History, Methodology and Identity for a 21st Century Social Economics (Routledge Advances in Social Economics) [1 ed.] 0367728680, 9780367728687

This book seeks to advance social economic analysis, economic methodology, and the history of economic thought in the co

139 28 3MB

English Pages 154 [165] Year 2020

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

History, Methodology and Identity for a 21st Century Social Economics (Routledge Advances in Social Economics) [1 ed.]
 0367728680, 9780367728687

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

History, Methodology and Identity for a 21st Century Social Economics

This book seeks to advance social economic analysis, economic methodology, and the history of economic thought in the context of twenty-­first-­century scholarship and socio-­economic concerns. Bringing together carefully selected chapters by leading scholars it examines the central contributions that John Davis has made to various areas of scholarship. In recent decades, criticisms of mainstream economics have rekindled interest in a number of areas of scholarly inquiry that were frequently ignored by mainstream economic theory and practice during the second half of the twentieth century, including social economics, economic methodology and history of economic thought. This book contributes to a growing literature on the revival of these areas of scholarship and highlights the pivotal role that John Davis’s work has played in the ongoing revival. Together, the international panel of contributors show how Davis’s insights in complexity theory, identity, and stratification are key to understanding a reconfigured economic methodology. They also reveal that Davis’s willingness to draw from multiple academic disciplines gives us a platform for interrogating mainstream economics and provides the basis for a humane yet scientific alternative. This unique volume will be essential reading for advanced students and researchers across social economics, history of economic thought, economic methodology, political economy and philosophy of social science. Wilfred Dolfsma is Professor and Chair of Business Management and Organisation at Wageningen University, The Netherlands, and Professor of Innovation and Strategy at the University of Curacao. He has published over twenty books and over one hundred articles in international academic journals. D. Wade Hands is Distinguished Professor of Economics at the University of Puget Sound, USA, and has taught history of economic thought for over thirty years. He has written on a wide range of topics in the history of economic thought and economic methodology. Robert McMaster is Professor of Political Economy in the Adam Smith Business School at the University of Glasgow, UK. He has published numerous academic articles and was a co-­editor of the Review of Social Economy from 2005 to 2017.

Routledge Advances in Social Economics Series Editor: John B. Davis Marquette University

This series presents new advances and developments in social economics thinking on a variety of subjects that concern the link between social values and economics. Need, justice and equity, gender, cooperation, work poverty, the environment, class, institutions, public policy, and methodology are some of the most important themes. Among the orientations of the authors are social economist, institutionalist, humanist, solidarist, cooperatist, radical and Marxist, feminist, post-­Keynesian, behaviouralist, and environmentalist. The series offers new contributions from today’s most foremost thinkers on the social character of the economy. Publishes in conjunction with the Association of Social Economics. Economics as Social Science Economics Imperialism and the Challenge of Interdisciplinary Roberto Marchionatti and Mario Cedrini Health Care Economics John B. Davis and Robert McMaster Economics and Other Disciplines Assessing New Economic Currents Ricardo Crespo Welfare Economics An Interpretive History Roger McCain History, Methodology and Identity for a 21st Century Social Economics Edited by Wilfred Dolfsma, D.Wade Hands and Robert McMaster For more information about this series, please visit: www.routledge.com/­ Routledge-­Advances-­in-­Social-­Economics/­book-­series/­SE0071

History, Methodology and Identity for a 21st Century Social Economics Edited by Wilfred Dolfsma, D. Wade Hands and Robert McMaster

First published 2020 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 52 Vanderbilt Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2020 selection and editorial matter, Wilfred Dolfsma, D. Wade Hands and Robert McMaster; individual chapters, the contributors The right of Wilfred Dolfsma, D. Wade Hands and Robert McMaster to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-­in-­Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-­in-­Publication Data Names: Dolfsma, Wilfred, editor. | Hands, D. Wade, editor. | McMaster, Robert (Political economist) editor. Title: History, methodology and identity for a 21st century social economics / edited by Wilfred Dolfsma, D. Wade Hands and Robert McMaster. Description: Abingdon, Oxon ; NewYork, NY : Routledge, 2019. | Series: Routledge advances in social economics | Includes bibliographical references and index. Subjects: LCSH: Economics—Philosophy. | Economics—History—21st century. | Economic development—Social aspects. | Davis, John B. Classification: LCC HB72 (ebook) | LCC HB72 .H58 2019 (print) | DDC 330.01—dc23 LC record available at https://lccn.loc.gov/2019008077 ISBN: 978-­0-­367-­11106-­9 (hbk) ISBN: 978-­0-­429-­20074-­8 (ebk) Typeset in Bembo by Apex CoVantage, LLC

Contents

List of figuresvii List of tablesviii List of contributorsix 1 Economics for the twenty-­first century: a celebration of John B. Davis’s contribution to economics

1

WILFRED DOLFSMA, D. WADE HANDS, AND ROBERT M C MASTER

PART I

History of economic thought and economic philosophy15 2 “Pets and favorites”: Keynes’s practice as investor in the stock exchange

17

MARIA CRISTINA MARCUZZO

3 Simulation and economic methodology

41

MARCEL BOUMANS

4 Identity: individual and social, and a theory of a ‘polymorphic’ but unique identity

51

RICARDO F. CRESPO

5 Reflections on reflexivity and complexity

67

J. BARKLEY ROSSER

PART II

Seeking justice: the promotion of care and capabilities87 6 Understanding the subaltern native middle class WILLIAM DARITY

89

vi  Contents

7 Max U, morality, and the future of economics

99

GEOFFREY M. HODGSON

8 The meso level in economics

124

IRENE VAN STAVEREN

9 Towards a caring economy

137

EVELYN L. FORGET

Index149

Figures

2.1 The evolution of Keynes’s sterling security portfolio in value (£), 1922–1945 2.2 Holdings by sectors, 1933–1939, in USD at 31 December of each year 5.1 Drawing Hands, Maurits Cornelis Escher, 1948 5.2 Las Meňinas, Diego Velásquez da Silva, 1656 5.3 Print Gallery, Maurits Cornelis Escher, 1956 7.1 Frequencies of key terms in leading economics journals 7.2 The relative decline of ‘morality’ in leading economics journals

21 30 69 72 78 108 116

Tables

2.1 Total number of companies traded by Keynes in sterling, 1911–194521 2.2 Weight by sector in Keynes’s sterling securities portfolio, 1933–194523 2.3 List of “pets” in sterling, 1911–1945 24 2.4 Weight of pets (%) on JMK sterling securities end-­of-­year 26 portfolio, 1933–1945 2.5 Weight of sectors (%) in Keynes’s US portfolio, 1933–1939 30 2.6 List of “pets” in USD, 1932–1945 31 2.7 Weight of “pets” (%) in Keynes’s US portfolio, 1933–1939 31

Contributors

J. Barkley Rosser, James Madison University, USA Marcel Boumans, Utrecht University School of Economics, The Netherlands Ricardo F. Crespo, IAE Business School, Universidad Austral, Argentina William Darity, Duke University, USA Wilfred Dolfsma, Wageningen University, The Netherlands, and University of Curaçao Evelyn L. Forget, University of Manitoba, Canada D. Wade Hands, University of Puget Sound, USA Geoffrey M. Hodgson, Loughborough University London, UK Maria Cristina Marcuzzo, Sapienza, Università di Roma, Italy Robert McMaster, University of Glasgow, UK Irene van Staveren, International Institute of Social Studies of Erasmus University Rotterdam, The Netherlands

1 Economics for the twenty-­first century A celebration of John B. Davis’s contribution to economics Wilfred Dolfsma, D.Wade Hands, and Robert McMaster Introduction John Bryan Davis has extraordinary credentials in economic thought and in promoting the idea of an economy that recognises individual dignity and flourishing. He has continued to contribute to a wide range of distinctive and complementary fields and topics, including economic theory, history of economic thought, economic methodology, ethics and economics, welfare economics and normative economics more generally, social economics, health­ economics, various types of heterodox economics, a wide range of policy studies, philosophy of mind and cognitive science, the theory of individual identity in economics, economic history, and reflexivity and complexity in economics. By any measure, this represents a substantial and sustained contribution to our understanding of complex socio-­economic phenomena and how we should study them. In this volume, the editors and contributors endeavour to critically examine key aspects of Davis’s thinking and approach to economic and social issues. For the purposes of this volume, the editors thought it would be useful to divide this vast array of research into a smaller and more manageable number of distinct “areas”. Many divisions were considered, and while none seemed to be perfect, we ultimately ended up with two broad categories: (1) history of economic thought and economic philosophy and (2) seeking justice: the promotion of care and capabilities, especially in the context of social economics. Both are themes that Davis has pursued for decades. The chapters generally reflect this distinction: Chapters 2 through 5 are primarily concerned with topics in the history of economic thought and economic philosophy, while Chapters 6 through 9 are primarily concerned with distinctive aspects of Davis’s investigation in the broad area of social economics. This introduction also reflects this separation, the first part discussing Davis’s work in history of economic thought and economic philosophy and the second part discussing his work in social economics and related topics. Of course, given the wide range of Davis’s contributions as well as his tendency to think holistically and in terms of complexity and co-­determination, spillovers between these two categories appear throughout

2  Wilfred Dolfsma et al.

(both in the editors’ introduction and in the chapters themselves). Nonetheless, the basic division seems to be useful in apprehending the nature of Davis’s extensive research. The next section provides a brief biographical outline of Davis’s career to date, listing some of his achievements and professional service. We then turn to the history of economic thought and economic philosophy, and outline the chapters in Part I of this book. The final section of the chapter examines aspects of Davis’s approach to “seeking justice: the promotion of care and capabilities” and outlines the chapters in Part II of this book.

John Bryan Davis John Bryan Davis is a scholar of international repute for his cutting-­edge and agenda setting work in the areas of economic methodology and philosophy, history of economic thought, social economics, and health economics. In particular, Davis has made significant contributions to the analysis of developments in recent economics, the philosophical approaches of Keynes and Sraffa, issues of identity in economics, capabilities, ethics and economics, and recently, stratification. John Davis attained PhD degrees in Philosophy at the University of Illinois at Urbana in 1983 and in Economics at Michigan State University in 1985. From 1984 to 1987, he was an assistant professor in Economics at the University of Dallas. In 1987, he commenced his thirty-­year ongoing association with the Department of Economics at Milwaukee’s Marquette University. Davis was promoted to full professor in Economics in 1999, holding this post until 2017, when the university conferred on him an emeritus professorship. From 2002 to 2012, he also held the Chair in History and Methodology of Economics in the Department of Economics at the University of Amsterdam. In 2012, he was appointed emeritus professor at Amsterdam. Over his illustrious career, John Davis has held visiting appointments at a host of European institutions, including Université de Paris I Panthéon-­Sorbonne, University of Cambridge, Université de Reims Champagne-­Ardenne, Erasmus University Rotterdam, Università di Torino, Université de Nice Sophia Antipolis, and École Normale Superieure, Cachan. He was also a British Academy Visiting Professor at the University of Aberdeen. In the United States, Davis was a visiting scholar at Duke University. Since 2002, he has served as a Fellow at the Tinbergen Institute and from 2017 as a Fellow at the Center for Global and Economic Studies at Marquette University. From 1987 until 2005, Davis served as editor of the Review of Social Economy. Since 2005, together with Wade Hands, he has co-­edited the Journal of Economic Methodology. He is also editor of the Routledge Advances in Social Economics series. Under Davis’s editorship, the Review of Social Economy evolved to reflect a broad compass of the relationship between ethics and economics. The Review was established in the 1940s under the auspices of the Catholic Economics

Economics for the twenty-­first century 3

Association (CEA), founded by Jesuits Thomas Devine and Bernard Dempsey. The CEA evolved into the Association for Social Economics (ASE) in 1970. This represented an important acknowledgement of the appeal of analysing the interface of ethics and economics from secular and non-­denominational perspectives. This was certainly a position endorsed and promoted by Davis. During his long-­running guardianship of the Review, the journal’s place as an outlet for heterodox economics was firmly established and its production professionalised. In 2005, Davis and Hands assumed the co-­editorship of the Journal of Economic Methodology from the team of Mark Blaug, Roger Backhouse, Kevin Hoover, and Uskali Mäki. The journal is an important outlet for research in economic methodology and philosophy of economics, and encourages diversity and publishes work from any area of economic inquiry as long as it contains a significant epistemological or methodological component. It also encompasses various subjects from the philosophy of the natural or social sciences as well as areas of philosophy such as ethics, as long as they have a direct bearing on debates within the two primary fields of inquiry. At the time of writing, Davis has published over twenty books and over 280 journal articles, book chapters, and reviews. Arguably, his monographs, The Theory of the Individual in Economics (2003) and Individuals and Identity in Economics (2011), confirmed Davis as one of the foremost scholars in the methodology and philosophy of economics. The European Association for Evolutionary Political Economy (EAEPE) awarded Davis the EAEPE-­Myrdal Prize1 for his 2003 work. His 2011 book has been translated into Chinese. His other books have investigated and developed theoretical, philosophical, and history of economic thought issues, such as Keynes’s Philosophical Development (1994), The Life and Economics of David Ricardo (with John Henderson; 1997), Economic Methodology: Understanding Economics as a Science (with Marcel Boumans; 2010), and Health Care Economics (with Robert McMaster; 2017). Davis also edited or co-­edited a number of handbooks and companions, such as The Handbook of Economic Methodology (with Wade Hands and Uskali Mäki; 1998), The Blackwell Companion to the History of Economic Thought (with James Biddle and Warren Samuels; 2001), and The Elgar Companion to Social Economics (with Wilfred Dolfsma; 2008, second edition 2015). This latter volume received the “Outstanding Academic Title” award in 2009 from the American Library Association. Recently, Davis has been at the forefront of debate on the evolution of mainstream (and heterodox) economics. Indeed, his contribution in this area has stimulated much discussion, as part of this volume demonstrates. He has published extensively on this important issue in Cambridge Journal of Economics, Journal of Institutional Economics, Review of Political Economy, and Revue Éthique et Économique, among others. Davis has served numerous professional associations with distinction. For example, he held several roles in the ASE, including on programme committees, as vice president, and as president in 2008. He was also president of

4  Wilfred Dolfsma et al.

the History of Economics Society in 2002, vice president of the European Society for the History of Economic Thought (2010–2012), and chair of the International Network for Economic Method from 2002 until 2004. He also served on various committees of the International Association for Feminist Economics. In his reflections of his career at Marquette University, Davis (2017a) emphasised the importance of interacting with students, in particular regarding the significance of ethics to the economy and business. Davis’s teaching interests, though, are wide-­ranging. He has contributed to courses in master’s and bachelor’s programmes and at all undergraduate levels. For example, he has taught behavioural development, labour, international trade, classical political economy, complexity, health, mathematical economics, and econometrics in addition to history of economic thought, ethics, methodology, and philosophy of economics. His teaching in many fields is research-­led, in that Davis’s pioneering research work is integral to the development of course curricula. Many students have benefitted from Davis’s diligent approach to teaching. Indeed, in his reflections, Davis (2017a) states: I was fortunate at Marquette to be in a department with many University teaching award winners. The dedication to teaching in the department was inspiring and set an example. . . . Teaching has a special personal value . . . because it is so wonderful to be involved with young people engaged in learning. In attempting to retain and promote important writing skills, Davis innovatively developed an online ethics and economics course. The aim was to offer students opportunities to develop communication skills in a way not offered by other courses. This, Davis observed, enabled him to “get back to engaging more directly with students on their work”. Davis also taught at the University of Amsterdam and at several summer schools provided by the ASE and the Portuguese Political Economy Association. Davis has supervised and served on examining committees of well over a dozen PhD students from Europe (including Turkey) and North and South America. The range of subjects these students have studied reflects Davis’s expertise in a host of fields. Some of his previous students have contributed to this volume. For us, this is recognition of the high regard and deep affection in which John is held by his students, colleagues, and the wider academic community.

History of economic thought and economic philosophy John Davis has exceptional credentials in both history of economic thought and economic philosophy.2 As noted in the biographical outline, he holds PhD degrees in both Economics and Philosophy, and has served as editor of both Review of Social Economy and Journal of Economic Methodology and as the head

Economics for the twenty-­first century 5

officer of key organisations in both fields as well as talking on many other responsibilities in these and related organisations. Given both the volume and the diversity of his contributions – and the fact that his interests and concerns have evolved over time – trying to briefly summarise his position on, or his approach to, the history of economic thought and economic philosophy seems to be less informative than briefly tracing his work in these fields historically from early in his career until his most recent work. Davis has always been concerned with path-­dependency and it is useful to view his own intellectual trajectory in that way as well. Although he was always writing extensively on other topics, his work in these two fields during the 1980s focused primarily on various topics in the history of economic thought – specific issues in the work of economists such as Keynes and Ricardo – often in comparison with neoclassical theory. This said, philosophical ideas – involving both epistemology and ethics – were often an important part of these historical inquiries. The 1990s were an extraordinarily productive period in which his research related to these two areas broadened out substantially. He expanded his writing on Keynes – including two edited volumes, a single-­authored book, and a large number of papers – and continued to work on Ricardo as well as Marx and Sraffa. Davis also began his work on individual identity in economics – with papers on identity and methodology, identity in game theory, the embedded individual of heterodox economics, and several other topics – while at the same time expanding his work on ethics, value-­ladenness, and normative economics. He also produced a substantial amount of research on economic methodology, including co-­editing the Handbook on Economic Methodology as well as producing a number of papers concerning the relationship between economics and ideas like postmodernism, post-­structuralism, and the rhetoric of science. Although much of his work in history of economic thought and economic philosophy continued into the 2000s, the research on traditional history of economic thought decreased to some extent as he increased his work on identity and economics, with the publication of The Theory of the Individual in Economics in 2003, Individuals and Identity in Economics in 2011, and numerous journal articles and chapters. He also expanded his writing on a number of other topics like health economics, social capital, economic development, and the historiography of economics. Two major research projects that came into full bloom during the mid-­2000s were his work on complexity/­reflexivity in economics and his writings on the “turn” in recent economic thought (the issue of whether mainstream economics has changed significantly from the neoclassical mainstream of the second half of the twentieth century). All in all, Davis has generated an extraordinarily prodigious scholarly output on an extremely wide range of important topics, and this only takes into account his work directly related to the history of economic thought and economic philosophy (the same will be said about social economics and related areas below). With this brief introduction to Davis’s work, we will turn to the four chapters that are most directly concerned with these topics.

6  Wilfred Dolfsma et al.

In Chapter 2 (“ ‘Pets and favorites’: Keynes’s practice as investor in the stock exchange”), Maria Cristina Marcuzzo employs extensive archival material to examine Keynes’s investment philosophy with particular attention to how his investment strategies related to his views on epistemology, probability, and rational decision-­making. Although it is clear that Keynes changed his baseline asset allocation over time – from a more speculative focus in the 1920s to a more long-­run investment focus later – Marcuzzo argues that this was because of changes in the economy and information rather than changes in his overall investment philosophy. In particular, she discusses four key features of Keynes’s philosophy of investment that remained remarkably consistent throughout his lifetime: its flexibility, the role of informed knowledge, the recognition of the ultimate limitations of knowledge, and the difference between the mentality of investment and speculation. Although Keynes did not ignore the conventions of market sentiment, they never dominated his judgment or kept him from staying focused on long-­run investment strategies. Marcuzzo concludes the chapter by pointing out how several aspects of Keynes’s investment philosophy connect up nicely with Davis’s research on Keynes: his early work on Keynes’s conception of conventions (Davis 1995) as well as his more recent work on Keynes and reflexivity (Davis 2017b). Marcel Boumans’s “Simulation and economic methodology” (Chapter 3) examines the impact that computer simulation has had on economic methodology. He contends that the computer is having a fundamental, qualitative impact on the practice of economic science. Computers are not simply faster and more convenient computational devices for traditional testing of deductive economic theories; they have transformed scientific practice and now provide simulations, which are virtual, or computational, experiments. This is a change in practice, he argues, that will in turn change (and ought to change) economic methodology – in particular, how methodologists and philosophers of economics think about scientific explanations in economics. Boumans argues that simulations have changed the character of explanations, from traditional covering-­law explanations providing answers to why questions to simulation-­based explanations providing answers to what-­if questions (also called how-­possibly or generative explanations). Covering-­law explanations are applications of universal scientific laws (one-­real-­world explanations), while simulations provide explanations in terms of many artificial worlds, each of which resembles the target in different ways. He argues that the validation process for such simulation-­based models – often called calibration in economics – is actually a version of a Turing test. One important feature of such simulations is that they can be valid and yet produce genuine novelty or surprise, which is not the case for traditional deductive models. Finally, these computer-­driven changes will have an effect on how “proofs” are constructed and understood in economics; instead of existence proofs (often proofs by contradiction), proofs in economic models will come from the actual construction (growing, building, generating) of a particular example of the object whose existence is in question. As Boumans emphasises, “methodology should reflect on and assess

Economics for the twenty-­first century 7

real practices of research, not ideal ones” and thus “if simulation is becoming a dominant research practice in economics, economic methodologists should pay attention to this”. Chapter 4 (“Identity: individual and social, and a theory of a ‘polymorphic’ but unique identity”), by Ricardo F. Crespo, explores the topic of individual identity. The chapter is in three parts. The first part examines Davis’s extensive work on individual and social identity presented in two books (Davis 2003, 2011) as well as numerous papers. The second part examines some other contemporary work on individual identity, including Akerlof and Kranton (2000, 2002, 2005, 2010), Kirman and Teschl (2004, 2006), and Sen (1999). Crespo is generally very supportive of Davis’s arguments about identity – his individuation and re-­identification criteria, concept of self-­narratives, characterisation of the relationship between identity and reflexivity, and critique of the neoclassical conception of the individual – and argues that Davis’s view is generally much richer than the other conceptions of identity within the contemporary economic literature. This said, Crespo suggests that Davis’s characterisation is not as clear as it should be about the deeper ontological conception of individual identity – “who is the individual that self-­narrates or reflects?” – and how it stabilises the relationship between individual and social identity. Providing such a deeper conception, grounded in Aristotelian ideas, is the subject of the third part of the chapter. Crespo builds his account on Aristotle’s hylemorphic theory of “explaining the constitution of natural things from ‘primitive’ principles that ‘disappear’ in the compound”. When applied to human beings, the soul is “the first actuality of a natural organic body”. He then combines this Aristotelian conception with John Finnis’s (2005) Aquinas-­inspired notion of individual identity to construct an account of individual identity that Crespo argues provides the appropriate ontological grounding and the proper relationship between individual and social identity, while still building on many of Davis’s insights. It is an identity that can “adequately carry out the decisions and actions of the economic agent, with all their characteristics”. J. Barkley Rosser’s “Reflections on reflexivity and complexity” (Chapter 5) focuses on topics that Davis has frequently discussed in recent work: reflexivity and complexity. The main purpose of the chapter is to organise some of the many different conceptions of reflexivity and complexity within the economic (and wider) literature in an effort to better understand how these two broad concepts are related. The literature on both topics is vast, expanding, and exhibits a breathtaking amount of diversity. These two sets of ideas emerge not only in economics and other sciences that formally model dynamic systems, but also in the computer science literature, in more traditional logical concerns such as Gödel’s incompleteness theorem; various ideas from sociology, political science, and social psychology; ideas about self-­fulfilling and self-­frustrating prophecies, cultural studies, art, and a host of other (interacting) fields of inquiry and contexts. Rosser begins with reflexivity, and in particular reflexivity before its recent discussion within economics. Here he employs the six broad types of reflexivity identified in Lynch (2000). Moving on to discussions of reflexivity

8  Wilfred Dolfsma et al.

in economics – particularly Davis and Klaes (2003) – Rosser examines the three types of reflexivity and the economic case studies Davis and Klaes discuss, as well as their use of particular artistic examples of the painter-­painting relationship. Rosser is ultimately concerned with the relationship between reflexivity and complexity – and in particular versions of reflexive complexity such as that of George Soros and Davis – but in the process he provides a useful taxonomy of complexity concepts based on three core notions of complexity: computational, dynamic, and hierarchical. He also provides an extended discussion of M. C. Escher’s print, Drawing Hands, in comparison with another Escher work, Print Gallery. Although Drawing Hands is frequently cited as the exemplar image conveying reflexivity, Rosser argues that Print Gallery actually captures reflexive complexity – particularly self-­referencing aspect – more “profoundly and effectively” than Drawing Hands.

Seeking justice: the promotion of care and capabilities John Davis’s roots in political economy and social economics suggest that as with other social sciences and science generally, economics is unavoidably value-­laden. Davis’s work has long reflected that scientific explanations are underpinned by the various values that people hold. Importantly, Davis’s approach is central to the social economics’ project of re-­integrating (and renewing) ethics into economic thinking: something that was lost in the transformation of political economy into economics. In adopting this perspective, Davis’s analysis in this area is informed by at least two major insights: 1 Davis refers to a distinction between methodological and substantive assumptions, which reflect his core-­periphery understanding of economic thought. Methodological assumptions refer to how to do economics, given one’s epistemological aims. Substantive assumptions are those fundamental commitments about the purposes of economic science, economic behaviour and institutions, and the economy in question. For Davis, these are inevitably conditioned by ideological commitments. In an interview with Jamie Morgan, Davis (2018) refers to a cleavage between mainstream and heterodox economics as revolving around the nature of relationships between agents. Mainstream economics’ position is that these may be explained in terms of economic, principally market relations. By contrast, heterodox economics subscribes to the belief that economic relations are explainable in terms of other social relationships. In effect, the economy is embedded in society. 2 Following this, Davis identifies dignity as an important emergent theme. In doing so, he draws on a Kantian understanding of the term: by virtue of their humanity, individuals possess intrinsic value or are ends in themselves. Individuals are thus worthy of respect, and accordingly there is a duty on agents and institutions to actively avoid the potential for humiliation. Davis’s position is a strong statement that precludes the possibility of

Economics for the twenty-­first century 9

an instrumental view of an individual, that is, as a means to an end (e.g., Davis and McMaster 2017). Describing dignity in this way represents an important departure from a consequentialist-­utilitarian line of reasoning in which an individual’s welfare may be of secondary concern to the maximisation of benefit for the greater number of people. These insights provide Davis with the platform to challenge the moral basis of the conception of the individual in mainstream economics. His pioneering work (2003, 2011) argues that the standard approach in economics invokes an atomistic model of the individual, one where the individual can be identified and characterised independently of the various social interactions in which the individual is engaged. By contrast, he promotes a socially embedded conception of the individual associated with institutionalism and other heterodox traditions, which contends that individuals are partly constituted, constrained, and enabled by their institutional contexts. One of Davis’s primary concerns is the notion of collective intentionality found in the distinction between “I” and “we” language. For Davis, “we” language is morally more demanding than “I” terminology. Shared intentions provide the basis for reciprocal obligations in a way that individual intentions do not. Hence, the analytical focus of economics should be relational. From this, Davis concludes that standard economics provides an inadequate account of the individual, and therefore he rejects Homo economicus. Again, in echoes of the institutionalist tradition, for Davis this conception implies that the individual is a prisoner of exogenously determined preferences through the dictum of utility maximisation and is accordingly lacking in autonomy or true agency. This is an inappropriate basis for either explaining social behaviour or for the promotion of the dignity of the individual. This argument enables Davis to develop Kenneth Boulding’s (1969) famous criticism of the economic mainstream’s embrace of the Pareto criterion as a guide of well-­being. Pareto represents a weak basis for the assessment of well-­ being that privileges the status quo. Moreover, later tests of well-­being, such as the Kaldor-­Hicks efficiency and Slutsky compensation criteria, do not escape the strictures of this version of welfarism. The standard approach evolved from hedonistic utilitarianism and conceives of well-­being solely in terms of individual preference satisfaction. This is not compatible with the rejection of an instrumentalist view of the individual (e.g., Boulding 1966; Sen 1987, 1993). Instead, Davis advocates a version of the capabilities approach associated with the work of Martha Nussbaum (2011) and Amartya Sen (2009). Capabilities draws from the Aristotelian notion of human flourishing and the ability to lead a fulfilling and meaningful life. Importantly, emphasising the relational dimension central to Davis’s approach, individuals’ capabilities are exercised in social settings in interaction with others. The capabilities literature identifies what an individual is able to do, or be, and “functionings”, actual being and doing (Davis and McMaster 2017). Functionings refer to various types of attainment. In different ways,

10  Wilfred Dolfsma et al.

Nussbaum (2011) and Sen (2009) seek to develop the capabilities framework through the reference of basic capabilities, such as education and health. Davis’s recent work in the area (e.g., Davis 2015a, 2015b, 2018; Davis and McMaster 2017) draws on this literature in examining the stratification of society and the provision of health care. Stratification economics is associated with the work of William Darity (2005). It assumes that the economy is embedded in society. Thus, the economy not only responds to market incentives but more importantly to social values. Given this, the stratification approach presents a divergent conception of equilibrium from standard economics. Instead, equilibrium is the persistence of dominant power structures between groups. In the event of some exogenous shock, the existing arrangement of social relations is likely to respond in such a way that distributes the impact of the shock in an asymmetric manner, such that the privileged are better protected. In this way, equilibrium is infused with ethical values that legitimise the persistence of power structures ensuring social advantage and disadvantage between groups (Davis 2015b, 2018). Davis views this in terms of the limitation of capabilities attainment of those individuals located in certain groups. Similarly, capability attainment may be restricted by a failure of care. Care has a fluid definition; nonetheless, scholars emphasise that it is essential to the functioning of society and our being-­in-­the-­world; it is truly a universal experience (e.g., Engster 2005; Tronto 2013). Care and caring can be manifest as attentive interest, concern, and subsequent activities and actions following such attention. For much of the literature, care is profoundly relational. Care has multiple and complex dimensions. It has psychological properties through emotional attachments and motivations. It possesses labour and work aspects through the functional delivery of care activities. Given its universality, care also invokes ethical importance. Care has at least three basic aims: individual survival, development, and social reproduction (Engster 2005). The first refers to care for the self and others in terms of addressing basic needs, such as secure food and clean water. The second involves enabling others to develop and sustain sufficient capabilities for basic functioning in society. The third is to help individuals avoid pain, or to relieve the suffering of those in pain in order to enhance their capabilities to live life as well as possible (Engster 2005). Fulfilling the aims of care is complex and sensitive to underlying institutional arrangements. Joan Tronto (2013) recognises the possibility of a care deficit: where the needs of care are not addressed by the provision of care (i.e., there are unmet needs). In the context of health care, Davis and McMaster (2017) argue that a care deficit implies a capabilities deficit. An institutional context that privileges the model of the atomistic individual driven by material incentives is not consistent with an array of institutions that facilitate a caring environment fostering the development of capabilities. John Davis’s important work arising from his insights about the economy nested in society, the notion of the socially embedded individual, and the

Economics for the twenty-­first century 11

subsequent emphasis on human dignity lead him to advocate a more just world through the lenses of capabilities and care. In different ways, Chapters 6 through 9 critically examine aspects of Davis’s agenda. In Chapter 6 (“Understanding the subaltern native middle class”), using stratification economics, William Darity explores the African American middle class in the United States. In doing so, Darity notes stratification economics’ affinity to colonial and subaltern studies in the social sciences. He draws on E. Franklin Frazier’s (1957) path-­breaking book, Black Bourgeoisie, and Davis’s (2015b) article on stratification and identity. As with Irene van Staveren’s analysis (Chapter 8), Darity’s focus is the meso level, which for the stratification approach lies in the “tribe”, identifiable by race, gender, class, and so forth. Davis’s contribution to the field emphasises the tension between multiple identities that transcend specific strata, such as gender and race. On this, Darity hypothesises that the greater the perceived distance between groups in society, the greater the importance the “average individual” will give to the relative position within their group. In his study, Darity argues that wealth is the most effective indicator of a group’s “subaltern status” (i.e., lower status and without the dominant power structure of a society). Darity notes that in the United States, African Americans constitute around 13%–14% of the population but possess less than 3% of the country’s wealth. Thus the African-­American middle class is essentially subaltern. This leads Darity to pose several questions about the characteristics of all subaltern middle classes, including: To what extent is this class proportionately smaller and financially less secure than the middle class of the dominant group? To what extent does the subaltern middle class demonstrate what Darity calls the “lower ladder” characteristic – where individuals at the apex of their group’s income distribution, at best, align with the middle of the distribution in the dominant group? New members of the subaltern middle class are more likely to engage in displays of ostentation and status. In the subaltern group, is there a greater downward mobility from the middle class? Other questions posed by Darity relate to the celebration of any accomplishment by the subaltern middle class, differences in health gradients, and complicity with the existing social power structure. Geoffrey M. Hodgson, in “Max U, morality, and the future of economics” (Chapter 7), focuses on theorising the individual in economics. For Hodgson, mainstream economics has been undergoing an impressive level of change since the 1980s with the “rehabilitation” of game theory and from the 1990s onwards with the “new behavioural turn” and experimental economics. Manifestations include applied studies in happiness economics and neuro-­ economics. This change has ushered an increased tendency for the publication of empirical articles in leading journals. Hodgson acknowledges the roles of the likes of David Colander (2005) as well as Davis (2006) in defining the nature of the changes in economics and in seeking to explain them. In doing so, Hodgson insightfully argues for definitional clarity and is critical of the term “neoclassical economics”. He contends that this is frequently wrongly associated with pro-­market policies by critics and proponents alike. Hodgson argues that some leading neoclassical

12  Wilfred Dolfsma et al.

economists, such as Oskar Lange and Kenneth Arrow, were in no way inclined to laissez-­ faire economics. Indeed, Leon Walras – the founder of general ­equilibrium – considered himself a “scientific socialist”. Instead, Hodgson advocates focusing on the central defining feature of mainstream economics utility maximisation – drawing from Deirdre McCloskey (2008), Max U. Importantly, Hodgson argues that Max U is unfalsifiable, which renders it vulnerable to Popperian accusations of not being scientific. This is an epistemic weakness. Hodgson explores weaknesses in Max U and poses the question as to whether it will be replaced. He is critical of Davis (and others) for underestimating the persistence of Max U in the behavioural turn and other recent innovations in mainstream thinking. From here, Hodgson notes theoretical limitations in the form of explaining the causes of behaviour, and the development of capacities and dispositions. Capacities, inter alia, can relate to moral motivation. Hodgson refers to a long train of thinking from Adam Smith to Amartya Sen and beyond in his argument. Moral judgements are universal and their manifestation is sensitive to institutional context. Here Davis’s work on the socially embedded individual is valuable in illuminating the centrality of morality. Chapter 8 (“The meso level in economics”), by Irene van Staveren, analyses the layer of reality located between the micro and macro. She refers to this as the level of reality in which individuals’ behaviour is guided. Van Staveren initially explores how various schools of thought conceptualise the meso level. She identifies the fallacy of composition and herding behaviour in Keynes’s thinking of the meso. Institutional economics, for van Staveren, demonstrates the importance of the meso level. Similar to Hodgson’s analysis in Chapter 7, van Staveren observes that the meso level is where institutions emerge as social rule systems. The capability approach is also meso-­level thinking in that institutions are important in providing “capability space”. Van Staveren then focuses on social economics, arguing that it centres on, first, what she terms the “community economy”, or social economy, and second, the value systems underpinning economic activities and institutions, such as the market. In social economics, van Staveren argues, the meso is manifest in social groups, norms, and interdependencies. She highlights Davis’s pioneering contribution in developing the theory of the socially embedded individual, which she considers prevents an overdetermination of agency. Individuals are influenced by society but can also influence society. The role of groups in Davis’s work establishes important links with stratification economics. Moreover, Davis’s analysis of norms resonates with the care literature. In outlining interdependencies, van Staveren refers to the use of the term “social capital” as a means of expressing social relatedness, such as norms and networks. She concludes by arguing that the meso level is the most significant in economic activity, and that social economics in general and Davis’s work in particular provide valuable insights. Van Staveren’s study shows the complementarities across social and institutional economics, and elements of Keynes’s approach. In Chapter 9 (“Towards a caring economy”), Evelyn L. Forget critically analyses Davis’s contribution to the conception of care in the context of

Economics for the twenty-­first century 13

health economics, in particular. Forget concentrates on Davis and McMaster’s (2017) Health Care Economics. Forget notes the profound difficulties in defining care in general. In the context of health care provision, she alludes to the deontological dimension of caregiving in the medical situation. In other words, caring is associated with social roles. On this, she emphasises two aspects of caregiving: the process of giving and the outcomes of caring. Mainstream health economics is inherently consequentialist and is therefore at risk of neglecting the process of caring. Indeed, Davis and McMaster argue that the standard approach to care as an externality renders it sensitive to issues of crowding out as self-­interested material incentives dominate other-­regarding relational aspects. While agreeing with the tenor of this argument, Forget suggests that Davis and McMaster’s critique is questionable on the grounds of potential trade-­offs in medicine and the possibility of conceptualising process utility distinctly from outcome utility. In effect, Davis and McMaster may present an exaggerated critique of the standard approach. Forget maintains that it is possible to measure process utility, and therefore establish its distinctness from consequentialism, yet also a bridge between process and outcome. Finally, Forget endorses Davis and McMaster’s argument concerning the tendency to relegate the importance of care, specifically in health care. She concludes that, overall, caring is a very human activity that is not readily open to machine delivery.

Notes 1 The EAEPE-­Myrdal Prize is awarded to the best published monograph “on a theme broadly in accord with EAEPE theoretical perspectives” (EAEPE, 2019: https://­eaepe. org/­?page=awards). 2 The term “economic philosophy” seems appropriate here, since the terms “economic methodology” and “philosophy of economics” seem to have multiple, and contested, definitions. Economic philosophy is much broader, subsuming all of the ways that economics and philosophy can come together: in the methodological discussions of economists, in the application of philosophy of natural science to economics, in the relationship between economics and other areas of science theory, and in the various connections between economics and other areas of philosophy such as ethics.

References Akerlof, G. and Kranton, R. (2000). Economics and identity, Quarterly Journal of Economics, 115(3): 715–753. Akerlof, G. and Kranton, R. (2002). Identity and schooling: Some lessons for the economics of education, Journal of Economic Literature, 40: 1167–1201. Akerlof, G. and Kranton, R. (2005). Identity and the economics of organizations, Journal of Economic Perspectives, 19(1): 9–32. Akerlof, G. and Kranton, R. (2010). Identity Economics: How Our Identities Shape Our Work, Wages, and Well-­Being. Princeton, NJ: Princeton University Press. Boulding, K. E. (1966). The concept of need in health services, Milbank Memorial Fund Quarterly, 44: 202–223. Boulding, K. E. (1969). Economics as a moral science, American Economic Review, 59: 1–12.

14  Wilfred Dolfsma et al. Colander, D. C. (2005). The future of economics: The appropriately educated in pursuit of the knowable, Cambridge Journal of Economics, 29: 927–941. Darity, W. Jnr. (2005). Stratification economics: The role of intergroup inequality, Journal of Economics and Finance, 29: 144–53. Davis, J. B. (1995). Keynes’s Philosophical Development. Cambridge: Cambridge University Press. Davis, J. B. (2003). The Theory of the Individual in Economics. London: Routledge. Davis, J. B. (2006). The turn in economics: Neoclassical dominance to mainstream pluralism? Journal of Institutional Economics, 2: 1–20. Davis, J. B. (2011). Individuals and Identity in Economics. Cambridge: Cambridge University Press. Davis, J. B. (2015a). Agency and the process aspect of capability development: Individual capabilities, collective intentions, Filosofia de la Economia, 4: 5–24. Davis, J. B. (2015b). Stratification economics and identity economics, Cambridge Journal of Economics, 39: 1215–1229. Davis, J. B. (2017a). Dr. John Davis reflects on his years at Marquette, Marquette Economics Newsletter, Fall. Davis, J. B. (2017b). The continuing relevance of Keynes’s philosophical thinking: Reflexivity, complexity, and uncertainty, Annals of the Fondazione Luigi Einaudi, 60: 5–17. Davis, J. B. (2018). Stratification as an economics of exclusion. Paper presented at the ASE World Congress, Colorado State University, June. Davis, J. B. and Klaes, M. (2003). Reflexivity: Curse or cure? Journal of Economic Methodology, 10: 329–352. Davis, J. B. and McMaster, R. (2017). Health Care Economics, Routledge: London. Engster, D. (2005). Rethinking care theory: The practice of caring and the obligation to care, Hypatia, 20: 50–74. Finnis, J. M. (2005). The thing I am: Personal identity in Aquinas and Shakespeare, Social Philosophy and Policy, 22: 250–282. Kirman, A. and Teschl, M. (2004). On the emergence of economic identity, Revue de philosophie économique, 9(1): 59–86. Kirman, A. and Teschl, M. (2006). Searching for identity in the capability space, Journal of Economic Methodology, 13: 299–325. Lynch, M. (2000). Against reflexivity as an academic virtue and source of privileged knowledge, Theory, Culture, and Society, 17: 26–54. McCloskey, D. N. (2008). Mr. Max and the substantial errors of manly economics, Economic Journal Watch, 5: 199–203. Nussbaum, M. (2011). Creating Capabilities: The Human Development Project. Cambridge, MA: Belknap Press. Sen, A. (1987). On Ethics and Economics. Blackwell: Oxford. Sen, A. (1993). Capability and wellbeing, in The Quality of Life, Nussbaum, M. and Sen, A. (eds), Oxford University Press: Oxford. Sen, A. (1999). Reason before Identity. New Delhi: Oxford University Press. Sen, A. (2009). The Idea of Justice. Penguin: London. Tronto, J. C. (2013). Caring Democracy: Markets, Equality, and Justice. New York: New York University Press.

Part I

History of economic thought and economic philosophy

2 “Pets and favorites” Keynes’s practice as investor in the stock exchange Maria Cristina Marcuzzo

Premise The aim of this chapter is to connect Keynes’s investment behavior in the stock exchange with his general views of choice in a context of uncertainty, characterized by limited knowledge, where his particular notions of probability and rationality apply. The work done by John Davis, among others, on Keynes’s general philosophy provides elements clarifying how Keynes’s epistemology is a key factor in the two-­way relationship between Keynes’s theoretical stances and his practical acquaintance with financial markets. The chapter begins with an overview of the sources and the literature on Keynes as investor and then reviews Keynes’s ideas on speculation as they evolved from the early Lectures on the Stock Exchange to the General Theory. The following sections examine Keynes’s investment choices in the London and New York Stock Exchanges during the years in which he was most active, focusing on the small set of shares he was keen on that had the largest share in his portfolio and to which he used to refer to his “pets” (see for instance Keynes 1971–1989, CWK XII: 78). I then summarize the key points in Keynes’s investment philosophy as revealed by his letters and memoranda, and I argue that Keynes’s behavior as an investor is to be understood in the context of his theory of decision-­making and choice, grounded on his notions of uncertainty, probability and conventions of which John Davis has put forward an authoritative interpretation.

Keynes as investor There is general agreement in the recent literature1 that the turning point in Keynes’s stock market investment policy came in the early 1930s, when Keynes had decidedly shifted from a credit-­cycle approach to a strategy of limited diversification and a highly idiosyncratic selection of a restricted set of shares. He summarized his views to the chairman of the Provincial Insurance Company,2 F. C. Scott, on 15 August 1934: As time goes on, I get more and more convinced that the right method in investment is to put fairly large sums into enterprises which one thinks one

18  Maria Cristina Marcuzzo

knows something about and in the management of which one thoroughly believes. (Keynes 1971–1989, CWK XII: 57) And a few years later, he made clearer how many companies he knew “something about” and how many there were in whose management he “thoroughly believed”: I myself follow very closely, or think I have some knowledge, of upwards of perhaps 200 investments[. . .] . Now out of the 200, which one tries to follow more or less, there are probably less than 50 in all classes about which, at any given time, one feels really enthusiastic. (Memorandum for the Provincial Insurance Company, 7 March 1938, in Keynes 1971–1989, CWK XII: 99) Keynes had an intense career as an institutional investor. He was also director of the National Mutual Life Insurance Company in 1919, and then chairman in 1921 until 1938; director of the Independent Investment Company (1923–1946), the A.D. Investment Trust (1921–1927) and the P.R. Finance Company (1924–1936, Chairman 1932–1936); and director and then chairman of the Financial Committee of the Provincial Insurance Company from 1923, remaining associated with the company in other capacities until his death. In 1921, Keynes became Second Bursar of King’s College, Cambridge, and then First Bursar in 1924 until the end of his life. To add to this experience as an institutional investor, there is of course his intense activity on his own account, which was a major engagement in Keynes’s life both in terms of duration and commitment, which he pursued among pressing public and academic commitments from his youth until his death. Keynes’s portfolio choices, both personal and institutional, were the result of the juxtaposition of his general vision and systematic analysis of the entire economic system with the information he could actually obtain on sectors and individual shares. He relied on a huge amount of detailed information on a selection of shares from a number of sources, providing him with professional (as distinguished from confidential) information and business analysis. The Keynes Papers3 reveal hundreds of documents showing Keynes’s need and passion to keep up to date with the markets by constantly gathering evidence. They also show that for Keynes, investment was a time-­consuming activity in which building up a network of reliable connections and collecting sound, relevant information was a costly but decisive task. While undoubtedly the evolution of Keynes’s economic thinking was influenced by the development of his business skills, achieved through an increase in his business contacts and sources of information, the direction of influence ran also in the opposite direction: his knowledge and method as an economist guided the selection of the sources of information and their assessment.

“Pets and favorites” 19

Speculation versus investment4 Before analyzing Keynes’s strategies as an actor in financial markets, it may be useful to review his ideas on speculation and investment. In his Lectures on the Stock Exchange (1910), when he did not have any practical experience as an investor, Keynes distinguished between gambling and speculation, applying the former term to situations in which risk is or is not calculable. The dividing criterion is in the amount of knowledge possessed by the actor in both cases: “the possession of superior knowledge [is] the vital distinction between the speculator and the gambler” (Keynes 1910: 98). Superior knowledge confers the speculator with an advantage over the market. To Keynes, this is a matter relevant not to measuring comparative success in gambling and in speculation, which may be dependent on other factors, but to evaluating the nature of the action in the two cases. The interesting point is that for Keynes, unlike speculation, gambling is a behavior that has no basis in knowledge, notwithstanding the fact that a gambler may at times be a winner and a speculator a loser. Does his “superior knowledge” allow the speculator to predict the future course of events? There are passages in the Lectures that seem to confirm it: • “Speculation [is a] reasoned attempt to gauge the future from present known data” (Keynes 1910: 95). • “The speculator [is] a person who endeavours to make a profit by means of a power of forecasting the future superior to the ordinary” (Keynes 1910: 95). • “Speculation consists in the use of superior skill in forecasting changes of value to take advantage of them by buying and selling” (Keynes 1910: 100).5 The next phase in Keynes’s thinking – as he became more closely acquainted with the working of markets – was the analysis of speculation in futures (currencies and commodities) presented in his “The Forward Market in Foreign Exchanges” (1922), incorporated in the Tract of Monetary Reform (Keynes 1971–1989, CWK IV) and in his 1923 article “Some Aspects of Commodity Markets” (Keynes 1971–1989, CWK XII: 255–265). The point of speculator as risk bearer is reiterated in the Treatise on Money, where he gave a more refined version of his theory. Not only is the speculator not a “gambler”, but also his ability, through superior knowledge, to forecast the future is downplayed. He is not “a prophet” (Keynes 1971–1989, CWK XII: 260), but rather a risk bearer: “The most important function of the speculator in the great organized ‘future market’ [is that of] a risk bearer” (Keynes 1971–1989, CWK XII: 260). At this stage of Keynes’s thinking, speculator’s profits are seen as the remuneration for risk bearing, not for forecasting skill. In chapter 12 of the General Theory, the analysis of speculation marks a departure from Keynes’s previous views, as the following quotations show: “the

20  Maria Cristina Marcuzzo

term speculation [is appropriated] for the activity of forecasting the psychology of the market”, and it is distinguished from enterprise, which is defined as the “activity of forecasting the prospective yield of assets over their whole life”. And Keynes adds: “As the organization of investment market improves, the risk of the predominance of speculation [. . .] does increase” (Keynes 1971– 1989, CWK VII: 158). The “beauty contest” (average market opinion) example explains why, according to Keynes, speculation does not promote price stability in those markets. Unlike the efficient market theory, according to which by buying low and selling high speculators push up the low prices and push down the high prices, Keynes points out its possible destabilizing nature. The destabilizing effects of speculation can be described as a sudden and large increase in optimism or pessimism propagated by herd behavior, unrelated to new information about fundamentals but triggered by market opinions. In the General Theory, Keynes made it clear how the energies and skill of the professional investor and speculator are mainly occupied [. . .] not with making superior long-­term forecasts of the probable yield of an investment over its whole life, but with foreseeing changes in the conventional basis of valuation a short time ahead of the general public. (Keynes 1971–1989, CWK VII: 154)6 Keynes, who had quite an experience as a speculator in currency and commodity markets, always collected information relative to each individual market and assets, weighing up the quality and reliability of that information through calculation of the relevant data, the advice of experts, and his own assessment of market conditions and of other participants’ opinions. While the grasp of “business psychology” was an important element both in his investment strategy and in his views on speculation, as he became more and more a long-­term investor he never lost sight of the fundamentals underlying the companies in which he invested. In conclusion, although Keynes’s investment philosophy seems to have changed in the early 1930s, following heavy losses in the commodity market and possibly progress in his new theoretical developments which culminated in the General Theory, the role of informed opinion about the relevant data never gave way to evaluation of market sentiment, conventions, and herd behavior. In the end the ability of the speculator rested, for Keynes, on individual judgment as opposed to the average market view: “My central principle of investment” – he explained in 1944 to a banker who was critical of his suggestions about how to manage Eton’s finances – “is to go contrary to general opinion, on the ground that, if everyone is agreed about its merits, the investment is inevitably too dear and therefore unattractive” (Keynes 1971–1989, CWK XII: 111). Let’s now turn to an overview of his practical experience as an investor.

“Pets and favorites” 21

London Stock Exchange (LSE)7 As can be seen from Figure 2.1, although Keynes had tried his hand in the London Stock Exchange since the early 1920s, he became a “serious” investor in shares and bonds only after 1932, when the LSE rebounded after the 1929 collapse. The number of companies (or local and national government issuers) in which Keynes invested was slightly over two hundred (see Table 2.1), 450,000 400,000 350,000 300,000 250,000 200,000 150,000 100,000

50,000

31/12/1945

31/12/1944

31/12/1943

31/12/1942

31/12/1941

31/12/1940

31/12/1939

31/12/1938

31/12/1937

31/12/1936

31/12/1935

31/12/1934

31/12/1933

31/12/1932

31/12/1931

31/12/1930

31/12/1929

31/12/1928

31/12/1927

31/12/1926

31/12/1925

31/12/1924

31/12/1923

31/12/1922

0

Figure 2.1  The evolution of Keynes’s sterling security portfolio in value (£), 1922–1945 Source: Our calculations on the basis of Keynes’s own evaluations at 31 December of each year, 1922–1945 (KP SE/­11/­4–7).

Table 2.1  Total number of companies traded by Keynes in sterling, 1911–1945 Number of Number of Number of pet Total number Number of pet companies companies of companies companies, 1922– companies appearing in not in the 1922–1945, traded in 1945 (companies his end-­of-­year evaluations only as longest sterling whose shares portfolio from (traded within holding (more 1911–1945 Keynes held for one to three the year in the than four more than four years period 1922– consecutive consecutive years years but below 1945) and accounting 1% of end-­of-­ for at least 1% of year portfolio) his end-­of-­year portfolio in each year) 205 38 12 49 106 Source: Our calculations on the basis of Keynes’s own evaluations, 1922–1945 (KP SE/­11/­4–7), statements of accounts in sterling securities by Buckmaster & Moore, 1920–1945 (KP SE/­8), and Keynes’s ledgers, 1911–1922 (KP SE/­11/­6/­18–30).

22  Maria Cristina Marcuzzo

with a great variety of instruments (regular shares, preferred, debentures, other fixed interest securities), but as far as his “valuations”8 are concerned (see endnote 5) we have data for approximately 130 companies. The 205 companies have been grouped into seven sectors: (1) mining, (2) industrials, (3) transport, (4) finance, (5) soft raw materials, and (6) government and corporate bonds, plus miscellanea and NA, which include utilities and new technology (see Table 2.2). The sectors in which we find his favorites shares are in fact less; if we take as a benchmark those he held for an average period of four years, with a weight in the portfolio of at least 1%, the number is slightly below forty (see Table 2.3). It can be seen that after Keynes became a substantial investor in the London Stock Exchange, on average 80% of his portfolio was made up of his “pets” (see Table 2.4). Let us now look into the characteristics of some individual company shares preferred by Keynes. Tin

Several of the tin companies whose shares Keynes held in his portfolio in the 1930s had Oliver Lyttelton as manager or director. From 1920, Keynes’s friend Oliver Lyttelton had been employed by the British Metal Corporation (BMC), where he was managing director between 1925 and 1939. BMC was a major shareholder of Anglo-­Oriental Mining and British Tin Corporation. He was also chairman of the London Tin Corporation and served on the boards of a number of foreign companies engaged in the metal trade. Turning now to individual shares, we note that Keynes’s largest amounts were invested in three companies – British Tin Investments, Southern Kinta, and Anglo-­Oriental Mining – which, however, he held only for two years. All of them had their main interests in Malaya. In order to understand why Keynes was so keen on investing in British-­managed tin companies, an observation in his memorandum to the Treasury in October 1940 may be of some help. When advising the British government on how to increase Britain’s dollar resources, Keynes declared that tin (and rubber) shares would not hold much appeal for the US private investors, mainly “for the reason that they are out of touch with the managements which are situated in London and in the East, and have no means of gauging their efficiencies” (Keynes 1971–1989, CWK XXIII: 14). Keynes clearly believed that he was able to do so. Gold

In a letter to F. C. Scott dated 21 June 1934, Keynes outlined the key reasons why he liked Union Corporation, the large South African mining company, one of his largest and most successful core holdings: he considered this asset as greatly undervalued on the market at that time and as having “enormous potentialities” (KP: PC/­1/­1/­365).

9.0 33.0 42.0

2.3 0.3 2.3

9.8 0.6 10.4

0

0

1.8

General industries Motor and airplane Industrials (total)

Railways Shipping Transport (total)

Investment trust Bank and insurance Finance (total)

Soft raw materials

Government bonds and stocks

Miscellanea and NA (included utilities and new technology) Total 100

1.3

30.6

1.0

7.0 4.0 11.0

3.5 0.0 3.5

2.9 19.6 22.5

5.3 22.1 2.7 0 30.1

1934

100

2.6

0

0

6.1 6.3 12.4

5.0 7.5 12.5

1.6 15.5 17.1

11.6 36.7 7.1 0 55.4

1935

100

7.4

0

0.9

9.5 5.1 14.6

3.0 5.3 8.3

1.2 7.6 8.8

17.0 31.1 11.9 0 60.0

1936

100

8.0

0

1.4

1.6 0 1.6

0 7.5 7.5

0.4 9.0 9.4

21.2 40.8 10.1 0 72.1

1937

100

7.6

0

1.8

4.0 0 4.0

11.0 7.8 18.8

0.6 7.2 7.8

21.5 27.6 10.1 0.8 60.0

1938

100

1.0

0

3.0

3.0 0 2.9

22.0 6.0 27

2.5 5.9 8.1

20.8 12.3 17.0 8.9 58

1939

100

8.3

1.2

2.2

6.0 0 6.0

25.0 5.7 30.7

9.3 7.9 17.2

16.1 1.9 16.4 0 34.4

1940

100

12.5

1.6

2.5

6.1 5.4 11.5

27.5 2.7 30.2

8.9 7.0 15.9

7.4 1.7 16.7 0 25.8

1941

Source: Our calculations on the basis of Keynes’s own evaluations at 31 December of each year, 1922–1945 (KP SE/­11/­4–7).

100

0 40.6 3.2 0 43.8

1933

Tin Gold Metals Oil Mining (total)

Sectors

Table 2.2  Weight by sector in Keynes’s sterling securities portfolio, 1933–1945

100

9.6

0

3.6

5.9 4.3 10.2

26.0 6.0 32

13.8 6.9 20.7

6.8 1.2 15.9 0 23.9

1942

100

13.1

0

3.0

8.3 3.5 11.8

21.0 4.5 25.5

18.4 9.3 27.7

5.4 5.0 8.5 0 18.9

1943

100

10.3

0

4.9

6.0 3.0 9

17.0 9.7 26.7

16.8 8.4 25.2

2.1 4.8 9.0 8.0 23.9

1944

100

18.5

0

0.5

8.7 2.7 11.4

13.5 15.5 29

15.6 10.3 25.9

0.4 4.9 9.0 0.4 14.7

1945

Table 2.3  List of “pets” in sterling, 1911–1945 MINING Tin Ampat* (1936–1945) British Tin Investment Corporation (1934–1944) Pahang Consolidated (1924–1927) Petaling Tin Mines (1934–1939) Southern Kinta (1934–1943, not 1935) Teja Malaya (1927–1931) Tekka Taiping Ltd. (1923–1926) Tronoh Mines (1934–1937) Gold Emperor Mines* (1936–1945) Gold Exploration & Finance Co. of Australia* (1934–1945) Loloma (Fiji) Gold Mines* (1936–1945) Randfontein Estates Gold Mining Co. (1933–1937) San Francisco Mines of Mexico (1934–1941) South African Torbanite Mining & Refining Co.* (1936–1941) Union Corporation (1933–1945) Metals Horden Collieries Ltd. (coal) (1916–1919, 1923–1936, not 1935) Mineral Separation (technology of ore extraction) (1936–1945) Richard Thomas & Co. (iron and steel) (1936–1945) Trepca Mines (lead and zinc) (1935–1941, not 1937) Wharncliffe Collieries* (coal) (1937–1942) Yorkshire Amalgamated Collieries (lime works and quarries) (1927–1933) Zinc Corporation (zinc) (1937–1945) INDUSTRIALS General industries Enfield Rolling Mills* (electric cables, UK) (1935–1945) Lancashire Cotton Corporation (merger of cotton firms, UK) (1935, 1940–1945) Lever Bros. (soap manufacturer, UK) (1939–1945) Mather & Platt (industrial machinery, UK) (1916–1934, not 1921) Motor and airplane Austin Motor (1927–1945) Bristol Aeroplane Co. Ltd.* (1939–1945) Hawker Siddeley Aircraft Co. Ltd. (1938–1945) Leyland Motors (cars) (1939–1945) Trojan Holdings (1934–1937) New technology P.I.V. Chain Gears* (1925–1932, 1935–1945) TRANSPORT Railways London & North Eastern (L.N.E.R.) (1928,1934–1936, 1938–1944) London Midland & Scottish (L.M.S.) Railway Ordinary (1934–1935, 1938–1945) Southern Railway Co. (1931–1933; 1936, 1939–1945) Shipping Hector Whaling Ltd. (1928–1945) Indo-­China Steamship Navigation Co. Ltd. (1923–1938) Ormes Ltd. Ordinary* (1935–1942) Tankers (1926, 1937–1940) Union Castle Mail Steamship Co. Ltd. Ord. (1937–1945)

“Pets and favorites” 25 MINING FINANCE Investment trusts AD Investment Trust (1921–1926) British Countries Trust* (1936–1943) Independent Investment Co. (1924–1945) London Trust (1911–1915) P.R. Finance Co. (1922–1935) Selection Trust Ltd. (1935–1945) Bank and insurance Provincial Insurance Corporation (1924–1935) SOFT RAW MATERIALS Gopeng Perak Rubber Estates Co. (1923–1926, 1937–1941) Straits Plantations (1936–1945) Sudan Plantations Syndicate Ltd. (1939–1944) *Starred items = below 1% of the portfolio in all the years. Sources: Keynes’s own evaluations, 1922–1945 (KP SE/­11/­4–7), statements of accounts in sterling securities by Buckmaster & Moore 1920–1945 (KP SE/­8), and Keynes’s ledgers, 1911–1922 (KP SE/­11/­6/­18–30).

He reiterated the point on 15 August 1934, adding “management has the highest standards and exercises extreme industry and care” (Keynes 1971– 1989, CWK XII: 56). The managing director was Henry Strakosch, of whom Keynes added: I feel that that part of one’s resources which one is willing to embark in gold and other metals cannot be better handled than by leaving him to manage them for one from his position of exceptional advantage. (Keynes 1971–1989, CWK XII: 56) A year later he wrote: “I reiterate that the concealed strength of Union Corporation is beyond computation” (JMK to F. C. Scott, 14 June 1935, KP: PC/­1/­3/­226). This judgment did not change in the following years, as testified by the following quotation of 1937: if I am wrong about Union Corporation in the long run I resign all claims to knowledgeability in these matters. If one is going simply on ground of investment safety I put that far and away ahead of most of our investment. (JMK to F. C. Scott, 7 June 1937, PC/­1/­4/­306) Once again, trust in the management seemed to have been the key factor in Keynes’s choice, once he also had formed a clear idea on the sector of investment. In the case of gold, obviously the events of 1931 (several countries went off the gold standard) and the Kaffir boom were relevant.9

60.1 249,317

86.5 70,465

82.5 222,575

1935 74.3 411,625

1936 83.8 199,012

1937 95.5 117,236

1938 95.7 139,325

1939 91.8 119,151

1940 89.0 151,799

1941

Source: Our calculations on the basis of Keynes’s own evaluations at 31 December of each year, 1922–1945 (KP SE/­11/­4).

% of pets Total value of end-­of-­year portfolio (£)

1934

1933

Table 2.4  Weight of pets (%) on JMK sterling securities end-­of-­year portfolio, 1933–1945

89.9 189,791

1942

86.4 232,293

1943

78.2 274,274

1944

78.5 329,720

1945

“Pets and favorites” 27 Metals

Keynes had heavily speculated in the non-­ferrous future and option markets, mainly in the 1920s, although he continued more sporadically throughout the 1930s (see Marcuzzo and Sanfilippo 2016), gathering knowledge of the sector. Zinc (spelter) was a commodity he knew very well from his investments in the 1920s and his work on the Memoranda on Stocks of Staple Commodities for the London and Cambridge Economic Service from 1923 to 1930 (Keynes 1971–1989, CWK XII: 267–648). On Zinc Co., Keynes wrote to F. C. Scott: The zinc corporation preference [shares] are certainly very good at present prices, since, as near as I can guess, they are earning 6 per cent at the present very low prices of the metals, prices that for many producers obliterate profits altogether. (JMK to F. C. Scott, 10 May 1938, KP: PC/­1/­5/­77) On Trepca Mines, producing zinc and lead, he wrote: “Lead mines are still, I think, attractive [. . .] I suggest we buy 5,000 Trepca [. . .] I know quite enough about it to be comfortable” (JMK to F. C. Scott, 30 January 1936, KP: PC/­1/­4/­281). In all these cases, Keynes appeared to be quite confident in the amount and quality of information he possessed on these companies, and this confirms that this was a key factor (much more than price) in determining his selection of assets. General industries

In 1920s Keynes often pointed to the difficulties of British industry and the British economy in general, mainly due to the increasing trade and industrial competition by US firms, and started a campaign in favor of a better efficiency to be accomplished through a deep rationalization and restructuring of the sector, especially as far as the management is concerned. His “pet” shares in this sector were either those that were technologically innovative (Enfield Rolling Mills and Mather and Platt) or catering to an expanding consumer market (Lever Bros.). Motorcars and airplanes

As far as the motor industry is concerned, Keynes’s most important “pet” in terms of dimension and longest holding was Austin Motor, which is a telling example of his lookout for “fundamental” rather than short-­term market price oscillations. He wrote to F. C. Scott: In addition to the Patent case there is the question of increased American competition in the export trade due to the depreciation of the dollar[. . .] .

28  Maria Cristina Marcuzzo

As you know, I have always been inclined to ignore market fluctuations of these shares, which are bound to be wide[. . .] . I hang on the facts (1) that their average earnings over the last five or six difficult years show an enormous yield, even on their present price, and (2) that their market capitalization is extraordinarily small for a company which has a third of the motor trade of this country. —(KP: PC/­1/­1/­260–1) Railways

Under the Railways Act 1921, almost all of the hundreds of existing rail companies were grouped together into four new companies: London, Midland and Scottish Railways (LMSR), London and North Eastern Railway (LNER), Southern Railway (SR), and Great Western Railway (GWR). This “grouping” lasted from 1 January 1923 to 31 December 1947, and Keynes bought all of them, being strongly in favor of such an investment, as he explained to F. C. Scott: [Railway shares] seem to me to be an almost ideal credit cycle security, in the sense of being good things to buy when one hopes it is somewhere near the bottom of the slump. (JMK to F. C. Scott, 19 August 1932, KP PC/­1/­1/­130) The first three railway companies appear as his “pet” holding. Keynes commented favorably of the reorganization of LMSR by Josiah Stamp on 14 March 1932 (Keynes 1971–1989, CWK XXI: 84), giving attention once again to the quality of the management. Shipping

The shipping sector went through a troubled time in the interwar period; Keynes was moderately invested in it, and only in later years. There are two exceptions, namely the companies Hector Whaling and Indo-­China Steamship Navigation – “pets” that he kept for a longer period. The first was owned and run by his former student, Rupert Trouton, later a business adviser at Buckmaster & Moore. The second was Hong Kong–based, owned by the British firm Jardine, Matheson & Co., and operated between Chinese ports and Japan; in the early twentieth century, more than quarter of all the ships on the Yangtze River were owned by ICSNC.10 When war came to China in 1937, the firm suffered heavily both in Hong Kong and in Mainland China. Keynes disinvested in it soon afterwards. Investment trusts, and bank and insurance

Keynes was a director of various British investment trusts, including the Independent Investment Company (from 1923 to his death), the A.D. Investment

“Pets and favorites” 29

Trust (1921–1927), and the P.R. Finance Company (1924–1936), at which he was also chairman from 1932 to 1936 (Keynes 1971–1989, CWK XII: 1; Moggridge 1983). Not surprisingly, he preferred to invest in these companies, which were also among his “pets”. So was the Provincial Insurance Company, with which he was closely associated (see Westall 1992). Soft raw materials

Keynes had made large speculation in rubber, suffering heavy losses in the late 1920s (Keynes 1971–1989, CWK XII: 15; Moggridge 1983), but certainly acquiring great knowledge of the commodity. It is not surprising, therefore, to find rubber-­related shares among his pets. In particular he was familiar with Gopeng Perak, in Malaya, a British-­run company in the Kinta Valley, where tin mines were also located. Now I turn to Keynes’s investment in Wall Street.

New York Stock Exchange (NYSE)11 Keynes tried some sporadic dealings in US companies between 1911 and 1913 (KP: SE/­11/­1/­3) and then again from 1925 to 1928. However, none of these attempts ever reached a noteworthy scale. It was during the 1930s that Keynes’s American dealings grew significantly. Until the beginning of 1934, Keynes’s inclination was that there was no immediate prospect of recovery in the US economy, and he became confident in the US market and the US economy only after the first months of the new Roosevelt administration. Four sectors stand out in his portfolio: utilities, investment trusts, mining, and oil. Other sectors in which Keynes invested from 1934 onward were the motor industry and a varying set of companies in the industrial and consumer goods industries (see Figure 2.2). By the end of 1934, utilities and investment trust shares together already represented about a third of Keynes’s dollar investments. From 1936 on, Keynes’s US portfolio did not change very much in terms of distribution among sectors, but it inevitably changed in value due to the recession, the fall of the market, and large liquidation by Keynes (see Table 2.5). Within the seven sectors in which Keynes invested even here we find the subset of twenty-­three companies, whose shares Keynes held for a period of at least four years (his American “pets”; see Table 2.6). They represented only a quarter of the total number of companies traded by him in this time span, but at the same time on average they accounted for about three-­quarters of the value of his portfolio, with the highest value in 1938 at about 87% and the lowest in 1934 at about 60% (see Table 2.7). Here again is an overview of his “favorites” and “pets”.

$

Utilities Oil Motors & Motors Related Investment Trusts

31/12/1939

31/12/1938

31/12/1937

31/12/1936

31/12/1935

31/12/1934

31/12/1933

1,400,000 1,300,000 1,200,000 1,100,000 1,000,000 900,000 800,000 700,000 600,000 500,000 400,000 300,000 200,000 100,000 0

Industrial & Consumer Goods Insurance Mining & Metallurgy Railways & Railways Related

Figure 2.2  Holdings by sectors, 1933–1939, in USD at 31 December of each year Source: Elaboration from Keynes Papers, SE/­11/­5–7.

Table 2.5  Weight of sectors (%) in Keynes’s US portfolio, 1933–1939 Sector Utilities Investment trusts Mining and metallurgy Oil Industrial and consumer goods Railways and railways related Motors and motors related Insurance TOTAL

1933

1934

1935

1936

1937

1938

1939

– – 69.13

14.17 18.71 19.51

24.02 32.31 12.06

44.91 33.67 6.05

52.53 27.89 16.31

41.67 35.50 4.84

33.09 42.97 7.35

– 13.72

25.30 7.49

13.77 5.90

4.78 4.47

– –

7.45 7.64

7.50 5.25

17.15

11.44

5.67

3.66

1.92

2.17

0.45



3.39

6.27

2.23

0.77



2.15

– 100.00

– 100.00

0.22 100.00

0.58 100.00

– 100.00

Source: Elaboration from Keynes Papers, SE/­11/­5–7.

0.74 100.00

1.22 100.00

Table 2.6  List of “pets” in USD, 1932–1945 UTILITIES American Cities Power & Light (1937–1945) American Power & Light (1934–1940) Central States Electric (1936–1945) Commonwealth & Southern (1934–1938) Electric Power & Light Corp. (1934–1944) General Realty & Utilities (1936–1940) International Hydro-­Electric Syst. (1934–1937) National Power & Light (1934–1938) INVESTMENT TRUSTS Chicago Corporation (1936–1945) General American Investors (1934–1940) Pennroad Corporation (1936–1940) Prudential Investors Corp. (1932–1943, not 1933) Tri-­Continental Corporation (1934–1945) US & Foreign Securities (1932–1937, not 1933) United Corporation (1935–1945) MINING American Metal Co. Ltd. (1935–1944, not 1936–1937) Climax Molybdenum Co. (1939–1945) Homestake Mining Co. (1933–1936) RAILWAYS Chicago Rock Island & Pacific R. (1934–1940) Erie Railroad Co. (1934–1938, not 1937) INSURANCE American General Corporation (1936–1945) INDUSTRIAL & CONSUMER GOODS Associated Dry Goods (1933–1945, not 1937) OIL United Gas Corporation (1934–1942, not 1937) Source: Elaboration from Keynes’s papers SE/­11/­5–7.

Table 2.7  Weight of “pets” (%) in Keynes’s US portfolio, 1933–1939

Value (USD) of Keynes’s portfolio at 31 December Pets

1933

1934

1935

1936

1937

1938

1939

34,988

242,318

774,325

1,359,503

590,851

459,846

253,436

82.85

60.63

70.11

76.35

68.24

86.74

85.30

Source: Elaboration from Keynes Papers, SE/­11/­5–7.

32  Maria Cristina Marcuzzo Utilities

Six utility companies (American Power and Light, National Power and Light, International Hydro-­Electric System, Electric Power and Light, Central States Electric, and Commonwealth and Southern) were operating in the production of electric and hydroelectric power. The years of Keynes’s holding coincided with the New Deal, when the process of electrification of the country was underway. This came after a period of electric utility mergers and acquisitions, the subsequent rise of a few large holding companies and their collapse in the Depression. Afterwards, the surviving companies had to reorganize themselves after the Public Utility Holding Company Act mandated dissolving non-­ contiguous firms with common ownership. Investment trusts

Investment trusts was an area in which Keynes was extremely confident. According to a study carried out in 1938, there were six hundred investment trusts operating in the United States between 1930 and 1936, most of which showed a poor record (see Allen 1938; Rutterford 2009). General American Investors and Tri-­Continental, in particular, not only represented the two investment trusts in which Keynes invested more in terms of value but also those he kept longer, as they both appear in his portfolio evaluations without interruption from 1934 to 1939. Except in the case of US and Foreign Securities, another of Keynes’s “pets” among the investment trusts, Keynes usually held common shares of these companies. Typical of Keynes’s dealings in this sector is their continuity over time throughout the 1930s. It seems that external factors, such as the general macroeconomic context or the policies implemented by the government, had no great impact on Keynes’s choices in this field, and that he was more influenced by the specific characteristics of solidity and performance of these companies. Mining

Homestake, the company that took the lion’s share of Keynes’s investments in the US mining sector, was old and established and appears to have been very solid. Keynes’s interest in the US mining sector, and gold in particular, always showed a penchant for this company, reflecting Keynes’s own specialization in gold mining. Climax Molybdenum was a company dealing with a metal (molybdenum) that enhanced the toughness, durability, corrosion resistance, and thermal performance of steel. Its success was based upon brilliant engineering and management under the leadership of President Max Schott. Climax survived the worst of the Great Depression on growing exports to Germany, Japan, and the Soviet Union. In 1935, the stock appreciation, together with net earnings

“Pets and favorites” 33

exceeding 50% of gross sales, marked the best performance of any Depression-­ era American company. Railways

Keynes probably felt quite soon that the outlook for the railways was not too good in the United States. Automobile travel had begun to cut into the rail passenger market, and as unemployment increased during the Depression, the net income of the railroads plummeted. Roosevelt promptly launched the Reconstruction Finance Corporation program, which spent hundreds of millions of dollars to pay off railway debts throughout the decade, and Keynes probably saw this as a good opportunity. However, his correspondence in 1937 shows that his hopes did not last for long, and annual valuation of his portfolio confirms it. In fact we find only two companies, Chicago Rock Island & Pacific Railways and Erie Railroad Co. among his pets. The first went bankrupt in 1933, but John Dow Farrington returned the railroad to profitability, proving to be the Rock Island’s greatest leader. The Great Depression and resulting economic downturn of the 1930s meant a long road of decline for the Erie, which went bankrupt in 1938, when Keynes disinvested. Insurance

American General Corporation, which would become AIG, was run during the period under consideration by Gus S. Wortham, a leader with a great influence on insurance regulation and legislation. Keynes’s deep knowledge of the insurance business enabled him to assess the quality of the management. Industrial and consumer goods

Associated Dry Goods Corporation owned a chain of department stores, which were becoming very popular and successful. Keynes rightly saw in it the core business model for the burgeoning consumer market. Oil

Keynes’s initial large investments in the oil sector might also have reflected the policy of the US government since 1934. During this period, the Roosevelt administration came to the rescue of industry, with a set of regulations setting production levels to match market demand. Keynes’s choice fell on one company in particular, United Gas Corporation, which appears to have been very solid and in expansion. Keynes kept this company in his portfolio, almost uninterruptedly, from 1934 to 1939.

34  Maria Cristina Marcuzzo

Key points in Keynes’s philosophy of investment In both his LSE and NYSE dealings, Keynes showed a remarkable consistency in his behavior as an investor. Factors like knowledge of the sector or company (often acquired through personal connections), the solidity of the ownership, and trust in the management, together with size and frequency of dividends may lie behind Keynes’s choice to hold on to a smaller subset of company shares more than the others. There are, in fact, several statements by Keynes that help us to understand his philosophy of investment. I present here a selection of his published and unpublished statements, presenting the key points of his approach.

Flexibility I am strongly opposed to rigidities . . . Fixed percentage – particularly within each group of industry etc. – is surely altogether opposed to having an investment policy at all. The whole art is to vary the emphasis and the center of gravity of one’s portfolio according to circumstances. Subject to a minimum in Government securities and a maximum in ordinary shares I would strongly urge the desirability of the greatest possible flexibility. (KP: PC 1.5.68)

Informed knowledge [T]he very best principles of investment, namely to pick a firm which one particularly lies [likes] within an industry which one believes to be doing well. . . . I should hate not be able to invest in those in which I heard good accounts without buying at the same time a mixed bag of stuff about which I knew nothing or had heard less satisfactory opinions. (KP: PC 1.2.21) Out of some 250 commercial, industrial and produce companies, you will have your work cut out to discover 50 which are prima facie attractive and about which you can acquire adequate information. . . . to back an industry rather than a particular business, dividing the [the] investment between all the leading firms in the business, even though we might not know much about them individually! (Keynes 1971–1989, CWK XII: 159–160)

Knowledge is limited I am quite incapable of having adequate knowledge of more than a very little range of investments. Time and opportunity do not allow more. Therefore, as the investible sums increase, the size of the unit must increase. I am in favor of having as large a unit as market conditions will allow and,

“Pets and favorites” 35

apart from a small group of securities, this generally means a smaller unit than would be made necessary by the size of the investible funds. (JMK to F. C. Scott, 6 February 1942, XII: 82)

Investment versus speculation mentality It sometimes seems to me that apart from the noble army of investors who never read the newspapers I am almost the only person left who has an investment rather than a speculative mentality! On every Board I sit on the great majority are influenced far more by the daily fluctuations which they read of in the newspaper than by any reasoned calculation of yield or ultimate prospects. (KP: PC 1.4.306) [I]t seems to me to be most important not to be upset out of one’s permanent holdings by being too attentive to market movements. Unless one believes the market movements to be well founded, taking a long view, I should like a case where I felt real confidence to disregard it. It is indeed awfully bad for all of us to be constantly revaluing our investment according to market movements. Of course, it would be silly to ignore such things, but one’s whole tendency is to be too much influenced by them. (JMK to F. C. Scott, 24 August 1934, XII: 58–59) In conclusion, the main feature of Keynes’s approach to investment is the evaluation of the amount of information available for each individual asset and the degree of uncertainty about the future course of the main factors underlying it. Collection of the relevant information as best available was the premise to evaluating the “weight” of any argument that could be inferred from it, according to the conceptual framework which Keynes used in his Treatise of Probability to illustrate any decision-­making process. Interestingly enough, Keynes’s focus on the evaluation of the information available remained a feature of his approach both in his “speculation” phase (currencies and commodities in the 1920s) and in his “investment” phase (UK and US shares, in the 1930). This continuity, I believe, is a reflection of his approach to decision-­making and epistemology, which while remaining fairly constant over the years evolved also as a result of his practical acquaintance with asset markets.

Keynes’s approach to decision-­making and choice In standard economic theory, decision-­making is assumed to be based on rationality and optimization. In recent times empirical investigations and experimental psychology applied to the actual behavior of individuals have presented a different picture altogether, and behavioral finance has argued that individuals’ behavior in the financial market is explained more by psychology than economic theory. Keynes has been considered by some interpreters as a

36  Maria Cristina Marcuzzo

precursor of behavioral finance, while others have stressed the particularity of the notion of rational choice in Keynes, that is, the view that individual choice can be considered as rational, but rational does not necessarily means internal coherence without paradoxes and anomalies. As Carabelli pointed out, in Keynes’s case, The rationality of an agent’s decision can be considered simply that of having some grounds, some reasons, some evidence in favor of a solution in his/­her mind while arguing in a non-­demonstrative way. Some situations may be indeterminate, ambiguous, imprecise, but they are not irrational for these reasons. (Carabelli 2002: 172) Thus for Keynes, decision-­making is based on individual judgment deriving from real cognitive grounds, although these grounds are limited and partial; it is a judgment which has – as its cognitive grounds – evidence or a reason in which one really believes. This follows upon Keynes’s approach to probability, which consists in inferring the degree of beliefs from the available evidence and his criticism of probability, identified with statistical frequency, since frequencies are only a special kind of evidence and do exhaust the set of information upon which decision-­making takes place (Zappia 2016). Keynes as an investor supports the interpretation of Keynes’s epistemology as grounded on his theory of probability and in particular on the notion of the weight of the argument, as much of the literature has convincingly argued. There is, however an additional point connecting Keynes’s epistemology and decision-­making theory to his practice as an investor, which I would like to bring to light. In the General Theory, Keynes distinguishes between two ways of acting in the stock exchange: (1) “the activity of forecasting the prospective yield of assets over their whole life”; and (2) “the activity of forecasting the psychology of the market” (Keynes 1971–1989, CWK VII: 158). Which kind of behavior best describes “Keynes the investor”? From what we have seen, he was an investor who stood against general conventions by bringing his knowledge and expertise in decision-­making, both when taking risks as speculator, guessing the direction of market prices and betting on market sentiments, and when patiently waiting for the yields of long-­term investments. The analysis here provided of Keynes’s practice shows that while he kept always an eye on the prevailing market sentiments (or conventions), he in fact pursued and defended the strategy of a long-­term investor. This strategy does not mean disregarding the role of conventions and guessing what the average market opinion was, but it means that “conventions” should be understood in a broader sense (see Dequech 2011). John Davis is among Keynesian scholars who has given a lot of thought to the meaning of conventions in Keynes and provided us with useful insights on the issue. Already in his early book (Davis 1995), we find the idea that in

“Pets and favorites” 37

Keynes speculation lato sensu implies an interdependent judgment that does not necessarily imply disagreeing with the current average opinion. Opinions that diverge from the average and assume a future change in the state of affairs are part of the convention, and, in this sense, are also conventional. This implies that conventional behavior includes both attempts to conform to the general opinion and attempts to avoid being caught by it, specifically by those (more skilled) agents who act as long-­term investors. Keynes’s description of convention as the belief “that the existing state of affairs will continue indefinitely, except in so far as we have specific reasons to expect a change”, should then be interpreted, following Davis, as meaning different things for different individuals according to their different perspectives on how they believe average expectation is likely to change. Or putting it differently: a convention is a form of practical interaction between individuals, where average opinion exercises regulative effects on individual opinion while still accommodating judgment and action that departs from this central reference. It can be seen that such a structure is normative in the most general sense in that it imposes an orientation upon individual behavior without at the same time making that orientation binding. (Davis 1995: 129) This interpretation of conventions tallies with another angle from which to see Keynes’s characterization of human (not just economic) behavior: the notion of “reflexive agents”, which according to Davis, is the key to understand the notion of uncertainty. Reflexive agents are characterized by “being able to constantly adjust to change in the basis on which they act”, and this is the reason we have uncertainty as an ontological property of the world in which we live, rather than just the epistemic framework in which individual decision-­making takes place (Davis 2017). In conclusion, it seems to me that Keynes the investor fits very well with the characterization of an agent who adapts to the circumstances, exploiting information and judgment in a flexible manner, pursuing a strategy which, even if it does not guarantee success all the time, is the outcome of a rational assessment of the evidence available. He was an innovative investor who left a legacy which was later much drawn upon, especially in institutional asset ­management – a valuable addition to the richness of his heritage.

Notes 1 “Keynes the investor” has only recently become a specific field of study; the extant literature includes studies of Keynes as a speculator in currencies (Accominotti and Chambers 2016), commodities (Fantacci et al. 2010, 2012; Cristiano and Naldi 2014; Foresti and Sanfilippo 2017; Marcuzzo and Rosselli 2018), options (Marcuzzo and Sanfilippo 2016), investments in Wall Street for King’s College (Chambers and Kabiri 2016) and for himself (Cristiano et al. 2018), and in the London Stock Exchange (Marcuzzo and Sanfilippo 2018). Others have focused on Keynes’s strategies and performance as

38  Maria Cristina Marcuzzo a stock market investor (Holder and Kent 2011; Chambers and Dimson 2013; Woods 2013; Cristiano and Marcuzzo 2018), and connected to these are the works on Keynes as an innovator in institutional portfolio management (Chambers et al. 2015; Morecroft 2017; Woods 2018), the management of King’s College portfolio being the best known case study. 2 Keynes became a member of the board of the Provincial Insurance Company in 1923, lessening his involvement only when he joined the Treasury in 1940. 3 Keynes’s Papers, kept at King’s College, Cambridge (henceforth KP, followed by the catalogue reference numbers) include statements of account sent by the brokers to Keynes (mainly files KP: SE/­9, and SE/­8); notes and computations handwritten by Keynes on dealings and profits on each company across time and correspondence with brokers (mainly files KP: SE/­2/­6–7); exchanges (published and unpublished) with other investors and businessmen, containing views or considerations on specific companies and/­or the US markets; weekly statements of the Tilton Company (mainly files KP: TC/­3/­4, TC/­3/­5, TC/­4/­3 and TC/­5), registering the open positions and valuations at each date; and finally Keynes’s manuscript ledgers (files KP: SE/­11/­5–7). Two other fundamental sources of information on Keynes’s holdings are represented by the “Valuations” of his security portfolio (including both British and US) at the end of each year, made by Keynes himself in his ledgers (files KP: SE/­11/­5–7) and the Buckmaster & Moore statements of account relating to Keynes’s positions at each date, recording the number of shares held for each company, the (market) stock prices (in sterling and in dollars for the US shares) and the total value of the stocks held, always in sterling (for the US shares too) (files KP: SE/­2/­6–7). Buckmaster & Moore was Keynes’s main broker for his investment activity in financial markets. From 1937 onwards Keynes traded in financial markets also through another broker, Laurence Keen and Gardener, and through the bank Barclay, which begun to act also as Keynes’s broker from 1939 onwards (KP SE/­11/­7/­42). 4 This section draws on Cristiano and Marcuzzo (2018). 5 Keynes’s approach in the Lectures is akin to what we would call today the “forecasting theory”, whereby there is no clear trend of price movements in futures markets and profits are determined by the ability of speculators to forecast prices accurately (Lee and Zhang 2009). 6 On the inside cover of a booklet in which Kahn kept a record of his Stock Exchange transactions was penned the following motto suggested to him by Keynes in 1934: “The principles of successful stock speculation are based on the supposition that people will continue in the future to make the mistakes they have made in the past” (RFK papers 15/­1). 7 This section draws on Marcuzzo and Sanfilippo (2018). 8 Keynes started to regularly register the value of his holdings (at market prices) at the end (or at the beginning) of each year only from 1922 onwards, when the dimension of his investment begun to reach a certain level (above £10,000). Therefore, all our elaborations relate to the period 1922–1945, although from other documents we know that his trading in sterling securities started well before (precisely in 1911). 9 Kaffirs was the name given to the South African gold shares quoted on the London market. When South Africa abandoned the gold standard in December 1932, the South African pound depreciated, thus boosting the local currency receipts of the South African gold-­mining companies. This led to a boom in “Kaffirs” in London. 10 http://­industrialhistoryhk.org/­indo-­china-­steam-­navigation-­company-­1882/­. 11 This section draws on Cristiano, Marcuzzo, and Sanfilippo (2018).

References Accominotti, O. and Chambers, D. (2016). If you’re so smart: John Maynard Keynes and currency speculation in the interwar years, Journal of Economic History, 76: 342–386. Allen, E. D. (1938). Study of a group of American management-­investment companies, 1930–36, Journal of Business of the University of Chicago, 11: 232–257.

“Pets and favorites” 39 Carabelli, A. (2002). Speculation and reasonableness: A non-­Bayesian theory of rationality, in S. Dow and J. Hillard (eds.), Keynes, Uncertainty and the Global Economy: Beyond Keynes, Vol. II. Aldershot: Edward Elgar, pp. 165–185. Chambers, D. and Dimson, E. (2013). John Maynard Keynes, investment innovator, Journal of Economic Perspectives, 27: 213–228. Chambers D., Dimson, E. and Foo, J. (2015). Keynes, king’s and endowment asset management, in J. Brown and C. Hoxby (eds.), How the Financial Crisis and Great Recession Affected Higher Education. Chicago: University of Chicago Press, pp. 127–150. Chambers, D. and Kabiri, A. (2016). Keynes and Wall Street, Business History Review, 90: 301–328. Cristiano, C. and Marcuzzo, M. C. (2018). John Maynard Keynes: The economist as investor, Review of Keynesian Economics, 6: 266–281. Cristiano, C., Marcuzzo, M. C., and Sanfilippo, E. (2018). Taming the Great Depression. Keynes’s personal investment in the US stock market, 1931–1939, Economia Politica, 35: 13–40. Cristiano, C. and Naldi, N. (2014). Keynes’s activity on the cotton market and the theory of the ‘normal backwardation’: 1921–1929, European Journal for the History of Economic Thought, 21: 1039–1059. Davis, J. B. (1995). Keynes’s Philosophical Development, Cambridge: Cambridge University Press. Davis J. B. (2017). The continuing relevance of Keynes’s philosophical thinking: Reflexivity, complexity, and uncertainty, Annals of the Fondazione Luigi Einaudi, 51: 5–17. Dequech, D. (2011). Financial conventions in Keynes’s theory: The stock exchange, Journal of Post Keynesian Economics, 33: 469–488. Fantacci, L., Marcuzzo, M. C., Rosselli, A., and Sanfilippo, E. (2012). Speculation and buffer stocks: The legacy of Keynes and Kahn, European Journal of the History of Economic Thought, 19: 453–473. Fantacci, L., Marcuzzo, M. C., and Sanfilippo, E. (2010), Speculation in commodities: Keynes’s practical acquaintance with future markets, Journal for the History of Economic Thought, 32: 397–418. Foresti, T. and Sanfilippo, E. (2017). Keynes’s personal investments in the wheat futures markets, 1925–1935, History of Economic Ideas, 25: 63–90. Holder, M. E. and Kent, R. J. (2011). On the art of investing according to Keynes, Journal of Portfolio Investment, Spring: 1–6. Keynes, J. M. (1910). Lectures on the Stock Exchange, 8 Lectures on Company Finance and Stock Exchange, Lent Term 1910, MSS, UA/­6/­3, Notebook. Keynes J. M. (1971–1989). The Collected Writings of John Maynard Keynes (CWK), Managing Editors E.A.G. Robinson and D. Moggridge, London: Macmillan. Keynes J. M. (1971–1989). CWK IV A Tract on Monetary Reform, Managing Editors E.A.G. Robinson and D. Moggridge, London: Macmillan. Keynes J. M. (1971–1989). CWK VII The General Theory of Employment, Interest and Money, Managing Editors E.A.G. Robinson and D. Moggridge, London: Macmillan. Keynes J. M. (1971–1989). CWK XII Economic Articles and Correspondence, Investment and Editorial, Managing Editors E.A.G. Robinson and D. Moggridge, London: Macmillan. Lee, J. W. and Zhang, Y. (2009). Evidence on normal backwardation and forecasting theory in futures markets, Journal of Derivatives & Hedge Funds, 15: 158–170. Marcuzzo, M. C. and Rosselli, A. (2018). Trading in the “Devil’s metal”: Keynes’s speculation and investment in tin (1921–1946), in J. Kregel, M. Corsi and C. D’Ippoliti (eds.), Classical Economics Today, London: Anthem Press, pp. 167–188. Marcuzzo, M. C. and Sanfilippo, E. (2016). Keynes and the interwar commodity option market, Cambridge Journal of Economics, 40: 327–348.

40  Maria Cristina Marcuzzo Marcuzzo M. C. and Sanfilippo E. (2018). “Not sound, but brilliant” Keynes’s personal investments in the London Stock Exchange, Presented at Eshet Conference, Madrid. Moggridge, D. E. (1983). Keynes as an investor, Editorial Introduction to the Vol. XII, in Economic Articles and Correspondence, Investment and Editorial, in The Collected Writings of John Maynard Keynes, London, Macmillan, pp. 1–113. Morecroft, N. E. (2017). The Origin of Asset Management from 1700 to 1960 Towering Investors, Cham: Palgrave Macmillan. Rutterford, J. (2009). Learning from one another’s mistakes: The Investment Trust in the UK and the US, 1868–1940, Financial History Review, 16: 157–181. Westall, O. (1992). The Provincial Insurance Company 1903–38, Manchester: Manchester University Press. Woods, J. E. (2013). On Keynes as an investor, Cambridge Journal of Economics, 37: 423–442. Woods J. E. (2018). J. M. Keynes’ influence on institutional portfolio management, Journal of Post Keynesian Economics, 41: 56–82. Zappia, C. (2016). Whither Keynesian probability? Impolite techniques for decision-­ making, European Journal for the History of Economic Thought, 23: 835–862.

3 Simulation and economic methodology Marcel Boumans

Introduction This chapter is a contribution to economic methodology, where economic methodology is understood as the philosophy of science for economics, as John Davis and I state explicitly in the opening lines of our jointly written textbook (Boumans and Davis 2016). As indicated by its subtitle, we consider that economic methodology aims at the understanding of economics as a science.1 Philosophy of science should indeed be about science, that is, the practice of science. Philosophy of science, therefore, is philosophy that analyzes knowledge in the making, which means the philosophical understanding of the daily practice of scientific research and everything that such practice entails (e.g., ideas, tools and techniques, processes of inquiry, institutional settings, and social dynamics among investigators). This chapter discusses a specific research practice, namely simulation, which became dominant in macroeconomics. There are numerous ways to define simulation, but here it is understood in the way how it originally has emerged as a new research practice since the 1950s: the use of the computer as a tool of exploration, of “virtual experimentation” similar to physical experimentation (Morgan 2004, 2008). The current generally shared view on simulations, both in economics and in economic methodology, is that simulations are experiments, not in the real world but in the world of the model, therefore also called “computational experiments” (see, e.g., Kydland and Prescott 1996). In this chapter, however, the direction of investigation will be turned around. A new practice will not be studied from the perspective of an existing methodological framework but whether developments in contemporary economics will change economic methodology, whether new research practices produce new economic methodology. In a recent paper,2 Davis (2018) raised these questions in relation to whether simulation influences the notion of what counts as a good explanation in social science, or in more general terms, does it produce a new conception of economic methodology? In that paper he is particularly interested in the relation between how we know something (core methodological concern) and whether we know something (core epistemological concern). He emphasizes that a close connection between both is required, that knowing something is intrinsically connected to how we came to know it.

42  Marcel Boumans

I expect one of the reasons of having this concern is that Davis considers the current dominant methods in economics, such as deductive logic, as too narrow for a social science as economics. Particularly for social science concepts such as reflexivity, deductive logic is not an appropriate method to capture these kinds of phenomena. Such methods are too narrow to analyze these more complex social relations. The investigation of this relationship is relevant because it questions the idea whether when in a simulation a phenomenon emerges, this simulation can be considered as an explanation of that phenomenon. Simulations of this kind readdress the question of what an explanation is. In Davis (2018: 12), this question is focused around Joshua Epstein’s claim, “If you didn’t grow it, you didn’t explain its emergence.” The question then is: why should having “grown” a phenomenon count as a necessary condition for its explanation? According to Davis, this question is relevant because this kind of “generative explanations” constitutes an innovation in economic methodology. This is because “they substitute the idea of explanation as a process and practice for the conventional idea that explanation is an epistemic logic subsumed under given universal rules” (p. 15). The conventional idea refers of course to Hempel’s deductive-­ nomological model of explanation. A covering-­law explanation is too static in comparison to a generative explanation. It is the outcome of a trial-­and-­error process in which one must be lucky or smart enough to find among the available universal rules the right one(s) from which the explanandum can be inferred.3 A simulation is much more a heuristic process in the sense that in an interactive way one gets feedback on various changes in conditions or in rules till the explanandum emerges, in a similar way to explorative experimentation. Whereas a covering-­law explanation is an answer to a why question, a generative explanation is an answer to a what-­if question. The latter explanations are therefore also called how-­possibly explanations. There is, however, another important difference between both types of explanations. A covering-­law explanation is a one-­real-­world explanation, while a generative explanation is a many-­artificial-­worlds explanation. Simulation methods construct artificial worlds that resemble or imitate our world in various ways, and consequently compare what could be the case in terms of conditions and rules with what could be the case in resulting phenomena were the real world to closely resemble the artificial one (Davis 2018: 18). This aspect of simulations, that is that a simulation can resemble our world in different ways, is particularly welcomed by Davis because a social world is not fixed as a natural world. The latter is ruled by eternal universal laws, the former by social rules and institutions, which are local as well as temporary. Because (scientific) knowledge influences our behavior (reflexivity), rules and institutions continuously change, and hence deductive methodologies are too rigid for social science. But if generative explanations are to be preferred for social science, and replace covering-­law explanations, it is relevant to reconsider the criteria that

Simulation and economic methodology 43

would enable us to assess what a good explanation is. For a covering-­law explanation, these are that the relationships should be sufficiently invariant (lawlike) and empirically validated and the rules of deductive logic should be properly applied. But for a simulation, other criteria have to be developed, otherwise only a very limited set of simulations could count as a proper explanation. This chapter will explore what kind of criteria could be considered for the assessment of simulations as explanations.

Models Economics is a model-­based science which does not refer to laws very often. And when it does, the relationships that are referred to as laws are not stable or exceptionless enough to be considered as the kind of laws that are prominent in Hempel’s deductive-­nomological model of explanation. In economics, explanations employ abstract mathematical models that account for stable and robust facts which are assumed to be the general characteristics of an economic system or phenomenon replacing those laws in explanations.4 Although these models are representations of the explananda, these representations need not be assessed in terms of truth, but can be evaluated in terms of validity. Validity means that the models should be useful with respect to some purpose. Models are built for a certain purpose, to answer a certain question, so their assessment is an evaluation of how well they perform for achieving this purpose. If models are built to answer what-­if questions, the assessment of these models should be in the same direction: for a well-­known if situation, the what outcome should be as expected. This means that a model should generate, replicate, familiar characteristics for real and known circumstances. If a model performs well for these circumstances, it induces faith that it also will perform reliable enough for other not yet explored, perhaps artificial, conditions. In other words, the model is assessed whether it mimics the relevant characteristics of an economic system or phenomenon. If so, it is validated to answer other what-­if questions. In economics, the historical origin of this kind of model validation is to be found in the first computer simulation of a model representing an economy, namely Irma and Frank Adelman’s (1959) study of the business-­cycle properties of the Klein-­Goldberger model of the US economy. Their simulation showed that the model cycles were remarkably similar to those described as being characteristics of the US economy by the National Bureau of Economic Research (NBER). From this they concluded that the Klein-­Goldberger model was “not very far wrong” (Adelman and Adelman 1959: 621). The Adelmans did not use the Klein-­Goldberger model to answer what-­if questions; they were investigating whether the model could explain the US business cycle, so to answer a why-­question. The turn toward what-­if questions and the connected kind of model validation was taken by Robert Lucas (1980). Lucas (1980: 697) considered a model to be an “imitation economy,”

44  Marcel Boumans

which is tested by putting questions to it that we are fairly certain about the answers. The more dimensions on which the model mimics answers that actual economies give to simple questions, the more we trust its answers to harder questions. A “good” model [. . .] will not be exactly more “real” than a poor one, but will provide better imitations. Of course, what one means by a “better imitation” will depend on the particular questions to which one wishes answers. (Lucas 1980: 697) In a paper (Boumans 1997) discussing the simulation of the Adelmans and Lucas’s response to it, I showed that Lucas’s suggestion for model validation is in fact a Turing test. A Turing test is an intelligence test for a machine. If a machine gives answers that are indistinguishable from answers by humans, the machine is intelligent. To test a machine in this way, one needs to feed it questions that one knows the answer to. If done well, one can trust the machine to be intelligent enough to answer the harder questions for which one does not know the answers. This way of testing a model soon became to be known as calibration. This term was first introduced into macroeconomics by Kydland and Prescott in their 1982 paper. However, since their application, calibration has been controversial in economics and has generated quite some literature discussing the meaning and role of calibration in economics. One reason for this is the ambiguous meaning Kydland and Prescott gave to calibration in their early work on this topic. In their later papers, they expressed it less ambiguously by putting it more in line with Lucas’s new model approach: “Generally, some economic questions have known answers, and the model should give an approximately correct answer to them if we are to have any confidence in the answer given to the question with unknown answer” (Kydland and Prescott 1996: 74). But what are the economic questions for which we have known answers? The dominant answer to this question is most explicitly given by Cooley and Prescott (1995: 3). They describe calibration as a selection of the parameter values for the model economy so that it mimics the actual economy on dimensions associated with long-­term growth by setting these values equal to certain “more or less constant” ratios. By these ratios they meant the so-­called stylized facts of economic growth. This latter specific interpretation of calibration, that the known answers must be “more or less constant” ratios, is highly problematic in economics because one could wonder whether there are aspects of economic phenomena that have such stable properties. Apart from the fact that this invariance requirement is too restrictive for economics, it is also not necessary for calibration. Allan Franklin (1997: 31) proposes a definition that although more liberal is not less strong: “the use of a surrogate signal to standardize an instrument.5 If an apparatus reproduces known phenomena, then we legitimately

Simulation and economic methodology 45

strengthen our belief that the apparatus is working properly and that the experimental results produced with that apparatus are reliable.” This interpretation is close to Lucas’s original idea of testing “imitation economies,” namely a Turing test. Calibration is a “behavior pattern test” and therefore has an enormous advantage for the validation of simulation models compared to “direct structure tests.”6 A direct structure test assesses the validity of the model structure by direct comparison with knowledge about the real system structure, and is therefore an appropriate way of validating models built to answer why questions. Most econometric tests are structure tests in this sense. It has the strong requirement of having a priori knowledge of the structure of a real-­world system or phenomenon. Behavior pattern tests do not evaluate whether models represent the real system structure well enough, but they measure how accurately they can reproduce the major behavior patterns exhibited by the real system. These tests do not require that models are realistic and therefore can evaluate the validity of models built to answer what-­if questions.

Surprise The requirement that explanations should be analytically tractable is often used against the acceptance of generative explanations as proper explanations. The emergent outcome of a simulation can, however, be unexpected or surprising. The common view is that if an emergent phenomenon is not logical deducible from its constituents, such as the model assumptions, then this simulation cannot be accepted as an explanation of this phenomenon. This requirement of analytical tractability has become problematic with the use of computers. Most if not all current simulations are run on computers. Because any computer algorithm can be translated into a Turing machine, any simulation run on a computer is in principle logically deducible. But does logical deducibility in this sense also mean analytical tractability? A Turing machine itself can be described in simple terms: it consists of a scanner and a limitless memory-­tape that moves back and forth past the scanner. The tape is divided into squares. Each square may be blank or may bear a single symbol – “0” or “1.” The scanner is able to examine only one square of tape at a time. The scanner contains mechanisms that enable one to erase the symbol on the scanned square, to print a symbol on the scanned square, and to move the tape to the left or right, one square at a time. In addition to these operations, the scanner is able to alter the state. A device within the scanner is capable of adopting a number of states, and the scanner is able to alter the state of this device whenever necessary. This device functions as a simple memory (Copeland 2004: 6–7). Turing’s machines are capable of computing anything that any current computer can compute. Because a Turing machine is not only the symbols on the tape and a set of states but also the basic operations erase, print, move, and change state, a

46  Marcel Boumans

simulation is not only the algorithm but also the running of the machine. The number of iterations in Monte Carlo simulations is in the thousands; the number of iterations in a Schelling simulation is in the tens; and the proof of the four-­color theorem needs millions of iterations. Which simulation (whether estimation, explanation, or proof) is acceptable? At what number do we draw the line of analytical tractability? Surprise can arise when the running of thousands of iterations leads to a different result than the one obtained by mathematical, hence analytically tractable derivation. An example of such a surprise in economics is the simulation of a well-­known model of the business cycle, Ragnar Frisch’s 1933 “Rocking Horse Model.” This is a system of three mixed difference-­differential equations that for decades was considered to be an explanation of the business cycle.7 In the 1930s, mathematical economists held the assumption that the mathematics of mixed difference-­differential equations is the most suitable formalism for business-­cycle models. In general, it is difficult to solve mixed differential-­ difference equations. Moreover, in the 1930s, there were hardly any systematic accounts available. Systematic overviews on mixed differential-­difference equations did not appear until the early 1950s. Therefore, they were studied as if they were the same as the more familiar differential equations. The general solution of this latter kind of equations is a finite weighted sum of trigonometric and exponential functions, so that their periodic behavior can easily be analyzed. In contrast, the general solution of a mixed difference-­differential equation is an infinite weighted sum of these functions. This is not necessarily a periodic movement if the weights are not further specified. In a more recent study by Stefano Zambelli (1992), Frisch’s 1933 “Rocking Horse Model” was analyzed and worked out using computer simulations. It appeared that this system was not a cycle model, because when it is subjected to an external shock it evolves back to equilibrium in a non-­cyclical manner. Why such a paradoxically result found in [Frisch 1933] went unnoticed for almost sixty years is an intriguing question that, in my opinion, should be of interest to scholars of mathematical economics, business-­cycle theory and history of economic thought. (Zambelli 1992: 53) This case illustrates that working with a mathematical model, in this case Zambelli’s computer simulation, can be confounding. This computer simulation did not tell where the former studies went wrong – the surprise could not be traced back – but it showed that the theory on mixed difference-­differential equations was not complete. This example of surprise shows that analytical tractability is not a necessary requirement for explanations. In my view, the underlying more essential requirement for explanations is that they should be constructed according to some accepted rules, whether those are of first-­=order logic or Turing machine.

Simulation and economic methodology 47

New constructivism The requirement of analytical tractability is closely related to the requirement of rigor. Today, rigor is equated with the rules or axioms of first-­order logic, such as consistency (¬(A∧¬A)). Historians of science Giorgio Israel (1981) and Roy Weintraub (1998), however, have shown that in the history of mathematics, rigor has not always been identified with these logical rules. In the nineteenth century, rigor meant that “the problematic phenomenon” could be connected to “a physical phenomenon by use of empirical data,” that it could be “grounded in real phenomena” (Weintraub 1998: 237). According to Israel, these rules should not be so strict that they empty mathematical research of any external determination and content to such an extent that relation to applications is lost; they “should define and guide the use of mathematics as an instrument for the description, interpretation and control of phenomena” (Israel 1981: 219). Simulations invite us to reconsider these rules. Because generative explanations must generate the explanandum, the rules should be construction rules such that the explanandum can be generated. And again, they should not be so narrow that they would make in advance certain constructions “impossible.” For example, the quadrature of the circle is impossible, even though we can calculate the surface of a circle (πr2) to any degree of preciseness, only depending how long one wishes to digitally expand π. This result is only “impossible” if one would use the classical Greek rules of construction. A construction consists of a sequence of operations performed on a given configuration, which results in a new element of the figure with certain required properties. In classical geometry, the only two legitimate operations were drawing with a ruler and a compass. In mathematics and logic, constructivism maintains a strict interpretation of the phrase “there exists” as “we can construct.” It means the re-­interpretation not only of the existential quantifier but all the logical connectives and quantifiers as instructions on how to construct a proof of the statement involving these logical expressions (Bridges and Palmgren 2018). For example, the indirect proof is forbidden, because it proofs existence not by construction but only in an indirect way: one assumes the negation of an existence claim and from that assumption one then derives a statement that is known to be false. Existence then is proven, but one does not know what it is, what it looks like. Here, in relation to simulations, we must consider the phrase “you explain it, if you can grow it.” What are proper rules for growth, that is, for emergence? I would suggest the universal Turing machine, that is, the stored-­program digital computer in abstract conceptual form. A universal Turing machine is defined by a table of instructions. A phenomenon is explained when a table of instructions can be written and placed on the tape, the universal machine reads the instructions and carries them out on its tape, such that the ultimate output of this machine is the explanandum.

48  Marcel Boumans

Conclusions Methodology should reflect on and assess real practices of research, not ideal ones. A practice consists of people working with ideas, tools, and techniques to tackle a specific set of problems. This means that besides the more traditional focus on ideas, we also must reflect on these tools and techniques. In current practices, the presence of computers is prominent and necessary and increasingly essential. This even becomes more acute when the computer is becoming more and more a tool of exploration, gradually replacing physical experimentation. In other words, if simulation is becoming a dominant research practice in economics, economic methodologists should pay attention to this. Due to these developments in economic research, Davis (2018) expects that it eventually will change economic methodology. In particular, it will change our conception of explanation and its assessment. Hempel’s nomological-­ deductive model of explanation will lose this role of being a model for explanation in economics because its two constituent elements (laws and deductive logic) are too narrow to capture explanations based on simulations. This latter kind of explanations are answers to what-­if questions, and not so much as the covering-­law explanations, answers to why questions. Simulations do not have to be based on realistic models and do not have to be analytically tractable to be rigorous explanations. We only must agree on new rules: rules of construction and rules of validation. I suggest that the first are Turing machines and the second are Turing tests. Considering the influence of the work of economists such as Lucas, Prescott, and Kydland, one can conclude that simulation has changed economic methodology in practice, but that this change has not yet been sufficiently acknowledged by economic methodology in preaching; a symptom already diagnosed in 1992 by Dan Hausman as “methodological schizophrenia.”

Notes 1 From now on in this chapter, economic methodology is defined as the philosophy of science for economics. 2 This paper was presented at a History of Economics session at the 2018 Allied Social Sciences Association (ASSA) meetings, of which I was the discussant. I enjoyed being the discussant of this paper because it contains many intriguing ideas which I invited him to explore further. When I received the invitation to contribute to this volume, I almost immediately decided that I, instead of Davis, should make the first attempts of these explorations and to offer them to Davis as an extensive commentary on his ASSA paper but now in writing. 3 Traditionally the context of discovery was not the concern of economic methodology, hence a matter of luck or genius. 4 For a more detailed exposition, see the discussion of universal knowledge in Boumans and Davis (2016: 31–33). 5 Franklin does not discuss economic cases but experiments run in natural science. I therefore suggest that “instrument” and “apparatus” should be read as “model.” 6 This terminology is drawn from Yaman Barlas’s (1996) discussion of different kinds of model validation.

Simulation and economic methodology 49 7 This example is explored in Boumans (2012) as a response to Mary Morgan’s claim that physical experiments have more epistemic power than experiments on models (here simulations) because a simulation can lead to “surprise” while experiments may “confound” the experimenter. However unexpected the model outcomes, Morgan assumes that they can be traced back to, and re-­explained in terms of, the model. For an unexpected outcome of an physical experiment this is not possible, hence the stronger denotation of surprise, namely confoundment. Confoundment in Morgan (2005) has thus the same meaning as surprise in this chapter.

References Adelman, I. and Adelman, F. (1959). The dynamic properties of the Klein-­Goldberger model. Econometrica, 27: 596–625. Barlas, Y. (1996). Formal aspects of model validity and validation in system dynamics. System Dynamics Review, 12(3): 183–210. Boumans, M. (1997). Lucas and artificial worlds. New Economics and Its History, edited by J. B. Davis. History of Political Economy, 29(Supplement): 63–90. Boumans, Marcel. (2012). Mathematics as Quasi-­matter to build models as instruments, in D. Dieks, W. Gonzalez, S. Hartmann, M. Stöltzner, and M. Weber (eds.), Probabilities, Laws, and Structures. Dordrecht: Springer, pp. 307–318. Boumans, M. and Davis, J. B. (2016). Economic Methodology: Understanding Economics as a Science, 2nd edition. London: Palgrave Macmillan. Bridges, D. and Palmgren, E. (2018). Constructive mathematics, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2018 Edition), forthcoming. https://­plato. stanford.edu/­archives/­sum2018/­entries/­mathematics-­constructive/­ Cooley, T. F. and Prescott, E. C. (1995). Economic growth and business cycles, in T. F. Cooley (ed.), Frontiers of Business Cycle Research. Princeton, NJ: Princeton University Press. Copeland, B. J. (2004). The Essential Turing: The Ideas that Gave Birth to the Computer Age. Oxford: Oxford University Press. Davis, J. B. (2018). Agent-­ based modeling’s open methodology approach: Simulation, reflexivity, and abduction. OEconomia [online] 8 (4). http://­journals.openedition.org/­ oeconomia/­4402; DOI: 10.4000/­oeconomia.4402. Franklin, A. (1997). Calibration. Perspectives on Science, 5: 31–80. Hausman, D. M. (1992). The Inexact and Separate Science of Economics. Cambridge: Cambridge University Press. Israel, G. (1981). “Rigor” and “Axiomatics” in modern mathematics. Fundamenta Scientiae, 2: 205–219. Kydland, F. E. and Prescott, E. C. (1982). Time to build and aggregate fluctuations. Econometrica, 50: 1345–1370. Kydland, F. E. and Prescott, E. C. (1996). The computational experiment: An econometric tool. Journal of Economic Perspectives, 19: 69–85. Lucas, R. E. (1980). Methods and problems in business cycle theory. Journal of Money, Credit, and Banking, 12(4): 696–715. Morgan, M. S. (2004). Simulation: The birth of a technology to create “evidence” in economics. Revue d’Histoire des Sciences, 57(2): 339–375. Morgan, M. S. (2005). Experiments versus models: New phenomena, inference and surprise. Journal of Economic Methodology, 12(2): 317–329. Morgan, M. S. (2008). Experiments without material intervention: Models experiments, virtual experiments and virtually experiments, in H. Radder (ed.), The Philosophy of Scientific Experimentation. Pittsburgh, PA: Pittsburgh University Press, pp. 216–235.

50  Marcel Boumans Weintraub, E. R. (1998). From rigor to axiomatics: The marginalization of Griffith C. Evans, in M. S. Morgan and M. Rutherford (eds.), The Age of Economic Measurement. Durham, NC: Duke University Press. Zambelli, S. (1992). The wooden horse that wouldn’t rock: Reconsidering Frisch, in K. Velupillai (ed.), Nonlinearities, Disequilibria and Simulation. Basingstoke: Macmillan, pp. 27–54.

4 Identity Individual and social, and a theory of a ‘polymorphic’ but unique identity Ricardo F. Crespo

Introduction Over more than two decades, John Davis has repeatedly stressed the relevance for economics of improving its conceptualization of the individual identity. He considers this topic ‘the central philosophical issue in economics’ (2012: 96), adding that ‘whether economists are able to do this will be a crucial test of the relevance of economics as a discipline in this century’ (2012: 97). His view that values pervade the world implies that the individual is not atomistic, as in the neoclassical economics’ conception, but socially embedded. For him, the central dilemma in the theory of the individual is how a person can be social and individual at the same time. His proposals for explaining this are still developing. He has established two ontological criteria of identity (individuation and re-­identification through change) and proposed a capabilities conception of the individual that satisfies both criteria (2011). Davis worries about how an individual’s multiple selves make up a single individual. He argues that ‘self-­narratives’, or ‘discursive accounts people keep of themselves’ (2011: 183), allow people to ‘construct personal identities for themselves in the form of autobiographies’. These autobiographies are dynamic: they influence their social context and are influenced by it. With a capabilities conception of the individual, people have a special identity capability that allows them to organize themselves through a self-­narrative (2011: 190). Davis has argued that the idea of reflexivity is linked with identity: individual behavior and identity need to be understood in terms of some sort of capacity to reflexively orient on that behavior and identity, a type of idea which has had little place in the theory of decision-­making in economics, with a few exceptions. (2016b: 23) In this chapter, I will first review other conceptions of identity in ­economics – George Akerlof and Elizabeth Kranton, Amartya Sen, and Alan Kirman and Miriam Teschl – highlighting the strengths and problems that Davis sees in them. Next I will present the development of Davis’s account of this subject.

52  Ricardo F. Crespo

Finally, I will introduce another theory of identity based on Aristotelian and neo-­Aristotelian concepts – a ‘polymorphic’ account of a unique identity – that may fit with Davis’s aims. Personal identity is not a negligible characteristic of the economic agent because it greatly influences people’s decisions, including their economic decisions. This is why Davis considers important to adopt a sound theory of the identity of the economic agent. My claim is that Aristotle’s theory of identity serves as a territory over which we can construct that theory of the identity of the economic agent that responds to many of Davis’s concerns. Aristotle applies his ‘hylemorphic’ conception about the character of natural things – a substantial ground that supports all the characteristics and actions of them – to the human being. This is for him the fundamental ground of personal identity. The Oxford philosopher John Finnis elaborates on this Aristotelian conception of personal identity and adds three levels that refine the definition of a particular individual identity, including her experiences, choices, relation with others, and cultural environment.

Theories of identity in economics Akerlof and Kranton

Akerlof and Kranton introduced the topic of identity in mainstream economics in their seminal article (2000). Through the tools of standard economic reasoning (utility functions, game theory), they show that some presumed anomalies of rational choice theory and expected utility theory can be explained by the identity of the agents. They assert (2002: 1168): An individual gains utility when her actions and those of others enhance her self-­image. Furthermore, selfimage, or identity, is associated with the social environment: People think of themselves and others in terms of different social categories. Examples of social categories include racial and ethnic designations, and in the school context include, for example, ‘jock’ and ‘nerd.’ Prescriptions give the ideal, or stereotypical physical attributes and behavior, of people in each category. Individuals then gain or lose utility insofar as they belong to social categories with high or low social status and their attributes and behavior match the ideal of their category. For them, ‘without a model that mirrors this sociology, economic analysis produces only partial answers to key questions’ (2002: 1168). Thus, they propose to consider two parts of the utility function: the traditional ‘standard utility’ and ‘identity-­utility’ (2005: 14; 2010: chapter 3, 17ff.), and they think that the individual balances these two parts (cf. 2010: 18). Although the idea of considering ‘non-­standard economic’ motivations is gaining ground, it does not escape the standard economic logic. I think that

Identity 53

this is an example of a process described by Davis. He asserts that ‘economics, as other sciences, has regularly imported other science contents in the past, and having subsequently “domesticated” them, remade itself still as economics’ (2008b: 365). In effect, Akerlof and Kranton consider psychosocial motivations for economic behavior, but they ‘domesticate’ them, incorporating them into the utility function. In addition, Davis has two explicit criticisms of Akerlof and Kranton’s position. First, he criticizes the insertion of identity in the utility function, arguing that this is a ‘circular explanation’: ‘the preferences-­utility conception of the individual says that if one has one’s own (well-­ordered) preferences, one can be represented with a utility function and then identified as an independent individual. This, however, only assumes what needs to be shown’ (2016a: 24). This circularity is a traditional claim of Davis referring more generally to neoclassical economics’ underlying conception of the economic agent identity. In effect, Davis (1995) already argues that using the individual’s own preferences for identity presupposes the very individual that is in question. Davis (1995: 42, 2001: 119) clarifies that an early antecedent to this argument is Butler’s criticism of Locke’s conception of identity: memory comes after the individual and not the reverse and, consequently, to define the individual by its memory is a circular argument.1 Previously, Davis (1989, 2001) had explained that the neoclassical theory of reference is Fregean, not the ‘causal theory’ of reference. In the Fregean theory, meaning determines reference and not the reverse, as in the causal theory where we designate things that were previously ontologically distinguished. The neoclassical theory, by defining individuals by their preferences, cannot distinguish them: if they are single individuals, families, firms, and so on. As Davis notes, neoclassical theory is reluctant to abandon its commitment to the preference/­utility function of its rational choice theory because it is a top-­down, ‘self-­isolated closed science’, ‘entirely embedded in its own logic as compared to more bottom-­up sciences’ (2016a: 17). The tendency is to sew patches onto its pants instead of changing it. In fact, as explained above, this is what Akerlof and Kranton do: they introduce a new variable in the same utility function (2006: 378). Davis (2011: 81–84) also stresses that Akerlof and Kranton avoid the ‘multiple selves’ problem: they do not define a criterion to unite the multiple social identifications of people. Davis (2006) develops the problem of this and other theories of identity concerning this problem that had previously presented in his 2003 book (chapter 4). Concerning Akerlof and Kranton’s theory, he shows that the concept of the self that they employ is just the individual utility function. He argues that their model does not explain how the multiple social identities of an individual are related to personal identity and also how different persons can constitute a social group with a shared social identity. Social identity is for them exogenous, not endogenous, and consequently they cannot explain it and its relation with personal identity.2 This is a problem that Davis has always emphasized. He asserts (2004: 27): ‘what seems to me to be key to this particular history of thinking about individuals in economics is a failure in

54  Ricardo F. Crespo

most of the literature on the individual to ask whether individuality is endogenous to the social economic process’. The conclusion is that although Akerlof and Kranton emphasize the idea of taking into account identity as a motivation of economic behavior, they do it in an ‘economic-­like’ way that does not allow reaching the very identity of the economic agent. This is why Davis (2006: 374–377) calls their position ‘the neoclassical strategy’.3 Amartya Sen

Sen develops his ideas on identity in his book Reason before Identity (1999). There (1999: 15–19) he answers the question ‘discovery or choice?’ asked by the communitarian philosopher Michael Sandel. While Sandel’s view of our identities is that we discover them, Sen thinks that we choose them. Yet, he considers that our choices are not unrestricted (1999: 17) and that sometimes we also make discoveries. However, he adds: ‘choices have to be made even when discoveries occur’ (1999: 19). This concept of identity correlates with a Kantian conception of practical reason. For Kant, like Sen, freedom has priority over practical reason, determining the way of acting, and consequently, practical reason is constructivist: it constructs our own identity.4 This is necessary for Kant because he does not recognize the possibility of a metaphysical or theoretical knowledge of human nature. Davis (2008a) argues that having an identity is the most important capability in Sen’s approach. For Sen, personal identity is constructed through the choices an individual makes. Then, it is essential for the election and development of an individual’s capabilities. For Sen, as interpreted by Davis (2006: 378), personal identity is the identity of a person in terms of how she sees herself by reasoning and self-­scrutiny, capable of evaluating their values and objectives, and it is linked with social identity in terms of how she chooses to make commitments to social groups. Davis agrees with Sen’s criticism of Sandel because ‘discovering’ annuls evaluation about commitment with social groups, and that by regarding the individual as capable of this evaluation, social identity becomes endogenous to its choice and action (2006: 379). Personal identity, in Davis’s interpretation as the capability of evaluating and choosing one’s own objectives, is a basic capability. In Davis (2008a), he argues that Sen regards having an identity as the most important capability. Given that, for Sen, identity is built by the choices an individual makes, it must be central to the development of other capabilities. In addition, Sen’s view of identity allows the single person to integrate different selves (although Sen does not explain how it does it). Kirman and Teschl

Alan Kirman and Miriam Teschl affirm that standard economics has answered the question what the economic agent is by describing her as a maximizing

Identity 55

being. They add that Akerlof and Kranton’s model considers what and where the person is by situating it in a specific network and society, but not who she is (2004: 73). They propose an answer about who she is: ‘a self-­reflexive human being who has the capacity of actively discovering and consciously creating her identity within a given social context’ (2004: 63; see also 2006). They rely on the ‘complex view of identity’ of Derek Parfit (1984) and on Pierre Livet’s (2004, 2006) account of personal identity. They characterize the identity of the economic agent as reflecting a process of continuity and change. They pose the idea of an identity triangle formed by what the person is and does, who they want to be, and where they choose to participate. The desired self-­image (who) and the social groups (where) can change and they influence each other. However, we may ask who is the ‘who’ underlying all these desires, decisions, and changes. Davis (2011: 197) argues that Kirman and Teschl reproduce Akerlof and Kranton ‘multiple selves’ problem in a dynamic context. Davis argues: ‘How can someone be said to have a personal identity if what that involves continually changes?’ (2011: 208); or, ‘is the concept of personal identity as involving change – since this is what serves to set the problem of the multiple selves aside – still a concept of personal identity?’ (2006: 383). Indeed, this is essentially the conclusion of Horst, Teschl, and Kirman, who say that ‘personal identity of individuals is relatively weak’ (Horst et al. 2007: 23). For them, people really don’t have personal identities when we think of them endogenously’ (Davis 2011: 208–209).5

Davis on identity In the previous sections, I advanced some characteristics of Davis’s approach to identity. I will now elaborate a bit more on his approach. In his 2001 (115) book chapter Davis states: ‘The agent identity approach [. . .] shares the general realist concern personal identity theorists have regarding the persistence and distinctness of a type of real being in the world’. He calls his approach to identity ‘an ontological-­criterial approach to identity’ (2011: 4). Davis proposes two criteria for defining an individual: first, the individuation criterion, representing individuals as ‘distinct and independent beings’ (2011: 5); and second, the re-­identification criterion, by which those individuals can be re-­identified as distinct and independent, despite changes in many of an individual’s characteristics over time (2011: 5). Referring to the first criterion, he thinks that ‘for individuals to be distinct and independent, they must hold together as single whole beings and cannot fragment or break up into multiple selves’ (2011: 9). Davis maintains that the individual has a special personal identity capability, ‘interpreted as a capability for maintaining and developing an account of oneself in changing interactions with others’ (2011: 188), calling this position ‘the capabilities conception of the individual’ (2011: 170; see also 2015). For Davis, social relations are embedded in individuals (2003: 130). Individuals are not alone but embedded in groups, and consequently their actions cannot

56  Ricardo F. Crespo

be explained in instrumental rational terms. Individuals’ shared intentions are influenced and influence the social structure of the groups to which they pertain. This collective intentionality account of individuals’ identity marks an essential difference with the atomistic mainstream economics conception of the individual. Davis upholds that individuals have several changing capabilities. How does he avoid the issue of multiple selves? He believes that it can be overcome with ‘self-­narratives’, or ‘discursive accounts people keep of themselves’ (2011: 183), that allow people to ‘construct personal identities for themselves in the form of autobiographies’ (2011: emphasis added). The identity capability comes from people’s ability to self-­organize through a self-­narrative (2011: 190), thus allowing people to have enduring personal identities (2011: 209). Self-­narratives are ‘evolutionary, open-­ended, and generally do not get resolved, because people are continually engaged in developing their capabilities and this continually creates new possibilities for how their narratives will proceed’ (2011: 209). Rather than an individual task, this implies a mutual influence of personal and social identities: ‘who they [individuals] are is socially influenced, while at the same time they are a part of the social world because they influence it as well’ (2011: 213). Consequently, for Davis, the relational social identity is pivotal for other social identities and unifies them (2011: 204). More recently, Davis (2016b: 23) has stressed the idea of reflexivity linked with identity: ‘individual behavior and identity need to be understood in terms of some sort of capacity to reflexively orient on that behavior and identity, a type of idea which has had little place in the theory of decision-­making in economics, with a few exceptions’ (see also 2008a: 94, 99 and 101). This view does look richer than others. However, I think that the question remains, who is the individual that self-­narrates or reflects? It seems that self-­ narration and reflection stand at an individual level, but isn’t it a deeper layer establishing identity? Who is this self-­narrating and reflecting individual, and how do we characterize it? The problems posed by Davis are how to integrate the different selves into one personal identity, or in other words, how to link personal identity to social identity, and how to conceive a shared social identity of many individuals. In the next section, I will introduce an ontological conception of identity with different but integrated levels that I think answers to the first of Davis’s concerns and to the questions formulated in the previous paragraph. I think that it will be complementary to Davis’s proposals. The second concern, on the other hand, is a matter of social ontology that is beyond the scope of the current chapter.

An alternative proposal for the definition of personal identity In a recent paper co-­authored with Ivana Anton Mlinar (2017), we proposed a phenomenological approach to identity which we think is a valid answer to

Identity 57

the question about who the person is. Here I would like to introduce another theory, grounded in Aristotelian ideas, which may provide an answer to this question. I will develop this alternative in two steps. First, I will introduce an ancient metaphysical position about the nature of the human being, namely Aristotelian hylemorphism, which defines it at an ontological level. This position serves as the basis for a complementary position developed by John Finnis that I will present in the following section. Aristotelian hylemorphism

The problems noted in the previous positions point to the following idea: there is something ontologically more basic in human identity, understood in the way that these positions do, that is human agency. These authors seem to conflate both concepts. Human identity comprises the ‘radical’ or ‘basic’ who the person is and also its subsequent characteristics and actions that complete this who. In effect, Davis (2001: 114) defines human agency ‘as the power to initiate and bring about events’. This does not exhaust the human identity of each person.6 We need first to ascertain the nature of the ontological basis over which the complete identity of each person is built. Another possibility is to define human agency in a broader way than Davis, a way that would equate it to human identity. I will come back to this at the end of this section. David S. Oderberg (2005) argues, picking up the criticisms of Joseph Butler and Thomas Reid to John Locke, ‘there is a vicious circularity in trying to analyze personal identity, as Locke does, in terms of memory or of consciousness in general, since these phenomena presuppose identity (i.e., that it is the same person who remembers or is conscious)’. This is the problem that Davis has noted in the previous theories of identity. Kirman and Teschl (2004: 303) are right when they observe that the criteria usually taken into account to underlie personal identity – the person’s body, memories, psychological features, the brain – change or can be damaged. This is indeed an ‘identity problem’. However, as Oderberg (2005) notes, it is a problem only to the materialist. He asserts that the ‘problem of personal identity’ is a relatively recent problem stemming from a metaphysical approach for which immateriality is highly problematic, conceived as something accidentally united to matter. He thinks that the problem of personal identity is a problem for materialists who want to defend the existence of the mental, of free will and rational agency. However, Oderberg does not defend a ‘Cartesian dualism’ in which the soul, only accidentally united to the body, would confer identity to the person. He defends Aristotelian hylemorphism, a doctrine sustaining that two ‘co-­principles’ – ‘form’ and ‘matter’ – are united in a unique substance. He proposes it as an account of personal identity. This position resembles the anthropological conception underlying the definition of health care proposed by Davis and Robert McMaster (2017: chapter 2). They criticize the Cartesian dualist conception of the human being subjacent to the instrumental and

58  Ricardo F. Crespo

reductive evaluations of biomedical care and health economics. Instead, drawing from Aaron Antonovsky’s notion of salutogenesis, they promote health care, which is based on a view of ‘people as “integrated organisms” where the mind and body are intimately and complexly related’ (Davis and McMaster 2007: 38). Mittelmark and Bauer (2017) find philosophical traces of salutogenesis – a vision focusing on health rather than on illness, considering the individual in her completeness – in Aristotle’s anthropological conceptions. Let us expand on the origin and meaning of hylemorphism.7 Pre-­Socratic philosophers, or physicoi, as Aristotle called them, searched for the principles or origins of natural phenomena, that is, the fundamental constituents of all natural objects. In Metaphysics, Aristotle reviews the different pre-­Socratic proposals about the nature of these principles and develops his own position with regard to them (which he has also developed in his Physics). A thing may be said to be in different senses, but the primary sense is the substance (book Z [VII], 1). In Physics, Aristotle deals with the movements and changes happening in substances. These are of different types: qualitative, quantitative, local (all three inhering in the substance), and a fourth type, the generation or corruption of the very substance. This fourth type calls for a different subject (from the substance). He concludes: If, then, we grant that the things of Nature have ultimate determinants and principles which constitute them, and also that we can speak of them ‘coming to be’ not in an incidental but in an essential sense [. . .] then it is obvious that they are composed, in every case, of the underlying subject and the ‘form’ which their defining properties give to it. (Physics I, 7, 190b 18–21) This is the hylemorphic doctrine, the theory explaining the constitution of natural things from ‘primitive’ principles that ‘disappear’ in the compound. This compound is not as a ring in which a diamond is mounted over the gold, but as bronze in which copper and tin disappear. Aristotle calls the underlying subject próte hýle – ultimate underlying or prime matter (Physics II, 1, 193a 29). For Aristotle, the substantial composition does not imply that matter and form are parts of the substance: they ‘cannot even exist if severed from the whole’ (Metaphysics VII, 10, 1035b 24–25). He also affirms: ‘the final matter and the form are one and the same thing, the one potentially, and the other actually’ (Metaphysics VIII, 6, 1045b 18–19). The form is the primary cause, the actuality of being, the unifier of the substance, and it is not an element but a principle (cf. Metaphysics VII, 17, 1041b 25–31). As Anna Marmodoro synthesizes, ‘the substance is a composite of matter and form, and yet one’ (2013: 20). Aristotle applies his theory of hylemorphism to living beings. For him, the psyche (soul) ‘is the first actuality of a natural organic body’ (Aristotle 1957, On the Soul II, 1, 412b 4) – namely, the form. It provides actuality and persistence to the living being (cf. Irwin 1990: 288). In Stephen Brock’s expression (2013: 342), it is its ‘ontological energy’. The body, according to Irwin’s interpretation

Identity 59

of Aristotle (1990: 285; see also 294), is not the ‘remote’ matter – the chemical stuffs, that is, what Descartes considers the body – but the ‘proximate’ matter, which is the ensouled (empsychos) body (Metaphysics VII, 1036a 17): the body does not exist outside an ensouled body. Body and soul only exist as a living being – an ensouled body – which is not body or soul or a composite of both, as if they were different entities, but an ontological unity. They are not united accidentally. Soul and body are one as form and matter are one and the actions are from the unity: it is not necessary to ask whether soul and body are one, just as it is not necessary to ask whether the wax and its shape are one, nor generally whether the matter of each thing and that of which it is matter are one. For even if one and being are spoken of in several ways, what is properly so spoken of is the actuality. (On the Soul II, 1, 412b 6–9) They are one at the level of the substance, which is the living being, and they are two at the level of substantial co-­principles, but the individual does not outlive if they are separated (cf. On the Soul I, 1, 413a 5). In a sense, we have a dualism, but it is not the Cartesian dualism in which two different substances are accidentally united, like in the example of the ring. Consequently, the activities of living beings – including human beings – are not of the soul or of the body, but they are activities of the whole: ‘human beings are psychophysical wholes’, as William Jaworski expresses it (2011: 307). David Bostock (2006: 97) maintains, ‘the soul is not “the ghost in the machine” [a Cartesian expression] that makes the machine work as it does, but is rather the working of the “machine” itself’. Even thinking cannot be performed without the body (On the Soul I, 1, 403a 10 and 403b 19). Aristotle (1957) affirms: still to say that the soul gets angry is as if one were to say that the soul weaves or builds a house. Probably it is better not to say that the soul pities, or learns, or thinks, but to say rather that the man with its soul does these things. (On the Soul I, 4, 408b 13–15) Michael Frede explains: ‘there is just one subject, the animate object, which in virtue of the particular kind of form or soul it has, is capable of all these things [actions]’ (1995: 97). While avoiding a dualistic view of the human being, Aristotle’s (1957) hylemorphic conception of the soul as the form of the body allows for two compatible explanations: the natural philosopher [the scientist] and the logician [philosopher, psychologist] will in every case offer different definitions, e.g., in answer to the question what is anger. The latter will call it a craving for retaliation, or

60  Ricardo F. Crespo

something of the sort; the former will describe it as a surging of the blood and heat around the heart. The one is describing the matter, the other the form or formula of the essence. (On the Soul 403a 39–403b 2) The form is closely related with Aristotle’s ‘cause for the sake of which’ (later called final cause). The final cause is the most excellent perfection at which the form tends for the very fact of being this form. Aristotle emphasizes the role of the form and the final cause because they provide an ontological root to the unity of the human being. We could be tempted to consider the soul (only) as the organization of the material constituents of the human being. However, as Frede remarks (1995: 98), Aristotle’s proposal goes beyond this. Aristotle himself argues this point (On the Soul I, 4). Frede indicates: we do not try to understand the configuration in terms of the material constituents and their properties, but rather the other way round; we try to understand the material constituents and their properties in terms of the form or organization. (1995: 99) The body is not simply matter; it is ensouled matter. ‘We understand why parts, organs, structure, and other bodily processes of animals are as they are when we understand them in psychic terms’ (Irwin 1990: 290): they are teleologically oriented by the form-­soul. At the same time, ‘we correctly believe that psychic states are not reducible to purely material states, because psychic states have a teleological role that resists reduction’ (Irwin 1990: 291–292). Coming back to the topic of identity, for hylemorphism the human soul is different from the person because both body and soul are essential for the person. Oderberg explains that the form is the root cause or bearer of identity. However, the form does not exhaust identity because it does not exist but in the compound. There is also a material cause of identity. In Oderberg’s (2005) words, ‘the identity of a substance is given by the form as instantiated in matter’. Aquinas was clear about the difference between the soul and the person: ‘my soul is not me (anima mea non est ego)’ (Commentary on I Corinthians, 15.2 (on v. 19), quoted by Finnis 2005: 254). Finnis (2005: 253, 255) asserts: In Aquinas’s account, the foundation of one’s identity as an individual person is one’s material, bodily reality as this body[. . .] . The principle – principium, fundamental element and source – of this life, giving unity and continuity to the shifting material components of the organism, is the factor that Aquinas, like Aristotle, calls soul: the very form and lifelong act(uality) by which the matter of one’s bodily makeup is constituted the unified and active subject/­object, this organism, this human being, this

Identity 61

person, from one’s outset until one’s death. [. . .] In short, one’s bodiliness is the foundation of one’s personal identity, a subsisting identity that precedes consciousness and lasts through even protracted terminal unconsciousness until death. This does not mean that the human person is a composite of an undifferentiated soul (a ‘generic human soul’) with a specific body. The soul is the soul of this person: the ‘generic soul’ does not exist but only as the soul of a particular individual.8 In this conception, given that the human being lives in the time, memory is the prime witness and testimony of identity but not itself the identity (see Finnis 2005: 257). Finally, I return to a pending question inchoate at the beginning of this section. There, I asserted that Davis’s concept of agency does not exhaust human identity. This is because Davis’s agency refers to actions of a who that had not been defined. This has been done here. However, we can postulate a broader concept of agency including the who. This would be consistent with the hylemorphic position because the form is a basic active principle that together with matter constitutes a particular active agent. Now, having expounded on this theory of the ontological basis of human identity, I will introduce in the next section a second step in the development of a theory of identity for economics. John Finnis’s idea of personal identity

Here I introduce an explanation of identity – John Finnis’s explanation – that I think deserves consideration in respect to the requirements asked by the economic affairs for identity, and that is based on hylemorphism. Only a rational ensouled body may naturally develop the characteristics and abilities that connect the physical and the intentional as the human being does. Finnis maintains that the four orders of reality explained by Aquinas in the Proemium of his Commentary to the Nicomachean Ethics apply in the case of personal identity. These orders are the theoretical, the logical, the moral, and the technical that originate the corresponding ways of explanation or sciences. Concerning the first order, Finnis proposes (2005: 251) that ‘one has one’s identity, one’s personal identity, as a given (independently of one’s thinking about it, or about anything else), as one’s subsisting and continuity as this animal’. This level corresponds to the hylemorphic constitutive condition of the human being as formulated in the previous section. By this way, the diachronic identity (the identity through time) and the synchronic identity (the difference between each individual of the same species) are saved. Form and matter together ensure diachronic and synchronic identity. We do not need self-­narration to assure these identity criteria. A person in a coma has an identity regardless of his/­her ability to narrate, think, and so forth. It is the ‘foundation’ of every human being.9 This is the basic identity ground on which other levels are built. This level is fundamental because it is the ultimate level of the

62  Ricardo F. Crespo

who. However, the other subsequent levels are those that define the identity of the economic agent. Concerning the second order considered by Aquinas, the logical, ‘one has one’s identity, one’s personal identity, in and by subsisting as knowing, as conscious of one’s remembered childhood ignorance, one’s present improved but incomplete understanding, and one’s envisaged and intended future increase of knowledge’ (Finnis 2005: 251). It is the identity of a knower (i.e., the subject of knowing all that can be known). It adds perfection to the first order, but it may fail. However, given that the first order remains, this failure does not imply losing one’s identity. The set of acquired knowledge during life adds to the identity of the person and influences her economic decisions and behavior. The third order, the moral, is that which is anticipated and shaped in deliberation, choice, intentional actions, and thus – because choices last in the chooser until negated by a contrary choice – is instantiated in one’s subsisting character [. . .] In this order one has one’s personal identity both as self-­determining and as self-­determined. (Finnis 2005: 251) It is the identity of a practical reasoning being that shapes himself according to this reasoning. Every choosing act is ‘intransitive’ or ‘immanent’ in the sense that it remains in oneself, strengthening own’s identity. Finally, concerning the fourth order, the technical, ‘the persona is an artifact, intelligible and assessable as a means of communication, and by all the criteria applicable to the arts’ (Finnis 2005: 252). The last three orders, especially the fourth, naturally link personal identity with social identity. Finnis (2005: 280) quotes Germain Grisez (1975: 349, 351): A person is in all four of the orders, and he embraces all of them in himself. In the person the four orders are distinct, irreducible, yet normally inseparable. [. . .] The four aspects of the person all involve and in a way include one another, as the four orders always do. . . . The person is the self who unifies these four distinct and irreducible but normally inseparable aspects. The self [the soul] is a unifying principle. Finally, Finnis (2005: 281) concludes: In short, there is the identity I have as the particular human being I have been since I was formed in my mother’s womb with all the radical capacities, and thus the nature, of a human person. There is the identity I have as a person conscious of much that I have experienced and thought since first I began to be aware of myself (my toes and my fears. . .) in relation to other things including people, and began to inquire and assert and deny

Identity 63

and argue. There is the identity I have as a person with the character I have made for myself by the choices I have made within the constraints, and between the opportunities, that have arisen from the nature of things, the choices of others, and the accidental conjunctions, by all of which factors, as Shakespeare and Aquinas each depict, divine providence ‘shapes our ends, /­rough hew them how we will’ (Hamlet, 5. 2.3509–10). And there is the identity I have as an object of cultural forms (a citizen of Australia) and a user of them (English-­speaking, tie-­wearing) and a projecter of new and partly new personas. These orders are at the same time distinct but united in the person. The hylemorphic theory corresponds to the first order. It is the basis on which the other orders are built. It is the ensouled body which is capable of developing the other orders constituting this polymorphic but unique identity. It provides ontological soundness to the individual, independently of her experiences, thoughts, decisions, education, and so on – that is, the dimensions that shape her identity as human (and economic) agent. I consider that this approach to personal identity clearly fits with the economic agent: a free agent, capable of innovating and creating, relying on the past and projected towards the future, self-­reflexive and learner, evaluator of his decisions, influenced by his social adherences, and potentially shaped by them – that is, a complete human agent, simply a human agent that performs economic actions. In effect, the basic constitution of the who is defined, and this who develops during its life through experience, knowledge, relation with other people, decisions, and social conditions. It is this who who is engaged in economic affairs deciding and acting with the complete charge of its identity.

Conclusion In this chapter, I have analyzed John Davis’s theory of identity and proposed another theory following these steps. First, I introduced other theories of identity in economics adding Davis’s appraisal of them. Then, I presented Davis position and expressed a concern about the need of a deep ontological foundation of his view of identity as a basic capability of evaluating our own aims and objectives and the capacity of self-­narrating ourselves through time. His worry about connecting personal and social identity is manifest: for him, an important challenge ‘is to explain how individuals can still be individual when socially embedded’ (2008a: 93). In the last section, I proposed another theory that I think ensures the unity of the personal identity linked to its social identity. This last theory posits four levels of identity. The first comprises the ontological hylemorphic constitution of the human being. This is the ground on which the other three levels rely. This whole ‘polymorphic’ identity can adequately carry out the decisions and actions of the economic agent, with all their characteristics.

64  Ricardo F. Crespo

Notes 1 Biological and psychological continuity (in all its possible realizations: memory, personality, projects, preferences, etc.) may be regarded as epistemic criteria for an individual diachronic identity, but it provides neither the necessary nor sufficient conditions for personal identity. 2 See Davis (2009) on the problems of exogeneity. 3 See also the critical article of Ben Fine with similar arguments (2009). 4 On the Kantian character of Sen’s conception of practical reason, see Petri Rasanen (2012). Herlinde Pauer-­Studer (2006) argues that Sen’s position on practical rationality and identity can be considered as a middle position between Humeanism and the Kantian line. 5 The right quotation of Horst et al. is at www.sss.ias.edu/­files/­papers/­econpaper78.pdf or http://­hp.gredeg.cnrs.fr/­maurizio_iacopetta/­HKT21-­1.pdf (p. 12). Horst et al. maintain that Akerlof and Kranton proposal consider an exogenously given identity and they try ‘to present a model that accounts for endogenous identity formation’ (2007: 2). 6 I find suggestive Soran Reader’s (2007) stress on the non-­agential features of the person or self, especially its ‘patient’ condition. 7 I draw some paragraphs about hylemorphism from Crespo (2017). 8 I am conscious of the difficulties implied in these statements. If the forms are different for each individual, they would constitute different species, as angels. But at the same time, how would the souls be differentiated once the person is dead? For animals, it is clear for hylemorphism that the criterion for the diachronic identity (the identity through time) is the form and for the synchronic identity (the difference between each individual of the same species) is the matter. But the human person seems to be a special case. Aquinas’s formula ‘forma dat esse’ can be a way of explaining these difficulties together with his idea that the individuation produced by matter persists after death (De ente et essentia 6, 2). There is an extensive literature on these topics. For example, see the reflections of Christopher Martin (1993), the book edited by Georg Gasser (2010), and the dissertation of Eduardo Zacchia (2013). I do not intend to delve here into these difficult topics. 9 One may wonder whether this ground is relevant to the economic agent. I agree with economist Joseph Cropsey who, in an old article dealing with welfare economics convincingly argues for the need for a metaphysical or ontological ground for it (1955). The article has been recently recalled by Eric Schliesser (2015). This is an important claim. However, it goes far beyond the reach of this chapter and consequently I leave it for another occasion.

References Akerlof, G. and Kranton, R. (2000). Economics and identity, Quarterly Journal of Economics, 115(3): 715–753. Akerlof, G. and Kranton, R. (2002). Identity and schooling: Some lessons for the economics of education, Journal of Economic Literature, 40: 1167–1201. Akerlof, G. and Kranton, R. (2005). Identity and the economics of organizations, Journal of Economic Perspectives, 19(1): 9–32. Akerlof, G. and Kranton, R. (2010). Identity Economics: How Our Identities Shape Our Work, Wages, and Well-­Being. Princeton, NJ: Princeton University Press. Anton Mlinar, I. and Crespo, R. (2017). Identity theories in economics: A Husserlian phenomenology approach, Paper presented at the XIII International Network for Economic Methodology Conference, San Sebastián, 28–30 August 2017. Aquinas, T. De ente et essentia, Retrieved 23 March 2015. http://­legacy.fordham.edu/­ halsall/­basis/­aquinas-­esse.asp.

Identity 65 Aristotle. (1957). On the Soul, Translated W. S. Hett, Loeb Classical. Cambridge, MA: Harvard University Press. Aristotle. (1995). The Complete Works of Aristotle. The Revised Oxford Translation, J. Barnes (ed.). Princeton, NJ: Princeton University Press, 6th printing with corrections. Bostock, D. (2006). Space, Time, Matter and Form. Essays on Aristotle’s Physics. Oxford: Clarendon Press. Brock, S. (2013). Spirituality and Hylemorphism, in Doctor Communis, Proceedings of the XII Plenary Session, Vatican City, pp. 340–345. Crespo, R. F. (2017). Aristotelian hylomorphism: A framework for non-­physicalist philosophers about philosophy of mind, in P. A. Gargiulo and H. L. Mesones (eds.), Psychiatry and Neuroscience Update. Vol. II: A Translational Approach. Dordrecht: Springer, pp. 37–46. Cropsey, J. (1955). What is welfare economics? Ethics, 65(2): 116–125. Davis, J. B. (1989). Axiomatic general equilibrium theory and referentiality, Journal of Post-­ Keynesian Economics, 11(3): 424–438. Davis, J. B. (1995). Personal identity and standard economic thinking, Journal of Economic Methodology, 2(1): 35–52. Davis, J. B. (2001). Agent identity in economics, in U. Mäki (ed.), The Economic World View. Cambridge: Cambridge University Press, pp. 114–131. Davis, J. B. (2003). The Theory of the Individual in Economics. London: Routledge. Davis, J. B. (2004). Identity and commitment: Sen’s conception of the individual, Tinbergen institute discussion paper TI 2004–055/­2. https://­papers.tinbergen.nl/­04055.pdf Davis, J. B. (2006). Social identity strategies in recent economics, Journal of Economic Methodology, 13: 371–390. Davis, J. B. (2008a). The conception of the socially embedded individual, in J. B. Davis and W. Dolfsma (eds.), The Elgar Companion to Social Economics. Cheltenham: Edward Elgar. Davis, J. B. (2008b). The turn in recent economics and the return of orthodoxy, Cambridge Journal of Economics, 32: 349–366. Davis, J. B. (2009). Competing conceptions of the individual in recent economics, in D. Ross and H. Kincaid (eds.), The Oxford Handbook of the Philosophy of Economic Science, Oxford: Oxford University Press. Davis, J. B. (2011). Individuals and Identity in Economics, Cambridge: Cambridge University Press. Davis, J. B. (2012). Identity problems: An interview with John B. Davis, Erasmus Journal for Philosophy and Economics, 5(2): 81–103. http://­ejpe.org/­pdf/­5-­2-­int.pdf Davis, J. B. (2015). Agency and the process aspect of capability development: Individual capabilities, collective capabilities, and collective intentions, Filosofía de la Economía, 4(1): 5–24. http://­bibliotecadigital.econ.uba.ar/­download/­filosofia/­filosofia_v4_n1_01.pdf Davis, J. B. (2016a). Economics, neuroeconomics, and the problem of identity. Schmollers Jahrbuch, 136: 15–32. Davis, J. B. (2016b). The Sea Battle tomorrow: The identity of reflexive economic agents. Paper presented at the 3rd International Conference Economic Philosophy, Aix-­en-­ Provence, June 15–16. Davis, J. B. and McMaster, R. (2017). Health Care Economics. Abingdon: Routledge. Fine, B. (2009). The economics of identity and the identity of economics? Cambridge Journal of Economics, 33: 175–191. Finnis, J. M. (2005). ‘The thing I am’: Personal identity in Aquinas and Shakespeare, Social Philosophy and Policy, 22(2): 250–282. Frede, M. (1995). On Aristotle’s conception of the soul, in M. Nussbaum and A. O. Rorty (eds.), Essays on Aristotle’s De Anima. Oxford: Oxford University Press, pp. 93–107.

66  Ricardo F. Crespo Gasser, G. (ed.) (2010). Personal Identity and Resurrection: How Do We Survive Our Death? Farnham: Ashgate. Grisez, G. (1975). Beyond the New Theism: A Philosophy of Religion. Notre Dame, IN: University of Notre Dame Press. Horst, U., Kirman, A., and Teschl, M. (2007). ‘Changing identity: The emergence of social groups’. GREQAM Groupement de Recherche en Economie Quantitative d’Aix-­ Marseille, Document de Travail n°2006–51. https://­pdfs.semanticscholar.org/­0eeb/­ cf02118bdad2e9a2beb4795ecb7fa18c8366.pdf?_ga=1.189480107.1675971206.1482015 747, Retrieved 18 December 2016. Irwin, T. W. (1990). Aristotle’s First Principles. Oxford: Clarendon Press. Jaworski, W. (2011). Philosophy of Mind: A Comprehensive Introduction. Malden, MA: Wiley-­Blackwell. Kirman, A. and Teschl, M. (2004). On the emergence of economic identity, Revue de Philosophie Economique, 9(1): 59–86. Kirman, A. and Teschl, M. (2006). Searching for identity in the capability space, Journal of Economic Methodology, 13(3): 299–325. Livet, P. (2004). La pluralité cohérente des notions de l’identité personelle, Revue de Philosophie Economique, 9(1): 29–57. Livet, P. (2006). Identities, capabilities and revisions, Journal of Economic Methodology, 13: 327–348. Marmodoro, A. (2013). Aristotle’s hylomorphism without reconditioning, Philosophical Inquiry, 36(1–2): 5–22. Martin, C. (1993). Tomás de Aquino y la identidad personal, Anuario Filosófico, 26(2): 249–260. Mittelmark, M. B. and Bauer, G. F. (2017). The meaning of salutogenesis, in M. B. Mittelmark, S. Sagy, M. Eriksson, G. F. Bauer, J. M. Pelikan, B. Lindström, and G. A. Espnes (eds.). The Handbook of Salutogenesis. Cham: Springer. Oderberg, D. S. (2005). Hylemorphic dualism, in E. F. Paul, F. D. Miller, and J. Paul (eds.), Personal Identity. Cambridge: Cambridge University Press, pp. 70–99. www.newdualism. org/­papers/­D.Oderberg/­HylemorphicDualism2.htm, retrieved 19 December 2016. Parfit, D. (1984). Reasons and Persons. Oxford: Oxford University Press. Pauer-­Studer, H. (2006). Identity, commitment, and morality, Journal of Economic Methodology, 13(3): 349–369. Rasanen, P. (2012). Kantian basis of Amartya Sen’s idea of the reasoned scrutiny of thinking, SATS, Northern European Journal of Philosophy, 12: 178–197. Reader, C. S. (2007). The other side of agency, Philosophy, 82(4): 579–604. Schliesser, E. (2015). On Joseph Cropsey’s “What is welfare economics?” Ethics, 125: 847–850. Sen, A. (1999). Reason Before Identity. New Delhi: Oxford University Press. Zacchia, E. I. (2013). Subsistent Parts Aquinas on the Hybridism of Human Souls, Thesis submitted to the Faculty of Graduate and Postdoctoral Studies in partial fulfillment of the requirements for the Ph.D. degree in Philosophy, University of Ottawa. www.newdu alism.org/­papers/­E.Zachia/­Isdra_Zachia_thesis2013.pdf, Retrieved 20 December 2016.

5 Reflections on reflexivity and complexity J. Barkley Rosser

Introduction There is no single agreed-­upon definition of reflexivity within the social sciences, and some meanings of the word do not fit with how it has generally been used in social science discussions. Thus it can mean some sort of automatic “reflex” response to something, as in a knee jerking when tapped. However, more generally the term has been used in ways tied more to the word “reflection,” as in an image being reflected back and forth, possibly many times. The idea has involved agents interacting with an other (or others) that then reflects back onto them in some way. While many have seen the idea operating earlier, with some pointing to Marx and his class struggle system as being an example of a reflexive system, probably the first to use the term reflexivity in social science was sociologist Robert K. Merton (1938). A leader of the functionalist sociology movement, he was concerned with how sociologists think about sociology and how their own roles and positions affect that. Merton would foreshadow ideas showing up later in economics, such as the idea of self-­fulfilling prophecies (1948). But much of his focus was on what would become the scientific study of knowledge, a methodological movement that provided an early use of the idea in economics by D. Wade Hands (1994a, 1994b, 1998). However, prior to economists getting into studying reflexivity, sociologists would consider a wide variety of meanings and applications of the term, and other social scientists and philosophers would get involved. Prior to Hands, successful financier and self-­ styled “failed philosopher” George Soros (1987) applied the term to “the alchemy” of financial markets, effectively highlighting Merton’s idea of the self-­fulfilling prophecy to promulgate how agents respond to each other in markets to bring about various dynamics including bubbles and crashes, with his work attracting much greater attention following the 2008 financial crash (Soros 2013). While Soros recognized the priority of Merton, his more important influences were Karl Popper (1959) in philosophy (Soros had been a student of his), as well as the financial arguments of J. M. Keynes, especially from chapter 12 of The General Theory (1936), where the beauty contest idea both implies the possibility of self-­ fulfilling prophecies and is certainly an early example of reflexivity thinking,

68  J. Barkley Rosser

even if Keynes did not use the term.1 As it was, Popper was almost a negative influence in the sense that his emphasis on falsifiability and fallibility led Soros to question standard rationalistic economic theory. In any case, for Soros the crucial element of reflexivity is not just the reflecting off others but the eventual self-­referencing that occurs as one considers also their reflecting off oneself, with Soros recognizing the link of this idea to old paradoxes in mathematical logic. Economic philosopher John Bryan Davis initially began discussing reflexivity along the lines of Hands as more of a methodological issue related to the scientific study of knowledge within economics (Davis and Klaes 2003). He posed a hierarchy of forms of reflexivity, with some sort of self-­referencing at the core but with other perspectives then being added on. However, in the wake of the financial crash and the attention paid to the arguments of Soros, Davis (2013, forthcoming a, forthcoming b) expanded his notion and brought it in line with other discussions and concerns he had been dealing with previously, including the philosophical foundations of Keynes’s work as well as the nature of evolutionary economics as conceived by Veblen and Hodgson. It was from this stew that Davis began linking reflexivity to complexity. As with reflexivity, there is no universally accepted single definition of complexity. Indeed, there are groups of types of complexity, with computational, dynamic, and hierarchical among those most appearing in economics. Themes cutting across these include the idea of wholes being greater than the sums of their parts. Complex systems often involve emergence with higher levels implied by or appearing out of lower levels. A centerpiece of thinking about the link between reflexivity and complexity is that the self-­referencing part of the former may play a role in the phenomenon of emergence of higher levels or forms or structures within a complex system. A particular theme we shall pursue here is that it may matter how this self-­ referencing may take place. This will draw on ideas of Douglas Hofstadter (1979, 2006), who has probably influenced sociologists’ thinking about these matters more than he has economists’. Inspired by the role that self-­referencing plays in the paradoxes of mathematical logic, which in turn are connected especially to questions of computational complexity, Hofstadter focuses on the nature of the feedback loops that can arise in such systems, with a focus on how these relate to the emergence of consciousness, a problem that also concerned Hayek (1952). Hofstadter emphasized that indirect, or “strange,” loops involving such reflexive self-­referencing may be particularly important, drawing on insights from the Gödel (1931) incompleteness theorem. This then may relate to dynamic complexity as well, with this approach offering a possible way to synthesize the main branches of complexity, often seen as distinct or even in competition with each other, with these strange self-­referencing loops possibly making it easier for emergence of higher order structures in dynamically complex systems. Thus the link between reflexivity and complexity may be deep indeed.

Reflections on reflexivity and complexity 69

Forms of reflexivity before economists got to it An excellent overview of what sociologists have had to say about reflexivity, as well as some philosophers, anthropologists, and political scientists, is due to Michael Lynch (2000). He identifies more kinds of reflexivity than economists have since discussed, although some of the ideas economists have had about it were foreshadowed at least to some extent by these non-­economists. He identifies six broad types of reflexivity, with most of these having sub-­categories, all of which together he labels “the Reflexivities.” The first broad category he calls mechanical reflexivity, which in turn has three sub-­categories: knee-­jerk, cybernetic loopiness, and reflections ad infinitum. The second (Bateson 1972) involves feedback loops in human communication systems that lead to self-­reflection. He identifies the third specifically with Hofstadter’s work, particularly his famous 1979 book, Gödel, Escher, Bach: An Eternal Golden Braid, mentioning such things as halls of mirrors, the Möbius strip, and the famous Maurits Cornelis Escher work, Drawing Hands (Figure 5.1), which shows two hands drawing each other with pencils and which has been

Figure 5.1  Drawing Hands, Maurits Cornelis Escher, 1948

70  J. Barkley Rosser

mentioned and reproduced by many others who discuss reflexivity. We shall discuss this drawing further below. The second broad category is substantive reflexivity, which in turn has two sub-­categories: systemic reflexivity and reflexive social construction. The first is more associated with political scientists (Beck et al. 1994) and such ideas as “reflexive modernization” that deals with interactions between public opinion and policy making, especially when this involves experts intervening in this supposedly reflexive process. The second seems more to involve “self-­reflection” and emphasizes how consensual beliefs and concerted practices give rise to objective social institutions, including specifically economic markets (Berger and Luckmann 1966). These socially constructed facts are real (such as prices), but their reality is said to depend upon and be continually sustained by reflexive subscription to that very reality. The third broad category is methodological reflexivity, which in turn has four sub-­categories: philosophical self-­reflection, methodological self-­consciousness, methodological self-­criticism, and methodological self-­congratulation. The first of these Lynch attributes to Descartes (1968 [1637]), whom he identifies as the father of Enlightenment self-­knowledge gained by philosophical introspection. Methodological self-­consciousness is “mundane” and involves getting students to recognize their own relations to what they study and how these can lead to distorting biases (Hammersley and Atkinson 1983). Extending this leads to methodological self-­criticism, attributed to Merton (1938) and Popper (1963), Soros’s main teacher and influence. This emphasizes the systematic self-­criticism that supposedly real sciences engage in continuously to rid themselves of false or flawed theories. This in turn, in the hands of functionalist sociologists of science (Bloor 1976; Merton 1978), led to methodological self-­congratulation, where these sociologists praised themselves for following the practices of “mature” sciences, making the sociology of science “self-­exemplifying.” Arguably the original studies of reflexivity in economics by Hands (1994a, 1994b, 1998) fit somewhat into this broad category of methodological reflexivity, as do the first discussions of it by Davis and Klaes (2003). Lynch’s fourth broad category is meta-­theoretical reflexivity, seen broadly as involving the ironic detachment of the intellectual who “steps back” (Mannheim 1936), and which has three sub-­categories: reflexive objectification, standpoint reflexivity, and breaking frame. Reflexive objectification involves trying to be “hyper-­objective” in stepping back, as advocated by Bourdieu and Wacquant (1992). Standpoint reflexivity involves more rigorous self-­criticism of one’s own standpoints in the effort to step back towards objectivity (Harding 1996). Finally, breaking frame invokes illusionism and theatrical frames to understand social standpoints (Goffman 1974). The fifth broad category is interpretative reflexivity, which has just two sub-­ categories: hermeneutical reflexivity and radical referential reflexivity. The former emphasizes a circle between texts and their interpretations that expands to theorize “the constitution of society” (Giddens 1993). The second focuses on representation and challenges the possibility of objectification, searching for deep

Reflections on reflexivity and complexity 71

preconceptions, with this supposedly especially important in science and technology studies (Woolgar 1991). Lynch spends much of his later discussion on this type of reflexivity, viewing it as disruptive and “problematical” in its challenge to more conventional modes of self-­reflection and methodological moves to objectivity. The final category is ethnomethodological reflexivity, with no sub-­categories, which was associated with a broader influential movement in sociology beginning in the 1960s (Garfinkel 1967). It involves both methodology and substance and emphasizes the importance of accounts of social phenomena. It warns against attempting to decontextualize such accounts. Lynch summarizes all of these by noting that they all involve some sort of “reflexive turning back,” with many identifying reflexivity with the more radical programs of attacking objective sociology, even as Lynch notes that many of them attempt to make sociology more scientific through rigorous self-­ reflection and understanding. In any case, these largely sociological approaches provided the framework for the first discussions of reflexivity by Hands (1994a, 1994b, 1998) and by Davis and Klaes (2003).

Davis on reflexivity before Soros As noted above, the first excursion by John B. Davis (Davis and Klaes 2003) focused largely on the sorts of methodological issues that had long had the attention of the sociologists discussed above, even citing some of them, including Mannheim, Popper, Giddens, Bloor, and Woolgar.2 Davis and Klaes developed a three-­tiered approach to reflexivity and then used this to analyze the work of four economists: D. Wade Hands (1994a, 1994b, 1998), Esther-­ Mirjam Sent (1998, 1999), Uskali Mäki (1998, 1999), and Philip Mirowski (2002). The three tiers of reflexivity for Davis and Klaes are immanent, epistemic, and transcendent. The paper opens by noting that reflexivity comes from the Latin reflectere, which means “bending back.” This can apply to light reflecting off a mirror or a person engaging in internal thought or self-­reflection. They also note a set theoretical definition such that a relation involves every element of a set mapping to itself. In linguistics, a reflexive pronoun refers back to the subject of a sentence. They also note recursive functions and close the paragraph by mentioning the ubiquitous “Drawing Hands” by Maurits C. Escher, although without reproducing it or discussing it further. Rather than Escher’s print, they focus on possibly the most famous painting in Spain, Las Meňinas (“The Ladies in Waiting” or “The Family of King Philip IV”), painted in 1656 by Diego Velásquez da Silva. This somewhat mysterious painting (see Figure 5.2) has front and center the royal infanta, who was later Holy Roman Empress, the young daughter of the king and queen of Spain, with some attendants around her, including a dwarf and a dog, with her governess and a male servant behind this group. On the left-­hand side is Velásquez himself, painting something, although we only see the back of part

72  J. Barkley Rosser

Figure 5.2  Las Meňinas, Diego Velásquez da Silva, 1656

of the canvas, not what is on it – a matter of much speculation. On the right is a male figure leaving the room, apparently the queen’s chamberlain. On the wall in back is a mirror with two somewhat murky figures reflected in it, with the traditional identification of these as being the king and queen, possibly located outside the painting, and also quite likely the subjects of the unseen painting in the painting that Velásquez shows himself painting. This certainly sets up various “possibilities for self-­referencing reflexivities of various sorts,” and that is what Davis and Klaes do, using the painting to describe their three types, while drawing heavily on interpretations of the painting by Foucault (1966). Their lowest tier, immanent reflexivity, is also the one that contains the core concept of reflexivity, namely self-­referencing. The key point is that this painting is about painting, the crucial self-­referencing. They note later in the paper that this level involves the possibility of “malign” logical paradoxes involving

Reflections on reflexivity and complexity 73

self-­referencing such as the liar’s paradox: “All Cretans are liars, and I am a Cretan,” which leads to an endless do-­loop, one of those reflections ad infinitum discussed by Hofstadter, and which Bertrand Russell tried to eliminate from mathematics by developing his theory of hierarchical types. Lynch had also noted “malign” radical reflexivity associated with Woolgar and associates as well as a more “benign” sort associated with taking account of one’s own biases and paradigms. It is a curious point that Davis and Klaes miss an even potentially deeper form of self-­referencing that might be implied in the painting. While most observers think that in the painting Velásquez is painting the king and queen who are reflected in the mirror on the back wall, and others simply say there is uncertainty here, another possibility is that the painting Velásquez is painting is the painting itself, the painting he is in. If somehow we could see that painting, then it would contain a potentially infinite set of paintings of paintings within paintings, all of each other, getting smaller and smaller. This would be a full-­blown self-­referencing of the sort Hofstadter makes so much of and is also the sort implied by the liar’s paradox. Such an interpretation of the painting was first made by André Gide in 1893 (Gide 1972) when he coined the term, mise en abyme, which refers to images that contain images of themselves, with implied infinite sequences.3 But Davis and Klaes do not go there. Their immanent reflexivity only self-­references in that this painting is about painting, at least partly. Epistemic reflexivity enters in because of Velásquez having himself in the painting, a higher form of self-­referencing. For Davis and Klaes it is crucial that the painter has made it explicit that it is himself in the painting. This supposedly brings to light the “foundations of the painting as a representation, attending to its perspective and point of origin.” The emphasis on representation links this type of reflexivity to the radical referential reflexivity of Woolgar and company, with its possibility for a “malign” interpretation. Indeed, this matter of malign versus benign reflexivity is a central theme of this paper, which has as its title “Reflexivity: Curse or Cure?” The malign “threatens to undermine itself,” while the benign “has a self-­reinforcing character” (Davis and Klaes 2003: 333). Transcendent reflexivity involves stepping outside the painting. So for the painting, this is possibly the royal couple who appear reflected in the mirror on the back wall, but are implicitly outside looking at the painting and putting it in its social and historical place in the world of art, with all observers of the painting also potentially part of this as well. They then proceed to apply this framework to the four thinkers they have selected, with their focus fundamentally on methodological issues. They focus on Hands (2001), who focuses on reflexivity and the economics of scientific knowledge. The crucial immanent reflexivity involves an argued Quinean turning, in which ultimately economic methodology derives from itself through other sciences. They argue that the others covered in this paper are also caught up in the debate over this matter. The shift to epistemic reflexivity

74  J. Barkley Rosser

arises with the commentary on this by Hands himself. The transcendent enters in with the commentary by others on these arguments and Hands’s responses. Regarding Sent (1998), the immanent is seen in her study of Sargent’s role and positions in the development of the rational expectations hypothesis. She sees him setting up a symmetry between economists forecasting agent behavior and those agents themselves. This become epistemic when she considers her own analysis of Sargent’s work as a symmetry between the two of them, with both of them having an ability to reconstruct their own views. The transcendent appears as others comment on all this, with Sent (as with Hands) losing control as outsiders enter in, although in the end this results in a call for pluralism. Mäki’s (1999) immanent reflexivity involves considering Coase’s call for an “economics of economics” that would itself obey free market principles, bringing about a possibly desirable “Consistency Supposition” where one’s object-­economics matches one’s meta-­economics. The epistemic level arises from Mäki’s own meta-­analysis of Coase forcing him to consider his own analysis. However, for him the transcendent level is more complicated in that it is not just how others interpret or comment on this, but the fact that Mäki himself helped organize a petition in 1992 that advocated a free market approach to economic methodology – albeit with some not signing who supported this position while some who did sign not obviously being likely supporters of such a position. Davis and Klaes argue that this shows that as one moves from the immanent into the epistemic and then the transcendent: “how we respond to reflexivity ends up depending on a variety of additional factors that explain how reflexive relationships are socially embedded” (Davis and Klaes 2003: 344). The case of Mirowski (2002) involves his analysis of “the Colonel’s dilemma” regarding the study of using game theory in the Cold War at the RAND Corporation in the early 1950s. The immanent reflexivity arises from using game theory to study using game theory, which came about from realizing an enigma involved that looks like a liar’s paradox hall of mirrors. On the one hand, to successfully use game theory in war requires that the enemy knows one is doing that, while on the other one does not want an enemy to know one is using randomized strategies involving bluffing and so forth. According to Mirowski, this de facto paradox put a “chilling effect” on the study of game theory for war purposes at RAND. Of course, the immanent reflexivity here is the use of game theory to study the use of game theory. The epistemic part comes from considering Mirowski’s own analysis of this: if it is caught in a sort of paradox, then is he not also? However, in this case the transcendent has another aspect than for the other three cases due to the fact that Mirowski was discussing matters long kept secret and known only to a few. Thus as outsiders come in through reading Machine Dreams, effectively a new history is created. In some respects, this final case of Mirowski most clearly provides the bottom line for Davis and Klaes. The problem of potentially destructive malignity with reflexivity largely lies at the immanent level, where the basic self-­referencing

Reflections on reflexivity and complexity 75

opens the door to liar’s paradox kinds of dilemmas that can lead to essentially a nihilistic outcome for methodological researchers. It is by moving on to the epistemic and especially the transcendent levels that one turns this destructive malignity into potential benignity. As a conclusion and a segue into our next sections, it should be noted that this hierarchy of increasingly benign “self-­reinforcing” reflexivities does not relate particularly to a complexity vision of reflexivity. When we contemplate emergence complexly arising from reflexivity in a system, this is usually thought of as an endogenous process arising from the dynamics or structure of the system itself. But in this triadic formulation, that is not what is happening. The epistemic and transcendent levels do not emerge endogenously from the immanent level. Rather they are exogenously imposed by bringing in extra parties, first the principal agent and then out and out outsiders. The saving of reflexivity from its malignant cursedness due to the paradoxes inherent in the self-­referencing at the immanent level essentially involves calling in a methodological cavalry of exogenous outsiders to save the day and make reflexivity a benign cure rather than a malignant curse.

Soros and the turn to reflexive complexity John B. Davis began discussing complexity in economics as early as 2006 (Davis 2006), but this was not in connection with reflexivity, rather in connection with debates triggered by Colander et al. (2004) over the relationship between neoclassical and heterodox economics, with Colander et al. arguing that complexity was becoming a major influence on modern economics. It would be after the Great Recession, and particularly increased attention paid to the ideas on reflexivity of George Soros (1987), with his ideas seen as helpful in explaining the financial crash that led to the Great Recession that Davis would begin linking the two ideas. This increase in attention culminated in a special issue in 2013 of the Journal of Economic Methodology that featured a paper by Soros (2013), who accepted that reflexivity was connected to complexity, although the argument was probably pushed more strongly by others in the issue (that came out of a symposium funded by Soros), including especially Beinhocker (2013), Caldwell (2013),4 Hommes (2013), and Davis (2013), who from this point on has been fully on board with the idea of the two concepts being deeply linked as seen in Davis (forthcoming a, forthcoming b). However, before getting into this further, let us very briefly review basic ideas of complexity, which have received much more attention in economics than reflexivity (Rosser 2004, 2009). More than forty types of complexity have been identified, although many are variations of each other. Three broader categories have had the most influence in economics, and all three have links with reflexivity. One is computational, another is dynamic, and the third is hierarchical. Velupillai (2009) provides a good discussion of computational complexity. Very roughly, the complexity of a computer program is correlated with how many steps it takes to solve it. A conventional view sees this as being in

76  J. Barkley Rosser

roughly four levels. The lowest involves linear systems that are not complex, whose solutions are quick and easy basically. The next level up is polynomial, or P, which take qualitatively longer than linear ones, but which are mostly able to be solved eventually by modern computers. It is a widely accepted but still unproven hypothesis that the P level is distinct from the non-­P (NP) level above it involving more complex nonlinear systems that are not polynomial. While in principal solvable, many of these take so long that even now they are nearly impossible to solve in any remotely reasonable time with current computers.5 Finally there are computer programs that cannot be solved, which end up in infinite do-­loops. Here we see the clearest connection with reflexivity in that the kinds of ad infinitum hall of mirrors effects can underlie such outcomes, as these ultimately can arise from self-­referencing paradoxes such as the liar’s paradox, with the earliest understandings of this in computer science ultimately referring to Gödel’s incompleteness theorem. Rosser (1999) identifies Richard Day (1994) as having provided a definition of dynamic complexity. A system is dynamically complex if it is nonlinear and endogenously generates dynamics that do not converge smoothly to an equilibrium point or a non-­oscillating growth path (possibly negative). This has included or excluded two-­period cycles, but we shall not delve into that minor issue here. Rosser (1999) identified four kinds of such dynamic complexity, the “four Cs”: cybernetics, catastrophes, chaos, and “small-­tent” or agent-­based complexity. The dynamics described by Soros in his reflexive view of financial markets is widely viewed as fully consistent with this form of complexity. Simon (1962) is widely seen as the seminal formulation of hierarchical complexity. While not absolutely necessary, he viewed the more interesting forms of this as involving a spontaneous or endogenous emergence of the higher levels in a system, such as we have observed in biological evolution, even though in some cases the emergence may have been triggered by some external pressure such as a climatic change for biological evolution. In any case, we can see elements of hierarchical complexity in the other two, with the hierarchies of computational complexity, and in dynamic complexity there are hierarchies of degrees of complexity. The possibility of emergence is also linked with the old “whole is greater than the sum of its parts” idea that dates back to Aristotle and that many see as a deep aspect of broader complexity. For now we shall hold off on how reflexivity relates to this type of complexity, although reminding that the hierarchy of immanent, epistemic, and transcendent reflexivity formulated by Davis and Klaes (2003) was essentially exogenously imposed rather than endogenously and spontaneously emergent. So let us return to the version of reflexivity in Soros (2013), which draws on much of his earlier work, heavily influenced by Popper (1959), although partly in reaction. Thus where Popper thought that one can have certainty in science through falsifiability that at least some propositions are not true, Soros decided after studying economics and becoming disillusioned by the rational agents of conventional theory, that when thinking agents are in a system there may

Reflections on reflexivity and complexity 77

be no definite standard of truth because of reflexivity. The very actions and thinking of the agents can change the system, which in turn can change their thinking and actions.6 He saw this reflexive interaction as a basically two-­way street: people perceive and think with their cognitive function while then they act on external reality using their manipulative function. The results of these then feed back through the cognitive function in an ongoing dynamic process that is nonlinear and self-­referential. He recognizes that self-­referencing leads to the logical paradoxes, but that in actual dynamic situations these simply imply indeterminacy and radical Knight-­Keynes uncertainty. He invokes the beauty contest of chapter 12 in Keynes’s General Theory and links this to Merton’s (1948) self-­fulfilling prophecy idea. Thus he describes boom-­bust cycles in financial markets as agents creating market conditions that they react to. They may even be believing economic theories, but their own actions undermine those theories, and in actual financial markets, there are no independent market fundamentals or long-­run equilibrium. He specifically applies this to the housing bubble and decline that led to the Great Recession. There is simply an ongoing interactive dynamic process, and he recognizes that this is consistent with the now large literature on nonlinear complex dynamics in financial markets (Hommes 2013). In his piece for the symposium, Davis (2013) mostly repeats these arguments while noting that they imply fat tails in financial market returns. He also notes connections to behavioral economics and the critiques of the efficient market hypothesis. Davis seems to have largely continued to hold to this essentially Sorosian view of reflexive complexity, but has applied it and expanded it further. Thus, long a student of the philosophical underpinnings of Keynes’s thought (Davis 1994), he (Davis forthcoming a) revisits philosophical aspects of Keynes’s thought in light of reflexive complexity. Unsurprisingly, like Soros, he sees the beauty contest of Keynes as a central place where Keynes profoundly foreshadows reflexive complexity. It also shows the deep link with uncertainty, as what is going on in the beauty contest involves agents trying to figure out what the other agents are thinking, with them realizing the other agents are trying to figure out what they are thinking. Rather than resolving to some nice game theoretic probabilistic equilibrium, the agents are potentially caught in an infinite regress that can go to higher and higher levels of thinking about thinking about thinking about what each are thinking about, as in the Holmes-­ Moriarty problem (Koppl and Rosser 2002). This is hall of mirrors reflexivity at its clearest. Somewhat further afield, Davis (forthcoming b) considers dynamics of Veblenian cumulative causation as studied by Geoffrey Hodgson (1998, 2010) in critiquing methodological individualism. He argues that the self-­reinforcing positive feedback of cumulative causation is ultimately a reflexive process involving reflexive agents. Given that Rosser and Rosser (2017) have tied cumulative causation to complexity, this suggests that this is yet another example of reflexive complexity.

78  J. Barkley Rosser

Drawing Hands versus the Print Gallery We have noted already that M. C. Escher’s Drawing Hands (Tekenen in original Dutch) from 1948 (see Figure 5.1) has been repeatedly cited by people discussing reflexivity as an image of it, including that Douglas Hofstadter (1979, 2006) has been among those citing it as an image of a strange loop that he considers to be the foundation of consciousness. In his 1979 book he also discusses another Escher print, Print Gallery (Prentententoonstelling in original Dutch) from 1956 (see Figure 5.3), as another example of a strange loop, but he does not discuss it in his 2006 book, I Am a Strange Loop, whereas Drawing Hands makes multiple appearances in that book. Nevertheless, I shall argue here that if we wish to see an artistic image that truly represents reflexive complexity, it is Print Gallery that does so more profoundly and effectively.

Figure 5.3  Print Gallery, Maurits Cornelis Escher, 1956

Reflections on reflexivity and complexity 79

It is not that Hofstadter dismisses Print Gallery as trivial or unimportant – far from it. In his 1979 Gödel, Escher, Bach, he notes that it has three forms of “in-­ness.” One is that the gallery is in the town depicted in the picture on the wall the man is looking at, which he labels “inclusion.” The second is that the town is in the picture on the wall that the man is looking at, which he labels “depiction.” Finally he argues that the picture is mentally in the person, which he labels “representation.” However, we can also put these together to see that the man is in the print gallery that he is looking at on the wall, thus proving the sort of indirectness Hofstadter repeatedly argues is the key to a strange loop. Indeed, it has this character of hierarchical emergence that we identify with complexity in a way that one does not find in Drawing Hands as one moves from man to picture to city to print gallery back to man contained in the print gallery.7 Drawing Hands is all at one level, even if it implies an infinite do-­loop or recursion as one hand draws the other drawing the first, ad infinitum. One could even argue that it is not really all that reflexive. Rather than a going back and forth, it might be depicting a simultaneous solution of a pair of systems, much like with the basic partial equilibrium of supply and demand in economics. The two are jointly determining each other simultaneously with no such reflecting back and forth. Of course, in the economy it is precisely that we see such an echoing back and forth between agents and parts that raises the reflexivity question. We do not know why Hofstadter has come to focus more on the possibly simpler Drawing Hands that seems both less complex and less indirect in its “strange loopiness” than Print Gallery, although it is possibly because the former is much more well-­known than the latter and has indeed been discussed by so many people dealing with reflexivity and related issues. But there is another possibility. In 1979 Hofstadter made an inaccurate speculation regarding Print Gallery. This involves the hole in its center (which happens to contain Escher’s signature). The question arises if the hole must be there for its grid/­ metric bending8 that brings about the multiple embeddings Hofstadter noticed to happen. Hofstadter speculated that it must be. However, in 2003, de Smil and Lenstra demonstrated that it was possible to eliminate the hole through a rotation across the complex number plane. The result of this was to eliminate the bending grid, leaving an image that has the Droste effect (or mise en abyme effect), with one looking at the city and seeing a print gallery in which a man is looking at the picture of the city that contains the print gallery with another image of the man and the picture, and so on – much like the Droste cocoa jar with an image of itself on it (and similar such images on US food item containers, such as a Morton’s salt container with an image of a girl holding it, and more recent Land O’Lakes butter packets showing an Indian woman holding an image of the packet). So, indeed, buried in Print Gallery is a more conventional infinite set of self-­referencing images, once one unravels it appropriately.9 But then Hofstadter could counter that it is no

80  J. Barkley Rosser

longer really the Print Gallery in this deconstructed form, and indeed in that form it loses the special indirect and self-­feeding hierarchy effect that implies complexity and emergence in this print in a way that one may not see in any other work of art there is. It is truly unique.

On the foundations of reflexive complexity We now see the possibility of linking our major schools of complexity through the subtle strange loopiness involved in indirect self-­referencing at the heart of a deeper form of reflexivity. The indirect self-­referencing at the heart of Gödel’s incompleteness theorem is deeply linked to computational complexity in that it leads to the infinite do-­loops of the highest level of computational complexity in which a program never stops. The way out of incompleteness involves in effect what Davis and Klaes invoked: moving to a higher hierarchical level in which an exogenous agent or program determines what is true or false, although this opens the door to incoherence (Landini et al. forthcoming). The indirect self-­referencing opens the door to dynamic complexity in its implications for market dynamics, with this also linking to hierarchical complexity as new levels of hierarchy can be generated. Let us consider briefly how this comes out of the fundamental Gödel (1931) theorem. The Gödel theorem is really two theorems. The first one is the incompleteness one: any consistent formal system in which elementary arithmetic10 can be carried out is incomplete; there are statements in the language of the formal system that can neither be proved nor disproved within the formal system. The second one addresses the problem of consistency:11 for any consistent formal system in which elementary arithmetic can be carried out, the consistency of the formal system cannot be proved within the formal system itself. So coherence implies incompleteness, but any attempt to overcome incompleteness by moving to a higher level involves one being unable to prove the consistency of this higher-­level system, with both parts of this failing due to paradoxes of (reflexive) self-­referencing leading to paradoxes. Hofstadter (2006) provides an excellent discussion of the nature of the indirectness involved in proving the main part of the theorem, which involves the use of “Gödel numbers.” These are numbers assigned to logical statements, and their use can lead to the creation of self-­referencing paradoxical statements even within a system especially designed to avoid such self-­referencing statements. The system that Gödel subjected to this treatment to eventually generate a statement equivalent to “This sentence is unprovable” was the logical system developed by Whitehead and Russell (1910–1913) specifically to provide a consistent formal foundation for mathematics without logical paradoxes. Russell in particular was much concerned about the possibility of paradoxes in set theory, such as that involving self-­referencing sets. The classic problem was “Does the set of all sets that do not contain themselves contain itself?” A famous simple version of this involves “Who shaves the barber in a town where the barber only shaves those who do not shave themselves?” Both of these involve similar

Reflections on reflexivity and complexity 81

endless do-­loops arising from their self-­referencing. Whitehead and Russell attempted to eliminate these annoyances by developing the theory of types that established hierarchies of sets in ways to avoid having them refer to themselves. But then Gödel pulled his trick of establishing his numbers, which he applied to the system of Whitehead and Russell so as to through indirection generate a self-­referencing statement that involved a paradox unresolvable within the system. It is rather like how the hole Escher put in the middle of his Print Gallery allowed for the man to look at a print on a wall in a gallery of a city that contains the gallery in which he is standing looking at it. Thus it is not surprising that the problem of self-­referencing has lain at the core of much of the thinking about reflexivity from an early point, and that this thinking took on a sharper edge when various figures thought about Gödel’s theorem, or even earlier about the paradoxes considered by Russell.12 Linking this to understanding provides a foundation for a reflexive complexity that encompasses all the major forms of complexity.

Conclusions We have reflected on reflexivity and its relation to complexity by considering the history of the development of the idea along with a consideration of its underpinnings and foundations. Ultimately involving some sort of self-­ referencing through reflecting off an other, with that other possibly something one is inside of, such as a community or profession, it links to complexity in its various main forms of computational, dynamic, and hierarchical in various ways, potentially showing how these can be linked to each other. The initial explicit discussion of reflexivity occurred mostly among sociologists, with a substantial number of varieties of it considered. Many of these involved having social scientists more broadly take seriously their own roles in their disciplines and how their interactions with what they study impact them and in turn impact what they study. This self-­study eventually led to radical critiques of doing social science, as the inability to escape from the reflexive self-­aware loops led to nihilism. Following methodological work on reflexivity by Hands and others, Davis and Klaes (2003) extol the bringing in of outsiders or higher-­level considerations when self-­referencing threatens to generate “malignant” paradoxes. They saw these self-­referencing paradoxes as arising for immanent reflexivity in which there is a direct self-­referencing, such as the painting in the Velásquez painting Las Meňinas. The next level up of epistemic reflexivity involves the agent contemplating their own role in the system, as with Velásquez inserting himself into the painting. Finally at the highest level is transcendent reflexivity, when other parties enter into the contemplation or discussion of the system, as symbolized by the images of the king and queen in a mirror on the wall in the painting, suggesting that they are outside looking in at it. This calling upon higher-­level outsiders to overcome malignant paradoxes to produce a benign reflexivity echoes somewhat the move to a higher level implied by Gödel’s

82  J. Barkley Rosser

theorem in order to resolve incompleteness in a system, even if doing so generates a possible inconsistency. Following the 2008 financial crash, ideas of reflexivity proposed by Soros (2013) became influential, which drew more explicitly on self-­referencing and how the interactions of heterogeneous agents trying to figure out what the other agents are thinking about what they are thinking can lead to complex, emergent, self-­organized. and self-­fulfilling prophetic dynamic outcomes in markets that can manifest themselves as speculative bubbles and crashes. Soros and others (Davis 1994, 2013, forthcoming a) clearly link these ideas to those of Keynes (1936) in his General Theory, chapter 12. In this, he poses the beauty contest in which the winner is not the person who selects the most beautiful woman but best guesses the selections made by the other participants in the game, with this possibly involving an infinite regress of endless levels of thinking about what others are thinking about each other’s guesses. As argued by Hofstadter (1979, 2006) a deep aspect of how emergence of higher levels occurs implying both dynamic and hierarchical complexity may involve indirect self-­referencing that also implies computational complexity as well as a deep form of reflexivity. While many observers of reflexivity have cited M. C. Escher’s Drawing Hands as a picture of such self-­referencing, it may be that his Print Gallery provides a deeper picture of this more important indirect self-­referencing that may generate the more profound form of reflexive complexity that involves all the major forms of complexity.

Notes 1 Another group drawing on Keynes to emphasize the importance of self-­ fulfilling prophecies and how they can arise from essentially random origins is the sunspot equilibrium approach (Azariadis 1981; Farmer 1999), who see the possibility of multiple such equilibria. 2 They also cite some not mentioned by Lynch, including anthropologist Clifford Geertz (1973) on people understanding their own positions in studying societies and Alvin Goldman (2001), a radical ally of Woolgar who may be more challenging of how to do the Scientific Study of Knowledge (SSK) than even Woolgar. 3 Gide took this term from medieval heraldry, where sometimes a heraldic image would have an image of itself within itself. It is usually translated in English as “into the abyss.” In Dutch discussions it has been called the Droste effect, after a 1904 cocoa company’s container that has an image of itself on it. An early painting showing it is Giotto’s Stefaneschi Tryptich from 1320, which depicts Cardinal Stefaneschi kneeling while presenting a copy of the tryptich to St. Peter. See Snow (2016) for a much deeper discussion of Gide’s analysis. A modified version of Las Meňinas that shows such a Droste effect image of itself with implied multiple such embedded images on the rear wall above the mirror is at www.flick.com/­photos/­centrasian/­5549818324/­in/­photostream. 4 Caldwell’s paper focused on Hayek, who had also known Popper as had Soros. Soros had long ignored Hayek, but Caldwell argues for his ideas being deeply tied to complexity (Hayek, 1952, 1967) and compatible with Soros’s view of reflexivity, with Soros largely accepting this argument in his paper in the special issue. 5 Whether or not P = NP is one of the leading unsolved problems in mathematics. It appears that it was first recognized by John Nash in a long classified letter to the US National Security Agency in 1955. He thought that they did not equal each other but

Reflections on reflexivity and complexity 83 stated he could not prove it. He identified it as important for cryptology, which is why his letter would become classified for many decades until only shortly before Nash died. 6 A variation of this is the idea of “performativity,” wherein a theoretical idea such as the Black-­Scholes options formula can become a crucial part of market dynamics when agents begin to believe in it and act on it (MacKenzie, 2006). 7 Another curious tidbit about Print Gallery is that most of the prints one sees on the walls aside from the one of the city with the print gallery curiously embedded in it are ones by Escher himself, suggesting that if this is a mental image by the man, that man might well be Escher himself – another subtle form of reflexivity in it. 8 In a letter to his son, Escher reported being partly inspired by the “bending” of space in Einstein’s general relativity theory in creating Print Gallery. 9 An alternative outcome is to leave the bending but to fill in the hole with an infinitely descending set of images of the print in the hole. These can be seen on the website escherdroste.math.leidenuniv.nl/­index.php?menu=im. 10 By “elementary arithmetic” is meant that which can be derived from Peano’s axiom set assuming standard logic of the Zermelo-­Frankel type with the Axiom of Choice (ZFC). 11 It should be noted that in his original theorem Gödel was only able to prove incompleteness for a limited form of ω-­consistency. A proof for a more general form of consistency was provided by Rosser (1936), who used the “Rosser Sentence” (or “trick”):“If this sentence is provable, then there is a shorter proof of its negation.” This has led some to refer to the combined theorem as the “Gödel-­Rosser theorem.” 12 Without directly addressing either reflexivity or complexity, some who saw Gödel’s theorem as implying deep limits for economic analysis include Simon (1955), who saw computational limits from this as one of the bounds on rationality, with both Albin (1982) and Winrich (1984) raising similar such issues involving self-­referencing statements in economics that imply unsolvable infinite regresses. As already noted, the beauty contest of Keynes (1936, chapter 12) also had the possibility of infinite regress through a clearly reflexive setup involving indirect self-­referencing.

References Albin, P. S. (1982). The metalogic of economic predictions, calculations, and propositions, Mathematical Social Sciences, 3: 329–358. Azariadis, C. (1981). Self-­fulfilling prophecies, Journal of Economic Theory, 25: 380–396. Bateson, G. (1972). Cybernetic explanation, in G. Bateson (ed.), Steps in an Ecology of Mind. New York: Ballantine Books, pp. 399–410. Beck, U., Giddens, A., and Lash, S. (eds.).1994. Reflexive Modernization: Politics, Tradition and Aesthetics in the Modern Social Order. Cambridge: Polity Press. Beinhocker, E. D. (2013). Reflexivity, complexity, and the nature of social science, Journal of Economic Methodology, 20: 330–342. Berger, P. and Luckmann, T. (1966). The Social Construction of Reality. New York: Doubleday. Bloor, D. (1976). Knowledge and Social Imagery. London: Routledge and Kegan Paul. Bourdieu, P. and Wacquant, L.J.D. (1992). An Invitation to Reflexive Sociology. Chicago: University of Chicago Press. Caldwell, B. (2013). George Soros: Hayekian? Journal of Economic Methodology, 20: 350–356. Colander, D., Holt, R.P.F., and Barkley Rosser, Jr., J. (2004). The Changing Face of Economics: Conversations with Cutting Edge Economists. Ann Arbor: University of Michigan Press. Davis, J. B. (1994). Keynes’s Philosophical Development. Cambridge: Cambridge University Press. Davis, J. B. (2006). The turn in economics: Neoclassical dominance to mainstream pluralism? Journal of Institutional Economics, 2: 1–20.

84  J. Barkley Rosser Davis, J. B. (2013). Soros’s reflexivity concept in a complex world: Cauchy distributions, rational expectations, and rational addiction, Journal of Economic Methodology, 20: 368–379. Davis, J. B. (forthcoming a). The continuing relevance of Keynes’s philosophical thinking: Reflexivity, complexity, and uncertainty, Annals of the Fondazione Luigi Einaudi: An Interdisciplinary Journal of Economics, History and Political Science. Davis, J. B. (forthcoming b). Hodgson, cumulative causation, and reflexive agents, in F. Gagliardi and D. Gindas (eds.), Institutions and the Evolution of Capitalism: Essays in Honour of Geoffrey M. Hodgson. Cheltenham: Edward Elgar. Davis, J. B. and Klaes, M. (2003). Reflexivity: Curse or cure? Journal of Economic Methodology, 10: 329–352. Day, R. H. (1994). Complex Economic Dynamics: An Introduction to Dynamical Systems and Market Mechanisms, Vol. I. Cambridge: MIT Press. Descartes, R. (1968 [1637]). Discourse on Method and the Meditations. London: Penguin. de Smil, B. and Lenstra, H. W. (2003). The mathematical structure of Escher’s print gallery, Notices of the American Mathematical Society, 50: 446–451. Farmer, R.E.A. (1999). The Macroeconomics of Self-­Fulfilling Prophecies. Cambridge: MIT Press. Foucault, M. (1966). Les Mots et les Choses. Paris: Gallimard. Garfinkel, H. (1967). Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-­Hall. Geertz, C. (1973). Interpretations of Culture. New York: Basic Books. Gide, A. (1972). Journals, 1889–1949, Translated by O’Brien, London: Penguin. Giddens, A. (1993). New Rules of Sociological Method, 2nd edition. Stanford, CA: Stanford University Press. Gödel, K. (1931). Über Formal Unentscheidbare Sätz der Principia Mathematica und Verwandte Systeme I, Monatshefte für Mathematik und Physik, 38: 173–198. Goffman, E. (1974). Frame Analysis: An Essay on the Organization of Experience. New York: Colophon. Goldman, A. I. (2001). Social epistemology, in E. N. Zaria (ed.), The Stanford Encyclopedia of Philosophy. Stanford, CA: Stanford University Press. Hammersley, M. and Atkinson, P. (1983). Ethnography: Principles in Practice. London: Tavistock. Hands, D. W. (1994a). Restabilizing Dynamics: Construction and Constraint in the History of Walrasian Stability Theory, Economics and Philosophy, 10: 243–283. Hands, D. W. (1994b). Blurred boundaries: Recent changes in the relationship between economics and the philosophy of science, Studies in the History and Philosophy of Science, 25: 754–772. Hands, D. W. (1998). Reflexivity, in J. B. Davis, D. Wade Hands, and U. Mäki (eds.), The Handbook of Economic Methodology. Cheltenham: Edward Elgar, pp. 413–416. Hands, D. W. (2001). Reflection Without Rules: Economic Methodology and Contemporary Science Theory. Cambridge: Cambridge University Press. Harding, S. (1996). Standpoint epistemology (a feminist version): How social disadvantage creates epistemic advantage, in S. Turner (ed.), Social Theory and Sociology: The Classics and Beyond. Oxford: Blackwell, pp. 146–160. Hayek, F. A. (1952). The Sensory Order: An Inquiry into the Foundations of Theoretical Psychology. Chicago: University of Chicago Press. Hayek, F. A. (1967). The theory of complex phenomena, in Studies in Philosophy, Politics and Economics. Chicago: University of Chicago Press, pp. 22–42. Hodgson, G. M. (1998). On the evolution of Thorstein Veblen’s evolutionary economics, Cambridge Journal of Economics, 22: 415–431.

Reflections on reflexivity and complexity 85 Hodgson, G. M. (2010). Choice, habit and evolution, Journal of Evolutionary Economics, 20: 1–18. Hofstadter, D. R. (1979). Gödel, Escher, Bach: An Eternal Golden Braid. New York: Basic Books. Hofstadter, D. R. (2006). I am a Strange Loop. New York: Basic Books. Hommes, C. (2013). Reflexivity, expectations feedback and almost self-­fulfilling equilibria: Economic theory, empirical evidence and laboratory experiments, Journal of Economic Methodology, 20: 406–419. Keynes, J. M. (1936). The General Theory of Employment, Interest and Money. London: Macmillan. Koppl, R. and Barkley Rosser, Jr., J. (2002). All that I have to say to you has already crossed your mind, Metroeconomica, 53: 539–560. Landini, S., Gallegati, M., and Barkley Rosser, Jr., J. (2019, forthcoming). Consistency and incompleteness in general equilibrium theory, Journal of Evolutionary Economics 29. Lynch, M. (2000). Against reflexivity as an academic virtue and source of privileged knowledge, Theory, Culture, and Society, 17: 26–54. MacKenzie, D. A. (2006). Is economics performative? Option theory and the construction of derivatives markets, Journal of the History of Economic Thought, 28: 29–55. Mäki, U. (1998). Is Coase a realist? Philosophy of the Social Sciences, 28: 5–31. Mäki, U. (1999). Science as a free market: A reflexivity test in an economics of science, Perspectives on Science, 7: 486–509. Mannheim, K. (1936). Ideology and Utopia. New York: Harcourt Brace. Merton, R. K. (1938). Science and the social order, Philosophy of Science, 5: 321–337. Merton, R. K. (1948). The self-­fulfilling prophecy, Antioch Review, 8: 193–210. Merton, R. K. (1978). The Sociology of Science, in R. K. Merton and J. Gaston (eds.), The Sociology of Science in Europe. Carbondale: Southern Illinois University Press, pp. 3–14. Mirowski, P. (2002). Machine Dreams: Economics Becomes a Cyborg Science. Cambridge: Cambridge University Press. Popper, K. (1959). The Logic of Scientific Discovery. London: Hutchinson Verlag von Julius Springer. Popper, K. (1963). Conjectures and Reflections. London: Routledge and Kegan Paul. Rosser, J. B. (1936). Extensions of some theorems of Gödel and Church, Journal of Symbolic Logic, 1: 87–91. Rosser, J. B., Jr. (1999). On the complexities of complex economic dynamics, Journal of Economic Perspectives, 13: 169–192. Rosser, J. B, Jr. (ed.), (2004). Complexity in Economics: The International Library of Critical Writings in Economics 174, Vol. 3. Aldershot: Edward Elgar. Rosser, J. B., Jr. (ed.). (2009). Handbook of Research on Complexity. Cheltenham: Edward Elgar. Rosser, J. B., Jr. and Rosser, M. V. (2017). Complexity and institutional evolution, Evolutionary and Institutional Economic Review, 14: 415–430. Sent, E.-­M. (1998). The Evolving Rationality of Rational Expectations. Cambridge: Cambridge University Press. Sent, E-­M. (1999). The economics of science: Survey and suggestions, Journal of Economic Methodology, 6: 95–124. Simon, H. A. (1955). A behavioral model of rational choice, Quarterly Journal of Economics, 69: 99–118. Simon, H. A. (1962). The architecture of complexity, Proceedings of the American Philosophical Society, 106: 467–482.

86  J. Barkley Rosser Snow, M. (2016). Into the Abyss: A Study of the Mise en Abyme. PhD Thesis, London Metropolitan University. Soros, G. (1987). The Alchemy of Finance. Hoboken, NJ: Wiley and Sons. Soros, G. (2013). Fallibility, reflexivity, and the human uncertainty principle, Journal of Economic Methodology, 20: 309–329. Velupillai, K. V. (2009). A computable economist’s perspective on computational complexity, in J. Barkley Rosser, Jr. (ed.), Handbook of Research on Complexity. Cheltenham: Edward Elgar, pp. 36–83. Whitehead, A. N. and Russell, B. (1910–1913). Principia Mathematica, Volumes I–III. London: Cambridge University Press. Winrich, J. E. (1984). Self-­reference and the incomplete structure of neoclassical economics, Journal of Economic Issues, 4: 987–1005. Woolgar, S. (1991). The turn to technology in social studies of science, Science, Technology & Human Values, 16: 20–50.

Part II

Seeking justice The promotion of care and capabilities

6 Understanding the subaltern native middle class William Darity

Stratification economics and critical theory Stratification economics provides a powerful bridge across critical discourse in its post-­colonial studies, subaltern studies, and critical race theory modes to quantitative social science inquiry. Economics is of particular significance as a cornerstone of the process of synthesis with critical theory and the social sciences. Taking as its starting point the “tribe”, whether identified by race, caste, class, gender, religious affiliation, or the intersections thereof, stratification economics posits that a central concern of human beings is relative position vis-­ à-­vis their tribal affinity group. Human beings engage in perpetual comparisons between the condition and status of the social group with which they identify against other social groups, as well as comparisons between their personal position against other members of their own social group. The former constitutes assessment of intertribal (or intergroup) relative position and the latter constitutes an assessment of intratribal (or intragroup) relative position. Economics’ idiosyncratic provocateur, Thorstein Veblen (1899), offered a typology of social systems in his classic work, The Theory of the Leisure Class. The earliest social system which Veblen labeled “savage culture” was characterized by little differentiation in the community on the basis of status or possessions. This, for Veblen, was fundamentally a communal mode of existence. However, in a successor social system, “barbarian culture”, both hierarchy and invidious comparisons become the norm. By conquest and appropriation, a dominant community can separate themselves from others both in terms of comparative position and comparative material well-­being. Here is the origin of the intertribal comparison as an object of human attention. Almost simultaneously differentiation arises within one’s own group with emergence of the Chief, the Big Man, the Boss, the Magnate. Conquest, power, and might – essential themes in barbarian culture – lie at the heart of the creation of hierarchy between and within social groups. Which comparison – the between group comparison or the within group comparison – carries greater weight? It is context specific and can vary across time and space. But I would like to propose a hypothesis about the relationship between the two types of assessments that warrants further consideration: the

90  William Darity

greater the perceived distance between the groups, the greater the importance the average individual will give to their relative position within their group. This pattern of emphasis can apply both to members of dominant and subordinate groups.

Black Bourgeoisie In what follows, I would like to explore the application of the premises of stratification economics to a particular type of intersection of identities – the race-­class or caste-­class dyad – by examining the specific case of subaltern middle classes. There are two entry points for this conversation. First, John Davis’s 2015 essay published in the Cambridge Journal of Economics poses the problem of the application of stratification economics to circumstances where individuals have multiple salient identities, in the context of this article the combination of racial and social class identities. The second entry point for this conversation is E. Franklin Frazier’s text, Black Bourgeoisie. Indeed, it is an appropriate time to reassess Black Bourgeoisie (1957), now more than sixty years after its publication in English, particularly in an allegedly “post-­racial” America. When it first appeared in the United States, Frazier’s study of the American black middle class stirred a hornet’s nest, particularly among the members of the social group that constituted its target. Today the book has approximately sixteen hundred citations on Google Scholar and has laid the foundation for all subsequent studies on America’s black middle class, as well as often guiding analyses of subaltern middle classes in other parts of the world. For example, Henning Melber’s edited volume, The Rise of Africa’s Middle Class (2016) includes essays that pay due notice to a host of issues raised in Frazier’s book. Frequently, Black Bourgeoisie has been dismissed as less “scientific” and “objective” than Frazier’s larger oeuvre of scholarship, especially the major study that established his reputation as one of the nation’s preeminent sociologists, The Negro Family in the United States (1939). Biographer Anthony Platt (1989: 188) has observed of Frazier, who, in 1948, was the first black scholar elected president of the American Sociological Association, “his more polemical and explicitly politicized work, such as Black Bourgeoisie, is generally regarded as forgettable and somewhat unreasonable, even slightly disreputable, detour from his ‘real’ accomplishments”. But there are rich insights that can be drawn from Frazier’s study, despite its angry timbre, that point toward development of a general theory of the subaltern middle class. Frazier self-­consciously viewed his interrogation of the position of the black middle class in the United States as a basis for the development of a general theory of subaltern middle classes. In the final chapter of Black Bourgeoisie, he observed that the position of America’s black middle class might parallel the experiences of the rising middle classes in nations emerging from colonial rule: Our purpose was to treat the black bourgeoisie as a case study of a middle-­ class group, which had emerged during the changing adjustment of a racial

The subaltern native middle class 91

minority to modern industrial society. From this standpoint our study may have a broader significance than the group which we have studied. It may have some relevance for the study of the emergence of a middle class in colonial societies, especially in African societies at present undergoing rapid changes. The characteristics of this class in the various societies will have to be studied in each case in relation to its history and the economic and social forces which are responsible for its development. (Frazier 1957: 233) What precisely is the “bourgeoisie”; indeed, what precisely is the “black bourgeoisie”? In classic Marxist theory, the bourgeoisie is the upper strata in a capitalist mode of production, originally was understood as interchangeable with the owners of the means of production. The definition widened to include the better remunerated sections of unproductive laborers, workers whose labor does not generate a profit for someone else, for example, the elite professionals – lawyers, doctors, artisans and entertainers, and civil servants. Note that “unproductive” is a technical term that does not, necessarily, connote uselessness on the part of this group of workers. Indeed, I have been among those who have argued that it is more interesting and useful to theorize “unproductive labor” as a separate social class, instead of as a fraction of the traditional working class, and that they are the heralds of a post-­capitalist mode of production, managerial society (Darity 2017). The core dimension of its character as a class is unproductive labor does not generate surplus value but, instead, is supported out of the social surplus. A distinctive fraction of the bourgeoisie has come to be known as the petty (or petit) bourgeoisie, the owners of smaller capital whose operations are not organized in large-­scale corporations. These would constitute the shopkeepers, owners of craft works, and other small business owners. Over time the petty bourgeoisie has become synonymous with “the middle classes”, a mixture of unproductive laborers and owners of small capital whose income and wealth typically leaves them below the rich and above the poor. The construction of the notion of the middle class, thus, borrows more from Max Weber’s (1978) approach to social stratification (relations of exchange) than from Marx (relations of production). Frazier’s concept of the bourgeoisie in Black Bourgeoisie fully embraces an equivalence with the middle classes. But even Weber’s influence is less apparent than the indirect inspiration from Thorstein Veblen’s Theory of the Leisure Class (1899) and the direct inspiration from C. Wright Mills’s White Collar (1951). Veblen’s study examines, in part, the aspirations of the rising nouveaux riches, particularly their intense status-­seeking behavior and conspicuous consumption. Mills’s study emphasizes the non-­manual labor of the workers who are the object of his inquiry. By combining Veblen and Mills, Frazier’s “black bourgeoisie” constitutes his specific black middle class – blacks with white-­collar jobs who engage in conspicuous consumption and status seeking.

92  William Darity

The empirical basis for subalternity Certainly, Frazier’s black middle class did not possess the political authority or the wealth that Weber ascribed to his elite. And their social prestige largely was due to their status within the black community, hence relative to other blacks. It was within-­group comparison that made the black bourgeoisie a “bourgeoisie”. Via the between-­group comparison with white Americans, Frazier’s “black bourgeoisie” is not a bourgeoisie proper because black Americans are negligible owners of the nation’s means of production. In Black Bourgeoisie, Frazier highlighted the minuscule scale of black-­owned business capital in the following passage: Negro businesses are those enterprises which are owned and operated by Negroes. These business enterprises come within the definition of small businesses; in fact, they fall within the lowest category of small businesses. When the first study was made of Negro business in 1898, it was found that the average capital investment for the 1,906 businesses giving information amounted to only $4,600. When the latest study of Negro businesses was made in 1944, it was revealed that the average volume of business of the 3,866 Negro businesses in twelve cities was only $3,260. (1957: 53–54) Today, the relative scale of black business remains much the same. The combined sales revenue of all the firms on Black Enterprise’s list of top two hundred black-­owned businesses consistently is less than the individual revenue of the top sixty to seventy firms on the Fortune 500 list. Moreover, the upper tier of blacks fall well below the upper tier of whites in income and wealth. The black ladder reaches a much lower height than the white ladder. Frazier discussed this while making direct reference to observations made by Mills in White Collar (1951: 72–73), when describing those at the apex of the black class structure: the black bourgeoisie is comprised essentially of white-­collar workers. The less than one per cent of Negroes with incomes between $4000 and $5000 who are at the top of the pyramid of the Negro bourgeoisie, have incomes about equal to the median income of white-­collar workers among whites. The extremely small proportion (five in every thousand) of Negroes whose annual incomes amount to $5000 or more includes principally physicians, dentists, lawyers, entertainers and businessmen. The relatively large incomes of the Negroes in the professions are due partly to their business activities. A study made some years ago showed that 32 percent of the Negro physicians and 13.5 per cent of the Negro dentists were engaged in business activities. [Carter G. Woodson The Negro Professional Man and the Community Washington: The Association for the Study of Negro Life

The subaltern native middle class 93

and History 1934 pp. 110–111, 175] In this same income bracket ($5000 or more), there are a few college professors, public school principals, and persons employed in the federal, state and municipal governments. The businessmen with incomes of $5000 or more are to be found in the occupational group designated as managers, proprietors, and officials. Some of these businessmen include Negroes who have been successful in reaching the top of the rackets in the United States. In Chicago, for example, where the “policy” racket is a big business, this business was until recently “organized” as a cartel in a syndicate of fifteen men including twelve Negroes. [Drake and Cayton Black Metropolis pp. 478–484]. (Frazier 1957: 52–53) Today, the subordinate relative economic position of blacks is exemplified best by the magnitude of the racial wealth (net worth) gap. Nationally, the most conservative estimates indicate that black households in the United States, at the median, have one cent for every ten cents in net worth held by whites. Black household heads with a college degree have two-­thirds of the median net worth of white household heads who never finished high school, and blacks in the third quintile of the income distribution have a lower level of median wealth than whites in the lowest quintile (Hamilton et al. 2015). Blacks own less than 3% of the nation’s overall wealth but constitute 13%–14% of the nation’s population. One in seven white families is worth one million dollars, but only one in fifty black families is worth one million dollars (Jan 2017). Wealth affords the most useful economic indicator of a social group’s subaltern status. It is a far superior indicator of economic well-­being than income. Greater wealth or net worth – in general, the larger the difference between the value of what you own and what you owe – enables households to meet emergencies without extreme stress. Emergencies can include loss of employment, collapse of a small business, or a catastrophic illness. Wealthier households also can ensure that their children receive high-­quality schooling, including access to tutors to enable them to pass the requisite examinations. Wealthier households have the capacity to live in more upscale households, to live in neighborhoods with desirable amenities, to leave bequests, to exercise political influence, and largely to be free of anxiety over their financial condition. There are striking differences in the information to be gained from examining income and wealth disparities across social groups. For example, in the United Kingdom in 2009, the estimated median income for white households was £25,300, for black Caribbean households £22,000, for black African households £20,000, and for Bangladeshi households £21,000. These disparities were quite modest when juxtaposed against the racial/­ethnic wealth gaps prevailing at the time. Median white household wealth was estimated at £221,000, while the estimate for black Caribbeans was £76,000, for black Africans £15,000, and for Bangladeshis £21,000. Sixty percent of black and Asian households had no

94  William Darity

savings, in contrast with 33% of white households. The gap in median income between white and black Caribbean households was £3,300, but the gap in median wealth was a staggering £145,000. Patently, wealth is a far more accurate index of subaltern status than income (Darity 2014). Frequently, both economists and sociologists use income cut-­offs for identifying who is located in the middle class. Often, they simply declare all persons whose incomes lie in the middle three quintiles to be “middle class”. The World Bank has declared the middle class on the African continent to consist of persons with incomes ranging between $2 and $20 per day. Even without considering the absence of a wealth standard, these criteria seem foolish. As Henning Melber has observed: [Former director of the World Bank’s research department Martin] Ravallion . . . is at least honest enough to admit that such a definition of middle class is at best precarious, since “the vulnerability of this new middle class to economic contraction is obvious: one-­in-­six people in the developing world now live on between $2 and $3 a day.” As Raphael Kaplinsky from the British Open University quipped in his lecture at the General Conference of the European Association of Development Research and Training Institutions (EADI) in mid-­2014, such a category means that everyone not starving qualifies as middle class. (2016: 2) The weakness of this approach only is compounded by its silence on wealth as a factor included in identification of a middle class. Perhaps a superior alternative is to construct an index with a mix of indicators of class status, for example, educational attainment, income, wealth, and occupational prestige (Marsh et al. 2007) – but like the income cut-­offs strategy, this is an operational definition that is not clearly connected to a conceptual definition of the middle class.

Locating the middle class My preferred concept of the middle class is to define it as those persons who functionally are between the two great classes that define a particular social mode of production. So under the slave mode of production characteristic of societies of European antiquity, the middle class consisted of those persons who were neither citizen nor slave. Traders and merchants from foreign dominions would be representative examples. Under the feudal mode of production, the middle class would consist of persons who were neither lord nor serf (e.g., the shopkeeper, midwife, shoemaker, housewife, or blacksmith) in emerging villages or towns set apart from the lord’s demesne. Under the bourgeois mode of production (or capitalism), the middle class comprises persons who are neither capitalists nor laborers, de facto persons who perform work that is unproductive in the technical sense of the term described above. Indeed, it was under the bourgeois mode of production that

The subaltern native middle class 95

unproductive labor, or the middle class, has gained its greatest degree of freedom and agency via an expanded role for the state (Darity 2017). From this perspective, today’s black middle class simply would be all black Americans who share the same functional social position as unproductive labor in general. But as noted above, this was not Frazier’s approach to constructing the boundaries of his “black bourgeoisie”, albeit a darker but pale imitation of the “white bourgeoisie”. In the process, his black middle class does not lie between labor and capital but consists of elements of both labor and capital – as long as the individuals live in a white-­collar world of employment and engage in conspicuous consumption. Of course, Frazier’s book inflamed black American high society. Its members were disturbed by his portrayal of their frivolous behavior, their failure to provide the leadership for the black community that DuBois hoped would come from “the Talented Tenth”, their efforts to replicate the activities of the white elite – despite lacking comparable resources and their descent into what Frazier described as a fantasyland, a “world of make believe”.

Contradictions of the subaltern native middle classes To conclude this commentary, I want to situate Frazier’s book in the context of two theoretical developments mentioned at the start of this article: the emergence of subaltern and post-­colonial studies. As Frazier suggested himself, we can read Frazier’s text as an examination of particular case of a subaltern native middle class – the black middle class in the United States. Subaltern studies is an interdisciplinary field that came into being about thirty years ago that focuses on the lived experiences of communities that are marginalized, denigrated, and assigned to a lower rank or status than other groups in a community. It is a far more useful concept than “minorities”, since a subaltern population can be a numerical majority (e.g., blacks in South Africa or Afro-­Brazilians) and a dominant population can be a numerical minority (e.g., Brahmins in India or Euro-­Brazilians). Subaltern communities constitute the global untouchables, and yet, within each “untouchable” community there is invariably a strata that is comparatively better off – a subaltern middle class. Post-­colonial studies introduce the concept of the “colonial gaze”, the practice of members of dominant groups of looking in on and looking down on the natives. That gaze sees the characteristic features and structures of the colonized generally as deformed and deficient – albeit, often, also as exotic – when compared with parallel features and structures in the communities of the colonizers. The gaze creates a process of “nativization”, so that there are uniquely “native” businesses, “native” schools, “native” villages, and “native” arts. And, of course, there is a “native” bourgeoisie. It is straightforward, in this context, to substitute the term “primitive” for “native” and “savage” for “primitive”. All of these terms do the same work at increasing degrees of disdain and distancing. And the native bourgeoisie, itself, often directs a middle-­class sneer toward their less well-­off co-­ethnics, frequently characterizing them as dysfunctional,

96  William Darity

unmotivated, and lazy. Looking down on the less “successful” among those who share the same social identity becomes routine among many members of the subaltern middle class. This type of intratribal comparison constitutes a convenient justification for intratribal stratification, masking both the structural practices of the colonizer and the complicity of the subaltern native middle class in maintaining its colonial status. If black Americans are construed as an internal colony in the United States (see, e.g., Pinderhughes 2011), then the black middle class can be seen as a native middle class – facing all the attendant issues and inclined toward all the general tendencies. Since black Americans patently are a subordinated community, then the black middle class also can be seen as a subaltern middle class. The combination yields a condition as a subaltern native middle class (SNMC). To what extent are black middle classes globally subaltern native middle classes? What attributes are shared by all subaltern middle classes? Some potential hypotheses are embedded in the seven themes that follow. First, to what extent is the subaltern middle class consistently smaller proportionately and far less financially secure than the middle class of the dominant group? Second, to what extent does the subaltern middle class feature the “lower ladder” phenomenon across all countries? This is the circumstance described in the passage quoted above, where Frazier asserted that American blacks at the apex of their intra-­group income distribution would be, at best, in the middle of the white income distribution. Third, I hypothesize that the flow of new persons into middle-­class status will be higher in subaltern communities, hence proportionately more members of the SNMC will undergo the nouveaux riches experience, which in turn will correspond to greater efforts to display their new affluence and to establish status credibility. Perhaps more established (multi-­generation) members of the SNMC may be less overt in their displays of status. Frazier (1957: 236) observed, “The single factor that has dominated the mental outlook of the black bourgeoisie has been its obsession with the struggle for status”. Stratification economics posits, instead, that the obsession with status is universal. The wider the gulf between the relevant social groups, the more internally directed the focus of interpersonal comparisons. Fourth, because of the greater fragility of their middle-­class position, we can predict that the SNMC will experience higher rates of exit from middle-­class status (more downward mobility) than members of the dominant middle class. Indeed, the degree of economic disadvantage faced by the SNMC may mean that moving from an environment of legal segregation to one where segregation is prohibited may not have much effect on the fundamental conditions of intergroup inequality that confront subaltern native middle classes. Wealth differentials, in particular, are large between SNMCs and the dominant middle class, making it harder for SNMCs to transfer their status to the next generation. Indeed, as suggested above, relative wealth may be the most useful measure for designating particular communities as subaltern or privileged (Darity 2014). Nevertheless, SNMCs frequently accomplish more with fewer resources, particularly with respect to educational attainment.

The subaltern native middle class 97

Fifth, racialization of the position of the SNMC will lead them to celebrate any accomplishment that appears to be a signal achievement among their members, especially racial “firsts”. Correspondingly, members of the dominant group will tend to denigrate all of their accomplishments by saying that those achievements occurred under lower standards of evaluation or assessment. This is especially evident in contexts where policies like affirmative action are used to address discriminatory practices directed against subaltern peoples. Furthermore, members of the SNMC who are highly successful (e.g., Dalit millionaires in India; famous individuals like Oprah Winfrey, President Barack Obama, or LeBron James in the United States) will find their stories trumpeted widely as indicators of substantial improvement among all the members of the subaltern community. Sixth, the standard gradient between health outcomes and socio-­economic status (SES) that has outcomes improving with higher SES will not operate for SNMCs. They may even experience worse outcomes in some categories (e.g., hypertension or infant survival). At minimum, I posit that middle-­class status does not have the same payoff in terms of improved health for SNMCs as it does for members of the dominant middle class. Finally, the SNMC is in a deeply socially contradictory position between serving as a vanguard of revolutionary change and complicity with the existing social order. They are in the complicated position of being “straddlers” – caught between the colonizers and the more depressed strata of their own colonized community (Darity 2014). This may lead to certain behaviors, psychological stress and accompanying health issues, political ambiguities, and contradictory postures. Frazier (1957: 236), while asserting that the black middle class distanced themselves from the black masses, complained, “When the opportunity has been present the black bourgeoisie has exploited the negro masses as have whites”. Ironically, however, it was the sons and daughters of the 1950s black bourgeoisie who were at the forefront of the civil rights activism of the 1960s, reinforcing the contradictory character of the black middle class. With respect to the latter observation, Amilcar Cabral posed the dilemma for the native middle classes in ferocious fashion: To retain the power which national liberation puts in its hands, the petty bourgeoisie has only one path: to give free rein to its natural tendencies to become more bourgeois, to permit the development of a bureaucratic and intermediary bourgeoisie in the commercial cycle, in order to transform itself into a national pseudo-­bourgeoisie, that is to say in order to negate the revolution and necessarily ally. In order not to betray these objectives the petty bourgeoisie has only one choice: to strengthen its revolutionary consciousness, to reject the temptations of becoming more bourgeois and the natural concerns of its class mentality, to identify itself with the working classes and not to oppose the normal development of the process of revolution. This means that in order to truly fulfill the role in the national liberation struggle, the revolutionary petty bourgeoisie must be capable of committing suicide as a class in order to be reborn as revolutionary workers,

98  William Darity

completely identified with the deepest aspirations of the people to which they belong. This alternative – to betray the revolution or to commit suicide as a class – constitutes the dilemma of the petty bourgeoisie in the general framework of the national liberation struggle. (Cabral 1966, emphasis added) While Cabral’s commentary was made in the context of the anti-­colonial revolutionary movement in Africa in the 1960s, it points toward a more general problem faced by the SNMCs. Will they act in their own interest as a class, regardless of the consequences for the wider subaltern community to whom they belong, or will they dissolve their own interest as a class into pursuit of the greater well-­being of the wider subaltern community to whom they belong? This is the fundamental question for the subaltern native middle classes, and, perhaps, there will be different answers at different times and in different places.

References Cabral, A. (1966). “The Weapon of Theory” Address delivered to the first Tricontinental Congress Conference of the Peoples of Asia, Africa, and Latin America, Havana, Cuba, January. www.marxists.org/­subject/­africa/­cabral/­1966/­weapon-­theory.htm Darity, W., Jr. (2014). Race, caste, class and subalternity, Journal of Asian Studies, 73(4), November: 1085–1090. Darity, W., Jr. (2017). “A revolution of managers” Nation, 4 May. https://­socialequity. duke.edu/­news/­what-­will-­kill-­neoliberalism-­roundtable-­its-­fate Davis, J. B. (2015). Stratification economics and identity economics, Cambridge Journal of Economics, 39(5): 1215–1229. Frazier, E. F. (1939). The Negro Family in the United States, Chicago: The University of Chicago Press. Frazier, E. F. (1957). Black Bourgeoisie, New York: Free Press.Hamilton, D., Darity, W., Jr., Price, A. E., Shridharan, V., and Tippett, R. (2015). Umbrellas Don’t Make It Rain: Why Studying and Working Hard Isn’t Enough for Black Americans, Insight CCED–Duke University, and the New School. Jan, T. (2017). 1 in 7 White families are now millionaires. For Blacks it’s 1 in 50. Washington Post, October 3. Marsh, K., Darity, W., Jr., Cohen, P., Casper, L., and Salters, D. (2007). The emerging black middle class: Single and living alone. Social Forces, 86(2), December: 735–762. Melber, H. (2016). ‘Somewhere above poor but below rich’: Explorations into the species of the African middle class(es), in H. Melber (ed.), The Rise of Africa’s Middle Class: Myths, Realities and Critical Engagements. London: Zed Books, pp. 1–16. Mills, C. W. (2002 [1951]). White Collar. New York: Oxford University Press. Pinderhughes, C. (2011). Toward a new theory of internal colonialism, Socialism and Democracy, 25(1), March: 235–256. Platt, A. M. (1989). E. Franklin Frazier reconsidered, Social Justice, 16(4), Winter: 186–195. Veblen, T. (2009 [1899]). The Theory of the Leisure Class. New York: Oxford University Press. Weber, M. (1978 [1922]). Economy and Society (Two Volumes). Berkeley: University of California Press.

7 Max U, morality, and the future of economics Geoffrey M. Hodgson

Introduction Mainstream economics has gone through several major changes since the Second World War.1 There was the ‘formalist revolution’ of the 1950s (Backhouse 1998; Blaug 1999, 2003; Weintraub 2002), which increased the frequency and prestige of mathematical techniques. There was the rise and 1970s theoretical breakdown of the ‘sound micro-­foundations’ project, followed by the vigorous rehabilitation of game theory in the 1980s (Kirman 1989; Rizvi 1994a, 1994b). By the 1990s, approaches in experimental economics (Smith 1992) and behavioural economics (Camerer et al. 2004) were well-­established in leading journals and prestigious departments of economics. Since then we have seen neuro-­economics (Glimcher 2003; Zak 2004; Camerer et al. 2005), happiness economics (Di Tella and MacCulloch 2006), identity economics (Akerlof and Kranton 2010), and much else. There has been a broad empirical turn in economics since the 1980s, leading to a higher preponderance of empirically grounded or applied papers in leading journals (Hamermesh 2013; Backhouse and Cherrier 2017). Finally, the old idea of ‘economic man’ as entirely greedy and self-­regarding has been challenged from within economics itself, by research in ‘other-­regarding’ and ‘social’ preferences, including discussions of altruism (Becker 1976b; Charness and Rabin 2002; Fehr and Fischbacher 2002; Bowles and Gintis 2011). The pace of change and growing diversity of the discipline are impressive. The rise of behavioural economics and other developments have made it possible to question some basic assumptions, when beforehand to raise such doubts was to exclude oneself from the discipline. These shifts pose greater challenges for critics and dissenters who have to fire at multiple, fast-­moving targets. Taking these various changes into account, the careful appraisals and judicious and informative interventions by John Davis (1995, 2003, 2007) in multiple sub-­ fields become even more impressive. There has been an important debate about the nature, extent, and evolution of the discipline since 1990 (Colander 2000a, 2000b, 2005a, 2005b; Colander et al. 2004; Davis 2006, 2008). A number of authors have argued

100  Geoffrey M. Hodgson

that the ‘neoclassical’ paradigm – involving its core concepts of rationality and ­equilibrium – is in decline within mainstream economics, and it has been replaced by a variety of different approaches. David Colander’s analysis of the discipline was based on interviews with graduate students in the most esteemed departments of economics, principally in the United States. He announced the ‘death of neoclassical economics’ as a useful label (Colander 2000a) and argued (Colander et al. 2004: 485; Colander 2005b: 930) that mainstream economics was ‘moving away’ from its ‘holy trinity’ assumptions of ‘rationality, selfishness and equilibrium’. There have been quarrels over whether or not neoclassical is a useful or appropriate term, or whether or not its current usage corresponds to that of Veblen, who coined it in 1900 (Aspromourgos 1986; Fayazmanesh 1998; Colander 2000a; Lawson 2013; Morgan 2016). A problem with the neoclassical label is that it might misleadingly suggest that the classical economists formed a coherent school or approach and that neoclassical economics is a new version of that older school. As Joseph Schumpeter (1954: 919) observed: ‘there is no more sense in calling the Jevons-­Menger-­Walras theory neoclassic than there would be in calling the Einstein theory neo-­Newtonian’. Another point against the use of the neoclassical label is that it is widely and wrongly assumed (often by critics and sometimes by exponents) that the economics to which it refers automatically leads to pro-­market policies.2 Yet leading proponents of ‘neoclassical’ economics with utility-­ maximizing agents and equilibrium models, including Oskar Lange, Kenneth Arrow, and John Roemer, favoured a socialist economy dominated by common ownership. Léon Walras called himself a ‘scientific socialist’ and advocated price regulation and the public ownership of natural monopolies, including of land. The focus on the utility-­maximizing individual does not necessarily imply market policies. Models with utility-­maximizing individuals have been developed in non-­market contexts. Indeed, mainstream economics has been inadequate in its grasp of basic trading institutions, such as private property and the market (Hodgson 1999: chapter 2; 2015b). It is too charitable to describe neoclassical economics as the economics of the market, as such analysis is deficient in its analysis of market institutions. Given this controversy and confusion, there are good reasons to adopt Deirdre McCloskey’s (2008, 2016) inspired ‘Max U’ label instead (except when neoclassical appears in quotes by others). Max U refers to the core ideas of utility maximization and equilibrium. Another prominent term is ‘the economic approach’. Hence Gary Becker (1976a: 5, 14) wrote: The combined assumptions of maximizing behavior, market equilibrium, and stable preferences, used relentlessly and unflinchingly, form the heart of the economic approach as I see it . . . all human behavior can be viewed as involving participants who maximize their utility from a stable set of preferences.3

Max U, morality, and the future of economics 101

Whatever the label, the key issue here is the theoretical core. Colander rightly highlighted the core assumptions of ‘rationality, selfishness and equilibrium’. He argued that mainstream economics was ‘moving away’ from these assumptions. Colander (2000a: 135–136) made even bolder (and highly questionable) assertions: ‘Few modern economists today accept utilitarianism; most see it as a quaint aspect of the past. One sees very little operational use of utility theory in modern economics’. Also, ‘general equilibrium models are seldom used’. In the following section, I contest the claim that these core ideas have been largely discarded, even if they are rivalled by other developments. But I agree with some important aspects of Colander’s argument. In particular, the shift towards game theory has radically changed microeconomics, the relaxations of the assumptions of perfect information and omniscience have been dramatic, and there has been a strong turn towards greater use of empirical data, partly driven by the emergence of new databases and the rapid advance of computer technology. But while economics has continued to innovate and change, the accounts by Colander and Davis overlook important elements of theoretical continuity at the core. I argue below that the degree of continuity of these core elements has been underestimated. Colander (2000a: 137) emphasized (with approval) that ‘the modelling approach to problems is the central element of modern economics’. But with disapproval, Tony Lawson (1997, 2006, 2013) famously argued that mathematical model-­building in economics is inappropriate for the ‘open systems’ that economics typically addresses. Both authors proposed that the prevalence of modelling in economics has eclipsed any other major continuity at the theoretical core. I dispute this shared conclusion, while fully accepting that model-­building (for good or ill) is well-­entrenched in economics. It is also evident that the persistence of particular kinds of model-­building have been facilitated by the enduring assumption of Max U. If these preliminary indications and arguments in this chapter hold water, then they have major implications, including for the efforts by Amartya Sen (1977, 1985a, 1985b, 2004), John Davis (1995, 2003), Deirdre McCloskey (2006, 2008, 2016), and others to promote a more nuanced and multifaceted picture of the economic agent, at least within the academic boundaries of economics as a discipline. They suggest more pessimistic conclusions concerning the possibility of changing the core assumptions of economics and thereby establishing a new paradigm (Davis 2008; Earl 2010).

The survival of Max U Colander (2005b) rightly identified the core of microeconomics as the zone where any fundamental change would occur in the discipline as a whole. Microeconomics sustains basic assumptions about individuals, their motivation, and their modes of interaction with others. Supporting his thesis that Max U

102  Geoffrey M. Hodgson

assumptions had already been largely abandoned, he claimed to identify a shift toward Herbert Simon’s version of behavioural economics. Here, according to his argument, we should expect further signs of development: Herbert Simon’s work . . . provides a map for the direction in which I am predicting economics will evolve. . . . Simon was neglected. . . . Today, however, his work and approach are beginning to be considered more carefully, as the profession catches up with his vision of the way in which one can understand the economy. (Colander 2005b: 930) But this overlooked the radical transformation of ‘behavioural economics’ itself. The original version of Simon (1957, 1959) adopted insights from psychology and subverted the notion of rational, optimizing, behaviour. By contrast, mainstream versions of ‘behavioural economics’ since 1990 are typically seen as modifications in the orbit of ever-­tenacious Max U. Herbert Simon was awarded the Nobel Prize in Economics in 1978 for his work in this area. But ‘behavioural economics’ did not take off in the highly ranked economics journals and departments until about twenty years later, and only after major modifications. Simon’s radical message had to be toned down and some of his major insights jettisoned (Sent 2004). Although Matthew Rabin (1998) announced the return of psychological insights into economics – after decades of shunning them through ‘as-­if’ reasoning – the ultimate acceptance of psychological ideas within economics proved to be highly limited. As Nathan Berg and Gerd Gigerenzer (2010) pointed out: For a research program that counts improved empirical realism among its primary goals, it is startling that behavioral economics appears, in many cases, indistinguishable from neoclassical economics in its reliance on as-­if arguments to justify ‘psychological’ models that make no pretense of even attempting to describe the psychological processes that underlie human decision making. (2010: 133) Sheila Dow (2011) considered in particular the highly limited use of psychology by behavioural finance, which misses some of the major psychological drivers of financial volatility. Peter Earl (2010) commented more generally: There is no systematic attempt to bring psychology into economics; instead, constrained optimization is modified to allow for preferences and/­or perceptions to be distorted by ‘heuristics and biases’ uncovered in empirical work. . . . Simon’s approach remains almost completely invisible within the new literature. (2010: 216–217)

Max U, morality, and the future of economics 103

As a witness, Earl (2010: 317) called the 740-­page reader on behavioural economics edited by Colin Camerer et al. (2004). Earl noted that within this hefty volume, Simon is mentioned only four times: the first three refer to his notion of ‘procedural rationality’ with no references. Only the last appearance cites any of his work and makes any connection with his concept of satisficing (for which there is no index entry). Bounded rationality appears on only three pages (all in the same paper). Camerer and his colleagues made it clear that the new behavioural economics was an adjustment to the reigning Max U paradigm rather than an attempt to overthrow it. Camerer et al. (2004: 1) insisted that behavioural economics ‘does not imply a wholesale rejection of the neoclassical approach to economics based on utility maximization, equilibrium, and efficiency’. Accordingly, Daniel Kahneman (2003: 1469) wrote: ‘Theories in behavioural economics have generally retained the basic architecture of the rational model, adding assumptions about cognitive limitations designed to account for specific anomalies’. The massive text on behavioural economics by Sanjit Dhami (2016) is a tour de force, and it mentions Simon’s work several times. The concept of satisficing appears on six of its seventeen hundred pages. But the text mostly replicates current practice of using utility maximization as a reference point. And there is no mention at all of the behavioural theory of the firm developed by Richard M. Cyert and James G. March (1963), which continues to receive attention from academics in business schools. One of the key offshoots of behavioural economics is the theory of ‘libertarian paternalism’, where key information is presented in a way to ‘nudge’ people to make more efficacious decisions (Thaler and Sunstein 2008). In a widespread interpretation, the ‘nudge’ is designed to help people to maximize their utility and overcome cognitive biases or other ‘errors’. Max U serves as the normative standard of behaviour, while deviations from that norm have to be ‘corrected’. Overall, there is little evidence that the new ‘behavioural economics’ has passed Colander’s (2005b: 930) ‘Herbert Simon Test’. So far, there is little evidence that ‘his work and approach are beginning to be considered more carefully’ or the profession are beginning to catch up ‘with his vision of the way in which one can understand the economy’. Crucially, Simon’s core concept of satisficing behaviour – which he adapted from psychology and administrative science, and which he regarded as more realistic than maximization or optimization, rarely appears in the new version of behaviouralism. Indeed, Max U has reconquered some territory conceded in the initial bubble of enthusiasm for the new behaviouralism. While promoting more sophisticated and accommodating Max U formulations, Herbert Gintis (2005) showed that experimental evidence by behavioural economists and psychologists that seemed to challenge utility-­maximizing rationality, does not in fact refute it, once more sophisticated (typically ‘social’ or ‘other-­regarding’) preference functions are introduced.

104  Geoffrey M. Hodgson

To see why Max U is capable of conquering unlimited behavioural territory, we need to appreciate that utility maximization is consistent with any data concerning the behaviour of any organism or machine. In other words, utility functions can be modified to fit data on any manifest behaviour. Particular utility or preference functions may be falsified because they are unable to predict some observed behaviour, but Max U in general terms is unfalsifiable (Steele 2014). Despite being repeated by leading authors, this argument is still insufficiently acknowledged. When the young Paul Samuelson (1937: 156) discussed utility maximization, he understood that ‘all types of observable behavior might conceivably result from such an assumption’. Many experiments show that people sometimes do not maximize monetary rewards. But that does not show that they are not maximizing their utility. Because utility is unobservable, all kinds of behaviour can be ‘expressed’ in terms of utility maximization, without fear of refutation. As Sidney Winter (1964: 309, 315), Lawrence Boland (1981) and others have argued, no evidence can possibly refute the theory that agents are maximizing some hidden or unknown variable (such as utility).4 Many experiments show that participants do not maximize expected monetary (or other visible) payoffs. But payoff maximization is not the same as utility maximization. Although it is often assumed that payoffs and utility are monotonically related, experimental economists such as Vernon Smith (1982: 929) admitted that this monotonic correspondence cannot be guaranteed, partly because preferences and utility are ‘not directly observable’ (Siakantaris 2000). Consequently, if we have evidence that payoffs are not being maximized, then this does not mean that utility is not being maximized. If experiments show that some consumers appear to prefer a monetary reward that is less than the expected outcome, or appear to have intransitive preference orderings, or defy the Von Neumann–Morgenstern independence axiom, then we can always get round these problems, and make the evidence consistent with utility maximization, by introducing other variables or states of the world into the utility function. For example, apparent preference inconsistency can be explained away by the fact that the apparently inconsistent choices are always made at different times and in (at least slightly) different circumstances.5 Given that we can never in principle demonstrate that some unobserved variable (like utility) is not being maximized, then an accommodating Max U is invulnerable to any empirical attack. The general core of expected utility theory is unfalsifiable. But it is not a tautology (in the logical sense) because it is conceivably false. Logical tautologies and unfalsifiable propositions are often confused, but they are quite different. Logical tautologies are true by logical inference or definition. By contrast, even if we cannot observe it, it might really be the case that utility exists, and it may be possible that an individual is not maximizing it. But no empirical evidence can ever establish this for certain. This does not necessarily mean that the utility-­maximization framework is useless or wrong. Neither tautological nor non-­falsifiable statements are necessarily

Max U, morality, and the future of economics 105

meaningless or unscientific.6 They can play a framing or organizing role in scientific thought, in relation to which falsifiable propositions are proposed and tested. In economics, Max U has played a prominent role in these terms, just as there are other key frameworks that structure enquiry in other sciences. But the unfalsifiable nature of Max U in general is rarely admitted. Is this truth avoided because of worries that it might risk Popperian accusations that Max U is not scientific? Yet Karl Popper himself did not require all science to pass the falsifiability test (Ackerman 1976: 30–31). Or is the truth eluded because it would undermine the claim that Max U leads to specific testable propositions? Frameworks that are unfalsifiable can predict a huge range of possibilities, because they can be made compatible with all of them. In fact, when predictions are made with Max U models, it is the auxiliary assumptions that constrain possibilities and do much of the predictive work. As Mark Blaug (1992: 232) put it: ‘The rationality hypothesis by itself is rather weak. To make it yield interesting implications, we need to add auxiliary assumptions’. In a critique of Gary Becker’s claim that Max U yielded useful predictions, Robert Pollak (2003) showed that Becker’s predictions always depended on auxiliary assumptions, which constrained the otherwise-­ unbounded possibilities for prediction with Max U. It seems that to provide it with predictive credentials, news of the unfalsifiability of Max U has to be suppressed. Consequently, devotees of Max U can set about their tasks of showing that apparently ‘anomalous’ results can be understood within a Max U framework, either as deviations from a particular Max U model or via complete incorporation in a more sophisticated Max U model that ‘explains’ the deviations as well. Camerer (2003: 101) explained the aim and practice of behavioural economics as ‘to find parsimonious utility functions, supported by psychological intuition, that are general enough to explain many phenomena in one fell swoop, and also make new predictions’. The parsimony criterion makes the task more difficult and the successful fitting of function to data more impressive. Then predictions are made to test the creation. But what is being tested here is not Max U in general but particular and parsimonious formulations of a utility function. If the test fails, then Max U is not abandoned, but the search is triggered to find another tolerably parsimonious function to fit the augmented data. And so it goes on. While the new behavioural economics thrived, the gravitational pull of Max U did not allow any escape from its orbit. If anything, the Max U force field has become more powerful in recent years. Gerardo Infante et al. (2016) documented how some behavioural economists have claimed to have rediscovered an inner Homo economicus within everyone. Deviations from this inner rationality are put down to mistakes, to misinterpretations of evidence, or to missing information. Max U eventually returns, once his or her mistakes and misperceptions are corrected. Max U has exerted influence over neuro-­economics as well. As soon as it became possible to garner extensive data on brain activity, pioneering

106  Geoffrey M. Hodgson

neuro-­economists (Platt and Glimcher 1999; Glimcher et al. 2005) claimed to discover the utility function exists as a physiological reality inside the brain. This prompted Gintis (2009: 2) to write: ‘Neuroscientists increasingly find that an aggregate decision process in the brain synthesizes all available information into a single unitary value’. Indeed, if utility could be directly measured in the brain, then utility would become observable and Max U would be falsifiable. But Jack Vromen (2010) argued that at best the neurological evidence exhibits consistency with the predictions of expected utility theory. There is no evidence of actual computation of utility in the brain. Given the argument here that any observed outcomes can be made consistent with some utility function, the consistency claim is hardly powerful or surprising. But the existence claims are unsupported. The whole argument shows that Max U still preoccupies influential economists. If they had abandoned utility functions, then they would not be looking for them. Note also the enduring use of the term ‘economic approach’ by mainstream economists, including in areas where there are contested modes of explanation, such as in studies of business organizations, legal systems, cultures, or even animal behaviour. This claimed ‘economic approach’ typically involves a focus on maximizing behaviour by individuals, taking into account costs and benefits. The persistence and specific use of this term suggests that numerous mainstream exponents still perceived that economics is defined by some Max U core. Max U is so capacious that it goes beyond the parameters of human decision. Experimental work with rats and other animals (Kagel et al. 1995) ‘revealed’ that animals have downward-­ sloping demand curves, supposedly just like humans. Becker (1991: 307) proposed that ‘economic analysis is a powerful tool not only in understanding human behavior but also in understanding the behavior of other species’. Similarly, Gordon Tullock (1994) claimed that organisms can be treated as if they have the same general type of utility function that is attributed to humans in the microeconomics textbooks. Seemingly, we now have ‘evidence’ of the ‘rationality’ of everything in evolution from the amoeba onwards. As a consequence, Max U cannot express what is specific to human nature, human motivation, or human society. Its very weakness, at least when applied to the human domain, stems from its excessive scope. This does not mean that economic theory as a whole has not changed. For example, the shift from general equilibrium to game theory in microeconomics has had a major impact on the discipline. As Davis (2008: 358) put it, game theory rejects the formerly standard ‘competitive model of indirect interaction solely through the price mechanism to focus on scenario-­driven direct interaction’. In some versions of game theory the notion of rationality becomes difficult to specify (Sugden 1991). Actors in agent-­based models and prominent versions of evolutionary game theory are formulated in terms of strategic algorithms rather than in conventional Max U terms. Thus Max U does not prevail over all contemporary economic theory. But game theory begets multiple equilibria, and agent-­based models illustrate multiple, often

Max U, morality, and the future of economics 107

radically different outcomes. Max U provides a less ambiguous and more secure theoretical world. Consequently it still prevails, and rumours of the death of Max U are grossly exaggerated. Matteo Richiardi and Roberto Leombruni (2005) pointed out that agent-­ based models are very rare in twenty highly rated journals in economics. They made up only 0.03% of published articles in these journals since 1988. Richiardi and Leombruni reported common arguments against agent-­based ­models – that they are often difficult to generalize, interpret, or estimate. Some high-­priesthood resistance to agent-­based modelling may also stem from its non-­reliance on Max U formulations. The growth of agent-­based models will be more significant for economics when it has achieved greater penetration into its top journals. Colander (2000a: 136) is also wrong to say that ‘general equilibrium models are seldom used’. Dynamic stochastic general equilibrium (DSGE) models (with their core Max U assumptions) are still widespread in macroeconomics. So too are the rational expectations and efficient markets hypotheses. These ideas persist even after the economic crash of 2008, when they were subjected to severe public criticism. As Alan Kirman (2009) noted: Thus both the development of the DSGE model and the evolution of the efficient market hypothesis share a common feature – despite the empirical evidence and despite their theoretical weaknesses, their development proceeded as if the criticism did not exist. Justin Fox (2009: 310) similarly remarked: ‘While behavioralists and other critics poked a lot of holes in the edifice of rational market finance, they haven’t been willing to abandon that edifice’. Max U has survived the Great Crash of 2008, and he still prospers in macroeconomics as well as in microeconomics. Colander was involved in two major surveys and sets of interviews, in 1985 and 2004, with graduate students in seven top-­ranking graduate economics programmes: the University of Chicago, Columbia, Harvard, Massachusetts Institute of Technology, Stanford, Yale, and Princeton. Similar questions were asked on both occasions to gauge how the discipline had changed. But some of Colander’s data undermine his own claims. Colander’s (2005a: 188, table 5) data show that those graduate students regarding the assumption of rational behaviour as ‘very important’ remained on average at 51% from 1985 to 2004. Those seeing it as ‘important in some cases’ increased on average from 41% to 43%. On the rational expectations hypothesis, those seeing it as at least ‘important in some cases’ increased on average from 70% to 83%. Colander’s own data show that, if anything, Max U had become more entrenched over the nineteen years between the two surveys (although the frequencies and changes are radically different at different leading universities). Figure 7.1 shows bibliometric data that are consistent with the proposition that Max U is still alive and influential. Using the JSTOR database, some key

108  Geoffrey M. Hodgson

70%

Frequencies of Key Terms

60% 50% 40% 30% 20%

ulity

maximi*

equilibrium

raonal*

10% 0% 1960–64 1965–69 1970–74 1975–79 1980–84 1985–89 1990–94 1995–99 2000–04 2005–09 2010–14

Figure 7.1  Frequencies of key terms in leading economics journals

words were searched in ten leading journals of economics from 1960 to 2014.7 The vertical axis shows the percentages of articles in any given period in which any one of the key terms appeared. In two cases the wild card (*) allowed for variations (maximise, maximize, maximization, etc.). Figure 7.1 shows that, contrary to rumours that Max U was on the way out, the use of terms such as utility, maximization, equilibrium, and rationality have increased in frequency from the 1960s to the most recent period. All four terms reach their highest frequencies in 2005–2009 or 2010–2014. There is no sign of any impact of the Great Crash of 2008 on the tenacity of Max U. Of course, some of the appearances of these terms may have been in criticism rather than by adoption, but they remain strongly on the agenda nevertheless. Furthermore, there is little sign of alternative terms making headway against them. Searches for ‘satisfic*’ found appearances below 1% frequency for the entire 1960–2014 period. Searches for (‘bounded rationality’ OR ‘boundedly rational’) never got above a frequency of 4%. The data in Figure 7.1 suggest that the Formalist Revolution of the 1950s led to a consolidation of the hard-­core assumptions of Max U, despite ongoing technical and formal innovations, such as game theory. Max U has survived the 2008 crisis without any wounds and seems alive and well. As Jason Potts (2000) has explained, Max U has often carried additional baggage. It has come with a ‘field’ ontology, where infinite adjustment of all or many variables and gross substitutability between them are possible. This allows, of course, the use of calculus. But game theory sheds this baggage, assuming a structured world without gross substitutability. Nevertheless, for much (but not all) contemporary game theory, Max U remains. Hence Max U covers a wider domain than the assumption of gross substitutability.

Max U, morality, and the future of economics 109

Will Max U be replaced? John Davis (2006: 2) rightly warned us that ‘the question of change in any science or system of ideas is an extremely complicated one, and that historians and philosophers of science are hardly in agreement over what constitute criteria for identifying change’. Among other outcomes, Davis (2006) considered the possibility of a ‘maturity’ scenario where ‘neoclassical’ economics persists but is no longer dominant. There may be something in his suggestion that Max U economics has completed its major contributions and theoretical returns to this paradigm are diminishing. But the case of behavioural economics gravitating back to Max U suggests that Max U survives not simply because it is a relic from the past. Furthermore, the limited bibliographic evidence in Figure 7.1 suggests that Max U is as dominant as it has ever been. Davis (2008) addressed the possibility of a shift from the current ‘pluralism’ toward a new orthodoxy. Again, the degree of persistence of Max U may have been underestimated, leading to some overstatement of the extent of current pluralism and an overestimation of the possibilities of the rise of a new orthodoxy. Like Colander, Davis pointed to behavioural economics as a harbinger of possible change. Like Colander, Davis seemed to have underestimated the degree to which behavioural economics has retained Max U, including as an enduring measuring rod to detect ‘deviations’ from it. To set the scene for a possible new orthodoxy, Davis (2008: 353) pointed to long-­run historic changes in the reigning paradigm of economics: ‘For example, classical economics was preceded by physiocracy and mercantilism, and succeeded by neoclassicism’. Is a major shift toward a new paradigm already underway and gaining ground in the twenty-­first century? I do not rule out this possibility, but I wish to bring another major factor into the argument: the professionalization of the economics in the modern research university. In the Anglophone world, this process of professionalization in modern, research-­oriented universities was underway by about 1880 and accelerated in the twentieth century. In economics, the top departments and journals became established and recognized (Maloney 1985; Coats 1993; Kadish and Tribe 1993; Fourcade 2006). Remarkably, for the entire period since 1880, despite enormous other changes, Max U has endured and dominated the core. It is possible that, as economics became professionalized and institutionalized, Max U served as its lodestone. The discipline defined itself in terms of Max U and his utilitarian trappings more than anything else. Even if homage to Max U becomes little more than ritual, it could remain totemistic for the discipline for the conceivable future. Crucially, Max U provides a framework for the organization and operation of particular kinds of mathematical formalism in economics. If this account of Max U’s defining function is true, then the chances of changing economics from within diminish significantly. Furthermore, given the strong additional empirical turn in the discipline, it may become even more difficult to challenge the Max U theoretical core. The

110  Geoffrey M. Hodgson

unfalsifiability of Max U makes it invulnerable to evidence. Empirical work thus has a steep uphill task trying to take on the Max U theoretical core as well. Earl (2010: 221) pointed out that ‘if empirical work is based on unfamiliar economic theory it will be disadvantaged, if only because papers may need to be far longer in order to introduce the theory to referees and readers’. To get the empirical study published, it is often best not to challenge hard-­core theoretical assumptions as well. Ultimately the empirical turn in economics may have made Max U stronger – at least in some ceremonial and discipline-­defining sense – rather than weaker. The increasing internal specialism within economics also frustrates challenges to any hard-­core assumptions or ceremonial shibboleths (Cedrini and Fontana 2018). The greater and deeper the specialisms, the more difficult it is to broaden the argument and shift the discourse away from the silos and toward general principles. A prominent account of the evolution of a science is that a theoretical core becomes weakened when accumulating data and failures of explanation stretch the reigning paradigm to its limits. Anomalies accumulate to the point where normal science is undermined, creating the conditions for a ‘scientific revolution’ (Kuhn 1962). The Darwinian revolution in biology, the Copernican revolution in cosmology, and the Newtonian and Einsteinian revolutions in physics are given as examples. But the non-­falsifiability of Max U at the core of mainstream economics places this discipline in a position different from others. Anomalies appear, not as a threat but as a challenge for those who will adapt preference functions in ever more complex ways to fit the data. They then declare in triumph that Max U lives on. With suitable ingenuity this always can be done. Neither utility nor preferences are directly observable, and they themselves are not directly subject to empirical scrutiny. I can think of no scientific paradigm in any other discipline that is largely defined in terms of unfalsifiable claims, and of no revolution that overthrew a scientific core that was largely unfalsifiable.8 In this respect, the defining characteristic of Max U economics seems unique. Its unfalsifiability is another reason why Max U lives on, and why a Kuhnian scientific revolution in economics is unlikely, at least in the foreseeable future. Consequently, while economics undeniably makes progress, relatively little advance is made by empirical interrogation of the Max U assumptions at the hard core. The enormous empirical challenges mounted by experimental and behavioural economics since the 1980s have led to major positive developments but have ultimately left the Max U hard core intact. The ‘anomalies’ have disappeared through adjustment of the utility functions, or their treatment as temporary deviations from the Max U norm. This is quite unlike other sciences. It had long been observed that the planet Mercury had a solar orbit that deviated from that predicted by Newtonian equations. This alone was not enough to dethrone Newtonian physics. Change began when it was shown that Albert Einstein’s alternative theory accurately

Max U, morality, and the future of economics 111

predicted the orbit of Mercury. This in turn led to Arthur Eddington performing a famous experiment in 1919 to test Einstein’s prediction that light rays were bent by the sun. Paradigm change was fuelled by empirical refutation of an old theoretical system when combined with an empirical confirmation of a new one. In economics there is no such dance of conjecture and refutation around its enduring Max U core. If at all, radical change at the heart of economics is more likely to come slowly, as accumulating empirical work and context-­ specific theory are gradually acknowledged as more important than any ceremonial attachment to Max U. The unfalsifiability and infinite plasticity of its core idea might eventually become more widely understood, and Max U might be seen less as triumphal vindication of disciplinary machismo and more as ceremonial and dispensable ritual. A deeper desire to understand human motivation and psychology may begin to eclipse ritual efforts to represent all human behaviour in Max U terms. Then things will change. But there is little sign of that happening yet. Despite the shock of the biggest global financial crisis since the 1930s, these questions are not yet asked. The current shift toward ‘complexity economics’ (Colander 2000c, 2000d; Gallegati and Kirman 2012) might move the discipline further away from simplistic formulations of interactive systems, heroic assumptions of representative agents, and much else. Much of this work may use agent-­based models and other devices that are consistent with Max U but not necessarily derived from it. But the evidence from Richiardi and Leombruni (2005) cited above showed that agent-­based models had made little penetration of the top-­rank journals. Some penetration of the citadels of the discipline may happen, but this process of change may be slow, and it does not necessarily follow that Max U will be removed from the high altar.

Bringing real humans back in The unfalsifiability of Max U sustains an epistemic critique. But it does not clinch the matter. One has also to consider the theoretical limitations of this stance. Max U falls down for at least two further reasons. First, it sidelines the problem of explaining the causes of behaviour. Second, it fudges the question of the individual development of capacities and dispositions. Both of these are vital in policy terms: they affect the design of incentives and popular education in aspects of public policy. They also affect the theoretical foundations of welfare economics (Steedman 1980). Max U is ‘good’ at prediction because, for reasons given above, it can predict any conceivable behaviour. But Max U falls down when it comes to explaining behaviour. As Ronald Coase (1977: 488) remarked: ‘To say that people maximize utility tells us nothing about the purposes for which they engage in economic activity and leaves us without any insight into why people do what they do’. Relatedly, Amartya Sen (1977: 325) pointed to the circularity of explaining behaviour ‘in terms of preferences, which are in turn

112  Geoffrey M. Hodgson

defined only by behaviour’. Sen (1987b: 73) noted elsewhere that the description of choices in terms of utility ‘does not give any independent evidence on what the person is aiming to do or trying to achieve’. This is why the use of utility maximization as a mere summary of behaviour is inadequate. Fitting utility functions to behavioural data does not amount to an explanation of motivation or action. Max U offers an ex-­post rationalization or ‘expression’ of behaviour, rather than an empirically grounded causal explanation. A utility function may serve a limited purpose as a formalized preference ordering. Such formal constructions can have some benefits in some contexts. They can be useful shortcuts for modelling or explanatory purposes. But they do not enhance our understanding of human motivation. The lack of an adequate explanation of human motivation, behaviour, and development diminishes the utility of Max U for policy design. If human motivation is irreducible to one variable, such as money, happiness or utility, then policy design should also focus on multiple incentives, including an appreciation of their possible interdependence and mutual development. If human motivation is multi-­dimensional and multi-­modular and driven by multiple, potentially conflicting habits (James 1890; Veblen 1914; Dewey 1922; Plotkin 1994; Hodgson 2004, 2010), then particular motivations that are triggered in particular circumstances must be identified. Hopefully, this point will become clearer when we discuss the issue of moral motivation shortly. Inspired by Adam Smith (1759), Nobel Laureate Amartya Sen (1977, 1985a, 1985b, 1987a, 1987b) argued that human agents are not purely self-­ interested but have a measure of sympathy for others. Sen also stresses the concept of commitment, defined as a practical reason for an action that is independent of any gains or losses for that actor. Sen (2002, 2004) later added the concept of identity to his multifaceted view of human nature. We choose an identity to make sense of our actions and to project a persona upon others. This cue was followed up by Davis (2007), Akerlof and Kranton (2010), and several others. For Sen, the reduction of human motivation to utility maximization neglects crucial issues surrounding individual capabilities, agency, and choice. The experimental economist Vernon Smith has become increasingly persuaded that the new behaviouralist game of fitting utility functions to data avoids the subtleties of human personality, including those revealed by Adam Smith (1759), who ‘saw the individual as not even defined except in a social context’ (Smith 2013: 6). By contrast, the new behaviouralists attempt to force relational phenomena between individuals into preference functions attributed to separable agents. Rejecting this, Vernon Smith wrote: In my view, utilitarianism cannot be rescued by adding arguments to the utility function without undermining human sociality as relationship. . . . Preferences cannot be ‘social’ (or ‘pro-­social’) because relationship processes matter, with judgment of conduct emerging from the context, and

Max U, morality, and the future of economics 113

from its propriety and ecological fitness with the order of rules defining social conventions. (2013: 4–5) For similar reasons, also addressing experimental evidence, Bart Wilson (2010) argued that ‘social preferences aren’t preferences’. Explanations of human behaviour cannot be devolved upon individuals (with or without preferences) alone. What matters too is social relations, and the rules embodied in social institutions. For the purposes of this chapter I focus on our capacities for moral feelings and judgements. As Adam Smith emphasized, these are important components of human motivation. To repeat: behaviour guided by moral motivations can be subsumed under and ‘predicted by’ Max U. But Max U fails to recognize the independent power and irreducibility of moral motivation, hence it offers an impoverished theoretical explanation and weakened policy guidance. Moral motivation is not necessarily the same as altruism or other-­regarding preferences. We may give to others, or show concern for others, for reasons other than morality. Furthermore, the Max U preference functions that accommodate altruism or other-­regarding behaviour still treat individuals as if they are maximizing their own utility, and for this reason they may be regarded as selfish too. By breaking from Max U, more genuine notions of moral motivation are able to overcome this limitation. Many controversies divide moral philosophers, including over the nature of a moral judgement. The best we can do here is to select a few prominent descriptions of the nature of moral judgment. The leading moral philosopher Richard M. Hare (1952) argued that morality was subject to reason and one cannot hold contradictory ethical judgments. As John L. Mackie put it in his classic account, a moral judgment is not purely descriptive, certainly not inert, but something that involves a call for action or for the refraining from action, and one that is absolute, not contingent upon any desire or preference or policy or choice, his own or anyone else’s. (1977: 33) In his impressive philosophical account of the Evolution of Morality, Richard Joyce (2006: 70) argued on the basis of considerations in the philosophical literature that morality has most or all of the following characteristics: 1 Moral judgments express attitudes (such as approval or contempt) and also express beliefs 2 The emotion of guilt is an important mechanism for regulating moral conduct 3 Moral judgments transcend the interests or ends of those concerned 4 Moral judgments imply notions of desert and justice

114  Geoffrey M. Hodgson

5 Moral judgments are inescapable 6 Moral judgments transcend human conventions 7 Moral judgments govern interpersonal relations and counter self-­regarding individualism. These characteristics do not establish a valid morality; they instead help us to identify what is a moral judgment, whether acceptable or otherwise. The argument here relies on descriptive rather than normative ethics: there is no attempt here to identify the ‘right’ morality but instead to identify the basic nature of a moral claim. Most religions uphold moral claims, but that does not make them all right or just. Like others, Joyce emphasized the role of the emotions as well as deliberation in making moral judgements. Joyce’s point (1) establishes that a moral judgment must involve both beliefs and sentiments and is not reducible to either alone. If an action is impelled purely by emotion and sentiment, then it cannot amount to moral motivation. Deliberations and beliefs are also vital, but are themselves insufficient because they must be backed by sentiments or emotions: acting morally is more than calculated conformity to moral rules. Moral judgments may be rationalized in various ways, but they are more than matters of reason or logic. Defiance of shared moral rules in a group is often met with emotional hostility. Conformity to them may sometimes bring an emotional glow. The emotional dimension of moral rules plays an important role in their evolution and their survival. Guilt (point (2)) is a particularly important emotion that sometimes emerges after breaches of moral rules, and it too plays a part in the evolutionary process. Joyce’s points (3) through (7) reveal the limitations of typical Max U approaches. Moral judgments are not simply expressions of an individual’s interests, preferences, sentiments, or beliefs. They are also universal claims, which are deemed to have force irrespective of the interests, preferences, sentiments, or beliefs of those to whom they are supposed to apply. As both Mackie and Joyce insisted, morality surpasses questions of preference. It is a matter of right or wrong, or of duty, of ‘doing the right thing’, irrespective of whether we like it or not. This is part of what makes us human: we are capable of considering moral rules and understanding that their observance is more than a matter of personal whim or satisfaction. As Charles Darwin (1871) pointed out, no other species has a developed morality, and its unique evolution is a vital matter for scientific research (Hodgson 2014). But Max U covers all species, and the genuinely moral dimension is missing in much of economics. Moral values are either ignored or subsumed under matters of utility or preference. This has important consequences for policy design (Hodgson 2013). For example, in creating policies for taxation, for health provision, or to deal with climate change, it is (possibly necessary but) insufficient to rely on pecuniary incentives alone. It is also vital to make a moral appeal, concerning the need for social solidarity and our duty to care for others, including for future generations. Such moral appeals transcend questions of individual benefit. They

Max U, morality, and the future of economics 115

speak to our senses of duty, solidarity, and nobility, rather than simply offering a ‘warm glow’ of extra utility, derived by making a donation and feeling satisfied. There is a fundamental difference between moral rules and other (normative) rules. ‘Murder is wrong’ does not carry the same connotations as ‘splitting infinitives in English is wrong’ or ‘in Britain, one must drive on the left side of the road’. Linguistic and traffic rules are matters of convention or local convenience. Murder, by contrast, is more than a breach of convention. Threat of punishment or respect for the law are each insufficient to explain the relatively low frequency of murder. Most of us abstain from murder not simply because the probability of severe punishment outweighs any expected benefit. Most of us refrain from murder because we believe that it is morally wrong; we would desist even if we lived in a country where murder went unpunished. Apart from a minority that may be morally insensitive, most people obey laws, when we perceive them as legitimate, for moral rather than instrumental reasons (Tyler 1990). It is a commonplace observation that what may be a moral rule for one culture may not be so for another. But this does not mean that moral rules are reducible to conventions. They become moral rules because many people believe in them as such, and they jointly uphold them as more than matters of convenience, self-­interest, or convention. The cultural specificity of some moral judgments does not justify a normative moral relativism, where one person’s morality is deemed as good as any other. In contrast to standard utilitarian approaches, a moral judgment is more than mere convention; it is inescapable and transcends individual preferences or interests (Smart and Williams 1973). Of course, when faced with moral dilemmas, people do often weigh up one option against the other. But to describe all this as utility maximization or a matter of preference misses the point. If people are always acting in a way that ends up maximizing their own utility, then they cannot be seen as truly altruistic or moral: instead they are psychopaths or sociopaths, manipulating circumstances for their own benefit, behind appearances of kindness. Drawing on what I believe is the majority view among moral philosophers, I maintain that moral dispositions cannot be adequately summarized by any preference function. It is thus no accident that Adam Smith resisted the utilitarianism of his friend David Hume. But as modern economics has developed, a lesser place has been given to independent moral considerations. Figure 7.2 shows how the usage of terms beginning in ‘ethic’ or ‘moral’ (but excluding appearances of ‘moral hazard’) has declined in articles in ten leading Anglophone journals of economics relative to the use of the term ‘utility’.9 The number of articles containing terms beginning in ‘ethic’ or ‘moral’ were searched, and the number of articles containing the word ‘utility’ were also found. The graph shows the former figure divided by the latter for each decade. Before 1910, terms such as ‘ethics’ or ‘morality’ were more common than ‘utility’. After the First World War, ‘utility’ took the lead. By the 1970s, articles with ‘moral’ terms were less than 20% as frequent as those with ‘utility’,

116  Geoffrey M. Hodgson 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s

Figure 7.2  The relative decline of ‘morality’ in leading economics journals

and from the 1980s they were around 10%. In an era where mathematical expression matters most, morality had been largely subsumed under utility.

Conclusion The principal argument in this chapter is that Max U has been more durable in recent decades than some scholars have argued. This is despite the major changes in theoretical approach, technique, and diversity in economics since 1980. Key developments – such as the kind of post-­Simon behavioural economics that emerged in the 1990s – still show the gravitational pull of Max U. There are exceptions, such as work on complexity, evolutionary game theory, or agent-­based modelling, but so far these have been insufficient to dislodge the dominance of Max U in the discourse of the most prestigious journals of economics. Neither has the increasing use of empirical methods and data undermined the reign of Max U. In part, Max U survives because the general proposition of utility maximization is strictly unfalsifiable. There is no way of demonstrating that agents are not maximizing the hidden variable of utility. But unfalsifiable propositions can have a framing or organizing role in science, and Max U plays this part to the full. To a large degree this explains its persistence. It has returned to behavioural economics, where it serves as a benchmark. This chapter cites bibliometric, survey, and other evidence to challenge the (Colander et al. 2004: 485; Colander 2005b: 930–931) claim that mainstream economics is moving away from its ‘holy trinity’ assumptions of ‘rationality, selfishness and equilibrium’. On the contrary, Max U is still high priest. He has donned new clothes, including ‘social’ and ‘other-­regarding’ preferences (all while his own selfish U is still being maximized).

Max U, morality, and the future of economics 117

To show the limitations of Max U, this article has considered the human capacity for moral judgement and the possibility of moral motivation. Although the concept of morality remains controversial within ethics, a number of moral philosophers suggest that a moral judgement is different from a convention or preference because it entails a claim to universality. For policy design, models such as Max U that predict behaviour are far from enough. Policies should be informed by richer explanations of motivational dispositions and behaviour. Moral motivation is important in this regard. Policy design can often appeal to moral values as well as pecuniary incentives. But apart from the work of Sen and a few others, there is little sign of economics moving in this direction and fully rehabilitating the ‘moral sentiments’ of Adam Smith. Not only is this development indiscernible in orthodox economics, but it is absent from much of heterodox economics as well. Fred Lee (2008) argued that heterodox economics comprises ‘a group of heterodox theories’ including Marxism, post-­Keynesianism, social economics, feminist economics, and institutionalism (he implied, but he did not make it clear, that this referred to the original rather than the new institutional economics). Missing from Lee’s article and his History of Heterodox Economics (2009) was any mention of behavioural economics, Herbert Simon, or Amartya Sen. Although Tony Lawson’s definition of heterodox economics is very different from Lee’s (Lawson sees orthodoxy as the inappropriate use of mathematical modelling), Lawson (2006) also failed to identify Simon or Sen as heterodox. Although self-­described ‘heterodox’ economists sometimes complain about Max U and make vaguely defined accusations about ‘methodological individualism’,10 rarely do they begin the task of constructing an alternative explanation of human behaviour. A notable exception here is the work of John Davis (2003, 2007). But generally, Max U is an important part of the enduring core of orthodox economics, and most heterodox critics have overlooked this prime target. Even if it were unlikely to divert mainstream economics from its tracks, a focused and sustained criticism of Max U by a network of scholars could have some positive influence and effect. But while such criticisms could have rich policy conclusions, they would not necessarily be aligned to what is currently described politically as the Left. The restoration of moral sentiments would be a project for traditional conservatives, as well as enlightened liberals or ethical socialists. The political ambiguity of the assault by moral sentiments on Max U is perhaps a reason why moral motivation is largely absent from existing ‘heterodox’ networks, which, from Cambridge henceforth, have generally inclined to the Left. Yet this would remain a viable and highly useful mission if it could find niches in some departments of economics and economics journals. For example, it could become the defining focus of the Review of Social Economy (for which, of course, John Davis has been a close and long-­standing mentor and associate).11 It could draw from psychology and philosophy as well as economics itself. Despite the Great Crash of 2008, we should discard any expectations

118  Geoffrey M. Hodgson

of an imminent scientific revolution in economics. But substantial, organized pressure from outside the citadel could have positive and permanent effects, including on other social sciences.

Notes 1 The author is very grateful to David Colander, Sheila Dow, Peter Earl, Deirdre McCloskey, Robert McMaster, and Wade Hands for comments on earlier versions of this chapter. 2 For example, the prominent Marxist academic David Harvey (2005, p. 20) conflated neoclassical economics with neoliberal politics. 3 Becker (1976a, p. 15) also described his work as ‘an application of neo-­classical economics’ to a wide range of phenomena, including the family. 4 See the useful methodological discussion in Dow (2013). 5 See Hodgson (2013, chapter 3) and my debate with Gintis and Helbing (2015) in Hodgson (2015a). 6 It is widely accepted in the philosophy of science that some unfalsifiable propositions are necessary. These include the principle of determinancy (every event has a cause) and the assumption of the uniformity of nature. Without these prior assumptions, science is impossible. 7 The journals were the American Economic Review, Econometrica, Economic Journal, Economica, Journal of Economic Literature, Journal of Economic Perspectives, Journal of Political Economy, Quarterly Journal of Economics, Review of Economic Studies, and Review of Economics and Statistics. The data were collected on 20 May 2018. In some cases, data were unavailable in the years 2013–2014 because of JSTOR access horizons. But as the frequency (rather than the absolute number) of appearances was measured, this deficit is less serious. 8 Popper (1976: 151, 168) once argued that Darwinism was unfalsifiable, but he then changed his mind (Popper 1978: 344–346). Darwinism is neither tautological nor unfalsifiable: the fittest are not defined as those who survive, and the fittest do not always survive. And, in any case, the term ‘survival of the fittest’ came from Herbert Spencer and was rarely used by Darwin. 9 This figure is taken from Hodgson (2013). The ten journals surveyed were the American Economic Review, Econometrica, Economic Journal, Economica, Journal of Political Economy, Oxford Economic Papers, Quarterly Journal of Economics, Review of Economic Studies, Review of Economics and Statistics, and the Southern Economic Journal. The oldest of these journals started in 1886 and the newest in 1938. Economists use the term ‘moral hazard’ to describe inefficiencies that occur when risks are displaced. It has minimal resemblance to the concept of morality in ethics. 10 Hodgson (2007, 2013) showed that ‘methodological individualism’ is an ambiguous term. Some versions of its substance are acceptable, save for the biased labelling of ‘individualism’. Other versions are unsustainable. 11 Some editors of the Review of Social Economy have attempted to shift its focus toward ethics. My suggestion is that, in addition, there should be a concerted critical focus on Max U and his problems.

References Ackerman, R. (1976). The Philosophy of Karl Popper. Amherst: University of Massachusetts Press. Akerlof, G. A. and Kranton, R. E. (2010). Identity Economics. Princeton, NJ: Princeton University Press. Aspromourgos, T. (1986). On the origins of the term “Neoclassical”, Cambridge Journal of Economics, 10(3), September: 265–270.

Max U, morality, and the future of economics 119 Backhouse, R. E. (1998). The transformation of U.S. economics, 1920–1960, Viewed through a survey of journal articles’, in M. S. Morgan and M. H. Rutherford (eds.), The Transformation of American Economics: From Interwar Pluralism to Postwar Neoclassicism, Annual Supplement to Volume 30 of History of Political Economy. Durham, NC: Duke University Press, pp. 85–107. Backhouse, R. E. and Cherrier, B. (2017). The age of the applied economist: The transformation of economics since the 1970s’, Béatrice Cherrier, History of Political Economy, 49(Supplement): 1–33. Becker, G. S. (1976a). The Economic Approach to Human Behavior. Chicago: University of Chicago Press. Becker, G. S. (1976b). Altruism, egoism, and genetic fitness: Economics and sociobiology. Journal of Economic Literature, 14(2), December: 817–826. Becker, G. S. (1991). A Treatise on the Family, 2nd edition. Cambridge, MA: Harvard University Press. Berg, N. and Gigerenzer, G. (2010). As-­if behavioral economics: Neoclassical economics in disguise? History of Economic Ideas, 18(1): 133–166. Blaug, M. (1992). The Methodology of Economics: Or How Economists Explain, 2nd edition. Cambridge: Cambridge University Press. Blaug, M. (1999). The formalist revolution or what happened to orthodox economics after World War II? in R. E. Backhouse, and J. Creedy (eds.), From Classical Economics to the Theory of the Firm: Essays in Honour of D. P. O’Brien. Cheltenham: Edward Elgar, pp. 257–280. Blaug, M. (2003). The formalist revolution of the 1950s, in W. J. Samuels, J. E. Biddle, and J. B. Davis (eds.), A Companion to the History of Economic Thought. Malden, MA: Blackwell, pp. 395–410. Boland, L. A. (1981). On the futility of criticizing the neoclassical maximization hypothesis, American Economic Review, 71(5), December: 1031–1036. Bowles, S. and Gintis, H. (2011). A Cooperative Species: Human Reciprocity and Its Evolution. Princeton, NJ: Princeton University Press. Camerer, C. F. (2003). Behavioral Game Theory: Experiments in Strategic Interaction. Princeton, NJ: Princeton University Press. Camerer, C. F., Loewenstein, G., and Prelec, D. (2005). Neuroeconomics: How neuroscience can inform economics, Journal of Economic Literature, 43(1), March: 9–64. Camerer, C. F., Loewenstein, G., and Rabin, M. (eds.). (2004). Advances in Behavioral Economics. New York and Princeton: Russell Sage Foundation and Princeton University Press. Cedrini, M. and Fontana, M. (2018). Just another niche in the wall: How specialization is changing the face of mainstream economics, Cambridge Journal of Economics, 42(2), February: 427–451. Charness, G. and Rabin, M. (2002). Understanding social preferences with simple tests, Quarterly Journal of Economics, 117(3), August: 817–869. Coase, R. H. (1977). Economics and contiguous disciplines, in M. Perlman (ed.), The Organization and Retrieval of Economic Knowledge. Boulder, CO: Westview Press, pp. 481–491. Coats, A. W. (1993). The Sociology and Professionalization of Economics: British and American Economic Essays, Volume II. London: Routledge. Colander, D. C. (2000a). The death of neoclassical economics, Journal of the History of Economic Thought, 22(2), June: 127–143. Colander, D. C. (2000b). New millennium economics: How did it get this way, and what way is it? Journal of Economic Perspectives, 14(1), Winter: 121–132.

120  Geoffrey M. Hodgson Colander, D. C. (ed.). (2000c). The Complexity Vision and the Teaching of Economics. Cheltenham: Edward Elgar. Colander, D. C. (2000d). Complexity and the History of Economic Thought. London: Routledge. Colander, D. C. (2005a). The making of an economist redux, Journal of Economic Perspectives, 19(1), Winter: 175–198. Colander, D. C. (2005b). The future of economics: The appropriately educated in pursuit of the knowable, Cambridge Journal of Economics, 29(6), November: 927–941. Colander, D. C., Holt, R.P.F., and Rosser, J. B., Jr. (2004). The changing face of economics, Review of Political Economy, 16(4), October: 485–499. Cyert, R. M. and March, J. G. (1963). A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-­Hall. Darwin, C. R. (1871). The Descent of Man, and Selection in Relation to Sex, 2 vols. London: Murray and New York: Hill. Davis, J. B. (1995). Personal identity and standard economic theory, Journal of Economic Methodology, 2(1), June: 35–52. Davis, J. B. (2003). The Theory of the Individual in Economics: Identity and Value. London: Routledge. Davis, J. B. (2006). The turn in economics: Neoclassical dominance to mainstream pluralism? Journal of Institutional Economics, 2(1), April: 1–20. Davis, J. B. (2007). Akerlof and Kranton on identity in economics: Inverting the analysis, Cambridge Journal of Economics, 31(3), May: 349–362. Davis, J. B. (2008). The turn in recent economics and return of orthodoxy, Cambridge Journal of Economics, 32(3), May: 349–366. Dewey, J. (1922). Human Nature and Conduct: An Introduction to Social Psychology. New York: Holt. Dhami, S. (2016). The Foundations of Behavioural Economic Analysis. Oxford: Oxford University Press. Di Tella, R. and MacCulloch, R. (2006). Some uses of happiness data in economics, Journal of Economic Perspectives, 20(1), Winter: 25–46. Dow, S. C. (2011). Cognition, market sentiment and financial instability, Cambridge Journal of Economics, 35: 233–249. Dow, S. C. (2013). Formalism, rationality and evidence: The case of behavioural economics, Erasmus Journal for Philosophy and Economics, 6(3) (Special Issue): 26–43. Earl, P. E. (2010). Economics fit for the queen: A pessimistic assessment of its prospects, Prometheus, 28(3): 209–225. Fayazmanesh, S. (1998). On Veblen’s Coining of the term ‘Neoclassical’, in S. Fayazmanesh and M. R. Tool (eds.), Institutionalist Method and Value: Essays in Honour of Paul Dale Bush Volume 1. Cheltenham: Edward Elgar, pp. 74–97. Fehr, E. and Fischbacher, U. (2002). Why do social preferences matter – the impact of non-­selfish motives on competition, cooperation and incentives, Economic Journal, 112, March: C1–C33. Fourcade, M. (2006). The construction of a global profession: The transnationalization of economics, American Journal of Sociology, 112(1): 145–194. Fox, J. (2009). The Myth of the Rational Market. New York: HarperCollins. Gallegati, M. and Kirman, A. (2012). Reconstructing economics: Agent based models and complexity, Complexity Economics, 1(1): 5–31. Gintis, H. (2005). Behavioral game theory and contemporary economic theory, Analyse & Kritik, 27: 48–72.

Max U, morality, and the future of economics 121 Gintis, H. (2009). The Bounds of Reason: Game Theory and the Unification of the Behavioral Sciences. Princeton, NJ: Princeton University Press. Gintis, H. and Helbing, D. (2015). Homo socialis: An analytical core for sociological theory, Review of Behavioral Economics, 2(1–2): 1–59. Glimcher. P. W. (2003). Decisions, Uncertainty and the Brain: The Science of Neuroeconomics. Cambridge, MA: MIT Press. Glimcher, P. W., Dorris, M. C., and Bayer, H. M. (2005). Physiologic utility theory and the neuroeconomics of choice, Games and Economic Behavior, 52: 213–256. Hamermesh, D. S. (2013). Six decades of top economics publishing: Who and how? Journal of Economic Literature, 51(1), March: 162–172. Hare, R. M. (1952). The Language of Morals. Oxford: Oxford University Press. Harvey, D. (2005). A Brief History of Neoliberalism. Oxford: Oxford University Press. Hodgson, G. M. (1999). Evolution and Institutions: On Evolutionary Economics and the Evolution of Economics. Cheltenham: Edward Elgar. Hodgson, G. M. (2004). The Evolution of Institutional Economics: Agency, Structure and Darwinism in American Institutionalism. London: Routledge. Hodgson, G. M. (2007). Meanings of methodological individualism, Journal of Economic Methodology, 14(2), June: 211–226. Hodgson, G. M. (2010). Choice, habit and evolution, Journal of Evolutionary Economics, 20(1), January: 1–18. Hodgson, G. M. (2013). From Pleasure Machines to Moral Communities: An Evolutionary Economics without Homo Economicus. Chicago: University of Chicago Press. Hodgson, G. M. (2014). The evolution of morality and the end of economic man, Journal of Evolutionary Economics, 24(1), January: 83–106. Hodgson, G. M. (2015a). A Trojan horse for sociology? Preferences versus evolution and morality, Review of Behavioral Economics, 2(1–2): 93–112. Hodgson, G. M. (2015b). Conceptualizing Capitalism: Institutions, Evolution, Future. Chicago: University of Chicago Press. Infante, G., Lecouteau, G., and Sugden, R. (2016). Preference purification and the inner rational agent: A critique of the conventional wisdom of behavioural economics, Journal of Economic Methodology, 23(1), March: 1–25. James, W. (1890). The Principles of Psychology, 2 vols. New York and London: Holt and Macmillan. Joyce, R. (2006). The Evolution of Morality. Cambridge, MA: MIT Press. Kadish, A. and Tribe, K. (eds.). (1993). The Market for Political Economy: The Advent of Economics in British University Culture, 1850–1905. London: Routledge. Kagel, J. H., Battalio, R. C., and Green, L. (1995). Economic Choice Theory: An Experimental Analysis of Animal Behaviour. Cambridge: Cambridge University Press. Kahneman, D. (2003). Maps of bounded rationality: Psychology for behavioral economics, American Economic Review, 93(5), December: 1449–1475. Kirman, A. P. (1989). The intrinsic limits of modern economic theory: The emperor has no clothes, Economic Journal (Conference Papers), 99: 126–139. Kirman, A. P. (2009). Economic theory and the crisis, Vox, 14 November. Retrieved 2 June 2018. www.voxeu.org/­index.php?q=node/­4208. Kuhn, T. S. (1962). The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lawson, T. (1997). Economics and Reality. London: Routledge. Lawson, T. (2006). The nature of heterodox economics, Cambridge Journal of Economics, 30(4), July: 483–505.

122  Geoffrey M. Hodgson Lawson, T. (2013). What is this ‘school’ called neoclassical economics? Cambridge Journal of Economics, 37(5), September: 947–998. Lee, F. S. (2008). ‘Heterodox economics’, in S. N. Durlauf and L. E. Blume (eds.), The New Palgrave Dictionary of Economics, 2nd edition. London: Palgrave Macmillan. www.hetecon. net/­division.php?page=about&side=heterodox_economics. Lee, F. S. (2009). A History of Heterodox Economics: Challenging the Mainstream in the Twentieth Century. London: Routledge. Mackie, J. L. (1977). Ethics: Inventing Right and Wrong. Harmondsworth: Penguin. Maloney, J. (1985). Marshall, Orthodoxy and the Professionalisation of Economics. Cambridge: Cambridge University Press. McCloskey, D. N. (2006). The Bourgeois Virtues: Ethics for an Age of Commerce. Chicago: University of Chicago Press. McCloskey, D. N. (2008). Mr. Max and the substantial errors of manly economics, Economics Journal Watch, 5(2), May: 199–203. McCloskey, D. N. (2016). Max U vs. humanomics: A critique of neo-­institutionalism, Journal of Institutional Economics, 12(1), March: 1–27. Morgan, J. (ed.). (2016). What is this ‘School’ Called Neoclassical Economics? Debating the Issues. London: Routledge. Platt, M. L. and Glimcher, P. W. (1999). Neural correlates of decision variables in parietal cortex, Nature, 400(6741), July 15: 233–238. Plotkin, H. C. (1994). Darwin Machines and the Nature of Knowledge: Concerning Adaptations, Instinct and the Evolution of Intelligence. Harmondsworth: Penguin. Popper, K. R. (1976). Unended Quest: An Intellectual Autobiography. LaSalle, IL: Open Court. Popper, K. R. (1978). Natural selection and the emergence of mind, Dialectica, 32(3–4): 339–355. Pollak, R. A. (2003). Gary Becker’s contributions to family and household economics, Review of Economics of the Household, 1(1–2), January: 111–141. Potts, J. (2000). The New Evolutionary Microeconomics: Complexity, Competence and Adaptive Behaviour. Cheltenham: Edward Elgar. Rabin, M. (1998). Psychology and economics, Journal of Economic Literature, 36(1), March: 11–46. Richiardi, M. and Leombruni, R. (2005). Why are economists sceptical about agent-­based simulations? Physica A, 355(1): 103–109. Rizvi, S.A.T. (1994a). The microfoundations project in general equilibrium theory, Cambridge Journal of Economics, 18(4), August: 357–377. Rizvi, S.A.T. (1994b). Game theory to the rescue? Contributions to Political Economy, 13: 1–28. Samuelson, P. A. (1937). A note on the measurement of utility, Review of Economic Studies, 4(2), February: 155–161. Schumpeter, J. A. (1954). History of Economic Analysis. Oxford: Oxford University Press. Sen, A. K. (1977). Rational fools: A critique of the behavioral foundations of economic theory, Philosophy and Public Affairs, 6(4): 317–344. Sen, A. K. (1985a). Commodities and Capabilities. Amsterdam: North-­Holland. Sen, A. K. (1985b). Goals, commitment, and identity, Journal of Law and Economic Organization, 1(2), Fall: 341–355. Sen, A. K. (1987a). On Ethics and Economics. Oxford: Basil Blackwell. Sen, A. K. (1987b). Rational behaviour, in, J. Eatwell, M. Milgate, and P. Newman (eds.), The New Palgrave Dictionary of Economics. London: Macmillan, Vol. 4, pp. 68–76. Sen, A. K. (ed.). (2002). Rationality and Freedom. Cambridge MA: Harvard University Press.

Max U, morality, and the future of economics 123 Sen, A. K. (2004). Social identity, Revue de Philosophie Economique, 9(1): 2–26. Sent, E-­M. (2004). Behavioral economics: How psychology made its (limited) way back into economics, History of Political Economy, 36(4): 735–760. Siakantaris, N. (2000). Experimental economics under the microscope, Cambridge Journal of Economics, 24(3), May: 267–281. Simon, H. A. (1957). Models of Man: Social and Rational. New York: Wiley. Simon, H. A. (1959). Theories of decision-­making in economic and behavioral sciences, American Economic Review, 49(2), June: 253–283. Smart, J.J.C. and Williams, B. (1973). Utilitarianism: For and Against. Cambridge: Cambridge University Press. Smith, A. (1759). The Theory of Moral Sentiments; or, An Essay Towards an Analysis of the Principles by which Men Naturally Judge Concerning the Conduct and Character, First of their Neighbours, and Afterwards of Themselves. London and Edinburgh: Millar and Kincaid and Bell. Smith, V. L. (1982). Microeconomic systems as an experimental science, American Economic Review, 72(5), December: 923–955. Smith, V. L. (1992). Papers in Experimental Economics. Cambridge: Cambridge University Press. Smith, V. L. (2013). Adam Smith: From propriety and sentiments to property and wealth, Forum for Social Economics, 42(4), pp. 283–297. Steedman, I. (1980). Economic theory and intrinsically non-­autonomous preferences and beliefs, Quaderni Fondazione Feltrinelli, 7(8): 57–73. Reprinted in Steedman, I. (1989). From Exploitation to Altruism. Cambridge: Polity Press. Steele, K. (2014). Choice models, in N. Cartwright and E. Montuschi (eds.), Philosophy of Social Science: A New Introduction. Oxford: Oxford University Press, ch. 10. Sugden, R. (1991). Rational choice: A survey of contributions from economics and philosophy, Economic Journal, 101(4), July: 751–785. Thaler, R. H. and Sunstein, C. R. (2008). Nudge: Improving Decisions About Health, Wealth, and Happiness. New Haven, CT: Yale University Press. Tullock, G. (1994). The Economics of Non-­Human Societies. Tucson, AZ: Pallas Press. Tyler, T. R. (1990). Why People Obey the Law. New Haven, CT: Yale University Press. Veblen, T. B. (1914). The Instinct of Workmanship, and the State of the Industrial Arts. New York: Macmillan. Vromen, J. J. (2010). On the surprising finding that expected utility is literally computed in the brain, Journal of Economic Methodology, 17(1), March: 17–36. Weintraub, E. R. (2002). How Economics Became a Mathematical Science. Durham, NC: Duke University Press. Wilson, B. J. (2010). Social preferences aren’t preferences, Journal of Economic Behavior and Organization, 73(1): 77–82. Winter, S. G., Jr. (1964). Economic “natural selection” and the theory of the firm, Yale Economic Essays, 4(1): 225–272. Zak, P. J. (2004). Neuroeconomics, Philosophical Transactions of the Royal Society of London, B, 359: 1737–1748.

8 The meso level in economics Irene van Staveren

Introduction There is surprisingly little attention in economics to the meso level. And when there is, the term is often not used. Rather, economists from a wide variety of schools of thought refer to it in terms of groups, regions, industries, social norms, institutions, interaction effects, fallacy of composition, or social capital. Four schools of thought in particular make reference to the relevance for economic analysis of this intermediate level in between micro and macro. But most of these make reference to only one dimension of the meso level. It is social economics in which we find the broadest understanding of the meso level. The purpose of this chapter is to argue, with the help of key contributions by social economists, and in particular by John Davis, the various ways in which social economics conceptualizes the meso level. In addition, connecting to other schools of thought, I will argue that this forms a coherent and meaningful understanding of the meso level for economics more generally.

The meso level in various schools of thought Keynes (1936) introduced two meso level concepts: the fallacy of composition and herd behaviour. The fallacy of composition explains why the macro level is only partially an aggregation of micro level units. Such aggregation is what economic geographers study when focusing on regions, or what is analyzed in the field of industrial organization as sectors. Regions nicely add up to countries and country groupings up to the world level, whereas sectors nicely add up to the whole of business activity in an economy. But not all of the macro level is an aggregation of micro level units. The concept of the fallacy of composition shows that various micro level variables do not simply add up to the macro level. These are variables, which are affected by interaction effects in the economic process. For such variables there is a mismatch between the individual level and the aggregate level. Keynes illustrated the problem of ignoring this mismatch in macroeconomics very clearly: “The classical theorists resemble Euclidean geometers in a non-­Euclidean world who, discovering that in experience straight lines apparently parallel often meet, rebuke the lines for

The meso level in economics 125

not keeping straight – as the only remedy for the unfortunate collisions which are occurring” (Keynes 1936: 16). The interaction effects that are responsible for the mismatch between the individual and the aggregate level involve positive and negative feedback effects so that at the macro level, the micro level units are not aggregated but transformed. This is particularly the case in financial markets, where for example derivatives are constructed from parts of individual assets and then traded and eventually put together in yet other packages for further trading. Another example from financial markets is the gap between the real value of corporations based on their balance sheets and their actual value as signalled by their share prices on the stock exchange. This gap is a consequence of expectations and speculation by investors and is strongly influenced by the interactions between investors, the media, financial supervisory agencies, and politics, so that depending on all these interaction effects the difference between the aggregate value of individual shares of a firm and its book value varies on a day-­to-­day basis. The other Keynesian concept that operates at the meso level is herd behaviour. It cannot exist without the meso level because it is caused entirely by the interaction effect of individual agents. We see it most clearly as well as with a frightening effect on asset values, in stock exchange crashes, currency crises, and bank runs. Herd behaviour is not an aggregation of independent individual behaviour, grounded in individual preference satisfaction, but it emerges from the interrelatedness of buying and selling decisions made by agents, largely on the basis of greed and fear. In behavioural economics, it has been referred to as irrational behaviour, or irrational exuberance (Shiller 2000). But this is a misleading term. It implies that each and every individual during financial turmoil would make irrational choices, which locates the behaviour at the individual level instead of the meso level. While it is not lack of individual rationality – when stock prices go down, it is still rational to sell shares before they decline even further – but something going on collectively, a group process (groupthink), as in a herd. Keynes understood that this occurs not at the micro level but somewhere in between the individual choice and the unintended aggregate result. So with two key concepts, Keynes showed that he had a clear understanding of the meso level, with the fallacy of composition showing that not all macro level phenomena are aggregated from micro level units, and with herd behaviour showing that under uncertainty, agents’ decisions tend to move collectively in the same direction. A different theory recognizes the relevance of the meso level in a much wider variety of markets than financial markets alone – from labour markets to consumer markets and from finance to trade: institutional economics. In institutional economics, the meso level is where most institutions emerge, consolidate, are challenged, and change or are replaced entirely by new institutions (Hodgson 1993). Various institutional economists recognize the existence of individual-­level institutions, such as rules of thumb in spending household income, for example. Other institutional economists focus on

126  Irene van Staveren

formal institutions in the shape of organizations or laws, which operate at the macro level. But even for such micro and macro institutions, they only function as institutions, and not merely as constraints (e.g., laws) or preferences (e.g., saving being valued more than non-­necessary consumption) because they influence individual decisions from beyond the individual agent herself. It is at the meso level that institutions guide the behaviour of a significant number of agents in a certain direction, as Thorstein Veblen (1899) has explained so well for conspicuous consumption patterns. In the branch of institutional economics known as evolutionary economics, the meso level is the key to understanding change and coordination in an economy. It is where, in the economy as an open system, the dynamics are located. According to Dopfer et al. (2004: 277): “Coordination and change are both endogenous in the evolutionary framework and inherently defined in terms of meso”. Therefore, the authors argue, the meso level has ontological status in evolutionary economics. In institutional economics, it is precisely at the meso level that institutions, as regular patterns in behaviour, or social norms, guide the behaviour of individuals and in their interactions they are contested and adapted. For example, through collective action by labour unions, in the case of labour market institutions such as minimum wages. Or through their normative character, as in the case of women’s higher share in unpaid work as compared to men’s share. Or because they reduce transaction costs, and thereby generate efficiency gains, for example through product standardization and work routines. Moreover, institutions are relatively resilient precisely because they generate patterns in behaviour, reducing the need for constant time-­consuming decision-­making by searching for and weighing alternative options. As a consequence, some institutions may survive even though they lose their original function, as is the case in path dependency. So, institutions operate largely at the meso level by guiding individual behaviour beyond preferences and constraints, because they refer to underlying values or group interests or they reduce transaction costs, even when they are no longer relevant. The third theory that addresses the meso level of the economy is actually not a theory but an approach: the capability approach. It puts not individual preferences and their maximization central but people’s capabilities – the doings and beings that they have reason to value (Sen 1987). Sen defines capabilities at the level of individuals but recognizes that their achievement is largely a socially embedded process. “The capability that we are concerned with is our ability to achieve various combinations of functionings that we can compare and judge against each other in terms of what we have reason to value” (Sen 2009: 233). Key in this process is how resources – individual and collective ones, including entitlements guaranteed by the state – are transformed to capabilities and eventually to functionings, and that this transformation process is influenced by institutions. These institutions are largely defined in the same way as in institutional economics and can be enabling or disabling capability development.

The meso level in economics 127

Enabling institutions would be, for example, equal rights in the labour market, whereas disabling institutions would be discrimination by employers. These opposite effects show that institutions affect the capability space for individuals and for social groups, so that the capability space of one group (e.g., men) may be larger than that of another group (e.g., women). So although capabilities are defined at the individual level, their realization requires that institutions are shaped in such a way that everyone can realize the capabilities that they have reason to value. The brief discussion in this section of how the meso level has been conceptualized in economics, apart from simply referring to the industry level or an economic sector, has provided a broad but sketchy view of what constitutes the meso level in economics. This picture is incomplete. Below, I will attempt to provide a more complete and coherent account of the meso level. I will do this with a discussion of the multifaceted understanding of the meso level in social economics.

The three forms of the meso level in social economics Social economics stands out for its analysis of the economy as a social and open system. It focuses on two related domains. The first is the community economy (also labelled as the third domain), or social economy, with its voluntary, cooperative, and non-­profit activities in and for communities. The second domain of analysis concerns the social values underlying the market and agents’ behaviour on markets and in mixed-­market economies. John Davis and Wilfred Dolfsma (2015: 4) connect these two domains by explaining that the social perspective on mixed market economies “is inspired by the original concern of social economics with the social economy, since their social values and social relationships are prominent and dominate economic values and economic relationships”. To be more precise, they argue that it is “social values and social relationships that underlie and drive all aspects of the market” (2015: 5). In other words, social economics regards the economy as embedded in society, and therefore it also considers economic agents – from individuals and households to firms – as socially embedded. This pervasiveness of the social, with its social values and relatedness of agents, is why social economics provides a key role for the meso level. Moreover, Davis et al. (2004) have argued in favour of a social ontology for economics. This implies that there is an ontological role for the meso level in social economics, as has also been argued for evolutionary economics as we have seen above. So the starting point for a discussion of the meso level in social economics is the recognition that it has an ontological status and is more than just a connection between the micro and the macro level, as it was for Keynes. In social economics, it is the meso level that enables individual agency and provides meaning to this, while at the same time it is the meso level that generates the dynamics observed at the macro level with its cycles, growth paths, and

128  Irene van Staveren

changing levels of inequality. From this starting point, I distinguish three forms through which the meso level manifests itself in social economics: 1 Social groups 2 Social norms 3 Interdependencies. I will elaborate each of these below, with a special reference to the work by John Davis, who has contributed to the conceptualization of each of these as well as to their connection. Moreover, my discussion will indicate that the three are closely related. Social groups

In his book The Theory of the Individual in Economics, John Davis (2003) has demonstrated that individuals are not independent, disengaged, autonomous beings driven by utility maximization but socially embedded persons with a social identity. Hence, individuals cannot be analyzed without the social positions they have in life – their roles and social identities. This embeddedness of individuals is enabled by the existence of social groups. Davis defines social groups as “collections of individuals whose shared characteristics constitute the basis for their interaction with one another” (Davis 2003: 133). This definition implies the following features, Davis continues: (1) a plurality of individuals tied to one another, (2) the principle of membership, and (3) a system of individual rights and obligations. Some social groups are given and rather fixed, for example sex or ethnicity or country of birth. Other social groups already make part of one’s social environment when growing up and living in a particular location. Hence, such groups are joined through one’s upbringing, peers, education, talents, or life events. Examples of such groups are lifestyle groups, neighbourhoods, religious communities, professional groups, or support groups for patients suffering from a particular disease. Finally, people make part of groups that are deliberately chosen, upon reflexion. Such groups include labour unions, the women’s movement, service clubs, or social activist groups. Some groups are formalized with memberships, fees, regulations, and assigned tasks. Other groups are informal and have vague boundaries, such as lifestyle groups (when are you a genuine organic consumer?) and movements (who defines who belongs to the women’s movement?). As a consequence, in some cases one is identified by outsiders as a member of a group, whereas in other instances one self-­identifies with a social group. Finally, people tend to be member of different groups at the same time, expressing the intersectionality of social life. Social groups not only influence who we are and the decisions we make, but they also allow us to change society rather than just accept it or being predisposed by it without being much aware of it. This feature of the embedded individual is what prevents an overdetermined perspective of agency, John

The meso level in economics 129

Davis argues: “individuals are not only influenced by society but also act upon and change society” (Davis 2003: 111). And this is precisely why the social embeddedness of individual agents must be located at the meso level and not at the macro level. It is a matter of interaction, not a one-­directional influence: it is where the social happens. This is another way of saying that the economy, as embedded in society, is an open system. Change happens at the meso level through the collective intentionality of group members, Davis argues. Such changes occur within groups but particularly also between groups, through alliances and conflicts that may occur between social groups. Hence there are interdependencies between members of social groups (e.g., union members) but also between social groups (e.g., employers and workers). In a recent article on stratification economics, Davis (2015) focuses on the conflicts between groups and within groups. His analysis helps to understand how change occurs at the meso level. In stratification economics, the focus of analysis “is group-­based inequality and the social economic hierarchies associated with individuals’ identification with social groups, allegiance to which is manifested in asymmetries between their pro-­own group and anti-­other group behaviours” (Davis 2015: 1215). From this perspective, social groups are not homogeneous but heterogeneous, which necessitates an intersectionality approach to the analysis of the meso level. This approach allows for an understanding of various types of group conflicts and hence, of conflictual forms of economic change. “Individuals can have pro-­own-­group allegiances to different social groups that on average practice anti-­other-­group behaviour towards one another and individuals can even find themselves be anti-­own-­group in circumstances when their other-­group-­loyalties are pre-­eminent” (Davis 2015: 1216). As a consequence, Davis notes, intragroup conflicts interact with intergroup conflicts. Such conflicts within and between groups tend to be reinforced by stigmatization of social identities. Think of phrases such as “women are always” or “immigrants all want”. This has serious consequences for the access to and control over resources for members of subordinated groups. They tend to be disadvantaged in labour markets, financial markets, and in access to affordable housing and education. As a consequence, Davis argues, the well-­being of members of subordinated groups is less than that of members of privileged groups. Moreover, and here comes the crucial insight from Davis’s combined analysis of stratification and social groups, the dynamics of the interaction of the two tends to reinforce inequality. “In an intersectional world, promotion ladders, social distance and stigmatisation consequently work to reinforce social stratification rather than weaken it, so that people having multiple identities fail to offset intergroup inequality” (Davis 2015: 1226). In other words, the change occurring at the meso level due to social interaction within and between groups tends to be in favour of already privileged groups. Such perverse dynamics inherent in the presence of social groups in an unequal world demonstrates the importance of giving more attention to the meso level of analysis in economics.

130  Irene van Staveren Social norms

Remember that one of the two key domains of analysis in social economics is the underlying values of markets and mixed-­market economies. This focus on values, which are intangible and difficult to measure, often leads to a focus on social norms. Just as in institutional economics, social norms are regarded as regular patterns in behaviour, while social economists emphasize that these norms are related to underlying values. Indeed, in social economics, these values are often found to have a moral character: they refer to right and wrong and imply how individuals and groups are supposed to behave in society at large and in the economy as it is embedded in society. Like norms, moral values go beyond an individual preference. They embody a collective intentionality, which Davis explains as “operating with a conception of embedded individuals rather than atomistic ones”, which makes it possible, he continues, “to include a sense of moral obligation” (Davis 2004: 399). Adam Smith already recognized the important role of values for a flourishing market economy. As I have discussed before (van Staveren 2001), Smith has stated not only in The Theory of Moral Sentiments but also in his Wealth of Nations that three values in particular underlie a flourishing economy: liberty, justice, and benevolence. In social economics, all three are still considered key values influencing the behaviour of individuals, households, firms, and the state. And since the community economy is so central in social economics, its underlying value of caring (in which we recognize Smith’s benevolence) is considered as a driving moral force for the behaviour of individuals and groups in the social economy. But, of course, a value like caring is not simply switched off when agents enter market relations or an engagement with the state. Cumbers et al. (2015) provide a good example of the attention to the value of care in social economic analysis. They argue that care is a neglected ethical value in economic analysis, although they acknowledge the exception to this by feminist economists (Folbre and Nelson 2000; van Staveren 2005; Himmelweit 2007). They argue for the importance of caring in human behaviour by referring to the institutional economist Thorstein Veblen, who labelled “the parental bent” as an instinct, or human motivation, for the betterment of society and the economy. He defined this instinct as other-­regarding and future-­ oriented. And it is here that caring, as a social norm, connects to social groups (with women providing, on average, more care than men due to a traditional gender division of labour) and to interdependencies (due to its reliance on taking up responsibility for others’ needs). Like Veblen did a century ago, Cumbers et al. (2015) acknowledge that the amount of care provided in an economy depends on the institutional setting enabling or stimulating a habit to care on the one hand or disabling and devaluing a habit to care on the other hand. This link between a value and an institutional setting, through habituation, is precisely how social norms operate at the meso level. If it were only virtue that some individuals may have and others not, its influence on the economy would remain at the micro level. But when

The meso level in economics 131

an ethical value, as for Smith and Veblen, is recognized to be an important and common human motivation, it will result in a widely shared social norm and be acted upon in a regular, patterned way. Hence such behaviour will become institutionalized through assigned roles, beliefs, and routines, and possibly also through laws and regulations. Now, observing the limited caring in todays’ society, Cumbers et al. (2015) conclude that the institutional setting around the value of care is limited and weak, leading to a care deficit. To illustrate the general weak support for the value of caring, they mention the low payment for typical caring jobs in health care (of nurses) as compared to curative jobs (of doctors), the assignment of many hours of unpaid care work to women in households, and the failing duty of care by financial institutions during the recent financial crisis. The example of the value of care also points out that social norms have an important coordinating function in the economy, when indeed the underlying ethical values and the institutions supporting these are aligned. These supporting institutions are not necessarily inclusive, as the examples mentioned above have shown. They may assign roles to particular groups and not others, and these roles may be low paid in the labour market, or undermined by market failures such as the too-­big-­to-­fail banks shifting the downside risk of their operations to clients and taxpayers. An important insight from the social economics of social norms, hence, is not only that the underlying values need sufficient institutional support for a coordinating role at the meso level, but also that such institutional support may create or reinforce economic inequalities. Interdependencies

The third dimension of the meso level is the least tangible. Social groups can be observed and measured in terms of their size, characteristics, and roles. Social norms can be observed indirectly, through surveys, focus group discussions, and ethnography for example. And the most formalized social norms are materialized in laws, organizations, and objects such as traffic lights, and therefore they are clearly visible. But interdependencies between individuals and groups tend to escape observation and measurement. They refer to the social content of how economic agents relate to each other in communities and markets. It is this intangible social relatedness that is behind the Keynesian concepts of the fallacy of composition and herd behaviour. The literature on social capital and social coherence has tried to make sense of this intangible factor in the economy. But very often, it has failed to do so by treating it as just another form of capital, a productive resource that individuals or firms can possess and control for their own benefit. This reductionism is precisely why social economists have criticized much of the social capital literature (see, for example, Dolfsma and Dannreuther 2003; Fine 2001; van Staveren and Knorringa 2008a). At the same time, they have not thrown the baby out with the bathwater. After all, the idea that social relatedness matters for the economy is central to social economics. So in social economics, social capital is a term used

132  Irene van Staveren

with caution, referring, according to one of the volumes mentioned above, to “the set of social relations that enable actors to gain, maintain or expand access to economic resources that may lead to the reinforcement of the productivity of these economic resources” (van Staveren and Knorringa 2008b: 112). To be honest, ten years after I wrote that down, I think it is still too instrumental. Now, I would emphasize the fact that social relatedness is first of all an end in itself – after all, human beings are social animals – and only in second instance interdependencies between individuals and groups can have benefits but also costs, depending on the power relations embedded in the relatedness. What is important in the critical, social economic analyses of social capital is the recognition that it is not a form of capital at all: it cannot be owned, controlled, bought and sold, and generate returns on investment for individual investors. Instead, it is affected by the whole range of aspects of social relations, with its positive sides such as collective action, coordination, cooperation, mutual trust and solidarity as well as its negative sides such as inequality, social exclusion, control, discrimination, and exploitation. The only thing that is retained in social economics from the mainstream literature on social capital is the distinction between bonding and bridging social capital. The first type refers to the social capital of groups through the strength of the ties between the members of the group. This can be based upon a common identity, a joint goal, or a shared value. The ties are therefore inward oriented and imply strong cooperation and mutual trust. In fact, bonding social capital combines the first dimension of the meso level (social groups) with the third dimension discussed in this subsection (interdependencies). Therefore, bonding social capital can be beneficial within a group but may turn against other groups, with limited cooperation between the groups, or even conflict, and low trust of members of other groups. The other type of social capital that is useful in social economic analysis is bridging social capital, which is based on weak ties between more or less strangers. The weak ties are however strong enough to generate sufficient mutual trust, referred to as generalized trust, and to sustain some shared values between heterogeneous individuals. So, bridging social capital cuts across groups and thereby enables markets, with trading relations, fiduciary money, and investments outside one’s own community or social groups. For the interdependency dimension of the meso level, it is bridging social capital, which is the most important of the two types of social capital. Without it, economies would not extend beyond the social economies of relatively isolated, or perhaps even warring, communities. Generalized trust and shared values enable communication and interaction with potential trading partners and with the gatekeepers and potential partners of distant investment opportunities. Their mutual trust and shared values are precisely what help them to overcome transaction costs in such distant economic relationships and which allow for reducing uncertainty by crafting joint rules. Again, we recognize a connection with one of the other dimensions that we distinguished at the meso level, in this case social norms. Trust is a kind of social norm, a shared expectation not to be taken advantage of by others. And joint rules to reduce uncertainty can

The meso level in economics 133

only be made when at least some social norms are shared, in particular norms about honesty, transparency, and responsibility or honour. In an edited volume about social capital and social economics, John Davis and Asimina Christoforou (2014) use the concept of social capital as a heuristic device, which nicely shows how it helps to define the meso level. In their introduction, they define social capital “in terms of social norms and social networks” and recognize that “it manifests itself in patterns of trust, reciprocity, and cooperation” (Davis and Christoforou 2014: 3). In his own chapter in the volume, concerned with trust and conflict, John Davis defines social capital as “the economic and non-­economic benefits produced by cooperation and trust in social networks” (Davis 2014: 98). But he also shows that it may have costs as well, when the benefits are distributed unequally, or when these are generated by the exploitation of others. Then he links the two types of social capital to the other two dimensions of the meso level that I have described above, but in a slightly different way, using the concept of identity. He links what he calls categorical social identity to social groups and hence to bonding social capital, whereas relational social identity is linked to bridging social capital. He then argues that bonding social capital automatically generates anti-­group behaviour toward other groups, with distrust and conflict. From this insight he then comes to a surprising policy implication. He argues that for more trust and cooperation in an economy, and less conflict, we should not focus on enlarging bonding social capital through expanding the size of groups, but we should rather try to stimulate bridging social capital. He argues that in the current social context “bridging social capital may increasingly be the default type of social capital . . . because it offsets the ways in which social group conflict break down trust and cooperation” (Davis 2014: 110). This is a plea for the third dimension of the meso level because it helps to reduce the negative side effects of social groups, the first dimension that we discussed. However, the relational social identity of bridging social capital has downsides as well, because it can reinforce the capitalist market instincts of fear and greed. When fear spreads through networks throughout the economy, it triggers distrust and panic in markets, particularly in asset markets, with economic crisis as a result. And when greed spreads, it triggers accumulation in markets, with growing power concentration through cartels, oligopolies, monopolies, and lobbyism for rules in favour of this power as a result. Hence, bridging social capital can be an influential force of the meso level, enabling the extension and deepening of markets with a reduction of uncertainty and transaction costs and an increase in generalized trust, but also reinforcing market volatility and inequality during across the economic cycle.

The meso level in economics What can we learn from this understanding of the meso level in social economics for the meso level in economics more generally? I like to answer this question by filling in the gaps between the three dimensions of the meso level described

134  Irene van Staveren

above, on the one hand, and linking this to how the meso level is perceived in the three other schools of thought mentioned earlier (post Keynesianism, institutional and evolutionary economics, and the capability approach), on the other hand. I will start with the connections between the three dimensions of the meso level: social groups, social norms, and interdependencies. As already indicated in the preceding discussion, these dimensions are woven together. Social groups tend to have goals and commitments of its members to these goals, which may be moral ends or morally neutral. In the case of moral goals, the economic behaviour of the group members will reflect a collective intentionality based on a sense of moral obligation within the group, and sometimes also extended to outsiders. This then sets a moral norm for their behaviour, to which they are committed and will hold others responsible for by reinforcing the norm through social pressure or even punishment of transgressors. In the case of morally neutral goals (e.g., collective action to defend joint interests), the collective intentionality is directly oriented towards a material goal (e.g., setting up joint management rules for a common pool resource), as Elinor Ostrom (1990) has described so meticulously in her fieldwork. Such rules emerging in collective action become social norms: they set the standard for the most effective behaviour towards achieving the goal and they also will include punishments for transgressors. When we take social norms as the starting point, instead of social groups, we see a connection to the interdependencies between individuals inherent in the social embeddedness of economic agents. Social norms can only play a role in economic behaviour when economic agents are not autonomous and independent individuals each on their own trying to maximize their subjective utility. Social embeddedness implies that behaviour is partially influenced, positively or negatively, by social norms. Again, such guiding social norms may have a moral character, as Davis often emphasizes, referring to values such as equality, fairness, and the positive freedom to achieve, or they may be of a more pragmatic nature. Hence, also without social groups, the behaviour of individual economic agents is guided by social norms, because they are shared by a critical mass of socially embedded individuals. How then is this understanding of the meso level related to the more sketchy conceptualizations of the meso level in other economic theories? On closer inspection, there is very little difference, I think. Take the fallacy of composition. This is just a different way of acknowledging that the interdependencies of socially embedded economic agents implies interaction effects of their behaviour so that the macro level can simply not be explained as an aggregation of the micro level. And for herd behaviour, Keynes argued that it is the result of uncertainty – the fact that economic agents cannot escape the reality of having to make decisions under conditions of unknown unknowns. Hence, they try to find anchor points for their behaviour in what others do, or what they expect others will do, or what they expect that others expect may be rational to do. So if we take this condition of fundamental uncertainty to hold for socially embedded individuals more generally, then there is no real

The meso level in economics 135

difference between the Keynesian concept of herd behaviour in financial markets and the social economic concept of a guiding social norm emerging in a more general market context. The role of institutions in economic change and coordination can be linked to social capital features of trust and cooperation on the one hand and distrust and group conflict on the other hand. Whereas the first features enable change and coordination, the others constrain them and can keep an economy locked­in in a situation of stagnation. Similarly, the role of institutions in the capability approach can be linked to constraints for the individual members of particular social groups. Such constraints may make it harder for some to achieve their capabilities due to discrimination, social exclusion, or stigmatization of the groups that they belong to. To conclude, the meso level of the economy is the level where the behaviour of individual economic agents is influenced at an ontological level, whether within groups, between groups, or between loosely interrelated individuals through dominant social norms, in such a way that individual behaviour does not necessarily aggregate up to the macro level. The meso level thus, is the level where it is perhaps the most significant, and the least visible, that the economy is inherently social. For this reason, it is time that economics, as a branch of the social sciences, begins to take the meso level more seriously. The work by social economists, and by John Davis in particular, has thoroughly prepared the grounds for this task.

References Cumbers, A., Davis, J., and McMaster, R. (2015). Theorizing the social provisioning process under capitalism: Developing a Veblenian theory of care for the twenty-­first century, Journal of Economic Issues, 49(2): 583–590. Davis, J. (2003). The Theory of the Individual in Economics. London: Routledge. Davis, J. (2004). Collective intentionality, complex economic behaviour, and valuation, in J. Davis, A. Marciano, and J. Runde (eds.), The Elgar Companion to Economics and Philosophy. Cheltenham: Edward Elgar, pp. 386–402. Davis, J. (2014). Social capital and social identity: Trust and conflict, in J. Davis and A. Christoforou (eds.), Social Capital and Economics: Social Values, Power, and Social Identity. London: Routledge, pp. 98–112. Davis, J. (2015). Stratification economics and identity economics, Cambridge Journal of Economics, 36: 1215–1229. Davis, J. and Christoforou, A. (2014). Social Capital and Economics: Social Values, Power, and Social Identity. London: Routledge. Davis, J. and Dolfsma, W. (2015). Introduction to the first edition, in The Elgar Companion to Social Economics. 2nd edition. Cheltenham: Edward Elgar, pp. 4–9. Davis, J., Marciano, A. and Runde, J. (2004). Introduction, The Elgar Companion to Economics and Philosophy. Cheltenham: Edward Elgar. Dolfsma, W., and Dannreuther, C. (2003). Globalization, Social Capital and Inequality. Cheltenham: Edward Elgar. Dopfer, K., Foster, J., and Potts, J. (2004). Micro-­meso-­macro, Journal of Evolutionary Economics, 14: 263–279.

136  Irene van Staveren Fine, B. (2001). Social Capital versus Social Theory: Political Economy and Social Science at the Turn of the Millennium. London: Routledge. Folbre, N. and Nelson, J. (2000). For love or money – or both?, Journal of Economic Perspectives, 14(4): 123–140. Himmelweit, S. (2007). The prospects for caring: Economic theory and policy analysis, Cambridge Journal of Economics, 31(4): 581–599. Hodgson, G. (1993). Economics and Evolution. Bringing Life Back into Economics. Cambridge: Polity Press. Keynes, J. M. (1936). The General Theory of Employment, Interest, and Money. London: Macmillan. Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. New York: Cambridge University Press. Sen, A. (1987). On Ethics and Economics. Oxford: Basil Blackwell. Sen, A. (2009). The Idea of Justice. Cambridge, MA: Harvard University Press. Shiller, R. (2000). Irrational Exuberance. Princeton, NJ: Princeton University Press. Van Staveren, I. (2001). The Values of Economics: An Aristotelian Perspective. London: Routledge. Van Staveren, I. (2005). Modelling care, Review of Social Economy, 63(4): 567–586. Van Staveren, I. and Knorringa, P. (2008a). Beyond Social Capital: A Critical Approach. London: Routledge. Van Staveren, I. and Knorringa, P. (2008b). Unpacking social capital in economic development: How social relations matter, in Beyond Social Capital: A Critical Approach. London: Routledge, pp. 107–135. Veblen, T. (1899). The Theory of the Leisure Class. An Economic Study of Institutions, edited by B. W. Huebsch. New York: Viking.

9 Towards a caring economy Evelyn L. Forget

Introduction Zora, the Robot Caregiver, was introduced to the world by the New York Times in November 2018 (Satariano et al. 2018). She1 made her debut earlier in the year at Jouarre, a nursing facility an hour outside Paris that cares for patients with dementia. ZoraBots, her Belgium-­based creator, has delivered over one thousand units to health facilities around the world. Zora’s role as a caregiver is limited: she doesn’t take blood pressure, feed patients, or dispense medication. Rather she provides company to lonely seniors, leads exercises, and plays games. By all reports, patients are enamored. They jostle to spend time with her, cuddle and caress her, show jealousy when others monopolize her time, and sometimes tell her things they will not tell human caregivers. Some see Zora’s incursion into hospital caregiving as sad or even dystopian. Nurse Nathalie Racine, for example, is quoted saying she wouldn’t let a robot feed patients even if it could: “nothing will ever replace the human touch, the human warmth our patients need”. Zora epitomizes the conflict at the heart of caregiving. Supremely competent at her assigned tasks and loved by her patients, Zora cannot respond to human patients with empathy, and she is unburdened by the ethical and moral conflicts that sometimes emerge in the context of caregiving. How is caregiving provided by a machine different from human caregiving, and is health economics capable of articulating and capturing the difference? This chapter unfolds as follows. The next section demonstrates how health economics struggles to capture the nuances of caregiving, resulting in a systematic undervaluation of its importance. The following section summarizes the alternative model of caregiving offered by John B. Davis and Robert McMaster, based on the social embeddedness of both caregivers and care receivers. The concept of process utility is then introduced, and the feasibility of using standard metrics to capture the deontological aspects of care is examined from the perspectives of both the provider and the recipient. The concluding section provides some final reflections on future research directions.

138  Evelyn L. Forget

Health economics and the valuation of care There is no standard and cohesive definition of caregiving, which has been approached by different scholars with different disciplinary backgrounds. It is by its very nature heterogeneous and has been defined to include everything from the nurture of young children, to non-­market or household production of goods and services, to being kind, or sharing empathy and/­or sympathy and/­or compassion in the activities undertaken by the helping professions. In the context of health care, we will define caregiving as the discharge of acts and duties associated with health services delivery by people who undertake these acts in the performance of a specific role. Thus, it can include services provided by physicians and nurses, but also services provided by informal caregivers, including family members, to individuals requiring assistance because of any number of disparate health challenges. A well-­recognized “market failure” occurs in the provision of services to patients by physicians. Patients often have much less information about their health and the alternative treatments potentially available to them than does their physician. Patients (or principals) thus rely on physicians (or agents) to ensure that they receive appropriate care. Yet the self-­interest of the physician might encourage the physician to provide more care than is appropriate in order to bolster her income, or to provide lower-­quality services than are billed and expected. The relationship between physicians and patients is, indeed, the classical principal-­agent problem associated with market failure in health economics. What would prevent a physician from providing less than appropriate care for her patient? One way that economists deal with the market failure implicit in the principal-­agent problem is to introduce incentives or constraints into the market relationship such that the physician is encouraged to provide the “appropriate” level and quality of care. That is, the “quality” of care is treated as an externality that must be internalized in order to correct market failure. A second response is through regulation of one type or another: there might be “best practice guidelines” that serve as soft regulations for physicians concerned about how they are perceived by colleagues or employers, or there might be oversight committees charged with ensuring that prescription guidelines, for example, are met in certain facilities. Again, the “appropriate” level and quality of care is treated as exogenously determined. A third response that seems, on the surface, to deal with the deontological aspect of care is to invoke “medical ethics”. While medical ethics are sometimes codified in practice guidelines or other formal regulations of institutions, often individual health care providers in a formal setting invoke their own understanding of medical ethics to shape their behavior. In the latter case, the well-­being of the patient enters into the utility-­maximizing behavior of the physician in a variety of ways that encourage other-­regarding decisions. The behavior of a physician might lead to optimal patient treatment because the physician expects some future benefit from behaving ethically or

Towards a caring economy 139

altruistically, or because the physician derives direct personal satisfaction from the perceived well-­being of the patient, or simply because the physician derives satisfaction from behaving in ways consistent with his or her own moral code (Khalil 2003). Whatever the motivation of the care provider, this explanation relies upon the internalization by the physician of an ethical code to ensure that patients receive optimal care. Health economics recognizes that unconstrained markets do not necessarily ensure that the “appropriate” level and quality of care are provided by those charged with the task. Therefore, either incentives must be built into markets to ensure that maximizing behavior on the part of the care provider is consistent with optimal patient care, or the market must be constrained to ensure that the appropriate level and quality of care is delivered. These constraints operate either through regulations of some sort that constrain physician behavior or through self-­regulation captured under the label of medical ethics. However, none of these approaches adequately captures one very significant characteristic of caregiving: the process of giving and receiving care might matter as much as the outcomes of care. In health economics as it is usually practiced, only the consequences of care seem to matter. The patient receives optimal care when the level and quality of care meet an exogenously determined “appropriate” standard. There is no room for the development of a changing social relationship between the carer and the cared-­for. The behavior of both patient and caregiver are determined only by individual optimization subject to constraints, with the acknowledgement that the well-­being of a patient might be an argument in the utility function of the caregiver. The consequentialist approach of standard analysis cannot capture the process of giving and receiving care, and this has important policy implications. If only the outcomes of care matter, resources are likely to be allocated on the basis of outcomes that can be measured. In the aggregate, measurable outcomes are optimized using standardized procedures. Standardized procedures govern the ways that caregivers can deliver services, shape educational practices that produce formal caregivers, and influence institutional arrangements that constrain and reward particular behaviors. Policy based on this approach, therefore, undervalues aspects of the process of caregiving when these cannot be easily measured or standardized. Consequently, the emotional, personal, and human aspects of caregiving will be crowded out by algorithms that optimize what we can measure.

The relational aspects of care and the socially embedded individual John B. Davis and Robert McMaster focus on the relational aspects of caregiving and argue that because individuals are socially embedded and care is provided in the context of a network of personal relationships and institutions, the process of care delivery must be explicitly considered (Davis and McMaster

140  Evelyn L. Forget

2013, 2017). They argue that none of the attempts to deal with the relational aspects of caregiving offered by health economics is adequate. They claim that patients and caregivers should be regarded as socially embedded individuals, reflecting their participation in family, community, and professional social groups that, together, take on responsibility for patient care. These social groups, in turn, are nested in social relationships that bring other social groups into indirect contact with the patient-­caregiver relationship. Davis defines such social groups formally as collections of individuals with shared characteristics that specify membership in those groups, which creates sets of rights and responsibilities based on individuals’ collective intentions (2003, 2011). By contrast, institutions are durable systems of rules that structure social interactions between and within groups, made manifest in enabling and constraining behavior (Hodgson 2008). Health care systems are networks of institutions and social groups. Consequently, individuals as members of many social groups interact directly and indirectly to provide care. The provision of care thus becomes a social act between socially linked individuals for whom it is rational to act on the basis of shared goals. Shared goals imply mutual obligations that structure decision-­ making. The rationality of this process, according to Davis and McMaster, is based not on the intended outcomes of these decisions, which reflects a consequentialist approach, but rather on the shared commitment of caregivers, families, and patients to the process of delivering a package of care to which they have freely agreed. In standard analysis, person-­centered care is ensured through constraints which are external to the (perhaps) interdependent utility functions of caregivers and patients. By contrast, when individuals are seen as socially embedded, person-­centered care becomes internal to the ongoing (and presumably evolving) relationships between individuals. Davis and McMaster argue that standard analysis imagines atomistic individuals who formulate intentions that apply only to themselves, and their choices are the instruments by which they maximize their individual utility. By contrast, socially embedded individuals formulate intentions that apply to themselves via the relationships they have with others. They go further and argue that these relationships govern care even in cases where patients are not competent to decide for themselves, because others in the patient’s social group act on behalf of the patient in ways consistent with their shared intentions. This model, they argue, captures important non-­instrumental qualities of care because it is based on a deontological rather than an instrumental rationality.

Process utility, relationships, and the quality of care Davis and McMaster focus on the deontological aspects of caregiving – the inherent morality of the process of caregiving itself. This they find primarily in the relational aspects of providing care, but there are other aspects of the process of caregiving not captured in the consequentialist approach of standard analysis. This section builds on the Davis-­McMaster analysis to ask whether

Towards a caring economy 141

the concept of process utility can capture these effects. If so, can we imagine a way to test whether process utility exists as an entity distinct from outcome utility in decision-­making, and to determine whether there are metrics that might capture it? Caregiving is as much about the way in which services are provided as it is about the services themselves. For example, Zora the robot caregiver can be programmed to “listen” and “respond” to patients who talk to her, but she cannot be programmed to be kind. She can be programmed not to physically harm patients, or to gesture in particular ways in response to certain verbal prompts, but she cannot be programmed to empathize, because that involves a human response. The services she provides will have particular consequences and can therefore be evaluated in terms of effectiveness and efficiency. These can even include scores that represent a patient’s or a family’s satisfaction with her services. However, there can be no relationship between caregiver and receiver – at least from Zora’s perspective. Care provided by humans also has two aspects. A nurse might provide all the necessary and expected services to a patient in a brusque and distant way. A loving husband might care deeply for an aged spouse, devoting all his time and energy to her care, yet perform specific tasks in ways that are demonstrably inferior to the same services provided by professionals. The consequences of the services provided, in both cases, may be at odds with the way the process of service provision is experienced by either or both the caregiver and the recipient. Utility may be derived from the consequences of caregiving as in the standard analysis or from the ongoing act of caregiving. This latter is referred to as process utility. Davis and McMaster acknowledge and reject attempts to model their relational analysis of caregiving by using the concept of process utility. They acknowledge Wiseman (1997), who argues that individuals can be conceived of as having two utility functions: one atomistic and the second from group or social activities. Wiseman claims that this goes beyond the standard analysis, which can only accommodate sympathy as an argument in the carer’s utility function. Mooney (2009) builds on this idea by arguing that a social welfare function may be conceived as something more than the aggregation of individual utility functions. He uses the social welfare function to explain altruistic acts that may seem irrational from an individual perspective but can nevertheless be viewed as rational from the perspective of social utility. Mooney foreshadows the language invoked by Davis and McMaster, arguing that care takes place in a community, and he argues that process utility can capture the contextual aspects of care. But Davis and McMaster ultimately reject Mooney’s argument for two reasons: first, they argue that in the case of dual utility functions, it remains possible to create a meta-­function that allows trade-­offs between the well-­being of the caregiver and the patient (cf. Brennan 1993), which they believe to be inconsistent with their analysis; second, they claim that Mooney’s approach relies exclusively on consequentialist reasoning.

142  Evelyn L. Forget

Both of these claims can be questioned. While Davis and McMaster reject trade-­offs between the well-­being of caregivers and patients, a model of caregiving that explicitly allows for the possibility of such trade-­offs makes it possible to examine critically the gendered and socially constructed aspects of the ever-­changing relationships at the heart of care provision. Moreover, it is possible to conceptualize process utility as distinct from outcome utility and even to develop tests for its existence using the tools of standard economic analysis. The possibility of measurement itself does not imply that what is being measured is an outcome; aspects of process can be measured as well, particularly if we recognize that the quality of care – the “value of care in its own right” – might be perceived quite differently by the caregiver, the patient, the interested bystander, and a hypothetical observer. Should we allow trade-­offs?

Davis claims that social groups are collections of individuals with shared characteristics that specify membership in those groups, which creates sets of rights and responsibilities based on individuals’ collective intentions (2003, 2011). However, a “collective intention” does not necessarily imply homogenous individual interests, either with respect to the process or to the outcome of caregiving. Nor have the “shared characteristics” of a social group been well articulated. If the process of care is conceptualized as unfolding in real time, the stability of group membership must also be considered. That is, even though group members freely agree to abide by the rights and responsibilities of group membership, each individual’s contribution to furthering the group’s “collective intentions” will be colored by the trade-­offs that the individual is prepared to make, at a particular point in time, between all the various factors that influence that individual’s lived experience. Membership in various social groups will also evolve over time, as individuals renegotiate their social context. Consider, for example, one very common caregiving activity. An elderly husband with dementia is being cared for at home by his equally elderly wife, and both have a range of physical health challenges. The grown children of the family are devoted to the well-­being of both parents, and professional caregivers in the form of physicians and community workers offer support within the contexts of their institutional roles. Should the father be institutionalized? The elderly parents are dependent on one another and the wife receives true satisfaction caring for her husband. The father resists placement in a facility, partly for purely egoistic reasons and partly because he knows, during moments of lucidity, how much his wife relies on his presence to maintain her own mental acuity and sense of well-­being. The daughter fears that her mother will suffer from the increasing physical burden of providing care and is conflicted by competing demands on her time that make it impossible for her to provide as much support as she would like to provide, while the son medicates his worry and mentally distances himself from the decision-­making. Is this a social group? Is the son “in” or “out”, and can he decide to change his status over time? The

Towards a caring economy 143

“collective intention” is to provide for the best care for both parents subject to a range of constraints. However, as soon as we make these intentions and constraints concrete, the tensions within and between each member of the social group become apparent. Davis and McMaster recognize the “gendered” aspects of caregiving, but this too requires unpacking. Most informal care in high-­income countries is provided by women, and women in these same countries have only in recent decades begun to close the gap in earnings in the paid labor market. Even if our example takes place in a country that provides a social safety net that makes it possible for an informal caregiver to take time off work to provide care for elderly parents or relatives with disabilities, should she do so? From a purely egoistic perspective, fewer demands on her time and attention would be desirable. How much current income is she prepared to give up to simplify her life? The well-­being of her parents is of paramount importance, but does she have the skills to care for her father as well as a professional caregiver can in a residential setting? Will her altruism risk his health, or that of her mother? Moreover, time off work today has negative implications for her own future career progress, and might also have collective implications. After all, every daughter who takes time off work to selflessly care for her aged parents reinforces the status quo of gender relations that reinforce social attitudes that describe informal caregiving as “women’s work”. This decision arguably reduces the collective well-­being of women workers – a social group that is indirectly related to the care decision of her father – while enhancing the well-­being of her parents and having an indeterminate effect on her own well-­being. It is essential to consider trade-­offs between the perceived well-­being of individuals within a social network as well as the tensions inherent in negotiating the multiple roles that anyone plays in society. Can process utility capture the value of “care in its own right”?

Davis and McMaster suggest that process utility is, like standard health economics, ultimately consequentialist because it relies on measurable outcomes. This, they claim, implies an instrumental approach in which care activities are reduced to a series of mechanical tasks that neglect individual circumstances and the broader social implications of caregiving. It is certainly the case that a focus on measurement draws attention to factors that can be measured and away from the less tangible aspects of care. However, measurement itself does not imply a narrow focus on ultimate consequences. One might imagine aspects of the process of caregiving that are measurable. For example, if the social aspects of caregiving are paramount, as Davis and McMaster claim, then one might imagine a periodic survey delivered to all members of the social groups directly involved in a patient’s care that asks explicitly about social relationships and social well-­being. There are a number of validated instruments that capture different aspects of social functioning, and one might measure, for example, changes in social well-­being as care unfolds

144  Evelyn L. Forget

in real time. The informal care provided by family members for a child with complex medical conditions might be expected to have ongoing effects on the resilience of a marriage, which will change over time. Similarly, such care will affect the perceived well-­being of siblings as well as the nature of the relationship between the child and his or her parents. These are difficult qualities to measure, but psychometrists have developed and validated a number of instruments designed to do precisely that. The various characteristics will depend on individuals and context, and no one would expect them all to indicate equal amounts of improvement or deterioration over time or even to necessarily be consistently answered by different parties to the relationship. Just as the well-­ being of different family members might be traded off against one another, so too might the process utility be experienced differently by different parties to the process, even though they all share in the collective intention that governs care. Social aspects of caregiving are prioritized by Davis and McMaster, but these are not the only aspects of process utility we might want to measure. Simple satisfaction with different interventions over the process of care would be usefully measured. In the case of care provided by a robot caregiver, for example, it would be interesting to know how the patient perceives the quality of care at different levels of dementia, as would it be useful to know if the misgivings of family members or onlookers are shared by care recipients. One could attempt to disentangle the complexities of loving but incompetent care provided by a family member, or the efficient but impersonal interactions between a patient and a busy health care worker. One way to peer into the complexity of process analysis is to imagine how one might use empirical evidence to determine whether process utility and outcome utility both exist. C. Donaldson and P. Shackley (1997) attempted to determine whether process utility actually exists by offering patients different descriptions of a laparoscopic versus a conventional cholecystectomy. They offered a random selection of patients on a hospital waiting list for cholecystectomy a description of the intervention that was either based solely on outcomes or one that emphasized process as well as outcomes. They then determined how much each patient was willing to pay to receive laparoscopic surgery. They followed up with a survey that asked patients to rate the reasons for their decision, some of which emphasized process and others outcomes. The idea was to determine whether those patients primed to consider process utility would be more conscious of such factors when making their decision, and whether such patients might be prepared to pay more for laparoscopic surgery based on a less onerous “process”. Their results did not support either hypothesis. On the basis of their results, they were unable to reject the hypothesis that process utility does not exist, but their discussion of the results yields interesting information. It is, they claimed, difficult in practice to distinguish between “outcomes” and “process”, particularly in this kind of an intervention. This difficulty is, perhaps, what motivates Davis and McMaster to claim that process utility is, ultimately, reductionist and consequentialist.

Towards a caring economy 145

Brouwer et al. (2005), however, used the concept of process utility to capture the benefits of caregiving to those who provide care and, unlike Donaldson and Shackley, they were able to demonstrate that process utility does exist for caregivers. Based on a sample of 950 Dutch volunteer caregivers, they measured process utility as the difference in perceived happiness between the current situation where the volunteer provides care and an alternative (hypothetical) situation where, all else equal, someone else provides care. The results show that process utility exists and is substantial. Almost half of the volunteers (48.2%) reported that they derive direct satisfaction from providing care, and on average happiness would decline if informal care tasks were handed over to someone else. Multivariate regression analysis shows that process utility especially relates to caregiver characteristics (age, gender, general happiness, relation to patient and difficulties in performing daily activities) and subjective caregiver burden, and also depends on the number of hours of care provided (objective burden). These attempts to isolate and measure process utility are interesting, and especially so because they demonstrate that the problem that Davis and McMaster address (caregiving) is particularly relevant to the concept of process utility. Surgery is often a one-­time intervention where the time lapse between the initiation and end of the intervention is negligible, especially from the point of view of a patient who spends most of it in an unconscious state. By contrast, protracted, ill-­defined, and informal caregiving is an intervention in which both care providers and care recipients can more easily distinguish between process and ultimate outcomes because caregiving takes place over a measurable period of time long enough for both caregivers and recipients to recognize and reflect upon changes in well-­being. A consideration that caregiving occurs in real time also has other analytical implications. The allocation of peoples’ time and attention is essential to some kinds of tasks, and reducing the time spent on them reduces the value of the work. Caring is a good example; the time spent on the process of care is directly correlated with both the quality of care received and also the quality of the caregiver’s experience. In the production of material goods, by contrast, the value of the work is embodied in the output produced. Therefore, reducing the amount of time spent producing the same output – the very definition of an efficiency improvement – makes us better off. This reasoning could even be extended to surgical interventions. More laparoscopic cholecystectomies and fewer conventional surgeries will improve the bottom line of the health care organization; in the aggregate, better outcomes will be achieved at lower cost. But for care workers, this is not the case. The time spent caring is the product of the labor; spending less time on care means that less care is provided, and the care provided yields a poorer experience for the caregiver as well as the patient. It makes sense to enhance and protect the value of care as well as the experience of the caregiver. A recognition of process utility for tasks that unfold over a protracted period of time is something health economics has not undertaken.

146  Evelyn L. Forget

Towards a caring economy Davis and McMaster develop important insights regarding the limitations of standard economic analysis of caregiving in the context of health services delivery, and they identify the relational setting that characterizes care delivery to and by socially embedded individuals. The purpose of this chapter was to identify some ways in which the Davis and McMaster model might be developed to better characterize the fraught nature of caregiving in a world of multiple relationships and multifaceted individuals simultaneously playing multiple roles. A more complex model can provide a better foundation for policy analysis and an entry into a gendered analysis. The Davis-­McMaster model recognized that care is provided and experienced by socially embedded individuals, but it was ultimately unable to capture the tension that is endemic in the provision of informal caregiving. Their characteristic representation of caregiving, while intended to capture aspects of caregiving more generally, was made concrete in a portrait of the delivery of health services to individual patients by health professionals. By making this choice, the complexity of decision-­making inherent in informal caregiving – where standards of care delivery are not well-­developed, and roles and relationships are created and transformed through the process of care delivery – is relegated to a secondary role. However, it is by considering informal caregiving in its full complexity that some ways forward emerge. Two aspects of the Davis-­McMaster model must be further developed. First, contrary to their claim, it must be possible to consider explicitly tensions and trade-­offs both within individuals and between individuals in the same social groups, as well as between social groups. It must be possible for a physician to be torn between optimizing work-­life balance and providing selfless care, or for the same physician to consider the trade-­offs between patient care, financial returns, and the regard with which that physician is viewed in the larger community. It must be possible for a caregiver to resent the emotional labor that simultaneously provides that caregiver with great satisfaction. We must be able to model a grown child who struggles with the decisions that pit the well-­being of family members against personal achievement. We are all, simultaneously, other-­oriented and egoistic. The idea of a “shared commitment to intentions regarding patient care”, freely entered into by members of a social group that then takes on responsibility for care, captures the social aspects of care, but does not fully capture its human complexity. Second, the Davis-­McMaster model needs to be extended to more fully consider the process of care that unfolds in real time. The social nature of care delivery is one aspect of the process of care, but there are others. By extending the range of considerations, we more fully recognize the importance of time spent in care delivery. While obvious at one level, a focus on the time spent in caring activities has important implications both for policy and for analysis. Analytically, process utility becomes a concept that might be better developed to explain caregiving. By articulating differences in the kinds of care

Towards a caring economy 147

provided, and the types of caregiving arrangements under consideration, Davis and McMaster might find a role for process utility that is analytically distinct from consequentialism. The Davis-­McMaster model goes a long way towards helping us to understand the limitations of existing analyses that lead, inevitably, towards a reductionist view of caregiving in our society. The essence of care is the time and social interaction spent in its delivery and, notwithstanding the current achievements and future potential of Zora, caregiving remains a very human activity.

Note 1 Actually, her gender is indeterminate. Some patients call Zora “her” and others “him”.

References Brennan, T. J. (1993). The futility of multiple utility. Economics and Philosophy, 9: 155–164. Brouwer, W. B., van Exel, N.J.A., van den Berg, B., van den Bos, G. A., and Koopmanschap, M. A. (2005). Process utility from providing informal care: The benefit of caring. Health Policy, 74(1): 85–99. Davis, J. B. (2003). The Theory of the Individual in Economics: Identity and Value. London: Routledge. Davis, J. B. (2011). Individuals and Identity in Economics. Cambridge: Cambridge University Press. Davis, J. B. and McMaster, R. (2013). Situating care in health economics. Adam Smith Research Foundation, 02. Glasgow. https://­pdfs.semanticscholar.org/­ebfe/­98247253a0e44 c384ed73c52a85d85de25ff.pdf. Davis, J. B. and McMaster, R. (2017). Health Care Economics. London: Routledge. Donaldson, C. and Shackley, P. (1997). Does “process utility” exist? A case study of willingness to pay for laparoscopic cholecystectomy. Social Science & Medicine, 44(5): 699–707. Hodgson, G. M. (2008). An institutional and evolutionary perspective on health economics. Cambridge Journal of Economics, 32: 235–256. Khalil, E. L. (2003). What is altruism? Journal of Economic Psychology, 25: 97–123. Mooney, G. (2009). Challenging Health Economics. Oxford: Oxford University Press. Satariano, A., Peltier, E., and Kostyukov, D. (2018). Meet Zora, the robot caregiver. New York Times (November 23). www.nytimes.com/­interactive/­2018/­11/­23/­technology/­ robot-­nurse-­zora.html. Wiseman, V. (1997). Caring: The neglected health outcome? Or input? Health Policy, 9: 43–53.

Index

Adelman, Frank 43 Adelman, Irma 43 African American middle class 11, 89 – 98 agency 57, 61, 128 – 129 agent-based complexity 76 – 77 agent-based models 106 – 107, 111, 116 agents 8, 104, 112, 128 – 129, 134; behaviour of 74; identity of 52 – 55, 62 – 63; reflexive 37, 67 – 82 airline industry 27 – 28 Akerlof, G. 52 – 54 altruism 99, 113, 141 analytical tractability 45, 46, 47 Antonovsky, Aaron 58 Aquinas, Thomas 60, 61, 62, 63 Aristotelian hylemorphism 57 – 61, 64n8 Aristotle 7, 9, 52, 58 – 61 Arrow, Kenneth 12, 100 Association for Social Economics (ASE) 3 automobile industry 27 – 28 “average individual” 11 banks 28 – 29 basic capabilities 10 Becker, Gary 100, 105 behaviour: explanations for 111 – 113; expression of 112 behavioural economics 77, 99, 102 – 105, 109 behavioural finance 36 behaviour pattern tests 45 Berg, Nathan 102 Black Bourgeoisie (Frazier) 11, 90 – 95 black-owned businesses 92 Blaug, Max 105 body 59 – 61 bonding social capital 132, 133 Bostock, David 59 Boulding, Kenneth 9 bourgeoisie 90 – 96 breaking frame 70

bridging social capital 132, 133 British industry 27 business cycle, model of 46 business psychology 20 Cabral, Amilcar 97 – 98 calibration 6, 44 – 45 Camerer, Colin 103, 105 capabilities approach 9 – 11, 55 – 56, 126 – 127 capitalism 94 – 95 care 11 – 13; health care 138 – 139 care/caregiving 13; aims of 10; deficit 10; duty of 114 – 115; failure of 10; gendered aspects of 143; health care 10, 12 – 13, 58; quality of 138, 140 – 145; relational aspects of 139 – 140; valuation of 130 – 131, 138 – 139, 143 – 145 caring economy 137 – 147 Cartesian dualism 57 – 58 catastrophes 76 Catholic Economics Association (CEA) 2 – 3 chaos 76 choice 111 – 112; Keynes’s approach to 35 – 37; rational 36, 52 circular explanations 53 class 90 – 91 classical economics 100, 109 Coase, Ronald 74, 111 cognitive function 77 coherence 80 Colander, David 11, 100 – 102, 107, 109 collective intentionality 9, 142 – 143 Colonel’s dilemma 74 – 75 colonial gaze 95 colonial studies 11, 95 communication skills 4 community economy 12, 127

150 Index complexity 5, 7 – 8, 68, 111; computational 75 – 76, 82; dynamic 76, 80; hierarchical 76, 82; reflexive 75 – 82 complex systems 68 computational complexity 75 – 76, 82 computational experiments 41 – 49 computer simulations 6 – 7, 41 – 49 consequentialism 9, 13 consistency 80 Consistency Supposition 74 constructivism 47 consumer goods 33 conventions 36 – 37 cooperation 135 covering-law explanations 6, 42 – 43 critical theory 89 – 91 Cropsey, Joseph 64n9 cumulative causation 77 cybernetic loopiness 69 cybernetics 76 Darity, William 10 Darwin, Charles 114 Davis, John Bryan: on behavioural economics 109; biographical outline of 2 – 4; on caregiving 139 – 147; on change 109; on complexity 75 – 77; contributions of 1 – 13; on human agency 57, 61; on identity 51 – 64; individual identity and 51 – 53, 128; on market conventions 36 – 37; on reflexivity 68, 71 – 75; on simulation 41 – 42; on social capital 133; on social groups 128 – 129; on social norms 130; on stratification economics 90 Day, Richard 76 decision-making, Keynes’s approach to 35 – 37 deductive logic 42 Dempsey, Bernard 3 Descartes, René 59, 70 Devine, Thomas 3 Dhami, Sanjit 103 dignity 8 – 9, 11 direct structure tests 45 diversification 17 – 18 Dow, Sheila 102 Drawing Hands (Escher) 8, 69 – 71, 78 – 80, 82 Droste effect 82n3 dynamic complexity 76, 80 dynamic stochastic general equilibrium (DSGE) models 107 EAEPE-Myrdal Prize 3, 13n1 Earl, Peter 102 – 103

economic methodology 5 – 7, 74; introduction to 41 – 43; simulation and 41 – 49 economic models 6 – 7, 43 – 48, 101 economic philosophy 4 – 8, 13n2 economics: behavioural 77, 99, 102 – 105, 109; classical 100, 109; complexity/reflexivity in 5, 7 – 8; debates in 99 – 101; evolutionary 126; experimental 11, 99; happiness 11, 99; health 12 – 13, 138 – 139; heterodox 8, 9; identity in 5, 7, 51, 52 – 55, 99; individual in 9, 11 – 12, 128; institutional 12, 125 – 127; John B. Davis’s contribution to 1 – 13; mainstream 3, 5, 8, 9, 11 – 12, 99 – 101; meso level in 12, 124 – 135; neoclassical 11 – 12, 51, 53, 100, 109; neuro-economics 11, 99, 105 – 106; professionalization of 109; social 8, 12, 127 – 133; specialism within 110; stratification 89 – 91, 96, 129; welfare 111 Eddington, Arthur 111 efficient market theory 20, 77 Einstein, Albert 110 – 111 emergence 68 emergent phenomenon 45 – 46 emotions 114 epistemic reflexivity 73 – 74, 81 epistemology 5 Epstein, Joshua 42 equilibrium 10, 100, 101; general 12, 101, 106, 107 Escher, M. C. 8, 69 – 71, 78 – 82, 83n7, 83n8 ethics 5, 114 – 116; medical 138 – 139 ethnomethodological reflexivity 71 European Association for Evolutionary Political Economy (EAEPE) 3 European Society for the History of Economic Thought 4 evolutionary economics 126 existence proofs 6 expectations 125 expected utility theory 52, 104, 106 experimental economics 11, 99 fallacy of composition 124 – 125, 131, 134 falsifiability test 76, 104 – 106 feedback loops 68 Finnis, John 7, 52, 60 – 63 first-order logic 47 flexibility, in investing 34 Formalist Revolution 99, 108 Fox, Justin 107

Index  151 Franklin, Allan 44 Frazier, E. Franklin 11, 90 – 93, 95, 96 Frede, Michael 59 – 60 Fregean theory 53 Frisch, Ragner 46 functionings 9 – 10 gambling 19 game theory 11, 74 – 75, 99, 101, 106 – 108 gender 11, 143 general equilibrium 12, 101, 106, 107 General Theory (Keynes) 19 – 20, 36, 67 – 68, 77, 82 generative explanations 42, 45, 47 Gide, André 73 Gigerenzer, Gerd 102 Gintis, Herbert 103 Gödel, Kurt 7, 76, 80 – 82, 83n11, 83n12 gold 22, 25 Great Recession 75 – 77, 107, 108, 117 Grisez, Germain 62 group conflicts 129 guilt 114 Hands, D. Wade 67, 71, 73 – 74 happiness economics 11, 99 Hare, Richard M. 113 Hausman, Dan 48 health care 10, 12 – 13, 58 health economics 138 – 139 health outcomes, socioeconomic status and 97 herd behaviour 12, 20, 124, 125, 131, 134 – 135 hermeneutical reflexivity 70 heterodox economics 8, 9 hierarchical complexity 76, 82 History of Economics Society 3 – 4 Hodgson, Geoffrey 77 Hofstadter, Douglas 68, 69, 73, 78 – 80, 82 Homo economicus 9, 105 human agency 57, 61, 128 – 129 human dignity 8 – 9, 11 human motivation 112, 113, 130 – 131 humiliation 8 hylemorphism 7, 57 – 61, 64n8 identity 51 – 64; Aristotelian approach to 52, 57 – 61; complex view of 55; Davis on 55 – 56; Finnis on 61 – 63; introduction to 51 – 52; Locke’s conception of 53, 57; phenomenological approach to 56 – 57; reflexivity and 51, 56; Sen’s theory of 54, 112; social 53 – 54, 56, 129; theories of, in economics 52 – 55

identity economics 99 “I” language 9 imitation economy 43 – 44 immanent reflexivity 72 – 74 incompleteness theorem 7, 76, 80 – 82, 83n11, 83n12 indirect proof 47 individual: “average” 11; capabilities approach to 55 – 56, 126 – 127; in economics 9, 11 – 12, 128; interdependencies and 131 – 133; socially embedded 10 – 12 individual dignity 8 – 9, 11 individual identity in economics 5, 7, 51, 52 – 55, 99; see also personal identity Individuals and Identity in Economics (Davis) 3, 5 individuation criterion 7, 55 industrial goods 33 infinite do-loops 73, 76, 79 – 81 informed knowledge 34 institutional economics 12, 125 – 127 institutionalism 9 institutions, role of 135 instrumentalism 9 insurance 28 – 29, 33 interdependencies 131 – 135 International Association for Feminist Economics 4 International Network for Economic Method 4 interpretative reflexivity 70 – 71 investment, vs. speculation 19 – 20, 35 investment philosophy: credit-cycle approach to 17; diversification 17 – 18; of Keynes 6, 17 – 38 investment trusts 28 – 29, 32 irrational behaviour 125 Israel, Giorgio 47 Jaworski, William 59 Journal of Economic Methodology 2, 3, 4 Joyce, Richard 113 Kahneman, Daniel 103 Kaldor-Hicks efficiency 9 Kant, Immanuel 54 Keynes, John Maynard 5, 6, 67 – 68, 77, 82, 82n1, 124 – 125, 134 – 135; approach to decision-making and choice 35 – 37; as institutional investor 18; investment philosophy of 6, 17 – 38; London Stock Exchange investments by 21 – 29; New York Stock Exchange investments by 29 – 33; on speculation vs. investment 19 – 20

152 Index Keynes Papers 18, 38n3 Kirman, Alan 54 – 55, 107 Klaes, M. 71 – 75 Klein-Goldberger model 43 knee-jerk 69 knowledge: informed 34; limits on 34 – 35 Kranton, Elizabeth 52 – 54 labour markets 126, 127 Lange, Oskar 12, 100 Las Meninas (Velásquez da Silva) 71 – 73, 81, 82n3 Lawson, Tony 101, 117 Lectures on the Stock Exchange (Keynes) 19 Lee, Fred 117 Leombruni, Roberto 107 liar’s paradox 73, 75 libertarian paternalism 103 Livet, Pierre 55 Locke, John 53, 57 logic: deductive 42; first-order 47 London Stock Exchange (LSE) 21 – 29 “lower ladder” phenomenon 96 Lucas, Robert 43 – 44 Lynch, Michael 69 – 71 Mackie, John L. 113 macro level 124, 125 – 126 mainstream economics 8, 9, 11 – 12, 99 – 101 Mäki, Uskali 71, 74 “malignant” paradoxes 81 manipulative function 77 market conventions 36 market economies 127 market relations 8 Marquette University 4 Marx, Karl 5, 67 Marxist theory 91 mathematical models 101 Max U 12, 99 – 118; introduction to 99 – 101; morality and 113 – 116, 117; replacement of 109 – 111; survival of 101 – 108; unfalsifiability of 104 – 105, 109 – 111, 116 McCloskey, Deirdre 12, 100, 101 McMaster, Robert 139 – 147 mechanical reflexivity 69 – 70 medical ethics 138 – 139 Melber, Henning 90, 94 Merton, Robert K. 67, 77 meso level 12, 124 – 135; in economics 133 – 135; forms of 127 – 133; schools of thought on 124 – 127

metals 27 meta-theoretical reflexivity 70 methodological assumptions 8 methodological reflexivity 70 methodological self-congratulation 70 methodological self-consciousness 70 methodological self-criticism 70 microeconomics 101 – 102 micro level 12, 124, 125 middle class: African American 11, 89 – 98; locating 94 – 95; subaltern 11, 89 – 98 Mills, C. Wright 91 mining 32 – 33 Mirowski, Philip 71, 74 – 75 mise en abyme 73 mixed difference-differential equations 46 mixed-market economies 130 Mlinar, Ivana Anton 56 – 57 models 6 – 7, 43 – 48, 101; agent-based 106 – 107, 111, 116 Monte Carlo simulations 46 morality 113 – 117, 134 moral judgments 12, 113 – 114, 117 moral relativism 115 moral rules 115 Morgan, Mary 49n7 motivation 112, 113, 130 – 131 multiple selves problem 53, 55, 56 Nash, John 82n5 National Bureau of Economic Research (NBER) 43 native bourgeoisie 95 – 96 nativization 95 neoclassical economics 11 – 12, 51, 53, 100, 109 neoclassical theory 5, 7 net worth 93 neuro-economics 11, 99, 105 – 106 “new behavioral turn” 11 new constructivism 47 Newtonian physics 110 – 111 New York Stock Exchange (NYSE) 29 – 33 normative rules 115 norms 12 Nussbaum, Martha 9, 10 objectification 70 – 71 Oderberg, David S. 57 – 58 oil sector 33 Pareto criterion 9 parsimony criterion 105

Index  153 payoff maximization 104 performativity 83n6 personal identity: construction of 51, 56; Davis on 55 – 56; Finnis on 61 – 63; introduction to 51 – 52; theories of 52 – 55; see also identity petty bourgeoisie 91 philosophical self-reflection 70 philosophy of science 41 physics 110 – 111 pluralism 109 policy design 114 – 115 political economy 8 Pollak, Robert 105 Popper, Karl 67, 68, 76 – 77, 105 post-colonial studies 95 Potts, Jason 108 power structures 10 principal-agent problem 138 Print Gallery (Escher) 8, 78 – 80, 81, 82, 83n7, 83n8 probability 36 process utility 140 – 145 proofs, in economic models 6 – 7 psychology 102 – 103 Rabin, Matthew 102 race 11, 90 – 91, 97 racial wealth gap 92 – 94 radical referential reflexivity 70 railways 28, 33 RAND Corporation 74 rational choice 36, 52 rational expectations hypothesis 74, 107 rationality 100, 101, 105, 106, 125 reflections ad infinitum 69 reflexive agents 37, 67 – 82 reflexive complexity 68, 75 – 82 reflexive objectification 70 reflexive social construction 70 reflexivity 5 – 8, 42; Davis on 71 – 75; definition of 67; epistemic 73 – 74, 81; ethnomethodological 71; forms of 69 – 71; hermeneutical 70; identity and 51, 56; immanent 72 – 74; interpretative 70 – 71; introduction to 67 – 68; mechanical 69 – 70; meta-theoretical 70; methodological 70; radical referential 70; standpoint 70; substantive 70; systemic 70; transcendent 73, 74, 81 re-identification criteria 7, 55 relationships 140 – 145 representation 70 – 71

Review of Social Economy 2 – 3, 4, 117 Ricardo, David 5 Richiardi, Matteo 107 rigor 47 Rocking Horse Model 46 Roemer, John 100 Rosser, Barkley 76 Russell, Bertrand 73, 80 – 81 salutogenesis 58 Samuelson, Paul 104 Sandel, Michael 54 Sargent, Thomas 74 Schelling simulation 46 Schumpeter, Joseph 100 scientific revolution 110 self-fulfilling prophecies 67, 77, 82, 82n1 self-interest 112 selfishness 101 self-narratives 7, 51, 56, 74 – 75 self-referencing 68, 72 – 73, 77, 80 – 82, 83n12 self-reflection 69, 70 Sen, Amartya 9, 10, 12, 54, 101, 111 – 112, 117, 126 Sent, Esther-Mirjam 71, 74 shipping sector 28 Simon, Herbert 76, 102, 103 simulation-based explanations 6 – 7 simulations 41 – 49 Slutsky compensation criteria 9 Smith, Adam 12, 112, 113, 115, 130 Smith,Vernon 104, 112 – 113 social capital 12, 131 – 133, 135 social coherence 131 social economics 8, 12, 124, 127 – 133 social economy 12 social groups 128 – 129, 134, 140, 142 – 143 social identity 7, 53 – 54, 56, 129 socially embedded individual 10 – 12, 139 – 140 social norms 126, 130 – 131, 134 – 135 social science 42 social stratification 10, 11 social systems 89 social values 10 socioeconomic status, health outcomes and 97 soft raw materials 29 Soros, George 8, 67 – 68, 75 – 77, 82, 82n4 soul 7, 59, 60 – 61 sound micro-foundations project 99 speculation 125; vs. gambling 19; vs. investment 19 – 20, 35

154 Index Sraffa, Piero 5 standpoint reflexivity 70 stock exchange, Keynes’ investment practices in 6, 17 – 38 strange loops 78 – 80 stratification economics 12, 89 – 91, 96, 129 subalternity 92 – 94 subaltern middle class 11, 89 – 98 subaltern native middle class (SNMC) 96 – 98 subaltern studies 11, 95 substantive assumptions 8 substantive reflexivity 70 sunspot equilibrium approach 82n1 surprise 45 – 46, 49n7 systemic reflexivity 70

US economy, model of 43 utilitarianism 9, 101, 115 utility companies 32 utility function 53, 104, 106, 110, 112 utility maximization 100, 103 – 105, 111, 112, 116 utility theory 52

Teschl, Mirian 54 – 55 Theory of the Individual in Economics,The (Davis) 3, 5, 128 tin 22 transcendent reflexivity 73, 74, 81 Treatise on Probability (Keynes) 35 tribal affinity groups 89 Tronto, Joan 10 trust 132 – 133, 135 Tullock, George 106 Turing machine 45 – 48 Turing test 44, 48

Walras, Leon 12, 100 wealth disparities 92 – 94 Weber, Max 91 Weintraub, Roy 47 “we” language 9 welfare economics 111 welfarism 9 well-being 9 what-if questions 6, 42, 43, 48 Whitehead, A. N. 80 – 81 why questions 6, 42, 43, 48 Wilson, Bart 113

uncertainty 37, 132 – 133 University of Amsterdam 4

Zambelli, Stefano 46 Zora 137, 141, 147

validity 43, 44 values 8, 10, 130 – 131 Veblen, Thorstein 89, 91, 100, 126, 130 Veblenian cumulative causation 77 Velásquez da Silva, Diego 71 – 73, 81 Von Neumann-Morgenstern independence axiom 104 Vromen, Jack 106