group-level approach with its focus on the collective beneﬁts of relationships in providing social capital in the form of norms, trust and reciprocity (Coleman, 1990; Putnam, 1995). Neither approach forgoes the other entirely, but the diﬀerence in deﬁnitional levels of analysis is ampliﬁed by seemingly contradictory predictions concerning the acquisition of social capital. At the individual level, connecting to disconnected others results in social capital; at the group level, connecting to others who are connected results in closure in the network and the social capital associated with trust, norms and group sanctions. The diﬀerence is in the focus on individual versus collective goods. For more information about the breadth of research on social capital’s deﬁnitions, see Ostrom in Chapter 1 of this book. The group-level deﬁnitions of social capital, which focus on norms, can present an ‘oversocialized’ view of human behavior that leaves little room for strategic action. Equally unlikely is the ‘undersocialized’ view that individuals act in isolation, disregarding the impact of social structure (Granovetter, 1985). Rather, I adopt a more multilevel deﬁnition that allows for strategic action without disregarding social structure. Incorporating both levels of analysis is essential because, as Lin (2001, p. 8) points out: ‘Whether social capital is seen from the societal-group level or the relational

262

Social capital: reaching in

(individual) level, all scholars remain committed to the view that it is the interacting members who make the maintenance and reproduction of this social asset possible.’ My focus in this chapter is on the strategies individuals might engage in to acquire social capital. Individual actions result in relationships that aggregate into social structure. While social structure may enhance or constrain individual actions, there is little doubt that such structure also arises from these actions. I view social capital, like other forms of capital, from an investment perspective with the expectation of future (often uncertain) beneﬁts (Adler and Kwon, 2002).

SOCIAL NETWORKS I approach the issue of social capital from a social network perspective. A social network is a set of three or more actors and ties representing some relationship, or lack of relationship, among the actors. Although social networks are built upon the dyadic interactions between actors, one of the basic assumptions of social network analysis is that dyadic relationships are aﬀected by third parties. For example, the dyadic relationship between a husband and wife is aﬀected by the birth of a child. A second child further aﬀects the family’s relationships, and we could easily add other actors to this family network (for example, mother-in-law). A lack of a relationship may aﬀect actors in the network. For example, a divorce between the parents aﬀects the children. Thus, we suggest that actors are embedded within a web or network of relationships with other actors. The focus is on the relationships, and the pattern of relationships, rather than on the attributes of the actors. It is this network of relationships, as the above deﬁnition notes, that is the basis for social capital. How then does one strategically build networks and how do strategic actions aﬀect network structure and social capital? We begin with the closure strategy, followed by the broker strategy.

THE CLOSURE STRATEGY The closure strategy, as advanced by Coleman (1990) and others, involves connecting to an acquaintance or friend of a friend; in other words, getting to know someone who is linked to a person you already know. We can trust the new acquaintance because they know someone we know and they are thus subject to our group’s norms, monitoring, and sanctions against inappropriate or opportunistic behavior. For example, the tight connections between the diamond merchants in New York enable them to

Connecting to brokers

263

exchange thousands of dollars worth of diamonds without worrying about theft (Coleman, 1990). There may be a natural tendency toward connecting to friends of friends and the resulting closure. The preference for interacting with others similar to ourselves – ‘homophily’ – drives this approach. ‘Birds of a feather flock together.’ People tend to connect with others who are similar (Brass, 1985; Ibarra, 1992; McPherson and Smith-Lovin, 1987; Mehra et al., 1998). The generally accepted explanation for homophily is ease of communication. The idea that we are more likely to trust people who are like us is equally probable. In this case, we may feel we can more easily predict their behavior. These homophily connections may not be ‘strategic’ connections, that is, connections made with some intentional investment, future value in mind. More strategic may be complementary connections. Complementarity is a criteria network members often use when they are trying to acquire resources, reduce uncertainty and achieve collective goals (Brass et al., 2004). The manufacturing organization becomes a complementary partner with the distribution company. However, even research on alliance formation, which one would expect to depend on a more strategic, investmentoriented approach, suggests that homophily is still relevant. Similar status and power are important, as well as trust, prior connections and thirdparty connections (Chung et al., 2000; Gulati, 1995; Gulati and Gargiulo, 1999). Homophily is also evident in investment banks that ally with others that have similar social characteristics (Podolny, 1994). I assume some complementarity at the interpersonal level, but further assume that such relationships are relatively short-lived unless the actors discover a basis for homophily. Strategically instrumental interactions seldom become enduring relationships without some basis for similarity. Ideally, we connect to others who are instrumentally complimentary and socially similar. Within the constraints of time and space, we build relationships with people who are similar and whom we like. If we think strategically, it probably involves little more than the generally accepted notion that bigger networks are better. Creating a Closure Strategy Homophily may be the seed of the closure strategy. Similarity breeds attraction; attraction breeds interaction; and interaction breeds more similarity (Umphress et al., 2003). This situation makes it relatively easy to create trust, norms of reciprocity and sanctions for inappropriate behavior. Yet, similarity is relative and increased similarity within groups fosters increased perceptions of dissimilarity across groups. Membership in

264

Social capital: reaching in

close-knit groups fosters in-group/out-group biases such as distrust of dissimilar others. Achieving social capital within the group may be at the expense of social capital across groups. Closure Requires Brokerage The irony of the closure strategy is that it may require brokerage to be successful. Closure, as discussed by Coleman (1990), does not require that everyone in the network be connected to everyone else in order for norms to develop or sanctions to be applied. Indeed, if everyone were to be required to connect to everyone else (maximum density), even moderately sized organizations or communities would never be able to develop norms. As the size of the group increases, the chances of everyone being connected to everyone else decrease considerably. For example, in a group of 100 actors, 4950 connections are required for everyone to be connected to everyone else (n (n1)2). In a medium-sized city, millions of connections are needed. Conversely, maximally connected groups may represent the small, isolated cliques that often result in in-group/out-group biases, distrust across groups, and the failure to transmit norms across groups. Closure and the resulting trust, norms of reciprocity, and so on require only that there be a network path connecting the members of the group. For example, in noting the beneﬁts of closure, Coleman (1990, p. 303) describes a mother who moves from Detroit to Jerusalem in order to ensure more security and freedom for her children. ‘In Jerusalem, the normative structure ensures that unattended children will be looked after by adults in the vicinity.’ In another example, Coleman (1990, pp. 302–3) describes the social capital of clandestine study circles – groups of radical student activists in South Korea. ‘To avoid detection, members of diﬀerent groups never meet, but communicate through an appointed representative.’ In this example of social capital, the representatives are brokers between otherwise disconnected groups. They are the crucial nodes through which information about norms and behavior travels in the network. Take away the broker’s connections and the larger, clandestine group (or the city of Jerusalem) becomes smaller, isolated groups that do not share trust and norms of reciprocity. This results in an ironic conclusion: to acquire closure and its associated beneﬁts, brokerage is necessary.

THE BROKER STRATEGY As Burt (1992) has argued, just making lots of connections at random is not very eﬃcient – your return on your time and eﬀort is not optimal.

Connecting to brokers

265

Rather, it is advisable to be the tertius gaudens – the third who beneﬁts. The beneﬁts are: (1) access to diverse, non-redundant information and resources; (2) control of information and resources that flow between disconnected others; and (3) ability to play one party oﬀ against the other. The third beneﬁt comes at the expense of the ﬁrst. The ability to play one contact oﬀ against the other (two potential buyers bidding for your services) requires redundancy between the contacts, and forfeits the beneﬁts of non-redundant information. Nevertheless, as the opening references illustrate, it is evident that the third can beneﬁt from the disconnection between the ﬁrst and the second. It is the broker strategy – connecting to diverse, disconnected others – that can integrate society and lead to a small world and six degrees of separation (Granovetter, 1973; Watts, 2003). The essence of the broker strategy is simple: build networks with people who are not themselves connected. However, achieving the broker strategy is not necessarily so straightforward. Homophily preferences and trust are likely to be tendencies to be overcome in our attempt to connect to diverse others. Indeed, it is these tendencies, along with proximity, that result in network actors clustering within relatively separate cliques (leading to the advantage to brokers who can bridge cliques). Ideas and information circulate within, rather than across, groups. By bridging groups, brokers not only acquire nonredundant information but are the ﬁrst to see synergistic opportunities across groups (Burt, 1992). Any connection to a friend of a friend, and the trust that results from closure in our network does not give us the advantage of the broker role. Thus, to become the tertius gaudens, we must connect to people who are not connected to someone we already know. Brokerage Requires Connecting to Brokers The irony of the brokerage strategy is that the new contact also becomes a broker between her own contacts and us. Consider Figure 10.1. Pat and Chris have the same number of direct and indirect connections but Pat has adopted a broker strategy, connecting to others who are not connected themselves. As a result, Pat has an advantage over Chris, who is connected to others who are themselves connected (Burt, 2005). However, each of Pat’s ﬁve connections is now a broker for the relationships between Pat and all her indirect ties (the 13 people connected to positions 1–5). This irony – connecting to disconnected others makes you a broker but also places your connections in a broker role relative to you – clouds the questions of whether it is advisable to be the ﬁrst or second if it is the third who beneﬁts. In becoming the third, you simultaneously also become the ﬁrst or second.

266

Figure 10.1

Social capital: reaching in

Brokerage requires connecting to brokers

Previous ‘ego’ network research has not included indirect ties, and ‘whole’ network research has considered both direct and indirect ties equally in measuring brokerage. This latter research indicates that persons 1–5, who occupy indirect brokerage positions, gain some advantage. Other recent evidence, however, suggests indirect, second-hand brokerage oﬀers little additional advantage over direct brokerage (Burt, 2006). We need to examine longitudinal data to clarify these issues. For example, connecting to brokers may provide opportunities to connect to the broker’s direct links (your indirect links) at a later time. Using a Broker Whom to use Whether we connect with Chris or Pat, we are using each as a broker. The broker strategy means connecting to brokers. But who would provide us with the most beneﬁt? In most situations, we want to connect to the person who has the most links – direct and/or indirect. Larger networks oﬀer more potential for transforming indirect links into direct ones. But, there is no assurance that either Pat or Chris will introduce you to their direct links. And, in this case, both Pat and Chris have the same number of direct and indirect links. Whose direct links are more valuable? The answer depends on the particular situation. If we face an unstructured task and need creativity, innovation and non-redundant information, Pat’s links look more valuable than Chris’s. This assessment does not necessarily require an accurate picture of the networks. Because Pat’s structural holes provide novel, creative ideas,

Connecting to brokers

267

Pat will likely appear more innovative than Chris as a potential partner. If creativity and innovation are important, Pat is the preferred connection. Connecting to brokers provides the possibility of more non-redundant partnering opportunities with the broker’s ties. Connecting to those with similar or more power can oﬀer signiﬁcant advantages (Brass and Burkhardt, 1992; Ostrom, 1990; Ring and Van de Ven, 1992; Brass, 1984; Kilduﬀ and Krackhardt, 1994). Power in organizations is often associated with the number of direct ties, access to indirect ties, and brokerage (Brass, 1984). While Chris and Pat are equally powerful in terms of direct ties and access to indirect ties, Pat has more power by virtue of brokerage. Therefore she appears to be the appropriate choice. Being the ﬁrst or second is strategic if the third is powerful. However, it is crucial to ensure that a cooperative situation is present – competing with other, powerful players can have negative consequences. Simply associating with a powerful other can also be advantageous (Kilduﬀ and Krackhardt, 1994). This is true for both individuals (Kilduﬀ and Krackhardt, 1994) and organizations (Podolny, 2001). When powerful actors spend time with you, it sends a signal to outside observers that you are worth their attention. It is a form of advertising, increasing your own perceived power and attractiveness. However, these same brokers also have more chances to indulge in opportunistic, even unethical behavior (Brass et al., 1998). Risk management You need to consider how likely it is that diﬀerent brokers may take advantage of you as you select which one to use. In this example, the ability to monitor unethical behavior and apply sanctions is much lower in Pat’s closure network than in Chris’s brokerage network (Coleman, 1990). However, because Chris’s network is closed he may be more likely to act unethically toward you when faced with a choice between you and his network (Fernandez-Mateo, 2007). The repercussions for him acting against his network’s norms will be much more severe than those he will probably face if he takes advantage of you. Pat’s choice, on the other hand, will not be as clear-cut; acting unethically toward you runs the same level of risk as acting that way toward one of her ﬁve clusters. If necessary, she can take advantage of one cluster and then move on and exploit another one. Chris cannot take unfair advantage of any of his connections as the action would quickly circulate among his partners. To take another perspective, you may also use a broker to take advantage of another party. Although the broker manages the information flow between you and the other party, you control the data the broker receives. You can withhold information from the third party by keeping it from the

268

Social capital: reaching in

broker. For example, when selling my house I do not tell the real-estate agent that the roof leaks. If a potential buyer asks the agent if the roof leaks, she can say that she has no knowledge of any leaks. Without the broker, I would have a moral, if not a legal, obligation to tell the buyer about the roof. Of course, with this approach you risk severing any future relation with the broker as well as the second party. Maintaining brokerage relationships Considering how the brokers maintain their positions may provide some additional insight into how to work with them. Unfortunately, we have very little longitudinal data that address maintenance. If brokers enjoy advantages, it seems logical that they would want to maintain their brokerage role and the social capital it provides. One strategy is to maintain the disconnect between the actors. The broker remains the source of the non-redundant information to each party. This is advantageous for him in the short run, but the broker must pay the rent of continually trying to provide novel information to each party. And, at some point the novelty wears oﬀ. Brokers may gain short-term advantage by playing the two actors oﬀ against each other, taking advantage of one or both. While this may work in one-time transactions with no further contact, such situations are rare, particularly within organizations and even industries. The downside is that the broker may acquire a negative reputation and be avoided by others. While brokers can avoid negative reputational eﬀects by seeking out new social clusters, it seems much more eﬃcient to foster a positive reputation and referrals. This is because reputational information seems to easily travel across boundaries – especially negative reputational information. Brokers may use a strong, positive reputation to sustain their position. If one acquires this reputation, others will seek him out. This provides the broker with many potential partners and allows her to select from a variety of connections. People come to the broker – and being the object of relations is more powerful than being the source of relations (Knoke and Burt, 1983). Referrals are a primary mechanism for building additional social capital. Closed networks often provide the monitoring that aﬀects reputation (Coleman, 1990) and spread information about it quickly. While gossip requires a closed network (the parties exchanging gossip must both know the third party being talked about), referrals do not require mutual knowledge of the third party. For instance, I recommend a doctor to a friend without the friend having any prior experience with the doctor. Indeed, this is the primary purpose of referrals – connecting two parties who have no prior knowledge of each other. Acting on a referral is equivalent to a

Connecting to brokers

269

closure strategy. I take the advice of my friend and connect to a person who is connected to my friend. When done well, the broker becomes the ‘go-toguy’ for getting things done in organizations, for partnering within an interorganizational alliance or for implementing an innovation in a company (Obstfeld, 2005). Venture capitalists – a deﬁnite broker role – are successful to the extent that they can connect investors with successful entrepreneurs. How does the broker acquire the positive reputation? This is accomplished by brokering synergies between the diﬀerent actors. Selecting beneﬁcial partners to recommend to others is important as no one wants to be referred to a bad potential partner. As Baker (1994) suggested, closing structural holes to create helpful synergies may cause the actors to reciprocate by referring the broker to their own contacts, and vice versa. But even without this reciprocity the two actors are reputation makers or breakers.

WHEN TO CONNECT TO BROKERS Rather than asking if it is wise to connect to brokers, we might consider when it may be wise. Below I note seven situations – all represent instances when it may be diﬃcult or inadvisable to make direct connections. In such cases, the use of a broker may be the best option. Search Costs When search costs are high, a broker may be advisable. As clearinghouses of information, they may be able to decrease costs. For instance, Pat may be able to condense and pass along valuable data obtained from her disconnected others. Sears brings a variety of products from diverse manufacturers together in one location. EBay provides a clearinghouse for a variety of hard-to-ﬁnd products. Uncertainty Brokers can also remove the uncertainty surrounding connections to a new thing or person. This is particularly true when the product, service or information is complex or uncertain. Brokers like Consumer Reports provide ‘grading’ – a popular recent activity on the Internet as well. Stockbrokers attempt to reduce the uncertainty of predicting stock market fluctuations, and venture capitalists reduce the risk of investing in new ventures where uncertainty is high.

270

Social capital: reaching in

Negative Ties Negative ties are a relatively unexplored area of social network research (Labianca and Brass, 2006). However, there is evidence of negative asymmetry where negative events may be more salient than positive ones. Negative information may also circulate more quickly through the network, may decay more slowly, and may be more attended to than positive information. Connecting to a broker may be particularly helpful when you need a resource to overcome the potential social liabilities created by negative ties (Labianca and Brass, 2006). Because of your negative relationship, it is unlikely that you can even approach the actor, let alone arrive at a cooperative synergy. In such cases, it would be advantageous to have a broker represent you in attempting to acquire information or resources. The broker may also be able to keep you anonymous. Expertise Disconnects between people may also be the result of a lack of the necessary expertise. In these cases, it may be more eﬃcient to hire a broker than to try and acquire the expertise. This is a common occurrence (for example, lawyers, foreign language translators). These expert brokers help us connect indirectly with others without the burden of becoming experts ourselves. They not only make our interests known to the other party, but they ﬁlter and translate the information from that actor. For example, lawyers translate legal jargon. We must keep in mind that we become dependent on the broker’s translation, which may include distortions and inaccuracies, and will need to absorb the costs of the expert’s services. Emotional Involvement Brokers can represent us in a detached, non-emotional manner that may be to our advantage in situations where rational decisions or presentations are diﬃcult or where we are emotionally involved (either positively or negatively). Consequently, brokers are often hired to handle emotionally charged exchanges. Lawyers represent people in bitter divorces, lobbyists promote causes, or agents bid on a coveted possession at an auction. Coordination Managers, who are the primary brokers within companies, coordinate tasks within organizations, and organizations hire brokers to coordinate

Connecting to brokers

271

interorganizational activities within industries or regions. This is necessary because communication becomes ever more diﬃcult as the number of people involved increases. In these cases a broker can coordinate activities and increase eﬃciency. Trust While trust decreases the transaction costs of any exchange and increases many of its beneﬁts (Krackhardt, 1992) it must be treated carefully with regard to brokers. As successful con artists know, the ﬁrst step in taking the mark is building trust. However, brokers are advisable when the ﬁrst actor feels they cannot trust the second one. The dilemma is how to know which brokers to trust and how to make arrangements. We are reticent to interact with people we do not trust and there is often little information upon which to base trust in a new connection. However, we may use a broker when we trust the broker more than the other actor. For example, PayPal helps eBay overcome the inherent mistrust of buying from, or selling to, a stranger. I trust eBay more than I trust the stranger.

CONCLUSIONS This chapter begins with the question of whether it is wise to be the ﬁrst or second actor in a network instead of the broker. This question becomes particularly important when comparing the broker and closure strategies, and their resulting eﬀects on social capital at both the individual and group level. The two strategies oﬀer diﬀerent beneﬁts, involve varying risks, and play to contrasting human tendencies. We just need to decide which one ﬁts our particular needs. The closure strategy, connecting to friends of a friend, appears to be the natural outcome of our preferences for homophily – connecting to similar others. Employing it seems to involve little more than recommending that we follow our instincts. It produces closed networks that provide for ease of monitoring, sanctions, and building trust and norms of reciprocity. While this strategy creates social capital within the group, it does little to promote social capital across groups. As a result, it may not provide the diverse, nonredundant information that may be necessary for change. By contrast, the broker strategy requires connecting to brokers, an action which can seem risky. Using this approach means we must deliberately step away from homophily; at the same time, we need to recognize that we may run into an unethical broker who will take advantage of us. Nevertheless, the broker strategy provides non-redundant information and opportunities

272

Social capital: reaching in

for creativity and can create bridges across groups that serve to integrate society. However, the possible beneﬁts of bridges across small groups, providing closure within larger groups, may depend on the actions taken by the broker. Opportunistic behavior of maintaining the disconnect and playing one party oﬀ against the other does not integrate society. The broker who chooses this opportunistic behavior must sever ties and move on to other clusters. On the other hand, connecting previously unconnected parties does promote integration and the beneﬁts of closure as well as innovation. And reputation and referrals may be the key mechanisms and motivation for brokers to connect others. Referrals do not require closed systems, but acting on referrals is a closure strategy. Both may be key mechanisms in promoting closure in the network and social capital at the collective level. Just as homophily may be less a ‘strategy’ than a natural tendency, brokerage may be more accidental than strategic. In our normal day-to-day activities, we meet new people. This is accidental brokerage and likely represents the majority of our brokerage opportunities. We then make choices about future interactions and introducing these new contacts to our friends and acquaintances. Such decisions may not be calculated, but they help create the social structure that becomes social capital.

REFERENCES Adler, P.S. and S. Kwon (2002), ‘Social capital: prospects for a new concept’, Academy of Management Review, 27, 17–40. Baker, W.E. (1994), Networking Smart: How to Build Relationships for Personal and Organizational Success, New York: McGraw-Hill. Brass, D.J. (1984), ‘Being in the right place: a structural analysis of individual influence in an organization’, Administrative Science Quarterly, 29, 518–39. Brass, D.J. (1985), ‘Men’s and women’s networks: a study of interaction patterns and influence in an organization’, Academy of Management Journal, 28, 327–43. Brass, D.J. (1995), ‘Creativity: it’s all in your social network’, in C.M. Ford and D.A. Gioia (eds), Creative Action in Organizations, Thousand Oaks, CA: Sage, pp. 94–9. Brass, D.J. and M.E. Burkhardt (1992), ‘Centrality and power in organizations’, in N. Nohria and R. Eccles (eds), Networks and Organizations: Structure, Form, and Action, Boston, MA: Harvard Business School Press, pp. 191–215. Brass, D.J., K.D. Butterﬁeld and B.C. Skaggs (1998), ‘Relationships and unethical behavior: a social network perspective’, Academy of Management Review, 23, 14–31. Brass, D.J., J. Galaskiewicz, H.R. Greve and W. Tsai (2004), ‘Taking stock of networks and organizations: a multilevel perspective’, Academy of Management Journal. Burt, R.S. (1992), Structural Holes: The Social Structure of Competition, Cambridge, MA: Harvard University Press.

Connecting to brokers

273

Burt, R.S. (2004), ‘Structural holes and good ideas’, American Journal of Sociology, 110, 349–99. Burt, R.S. (2005), Brokerage and Closure: An Introduction to Social Capital, Oxford: Oxford University Press. Burt, R.S. (2007), ‘Second-hand brokerage: evidence on the importance of local structure for managers, bankers, and analysts’, Academy of Management Journal, 50 (1), 119–48. Chung, S., H. Singh and K. Lee (2000), ‘Complementarity, status similarity, and social capital as drivers of alliance formation’, Strategic Management Journal, 21, 1–23. Coleman, J. (1990), Foundations of Social Theory, Cambridge, MA: Harvard University Press. Fernandez-Mateo, I. (2007), ‘Who pays the price of brokerage? Transferring constraint through price-setting in the staﬃng sector’, American Sociological Review, 291–317. Granovetter, M. (1973), ‘The strength of weak ties’, American Journal of Sociology, 78, 1360–80. Granovetter, M. (1985), ‘Economic action and social structure: the problem of embeddedness’, American Journal of Sociology, 91, 481–510. Gulati, R. (1995), ‘Social structure and alliance formation patterns: a longitudinal analysis’, Administrative Science Quarterly, 40, 619–52. Gulati, R. and M. Gargiulo (1999), ‘Where do interorganizational networks come from?’, American Journal of Sociology, 104, 1439–93. Ibarra, H. (1992), ‘Homophily and diﬀerential returns: sex diﬀerences in network structure and access in an advertising ﬁrm’, Administrative Science Quarterly, 37, 422–47. Kilduﬀ, M. and D. Krackhardt (1994), ‘Bringing the individual back in: a structural analysis of the internal market for reputation in organizations’, Academy of Management Journal, 37, 87–108. Knoke, D. and R.S. Burt (1983), ‘Prominence’, in R.S. Burt and M.J. Miner (eds), Applied Network Analysis: A Methodological Introduction Beverly Hills, CA: Sage, pp. 195–222. Koput, K. and W.W. Powell (2003), ‘Organizational growth and alliance capability: science and strategy in a knowledge-intensive industry’, unpublished paper, University of Arizona. Krackhardt, D. (1992), ‘The strength of strong ties: the importance of philos in organizations’, in N. Nohria and R. Eccles (eds), Networks and Organizations: Structure, Form and Action, Boston, MA: Harvard Business School Press, pp. 216–39. Labianca, G. and D.J. Brass (2006), ‘Exploring the social ledger: negative relationships and negative asymmetry in social networks in organizations’, Academy of Management Review, 31, 596–614. Lin, N. (2001), ‘Building a network theory of social capital’, in N. Lin, K. Cook and R.S. Burt (eds), Social Capital, New York: Aldine de Gruyter, pp. 3–30. McEvily, B. and A. Zaheer (1999), ‘Bridging ties: a source of ﬁrm heterogeneity in competitive capabilities’, Strategic Management Journal, 20, 1133–56. McPherson, J.M. and L. Smith-Lovin (1987), ‘Homophily in voluntary organizations: status distance and the composition of face-to-face groups’, American Journal of Sociology, 52, 370–79.

274

Social capital: reaching in

Mehra, A., M. Kilduﬀ and D.J. Brass (1998), ‘At the margins: a distinctiveness approach to the social identity and social networks of underrepresented groups’, Academy of Management Journal, 41, 441–52. Mehra, A., M. Kilduﬀ and D.J. Brass (2001), ‘The social networks of high and low self-monitors: implications for workplace performance’, Administrative Science Quarterly, 46, 121–46. Nahapiet, J. and S. Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23, 242–66. Obstfeld, D. (2005), ‘Social networks, the tertius iungens orientation, and involvement in innovation’, Administrative Science Quarterly, 50, 100–130. Oh, H., M. Kilduﬀ and D.J. Brass (2006), ‘Network ties and business success: the case of Korean entrepreneurs’, paper presented at Networks and Entrepreneurs conference, July, Amsterdam. Ostrom, E. (1990), Governing the Commons: The Evolution of Institutions for Collective Action, New York: Cambridge University Press. Podolny, J.M. (1994), ‘Market uncertainty and the social characteristics of economic exchange’, Administrative Science Quarterly, 39, 458–83. Podolny, J.M. (2001), ‘Networks as the pipes and prisms of the market’, American Journal of Sociology, 107, 33–60. Provan, K.G. and H.B. Milward (1995), ‘A preliminary theory of interorganizational network eﬀectiveness: a comparative study of four community mental health systems’, Administrative Science Quarterly, 40, 1–33. Putnam, R.D. (1995), ‘Bowling alone: America’s declining social capital’, Journal of Democracy, 6, 65–78. Ring, P.S. and A.H. Van de Ven (1992), ‘Structuring cooperative relationships between organizations’, Strategic Management Review, 13, 483–98. Rowley, T.J. and J.A.C. Buam (2004), ‘Sophistication of interﬁrm network strategies in the Canadian investment banking industry’, Scandinavian Journal of Management, 20, 103–24. Seibert, S.E., M.L. Kraimer and R.C. Liden (2001), ‘A social capital theory of career success’, Academy of Management Journal, 44, 219–37. Simmel, G. (1950), ‘Individual and society’, in K.H. Wolﬀ (ed.), The Sociology of Georg Simmel, New York: Free Press. Stuart, T.E. and J.M. Podolny (1999), ‘Positional causes and correlates of strategic alliances in the semiconductor industry’, Research in the Sociology of Organizations, 16, 161–82. Umphress, E.E., G. Labianca, D.J. Brass, E. Kass and L. Scholten (2003), ‘The role of instrumental and expressive social ties in employees’ perceptions of organizational justice’, Organization Science, 14, 738–53. Watts, D.J. (2003), Six Degrees: The Science of a Connected Age, New York: W.W. Norton.

11.

Trust, distrust and building social capital Roy J. Lewicki and Chad T. Brinsﬁeld

The relationship between trust and social capital oﬀers signiﬁcant insights that are often actionable. While it is hard to challenge the existence of a strong, positive relationship between the two, a closer examination of trust research reveals multiple issues. Trust has been deﬁned in many diﬀerent ways: in some cases it focuses on the trustors’ perceptions, intentions, attitudes, expectations and behaviors; in others, the qualities and conduct of the trustee; in yet others, the type and level of trust judgments or the context in which people decide to trust. The question also exists as to whether trust and distrust are opposite ends of a single construct or distinctly diﬀerent entities that can coexist. Similar diﬃculties occur when we try to relate trust to social capital. As this volume clearly shows (Ostrom, Chapter 1; Kramer, Chapter 9; Davis and Bartkus, Chapter 13), the delineation of social capital and its related constructs is still the subject of much scrutiny and academic discussion, often yielding far more heat than light. Our argument is that diﬀerent types of trust exist and that trust and distrust are conceptually and empirically distinct. However, we also assert that the same relationship may contain both trust and distrust, and that such a combination can be healthy. In fact, the balance of the two forms the basis of social capital in many relationships (interpersonal, group or organizational). But if the relationship is going to remain productive, the parties involved must continuously consider how to build trust and manage distrust – two distinct sets of actions. Our chapter begins by briefly examining social capital’s core concepts, then reviewing the theoretical and empirical work on trust. We focus on discriminating between diﬀerent types of trust and diﬀerentiating between trust and distrust. At the end, we discuss the implications of both processes for building and sustaining social capital.

275

276

Social capital: reaching in

WHAT IS SOCIAL CAPITAL? The term ‘social capital’ has often been used so broadly that confusion is rampant relative to what speciﬁc uses mean (Knorringa and van Staveren, 2006). Although diﬃcult to deﬁne consistently, social capital is an important and complex concept, referred to by some as the basic raw material of a civilized society (Onyx and Bullen, 2000). It stems from several theoretical traditions. Putnam, who drew the concept from sociologist James Coleman, deﬁned social capital as ‘those features of social organization, such as trust, norms and networks, that can improve the eﬃciency of society by facilitating coordinated actions’ (Putnam, 1993). Nooteboom, in 2007, stated that social capital denotes almost anything related to ties between people; however, these ties are usually informal and voluntary rather than formal (for example, hierarchical or contractual) and mandated (see Latham, 1997). Researchers have also analyzed social capital at the macro (countries, regions), meso (neighborhoods), micro (social networks) and individual psychological (trust) levels (Macinko and Starﬁeld, 2001; Mohseni and Lindstrom, 2007). Social capital is clearly extremely complex. Ostrom (Chapter 1 in this volume) refers to it as ‘a set of relationships and shared values created by multiple individuals to solve collective action problems in the present and future’. As such, social capital encompasses ‘how cultural, structural and institutional aspects of small to large groups in a society interact and aﬀect individual incentives and behavior, and resultant economic and political change’ (Ostrom, Chapter 1 in this volume; Pasotti and Rothstein, 2002). It does this by creating shared understandings about norms, rules and expectations among groups (and institutions) (Coleman, 1990; Ostrom, Chapter 1, this volume). These norms emerge as individuals interact with each other over time. Structurally, these redundant exchanges become ‘networks’, or webs of activities that occur both within and across organizations. As the exchanges become routine, implicit and explicit ‘rules’ take shape. Examples include families, friendship and information networks, communities, and various organizations that use such norms or ‘rules of the game’ to dictate actions that are necessary and appropriate, or unnecessary and inappropriate (North, 1990). Social capital also strengthens productive exchange relationships, reduces transaction costs, and fosters local ﬁnancial development and general economic growth. It enhances civic engagement, promotes greater social coordination (for example, matches people to economic opportunities) and enhances social stability (Chou, 2006; Coleman, 1990; Helliwell and Putnam, 1995; Putnam, 1993).

Trust, distrust and building social capital

277

Not all of social capital’s outcomes are positive. Several potential negative consequences have surfaced from recent research, including exclusion of outsiders, excess claims on group members and restrictions on individual freedoms. The strong ties that help group members identify with each other and that convey beneﬁts to members may prompt the group to bar access to others. Furthermore, these close, cohesive intra-group relations can give rise to signiﬁcant free-riding as less diligent members demand more from more successful ones. The shared norms often support these requests. Strong social capital may also exert pressure to conform (Portes, 1998). Composition of Social Capital Researchers often treat social capital as a combination of two separate, related components. These aspects are structural and relational (for example, Nahapiet and Ghoshal, 1998; Sandefur and Laumann, 1998; Walker et al., 1997). The structural dimension includes the more formalized patterns of connections between exchange partners, such as network conﬁgurations, participation in voluntary organizations and formal network ties (Mitchell and Bossert, 2007; Nahapiet and Ghoshal, 1998). The scope and extent of social interaction between the actors is an important aspect of this dimension (Tsai and Ghoshal, 1998). In contrast, the relational dimension refers to the human assets within these network relationships. These assets’ value reflects the quality of the individual ties that make up the broader network (Coleman, 1990). This dimension focuses on relational phenomena like trust, fairness and social norms – which lead to obligations and expectations about the extent and nature of cooperation with others (De Clercq et al., 2007; Uzzi, 1997). Reciprocity, which involves an obligation of future repayments for beneﬁts received from others, is a fundamental part of this dimension. This norm is distinct from formal exchanges as a normative reciprocated beneﬁt is not explicitly required even if it is expected. Such behaviors are often grounded in short-term altruism and long-term self-interest (Onyx and Bullen, 2000; Tansley and Newell, 2007; Taylor, 1982). In summary, social capital influences peoples’ perception of the psychological contract, relative to what is expected of them and what they expect from others, and the fairness and balance of the exchange (Ho et al., 2006; Tansley and Newell, 2007). The Key Role of Trust Because social capital is a relational construct, it depends heavily on trust. For many researchers, trust is the ‘glue’ that holds relationships together.

278

Social capital: reaching in

Indeed, norms of reciprocity are based on trust. Furthermore, according to Putnam (1993), the combined eﬀect of trust, social networks and norms of reciprocity creates a strong community, with a sense of shared resource ownership. This ownership keeps opportunists from overusing shared resources and avoids reductions in resource value. But doing this requires strong norms for trust and reciprocity, as well as eﬀective informal social sanctions against free-riders (Onyx and Bullen, 2000). In the next section, we briefly review the concept of trust and its implications for the development and maintenance of social capital.

DISTINCT CONCEPTUALIZATIONS OF TRUST Trust is conceptualized and deﬁned in multiple ways. The literature treats it as a type of behavior (Deutsch, 1962; Mayer et al., 1995, p. 712), an underlying disposition (Rotter, 1967, 1971; Worchel, 1979) and/or a psychological state (Lewicki et al., 1998, p. 440; Rousseau et al., 1998). Although some diﬀerences exist across these and other deﬁnitions (see Bigley and Pearce, 1998 for a complete review), all of them discuss beliefs and expectations about the intentions of another party and the willingness to accept vulnerability. Deutsch (1962) said trusting behavior consisted of actions that increased a person’s vulnerability to another whose behavior one did not control where the penalty one suﬀers if the vulnerability is abused exceeds the beneﬁt gained if the other does not abuse it. Worchel (1979) deﬁned trust as a belief, expectancy or feeling deeply rooted in the personality, with origins in the individual’s early psychosocial development (Erickson, 1963; Rotter, 1971). Rousseau et al. (1998) described trust as ‘a psychological state comprising the intention to accept vulnerability based on the positive expectations of the intentions or behavior of another’ (p. 395). Since the early 1990s, researchers have further reﬁned the concept of trust. Some have delineated its structural components, whether these are behaviors (see Deutsch, 1958; Flores and Solomon, 1998; Berg et al., 1995) or a complex psychological state, with cognitions, emotions and intentions (see Worchel, 1979; Rousseau et al., 1998; Lewicki et al., 1998; Bigley and Pearce, 1998; Mayer et al., 1995). Exploration has also focused on the qualities of the person being trusted (see Mayer et al., 1995). We will not explore these diﬀerent perspectives (see Lewicki et al., 2006, for one review), but presume that all of them are relevant to understanding the role of trust in social capital. However, we will discuss three other more recent approaches that appear to have particularly important implications for social capital: (1) that the

Trust, distrust and building social capital

279

evolution of relationships over time may create diﬀerent ‘forms’ of trust – in other words, trust may transform over time from an approach grounded in the explicit costs/beneﬁts of an interpersonal transaction to an approach based on shared values and attributes; (2) that trust and distrust are separate constructs which vary independently from ‘low’ to ‘high’ levels and which coexist with each other; and (3) that theory and research suggest that trust does not necessarily begin at zero. Trust Has Diﬀerent Forms Researchers have argued that many of the issues surrounding trust appear to be based on an overly simplistic view of both trust and relationships. Many early theories viewed trust as part of simple economic transactions between parties with little or no relationship history or experience with each other. They then extrapolated ﬁndings from these situations to describe how trust develops and functions in richer, more complex relationships. However, we posit that one cannot assume trust and its development are the same in all types of relationships. For example, professional relationships are task-oriented and focus on achievement of goals external to the relationship. Personal or relational relationships are social and emotional, and deal with the relationship itself and the persons in it (Deutsch, 1985; Sheppard and Sherman, 1998). We will now examine professional relationships in more detail. Shapiro et al. (1992) proposed that trust in professional relationships divides into three types: deterrence-based, knowledge-based and identiﬁcation-based. In recent papers, Lewicki and Bunker (1995, 1996), Lewicki and Wiethoﬀ (2000) and Lewicki (2006) modiﬁed these three types into two: calculus-based (deterrence-based trust renamed) and identiﬁcation-based. Calculus-based trust ‘Calculus-based trust’ is the ﬁrst form of trust in professional relationships. It ensures consistent behavior – in other words, that individuals do what they promise to do. Shapiro et al. (1992) called this deterrence-based trust because they argued that it was grounded in fear of consequences or punishment and that this threat was more motivating than the promise of reward. Trust is sustained only to the degree that the deterrent (punishment) is clear, possible and likely to occur if the trust is violated. Lewicki and Bunker (1995, 1996) renamed this as calculus-based trust (CBT); their deﬁnition, which focuses on vulnerability, is similar to those of Deutsch (1962) and Rousseau et al. (1998). However, they value the beneﬁts of fulﬁlling trust equally with the deterrent consequences of not

280

Social capital: reaching in

fulﬁlling it. Trust is now an ongoing, market-oriented, economic calculation that weighs the outcomes of sustaining the relationship against the costs of maintaining or severing it. Compliance is often ensured both by the rewards of trusting and being trustworthy and by the threat of direct negative consequences if trust is violated (for example, damage to reputation). In this scenario, network members calculate and balance the shortterm gains from untrustworthy behaviors against the long-term beneﬁts of a good reputation. While trustors focus on risk and vulnerability, and initiate only moderate-level trust gestures, trustees focus on beneﬁts and are more likely to reciprocate when reward levels are high (Malhotra, 2004; Weber et al., 2005). For instance, ‘moderate initial trust’ – which does not make economic sense but is crucial to creating trust – may occur because some trustors manage their anxieties about vulnerability by placing high initial trust in others or by focusing more on the potential upside. Diﬀerences in how disposed diﬀerent personalities are to trust may explain this dynamic (Malhotra, 2004; Weber et al., 2005). It is also important to watch for signal miscues between trustors and trustees, which can occur easily based on their diﬀerent perspectives. Identiﬁcation-based trust (IBT) CBT is usually the ﬁrst, early stage in trust. The next stage, called identiﬁcation-based trust, centers around recognitions and empathy with the other’s desires and intentions. Parties must understand and appreciate one another’s wants and needs. This mutual understanding eventually develops to the point that each person can act for the other and can, in some cases, serve as the other’s agent or substitute in interpersonal transactions (Deutsch, 1949). Both parties feel conﬁdent their interests are protected; no monitoring is necessary. A truly strong IBT takes place when one party acts more zealously or goes out of their way to advocate more for the other than the situation warrants. As parties get to know and identify more closely with each other, they understand more clearly what they need to do to sustain the other’s trust – an example of ‘second-order’ learning. One learns what really matters to the other and begins to place the same importance on those behaviors. The person starts to think and respond like the other and, in many cases, to anticipate the other’s reactions. A collective identity develops where strong empathy exists and each person incorporates some of the other’s needs, preferences, thoughts and behavior patterns into themselves. Such trust provides multiple beneﬁts. When present in working relationships it enhances performance (for example, a cappella choir with unverbalized, synchronous chemistry, improvised jazz groups, cohesive work

Trust, distrust and building social capital

281

Relational trust

Calculative trust

Early

Middle

Late

Developmental time Source:

Rousseau, Sitkin, Burt and Camerer (1998)

Figure 11.1

Integrating trust and distrust: alternative social realities

groups or championship basketball teams). Relationships are more stable during conflict and negative events; when high-trusting parties disagree, they tend to see the best in their partner’s motives because they make more positive assumptions about the partner’s motives (vs low-trusting parties). It is possible this attribution may be the deciding factor in whether relationships survive conflict (Miller and Rempel, 2004). Rousseau et al. (1998) proposed a comparable, somewhat simpler model of these two trust forms. The model provides a way to organize multiple theoretical articles on trust (see Figure 11.1). It is also based on the idea that trust ‘takes diﬀerent forms in diﬀerent relationships – from a calculated weighting of gains and losses to an emotional response based on interpersonal attachment and identiﬁcation’ (Rousseau et al., 1998, p. 398). They also propose that repeated interactions over time changes trust from a more calculative type to a more relational one, one which is grounded in stronger aﬀect and identiﬁcation with the other. Trust and Distrust as Separate Constructs Recently, scholars have recognized the value of examining trust and distrust as distinct constructs which may exist simultaneously in a given relationship (Lewicki et al., 1998). Before this, research assumed trust was a unidimensional construct ranging from ‘high trust’ at one end of the

282

Social capital: reaching in

spectrum to either ‘low trust’ or ‘distrust’ at the other (Mayer et al., 1995; Schoorman et al., 2007; Rousseau et al., 1998; Cummings and Bromiley, 1996). In the new paradigm, distrust is a distinct concept with more meaning than ‘the absence of trust’. It is deﬁned as ‘conﬁdent negative expectations regarding another’s conduct’ (Lewicki et al., 1998, p. 439). This phrase means ‘a fear of, a propensity to attribute sinister intentions to, and a desire to buﬀer oneself from the eﬀects of another’s conduct’ (Lewicki et al., 1998, p. 439). Trust, on the other hand, is deﬁned as ‘conﬁdent positive expectations regarding another’s conduct’ (Lewicki et al., 1998, p. 439). Both dimensions – trust and distrust – fall along a continuum from ‘high’ to ‘low’ – see Table 11.1. But low trust is not the same as high distrust; low trust suggests a lack of hope, an unsure assessment of the other’s behavior and hesitancy to act, while high distrust implies fear, strong skepticism and a commitment to monitor the other vigilantly. Conversely, high trust leads to hope, faith and conﬁdence, whereas low distrust suggests an absence of fear, skepticism and a reduced need to monitor the other. Lewicki et al. (1998) argue that this framework reflects the ways people make complex judgments around trust or distrust more accurately. The unidimensional model works best if one always adopts a single or aggregate level of trust or distrust for another party, no matter the level of experience with that other. But the authors posit that such unidimensional models oﬀer a simplistic view of parties’ perceptions and judgments of trust. Because most interpersonal relationships are constantly in flux, with broad ‘bandwidth’, the model needs to reflect this evolution. As a relationship develops, an actor encounters the other party in a variety of contexts and settings. Each interaction leads to an implicit or explicit judgment about the other (which the authors call a ‘facet of experience’). As relationships become deeper, these facets accumulate into bands. Some of the facets may be judgments of trust, others of distrust; some facets may have greater weight due to the relative importance of the experience. In such a relationship, the critical question is not, ‘Do you trust or distrust this other party?’ but ‘Do I trust them to do what?’ Let us look at an example. In a simple relationship, such as the one with my coﬀee barista, I drive through in the morning, order my large latte, smile and say ‘Good morning’, pay for my beverage, and drive oﬀ. The relationship is long-lasting but single faceted, as the barista and I have no exchange other than greetings and an exchange of money for product. Complex, multifaceted relationships are radically diﬀerent. My spouse may trust me to be up on current events because I read the newspaper for an hour in the morning; they may not trust me to remember to water the plant next to my reading chair given my previous history of letting plants die.

Trust, distrust and building social capital

Table 11.1

283

Integrating trust and distrust: alternative social realities

High Trust Characterized by: Hope Faith Confidence Assurance Initiative

High value congruence Interdependence promoted Opportunities pursued New initiatives

2 1 Low Trust Characterized by: No Hope No Faith No Confidence Passivity Hesitance

Trust but verify Relationships highly segmented and bounded Opportunities pursued and downside risks/vulnerabilities continually monitored

4 3

Casual acquaintances Limited interdependence Bounded, ‘arm’s-length’ transactions Professional courtesy

Undesirable eventualities expected and feared Harmful motives assumed Interdependence managed Pre-emption: Best offense is a good defense Paranoia

Low Distrust Characterized by:

High Distrust Characterized by:

No Fear Absence of Skepticism Absence of Cynicism Low Monitoring Non-vigilance

Fear Skepticism Cynicism Wariness and Watchfulness Vigilance

Trust in complex relationships has to be multifaceted if it is going to reflect multiple, diﬀerent experiences. Many relationships may ultimately be ‘ambivalent’ (that is, a set of intricate conditions or contexts where one trusts or distrusts the other party). In truth, such multifaceted relationships are far more functional than attempting a simple unidimensional approach that misses varied interactions. If one does use a unidimensional measure, the ‘meaning’ of any single judgment requires that we unpack, disentangle and evaluate a widely diverse set of past experiences. To use a simple metaphor, the single view of trust assumes that multiple experiences can be blended into one flavor and still retain a meaning. The second approach

284

Social capital: reaching in

treats the blend as a multiflavored dish where each ingredient’s taste and texture is still strongly distinguishable. For example, a high-trust, lowdistrust relationship combines many trust-related facets with a few distrustrelated ones. The parties have conﬁdent positive expectations of each other’s conduct; they have had few experiences that created strong negative expectations and led to distrust. They probably interact frequently and believe that the other is pursuing their common goal and is committed to the relationship. As noted by Lewicki et al.: The facet elements, bands and bandwidth of the relationship reflect a large number of positive experiences, in which the aggregate experience has been trust reinforcing. This experience creates social capital that enables the trusting party to exercise initiative, assured of the support of the trusted party. Parties are likely to seek ways to continually develop and enrich this relationship and to expand their mutually beneﬁcial interdependencies. Conversation is complex and rich, reflecting each party’s awareness of the other. (p. 446)

Separating trust and distrust also reveals that (positive) trust can be dysfunctional and (negative) distrust can be functional. McAllister (1997) extensively documented positive trust’s dysfunctionality. McAllister divides the ‘dark side’ of trust into four types of risk: (1) others may easily manipulate and abuse one’s trust; (2) one can make serious errors in judgment about whether others are, or will be, trustworthy; (3) parties may continue trusting one another despite evidence that the trust is failing; and (4) after one recognizes that trust has been betrayed, a destructive cycle of behaviors may occur, a cycle that requires signiﬁcant eﬀort to overcome. Distrust can prevent some of these problems and protect us from the naivety that keeps us from detecting others’ deception and exploitation. It helps us remain vigilant and may prompt the creation of formal contracts that aﬃrm and clarify verbal ‘handshake’ pacts – a step that may avoid later conflict. As Tomlinson and Lewicki note: ‘In short, it is possible (and even advisable) to have a healthy dose of distrust of others, particularly with people whom we do not know well or those who indicate that they may not be trustworthy’ (Tomlinson and Lewicki, 2006, p. 221). Lewicki (2006) has combined the two-stages model (CBT and IBT) with the positive–negative spectrum in an attempt to create an overarching framework for examining trust. The model speciﬁes four types of trust: ●

Calculus-based trust (CBT) is a conﬁdent positive expectation regarding another’s conduct. It is grounded in impersonal transactions; the overall anticipated beneﬁts from the relationship are assumed to outweigh any anticipated costs.

Trust, distrust and building social capital ●

●

●

285

Calculus-based distrust (CBD) is deﬁned as conﬁdent negative expectations regarding another’s conduct. It is also grounded in impersonal transactions; the overall anticipated costs are assumed to outweigh the anticipated beneﬁts. Identiﬁcation-based trust (IBT) is deﬁned as conﬁdent positive expectations regarding another’s conduct. It is grounded in perceived compatibility of values, common goals and positive emotional attachment to the other. Identiﬁcation-based distrust (IBD) is deﬁned as conﬁdent negative expectations regarding another’s conduct, grounded in perceived incompatibility of values, dissimilar goals and negative emotional attachment to the other (Lewicki, 2006, p. 100).

The implications of these four types of trust will be explored further in the last section of this chapter. Trust May Not Begin at ‘Zero’ Most research has assumed that trust of another begins at a relative ‘zero’ level and develops gradually over time (for example, Rempel et al., 1985; Jones and George, 1998). Other approaches have talked about the presence of a baseline of initial distrust. Various factors could account for this. They include: a distrusting personal disposition; cultural factors that bias individuals toward initial distrust (for example, negative social capital); prior information about another’s untrustworthiness; perceived mismatch of values; or a social context that heavily monitors and controls its employees (Kramer, 1999; Kelly and Stahelski, 1970; Cialdini, 1996; Sitkin and Stickel, 1996). Since 1990, researchers have challenged both assumptions, pointing out that parties with no information or history about each other nevertheless show a remarkable amount of initial trust (for example, Kramer, 1994; McKnight et al., 1998). A non-zero, positive initial disposition to trust others may be due to several factors. First, diﬀerences in individual personality may dispose some to be more trusting. Second, individuals may rapidly process information about the other party and use this to determine whether they are trustworthy – for example, self-presentation, dress, mannerisms (Gladwell, 2000). Third, institutions may provide formal and normative structures that provide protection against distrusting actions by another, thus raising the level of social capital and increasing the likelihood of a trusting judgment (Sitkin, 1995). Finally, some individuals and groups possess commonly understood and agreed-upon procedures; because all parties adhere to these practices, trust increases quickly. Such ‘swift trust’

286

Social capital: reaching in

in surgical teams, disaster rescue squads and airline cockpit crews is facilitated by the parties’ mutual commitment to work together according to the prescribed procedures (for example, Meyerson et al., 1996). Multistage models of trust tend to adopt the perspective that relationships often start with a moderate level of trust. Lewicki and Bunker (1995, 1996) indicate that trust begins with some moderate level of CBT. Repeated interactions, the degree of interdependence and reputational dynamics can all strengthen CBT. Rousseau et al. (1998) also imply that trust begins with a moderate level of calculative trust. While both approaches specify that trust develops from more calculative to more identiﬁcation-based (or relational) over time, they acknowledge that this evolution may be very slow or very rapid. How quickly the trust shifts form may depend on the type and level of aﬀect involved as well as the relationship’s purpose (for example, structured around transactional vs ‘pure’ relational objectives). Summary In this section, we identiﬁed major research themes in trust and reviewed areas of particular importance to social capital. Our primary points included: ●

●

●

Several forms of trust exist: a more calculative, economic-based form of trust (CBT) and a more relational, identity-based form of trust (IBT). Trust and distrust are conceptually distinct, and can coexist in one relationship; they may also combine with the more calculative and identity-based approaches. While the more economically based approaches to trust assumed that trust begins at ‘zero’ in any relationship, research has consistently shown that parties may express moderate levels of trust in another party, prior to any direct or indirect evidence of the other’s trustworthiness.

HOW TRUST AND SOCIAL CAPITAL WORK TOGETHER As we noted in the chapter’s introduction, the role of trust in social capital has been viewed in many diﬀerent ways. Some view trust as an important component of or necessary prerequisite for social capital (for example, Putnam, 1993). Others see the two in a reciprocal relationship where trust aids in social capital’s formation and the successful utilization of social

Trust, distrust and building social capital

287

capital increases trust between participants (see Nooteboom, 2007). Trust’s role in bringing about positive outcomes is also a point of debate, ranging from a direct cause to an approach where it facilitates other factors that create the beneﬁts (that is, is a moderating variable). More positive attitudes, higher levels of cooperation and superior levels of performance could produce direct beneﬁts – this view of trust, as a driver of gains, dominates the research. Despite the dominance of the ‘direct’ viewpoint, trust’s eﬀect on social capital may also be indirect. Although the literature and support for the indirect model are relatively underdeveloped (Dirks and Ferrin, 2001), the concept warrants further investigation. One proposition suggests that trust facilitates the eﬀects of motivation on workplace behaviors and outcomes by shaping expectations of another person’s future behavior. Models that treat trust as a predictor (for example, O’Reilly and Roberts, 1974) or that position it as a mediator (see Dirks and Ferrin, 2001) see a sequence where trust leads to risk-taking which then leads to positive outcomes. As a result, trust directly aﬀects the formation of reciprocity norms. A second proposition suggests that trust moderates the relationship between a partner’s action and the trustor’s response by influencing the interpretation of the behavior (Dirks and Ferrin, 2001). High trust will probably form closer network bonds as relevant individuals are perceived more favorably. This model suggests that trust provides the conditions for certain outcomes, for example cooperation and higher performance, to occur. In sum, trust is a necessary but not suﬃcient ingredient for creating and sustaining social capital. Trust requires a willingness to accept vulnerability based on positive expectations of a trustee’s behavior (Rousseau et al., 1998); however, it does not imply that the trustee will act in a way that will beneﬁt the trustor. To clarify this diﬀerence, we will now outline some crucial distinctions between trust and social capital. Trust as a Reservoir Trust is generally treated at an individual, unilateral level (that is, A’s trust for B). But we must also examine trust across time, whether at a single point in time (bilateral trust between A and B) or added over time (aggregate bilateral trust between A and B). We refer to the aggregate bilateral trust for one relationship as a ‘reservoir of trust’; this label also applies to the aggregate of trust across multiple relationships. Either party can draw on this reservoir as needed. For example, you may have a high degree of trust in a colleague, but not feel comfortable asking them, nor expect them, to help you with personal family needs. In other words, you may expect a person you trust not to harm you intentionally or to do what they promise

288

Table 11.2

Social capital: reaching in

Contributors to enhanced or reduced social capital Enhanced social capital

Reduced social capital

Definition

A strong, positive, productive reservoir of good will in a relationship

A low reservoir of good will, and perhaps evidence of bad will in a relationship

Network ties

Positive network ties

Negative network ties

Norms

Positive, productive norms

Values

Relationship-supporting (honesty, integrity, cooperation)

Negative, dysfunctional norms Relationship-debilitating (dishonesty, low integrity, destructive competition)

Expectations

Benevolent and anticipating trusting behavior from the other

Malevolent and anticipating distrusting behavior from the other

Trust dynamics

Dominantly trust-based

Dominantly distrust-based

to do, but not anticipate that they would support you during an emergency. However, relational norms in some relationships extend beyond the collegial context; as a result, support, help and assistance are available on a more personal level. Factors that work with the trust reservoir to enhance social capital Several network characteristics help build and expand trust reservoirs. In organizations these elements also work with reservoirs to create strong social capital. They include: productive network ties; positive and productive norms; benevolent, positive expectations; and relationship-supporting values. We will now examine each of these in turn. a. Productive network ties Social networks are a fundamental aspect of social capital. They consist of nodes and connections between the nodes. In social capital, the nodes are typically other individuals, and the connections are one’s relationships with those other network members. Networks contribute to social capital by oﬀering access to other individuals who control valued resources and with whom one can exchange resources or contributions. Such networks help form trust quickly in the beginning of a relationship, particularly when direct experience with the other party is limited. Network members are often able to form perceptions of other members despite incomplete information based on reputation and other network information. Their willingness to trust early then raises the level of generalized interpersonal trust (Bartkus and Davis, 2008). The

Trust, distrust and building social capital

289

overall level of trust can also facilitate cohesion and connectivity within the network (see Krackhardt and Hanson, 1993; Davis and Bartkus, (Chapter 13 in this volume; Lin, 2001; Putnam, 1995) and help spread trust beyond direct individual experience (see Kramer, Chapter 9 in this volume; Putnam, 1993). b. Positive, productive norms As previously discussed, networks serve as the backbone, or infrastructure, through which social capital manifests and operates, but the existence of a network does not necessarily result in the existence of social capital. Positive, productive norms must also exist, likely moderating the relationship between network members. When positive, productive norms are high, these network relationships may give rise to social capital; when positive, productive norms are low, network relationships are less inclined to facilitate social capital. Trust impacts upon social capital through enhancing both positive norm development directly, and indirectly by supporting the structural integrity of the network in which positive norms can develop (that is, network eﬃciency is diminished in a state of low trust or high distrust). For example, norms are important for the formation of organizational citizenship behaviors (that is, employee discretionary behaviors that, in aggregate, promote the organization’s eﬀective functioning; Ehrhart and Naumann, 2004; Organ, 1988). Research on team functioning has also revealed the importance of norms. These help members anticipate behaviors from their teammates (Feldman, 1984). In this sense, norms can increase conﬁdence in the positive exchange required to form and maintain strong social capital. c. Benevolent, positive expectations Based on the deﬁnitions discussed earlier, trust entails beliefs and expectations about another party’s intentions and the willingness to accept vulnerability based on those expectations (see Lewicki et al., 1998; Mayer et al., 1995; Rousseau et al., 1998). Trust therefore implies a degree of certainty about another’s behavior, but no assurance that the other party will act on your behalf when needed. Strong social capital, however, is based on trust in another’s willingness to act on your behalf in a speciﬁc manner and context, even when details have never been explicitly discussed. d. Relationship-supporting values Values are personally held, internalized guides of behavior (Hall, 1995). They represent beliefs about optimal conduct and often form the basis upon which individuals make life choices (Rokeach, 1973). Certain mutually held values (for example, honesty, integrity, cooperation and mutual support) are critical for healthy interpersonal relationships (Rogers, 1961). Although behaviors inconsistent

290

Social capital: reaching in

with these values can occur when motivation is strong enough, actions will tend to align with values over the long term. Given the fact that social capital takes some time to develop and that true values drive long-term behaviors, a group with mutually held, positive values would develop social capital more eﬀectively. Summary The trust reservoir can help create strong social capital. This occurs when it is coupled with positive network ties, functional and promotive norms, positive expectations of others, and relationship-strengthening values. The cycle then becomes self-reinforcing as strong social capital emerges and trust grows at a group level, raising the level of the reservoir even higher. Factors that detract from trust reservoirs and weaken social capital The opposites of the elements above can actively reduce social capital: negative or unproductive network ties; negative, unproductive norms; malevolent, negative expectations; and relationship-debilitating values. These factors may also then drain the trust reservoirs for individual relationships. a. Negative network ties Not all relationships with other individuals in an organization are positive. We would prefer to avoid interacting with many people but have to, based on work context. These negative ties may result from past interactions where parties have acted expediently, or exploited benevolent gestures, or betrayed good faith gestures and violated trust. Unfortunately, every negative occurrence not only weakens social capital but diminishes the total reservoir. b. Dysfunctional norms Dysfunctional norms in organizations – for example, discrimination, excessive careerism, incivility – diminish social capital. Unfortunately, such norms are pervasive and motivated by a strong desire to conform – a quality that can, in diﬀerent circumstances, strengthen social capital (Lewicki et al., 2008). In contrast to positive norms, these increase negative expectations and may present obstacles to relationships. c. Negative expectations Negative expectations of others’ behavior increase distrust and diminish social capital. Lewicki et al. (1998) deﬁne distrust in terms of conﬁdent negative expectations regarding another’s conduct. In reciprocal terms, scholars describe distrust as the expectation that others will not act in one’s best interests or will not produce capable and responsible behavior. They may also engage in potentially harmful behavior (Barber, 1983; Govier, 1994).

Trust, distrust and building social capital

291

d. Relationship-debilitating values Whereas certain mutually held positive values are critical for healthy interpersonal relationships, negative values often debilitate relationships and diminish social capital. Such values include dishonesty, low integrity, a highly utilitarian and egoistic view of the world, and bigoted beliefs about race, gender or the value of community. Summary Social capital and trust reservoirs are often weakened by negative network ties, dysfunctional norms, negative expectations of others and relationshipdebilitating values. These qualities foster distrust, prompt distrust in the absence of direct relationships, and are enhanced by distrusting, untrustworthy actions.

THE IMPLICATIONS OF DIFFERENT FORMS OF TRUST AND DISTRUST FOR THE DEVELOPMENT OF SOCIAL CAPITAL This ﬁnal section looks at how trust can be created and distrust managed so that the reservoir that enhances social capital stays as high as possible. The summary follows the structure below: (1) understand factors that contribute to ‘quick’ calculus-based trust (CBT); (2) identify further factors to strengthen calculus-based trust; (3) develop identiﬁcation-based trust (IBT); (4) manage calculus-based distrust (CBD); and (5) manage identiﬁcation-based distrust (IBD). Contributors to ‘Quick’ Calculus-Based Trust and Social Capital The literature has revealed that trust often begins above a ‘no trust’ or ‘zero’ level at the beginning of a relationship. Several factors, otherwise known as ‘swift trust’ contributors, appear to drive this initial trust. They include positive reputations, category-, role- and rule-based trusts, and past history of interaction within the group. Others’ positive reputation Some initial CBT is based on the other’s reputation (Gabarro, 1978; Butler, 1991). This reputation may be grounded in what one has heard through social contacts, professional and personal relationships, or through direct experience with the other’s trustworthiness (Mayer et al., 1995). These researchers argued that three items precede trust: the trusting party’s perception of the other party’s ability to do what they say they can do; their

292

Social capital: reaching in

integrity (that is, their honor in doing what they say they will); and their benevolence toward the trusting party. Category-based trust If the other belongs to a social or organizational category that we believe is relevant when assessing trustworthiness, our presumptive trust may increase. This trust may be either CBT or IBT, depending on the basis we use. For example, if we walk into a doctor’s oﬃce and greet a person in a white coat with a stethoscope around his neck, we place CBT in that individual based on the presumption that the uniform conveys professional training and competence as a doctor. In contrast, if we are walking down the street and encounter a person who is wearing a hat from our undergraduate alma mater, we probably trust that individual more because of the presumed common ‘identiﬁcation’ with institutional loyalty, undergraduate experience, common values, and so on. Research by Brewer (1981, 1991) indicates that if the other belongs to our social category – for example, gender, race, nationality, political disposition – we tend to assume some level of trustworthiness exists (relative to others who are not of the same category). Moreover, we tend to attribute other positive characteristics to such people, including honesty and cooperativeness. Role-based trust If the other occupies a particular professional role – a physician, an accountant, a dentist, a law enforcement oﬃcer – it may increase our initial trust. First, we tend to assume that it takes signiﬁcant skill and competence to fulﬁll this role, including thorough screening, rigorous training, extensive testing and certiﬁcation. Second, most of those professions require members to abide by a publicly understood code of professional conduct that aﬃrms their commitment to professional values and to treating customers and clients according to this code (cf Dawes, 1994; Meyerson et al., 1996). In short, trustors are often willing to trust members in a role that embodies ability, benevolence and integrity (Mayer et al., 1995) without knowing whether the trustee deserves trust. Rule-based trust If rules, laws, boundaries or explicitly articulated norms about what constitutes appropriate behavior constrain others’ behavior, we are also more willing to trust. Airlines can assemble and reassemble crews of pilots and flight attendants who may not have worked together before because they all agree to follow standard operating procedures for flight, safety, and so on. Medical teams practice emergency and triage so that they can work together quickly and eﬀectively with very little discussion and miscoordi-

Trust, distrust and building social capital

293

nation when the true need arises. Fine and Holyﬁeld (1996) note that such rules function in three ways: they allow parties to ‘trust’ another even when their personal behavior may not inspire conﬁdence; they sustain trust by assuring that new members to such teams are rigorously trained and monitored (role-based trust); and they ‘transform’ trust by demonstrating that trust without intense personal scrutiny is acceptable and viable. Past history of interaction If the parties have enjoyed successful social interactions in the past, they will probably be more willing to presume trust. ‘Successful’ means that both parties beneﬁted from the interaction and that each made a commitment to reciprocate in the relationship (for example, Deutsch, 1958; Lindskold, 1986). Kramer (1999) notes that the deﬁnition of success is based on the a priori expectations that the other will act trustworthily, and that the expectations will change to the degree that experience validates or invalidates them. Further Developing Calculus-Based Trust CBT develops over time as we interact with the other party and identify their behavior patterns. Forward progress is often slow but steady. People prove their trustworthiness through the simple actions of making and meeting commitments; at the same time, they test the other’s trustworthiness and trust. Lewicki and Bunker (1995) suggest that the children’s game ‘Chutes and Ladders’ is an appropriate metaphor for CBT’s growth. Players advance on the game board by throwing the dice and moving ahead (‘up the ladder’) in a series of steps. In contrast, landing on a chute may move the player back one or more steps or, in the worst case, back to the beginning. In the same way, a single inconsistent or unreliable experience may move CBT back one or more steps – or, if the violation is serious enough, back to the beginning. As this example demonstrates, CBT can be fragile. For strong CBT to develop, the recipient must signal that they will receive more trust well, and the actor must be clear that he will take these actions only if they are strongly reciprocated and not abused. More formalized agreements are sometimes used to reinforce these messages. Three factors facilitate the process, as outlined below. Predictability, consistency and reliability Previous research has demonstrated that eﬀective business relationships are based on predictability (Jennings, 1971). CBT can be enhanced if we can anticipate the other’s actions (that is, know how they will respond to

294

Social capital: reaching in

commonplace experiences), and the other responds consistently to similar stimuli (Gabarro, 1978). CBT also increases if the other can be relied upon to do what they say they will do (that is, perform as expected, meet expectations and follow through on commitments) (McAllister, 1995). Positive emotions While we often view calculus-based trust as highly rational and ‘computational’, positive emotional displays – for example, happiness and gratitude – can also build trust in this form. Viewing another as ‘human’ and witnessing a genuine emotional display can increase trust signiﬁcantly (Dunn and Schweitzer, 2005). Third parties as mediators of trust Trusted third-party intermediaries can ‘leverage’ trust to others who may not be trusted. Mediators, facilitators and other ‘third siders’ (Ury, 2000) create trust by brokering open, honest communication; they also uncover common ground and prevent trust violations by monitoring each side’s behavior. These third parties use the trust placed in them to help others engage in behaviors which, over time, build CBT between them. Developing Identiﬁcation-Based Trust Collective identity Certain activities strengthen IBT (Shapiro et al., 1992; Lewicki and Bunker, 1995, 1996; Lewicki and Stevenson, 1998). Such actions can involve developing a collective identity (for example, a joint name, title or logo), colocating in the same building or neighborhood, creating joint products or goals (for example, a new product line or a new set of objectives), or committing to commonly shared values (so that the parties can substitute for each other in external transactions). At the organization level, Kramer (2001) argued that identiﬁcation with the organization’s goals leads individuals to trust the organization and others within it. Furthermore, he argued that common group and organizational membership could increase trust far more than ‘reputation’ or calculative-based considerations. Rather, common group membership creates actions with symbolic meanings. As a result: ‘engaging in acts of trust thus provides organizational members with an opportunity to communicate to others the symbolic value they attach to their organizational identity. From this perspective, the psychological signiﬁcance of trust acts resides . . . in the social motives and aﬃliative needs of group members that are met through such actions’ (Kramer, 2001, p. 171).

Trust, distrust and building social capital

295

Shared values and goals IBT develops as one knows and predicts the other’s needs, preferences and choices, and as one also starts to ‘own’ some of those same needs, preferences and choices. If we believe that the other shares our concerns and goals, or we observe the other reacting as we believe we would react, IBT is enhanced (Lewicki and Stevenson, 1998). Other research supports this statement. Gabarro (1978) found that spending time together sharing personal values, goals, motives and perceptions enhanced trust. Lewicki and Stevenson (1998) proposed that IBT develops in situations where the parties possess the same values and principles, which they interpret as integrity. Moreover, organizational studies have indicated that one component of managers’ trust in their subordinates is the degree to which the employee demonstrates that they have the best interests of the manager or the organization (or both) at heart (Schoorman et al., 1996; Butler, 1991). Enabling emotional ‘connections’ IBT also has a signiﬁcant aﬀect component (Lewicki and Bunker, 1995, 1996; McAllister, 1995). How we respond to others often depends on our immediate aﬀective reactions to their self-presentation (Chaiken, 1986), the situation and circumstances under which we meet (Jones and Brehm, 1976), and/or our mood at the time of the encounter (Forgas and George, 2001). Given this reality, IBT is likely to occur only with those whom we believe share our goals, interests, perceptions and values, and whom we encounter in circumstances that provide a positive aﬀective impression. Managing Calculus-Based Distrust CBD may also begin at a non-zero value. Such ‘non-zero, negative trust’ (distrust) often indicates low social capital. Factors that contribute to an initial negative judgment comprise the other’s negative reputation and/ or a negative history of interaction with other parties, including unfulﬁlled commitments, broken agreements, and even dishonesty, deception and ‘cheating’ on relationship expectations. How can we manage – and minimize – this distrust? Recognize the existence of strong distrust The ﬁrst step is recognizing the typical indicators of dysfunctional distrust. They consist of: (1) a strong aversion to working with others because the perceived vulnerability of trust outweighs the perceived beneﬁts of cooperation; (2) suspicion about the other’s motives and intentions; (3) a past history of broken agreements and unkept promises on the part of the trustee; and (4) negative characterizations of the other. Analyzing the

296

Social capital: reaching in

origins of this distrust and trying to understand the situation from the other party’s point of view may ameliorate the situation (Kemp and Smith, 1994). Create explicit agreements for managing the relationship Probably the most important aspect of managing distrust is the creation of explicit agreements for conducting future transactions, with penalties for non-compliance. These explicit agreements and ‘boundaries’ will probably reduce the fear parties may have about conducting transactions with each other. Agree on procedures for monitoring and verifying the other’s actions Procedures and processes need to be in place to monitor distrustful actions and ensure they do not occur. Osgood (1962), writing about disarmament procedures during the Cold War, suggested ways parties could monitor each other while initiating small, clear, unilateral steps that would signal trustworthiness and ultimately rebuild trust. Eﬀective coordination of these steps with statements keeps words and actions consistent and demonstrates commitment to keeping one’s word. Insisting on reciprocal actions from the other indicates that this must be a mutually beneﬁcial process (Lindskold, 1986). Third parties sometimes oversee adherence to rules as well as monitor and enforce rule infringements. Provide for suﬃcient deterrence (punishment) for violating agreements Parties need to agree to ﬁnes and punishments for trust violations. If such penalties are sizeable, they provide an eﬀective deterrent to future violations. The threat of publicizing these infringements and damaging the other’s reputation may also serve as a restraint. Apologize and attempt to repair trust violations When trust is violated, Tomlinson et al. (2004) found that victims are more willing to reconcile with oﬀenders who oﬀer an explicit, sincere and timely apology rather than those who merely attempt to placate them. Furthermore, they found that if the violator has a good relationship history with the victim, any digression is more likely to be seen as the exception rather than the rule. Cultivate ways to minimize interdependence and future interaction with the other while still meeting one’s needs Distrust can also be managed by minimizing the parties’ interdependence and ﬁnding alternative ways to meet their needs. In negotiations, Fisher et al. (1991) refer to this as a ‘best alternative’. While this approach does not necessarily strengthen trust, it also does it no further harm; it may also help

Trust, distrust and building social capital

297

the parties ‘cool oﬀ’ and restore a modicum of civility to their relationship. Third parties are also relevant here. They can help structure, manage, limit and control the communication and interaction between parties so that inflammatory and distrust-building actions by one or both parties are supervised and monitored. Managing Identiﬁcation-Based Distrust Because IBD is considered to be deeper and more complex than CBD, it may prove more problematic. When IBD exists, it is often the result of highly unanticipated, inappropriate or incorrect actions on the part of the trustee, or of a climate of suspicion, distrust or cynicism. In some cases, the trustor may begin to question whether or not she truly belongs in this relationship or group. Even though the situation is much more diﬃcult, a few techniques can help manage the consequences of identiﬁcation-based distrust, as outlined below. Explore the validity of identity-based diﬀerences One strategy is to investigate the true source of perceived identity-based diﬀerences, which may be more imagined than real. As a wealth of research on the nature of identity has shown (for example, Kramer, 2001), social categorization within groups and organizations can be both functional and dysfunctional. The awareness that one is ‘diﬀerent’ from others can lead to stereotyping and negative judgment, which may lead to paranoid cognitions and signiﬁcant distrust of others (see Kramer, 1999 for one review). Eﬀorts to discuss the perceived diﬀerences (often with a facilitator) may help parties recognize commonalities of which they were not previously aware. Their initial negative judgments may change as they get to know the other, share perceptions and test the validity of those judgments. Openly acknowledge areas of contention Openly acknowledging key areas of contention and avoiding situations that highlight these points may also be an eﬀective management strategy. If one recognizes that repeatedly discussing areas of clear ideological disagreement with the other only worsens the situation, attempt to avoid the topic. Both parties should try to ‘agree to disagree’ and ﬁnd alternative topics of conversation. Develop a mutually acceptable process for controlled interaction Although IBD may never be completely alleviated, developing mutually acceptable processes to manage required interaction and steer around areas of contention may minimize some of its negative impact. Parties may then

298

Social capital: reaching in

be able to see that they share some core values – for example, justice, civility, respect – despite strong ideological diﬀerences around other values, principles and commitments. They may then be able to create a carefully structured, civil working relationship. Third parties can once again play a critical role in shaping and monitoring future interactions (Burgess and Burgess, 1996).

SUMMARY In this chapter, we have explored the relationship between trust and social capital. While the two constructs are seen as strongly related to each other and even talked about as synonymous with each other, we have worked here to describe a more complex interrelationship. We reviewed the recent developments in trust research, and indicated some important developments in conceptualizing trust: that trust can take diﬀerent forms (calculus-based or identiﬁcation-based, or calculative and relational); that trust and distrust can be treated as separate constructs; and that even in ‘beginning’ relationships, trust may not begin at zero. We then addressed the implications of each of these developments for a more complex understanding of trust and social capital. The challenge for developing and sustaining social capital requires a complex orchestration of building trust and managing distrust. While eﬀorts to build trust are important and admirable – and we have described a number of courses of action to accomplish this – speciﬁc (and often diﬀerent) actions must be taken to manage and reduce distrust. In situations of low social capital, stabilizing and managing distrust may be the ﬁrst, and most important, step toward developing and strengthening trust and enhancing the complex fabric of social capital.

REFERENCES Barber, B. (1983), The Logic and Limits of Trust, New Brunswick, NJ: Rutgers University Press. Berg, J., J. Dickhaut and K. McCabe (1995), ‘Trust, reciprocity, and social history’, Games and Economic Behavior, 10, 122–42. Bigley, G.A. and J.L. Pearce (1998), ‘Straining for shared meaning in organization science: problems of trust and distrust’, Academy of Management Review, 23, 405–22. Brewer, M.B. (1981), ‘Ethnocentrism and its role in interpersonal trust’, in M.B. Brewer and B.E. Collins (eds), Scientiﬁc Inquiry and the Social Sciences, New York: Jossey-Bass, pp. 214–31.

Trust, distrust and building social capital

299

Brewer, M.B. (1991), ‘The social self: on being the same and diﬀerent at the same time’, Personality and Social Psychology Bulletin, 17, 475–82. Burgess, H. and G. Burgess (1996), ‘Constructive confrontation: a transformative approach to intractable conflicts’, Mediation Quarterly, 13, 305–22. Butler, J.K. (1991), ‘Toward understanding and measuring conditions of trust: evolution of a conditions of trust inventory’, Journal of Management, 17, 643–63. Chaiken, S.L. (1986), ‘Physical appearance and social influence’, in C.P. Herman, M.P. Zanna and E.T. Higgins (eds), Physical Appearance, Stigma, and Social Behavior: The Ontario Symposium, Mahwah, NJ: Erlbaum, pp. 142–77. Chou, Y.K. (2006), ‘Three simple models of social capital and economic growth’, Journal of Socio-Economics, 35, 889–912. Cialdini, R.B. (1996), Influence: Science and Practice, 4th edn, New York: HarperCollins. Coleman, J.S. (1990), Foundations of Social Theory, Cambridge, MA: Harvard University Press. Cummings, L.L. and P. Bromiley (1996), ‘The organizational trust inventory (OTI): development and validation’, in R.M. Kramer and T.R. Tyler (eds), Trust in Organizations: Frontiers of Theory and Research, Thousand Oaks, CA: Sage Publications, pp. 302–30. Davis, J.H. and V. Bartkus (2008), ‘Organizational trust and social capital’, Chapter 13 in this volume. Dawes, R.M. (1994), House of Cards: Psychology and Psychotherapy Built on Myth, New York: Free Press. De Clercq, D., N. Thongpapanl and D. Dimov (2007), ‘The role of conflict and social capital in cross-functional collaboration’, paper presented at the 4th Workshop on Trust Within and Between Organizations, 25–26 October. Deutsch, M. (1949), ‘A theory of cooperation and competition’, Human Relations, 2, 129–51. Deutsch, M. (1958), ‘Trust and suspicion’, Journal of Conflict Resolution, 2, 265–79. Deutsch, M. (1962), ‘Cooperation and trust: some theoretical notes’, in M.R. Jones (ed.), Nebraska Symposium on Motivation, Lincoln, NB: University of Nebraska Press, pp. 275–319. Deutsch, M. (1985), Distributive Justice: A Social Psychological Perspective, New Haven, CT: Yale University Press. Dirks, K.T. and D.L. Ferrin (2001), ‘The role of trust in organizational settings’, Organization Science, 12, 450–67. Dunn, J.R. and M.E. Schweitzer (2005), ‘Feeling and believing: the influence of emotion on trust’, Journal of Personality and Social Psychology, 88, 736–48. Ehrhart, M.G. and S.E. Naumann (2004), ‘Organizational citizenship behavior in work groups: a group norms approach’, Journal of Applied Psychology, 89, 960–74. Erickson, E.G. (1963), Childhood and Society, New York: W.W. Norton. Feldman, D.C. (1984), ‘The development and enforcement of group norms’, Academy of Management Review, 9, 47–53. Fine, G.A. and L. Holyﬁeld (1996), ‘Secrecy, trust, and dangerous leisure: generating group cohesion in voluntary organizations’, Social Psychology Quarterly, 59, 22–38. Fisher, R., W. Ury and B. Patton (1991), Getting to Yes, 2nd edn, New York: Penguin. Flores, F. and R.C. Solomon (1998), ‘Creating trust’, Business Ethics Quarterly, 8 (2), 205–32.

300

Social capital: reaching in

Forgas, J.P. and J.M. George (2001), ‘Aﬀective influences on judgments and behavior in organizations: an information processing perspective’, Organizational Behavior and Human Decision Processes, 86, 3–34. Gabarro, J.J. (1978), ‘The development of trust, influence and expectations’, in A.G. Athos and J.J. Gabarro (eds), Interpersonal Behavior: Communication and Understanding in Relationships, Upper Saddle River, NJ: Prentice Hall, pp. 290–303. Gladwell, M. (2000), The Tipping Point, Boston, MA: Little Brown. Govier, T. (1994), ‘Is it a jungle out there? Trust, distrust, and the construction of social reality’, Dialogue, 33, 237–52. Helliwell, J. and R. Putnam (1995), ‘Economic growth and social capital in Italy’, Eastern Economic Journal, 21, 295–307. Ho, V.T., D.M. Rousseau and L.L. Levesque (2006), ‘Social networks and the psychological contract: structural holes, cohesive ties, and beliefs regarding employer obligations’, Human Relations, 59, 459–81. Jennings, E.E. (1971), Routes to the Executive Suite, New York: McGraw-Hill. Jones, G.R. and J.M. George (1998), ‘The experience and evolution of trust: implications for cooperation and teamwork’, Academy of Management Review, 23, 531–46. Jones, R.A. and J.W. Brehm (1976), ‘Attitudinal eﬀects of communicator attractiveness when one chooses to listen’, Journal of Personality and Social Psychology, 6, 64–70. Kelley, H.H. and A.J. Stahelski (1970), ‘Social interaction basis of cooperators’ and competitors’ beliefs about others’, Journal of Personality and Social Psychology, 16, 66–91. Kemp, K.E. and W.P. Smith (1994), ‘Information exchange, toughness and integrative bargaining: the roles of explicit cues and perspective-taking’, International Journal of Conflict Management, 5, 5–21. Knorringa, P. and I. van Staveren (2006), Social Capital for Industrial Development: Operationalizing the Concept, Vienna: UNIDO. Krackhardt, D. and J. Hanson (1993), ‘Informal networks: the company behind the chart’, Harvard Business Review, July, 71 (4), 104–11. Kramer, R.M. (1994), ‘The sinister attribution error: paranoid cognition and collective distrust in organizations’, Motivation and Emotion, 18, 199–230. Kramer, R. (1999), ‘Trust and distrust in organizations: emerging perspectives, enduring questions’, Annual Review of Psychology, 50, 569–98. Kramer, R. (2001), ‘Identity and trust in organizations: one anatomy of a productive but problematic relationship’, in M. Hogg and D.J. Terry (eds), Social Identity: Processes in Organizational Contexts, Philadelphia, PA: Psychology Press, pp. 167–234. Kramer, R.M. (2008), ‘Social capital creation: collective identities and collective action’, Chapter 9 in this volume. Latham, M. (1997), ‘The search for social capital’, paper presented at the There Is Such a Thing as Society, Sydney, January. Lewicki, R.J. (2006), ‘Trust, trust development, and trust repair’, in M. Deutsch and P.T. Coleman (eds), The Handbook of Conflict Resolution: Theory and Practice, 2nd edn, San Francisco, CA: Jossey-Bass, pp. 92–109. Lewicki, R.J. and B.B. Bunker (1995), ‘Trust in relationships: a model of trust development and decline’, in B.B. Bunker and J.Z. Rubin (eds), Conflict, Cooperation and Justice, San Francisco, CA: Jossey-Bass, pp. 133–73.

Trust, distrust and building social capital

301

Lewicki, R.J. and B.B. Bunker (1996), ‘Developing and maintaining trust in work relationships’, in R.M. Kramer and T.R. Tyler (eds), Trust in Organizations: Frontiers of Theory and Research, Thousand Oaks, CA: Sage Publications, pp. 114–39. Lewicki, R.J. and M. Stevenson (1998), ‘Trust development in negotiation: proposed actions and a research agenda’, Journal of Business and Professional Ethics, 16, 99–132. Lewicki, R.J. and C. Wiethoﬀ (2000), ‘Trust, trust development, and trust repair’, in M. Deutsch and P.T. Coleman (eds), The Handbook of Conflict Resolution: Theory and Practice, San Francisco, CA: Jossey-Bass, pp. 86–107. Lewicki, R.J., D.J. McAllister and R.J. Bies (1998), ‘Trust and distrust: new relationships and realities’, Academy of Management Review, 23, 438–58. Lewicki, R.J., E. Tomlinson and N. Gillespie (2006), ‘Models of interpersonal trust development: theoretical approaches, empirical evidence and future directions’, Journal of Management, 32 (6), 991–1022. Lin, N. (2001), Social Capital: A Theory of Social Structure and Action, Cambridge: Cambridge University Press. Lindskold, S. (1986), ‘GRIT: reducing distrust through carefully introduced conciliation’, in W.G. Austin and S. Worchel (eds), The Social Psychology of Intergroup Relations, Chicago, IL: Nelson-Hall Publishers. Macinko, J. and B. Starﬁeld (2001), ‘The utility of social capital in research on health determinants’, Milbank Quarterly, 79, 387–427. Malhotra, D. (2004), ‘Trust and reciprocity decisions: the diﬀering perspectives of trustors and trusted parties’, Organizational Behavior and Human Decision Processes, 94, 61–73. Mayer, R.C., J.H. Davis and F.D. Schoorman (1995), ‘An integrative model of organizational trust’, Academy of Management Review, 20, 709–34. McAllister, D. (1995), ‘Aﬀect- and cognition-based trust as foundations for interpersonal cooperation in organizations’, Academy of Management Journal, 38 (1), 24–59. McAllister, D. (1997), ‘The second face of trust: reflections on the dark side of interpersonal trust in organizations’, in R.J. Lewicki, R. Bies and B.H. Sheppard (eds), Research on Negotiation in Organizations, Vol. 6, New York: JAI Press, pp. 87–111. McKnight, D.H., L.L. Cummings and N.L. Chervany (1998), ‘Initial trust formation in new organizational relationships’, Academy of Management Review, 23, 473–90. Meyerson, D., K. Weick and R. Kramer (1996), ‘Swift trust and temporary groups’, in R.M. Kramer and T.R. Tyler (eds), Trust in Organizations: Frontiers of Theory and Research, New York: Sage Publications, pp. 166–95. Miller, P.J. and J.K. Rempel (2004), ‘Trust and partner-enhancing attributions in close relationships’, Personality and Social Psychology Bulletin, 30 (6), 695–705. Mitchell, A.D. and T.J. Bossert (2007), ‘Measuring dimensions of social capital: evidence from surveys in poor communities in Nicaragua’, Social Science and Medicine, 64, 50–63. Mohseni, M. and M. Lindstrom (2007), ‘Social capital, trust in the health-care system and self-rated health: the role of access to health care in a populationbased study’, Social Science and Medicine, 64, 1373–83. Nahapiet, J. and S. Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23, 242–66.

302

Social capital: reaching in

Nooteboom, B. (2007), ‘Social capital, institutions and trust’, Review of Social Economy, 65, 29–53. North, D.C. (1990), Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press. Onyx, J. and P. Bullen (2000), ‘Measuring social capital in ﬁve communities’, Journal of Applied Behavioral Science, 36, 23–42. O’Reilly, C.A. and K.H. Roberts (1974), ‘Information ﬁltration in organizations: three experiments’, Organizational Behavior and Human Performance, 11, 253–65. Organ, D.W. (1988), Organizational Citizenship Behavior: The Good Soldier Syndrome, Lexington, MA: Lexington Books. Osgood, C. (1962), An Alternative to War or Surrender, Urbana, IL: University of Illinois Press. Pasotti, E. and B. Rothstein (2002), ‘In the market for ideas: a quest for a uniﬁed conceptual approach in political science’, paper presented at the Annual Meeting of the American Political Science Association, Boston, MA, 29 August–1 September. Portes, A. (1998), ‘Social capital: its origins and applications in modern sociology’, Annual Review of Sociology, 24, 1–24. Putnam, R. (1993), Making Democracy Work: Civic Traditions in Modern Italy, Princeton, NJ: Princeton University Press. Putnam, R. (1995), ‘Bowling alone: America’s declining social capital’, Journal of Democracy, 6, 65–78. Rempel, J.K., J.G. Holmes and M. Zanna (1985), ‘Trust in close relationships’, Journal of Personality and Social Psychology, 49, 95–112. Rogers, C.R. (1961), On Becoming a Person: A Therapist’s View of Psychotherapy, Boston, MA: Houghton Miﬄin. Rokeach, M. (1973), The Nature of Human Values, New York: Free Press. Rotter, J.B. (1967), ‘A new scale for the measurement of interpersonal trust’, Journal of Personality, 35, 651–65. Rotter, J.B. (1971), ‘Generalized expectancies for interpersonal trust’, American Psychologist, 26, 443–52. Rousseau, D.M., S.B. Sitkin, R.S. Burt and C. Camerer (1998), ‘Not so diﬀerent after all: a cross-discipline view of trust’, Academy of Management Review, 23, 393–404. Sandefur, R.L. and E.O. Laumann (1998), ‘A paradigm for social capital’, Rationality and Society, 10, 481–501. Schoorman, F.D., R. Mayer and J. Davis (1996), ‘Empowerment in veterinary clinics: the role of trust in delegation’, paper presented to the 11th Annual Meeting of the Society for Industrial and Organizational Psychology, San Diego. Schoorman, D., R. Mayer and J. Davis (2007), ‘An integrative model of organizational trust: past, present and future’, Academy of Management Review, 32 (2), 344–54. Shapiro, D.L., B.H. Sheppard and L. Cheraskin (1992), ‘Business on a handshake’, Negotiation Journal, 8, 365–77. Sheppard, B.H. and D.M. Sherman (1998), ‘The grammars of trust: a model and general implications’, Academy of Management Review, 23 (3), 422–37. Sitkin, S. (1995), ‘On the positive eﬀects of the legalization of trust’, in R.J. Bies, R.J. Lewicki and B.H. Sheppard (eds), Research on Negotiation in Organizations, Vol. 5, Greenwich, CT: JAI, pp. 184–217.

Trust, distrust and building social capital

303

Sitkin, S.B. and D. Stickel (1996), ‘The road to hell: the dynamics of distrust in an era of quality’, in R. Kramer and T. Tyler (eds), Trust in Organizations: Frontiers of Theory and Research: Newbury Park: Sage, pp. 216–45. Tansley, C. and S. Newell (2007), ‘Project social capital, leadership and trust: a study of human resource information systems development’, Journal of Managerial Psychology, 22, 350–68. Taylor, M. (1982), Community, Anarchy and Liberty, Cambridge: Cambridge University Press. Tomlinson, E.C. and R.J. Lewicki (2006), ‘Managing distrust in intractable conflicts’, Conflict Resolution Quarterly, 24, 219–28. Tsai, W. and S. Ghoshal (1998), ‘Social capital and value creation: the role of intraﬁrm networks’, Academy of Management Journal, 41, 464–76. Ury, W. (2000), The Third Side, New York: Penguin. Uzzi, B. (1997), ‘Social structure and competition in interﬁrm networks: the paradox of embeddedness’, Administrative Science Quarterly, 42, 35–67. Walker, G., B. Kogut and W. Shan (1997), ‘Social capital, structural holes and the formation of an industry network’, Organization Science, 8, 109–25. Worchel, P. (1979), ‘Trust and distrust’, in W.G. Austin and S. Worchel (eds), The Social Psychology of Intergroup Relations, Belmont, CA: Wadsworth, pp. 174–87.

12. Understanding social capital: in whom do we trust? Darryl Stickel, Roger C. Mayer and Sim B. Sitkin Since the 1990s there has been a growing interest in social capital, as evidenced by the chapters in this volume. In tandem with the interest has been a concern: empirical evidence shows trends that people are less likely to join groups, invest time and energy in community projects, or volunteer for a wide variety of causes. Overall, US society seems to be experiencing a more general decline in social capital being. That is, people are investing less time to create ‘the set of trust, institutions, social norms, social networks, and organizations that shape the interactions of actors within a society and are an asset for the individual and collective production of well-being’ (Sabatini, 2006). Social capital’s decline is of particular concern because of the wideranging beneﬁts associated with high levels of social capital, including improved children’s welfare, education, safety, economic prosperity, public health, individual well-being and democracy. With healthy social capital, we can also expect more flexible and timely responses to new issues because: ‘[a]ccumulated social trust allows groups and organizations, and even nations, to develop the tolerance sometimes needed to deal with conflicts and diﬀering interests’ (Cox, 1995). High levels of social capital also seem to correlate with more eﬃcient and eﬀective organizations, communities and governments. Despite the growing body of literature on this topic and increasing levels of concern expressed by scholars and government leaders, little progress seems to have been made in understanding how to slow or reverse the ongoing negative trend. In many treatments of social capital, trust seems to be a core issue. Since the late 1990s there has been an explosion of attention paid to it in the literature (Mayer, 2007). However, these developments have not been carefully examined in the social capital arena. We believe that such a more thorough consideration will build the foundation for an improved understanding of trust’s relationship with social capital. In this chapter we explore both the deﬁnition of trust and the links between social capital and trust. While building on existing deﬁnitions of 304

Understanding social capital

305

social capital, we ﬁrst propose a new way to conceptualize it which diﬀerentiates it from its antecedents and outcomes, and then proceed to argue that investments in social capital are a trust decision. We explore how individuals make the decision to trust others, and then examine social changes which may be leading to declines in trust in hopes of providing some insight into the ongoing declines in social capital production.

SOCIAL CAPITAL AND TRUST While the importance of social capital seems to be well accepted among researchers, there is less agreement about how to deﬁne it. Deﬁnitions in the extant literature vary considerably, but many tend to share certain characteristics. Many researchers include values in their deﬁnitions; others include networks or trust (Davis and Bartkus, 2006). We fully agree that these are important, but conceptualize it diﬀerently. We believe that the attributes of values, networks and trust do not constitute social capital per se. Rather, they are causal antecedents. Each of them contributes to a more global construct that aﬀects the ability of a collective to function eﬀectively. This construct’s outcomes are behaviors and investments, for example putting in time and eﬀort to help the collective be successful. These behaviors and investments are likewise not social capital, but often drive our interest in social capital. Values, networks and trust do not improve the collective but rather make it more likely that its members will behave in ways that do improve it. Our conceptualization of social capital is the bridge between variables like values, networks and trust, and outcomes they facilitate – cooperative and helping behaviors. It is an asset like ﬁnancial capital or human capital. It can be thought of as a resource or a capability which is distinct from how it is derived or how it is used. As such, social capital is the socially derived potential for actions that contribute to the collective. It is a resource that is distinct from both how it is created and how it is used. In this way, social capital increases the likelihood that a member or members will contribute to address a collective need. Like potential energy in physics, social capital is the capacity to produce action, whether or not any action is produced. Social capital’s antecedents – values, networks and trust – have strong eﬀects. However, we believe that a three-way interaction among the three main components of Putnam’s (1993) deﬁnition – networks, values and trust – is essential if a collective’s members are going to invest in social capital. For example, networks provide members with information and access to other members. A network’s ability to increase social capital,

306

Social capital: reaching in

however, also depends on values. Merely having access to others will not encourage investments of energy unless individual and/or collective values stress the importance of contributing to the collective. Even this combination may not be suﬃcient without trust. We posit that individuals must trust prior to making an investment in social capital. This deﬁnition, consistent with prior research (for example, Putnam, 1993), depicts a clear link between trust and social capital. Researchers appear to agree that trust is a key ingredient. It is probably not a coincidence that recent declines in social capital appear to be occurring at the same time as drops in the level of trust in a range of institutions (see Schoorman et al., 2007). Given our claim that trust is essential if individuals are going to act for the collective as well as data showing that trust is declining, we need to develop a clear understanding of trust’s characteristics and origins. If we hope to reverse the current downward trend in social capital, we also need to determine how to create more trust within various collectives. It is important that construct deﬁnitions be isomorphic – that is, consistent across levels of analysis (Mossholder and Bedeian, 1983; Rousseau, 1985). Without isomorphic deﬁnitions it becomes diﬃcult if not impossible to specify, for example, how trust at an individual level aﬀects trust at an interorganizational level. Although trust is considered at multiple levels (Schoorman et al., 2007), it is often deﬁned diﬀerently at those various levels. We argue that an understanding of trust is important to understanding social capital and, thus, an isomorphic deﬁnition of trust is fundamental to advancing our understanding of social capital. In an attempt to create a deﬁnition of trust that could be applied across multiple levels, Mayer et al. (1995) reviewed the literature and pointed out a number of diﬃculties with trust’s conceptualization. They illustrated that it was often confused with cooperation, predictability and various qualities of the trustee. Furthermore, the critical relationship between trust and risk was unclear. In addition, trust needed to be diﬀerentiated from its causes and outcomes. To respond to these issues, Mayer et al. (1995) developed a parsimonious model of trust, its causes and its outcomes. We briefly review this model in the next section and then examine its applicability for understanding trust in the context of social capital.

THE DEVELOPMENT OF TRUST Based on several prior theorists (for example, Coleman, 1990; Deutsch, 1958; March and Shapira, 1987), Mayer et al. (1995) recognized that trust is clearly related to risk. As a result, they deﬁned trust as a willingness to

Understanding social capital

307

be vulnerable to another party which cannot be monitored or controlled, based on the expectation that the party will perform a particular action important to the trustor. Trust is strongly connected to vulnerability in this model. Risk is inherent in vulnerability, which makes risk inherent in their deﬁnition. Mayer et al. then summarized the literature with respect to the qualities that made a trustee worthy of trust. Unlike prior models that included up to ten characteristics, Mayer et al. concluded that three were most often cited and were suﬃcient. They argued that these three – ability, benevolence and integrity – seemed to account for most of the variance in trust. Ability is the set of relevant skills and competencies for a particular situation or task. Because these vary, the level of trust in the same trustee may be diﬀerent for diﬀerent tasks or situations. For instance, a computer analyst may have excellent expertise in their technical domain but lack interpersonal and communication skills. They would garner a very high level of trust when there are glitches in the operation of the computer system, but would not be trusted to represent the organization to a key client. Benevolence is the perception that the trustee cares about the trustor. This characteristic captures the individual relationship between the two parties. It depicts the extent to which a party believes that a trustee has the trustor’s best interests at heart and would go out of their way to help the trustor. Clearly, a trustor is likely to be more willing to be vulnerable to someone who is perceived to care about the trustor and the trustor’s interests. Integrity is the perception that the trusted party follows a set of values that the trusting party accepts. The trustor believes that the trustee has acceptable values and has the courage to follow those values even in diﬃcult situations. A trustor is less willing to be vulnerable if the trustee lacks integrity. If the trustor does not agree with the trustee’s values, trust is unlikely. If the trustee espouses an acceptable set of values but does not always follow them, the trustor cannot assume on a given occasion that the trustee will act in a predictable way. Either situation will diminish the perception of integrity. Ability, benevolence and integrity are the foundation of the perception of a party’s trustworthiness. These three combine in complex, idiosyncratic ways. Their weighting and combinations vary from party to party and from situation to situation. Early in a relationship, it is easier to gather information about someone’s ability and integrity based on credentials or reputation. It takes longer to determine benevolence (Mayer et al., 1995), which cannot usually be surmised from external sources as it is speciﬁc to each relationship. In a longitudinal study of a change of performance appraisal

308

Social capital: reaching in

systems, Mayer and Davis (1999) found that while ability, benevolence and integrity all rose when a new appraisal system was introduced, the speciﬁc increases were not signiﬁcant. In contrast, trust, which was aﬀected by all three, rose signiﬁcantly. According to this model, trust leads to risk-taking in the relationship (RTR) with the other party. Regardless of the particular trustee, there is always some level of risk in a speciﬁc action. This risk must be overcome by trust before an action will occur. If the level of trust in the other party exceeds the amount of inherent risk, the trustor will engage in the risktaking behavior. If, on the other hand, the level of trust is lower than the inherent risk, the trustor will not perform the action. For example, a scholar may consider lending some rare books to a colleague. Loss or damage of the books would be a signiﬁcant loss to the owner. This is a risk that must be overcome. It matters not who damages or loses the books, the loss is the same. If the potential borrower is a long-time trusted collaborator or colleague, there may be suﬃcient positive data about that person to warrant taking the risk and lending the books. On the other hand, if there are reasons to question any of the trustworthiness factors or if there is simply not enough information about them (for example, a less established relationship), then the level of trust may not be suﬃcient. In this case the scholar would not lend the books. This deﬁnition of trust accomplishes several goals. Firstly, it diﬀerentiates trust from its antecedents. Trust is distinct from trustworthiness – a characteristic of the trustee – because it is the trustor’s reaction to their perceptions of the trustee and the resulting willingness to accept risk. Secondly, it diﬀerentiates trust from its outcomes. While trust is an intention to behave – that is, to take risk – RTR is actual risk-taking. Thirdly, it can be applied across multiple levels. As Schoorman et al. (1996, 2007) explain, the model was designed for use at multiple levels. The deﬁnitions of trust, ability, benevolence and integrity can all be applied to a person, a group or an organization. This flexibility allows us to consider trust issues within and between organizations and larger social entities while maintaining isomorphism of the constructs involved. This is a critical feature if we are to understand the relationship between trust and social capital.

INVESTMENT IN SOCIAL CAPITAL AS A TRUST DECISION We now consider the relationship between the trust and social capital. Putnam and others have questioned why people do not want to invest their

Understanding social capital

309

time and eﬀort in social capital. We reframe the question as follows: why would someone invest in social capital, and what might stop them? Although the answers to these questions are complex, we will focus here on the role that trust plays. We believe that the decision to invest in social capital is essentially a trust decision. Trust is a willingness to be vulnerable to another party when that party cannot be monitored or controlled. Similarly, one invests scarce resources for an uncertain future pay-oﬀ when building social capital. Consider, for example, a small group of people who have known each other well for many years and interact regularly. Person A is considering taking a risk and expending time, energy and other resources to assist a project that would beneﬁt the whole group. If Person A perceives that relationships in the group are strong, they are more likely to assume that other members will likewise expend resources for other group needs in the future. However, Person A’s decision is not based on an economic quid pro quo. Person A and the other members of the group do not form a ‘contract’. Furthermore, while the initial need is known, the future one to which another person may contribute is unknown. Person A is clearly taking a risk if they spend time, eﬀort and perhaps money on the initial project. If, later on, another person spends only those resources necessary to help themselves when the whole group has a need, Person A’s interests are damaged. Although they previously invested resources to take care of the rest of the group, A receives nothing in return and must deal with the new eﬀort as it pertains to themselves. As we move from relationships among small numbers of known parties to larger numbers of lesser-known parties, the willingness to trust also falls oﬀ. The cohesiveness of the small group is likely to drive members’ decisions to invest in social capital. However, as we move to a larger collective, the trust problem becomes more diﬀuse. Social exchange theory refers to this as moving to a generalized exchange. Malhotra and Murnighan (2002) point out that risk increases as the exchange generalizes. The diﬃculties in monitoring multiple parties are part of the issue. The problems with monitoring also exacerbate the ‘free-rider’ or ‘commons’ issue. This is the situation where a shared resource needs to be maintained by every member of the group. However, while every member beneﬁts from the resource, the cost of maintaining it is not allocated to any particular person or set of persons. As a result, some people will use the resource without contributing to it – a tactic known as ‘free-riding’. Social capital is very sensitive to the ‘free-rider’ problem in larger groups. Not only is monitoring more diﬃcult, but it is more likely that some members will coast along on their colleagues’ eﬀorts. If other group members see that others are ‘free-riding’, they may also become cynical and

310

Social capital: reaching in

stop contributing. This leads to a spiral of declining contribution and may completely halt social capital development. Another issue with large groups is the fact that it becomes harder to evaluate the trustworthiness of individual members in larger groups. In a very large organization, one might have no knowledge of many, most or even virtually all of the members. Mayer et al. (1995) accounted for this by considering the propensity to trust. Based on the work of Rotter (1967), they described propensity as something similar to a personality trait carried by a trustor from one situation to the next. It describes how trusting a person is in the absence of information about another person. People with high propensity are generally trusting – they tend to trust others even with little information. If a collective were ﬁlled with people with high propensity to trust, it would be more likely that members would invest in social capital. Members would assume that others in the collective would reciprocate when the appropriate time arises. Not everyone, however, has high propensity. If someone has low propensity and lacks knowledge about others in the collective, that person will probably not take risks. While propensity can ameliorate the need for high levels of reliable information, there are other means of addressing this requirement and other contextual factors that may allow for risk-taking. We will now explore these issues.

NETWORKS AND TRUST Networks are a diﬀerent area of research within social capital. More social capital appears to be present when networks are more extensive and stronger. We submit that trust is a key reason for this relationship. However, developed networks can build trust by allowing more people in the system to know one another, providing knowledge of unknown others through trusted sources, and oﬀering increased opportunities to interact with those others. The network thus reduces the uncertainty and risk inherent in investing in social capital. While contributing to social capital is still a riskier type of generalized exchange, as explained earlier, the access to more reliable information from a strong network mitigates that risk somewhat. However, if networks are weak or non-existent, relationships are limited to members with whom an individual has direct contact. Few or no ways exist to get information about others’ trustworthiness. In these cases trust is less likely to develop. It is important to note that networks and trust are interdependent. Parties who trust others more are more likely to form networks and share relevant information than members who do not trust others. Likewise, the

Understanding social capital

311

more developed the network, the more complete the information one can collect about others’ trustworthiness. There are a few important caveats that merit consideration. While networks can contribute to trust’s development, they provide no guarantee that trust will be enhanced. Being able to get accurate, credible information about others does not mean that one will trust them. If the others in the collective are not trustworthy, more information from the network could lower trust. Likewise, a high level of trust among a clique’s members might lead this group to isolate itself from the larger collective. While the clique’s development strengthens trust among its members, it may hinder their trust of members of the larger collective. Likewise, trust can encourage the development of networks, but does not necessarily cause them. For example, when trust substitutes for feeling connected, individuals may view networking as a less necessary use of scarce resources (for example, time and eﬀort) – even if it is still desirable (Sitkin, 1995).

CONTEXTUAL FACTORS Context forms the rules of the game by which people play. When little or nothing is known about the other parties in a relationship, context enhances one’s ability to predict how others will behave. It lets us understand how people’s behavior is constrained or encouraged in a given situation. The factors may include people’s reputations, motivations, and what they stand to gain or lose. This knowledge reduces uncertainty about what to expect from those parties and may make it easier to take risks. However, trust is diﬃcult when context is unclear because parties do not know how big a risk they are taking or how others are likely to act (Sitkin, 1995). This can occur if context is unknown or when people are subject to multiple roles. Not only might an individual be a professional, but they may also be a parent and member of a political cause, each of which could push the person’s actions in a diﬀerent direction (George and Brief, 1996). People then ﬁnd it more challenging to predict how they may be personally aﬀected and thus perceive more uncertainty. The role of context is pervasive. There are many examples of situations where we trust people about whom we have little or no personal knowledge. Take our relationships with our doctors. In the safety of a doctor’s oﬃce, many of us will remove our clothes, disclose personal, possibly embarrassing information, and allow the doctor to poke and prod at us. There are few other situations where we give a relative stranger such personal contact and knowledge. Similarly, we trust our lives to strangers if they are, for example,

312

Social capital: reaching in

the pilots of our plane. In the absence of a high level of propensity, why would one take such major risks with strangers? A helpful approach to understanding the choices people make in such situations is through the use of Stickel’s (1999) matrix. In it, Stickel focuses on various sources of influence that constrain others’ actions and make their behavior more predictable. These factors reduce our uncertainty by helping us determine how helpfully we expect the other party to behave or, conversely, how much damage the other party might inflict. He asserts that contextual influences can be classiﬁed by whether they are within the immediate situation (micro) or embedded in the larger outside environment (macro), and by whether the mechanisms are formal or informal. Both the micro and macro levels have rules and norms, as well as expected beneﬁts if people play by the rules and expected consequences if they do not. Stickel’s work directs our attention to four broad categories of sources that make it easier to predict a person’s behavior in a situation. Trusting parties can then better manage the risk in dealing with that individual. The matrix in Figure 12.1 was developed for a work environment, which governed the speciﬁc examples within each cell. However, these line items can be adapted to many other contexts. We will consider each of the four cells below. Degree of Formalization The bottom left cell in the matrix lists micro structures that influence the decisions and behaviors of individual employees. They should therefore make employee behaviors more predictable. Clearly, an employee can choose to deviate from his or her job description either by not contributing in the prescribed ways or by engaging in activities that are inconsistent with policies. Financial and other incentives can be ignored as well. In each case, however, negative consequences will occur. Ignoring the directions of the incentive system will bring fewer ﬁnancial rewards. Deviating from the job description or from company policies can reduce one’s chances for promotion. The outcomes can be even more extreme – deviation could cost the employee his or her employment. The bottom right quadrant outlines informal micro mechanisms that aﬀect individuals. Informal forces such as company culture and values, social networks, personal reputations, and individuals’ assumptions or expectations can have incredibly powerful eﬀects on behavior. Violation of an organization’s unwritten rules can cause an employee to be isolated or even purged or sabotaged by other employees. Regardless of people’s individual qualities, these mechanisms constrain what they will or will not do to the extent that they want acceptance and approval within the social network of the organization.

313

National norms, values Socially accepted norms Industry reputation Societal stereotypes

Informal

• Company culture, values • Social networks • Personal reputation • Idiosyncratic preconceptions

• • • •

Degree of Formalization

Formal

• Job/role descriptions • Incentive structures • Company policies

• Industry regulations • Laws and rules • Codes of ethics

Figure 12.1 Stickel matrix of forms of social control

(company/ personal)

Micro

Scale

(nation/ industry)

Macro

314

Social capital: reaching in

Scale Sources outside the organization (macro forces) also exert signiﬁcant influence on employees and make their behavior more predictable. Formal ones, which are shown in the upper left quadrant, include industry regulations, laws and rules, and codes of ethics. For instance, if an employee violated a professional code of conduct (for example, American Medical Association, American Psychological Association) he or she could be professionally sanctioned and lose his or her license – a constraint that reaches far beyond the current employer. Violation of industry regulations or legal boundaries can lead to ﬁnes or imprisonment. These standards and punishments can change with time, for example the Sarbanes–Oxley Act which substantially increased the penalties for an executive who willfully misrepresents information in ﬁnancial statements. More nebulous, but still powerful influences are the ones that are both macro and informal. These are in the upper right quadrant and include national norms and values, socially accepted norms, industry reputations and societal stereotypes. These mechanisms may not even be conscious. For example, employees tend to avoid behavior if they believe it will give them a lasting social stigma. Aside from serving time in prison, executives convicted of wrongdoing in the Enron debacle will face skepticism from the business community long after they are released. What is more, their families may be subjected to abuse as well.

TRUST AND SOCIAL CAPITAL: WHAT HAS CHANGED? Perhaps the driving reason that so many scholars are interested in social capital is that higher levels of social capital beneﬁt members in a variety of ways. When something unexpected and negative happens, social capital facilitates a solution. At a small-group level, when a parent has car problems and cannot take a child to a practice session for a sports team, other parents ﬁll in to make sure the child does not miss practice. On a larger scale, a subdivision might use a resident network to stop a landﬁll from being built near their neighborhood. As mentioned earlier, researchers have seen a decline in social capital. Putnam (2000) has raised the question of why people do not invest in social capital as much as they used to. We ﬁnd it useful to reframe this question as: why would people invest in social capital? Having already established the link between trust and context, we can now examine how recent social changes correlate to decreases in social capital. We will also investigate how

Understanding social capital

315

the distribution of the positive outcomes of social capital may be aﬀecting trust and people’s willingness to invest in social capital. First, we must recognize that broad economic forces are having a negative impact on people’s decisions to invest in social capital. While we enjoy a higher standard of living than a few decades ago, it comes at a price. Most adults in today’s society have very little truly discretionary time. On average, people in the US work more hours than they used to. In addition, they spend more hours commuting between home and work. Together, these trends reduce the amount of discretionary time people possess. The growth of dual-career couples exacerbates this shortage. In these situations nobody is home to deal with tasks like cooking, cleaning, grocery shopping and paying bills. Thus, in addition to longer days at work and in transit, more home time must be spent in mundane but necessary tasks. Given that people have less free time, our earlier point that investment in social capital is essentially a trust decision takes on new meaning. Risk becomes more important than it was previously because people’s time is such a precious commodity that they do not want to waste it. When deciding whether or not to invest in social capital, anything that makes people question how much they can trust others in the collective may have a disproportionately negative aﬀect on their decision to trust and invest. There are many potential sources of declines in trust. In the macro context, recent government and industry scandals have drastically lowered trust and reduced the incentive to participate in macro-level building of social capital (for example, the non-competitive Iraqi rebuilding contracts granted to Halliburton and the scandals at Enron, WorldCom, and the demise of the accounting ﬁrm Arthur Anderson). The question of who will reap the beneﬁts also becomes very relevant. If the party perceives that influential others will reap the lion’s share of the beneﬁts – despite the party’s investment – they are less likely to invest resources. For example, employees produce proﬁtable performance and considerable goodwill. Research has recently focused on those nonprescribed behaviors that beneﬁt customers (Schneider et al., 2005). When customers perceive that employees are going beyond what is required of them and putting in extra eﬀort, it produces goodwill. These customers are more likely to return with more business, generating more proﬁts for the company. When employees socialize together outside of work, they build social capital. The strengthened bonds and mutual understanding increase their ability to work together eﬀectively. If this improves performance via enhanced cooperation or innovation, the company beneﬁts through higher performance and proﬁts. However, the division of outcomes – proﬁts and performance – does not reflect the realities of who builds the social capital. By sheer numbers,

316

Social capital: reaching in

lower-level employees generate most of the outcomes. In spite of this, the well-documented disparity between executive compensation and the income of ﬁrst-line workers suggests that executives will receive the vast part of the social capital’s beneﬁts. When employees see this happening, they are likely to revisit their decision about continuing to invest their eﬀorts and energies, which March and Simon (1958) refer to as the decision to produce. If employees see that they have made themselves vulnerable and that the outcomes of their risk-taking are unfavorable, they will trust the company and its leadership less (Mayer et al., 1995). Once this happens, employees often reduce their risk-taking behavior and their investments in building social capital (for example, exceptional service to customers, outside social activities with other employees). This argument has been developed for a single company and its employees, who make ongoing decisions about how much eﬀort to invest in social capital. This same rationale can be applied to a country and its government. For governments, those in power as well as those connected to those in power often reap undue beneﬁts from tax resources and opportunities. Those outside this privileged inner circle were not aﬀorded the same access. Recent history provides an example of this. After the invasion of Iraq, the rebuilding eﬀort began. Halliburton, where the US Vice-President had worked and still owned stock, received billions of dollars of contracts with the US government without any bidding process (New Yorker, 16 February 2004). A story of a group of residents in a suburb which stopped a landﬁll provides a telling example on a smaller scale. While everyone in the subdivision beneﬁts, those in leadership roles now have greater access to government oﬃcials, developers and other constituencies with whom the neighborhood interacts. They thus have greater knowledge of the issues, clearer insights into plans of local oﬃcials, and other such beneﬁts related to knowledge and access to those in power. The leaders are also now power brokers with the other outside constituencies. In some cases leaders may share this information and access as completely as possible; however, they may instead use these resources for their own personal advantage – even if that is not in the best interest of their neighborhood. In the latter case, as other members experience more known abuses of power, influence and access to resources, they will increasingly interpret ambiguous actions as being self-serving. Many people may already hold this cynical point of view in general, based on widely publicized corporate and government scandals. Increased mobility and access to groups and ideas outside one’s geographical area also aﬀects one’s willingness to invest in social capital. While globalization and mobility provide many beneﬁts, they also make context hard to deﬁne and often unstable. Mobility, which increases diversity, can

Understanding social capital

317

destabilize a community’s context. If turnover is small in a ‘community’, the community remains stable and those who move in are assimilated into it. Alternatively, when turnover is widespread and rapid, the community becomes a more heterogeneous mix. As a result, individuals are less certain of how they ﬁt and what the potential beneﬁts are of investing in social capital. People who understood how to interact with others in the stable environment face a diﬀerent situation, where a once acceptable and even appreciated gesture (for example, giving someone a hug) may become oﬀensive. In such a context, uncertainty is high. The risks of reaching out and networking become even higher. Consequently, people are less willing to invest in either networking or social capital.

CONCLUSION Since the 1990s there have been many publications about trust. The literature in the social capital arena has not previously paid suﬃcient attention to developments in this topic. This chapter has focused on deﬁning trust and then more clearly articulating its relationship with social capital. We have also developed a new deﬁnition of social capital that diﬀerentiates it from both its antecedents and its outcomes to clarify social capital’s relationship with trust and networks as well as desired outcomes that involve risk-taking behaviors. We have discussed the importance of risk, arguing that trust enables a person to take risks with another party or parties. Sometimes the ‘other’ is a known individual, but can be a group or individual who is less known. Regardless of who the trustee is, an individual who invests time and energy in social capital is taking a risk. For someone to take those risks and build social capital, the person must believe in future pay-oﬀs. What is more, the person must perceive that the group allocates the ‘spoils’ of the increased social capital fairly. We argue in this chapter that factors that either decrease trust in others in the collective or increase the perception of risk of getting future pay-oﬀs from investments in social capital will reduce the investment that people make in social capital. Conversely, factors that control the level of risk to the individual will foster the investment of social capital. Among other factors, we suggest that abuses of power and influence in both government and business reduce the level of trust that people have in leaders in general. Such factors as mobility increase uncertainty in the context, and as the level of risk inherent in the context increases, the investments in social capital will decline.

318

Social capital: reaching in

REFERENCES Coleman, J.S. (1990), Foundations of Social Theory, Cambridge, MA: Harvard University Press. Davis, J.H. and V. Barktus (2006), ‘Networks, norms, and trustworthiness in the creation of social capital: an integrated framework’, paper presented at the Notre Dame Conference on Social Capital. Deutsch, M. (1958), ‘Trust and suspicion’, Journal of Conflict Resolution, 2, 265–79. George, J.M. and A.P. Brief (1996), ‘Motivational agendas in the workplace: the eﬀects of feelings on focus of attention and work motivation’, in B.M. Staw and L.L. Cumings (eds), Research in Organizational Behavior, Vol. 18, Greenwich, CT: JAI Press, pp. 75–109. March, J.G. and Z. Shapira (1987), ‘Managerial perspectives on risk and risk taking’, Management Science, 33, 1404–18. March, J.G. and H.A. Simon (1958), Organizations, New York: Wiley. Mayer, R.C. (2007), ‘Employee loss of trust in management: surviving in a new era’, in Janice Langan-Fox, Cary Cooper and Richard Klimoski (eds), Research Companion to the Dysfunctional Workplace: Management Challenges and Symptoms, New Horizons Management Series, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 125–35. Mayer, Roger C., James H. Davis and F. David Schoorman (1995), ‘An integrative model of organizational trust’, Academy of Management Review, 20, 709–34. Mossholder, K.W. and A.G. Bedeian (1983), ‘Cross-level inference and organizational research: perspectives on interpretation and application’, Academy of Management Review, 8, 547–58. Putnam, R.D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Sabatini, Fabio (2006), ‘The role of social capital in economic development’, Conference on Social Capital, Sustainability, and Sociology, www.lse.ac.uk, p. 4. Schneider, B., M.G. Ehrhart, D.M. Mayer, J.L. Saltz and K. Niles-Jolly (2005), ‘Understanding organization–customer links in service settings’, Academy of Management Journal, 48, 1017–32. Schoorman, F.D., R.C. Mayer and J.H. Davis (2007), ‘An integrative model of organizational trust: past, present, and future’, Academy of Management Review, 32, 344–54. Sitkin, S.B. (1995), ‘On the positive eﬀect of legalization on trust’, Research on Negotiation in Organizations, 5, 185–217.

13.

Organizational trust and social capital James H. Davis and Viva Ona Bartkus

● ● ●

Does trust create social capital? Does social capital foster trusting relationships? Are trust and social capital synonymous?

Controversy persists regarding the relationship between trust and social capital. Opinions vary widely. At one end of the spectrum, scholars argue passionately that trust is a necessary ingredient for – or antecedent of – social capital,1 while others claim trust is an outcome.2 Some scholars declare that trust and social capital are essentially the same thing.3 This chapter represents our entry in this vigorous debate. We first argue that social capital is impossible without prior organizational trust. Such trust creates the environment community members need if they are going to engage in activities together. This atmosphere also mitigates the risks that individuals take when they choose to act as part of a community, and makes them more open to being vulnerable. The resulting collaboration is how members obtain social capital’s benefits. Yet how does a community foster the organizational trust so needed to build social capital? We argue that a community’s level of organizational trust depends directly on the strength of its networks, the breadth and depth of its shared norms (for example, reciprocity, helpfulness), and the abilities of the members and the group itself. Our model is firmly grounded on existing trust literature and empirically tested with a sample of students and study teams in the business college of a major university. More specifically, our research investigates bonding social capital – in other words, relationships within a closed and tightly knit community – a reasonable assumption given the nature of most graduate business programs. By providing both a theoretical model and empirical evidence for organizational trust’s role in creating social capital, we hope to provide direction as to how social capital may be better understood and measured in the future. The concluding discussion outlines possible implications of this model for subsequent investigations into both bonding and bridging social capital. 319

320

Social capital: reaching in

DEFINITIONS OF SOCIAL CAPITAL AND ORGANIZATIONAL TRUST The lack of a commonly agreed-upon definition of social capital has handicapped its study over the years. An examination of social capital’s roots reveals early clues that pinpoint the challenges presented by different definitions. For instance, Coleman focused on children’s education in his definition of social capital: ‘the norms, the social networks, and the relationships between adults and children that are of value for the child’s growing up; social capital exists within the family, but also outside the family, in the community’ (Coleman, 1990, p. 334). Putnam, on the other hand, framed social capital as: ‘[the] features of social life – networks, norms and trust – that enable participants to act together more effectively to pursue shared objectives’ (Putnam, 1995, p. 56). These two definitions already reveal one of the major problems. They, like many others, are functional in nature, and are not easily measured. Despite these issues with definitions, the core concept of social capital is generally accepted: interpersonal relationships matter and provide value to individuals and groups. At its most basic, any capital can be thought of as assets capable of generating benefits (or value) for some individuals in the future (Lachman, 1978). To keep it simple and create a measurable model of organizational trust and social capital, we adopt Nahapiet and Ghoshal’s (1998) and Adler and Kwon’s (2002) proposed definition: social capital represents the resources or goodwill inherent in relationships. This chapter must also identify a useful definition of organizational trust. While scholars have dedicated considerable effort to investigating trust in the 1990s, a commonly accepted definition still eludes them. Most research has focused upon micro, dyadic relationships,4 although emerging research is investigating trust within and across organizations.5 We define organizational trust as: the cumulative willingness of members of a group to be vulnerable to the actions of that group, even if they do not know all the other members of the group and even if the actions of other members cannot be monitored or controlled. This definition mirrors the most commonly accepted one for dyadic trust, which is ‘the willingness of an individual to be vulnerable toward another individual, even if the other’s action can be neither monitored nor controlled’ (Mayer et al., 1995). The important difference between these two definitions is that in the one for organizational trust community members are willing to be vulnerable to the actions of the community. Why is another definition of trust necessary? Why is dyadic trust not sufficient? Unfortunately, scaling dyadic trust to a group level does not truly represent the idea of organizational trust. In dyadic trust, the trustor is

Organizational trust and social capital

321

willing to be vulnerable toward one trustee to do some specific action. Further, the trustor knows the trustee personally. Within a larger group or community, an individual cannot know every other group participant. As a result, organizational trust requires that the trustor be vulnerable toward the trustee for some specific action, even when the trustor does not know the trustee, primarily because the trustee is a member of the group.

ORGANIZATIONAL TRUST AS ANTECEDENT OF SOCIAL CAPITAL If organizational trust is the willingness of members to assume the risk that is part of the group’s activities, then social capital reflects the actual risks taken by the group. Once again, organizational trust and dyadic trust parallel each other. ‘Risk-taking in relationships’ reflects the level of trust between two individuals, according to Mayer et al. (1995). The overall level of organizational trust, and thus social capital, is therefore visible in the combined group members’ willingness to take risks within their community. Where organizational trust is high, the group’s social capital and risktaking behavior should be greater. Where organizational trust is low, the group’s social capital and risk-taking behavior should be lower. Ultimately, social capital can be indirectly measured as the level of organizational trust within the network. This suggests the following hypotheses for future empirical testing: networks with higher organizational trust have higher social capital reflected in higher risk-taking in the community; networks with lower organizational trust have lower social capital reflected in lower risk-taking in the community. To illustrate further, take the business students who participated in our research at a major university. Students constantly make decisions as to how to invest their most valuable commodity – time – based on the returns they may get: for example, improved learning of course material or better access to job opportunities. Whether the choices are within their academic program or their specific study teams, they still face them. For instance, they can study on their own, obtain a tutor or study with teammates. But in this model, they need an environment of organizational trust within their teams if they want to gain the benefits of social capital; this situation enables them to collaborate effectively with teammates on course projects and perhaps job searches. In essence, the more organizational trust within a study group, the higher the social capital.

322

Social capital: reaching in

NETWORKS, NORMS AND ABILITY AS THE ANTECEDENTS OF ORGANIZATIONAL TRUST But what creates organizational trust? What are outcomes of and inputs to organizational trust and social capital? How do they fit together? The last section argued that organizational trust is an antecedent of social capital. Other scholars have emphasized the importance of networks, norms and ability as antecedents of social capital.6 This section posits an extra theoretical step. It posits that these items – the strength of social networks, collectivist norms such as reciprocity, and the abilities of the group and its members – are the antecedents of organizational trust, which in turn creates the environment necessary to foster social capital. This work also contrasts with other research that contends that when networks, norms and ability are in place, these conditions can directly lead to individual organizational citizenship behavior and collective action as ultimate outcomes. However, our research contends that individual citizenship behavior and collective action are outcomes of social capital, not part of social capital. Likewise, individual and shared perceptions of networks’ characteristics, group norms and ability are antecedents to organizational trust, not social capital itself. Figure 13.1 summarizes our proposed theory of organizational trust and social capital. What relevance might these observations have for the firm? When such interactions repeat over time, they reinforce the strength of existing relationships and networks and foster new ones. The aggregate of these interactions also reinforces commonly held norms, and perhaps even builds the group’s abilities. This sequence sets up a self-reinforcing feedback loop that strengthens both organizational trust and social capital. The main benefits from organizational trust and social capital lie in advantageous interpersonal interactions that accumulate knowledge and expertise in the organization over time (Nahapiet and Ghoshal, 1998). Social capital is the basis for the exchange and combination of information that then creates intellectual capital, which can in turn produce competitive advantage. Why Are Networks an Antecedent for Organizational Trust? There are numerous characteristics of networks – many well beyond the scope of this investigation – which focus on the characteristics that explain the formation of organizational trust. This work examines the more social aspects of networks rather than the overemphasized concept of them as pipes. Social identity perspectives support this approach. This perspective argues that the group must first be defined cognitively, in that the individual must see themself as a group member (Tajfel, 1959, 1969, 1984; Hogg,

323

Figure 13.1

Feedback loop

Organizational trust

Model of organizational trust and social capital

Ability

Norms

Network strength

Social capital Risk-taking in community

Outcomes

324

Social capital: reaching in

2001). A group or network only exists psychologically when three or more individuals define themselves as sharing characteristics, which may also differentiate them from others. Although this was initially an approach for analyzing intergroup relations – particularly prejudice, discrimination and social conflict (Tajfel, 1969) – it now provides a basis for group membership and commitment within the group. Social identification literature also posits that identification is closely related to a sense of shared fate among group participants, as well as a psychological acceptance of membership with all of its associated obligations. An individual is vulnerable to, depends upon and experiences solidarity with any group with which they identify strongly. This identification, which entails vulnerability, thus forms one of organizational trust’s foundations. Given this reality, our research uses the social identity an individual feels for a group as a proxy for the influence of networks on organizational trust. Furthermore, the number of business students within each study team and within the entire program is already determined. Size of the network is frequently a variable determining the strength of the network. Because network size is fixed within this specific study, the level of identification with the network can also be a proxy for network strength. These points suggest the following hypothesis: Hypothesis 1: The strength of networks as measured by identification of individuals with the group will explain a significant portion of organizational trust variance. Why Are Norms an Antecedent for Organizational Trust? Coleman (1988) originally argued that norms express behaviors accepted by others within a network or organization and are used to regulate actions. They arise when actions affect other actors who do not make decisions. Communicated throughout an organization both formally and informally, norms become a permanent part of the organizational culture. Positive and collective norms based on reciprocity and helpfulness mitigate the risks that individual group members take as they engage in collaborative activities. The norms limit the downside of such activities, because they hinder exploitation, facilitate exchange, and control individual and group behavior (Lin, 2001). They enable repeated congenial interactions, which in turn foster organizational trust. But the nature of these norms is critically important when considering organizational trust. When group members follow norms such as benevolence between members, openness and teamwork, they act in a more trusting manner toward the organization and are more willing to take risks on

Organizational trust and social capital

325

behalf of fellow members. But close-knit communities can also foster norms that lead to discrimination, prejudice and even possible ostracism. When norms like dishonesty, lack of communication, opportunistic behavior and individualism persist, one would expect lower organizational trust and fewer collective actions. These ideas suggest the following hypothesis: Hypothesis 2: Collective norms such as reciprocity and helpfulness among members will explain a significant portion of organizational trust variance. Why Are Abilities an Antecedent for Organizational Trust? Individual and collective abilities are the third critical antecedent – along with networks and norms – of organizational trust. Ability means taskrelated competence. In a similar argument, Mayer et al. (1995) contend that the trustor’s perception of the trustee’s ability to do what they say they can do is an antecedent to dyadic trust. At the organizational level, members are unlikely to trust the group if they perceive that the group is unable to carry out its allotted tasks. Even if the network and norms are strong, a group with limited ability to perform and reciprocate will produce low organizational trust. In this situation, any collective activity on the part of the individual is an act of altruism or charity. The expectation of future benefits from relationships also falls off. However, abilities are not enough by themselves. Groups with significant task-related ability but limited social relations – that is, inadequate networks and norms – will often use contracts between parties to render services. Whether charity or contract, however, the actions would probably not foster organizational trust. This suggests the following hypothesis: Hypothesis 3: Collective abilities among members will explain a significant portion of organizational trust variance.

THE ORGANIZATIONAL TRUST MODEL Members of an organization must be willing to take risks in dealing with each other – to be vulnerable to the actions of others within the group – for organizational trust to exist. How does this happen? In general, people tend to build connections with others whose values they share. They begin to take more risks as they know the community better and their network relationships strengthen. They perceive that other members are benevolent, have integrity, and have the ability to follow through and help the group

326

Social capital: reaching in

(Mayer et al., 1995). They may also see that others’ actions are directed to the collective good of the organization and to each other. This research argues that organizational trust rests on the confluence of three antecedents: the strength of networks among group members, shared norms such as helpfulness and reciprocity, and the abilities that each member brings to the group. Collective perceptions of the organization’s norms, networks and abilities then lead to organizational trust (see Figure 13.1). This suggests the following hypothesis: Hypothesis 4: The combination of network strength, collective norms and abilities will explain a significant portion of the variance in organizational trust. Methodology Our research tested the proposed model of organizational trust and social capital with a sample of graduate students in the College of Business in a major Midwestern university. Student perceptions of organizational trust and its antecedents were collected with respect to two different networks: their study group and their graduate program. We distributed surveys to a population of 315 graduate students. Of the 173 surveys submitted, 167 were usable for testing purposes, a 53 percent response rate from the original cohort. Some background information on the respondents may be helpful. The students who completed the survey had an average salary of $51 320 prior to their graduate studies and anticipated an average salary of $86 050 at the conclusion of their studies. The average student started the program with 6.56 years’ work experience and a GMAT (Graduate Management Admissions Test) score of 663 out of a possible 800 points. The next section describes the control and independent variables and their measurement. Control Variables Although any number of variables can explain organizational trust, our research focuses on those variables that both explain the most variance and that business managers can influence. Our study attempted to control for variables that are outside managers’ control, including familial influences that might frame social values, psychology and motivations. Research has shown that early family relationships influence an individual’s psychology later in life.7 While childhood experiences are important and their variance must be accounted for, they are outside the firm’s control and are

Organizational trust and social capital

327

therefore treated as control variables. Three variables accounted for familial influences on perceptions of organizational trust, social capital and its antecedents: 1. The number of individuals living in the respondent’s childhood home. 2. Whether their family is active in community affairs or not (0 yes, 1 no). 3. Whether their family is involved in community charitable causes (0 yes, 1 no). Independent Variables This section explains how each antecedent was measured. Network strength Organizational trust exists in the context of networks, which are a crucial antecedent for it. However, networks are incredibly complex and any attempt to measure all of their dimensions is well beyond the scope of this research.8 Scholars have shown that social identification with a group fosters feelings of vulnerability, dependence and solidarity within the individual. This social identity approach simplifies the complexities of dealing with networks and their potential influence on organizational trust (Hogg, 2001). Such research defines the group cognitively – that is, individuals define themselves as members of the group – rather than through structure and types of relationships. When groups have a fixed size, the level of social identification with the group could be a proxy for network strength. The hypothesis is that a positive relationship will exist between the level of member identification with the network and the level of organizational trust. Our study examined two different networks. First, it explored respondents’ study teams. Students belonged to study groups and project teams during their entire program of graduate study; these were the focal network for this research. Different teams have very different levels of organizational trust. Second, the research looked at the entire graduate program. This network included all graduate students, faculty and staff in the respondents’ program of study. A ten-item scale measured social identification, and thus network strength. It included four items from two scales developed in previous research (Green, 2002; Weisberg and Hasecke, 1999) with six developed for this study (see Table 13.4). Respondents indicated on a five-item Likert scale how much they agreed with each statement (1 strongly disagree to 5 strongly agree) such as: ‘when someone praises the group it feels like a

328

Social capital: reaching in

personal compliment’ and ‘I like their values’. The Chronbach alphas for the scale on the two networks were in the acceptable range for this research, 0.785 for teams and 0.763 for the program (see Table 13.1). Collective norms Collective norms help regulate behaviors within a network by limiting exploitation and facilitating cooperation. A 12-item Likert scale measured the perceptions of team and program norms including helpfulness, benevolence and willingness to consult with each other. Items measuring benevolence came from a previously published scale developed by Davis et al. (2000). Respondents indicated the extent to which they felt their team and program were concerned about their welfare, watched out for them and protected what was important to them. The reliability measures were found to be in an acceptable range: team, 0.847 and program, 0.847 (see Table 13.1). Ability One can consider the collective’s ability as ultimately contributing the ‘capital’ part of social capital. In a graduate program of study it includes group aptitude, skills and competencies that members of the group can bring to bear. A six-item scale developed by Davis et al. (2000) measured individual perceptions of abilities within the groups. The items used a five-point Likert scale, asking respondents: ‘how capable the group was at performing a job’, ‘how confident they were about the group’s skills’, ‘the group’s knowledge about the work that needed to be done’ and ‘the group’s specialized capabilities that might increase its performance’. The scale had strong levels of reliability with Chronbach alpha scores of 0.905 (team) and 0.882 (program) (see Table 13.1). To view all of the survey items used to measure perceived network strength, collective norms and ability, see Box 13.1. Dependent Variable: Organizational Trust Organizational trust can be treated as the generalized vulnerability of network membership. At the group level, this vulnerability defines the willingness of members, as a whole, to engage in collective behavior for the good of the group and its members. A six-item scale developed by Davis et al. (2000) assessed trust (see Box 13.1), or, in other words, group members’ willingness to be vulnerable to the group. Organizations with low levels of willingness are unlikely to engage in behaviors for the good of the group, and therefore have low organizational trust. By contrast, groups with high levels of collective vulnerability have high organizational trust and will probably act for the collective good. Items

329

Team network strength Team norm Team ability Team trust Program network strength Program norm Program ability Program trust

Table 13.1

0.722 0.526

0.390

0.414

0.360

0.847

0.882

0.743

0.742 0.718

0.847 0.905

0.786 0.763

1

Team network strength

0.785

Chronbach alpha

0.339

0.385

0.540

0.749 0.365

1 0.695

Team norms

Chronbach alphas and correlations

0.422

0.545

0.417

0.815 0.353

1

Team ability

0.490

0.395

0.355

1 0.300

Team trust

0.690

0.668

0.632

1

Program network strength

0.613

0.606

1

Program norms

0.765

1

Program ability

1

Program trust

330

Social capital: reaching in

BOX 13.1

ORGANIZATIONAL TRUST AND SOCIAL CAPITAL: SCALES ITEMS

Network strength: identity 1. When someone praises this group it feels like a personal compliment.** 2. When I talk about this group, I usually say ‘we’ rather than ‘they’.** 3. When the media criticize this group, it feels like a personal insult.* 4. This group’s successes are my successes.* 5. They have a strong sense of justice. 6. I never have to wonder whether they will stick to their word. 7. They try hard to be fair in dealings with others. 8. Their actions and behaviors are not very consistent (reverse scaled). 9. Sound principles seem to guide their behavior. 10. I like their values. ** items used in Weisberg and Hasecke (1999); * items used in Greene (1999) Network ability 1. I feel very confident about their skills. 2. They have much knowledge about the work that needs to be done. 3. They have specialized capabilities that can increase our performance. 4. They are well qualified. 5. They are very capable of performing their job. 6. They are known to be successful at things that they try to do. Network norms 1. Members of this network ask each other for help. 2. I am pleased to be associated with this group of students due to its values. 3. Even if I do not know everyone in my group personally, I would go out of my way to help them.

Organizational trust and social capital

331

4. Students in the network would do what is in the best interests of the network. 5. Because I help other students, I expect them to help me. 6. If someone helps me, I would help them. 7. Students work toward the interests of the network rather than themselves. 8. They really look out for what is important to me.* 9. They are very concerned about my welfare.* 10. They would not knowingly do anything to hurt me.* 11. My needs and desires are very important to them.* 12. They go out of their way to help me.* * items used in Davis, et al. (2000) Organizational trust 1. I would be comfortable giving this group a task or problem which was critical to me, even if I could not monitor their actions. 2. Even though I do not know everyone personally, I am willing to let them have influence over my education. 3. I would be willing to let this group have complete control over my program of study. 4. This group is generally thought of as being trustworthy. 5. If I had my way, I would not let them have any influence over issues that are important to me (reverse scaled). 6. I really wish I had a good way to keep a better eye on this group (reverse scaled).

in the scale asked participants: ‘whether they would be willing to allow their collective to have complete control over issues and outcomes that are important to them personally’, ‘their comfort level with giving their network complete control over a task or problem critical to them’, and ‘whether they thought their network was trustworthy’. The reliability for the organizational trust measure was acceptable for both networks with Chronbach alphas of 0.786 for team and 0.743 for program (see Table 13.1). Readers may pose a few questions about the study methodology based on response bias and levels of analysis problems. A response bias does exist. On the levels question, one could point out that although the study looks at group-level phenomena, it surveys individuals. Both organizational trust

332

Table 13.2

Social capital: reaching in

Regression – dependent variable: organizational trust in team Model 1 Team/network strength

(Constant) Family size Family involved in community affairs Family involved in charity Team/network strength Team norms Team ability Adjusted R square

Model 2 Team norms

Model 3 Team ability

Model 4 Complete model

1.624 0.143 0.496

9.820 0.114 0.379

0.785 0.139 0.554

0.222 0.052 0.808

0.007

0.158

0.076

0.028

0.836 0.667

0.087 0.011 0.493 0.741

0.522 0.025 . 0.516

0.556

and social capital research constantly face this issue. In this case, however, the survey asks individuals about their perceptions of the collective as a whole, not about other individuals. Thus, when the individual perceptions are combined they become the collective perception of network attributes. It is unlikely that outside observations by an external panel would be more accurate. The Conclusion returns to this discussion, but the chapter now turns to the study’s results.

RESULTS Strong, significant and positive correlations exist between all variables – network strength, collectivist norms and ability – and organizational trust, particularly at the study team level. Table 13.1 displays these results. Because the correlations between the antecedent variables exceeded 0.70, multi-colinearity needed to be investigated. A regression analysis would show whether or not each variable explained independent, significant variance in organizational trust. Hypothesis 1 – the strength of networks as measured by identification of individuals to the network will explain a significant portion of organizational trust variance – was confirmed for both the team members and the program network. Model 1 (for team members) explained 51.6 percent of the variance in organizational trust (see Table 13.2). This occurred even

333

Organizational trust and social capital

Table 13.3 Regression – dependent variable: organizational trust in program Team/network strength (Constant) Family size Family involved in community affairs Family involved in charity Program/network strength Program norms Program ability Adjusted R square

Team norms

Team ability

Complete model

3.161 0.353 0.077

10.830 0.140 0.033

0.395 0.057 0.608

4.240 0.179 0.345

0.502

0.700

0.503

0.516

0.222 0.024

0.495

0.641

0.371

0.805 0.583

0.498 0.006 0.646

when family size and family activity with community groups and charities were controlled. The number of members in the respondent’s family also approached significance. Network strength also explained a significant portion of organizational trust within the larger, program network. Program network strength explained 49.5 percent of the variance in program trust (see Table 13.3). The only control variable that was significant was family size. Hypothesis 2 – collective norms such as reciprocity and helpfulness among members will explain a significant portion of organizational trust variance – was also supported for both teams and the larger program. The second model in Table 13.2 shows that norms explain 55.6 percent of trust variance with team networks (see model 2 in Table 13.2); model 2 in Table 13.3 shows that norms explained 37.1 percent of the variance in program trust. The only control variable that approaches significance is family activity in charities with program trust. Hypothesis 3 – collective abilities among members will explain a significant portion of organizational trust variance – was supported for both team and program levels. It actually had the strongest outcomes of the three individual variables, explaining 66.7 percent of the variance with teams (see model 3 in Table 13.2) and 58.3 percent with program networks (see model 3 in Table 13.3). While the control variables in the model were not found to be statistically significant, family involved in community affairs approached significance with program trust (p < 0.10). The final hypothesis tests the complete model – that the combination of network strength, collective norms and abilities will explain a significant

334

Social capital: reaching in

portion of the variance in organizational trust. This hypothesis explained more of the variance for the teams and the program than any of the three individual variables. It accounted for 74.1 percent of the variance with teams and 64.6 percent of the variance with the program. Model 4 in Table 13.2 shows the results of testing hypothesis 4 with teams. None of the control variables are significant. All three hypothesized antecedents to team trust were. Program trust’s results are in model 4, Table 13.3.

CONCLUSION AND FUTURE RESEARCH DIRECTIONS It is important to address the implications and limitations of the current study and identify future research directions. Our empirical tests of the proposed model of organizational trust and social capital are but a work in progress. While the survey design successfully tested the front end of the model, the remainder of the model – how social capital creates both individual and collective benefits – is critically important and still needs to be examined. Further research should test the entire model from its beginning through to social capital’s outcomes, for example collective actions and individual citizenship behaviors. Assessments should also occur in other non-university contexts, or at the very least at another university, and possibly include additional design parameters that address response bias and levels issues. If social capital is important to businesses, as other scholars have argued, in fostering innovation, collaboration and competitive advantage, then organizational trust is equally crucial. Such trust underpins the willingness and capacity of group members to cooperate on and coordinate their activities. These behaviors then generate a pay-off in terms of the productivity of the group. The research findings that network strength, widely held norms and group abilities directly affect the level of organizational trust highlight the need for organizations to build and sustain these qualities. Furthermore, the model should be extended to a ‘bridging social capital’ situation. It was built for a bonding social capital one, where strong closed networks of relationships exist and many community members know each other. However, it could be modified and developed to encompass a ‘bridging’ milieu. Some of the new important factors might include other network attributes like the extent of structural holes that relationships cover (Burt, 2005). The nature of norms may also differ, shifting from helpfulness to information transparency and sharing. However, some aspects of

Organizational trust and social capital

335

the original model of organizational trust and social capital would probably remain important: for example, the strength of the network, the norm of reciprocity. The longer-term, theoretical implications of this model and research may be even more intriguing as they begin to address why we should care about organizational trust, and ultimately, social capital in business. The most entrenched theory of the firm presumes antagonistic relationships between principals and agents, with principals focusing solely on shareholder value creation. Yet common sense dictates that the strength of many companies relies on the results of their relationship-building efforts in the broadest sense. The success and even survival of many firms depends on their relationships not only internally with their employees, but also externally with suppliers, customers, creditors, regulators, government and society in general. The study of such company relationships and social capital is timely and needed for a number of reasons. First, the complexity, velocity and fragmentation of today’s business world present heretofore unseen uncertainties and risks. These new uncertainties derive from massive changes due to increased globalization, demographic changes and technological changes that are driving some transaction costs nearly to zero. Organizational trust and social capital may provide companies with needed capabilities to address these major new uncertainties. Second, Putnam’s monumental study, Bowling Alone (2000), described the general decline of social capital in the United States, and offers further reason for examining the importance of relationships in the business world. So what is a manager to do? When might organizational trust and social capital become more or less important to a firm? They are always critical for a successful organization, but never more so than when uncertainty is high. Companies can benefit from investigating and then fostering greater internal networks among employees and reinforcing norms based on reciprocity. Attracting, retaining and training employees with the appropriate norms would further serve to build task-related abilities. Building effective relationships has become ever more critical to performance. Contracts are insufficient, as unexpected circumstances inevitably arise. An environment of organizational trust, on the other hand, diminishes transaction costs and minimizes defensive actions. The overall economic performance of the firm improves if employees expect others to act reasonably no matter what the contingency. As the evolution of the current business environment accelerates, and presents new risks and uncertainties to the managers, groups may need even more organizational trust and social capital if they are going to overcome their collective challenges.

336

Social capital: reaching in

NOTES 1. Tsai and Ghoshal (1998), Chung et al. (2000), Nahapiet and Ghoshal (1998), Leana and Buren (1999), Pretty and Ward (2001). 2. Fox and Gershman (2000), Putnam (1995a), Woolcock (2001). 3. Knack and Keefer (1997), Woolcock (1998), Glaeser et al. (2002), Cohen and Prusak (2001), Bowles and Gintis (2002). 4. Whitener et al. (1998), Elangovan and Shapiro (1998), Mayer et al. (1995). 5. Zaheer et al. (1998), McEvily et al. (2003), Perrone et al. (2003). 6. Chung et al. (2000), Tsai (2000), Leana and Buren (1999), Nahapiet and Ghoshal (1998), Burt (2000, 1997), Adler and Kwon (2002). 7. Matheis (2004), Fullinwider-Bush (1993). 8. Network characteristics have been studied for some time now with respect to their type (Burt, 1992); ties, configuration and structure (Adler and Kwon, 2002); hierarchy, density and connectivity (Krackhardt, 1992); and size, density, stability, centrality linkages and transactional content (Tichy et al., 1979).

REFERENCES Adler, P.S. and S.W. Kwon (2002), ‘Social capital: prospects for a new concept’, Academy of Management Review, 27 (1), 17–40. Bowles, Samuel and H. Gintis (2002), ‘Social capital and community governance’, Economic Journal, 112, F419–F436. Burt, Ronald S. (1992), Excerpt from ‘The social structure of competition,’ in Structural Holes: The Social Structure of Competition, Cambridge, MA and London, UK: Harvard University Press, pp. 8–30. Burt, Ronald S. (1997), ‘The contingent value of social capital’, Administrative Science Quarterly, 42, 339–65. Burt, Ronald S. (2000), ‘Network structure’, Research in Organizational Behavior, 22, 345–423. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Capital, Oxford: Oxford University Press. Chung, S.A., H. Singh and K. Lee (2000), ‘Complementarity, status similarity and social capital as drivers of alliance formation’, Strategic Management Journal. Cohen, D. and L. Prusak (2001), In Good Company: How Social Capital Makes Organizations Work, Boston, MA: Harvard Business School Press. Coleman, James (1990), Equality and Achievement in Education, Boulder, CO: Westview. Davis, J.H., D. Schoorman, R.C. Mayer and T.H. Hoon (2000), ‘The trusted general manager and business unit performance: empirical evidence of a competitive advantage’, Strategic Management Journal, 21 (5), 563–76. Elangovan, A.R. and D.L. Shapiro (1998), ‘Betrayal of trust in organizations’, Academy of Management Review, 23 (3), 547–66. Fox, J. and J. Gershman (2000), ‘The World Bank and social capital: lessons from ten rural development projects in the Philippines’, Policy Sciences, 33, 399–419. Fullinwider-Bush, Nell and D.B. Jacobitz (1993), ‘The transition to young adulthood: generational boundary dissolution and female identity development’, Family Process, 32 (1), 87–103.

Organizational trust and social capital

337

Green, Stephan (2002), ‘Social psychological measurements of participation’, Political Behavior, 24 (3), 171–97. Hogg, Michael A. (2001), ‘Social categorization, depersonalization, and group behavior’, in M.A. Hogg and R.S. Tindale (eds), Blackwell Handbook of Social Psychology: Group Processes, Oxford: Blackwell, pp. 56–85. Inkpen, A. and E. Tsang (2005), ‘Social capital, networks, and knowledge transfers Academy of Management Review’, 30 (1), 146–65. Knack, Stephen and Philip Keefer (1997), ‘Does social capital have an economic payoff ? A cross-country investigation’, Quarterly Journal of Economics, 112 (4), 1251–88. Krackhardt (1992), ‘Assessing the political landscape: structure, cognition, and power in organizations’, Administrative Science Quarterly, 35 (2), 342–69. Lachman, Ludwig M. (1978), Capital and Its Structure, Kansas City, MO: Sheid Andrews & McMeel. Leana, C.R. and H.J. van Buren III (1999), ‘Organizational social capital and employment practices’, Academy of Management Review, 24 (3), 538–55. Lin, Nan (2001), Social Capital: A Theory of Social Structure and Action, Cambridge: Cambridge University Press. Matheis, Shawn (2004), ‘Family climate and identity style during late adolescence’, Identity, 4 (1), 77–95. Mayer, R.C., J.H. Davis and D. Schoorman (1995), ‘An integrative model of organizational trust’, Academy of Management Review, 20 (3), 709–34. McEvily, B., V. Perrone and A. Zaheer (2003), ‘Trust as an organizing principle’, Organization Science, 9 (2), 91–103. Nahapiet, Janine and Sumantra Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23 (2), 242–66. Perrone, V., A. Zaheer and B. McEvily (2003), ‘Free to be trusted? Organizational constraints on trust in boundary spanners’, Organization Science, 14 (4), 422–39. Pretty, J. and H. Ward (2001), ‘Social capital and the environment’, World Development, 29 (2), 209–27. Putnam, R.D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Tajfel, H. (1959), ‘Quantitative judgment in social perception’, British Journal of Psychology, 50, 16–29. Tajfel, H. (1969), ‘Cognitive aspects of prejudice’, Journal of Social Issues, 25, 79–97. Tajfel, H. (ed.) (1984), The Social Dimension: European Developments in Social Psychology, Cambridge: Cambridge University Press. Tichy, N.M., M.L. Tushman and C. Fombrun (1979), ‘Social network analysis for organizations’, Academy of Management Review, 14 (4), 507–19. Tsai, W. (2000), ‘Social capital, strategic relatedness, and the formation of intraorganizational linkages’, Strategic Management Journal, 21, 925–39. Tsai, W. and S. Ghoshal (1998), ‘Social capital and value creation: the role of intrafirm networks’, Academy of Management Journal, 41, 464–76. Weisberg, H.F. and E.B. Hasecke (1999), ‘Multidimensional conceptualization of party identification’, Political Behavior, 21 (2), 33–60. Whitener, E.M., S.E. Brodt, M.A. Korsgaard and J.M. Werner (1998), ‘Managers as initiators of trust: an exchange relationship framework for understanding

338

Social capital: reaching in

managerial trustworthy behavior’, Academy of Management Review, 23 (3), 513–30. Woolcock, Michael (1998), ‘Social capital and economic development: toward a theoretical synthesis and policy framework’, Theory and Society, 27 (2), 151–208. Woolcock, Michael (2001), ‘The place of social capital in understanding social and economic outcomes’, Canadian Journal of Policy Research/Revue canadienne de recherche sur les politiques, Government of Canada’s Policy Research Secretariat, 2 (1), 1–8. Zaheer, A., B. McEvily and V. Perrone (1998), ‘Does trust matter? Exploring the effects of interorganizational and interpersonal trust on performance’, Organization Science, 9 (2), 141–59.

Conclusion: frontiers of social capital research Viva Ona Bartkus and James H. Davis Social capital scholarship has made an ongoing and lasting contribution to the social sciences, despite the many questions and criticisms that circle around it. This framework has ensured that social elements are no longer neglected but rather are incorporated into the heart of analysis of individual and collective action. This Conclusion closes our book by providing a succinct overview of social capital’s wide-ranging contributions, by highlighting its contradictions and valid criticisms, and, most importantly, by charting out a path for future social capital research.

CONTRIBUTIONS OF SOCIAL CAPITAL RESEARCH The sheer scope of social capital scholarship is breathtaking. As Woolcock (1998) demonstrated, scholars use social capital to address questions in sociology, anthropology, political science, economics and management theory. They also employ social capital perspectives as a powerful analytical tool in examining numerous themes including: education and youth problems, democracy and governance, economic development, collective action and social class evolution. Worried about the uncritical presumption of social capital’s positive impact, some authors focus on its negative eﬀects – for example, exclusion, reduction of personal freedom or loss of opportunities for those outside the collective. Table C.1 summarizes these broad research themes on the eﬀects of social capital. In addition to these areas, business leaders are showing more interest in social capital as research connects it to beneﬁts at the individual, organizational and interorganizational levels. By building their own social capital, individuals may beneﬁt through more eﬀective job searches, enhance their compensation, and even increase their overall career success. Further, scholars have shown that social capital at the organizational level can have a positive eﬀect in multiple and diverse areas of business including intellectual capital creation, improved scarce resource distribution, increased 339

340

Table C.1

Social capital

Benefits (costs) of social capital

Theme

Scholar

Education/youth

Coleman (1987, 1988) Fukuyama (1995)

Democracy and governance

de Tocqueville ([1840] 1945) Fukuyama (1995)

Economic development

Knack and Keefer (1997) Helliwell and Putnam (1995 [2000]) Narayan (1999) Narayan and Pritchet (2000) Woolcock and Narayan (2000) Granovetter (1973, 1985) Lin (2001) Bourdieu (1985) Burt (2005) Coleman (1987, 1988) Lowry (1992) Portes and Sensenbrenner (1993) Fine (2001) Portes and Landolt (1996)

Collective action Anthropology/ community life

Negative impacts/ exclusion/harm

Campbell (2000, 2002) Hallinan (2007), Chapter 5 of this volume Putnam (1993a and 1993b, 1995, 2000, 2002) Campbell (2002) Becker (1996) Helliwell (1996) Stiglitz (1999) Burt (1992, 1997) Greif (1993, 1994)

Ostrom and Ahn (2003) Ostrom (1990, 1992, 1998) Ostrom (1990, 1992, 1998) Portes (1994, 1995, 1996, 1998)

Putnam (2000)

product innovation, enhanced knowledge-sharing and cross-functional team eﬀectiveness, improved pool of talent for ﬁrm recruiting and turnover rate reduction. Scholars have also shown the influence of social capital across organizations, including strengthened supplier relations, strengthened interﬁrm learning, facilitation of entrepreneurship and start-up formation, and reduction in organizational dissolution. Table C.2 summarizes scholarship on the beneﬁts (and costs) of social capital within the management arena.

CRITICISMS OF SOCIAL CAPITAL RESEARCH Social capital, however, has not escaped scathing criticism. Perhaps the enthusiastic response with which many social scientists embraced the

341

Conclusion

Table C.2

Benefits (costs) of social capital within management arena

Theme

Examples

Scholars

Individual benefits within organization

• Search for job opportunities

Bourdieu (1985), Bourdieu and Wacquant (1992), Granovetter (1973, 1985), Burt (1992, 2005), Lin (2001), Bourdieu, (1985), Burt (1997) Burt (1992), Gabbay and Zuckerman (1998), Podolny and Baron (1997)

• Executive compensation • Overall career success

Inter-organizational benefits

• Competitive advantage • Intellectual capital creation • Improved resource exchanges • More product innovation

Nahapiet and Ghoshal (1998) Nahapiet and Ghoshal (1998) Tsai (2000) Gabbay and Zuckerman (1998) Tsai and Ghoshal (1998), Tsai (2002)

• Knowledge-sharing/crossfunctional team effectiveness • Scarce resource distribution Kilduff and Tsai (2003), • Improved pool of Fernandez et al. talent for firm recruiting (2000) • Turnover rate Krackhardt and reduction Hanson (1993) Intra-organizational benefits

• Reduction in Pennings et al. (1998) organizational dissolution • Entrepreneurship/ Chong and Gibbons, facilitation start-up (1997) • Strengthens supplier Asanuma (1985), Baker relations (1990), Dore (1983), Gerlach (1992), Smitka (1991), Uzzi (1997) • Strengthens interfirm Kraatz (1998) learning

342

Social capital

concept sowed some of the seeds of its inevitable critique. For, as Narayan and Pritchett assert: ‘Social capital, while not all things to all people, is many things to many people’ (1997, p. 2). As outlined in the Introduction, the diﬃculty starts with the lack of a well-accepted and well-adhered-to deﬁnition of social capital. As a result, numerous challenges arise: two of the most important are whether social capital should be conceptualized as a universal phenomenon or one that is deﬁned by its unique history and context; and which factors should be considered endogenous versus exogenous to social capital formation and use. As Burt colorfully describes the existing state of aﬀairs: ‘Social capital is the Wild West of academic work. There are no skill or intellectual barriers to entry . . . The variety is as interesting and exciting as it is corrosive to cumulative work’ (2005, p. 5). These concluding reflections do not try to tame this Wild West; however, we will attempt to map the frontiers of existing scholarship and indicate possible future research directions. Contextually Bounded vs Universally Conceived Social Capital Whether social capital should be understood as a contextually bounded phenomenon – a product of the confluence of unique circumstances – or a universally conceived one remains a major challenge. Bourdieu, Granovetter, Fine and many other social scientists argue that social capital is both created and limited by the unique circumstances of its history. Consequently, social capital is both contextual and constructed by members of speciﬁc collectives. To illustrate this point, let us turn to Bourdieu’s original research. He explored the diverse foundations of social order to understand better the drivers behind social, economic and power inequality. Social capital represented the resources individuals could call upon based on the speciﬁc density and durability of their particular social ties (Bourdieu, 1985). Certainly relationships required work, particularly to be transformed into viable resources based on mutually felt obligations. Ultimately, Bourdieu’s theories of social capital attempted to explain the persistence of ﬁnancial resources within social classes across generations, thereby linking social capital to class stratiﬁcation. Other proponents of social capital as contextually bounded may diﬀer in their speciﬁc details but reinforce the basic concept. Granovetter (1973, 1985) claims that social theory must be embedded within its social context. He argues that ﬁrst, an individual’s actions always occur within a particular social situation and cannot be explained without reference to that situation, and second, that social institutions are built within speciﬁc social circumstances rather than arising automatically in some general and

Conclusion

343

inevitable form. Fine observes that the preponderance of social capital deﬁnitions cannot be separated from its uses, which means that social capital must depend on the unique context of its creation (Fine, 2001, p. 103). Woolcock further argues that two critical elements of social capital – trust and reciprocity – ‘are nurtured in and by particular combinations of social relationships . . . but they do not exist independently of social relationships’ (Woolcock, 1998, p. 185). By contrast, other scholars such as Becker, Coleman and Putnam propose an understanding of social capital as a universally conceived phenomenon. At its most basic and universal deﬁnition, Becker (1996) contends that social capital represents any social (read non-market) interaction with continuing eﬀect. Given the multitude of imperfect markets, economists could use social capital to help explain persistent improvements in economic performance. Within this paradigm, Coleman may have done more to adopt rational choice assumptions from neoclassical economics and introduce them into the broader social sciences than any other scholar. By applying rational choice theory to contemporary sociology, he provided broader social theory with a universal foundation. Social capital became the by-product of each individual rationally pursuing his or her own self-interest. This theoretical by-product of universal assumptions became a way to explain why people cooperate. Through his massive investigation of the drivers of academic achievement of school children in the US, Coleman used the concept of social capital to explain the ﬁnding that students tended to perform better in Catholic schools than public (that is, state-funded) schools, especially when socio-economic, ethnic, ﬁnancial and other factors have been taken into account. He focused on a particular case study – family and community relationships that support academic achievement – prior to developing a theory of how human capital and social capital reinforce each other. Putnam and his colleagues, like Coleman, employed a speciﬁc case study to develop their more general theories of social capital. Putnam’s initial study concerned the role of associations and civic engagement in explaining diﬀerences in the performance of democratic institutions between Northern and Southern Italy. Following this research, Putnam turned his attention to the state of the social capital of the US. For Putnam, ‘Social capital here refers to features of social organization, such as trust, norms, and networks, that can improve the eﬃciency of society by facilitating coordinated actions’(1993, p. 167). More practically, social capital reinforces collective action in many ways – by sanctioning behavior inconsistent with community norms, fostering reciprocity through repeat interactions, and facilitating information flows including transmission of individual reputations.

344

Social capital

Both Coleman and Putnam ﬁrst narrowly circumscribed their deﬁnitions of social capital to ﬁt the needs of their speciﬁc empirical case studies. Coleman identiﬁes family and community relationships that support academic achievement. Putnam focuses on manifestations of civic engagement such as associational membership. It is perhaps interesting to note that both Coleman’s ﬁndings and Putnam’s initial research – upon which they subsequently framed their more general theories – either have not been replicated or have been scathingly criticized on methodological grounds. Indeed, Hallinan’s own massive longitudinal study of Chicago public and Catholic schools described in Chapter 5 could not recreate Coleman’s earlier results. Yet, despite the possible flaws in their original empirical research, their theories of social capital as a universal phenomenon have survived and continue to be embraced by scholars and policy-makers alike. The theoretical and empirical implications arising from adopting either the contextual or the universal perspective of social capital pose substantive questions for future research in both directions. Since scholars will inevitably continue to debate such a fundamental question as the embeddedness or universality of social capital, perhaps a way forward is to call both for clarity of perspectives on this dimension and then an accumulation of scholarship speciﬁcally within each stream of research. Endogenous vs Exogenous Variables As with any complex, dynamic phenomenon, confusion frequently arises as to whether certain critical variables should be considered endogenous or exogenous to the model. The role of trust is the most vexing issue in this category. Scholars have adopted contradictory methodologies – some assuming that trust is exogenous, while others consider it endogenous. This confusion leads to even more questions about how the eﬀects of social capital might influence the further creation, change or maintenance of social capital. To illustrate, several economists and political scientists assume that trust is exogenous. Their research ties social capital outcomes such as local government eﬀectiveness (Putnam, 1993) and the performance of society’s institutions (Fukuyama, 1995) to levels of trust within a community. Knack and Keefer (1997) adopt a similar approach, employing the University of Michigan’s World Values Survey to gauge a society’s level of trust and then correlate it to economic growth. By contrast, other scholars, primarily sociologists such as Coleman, Bourdieu and Granovetter, argue that norms such as trust arise from the relationships embedded within the community and that therefore trust is endogenous. As described earlier, for Coleman, trust and other norms are by-products of the rational choices self-interested individuals make.

Conclusion

345

Bourdieu contends that trust arises from the interwoven and durable interactions of members of a similar social class. Reconciling these diﬀerences regarding trust’s role is crucial for social capital and presents a third avenue of potential research.

FRONTIERS OF SOCIAL CAPITAL RESEARCH Identifying and describing all the possible future paths of research regarding social capital lies well beyond the scope of these closing reflections. Instead, we will limit ourselves to three themes that may provide some of the more interesting insights in the coming years. Contextually Bounded Social Capital: The Structural Hole between Individual Personality and Collective Social Capital The contextual perspective on social capital leads to the rather simplesounding observation that individuals shape and influence the communities in which they are members, and therefore their associated social capital. Although this observation may seem obvious, it leads directly to interesting and potentially ground-breaking questions: how do individuals’ character traits and their will aﬀect the networks of which they are a part? Questions of psychology – and in particular, cognition, personality and agency – come to the fore in exploring how individuals and collectives create and employ social capital. Knitting together micro-psychological approaches to macro-sociological phenomena may yield further interesting insights. Kilduﬀ and Tsai (2003) outline how an individual’s understanding of circumstances directly aﬀects the content of his or her immediate relationships. Basic cognition – cognitive balance, cognitive accuracy, cognitive maps – has a direct impact on an individual’s motivation. In Chapter 3 of this book, Fishman makes a similar and eloquent case that more can be learned regarding behavior from the study of the speciﬁc content of particular social ties. This speciﬁcity helps us understand the potential impact of those unique relations. In both cases, the underlying assumption is that how individuals understand the ties that bind them substantially influences how they think about relationships and their subsequent actions – which relations they consider trustworthy and which personal interactions they continue to invest in. These factors also determine which relationships will have broader impact in the community. Further, Kilduﬀ and Tsai (2003) also point out that personality dispositions – which may be as simple as being introverted or extroverted – aﬀect

346

Social capital

how communities are formed. This influence may in fact take the form of a feedback loop within the community. Not only do individual psychological traits aﬀect one’s position in that community, but one’s position within the community can also influence one’s disposition. Moreover, such contextual perspectives can help explore how communities create, build, change and utilize their social capital. Returning to basic principles, serendipity and human agency can drive such changes. On the one hand, new relationships arise from serendipity: two people meet, ﬁnd they have values, interests and experiences in common, and build a friendship. On the other hand, human agency may be required. Moreover, as Fishman points out, exploring the nature of speciﬁc human ties helps one begin to unravel their motivation and consequences. Sometimes we spend time with friends, neighbors and colleagues for the sole reason that we enjoy their company. Yet these same friends, neighbors and colleagues also provide us with new opportunities or assistance to get through hard times. At other times, we get to know others speciﬁcally to gain access to those opportunities or assistance. Indeed, diﬀerent types of social capital may be associated not only with diﬀerent types of people but also with various types of intrinsic or instrumental motivations. In summary, further research into the formation, change and use of collectives and their social capital may beneﬁt from greater attention to the individual contingencies and contexts on which the underlying relationships are founded. Employing Burt’s language on social capital, there remains a unique ‘structural hole’ between the study of individual psychology and collective outcomes. Substantial further insights into human behavior, network formation and ultimately social capital may arise from micro- and macro-scholars systematically bridging this structural hole. Universally Conceived Social Capital: The Question of Measurement Solow (1995, 2000) lays down the challenge to measure the stock of social capital: ‘If social capital is to be more than a “buzzword”, its stock should somehow be measured, even inexactly . . . but measurement seems very far away.’ To address Solow’s challenge, scholars would need to abandon the previously described discussion on the rich contextual aspects of social capital and embrace a more universal approach. Numerous scholars have already attempted to address this question. Yet measurement has proven elusive, primarily because comprehensive deﬁnitions of social capital are multidimensional and incorporate diﬀerent levels and units of analysis. As a consequence, scholars have applied a range of methods to its measurement. As Fine (2001) points out with some

Conclusion

347

caveats, there are at least two theoretically possible approaches to measuring social capital. First, a direct approach that calculates the number of and the strength of relationships within a collective and then employs regression analysis to derive the eﬀects on a social capital model. These eﬀects, for example, could include per capital gross domestic product (GDP) growth or household wealth. Nevertheless, such an approach consciously rests on the assumption that some conceptualizational of social capital remains universally valid. A second, indirect methodological approach is represented by Hallinan’s factor analysis that identiﬁes some of the characteristics of social capital and then correlates those to eﬀects. This method attempts to use the positive or negative externalities that social capital causes to infer its existence. Within both of these approaches, scholars diverge on whether their investigations focus on macro-units of analysis – countries, economies, societies – or micro ones such as individual communities, families, institutions and households. Measurement methodologies also vary based on whether the scholar holds a more horizontal or vertical perspective on social capital. Recent World Bank initiatives focus on horizontal deﬁnitions of social capital in terms of the aggregate of formal and informal associations and institutions either at the level of the entire society or at more micro levels. By contrast, other scholars such as Ostrom view social capital from a more vertical perspective as the collaboration among wellfunctioning public and private institutions in order to co-produce the public goods needed for economic development. The insightfulness of direct measurement at the macro level frequently rests on the strength of the proposed indices. In his study of Italian regional government eﬀectiveness, Putnam (1993) measured social capital as a composite of voter turnout in referenda, newspaper readership and membership in cultural associations. In subsequent research Putnam (2000) treats the aggregate number of associations and their membership as representing a society’s social capital. These associations range from bowling leagues and sporting clubs to political parties and civic organizations. Fukuyama (2000) modiﬁes Putnam’s approach by introducing three additional variables: the cohesiveness of the groups, and their positive or negative externalities such as the radius of trust or distrust. Other scholars have employed variations of Putnam’s and Fukuyama’s approaches; for instance, Onyx and Bullen (2000) created an index of social capital that included variables such as connections with work, family, friends, neighborhood, tolerance of diversity and feelings of trust. Krishna and Shrader (1999) have launched a comparable major social capital initiative under the auspices of the World Bank to measure the phenomenon based on a composite index with pilots in Panama and India.

348

Social capital

Nevertheless, a comprehensive catalogue of the whole number of a society’s informal networks and associations has so far exceeded our reach. In contrast to the approaches above, some scholars have attempted to create a direct measurement of social capital at the micro level. Narayan and Pritchett (1997) surveyed 1400 households in 87 Tanzanian villages to create an index of social capital at the family and community levels based on the density and characteristics of formal and informal networks. The World Bank has supported similar detailed surveys of speciﬁc local institutions to create a composite index resting on density of associations, member heterogeneity and degree of active participation (Bolivia, 2000; Burkino Faso, 1999; Indonesia, 1999). In contrast, Dasgupta (2000b) uses indirect measures at the macro level for social capital; he tries to assess social capital from changes in total factor productivity, rather than as a factor of production comparable to physical and human capital. As another example of this approach, Coleman (1988, 1990) utilized the concept of social capital to explain the network eﬀects leading to Catholic high schools achieving much higher graduation rates than public or other private schools. Proxies for social capital such as trust and civic engagement have also been proposed. Although not directly studying social capital, Hofstede (1980) relates psychological factors, like reactions to inequality, uncertainty and individualism, to economic and political outcomes in 40 nations; Inglehart (1997) examines the impact on similar outcomes from changes in religion, family and other values across 43 countries. Other scholars, such as Knack and Keefer (1997), have employed the University of Michigan’s World Values Survey to relate a nation’s level of trust to investment levels and economic results. Additional researchers who have studied trust in this context include Arrow (1972, 2000), Dasgupta (2000b), Gambetta (2000) and LaPorta et al. (1997). On the micro scale, scholars look for an indication of social capital through the study of networks or norms. Greif (1993, 1994, 1997), Helliwell (1996) and Serageldin (1996) researched horizontal associations such as credit cooperatives, savings associations, commercial guilds and civic associations as proxies for social capital. Others have tried to measure social capital by examining the formation of norms. An iterated Prisoner’s Dilemma can produce rudimentary cooperation. Ostrom’s (1992, 2000) study of Nepalese irrigation systems revealed how informal institutions arise to deal successfully with common pool resource problems in an n-sided Prisoner’s Dilemma game. Varshney (2002) investigated the role of the inter-communal networks in anticipating and dispelling Hindu–Muslim violence in three matched pairs of cities in India.

Conclusion

349

No simple measure can capture a phenomenon as inherently multifaceted and complex as social capital. Woolcock suggests one way to strike the balance between parsimony and complexity, which is to ‘unbundle social capital into its dimensions’. This has been a call to arms for many scholars including Davis and Bartkus in Chapter 13 of this volume: to propose an analytically rigorous, formal model of social capital and then to test its antecedents empirically in order to make some progress toward the ultimate measurement of this phenomenon. Davis and Bartkus argue that social capital is not possible without organizational trust. Within a community or group, organizational trust creates the essential environment for community members to be willing to engage in activities together by mitigating the risks associated with individual and collective action. The authors argue further that the level of organizational trust within a community depends directly on the strength of the networks among its members, the degree of shared collectivist norms such as reciprocity and helpfulness, and the abilities of its members and the group itself. The authors then test this model empirically with a sample of students, and their study teams, in the college of business of a major university. By providing empirical evidence for organizational trust, Davis and Bartkus begin to provide some direction as to how social capital may be better understood and measured in the future. Given the amount of research currently occurring in this arena, it is possible that scholars will successfully create a well-accepted social capital index both at the level of an entire economy or society and at the level of a community or organization. Nevertheless, there is much theoretical and empirical richness here still to explore. Social Capital as Potential Organizing Principle Most scholars acknowledge the importance of trust as a foundational element of economic transactions and social interactions (Arrow, 1972), regardless of whether they consider trust as an endogenous or exogenous variable in their models of human behavior. However, given the complexity inherent in the concept of trust, scholars often only describe a single snapshot – focusing on their empirical case study or their social capital model at time t – within the broader moving picture that represents the dynamic of social capital. Yet incorporating trust into a phenomenon that covers an entire community – which we would argue represents the essence of social capital – presents a potential opportunity to employ social capital as a fundamental organizing principle of collectives. As Ouchi (1980) argued, some type of organization is necessary to coordinate eﬀorts across individuals to achieve objectives under conditions

350

Social capital

of interdependence and uncertainty. Further, the division of work among individuals – without the ability to monitor and sanction others’ actions and without the assumptions of a confluence of interests, intents and goals – implies that the individuals must exchange information and rely on others to do their part. Faced with these challenges, scholars have proposed three organizing mechanisms: market, hierarchy and clan (Ouchi, 1980) or alternatively, price, authority and norms (Adler, 2001; Bradach and Eccles, 1989). The market uses price as a clearing mechanism for transactions between anonymous sellers and purchasers. Hierarchy coordinates labor through acknowledged authority and is especially useful in situations needing command and control. The principle of norms and of self-organizing community, or ‘collaborative community’ (Heckscher and Adler, 2006), presents a third possible organizational alternative. Ostrom argues that norms are internal commitments within a community of which actions should be undertaken and which should not. When individuals adopt shared norms, and accept a commitment toward each other, they can frequently organize a wide variety of activities without using market or hierarchical institutions. Indeed, norms can informally coordinate activities through kinship or mutual commitment, thereby subjugating individual interests to collective needs. Yet scholars have struggled to come to grips with this principle. Perhaps Heckscher and Adler (2006) have come closest. Their argument is that a new form of collaborative community within the ﬁrm has evolved to address the challenges posed by recent major economic transformations. Collaborative communities provide the basis for conﬁdence, trust and mutually reinforcing expectations. The critical question that persists, however, is how to scale up the dyadic trust enjoyed by two people described by Mayer et al. (1995) to this level of an organization so that interdependent activities can be coordinated under uncertain conditions. Based on Tonnies’ original scholarship on Gemeinschaft and Gesellschaft, Heckscher and Adler hypothesize that ‘thick’ and ‘thin’ trust exist, respectively. The former represents a high degree of conﬁdence in others’ actions due to the long-term stability of relationships, while the second represents rather less trust, with one merely expecting others to act in their own self-interest. Yet the question remains, how can scholars better describe organizations that ﬁt neither market nor hierarchical models? And how might recent scholarship on social capital inform this important debate regarding organizing principles? Organizational scholars need to take into account the rapid changes in technology and the economy. The innovation imperative in the new knowledge economy, as well as changes in technology and physical supply chain improvements, have driven transaction costs down,

Conclusion

351

sometimes almost to zero. These economic and technological trends impact upon hierarchy and market relations diﬀerently. Perhaps more interestingly, they also add more ambiguity in performance and drive-increased interdependence – critical aspects that Ouchi argues support non-market and nonhierarchical mechanisms. How might social capital help create a more satisfactory description of the current situation and help reﬁne the organizing paradigm of norms? Some scholars argue that aspects of social capital such as trust provide the mechanism to structure and mobilize interdependent activities (McEvily et al., 2003). Others emphasize the communication and surveillance and sanctioning mechanisms that social capital provides which would directly assist in coordination (Burt, 2005). Exploration of social capital as organizing paradigm also leads directly to questions of the boundaries of organizations. Coase (1937) and Williamson (1975, 1985) argue that ﬁrms are formed when markets fail to manage and reduce high transaction costs. By contrast, Kogut and Zander (1996) and Nahapiet and Ghoshal (1998) outline positive reasons for the formation of ﬁrms, namely to drive competitive advantage by creating intellectual capital. Social capital within a ﬁrm enables the combination and exchange of information and knowledge, thereby leading to formation of intellectual capital. Given the challenges posed by changes in technology and knowledge, both of which increase the reliance on mechanisms other than price and authority to coordinate complex activities (Adler, 2001), further exciting new insights may lie at the crossroads of the study of organizational principles and social capital. Social capital research has made a signiﬁcant impact on the broader academic community. It has ensured that scholars acknowledge the critical importance of the social – regardless of how they classify or characterize it – and incorporate it into their analyses. Given its interdisciplinary nature, social capital also presents a natural platform on which scholars of diﬀerent disciplines can interact, discuss and debate their research. This book has been dedicated to fostering further this critical dialogue among disciplines.

REFERENCES Adler, P.S. (2001), ‘Market, hierarchy, and trust: the knowledge economy and the future of capitalism’, Organization Science, 12 (2), 215–34. Adler, P.S. and B. Borys (1996), ‘Two types of bureaucracy: enabling and coercive’, Administrative Science Quarterly, 41 (1), 61–89. Adler, P.S. and S.W. Kwon (2002), ‘Social capital: prospects for a new concept’, Academy of Management Review, 27 (1), 17–40.

352

Social capital

Arrow, Kenneth (1972), ‘Gifts and exchanges’, Philosophy and Public Aﬀairs, 1 (Summer), 343–62. Asanuma, B. (1985), ‘The organization of parts purchases in the Japanese automotive industry’, Japanese Economic Studies, 13 (Summer), 32–78. Baker, W. (1990), ‘Market networks and corporate behavior’, American Journal of Sociology, 96, 589–625. Baker, W.E. and R.R. Faulkner (1993), ‘The social organization of conspiracy: illegal networks in the heavy electrical equipment industry’, American Sociological Review, 58, 837–60. Becker, Gary (1996), Human Capital, 3rd edn, Chicago, IL: University of Chicago Press. Bourdieu, P. (1985), ‘The forms of capital’, in J.G. Richardson (ed.), Handbook of Theory and Research for the Sociology of Education, New York: Greenwood, pp. 241–58. Bourdieu, P. and L.J.D. Wacquant (1992), An Invitation To Reflexive Sociology, Chicago, IL: University of Chicago Press. Bradach, J.L. and R.G. Eccles (1989), ‘Price, authority, and trust: from ideal types to plural forms’, Annual Review of Sociology, 15, 97–118. Burt, Ronald S. (1992), Structural Holes: The Social Structure of Competition, Cambridge, MA and London, UK: Harvard University Press, pp. 8–30. Burt, Ronald S. (1997), ‘The contingent value of social capital’, Administrative Science Quarterly, 42, 339–65. Burt, Ronald S. (2005), Brokerage and Closure: An Introduction to Social Capital, Oxford: Oxford University Press, pp. 333–63. Burt, Ronald S. and M. Knez (1996), ‘Trust and third-party gossip’, in R.M. Kramer and T.R. Tyler (eds), Trust in Organizations: Frontiers of Theory and Research, Thousand Oaks, CA: Sage, pp. 255–92. Campbell, David (2000), ‘The proactive employee: managing workplace initiative’, Academy of Management Executive, 14 (3), 52–66. Campbell, David (2002), ‘Vote early, vote often’, Hoover Institute Paper, Palo Alto, CA: Stanford University. Coase, R.H. (1937), ‘The nature of the ﬁrm’, Economica, 4 (16), 386–405. Coleman, James (1987), ‘Norms as social capital’, in Gerard Radnitzky and Peter Bernholz (eds), Economic Imperialism: The Economic Approach Applied Outside the Field of Economics, New York: Paragon House Publishers, pp. 133–55. Coleman, James (1988), ‘Social capital in the creation of human capital’, American Journal of Sociology, 94 (Supplement: Organizations and Institutions: Sociological and Economic Approaches to the Analysis of Social Structure), S95–S120. Coleman, James (1990), Equality and Achievement in Education, Boulder, CO: Westview. Dasgupta, Partha (2000a), ‘Social capital and economic performance analytics’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 1–31. Dasgupta, Partha (2000b), ‘Economic progress and the idea of social capital’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 325–424. de Tocqueville, Alexis ([1840] 1945), ‘Of the use which the Americans make of public associations in civil life’, ‘Of the relation between public association and the news-

Conclusion

353

papers’, ‘Relation of civil to political association’ and ‘How the Americans combat individualism by the principle of self-interest rightly understood’, in Phillips Bradley (ed.), Democracy in America, Volume 2, New York: Random House. Dore, R. (1983), ‘Goodwill and the spirit of market capitalization’, British Journal of Sociology, 34, 459–82. Fernandez, R.M. (1991), ‘Structural bases of leadership in intraorganizational networks’, Social Psychology Quarterly, 54, 36–53. Fine, Ben (2001), Social Capital Versus Social Theory: Political Economy and Social Science at the Turn of the Millennium, London: Routledge. Fine, Ben (1999), ‘The developmental state is dead – long live social capital?’, Development and Change, 30 (1), 1–19. Fukuyama, Francis (1995), ‘Social capital and the global economy’, Foreign Aﬀairs, 74 (5), 89–103. Fukuyama, Francis (2000), ‘Social capital and civil society’, International Monetary Fund Working Paper, WP/00/74, 1–18. Gabbay, S.M. and E.W. Zuckerman (1998), ‘Social capital and the opportunity in corporate R&D: the contingent eﬀect of contact density on mobility expectations’, Social Science Research, 27, 189–217. Gambetta, Diego (2000), ‘Can we trust trust?’, in Diego Gambetta (ed.), Trust: Making and Breaking Cooperative Relations, Electronic Edition, Oxford: Department of Sociology, University of Oxford, pp. 213–37, http://www. sociology.ox.ac.uk/papers/gambetta 213-237.pdf. Gerlach, M.L. (1992), Alliance Capitalization: The Social Organization of Japanese Business, Berkeley, CA: University of California Press. Granovetter, Mark S. (1973), ‘The strength of weak ties’, American Journal of Sociology, 78 (6), 1360–80. Granovetter, Mark S. (1985), ‘Economic action and social structure: the problem of embeddedness’, American Journal of Sociology, 91 (3), 481–510. Greif, Avner (1993), ‘Contract enforceability and economic institutions in early trade: the Maghribi traders’ coalition’, American Economic Review, 83, 525–48. Greif, Avner (1994), ‘Cultural beliefs and the organization of society: a historical and theoretical reflection on collectivist and individualist societies’, Journal of Political Economy, 102, 912–50. Greif, Avner (1997), ‘Cultural beliefs as a common resource in an integrating world: an example from the theory and history of collectivist and individualist societies’, in P. Dasgupta, K.G. Maler and A. Vercelli (eds), The Economics of Transnational Commons, Oxford: Clarendon Press. Heckscher, Charles and Paul S. Adler (2006), The Firm as Collaborative Community: Reconstructing Trust in the Knowledge Economy, Oxford: Oxford University Press. Helliwell, John H. (1996), ‘Economic growth and social capital in Asia’, NBER Working Paper 5695, National Bureau of Economic Research, Cambridge, MA. Helliwell, John H. and Robert Putnam (1995), ‘Economic growth and social capital in Italy’, reprinted in P. Dasgupta and I. Serageldin (eds) (2000), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 253–68. Hofstede, Geert (1980), Culture’s Consequences: Comparing Values, Behaviors, Institutions, and Organizations Across Nations, Thousand Oaks, CA: Sage Publications. Inglehart, Ronald (1997), Modernization and Post-Modernization, Princeton, NJ: Princeton University Press.

354

Social capital

Kilduﬀ, M. (1990), ‘The interpersonal structure of decision making: a social comparison approach to organizational choice’, Organizational Behavior and Human Decision Processes, 47, 270–88. Knack, Stephen and Philip Keefer (1997), ‘Does social capital have an economic payoﬀ? A cross-country investigation’, Quarterly Journal of Economics, 112 (4), 1251–88. Kogut, B. and U. Brass Zander (1996), ‘What do ﬁrms do? Coordination, identity, and learning’, Organizational Science, 7 (5), 502–18. Kraatz, M.S. (1998), ‘Learning by association? Interorganizational networks and adaption to environmental change’, Academy of Management Review, 41, 621–43. Krackhardt, D. and J.R. Hanson (1993), ‘Informal networks: the company behind the chart’, Harvard Business Review, 71 (4), 104–11. LaPorta, R., F. Lopez-de-Silanes, A. Shleifer and R.W. Vishny (1997), ‘Trust in large organizations’, American Economic Review: Papers and Proceedings, 87, 333–8. Lin, Nan (2001), Social Capital: A Theory of Social Structure and Action, Cambridge: Cambridge University Press. Lin, Nan, W. Engel and J.C. Vaughn (1981), ‘Social resource and strength of ties: structural factors in occupational status attainment’, American Sociological Review, 46, 393–405. Mayer, Roger C., J.H. Davis and F.D. Schoorman (1995), ‘An integrative model of organizational trust’, Academy of Management Review, 20 (3), 709–34. McEvily, B., V. Perrone and A. Zaheer (2003), ‘Trust as organizing principle’, Organizational Science, 14 (1), 91–103. Mehra, A., M. Kilduﬀ and Dan Brass (2001), ‘Social networks of high and low selfmonitors: implications for workplace performance’, Administrative Science Quarterly, 35, 121–46. Nahapiet, Janine and Sumantra Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23 (2), 242–66. Narayan, D. (1999) ‘Bonds and bridges: social capital and poverty’, World Bank papers. Narayan, Deepa and Lant Pritchett (1997), ‘Cents and sociability: household income and social capital in rural Tanzania’, Washington, DC: World Bank. Narayan, Deepa and Lant Pritchett (2000), ‘Social capital: evidence and implications’, in P. Dasgupta and I. Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 269–95. Onyx, Jenny and Paul Bullen (2000), ‘Measuring social capital in ﬁve communities’, Journal of Applied Behavioral Science, 36 (1), 23–42. Ostrom, Elinor (1990), Governing the Commons: The Evolution of Institutions for Collective Action, Cambridge: Cambridge University Press. Ostrom, Elinor (1992), ‘Institutions as rules-in-use’, in Crafting Institutions for Self-Governing Irrigation Systems, San Francisco, CA: ICS Press, pp. 19–39, references. Ostrom, Elinor (1998), ‘A behavioral approach to the rational choice theory of collective action,’ American Political Science Review, 92 (1), 1–22. Ostrom, Elinor (2000), ‘Social capital: a fad or a fundamental concept’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 172–214.

Conclusion

355

Ostrom, Elinor and T.K. Ahn (2003), Foundations of Social Capital, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Ouchi, W.G. (1980), ‘Markets, bureaucracies, and clans’, Administrative Science Quarterly, 25, 129–41. Pennings, J.M., K. Lee and A. Van Witteloostuijn (1998), ‘Human capital, social capital, and ﬁrm dissolution’, Academy of Management Journal. 41, 425–40. Podolny, J.M. and J.N. Baron (1997), ‘Resources and relationships: social networks and mobility in the workplace’, American Sociological Review, 62, 673–93. Portes, A. (1984), ‘The rise of ethnicity: determinants of ethnic perceptions among Cuban exiles in Miami’, American Sociological Review, 49, 383–97. Portes, A. (1987), ‘The social origins of the Cuban enclave economy in Miami’, Sociological Perspectives, 30, 340–72. Portes, A. (1988), ‘Social capital: its origins and applications in modern sociology’, Annual Review of Sociology, 24, 1–24. Portes, A. (1994), ‘The informal economy and its paradoxes’, in N.J. Smelser and R. Swedberg (eds), The Handbook of Economic Sociology, New York: Russell Sage Foundation, pp. 426–49. Portes, A. (1995), ‘Economic sociology and the sociology of immigration: a conceptual overview’, In A. Portes (ed.), The Economic Sociology of Immigration: Essays on Networks, Ethnicity and Entrepreneurship, New York: Russell Sage Foundation, pp. 1–41. Portes, A. (1996), ‘A dissenting view: pitfalls on focusing on relative returns to ethnic enterprise’, National Journal of Sociology, 10, 45–7. Portes, A. (1997), ‘Los Angeles in the context of the new immigration’, Working Paper No. 18, Lewis Center for Regional Policy Studies, School of Public Policy and Social Research, University of California, Los Angeles. Portes, A. (1998), ‘Social capital: its origins and applications in modern sociology’, Annual Review of Sociology, 24, 1–24. Portes, A. (2000), ‘The meanings of social capital’, Sociological Capital Forum, 15, 1–12. Portes, A. and P. Landolt (1996), ‘The downside of social capital’, American Prospect, 26 (7), 18. Portes, A. and J. Sensenbrenner (1993), ‘Embeddedness and immigration: notes on the social determinants of economic action’, American Journal of Sociology, 98 (6), 1320–50. Portes, A., J.M. Clark and M.M. Lopez (1982), ‘Six years later: the process of incorporation of Cuban exiles in the United States, 1973–1979, Cuban Studies, 11–12, 1–24. Putnam, Robert D. (1993), ‘The prosperous community: social capital and public life’, American Prospect, 13, 35–42. Putnam, Robert D. (1995a), ‘Bowling alone: America’s declining social capital’, Journal of Democracy, 6 (1), 65–78. Putnam, Robert D. (1995b), ‘Tuning in, tuning out: the strange disappearance of social capital in America’, PS: Political Science and Politics, 28 (4), 664–83. Putnam, Robert D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Putnam, Robert D. (ed.) (2002), Democracies in Flux: The Evolution of Social Capital in Contemporary Society, New York: Oxford University Press. Serageldin, Ismail (1996), ‘Sustainability as opportunity and the problem of social capital’, Brown Journal of World Aﬀairs, 3, 187–203.

356

Social capital

Smelser, Neil (ed.) (1988), Handbook of Sociology, Beverly Hills, CA: Sage. Smitka, M. (1991), Competitive Ties: Subcontracting in the Japanese Automotive Industry, New York: Columbia University Press. Solow, Robert (1995), ‘But verify’, The New Republic, 11 September, pp. 36–9 (review of F. Fukuyama, The Social Virtues and the Creation of Prosperity). Solow, Robert (2000), ‘Notes on social capital and economic performance’, in Partha Dasgupta and Ismail Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 6–12. Stiglitz, J. (1999), ‘Whither reform? Ten years of the transition’, World Bank Annual Conference on Development Economics, Keynote Address. Stiglitz, J.E. (2000), ‘Formal and informal institutions’, in P. Dasgupta and I. Serageldin (eds), Social Capital: A Multifaceted Perspective, Washington, DC: World Bank, pp. 59–70. Tsai, W. (2000), ‘Social capital, strategic relatedness, and the formation of intraorganizational linkages’, Strategic Management Journal, 21, 925–39. Tsai, W. (2001), ‘Knowledge transfer in intra-organizational networks: the eﬀects of network position are absorptive capacity on business and innovation and performance’, Academy of Management Journal, 44, 996–1004. Tsai, W. (2002), ‘Social structures of “Competition” within a multi-unit organization: coordination, competition, and intra-organizational knowledge sharing’, Organization Science, 13, 79–90. Tsai, W. and S. Ghoshal (1998), ‘Social capital and value creation. the role of intraﬁrm networks’, Academy of Management Journal, 41, 464–76. Uzzi, B. (1997), ‘Social structure and competition in interﬁrm networks: the paradox of embeddedness’, Administrative Science Quarterly, 42, 35–67. Varshney, Ashutosh, (2002), Ethnic Conflict and Civic Life: Hindus and Muslims in India, New Haven, CT: Yale University Press. Williamson, Oliver (1975), Markets, Hierarchies, Analysis and Anti-Trust Implications: A Study in the Economics of Internal Organization, New York: Free Press. Williamson, Oliver (1985), The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting, New York: Free Press. Woolcock, Michael (1998), ‘Social capital and economic development: toward a theoretical synthesis and policy framework’, Theory and Society, 27 (2), 151–208. Woolcock, Michael (2001), ‘The place of social capital in understanding social and economic outcomes’, Canadian Journal of Policy Research/Revue canadienne de recherche sur les politiques, Government of Canada’s Policy Research Secretariat, 2 (1), 1–8. Woolcock, Michael and Deepa Narayan (2000), ‘Social capital: implications for development theory, research, and policy’, World Bank Research Observer, 15 (2), 225–49. World Development (1996), Special edition on social capital.

Index ability 307–8, 323, 325, 328 Abrams, D. 247 academic achievement and social capital study approach 148–9 background 145–9 data sample 149–50 descriptive analyses ﬁndings 151, 152 variables 150–51 inferential analyses ﬁndings 153–7 model 154 variables 151, 153–4 accidental brokerage 272 adaptive eﬃciency 225–6 Adler, P.S. 3, 207, 211, 213, 223, 262, 320, 350, 351 Advanced Institute of Management Research 215 aesthetic networks 78, 80 aﬀective transformations 248–50 agency-managed irrigation systems (AMIS) 28–30 Ahn, T.K. 3, 173, 340 Ahuja, G. 210, 224 Allen, T.J. 228 alliance function 227 Allport, G.W. 188 AMIS (agency-managed irrigation systems) 28–30 Apple 63 appropriability 211, 213–14, 226 Araral, E. 31 Arrow, K. 10, 18, 348, 349 Asanuma, B. 341 Ashforth, B.E. 241, 247 Asian Development Bank 28 associational membership 71 associative systems of exchange 212 Atlanta Constitution 162

Baiocchi, G. 80 Baker, W.E. 209, 213, 214, 269, 341 Baldwin, C.Y. 220 banker reputation 47–9 Barber, B. 290 Barney, J. 208 Baron, J.N. 341 Bartkus, V. 288, 305 Baum, J.A.C. 260 Bearman, P. 77 Beck, P.A. 172–3 Becker, G. 21, 340, 343 Bedeian, A.G. 306 Benetton 211 benevolence 307–8 Benjamin, P. 28 Biggart, N. 212, 213 biotechnology industry innovation research 211 Birkinshaw. J. 223, 227 Blader, S.L. 243 Boland, R. 218, 221 bonding social capital 2, 90, 244, 251, 334 Bonds of Civility (Ikegami) 78 Bossert, T.J. 277 boundary-crossing ties 78–9, 80 Bourdieu, P. 3, 212, 214, 226, 340, 341, 345 Bouty, I. 213, 224 Bowling Alone (Putnam) 335 Bradach, J.L. 350 Brass, D.J. 270 Brazil, participatory study 80 Brehm, J.W. 295 Brennan, R.L. 153 Brewer, M.B. 242, 247, 251, 252, 253, 292 bridge relations 50, 51, 60 bridging social capital 2, 90, 244, 251, 334 bridging ties 219 357

358

Index

Brief, A.P. 311 Brock, T.C. 194 brokerage 39, 60–62, 73, 264–9 accidental brokerage 272 choice of broker 266–7 connection to brokers 265–6 and growth 42–5, 216 maintaining relationships 268–9 as requirement of closure 264 risk management 266–8 in virtual organization 56 see also brokers Brokerage and Closure (Burt) 73 brokers 43, 228, 260–61, 265–6 choice of 266–7 reasons to connect to 269–71 reputation of 268–9 see also brokerage Bromiley, P. 282 Brown, J.S. 40, 62, 215, 218, 221, 223 Brown, S. 222 Bryk, A. 148 Bullen, P. 276, 277, 278, 347–8 Burgess, G. 298 Burgess, H. 298 Burt, R.S. 2, 3, 5, 18, 19, 20, 73, 74, 145, 165, 186, 207, 209, 223, 228, 260, 265, 268, 340, 341, 342, 351 Butler, J.K. 291, 295 calculus-based distrust (CBD) 285, 295–7 calculus-based trust (CBT) 279–80, 284, 286, 291–4 Caldeira, G.A. 161 Campbell, D.E. 340 Campbell, D.J. 340 Canada Equality, Security and Community Survey (ESC) 90, 91–2, 93–100, 111–12, 113, 116–21, 131, 136–9 Ethnic Diversity Survey 92 General Social Survey-17: Social Engagement Survey (GSS) 90, 92–5, 101–6, 112, 114, 122–5, 131–2, 142–4 Statistics Canada 93 capital deﬁnition 20 negative eﬀects 23

capital assets 22, 23, 24, 67–8 Cappeli, P. 253 CARE 28 Carlile, P. 221 Castells, M. 206 category-based trust 292 Catholic school advantage 7–8, 147–8, 343 Catholic schools, achievement and social capital study approach 148–9 background 145–9 data sample 149–50 descriptive analyses ﬁndings 151, 152 variables 150–51 inferential analyses ﬁndings 154–7 model 154 variables 151, 153–4 CBD see calculus-based distrust (CBD) CBT see calculus-based trust (CBT) Chaiken, S.L. 295 Chicago School Study 149 China, structural holes 229 choice behaviour in collective contexts 245–50 cognitive transformations 246–7 hedonic/aﬀective transformations 248–50 motivational transformations 247–8 Chou, Y.K. 276 Christensen, C.M. 39 Chung, S. 263 Cialdini, R.B. 285 Citizenship Participation Study 169–70 civic norms 163–7, 174, 175–8, 178, 180 civics instruction 172–3 Clark, K.B. 220 closure 39, 60–62, 73, 217, 262–4, 271 brokerage requirement 264 creation of 263–4 and stability 46–50, 51 in virtual organization 56 see also intergenerational social closure closure networks 5, 145, 146 Coase, R.H. 351

Index cognitive social capital 218 cognitive transformations 246–7 Cohen, S.S. 222 Coleman, J. 3, 5, 7–8, 18, 49, 72, 145, 146–7, 164, 172, 176, 186, 207, 212, 213, 214, 226, 227, 263, 264, 267, 276, 277, 320, 324, 340, 343, 344, 348 collaborative advantage 209–15, 220, 230 collaborative community 350 collective action 160, 161, 169–70, 181 see also voter turnout collective commitments 223–4 collective identity 241–2, 251, 252, 294 cognitive transformations 246, 247 hedonic/aﬀective transformations 249–50 motivational transformations 247–8 self-aﬃrmation 248 self-categorization 243, 246 and social capital, causal relationships 253 and social capital creation 243–5 collective norms 174, 324–5, 328 common space 63 communal systems of exchange 212–13 community homogeneity 165–7, 170 community norms 172 competitive advantage 209–10, 211 complementary connections 263 Conover, P.J. 172 consensus 57–9 consensus premature 57–9, 62 contagion 198–200, 201 contextually bounded social capital 342–3, 344, 345–6 conversational themes 77–8 conversational ties 78–9 convertibility 211, 214–15 Conway, M.M. 162 Cook, S. 221 Cool, K. 208 cooperation 4–5 cooperative advantage see collaborative advantage Coward, E.W. 33 Cox, G.W. 162 cross-cultural research 229–30

359

Cuervo-Cazurra, A. 218 cultural associations and ties 78 Cummings, L.L. 282 Dasgupta, P. 10–11, 348 Davis, J.H. 288, 305 Davis, J.L. 194 Dawes, R.M. 292 De Clercq, D. 277 de Tocqueville, A. 3, 208, 340 dead capital 23 DeFillippi, R. 209 Delbridge, R. 212, 213, 224 Democracy’s Voices (Fishman) 66, 76, 78 density of networks 25, 146, 149, 153, 158, 240–41 depersonalized trust 252–3 deterrence-based trust 279 Deutsch, M. 278, 279, 280 development agencies 28, 30, 31–3 Dhanaraj, C. 228 Dierickx, I. 208 diﬀerentiating network 60, 62 Digital Millennium Copyright Act 63 DiMaggio, P. 206 Dirks, K.T. 287 distrust 275, 281–5 calculus-based 285, 295–7 identiﬁcation-based 285, 297–8 Dore, R. 341 Downs, A. 161 Doz, Y. 216, 222, 223, 224 Drucker, P. 215, 216, 230 Duguid, P. 215, 218 Dunn, E.W. 116 Dunn, J.R. 294 Durlauf, S.N. 24 dyadic trust 252, 320–21 dysfunctional norms 290 Easterly, W. 31 eBay 271 Ebers, M. 209, 214, 219, 226 Eccles, R.G. 350 economic capital 68, 70, 74, 214–15 economic exchange 212 Ehrhart, M.G. 289 Eisenhardt, K. 206, 222

360

Index

elections, voter turnout 160 adolescence and adult voting link 177–8 and electoral competition 161–3, 167–9 homo civicus 163–7 homo politicus 161–3 school civics instruction, eﬀect of 172–3 school experience in adolescence, eﬀect of 173–7 and social networks 194–7 voters’ motivations 169–71 Elfring, T. 219 Elster, J. 249 embedded ties 211 embeddedness 211, 212 embedding network 60, 62–3 Enron 314 Equality, Security and Community Survey (ESC) 90, 91–2, 93–100, 111–12, 113, 116–21, 131, 136–9 Erickson, E.G. 278 ESC survey see Equality, Security and Community Survey (ESC) Ethnic Diversity Survey 92 ethnicity and workplace trust 112, 113 Evans, P. 32 expectations 289, 290 expertise and use of broker 270 farmer-managed irrigation systems (FMIS) 28–30, 32 Farr, J. 208 Federal Communications Commission 63 Ferejohn, J.A. 162 Ferlie, E. 218 Fernandez-Mateo, I. 267 Fernandez, R.M. 341 Ferrin, D.L. 287 ﬁeld experiments 188–9 Fields, G. 222 Fine, B. 340, 343, 347 Fine, G.A. 20, 293 Fiore, F. 162 Fiorina, M.P. 162 Fisher, R. 297 ﬂexible embedding networks 62–3

Florin, J. 215 ﬂow assets 22 FMIS (farmer-managed irrigation systems) 28–30, 32 Follett, M.P. 210 Ford, D. 227 Forgas, J.P. 295 formal rules (rules-in-form) 27, 32 Foster, C.B. 162 Fourtou, Jean-René 39 free-riding 250, 277, 309–10 Frey, B.S. 115 friendship network 22 Fukuyama, F. 213, 340, 344, 347 Gabarro, J.J. 291, 294, 295 Gabbay, S.M. 341 Galaskiewicz, J. 210 Gambetta, D. 20, 23, 348 Gardner, W. 242 Gargiulo, M. 263 General Electrics 39 General Social Survey-17: Social Engagement Survey (GSS) 90, 92–5, 101–6, 112, 114, 122–5, 131–2, 142–4 generalized reciprocity 213 George, J.M. 295, 311 Gerber, A.S. 194, 198 Gerlach, M.L. 341 Get Out The Vote (GOTV) message 198–9 Ghoshal, S. 3, 4, 19, 20, 205, 226, 240, 241, 277, 320, 322, 341, 351 Gibson, C. 227 Gilovich, T. 249 Gladwell, M. 285 Global System for Mobile Communication (GSM) 63 Golembiewski, R.T. 252 Gomes-Casseres, B. 213 GOTV message 198–9 Govier, T. 290 Grabher, G. 39 Granovetter, M. 67, 207, 212, 261, 265, 340, 341 Gray, V. 161 Green, D.P. 194, 198 Green, S. 327 Greif, A. 340, 348

Index

361

group identiﬁcation see collective identity GSM (Global System for Mobile Communication) 63 GSS survey see General Social Survey17: Social Engagement Survey (GSS) Gulati, R. 220, 263 Gumbel, A. 162 Gupta, D.K. 239, 243 Gurin, P. 251

human capital 18, 21, 24, 27, 216 and reputation stability 47 human-made capital deﬁnition 20–21 human capital 21 measurement 24–6 physical capital 21, 22, 23–4 shared characteristics 22–4 social capital 22 types 21–4 Hume, D. 3

Hagedoorn, J. 212 Hagel III, J 40, 62 Hagel, J 223 Halliburton 315, 316 Hamel, G. 206 Hamill, H. 20 Hanifan, L.J. 3 Hansen, J.M. 177 Hansen, M.T. 218 Hanson, J.R. 341 Hardin, G. 19 Hardin, R. 21 Hargadon, J. 218 Hasecke, E.B. 327, 330 health sector innovations study 218 Hechter, M. 164 Heckscher, C. 207, 223, 350 hedonic anticipation 249 hedonic transformations 248–50 Helliwell, J. 276, 340, 348 high distrust 282, 283 High School and Beyond survey 147 high trust 282, 283, 287 Hilton, R. 32 Ho, V.T. 277 Hochberg, Y.V. 215 Hoﬀman, E. 23 Hoﬀstetter, C.R. 161 Hofstede, G. 348 Hogg, M.A. 247, 322, 327 Holyﬁeld, L. 20, 293 homo civicus 160, 161, 163–7, 170–71 homo economicus 5, 161 homo politicus 160, 161–3, 170–71 homophily 263, 265, 272 horizon-expanding schools 147, 148 Huckfeldt, R. 180, 189, 194, 200

IBM 220 identiﬁcation-based distrust (IBD) 285, 297–8 identiﬁcation-based trust (IBT) 253, 280–81, 285, 286, 294–5 identity see social identity IDEO 211 immigration status and workplace trust 112, 113 in-group identity 242–3, 244, 245, 246, 247, 249 Inarra, H. 263 income and workplace trust 92, 93, 94–110, 111, 112 Independent, The 162 India, Rajasthan, study 25–6 indirect brokerage 266 individual personality 345–6 individual social capital 219, 220 informal networks 25 informal rules (with rules-in-use) 27 informal social control 312–14 Inglehart, R. 348 Inkpen, A. 218–19 innovation 218, 220, 228 institutional monocropping 32 institutions as social capital 26–7, 31–3 Nepali irrigation systems 27–30 integrated diversity 39 integrity 307–8 intellectual capital 215–16, 219, 221 intergenerational networks 244 intergenerational social closure 145, 146–7, 148, 156, 158, 241 international development agencies 28, 30, 31–3 International Labor Organization 28

362

Index

interorganizational relations 210–11, 218 interpersonal inﬂuence 190, 193, 195–6, 200 interpersonal trust 252, 288 irrigation projects 31–2, 33 Nepali irrigation systems study 27–30 Iyengar, S. 188 Jarzabkowski, P. 222 Jennings, E.E. 293 Johnson, S.A. 241 Jones, R.A. 295 Jorstad, C.M. 23 Joshi, N.N. 28, 29, 30 Junn, J. 172, 173 Kale, P. 227 Keefer, P. 340, 344, 348 Kelley, H.H. 245, 248, 285 Kelly, C. 243 Kelly, J.E. 243 Kelly, T. 59 Kent Jennings, M. 172–3 Kilduﬀ, M. 207, 267, 341, 345, 346 Kinder, D.R. 188 Knack, S. 164–5, 340, 344, 348 Knoke, D. 268 Knorringa, P. 276 knowledge economy 215–17 knowledge processes and social capital 217–19 Kogut, B. 224, 351 Kolen, M.J. 151 Koput, K. 260 Kosonen, M. 222, 223, 224 Kraatz, M.S. 341 Krackhardt, D. 267, 271, 341 Kramer, R. 285, 293, 294 Krishna, A. 25–6, 348 Kulatilaka, B. 224 Kulnych, J. 71 Kwon, S.W. 3, 211, 213, 262, 320 Labianca, G. 270 laboratory experiments 188 Lachmann, L.M. 20, 320 Lam, W.F. 28 Landolt, P. 340

Langbein, L. 23 LaPorta, R. 348 Larson, A. 217 latency 211 latent ties 214, 225 Laumann, E.O. 277 lawyers as brokers 270 Layman, G. 167 Leary, M.R. 248 Li & Fung 211, 228–9 Libecap, G. 23 Lichterman, P. 80 life satisfaction and workplace trust 91–3, 94–5, 111, 112, 113, 115, 130 Likert, R. 228 Lin, N. 18, 19, 145, 261–2, 324, 340, 341 Lind, E.A. 248 Lindskold, S. 252, 296 Lindstrom, M. 276 linking social capital 90–91 Loewenstein, G. 249 Los Angeles Times 162 Loury, G. 3 Lovas, B. 226 low distrust 282, 283 low trust 282, 283 Macinko, J. 276 Mackie, D.M. 194 Mael, F. 247 Making Democracy Work (Putnam) 17 Malhotra, D. 280 manager network 40, 41 March, J.G. 248, 250, 251, 316 Marcus, A. 218 Mariotti, F. 225 Maurer, I. 206, 214, 219, 226 Mayer, R.C. 9, 278, 282, 291, 292, 320, 321, 325, 326 Mayer-Schönberger, V. 62 McAdams, R.H. 164 McAllister, D. 284, 294, 295 McConkie, M. 252 McDonald, M.P. 162, 163 McEvily, B. 217, 218, 225, 260, 351 McNally, D. 3 McPherson, J.M. 263 measurement of social capital 10–11, 69–73, 346–9 Mehra, A. 260, 263

Index Merriam, C. 172 metaphor-only social capital approach 68–9, 73–5 Meyerson, D. 292 Miller, N. 251 Miller, P.J. 281 Milward, H.B. 260 Mitchell, A.D. 277 mobile phone market 63 mobility and decline in social capital 316–17 Mohseni, M. 276 monopoly embedding networks 62, 64 moralistic trust 250 Moran, P. 206, 219 more-than-metaphor social capital approach 68, 69–73 Morgan, S.L. 147–8 Mossholder, K.W. 306 motivational transformations 247–8 motor sport industry, knowledge and networks study 224 Munger, M.C. 162 Musick, M. 188 Nahapiet, J. 3, 4, 19, 20, 240, 241, 277, 320, 322, 341, 351 Narayan, D. 208, 340, 342, 348 National Education Longitudinal Study (NELS) 147–8 natural capital 20 Naumann, S.E. 289 negative ties 270 negative values 291 NELS (National Education Longitudinal Study) 147–8 Nepal, irrigation systems 27–30, 32, 348–9 network centrality 217, 219, 228 network density 25, 146, 149, 153, 158, 240–41 network duality 39, 60–64 brokerage and growth 42–5 closure and stability 46–50, 51 failure modes 60–62 structural holes 40–43 virtual organization 50, 52–5, 62, 63 evaluation 56–9 people turnover 56–7 premature consensus 57–9

363

network orchestration 228–9 network strength 327–8, 330 network ties 78, 288–9, 290 networks 4–5, 20, 262, 288–9, 305–6 as antecedent of organizational trust 322–4, 327–8, 330–31 with intergenerational social closure 145, 146–7, 158 as measure of social capital 348 in norm diﬀusion 186 in research 19 in social capital deﬁnitions 18 with social closure 145, 156, 157, 158 and trust 310–11 and voter turnout 194–6 New York City apparel industry research 211 New Yorker 316 Newell, S. 214, 277 Niemi, R.G. 172, 173 9/11 Commission Report Investigating Terrorist Attacks 239 Nooteboom, B. 276 norm diﬀusion, experimental approaches 186–7, 200–201 exogenous shocks to a network 197–200 randomized experiments, beneﬁts of 187–9 randomized network, creation and measurement 189–93 randomized structured interactions within a network 193–6 norm-enforcing schools 147, 148 norms 4–5, 147, 164–5, 180, 276, 289, 350 as antecedent of organizational trust 322, 324–5, 328 civic norms 163–7, 174, 175–8, 178, 180 dysfunctional norms 290 internalization of 172, 176 as measure of social capital 348 of reciprocity 5, 278, 287 schools and voter turnout study 160, 164–5 adolescence and adult voting link 177–8 and electoral competition 161–3, 167–9

364

Index

homo civicus 163–7 homo politicus 161–3 school civics instruction, eﬀect of 172–3 school experience in adolescence, eﬀect of 173–7 and social networks 194–6 voters’ motivations 169–71 see also norm diﬀusion, experimental approaches North, D.C. 27, 225–6, 277 Obstfeld, D. 228, 230, 269 O’Donoghue, T. 130 Oh, H. 211, 260 Olson, M. 19, 160, 252 Onyx, J. 276, 277, 278, 347–8 Opp, K.-D. 164 opportunistic behavior 272 Organ, D.W. 289 Organisation for Economic Cooperation and Development (OECD) 88 organizational social capital 219, 220 organizational trust 11, 319, 323, 334, 349 as antecedent of social capital 321, 323 antecedents of 322–5 ability 323, 325 networks 322–4 norms 323, 324–5 deﬁnition 320–21 organizational trust model 325–32 control variables 326–7 dependent variable: organizational trust 328, 330–32 future research 334 hypothesis 1: network strength 324, 332–3 hypothesis 2: collective norms 325, 333 hypothesis 3: collective abilities 325, 333 hypothesis 4: network strength, collective norms and abilities 326, 333–4 implications 335 independent variables 327–8 ability 328

collective norms 328 network strength 327–8 methodology 326 results 332–4 Orlikowski, W.J. 221 Orr, J.E. 221 Osgood, C. 296 Osterlund, C. 220 Ostrom, E. 3, 173, 267, 340, 348–9, 350 Ouchi, W.G. 350 Parigi, P. 77 Parkhe, A. 228 partisanship 166–7, 180 Pasotti, E. 22, 27, 240, 276 Patterson, S.C. 161 PayPal 271 peer-to-peer mobilization 195–6 Pennings, J.M. 341 personal identity 242, 245, 246, 247, 249 personal social capital 244 Petroski, H. 24 Pettigrew, A. 206 photocopier technicians study 221 physical capital 18, 21, 22, 23–4, 26, 27, 67–8 Pitman, G. 31 Podolny, J.M. 260, 263, 267, 341 Polanyi, K. 212 political elections, voter turnout 160 adolescence and adult voting link 177–8 and electoral competition 161–3, 167–9 homo civicus 163–7 homo politicus 161–3 school civics instruction, eﬀect of 172–3 school experience in adolescence, eﬀect of 173–7 and social networks 194–7 voters’ motivations 169–71 Popkin, S.L. 162 Porter, M. 210 Portes, A. 18, 277, 340 Postman, L.J. 188 Powell, W. 211, 219, 260 practice perspective on social capital 220–21

Index Prahalad, C.K. 206 premature consensus 60–62, 65 present-biased preferences 130 presumptive trust 252 Prisoner’s Dilemma game 245–6, 348–9 Pritchett, L. 32, 340, 342, 348 Provan, K.G. 260 psychological embeddedness 240, 241–5 psychological transformations of identity 245–50 cognitive 246–7 hedonic/aﬀective 248–50 motivational 247–8 Putnam, R.D. 2, 3, 4, 5, 10, 17, 18, 25, 26–7, 72–3, 157, 164–5, 207, 213, 244, 252, 253, 276, 320, 340, 343–4 Quinn, J.B. 206 Rabin, M. 118, 130 Rajasthan, India, study 25–6 randomized experiments 188–9 rational choice 343 reciprocity 5, 167, 211, 212–13, 217, 277, 324 generalized reciprocity 213 reciprocity norms 287 Reed, R. 209 referrals 268–9, 272 Regmi, A. 28 relational economy 215–17 relational embeddedness 19, 20, 219 relational lock-in 219 relational quality 217–18 relational social capital 217–18, 228, 241, 277 relationship quality 223–4 reputation 46, 49, 56, 225, 272, 291–2 of brokers 268–9 reputation stability 47–9 reservoir of trust 287–81 social capital enhancement 288–90 social capital reduction 288, 290–91 resource ﬂuidity 222, 223 Rhone-Poulenc 39 Ring, P.S. 267 risk 306–7 risk-taking in relationships 308, 321 Robinson, J.A. 162

365

Rogers, C.R. 289 Rokeach, M. 289 role-based trust 292 Rosenstone, S.J. 177 Rothstein, B. 22, 27, 240, 276 Rotter, J.B. 250, 252, 278, 310 Rousseau, D.M. 278, 279, 281, 282, 286, 287, 306 Rowley, T.J. 260 rule-based trust 292–3 rules-in-form (formal rules) 27, 32 rules-in-use (informal rules) 27 Rusbult, C.E. 194 Sabatini, F. 304 safe-harbor regulations 63–4 Saltzer, J.H. 63 Sandefur, R.L. 277 Santos, F.M. 206 Saxenian, A. 39 Schneider, B. 315 schools, achievement and social capital study approach 148–9 background 145–9 data sample 149–50 descriptive analyses ﬁndings 151, 152 variables 150–61 inferential analyses ﬁndings 154–7 model 154 variables 151, 153–4 schools and voter turnout, social norms study 160 adolescence and adult voting link 177–8 and electoral competition 161–3, 167–9 homo civicus 163–7 homo politicus 161–3 school civics instruction, eﬀect of 172–3 school experience in adolescence, eﬀect of 173–7 and social networks 194–7 voters’ motivations 169–71 Schoorman, F.D. 282, 295, 306, 308 Schuller, T. 208, 213 Schultz, T.W. 21

366

Index

Schweitzer, M.E. 294 search costs 269 Searing, D.D. 172 Seibert, S.E. 260 self-aﬃrmation 248 self-categorization 246 Sensenbrenner, J. 340 Serageldin, T. 10, 348 shallow trust 252 Shapiro, D.L. 279, 294 shared social identity see collective identity Sheppard, B.H. 279 Sherer, P.D. 253 Sherman, D.M. 279 Shivakoti, G. 28 Shrader, E. 348 Silicon Valley 39, 222, 244 Simmel, G. 260 Simon, H.A. 249, 316 Sitkin, S. 285 Skandia 223 Slater, R. 39 Smith-Doerr, L. 219 Smith-Lovin, L. 263 Smith, S.S. 71 Smitka, M. 341 social capital 1, 17, 21, 24, 88, 239–40, 261–2, 276–7 composition of 277 criticisms 5–7, 19 decline in 304, 314–17 deﬁnitions 2–4, 18, 22, 305–6, 320 measurement 10–11, 69–73, 346–9 negative eﬀects 23, 277 role of trust 277–8 Social Capital Community Benchmark Survey (US Benchmark) 90, 93–5, 107–10, 111, 112, 114, 126–9, 132, 140–41 social capital research 339–41 criticisms 342–5 contextually bounded vs universally conceived 342–4 variables, endogenous vs exogenous 344–5 future paths 345–51 individual personality and collective social capital 345–6 measurement 346–9

social capital as organizing principle 349–51 management arena themes 341 themes 340 social categorization 246–7 social connections see social ties social control degree of formalization 312–13 scale 314 social identity 165, 239, 241–5, 253, 322, 324, 327 cognitive transformations 246–7 collective identity 241–2, 251, 252, 294 cognitive transformations 246, 247 hedonic/aﬀective transformations 249–50 motivational transformations 247–8 self-aﬃrmation 248 self-categorization 243, 246 and social capital, causal relationships 253 and social capital creation 243–5 hedonic/aﬀective transformations 248–50 in-group identity 242–3, 244, 245, 246, 275, 249 motivational transformations 247–8 personal identity 242, 245, 246, 247, 249 social networks see networks social norms, schools and voter turnout study 160, 164–5 adolescence and adult voting link 177–8 and electoral competition 161–3, 167–9 homo civicus 163–7 homo politicus 161–3 school civics instruction, eﬀect of 172–3 school experience in adolescence, eﬀect of 173–7 and social networks 194–7 voters’ motivations 169–71 social sanctions 164–5, 186 social ties 66–7 metaphor-only social capital approach 73–5

Index more-than-metaphor social capital approach 69–73 ways of studying 75–80 socialization 172 Soda, G. 219 Solow, R.M. 19, 346–9 Sorenson, A.B. 147–8 Spee, A.P. 222 Sprague, J. 180, 189, 194, 200 Sproule-Jones, M. 27 stability and closure 46–50 Stahelski, A.J. 285 Stam, W. 219 Standing, W.H. 162 Starbuck, W.H. 62 Starﬁeld, B. 276 start-ups and business ﬁnancing research 214 Statistics Canada 93 Steele, C. 248 Stepp, D.R. 162 Stewart, T.A. 39 Stickel, D. 285 Stiglitz, J. 340 stockbrokers 269 strategic management 205–7, 229–30 ambidexterity 227 cognitive social capital 218 collaborative advantage 209–15, 220 innovation 220 knowledge economy as relational economy 215–17 knowledge processes and social capital 217–19 network-building 227–8 relational social capital 217–18 and social capital 207–9, 217–19 social capital dimensions, interrelations 218–19 social capital, organizing for 226–9 strategy as practice 222 structural social capital 217, 218 and uncertainty 222–26 strategic opportunities 224–5 strategic sensitivity 222–3 structural embeddedness 19–20, 219, 240–41 structural holes 40–43, 60, 62, 64, 145–6, 219, 260 in diﬀerent cultural contexts 229

367

structural social capital 217, 218, 278 Stuart, T.E. 225, 260 student achievement and social capital study approach 148–9 background 145–9 data sample 149–50 descriptive analyses ﬁndings 151, 152 variables 150–51 inferential analyses ﬁndings 154–7 model 154 variables 151, 153–4 Stutzer, A. 115 subgroup identities see in-group identity Subramaniam, M. 207, 217 ‘sucker’s pay-oﬀ’ 250 Swan, J. 211 swift trust 285–6 Szulanski, G. 216 tacit knowledge 42, 218 Tajfel, H. 322, 324 Tansley, C. 277 Taylor, M. 277 Teece, D. 206 Tenkasi, R. 218, 221 thick trust 350 thin trust 350 ties-as-brokerage 78, 79 ties-as-conversation 78–9 Tomlinson, E.C. 296 Townsend, A. 251 Toyota 211, 229 trust 4, 9–10, 147, 217, 252–3, 305–6 and brokers 271 calculus-based 279–80, 284, 286, 291–4 category-based 292 concept of 278–9 and context 311–14 decline in 315–16 development of 306–8 diﬀerent forms of 279–81 and distrust 281–5 as endogenous or exogenous 344–5 identiﬁcation-based 280–81, 285, 286, 294–5

368

Index

as measure of social capital 348 moralistic trust 250 and networks 310–11 from non-zero start 285–6 as a reservoir 287–91 social capital enhancement 288–90 social capital reduction 288, 290–91 role-based 292 role in social capital 277–8, 286–7, 308–10 rule-based 292–3 transferability 225 see also distrust; organizational trust; workplace trust and social capital trustworthiness 5, 27, 88, 293, 307, 308, 310 Tsai, W. 207, 277, 341, 345, 346 Tsang, E. 218–19 Tsui, A. 229 Tullock, G. 161 Turner, J.C. 246 Turral, H. 31 Tyler, T.R. 243, 248 UK health sector innovations study 218 Umphress, E.E. 263 Un, C.A. 218 uncertainty and strategic management 222–6 and use of brokers 269 union membership and workplace trust 90–91, 111–12, 113, 114 United States church groups analysis 80 Social Capital Community Benchmark Survey (US Benchmark) 90, 93–5, 107–10, 111, 112, 114, 126–9, 132, 140–41 United States Agency for International Aid (USAID) 31–2 universally conceived social capital 342–3, 344, 346–9 University of Michigan World Values Study 5 Uphoﬀ, N. 25 Ury, W. 294

US Benchmark survey see United States: Social Capital Community Benchmark Survey USAID 31–2 Uzzi, B. 211, 212, 213, 224, 277, 341 value creation 206, 207, 220, 228–9 values 289–90, 291, 295, 305–6 negative values 291 Van de Ven, A.H. 267 van Staveren, I. 276 Varshney, A. 349 Venktraman, N. 207, 217 venture capitalists 269 le vide 39, 64 Virginian tobacco farmers 64 virtual organization, network duality 50, 52–5, 62, 63 evaluation 56–9 people turnover 56–7 premature consensus 57–9 Voss, T. 164 Vote for America 195 voter contagion 198–200 voter turnout, social norms study 160 adolescence and adult voting link 177–8 and electoral competition 161–3, 167–9 homo civicus 163–7 homo politicus 161–3 school civics instruction, eﬀect of 172–3 school experience in adolescence, eﬀect of 173–7 and social networks 194–6 voters’ motivations 169–71 vulnerability 307, 320, 324 Wacquant, L.J.D. 3, 341 Walker, G. 210, 277 Walsh, J.P. 206 Wasserman, S. 213 Watts, D.J. 265 Weisberg, H.F. 327, 330 well-being and workplace trust 112–13 Wellman, B. 67 Welsh, Jack 39

Index Why We Vote: How Schools and Communities Shape Our Civic Life (Campbell) 172 Williamson, O. 351 Wilson, J. 188 Winpenny, J.T. 31 Wolﬁnger, R.E. 177 Woolcock, M. 3, 32, 208, 339, 340, 343, 349 Worchel, P. 278 workplace trust and social capital 87–8 diﬀerences in trust, explanations for 113–14, 116–29 ethnic status 112, 113 immigration status 112, 113 improvement, strategies for 130 income 92, 93, 94–110, 111, 112 life satisfaction 91–3, 94–5, 111, 112, 113, 114, 130 sample 90 surveys used 90

369

trust among colleagues 90, 111, 114 trust in management 90–91, 111, 113 underinvestment in 114–15, 130 union membership 90–91, 111–12, 113, 114 values of evidence of 90–112 methodology 88–90 well-being 89, 111, 112–13 World Bank 11, 28, 31, 210, 347, 348 Yamagishi, K. 252 Yamagishi, M. 252 Yli-Renko, H. 218, 220 Youth Parent Socialization Study (YPSS) 174–5 Zaheer, A. 210, 260 Zander, U.B. 351 Ziao, Z. 229 Zuckerman, E.W. 341