On Kolm's Theory of Macrojustice: A Pluridisciplinary Forum of Exchange (A Pluridisciplinary Appraisal)

On Kolm’s Theory of Macrojustice

•

Claude Gamel

•

Michel Lubrano

Editors

On Kolm’s Theory of Macrojustice A Pluridisciplinary Forum of Exchange

123

Editors Professor Claude Gamel GREQAM-IDEP Université Paul Cézanne Centre Forbin 15-19, allée Claude Forbin 13627 Aix-en-Provence Cedex 1 France [email protected]

Professor Michel Lubrano GREQAM-CNRS Centre de la Vieille Charité 2 rue de la Charité 13236 Marseille Cedex 02 France [email protected]

ISBN 978-3-540-78376-3 e-ISBN 978-3-540-78377-0 DOI 10.1007/978-3-540-78377-0 Springer Heidelberg Dordrecht London New York Library of Congress Control Number: 2010937679 c Springer-Verlag Berlin Heidelberg 2011  This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Cover design: WMXDesign, GmbH Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

“Discourse ethics rests on the intuition that the application of the principle of universalization, properly understood, calls for a joint process of “ideal role taking”.[. . . ..] Under the pragmatic presuppositions of an inclusive and non-coercive rational discourse among free and equal participants, everyone is required to take the perspective of everyone else, and thus projects herself into the understandings of self and world of all others; from this interlocking of perspectives there emerges an ideally extended we-perspective from which all can test in common whether they wish to make a controversial norm the basis of their shared practice; and this should include mutual criticism of the appropriateness of the languages in terms of which situations and needs are interpreted. In the course of successively undertaken abstractions, the core of generalizable interests can then emerge step by step.” Jürgen Habermas (1995). Reconciliation through the public use of reason: Remarks on John Rawls’s political liberalism. The Journal of Philosophy 92 (3), pp. 109–131.

•

Foreword

This is a book about both economics and philosophy. It should be read by both economists and philosophers. As Serge-Christophe Kolm himself writes on page 38, “the association of economics and philosophy . . . is not only a laudable (and probably too rare) aspect of scholarship: it is . . . simply unavoidable for making sense and progress”. Indeed, he argues that we should make no distinction between the two fields. This is a book about a book: Kolm’s Macrojustice, published in 2005. In philosophy, it is common for there to be books about books. In economics, it is rare. In part, this is because economists write books less frequently, and books play a smaller role in the economics profession. This is, in my view, much to be regretted. Original ideas in economics today are communicated scientifically through peer-reviewed journal articles. This is in many respects an admirable format, but in my view it has serious limitations. The analysis has to be presented in a circumscribed form, often without a full exploration of the underlying assumptions. Within the compass of an article, it is not usually possible to draw out the inter-connections with other branches of the literature or with other disciplines. In writing a book, an author has both the opportunity and the challenge of presenting ideas on a broader canvas. This challenge is one to which SergeChristophe Kolm and the authors of this volume have risen magnificently. Macrojustice is a tour de force, giving centre place to the concept of “equal labour income equalization” (first developed by Kolm many years ago) and demonstrating its power and reach. The present volume brings together the author of Macrojustice and some of its “first and most attentive readers” in a highly productive interchange, which is essential complementary reading. In part, the absence of this kind of volume in economics reflects the fact that today’s economists are less likely to read books. This brings me to a central issue. I have said that the book should be read by economists, but how can this be ensured? It is not just the format but also the subject matter that is unfashionable. As an undergraduate, I studied courses on

vii

viii

Foreword

“welfare economics”, whose dull treatment of issues such as compensation tests would have been greatly enlivened by the present volume. But these courses are no longer taught, and the idea of economics as a “moral science” has been lost from sight. All that I can say is that the present volume offers a wonderful antidote. Every student should be required to read the Introduction. They will then be convinced as to why we should debate the theory of macrojustice. Tony Atkinson Nuffield College Oxford and London School of Economics

Contents

1

Why Should We Debate the Theory of Macrojustice? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Claude Gamel and Michel Lubrano

1

Part I The Macrojustice of Serge-Christophe Kolm 2

General Presentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Serge-Christophe Kolm

3

Economic Macrojustice: Fair Optimum Income Distribution, Taxation and Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Serge-Christophe Kolm

Part II Philosophical Aspects of Macrojustice 4

ELIE and the Emotions Related to Social Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133 Pierre Livet

5

Basic Income and ELIE Transfers: Argument for Compatibility Despite Divergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145 Claude Gamel

Part III Economic Analysis of Macrojustice 6

An Axiomatic Study of the ELIE Allocation Rule . . . . . . . . . . . . . . . .189 François Maniquet

7

An Exploration of Incentive-Compatible ELIE . . . . . . . . . . . . . . . . . . .207 Laurent Simula and Alain Trannoy

ix

x

Contents

Part IV Combinations of ELIE with Other Targeted Transfers 8

Is ELIE a Wasteful Minimum Income Scheme? . . . . . . . . . . . . . . . . .235 Erwin Ooghe and Erik Schokkaert

9

ELIE-Minating Poverty? Limits of the Mechanism and Potential Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257 Alain Leroux and Justin Leroux

Part V Econometric Evaluations of ELIE 10

The Redistributive Aspects of ELIE: A Simulation Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275 Michel Lubrano

11

The Trade-off Between Growth and Redistribution: ELIE in an Overlapping Generations Model . . . . . . . . . . . . . . . . . . . . . .305 David de la Croix and Michel Lubrano

Part VI Selective Comments by Serge-Christophe Kolm 12

Macrojustice in Normative Economics and Social Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .341 Serge-Christophe Kolm

Contributors

David de la Croix IRES and CORE, Université catholique de Louvain, Louvain-la-Neuve, Belgium, [email protected] Claude Gamel GREQAM-IDEP, Université Paul Cézanne, Centre Forbin, 15-19 allée Claude Forbin, 13627 Aix-en-Provence Cedex 1, France, [email protected] Serge-Christophe Kolm EHESS, CREM, 20 Rue Henri-Heine, 75016 Paris, France, [email protected] Alain Leroux GREQAM, Université Paul Cézanne, Centre Forbin, 15-19 allée C. Forbin, 13627 Aix-en-Provence Cedex 1, France, [email protected] Justin Leroux HEC Montréal, CIRANO and CIRPÉE, 3000 chemin de la Côte-Ste-Catherine, Montréal, QC H3T 2A7, Canada, [email protected] Pierre Livet Université de Provence, CEPERC, 29 avenue Robert Schuman, 13621 Aix-en-Provence Cedex 1, France, [email protected] Michel Lubrano GREQAM and CNRS, Marseille, Centre de la Vieille Charité, 2 rue de la Charité, 13236 Marseille Cedex 02, France, [email protected] François Maniquet CORE, Université catholique de Louvain, 34 Voie du Roman Pays, 1348 Louvain-la-Neuve, Belgium, [email protected] Erwin Ooghe Department of Economics, UCBrussels and KULeuven, Belgium, [email protected] Erik Schokkaert Department of Economics, KULeuven and CORE, UCLouvain, Louvain-la-Neuve, Belgium, [email protected] xi

xii

Contributors

Laurent Simula Uppsala University and Uppsala Center for Fiscal Studies, Department of Economics, 75120 Uppsala, Sweden, [email protected] Alain Trannoy EHESS and IDEP-GREQAM, Centre de la Vieille Charité, 2 rue de la Charité, 13236 Marseille Cedex 02, France, [email protected]

•

Chapter 1

Why Should We Debate the Theory of Macrojustice? Claude Gamel and Michel Lubrano

Abstract In this introductory chapter, we give a subjective account of the content of Kolm’s book Macrojustice (2005) that gave rise to the idea of organising in 2006 a round table where this book was discussed by different authors coming from a large variety of horizons: philosophers, economists, econometricians. We leave Serge-Christophe Kolm the task of presenting his theory in the first part of this book. Macrojustice is concerned about social justice proposing a comprehensive redistributive scheme. Of course, any distributive proposal always raises questions at the ethical, theoretical and practical levels. These questions are at the core of the discussions that are presented in this book, which is designed as a forum for multidisciplinary exchange.

1.1 A Multidisciplinary Forum of Exchange In April 2006, l’Institut d’Economie Publique of Marseille (IDEP) organised a workshop devoted to the book Macrojustice published the previous year, with the participation of its author. In the presence of Serge-Christophe Kolm, this one day workshop gathered economists, philosophers and C. Gamel (B) GREQAM-IDEP, Université Paul Cézanne e-mail: [email protected] M. Lubrano GREQAM and CNRS, Marseille e-mail: [email protected] The writing of this introductory paper benefited from the numerous and precise comments of SergeChristophe Kolm and of François Maniquet who helped a great deal in improving on a previous version. The cooperation of Steve Bazen was precious for correcting the English. Remaining errors are solely ours.

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_1, 

1

2

C. Gamel and M. Lubrano

econometricians from the universities of Aix-Marseille and Louvain. The quality and the intensity of the discussions encouraged us to widen the circle of participants by publishing a book devoted to discussion of Kolm’s theory of macrojustice. Beyond the original participants, five other authors or co-authors agreed to collaborate on the project. We are grateful to all of them, starting obviously with Serge-Christophe Kolm. As with the workshop, this book has the aim of being a multidisciplinary exchange between Serge-Christophe Kolm and some of the first and most attentive readers of Macrojustice: – In addition to the introduction (Chap. 1), the reader will find a contribution by Serge-Christophe Kolm reiterating the latest state of his thought on macrojustice (Part 1). It also provides the author with an opportunity to highlight certain aspects of his theory in the light of the remarks that were made here and elsewhere since the publication of his book.1 – The main corpus of the book is made up of a series of eight contributions and is organised in four parts. The first two parts are theoretical and include philosophical (Part 2) and economic (Part 3) contributions. The two following sections are more applied in nature. The first contributions are devoted to studying a combination of ELIE (Equal-Labour Income Equalisation) transfers with other more targeted redistributive tools (Part 4). Others contain econometric simulations of the implementation of ELIE transfers in a real world context (Part 5). – The final section of the book is written by Serge-Christophe Kolm who comments on the contributions. From our point of view, this is not a conclusion, not even a provisional one, about macrojustice. It is natural to allow the author of the original theory to comment freely on the research inspired by his theory. It is also of interest to the reader to find in the same volume, both contributions made by peers and Kolm’s reaction to them.

1.1.1 The Main Corpus of This Book Concerning the various contributions that constitute the central part of this book, the reader will also find an abstract at the beginning of each chapter. We emphasise here in a few words the pluridisciplinary nature of contributions to which Kolm’s theory has given rise:

1

See in particular the reviews of the book Macrojustice which were published in different journals: Roemer (2005) and the reply by Kolm (2006b), Gamel (2005), Fleurbaey (2006), Sturn (2008), Schokkaert (2009) and the response by Kolm (2009).

1

Why Should We Debate the Theory of Macrojustice?

3

– In the section concerned with Philosophical aspects of macrojustice, Pierre Livet deepens Kolm’s analysis of “endogenous social choice”, starting not from the satisfaction of individual interests but from the emotional bonds that can be generated by social recognition: in order for a consensus about ELIE transfers to emerge, all individuals must feel that they have the regard of others, not only the most productive and net contributors, but also the less productive and net recipients of transfers (Chap. 4). Following this, Claude Gamel compares Kolm’s ELIE transfers with Van Parijs’ proposal concerning basic income, both projects stemming from a discussion of the notion of freedom. Their theoretical differences could be fruitful: TECIE transfers on capital, modeled on Kolm’s ELIE transfers, could provide both a means of funding basic income and an extension to the theory of macrojustice (Chap. 5). – In the section Economic analysis of macrojustice, François Maniquet draws on his work with Marc Fleurbaey on the axiomatic definition of allocative rules which could provide first-best solutions to the issue of income taxation. The axiomatic investigation of ELIE transfers confirms that the ELIE transfers are members of the class of first-best solutions and that the non-axiomatic arguments of Kolm are consistent with Maniquet’s own presentation (Chap. 6). Next Laurent Simula and Alain Trannoy investigate second-best fiscal solutions that could satisfy ELIE transfer properties, when information on individual working time is not available for establishing such transfers. These second-best solutions are sought in two ways: a welfarist version with a system of social weights and a non-welfarist one with a system of social transfers and taxes (Chap. 7). – In the section Combinations of ELIE with other targeted transfers, Erwin Ooghe and Eric Schokkaert seek to extend the theory of macrojustice to individuals who voluntarily work less than the ELIE equalisation parameter k, or do not work at all. This leads one to consider ELIE transfers as a system of minimum income. Such an extension is a source of waste because of the impossibility of distinguishing between voluntary and involuntarily unemployment. However this waste seems to be small for realistic levels of minimum income (Chap. 8). Next Alain and Justin Leroux seek to assess the efficiency of ELIE transfers for fighting poverty. As ELIE transfers were not designed for this purpose, their relatively limited efficiency in this domain can be improved by associating them with the “personal allowance mechanism” presented by the two authors. Such a combination could help to eradicate poverty while using only a small share of the resources collected by ELIE transfers (Chap. 9). – In the section Econometric evaluations of ELIE, Michel Lubrano uses a calibrated simulation model to measure the comparative static redistributive effects of ELIE transfers, when compared to the current French

4

C. Gamel and M. Lubrano

system of redistribution. Their impact depends on the initial income distribution. He considers the effect of imperfect information about wage rates as a measure of productive capacities, the role of family allowances and that of capital accumulation in a simple Solow growth model (Chap. 10). In the following chapter, David de la Croix and Michel Lubrano investigate the long-term impact of ELIE transfers on growth and redistribution, through their impact on individual investments in education, human capital accumulation and growth. They use an overlapping generations model of endogenous growth to underline the trade-off between growth and redistribution. The terms of the trade-off can be inverted, depending on the basis chosen for taxation, in particular whether ELIE transfers are intended to subsidise education or not (Chap. 11).

1.1.2 Reading Macrojustice (2005): A Valuable Challenge We now present our own view of what Macrojustice achieves and the key position it occupies in the work of Serge-Christophe Kolm. Following this presentation, we raise a number of questions, some of them developed by the contributors to this book, suggesting that researchers should debate macrojustice and extend this theory, which is precisely what all the book is about. One of the characteristics of really innovative books is probably that they are initially misunderstood, or at least difficult to understand and to interpret. In this respect, A Theory of Justice by Rawls (1971) represents a famous precedent and an illuminating example: the degree of misunderstanding was so great that, at the beginning of the seventies, the book was sometimes considered as a new apology for socialism, before being gradually recognised for what it really was: the seminal book on “liberal egalitarianism” for some academics or on “post-welfarism” for the others. To be fair with the first readers, Rawls’ A Theory of Justice was certainly not easy to read. Three decades later, a similar reaction appears to have happened in 2005 with Kolm’s book Macrojustice. At the confluence of two disciplines (philosophy and economics), the question is again to figure out among several hundred dense pages which are the major innovations that justify the intellectual investment required to read Macrojustice. The aim of these introductory comments is to convince the potential reader of the excellent return to this investment. We would like to supply him with a first comprehensive overview of this elegant theoretical construction. We would also like to encourage him to form his own opinion, because the questions and the stakes which arise from Kolm’s innovative thought are

1

Why Should We Debate the Theory of Macrojustice?

5

both numerous and diverse. This is not unsurprising. As Kolm’s bibliography contains already more than three hundred references in the fields of public economics and economic philosophy, it is difficult to imagine which new elements could be added to an already very impressive amount of work. Before providing an answer to this question, it is worth underlining that we are referring to someone who has made fundamental contributions to the scientific literature. His innovations include introducing as early as 1963 the expression “public economics”,2 and as early as 1966, integrating the notion of social justice into the field of economic theory3 ; he also contributed in 1971, in his monograph Justice and Equity, to the emergence of the economic theory of envy-freeness as well as to the introduction of the notion of leximin as a criterion for practical justice. The fact that the major part of his early work was written in French did not prevent him from acquiring an international reputation, a reputation that the publication of Macrojustice in English will strengthen. With Macrojustice, Kolm builds on his previous work, by pursuing two further goals: – He provides the keystone to the architecture of his past work which was perhaps missing. That is to say, he orders his multiple and analytical works in a synthetic manner allowing one to discern his general conception of man and society. – More explicitly, he contributes to the contemporary debate on income distribution, in which philosophers and economists try to extract from the theories of social justice arguments that can provide justifications for specific reforms (negative income tax, basic income, ...), to which Kolm adds his original contribution with ELIE transfers.4 In this introduction, we first want to give an idea of the way in which these two objectives fit together. Three arguments define the comprehensive architecture of Kolm’s work, where justice intervenes only as a third-best palliative. This constitutes an important landmark and contributes to an understanding of the depth of his thought (Sect. 1.2). Two further arguments relate to the dual consensus on the structures of ELIE redistributive 2

See the book Les fondements de l’économie publique: Introduction à la théorie du rôle économique de l’État (Kolm 1963). 3 See the communication made during that year at the famous meeting on public economics held in Biarritz and entitled The optimal production of social justice. The proceedings of this conference were published in book form in French and in English (Guitton and Margolis 1968, 1969). 4 The basic idea of ELIE redistributive schemes is however not new in Kolm’s work, because we can find its origin in 1966 as cited in Kolm (1996), and then again in several texts of the 1990s; on this point, see Kolm (1996, pp. 128–130). The originality of Macrojustice is to use this idea as the central proposition which structures the whole book.

6

C. Gamel and M. Lubrano

schemes which constitute the core of the theory of macrojustice (Sect. 1.3). We then set out what we consider to be the most innovative components in this highly original theory (Sect. 1.4) and then raise some major questions which illustrate the fruitfulness of the debate (Sect. 1.5).

1.2 A Landmark in Kolm’s Thought: Justice as a Third-Best Palliative Contrary to a tendency well established since Aristotle, Kolm (2005, p. 10) does not consider justice as the cardinal virtue of a society: “If people were sufficiently able to control the birth of their desires, the desires they would choose for avoiding dissatisfaction caused by the scarcity of goods, would ipso facto elicit no conflict about scarce goods”. He then refers to his book published in 1982 Le Bonheur-Liberté, which earlier explores the ways in which modern Buddhism could help to overcome the inconveniences of economic constraints by mastering the birth of one’s desires. However, such a first-best solution is difficult to design in large-scale modern societies, for obvious reasons of information and of the personal conscious involvement in such an ethical position. Therefore, a second-best objective should be preferred, if a form of absolute altruism (“general reciprocity”) were able to supplant the two principles of behaviour that are most commonly observed, i.e. self-interested market exchange and enforced compulsory redistribution. They have indeed succeeded in transforming a situation of complementarity and cooperation, in which men need each other, in a series of hostile relations that divide men on the basis of either “blackmail” (exchange) or coercion (redistribution). At this point, Kolm stresses the comparative attractiveness of reciprocal behaviour as a founding stone for La Bonne Economie – the title of another of his important works published in 1984. Again, the main difficulty lies in the generalisation of the notion of “general reciprocity” to a large-scale society: despite the undeniable qualities of such a system, it is difficult to promote this common ethical framework because the use of market exchange or of coercion to achieve this goal is naturally excluded (a gift, by nature and definition, can be realised neither by coercion nor by selfish exchange). Accordingly, neither the modification of individual preferences nor the generalisation of reciprocity are able to solve the issue of social justice. If the two previous solutions are not possible to implement, then justice only offers a third-best palliative solution. Justice is both the only conceivable objective in modern societies which are still dominated by the pursuit of

1

Why Should We Debate the Theory of Macrojustice?

7

personal interest, and the necessary condition for human progress to develop personal values and social relations of the highest quality. Of course, Kolm has devoted several of his works to this critical issue,5 but in Macrojustice, we find the heart of the problem. Macrojustice could be defined as the product of two elements: the basic general rules of society (classical basic rights including the right of free exchange) and the consequences that distributive justice derives from them (the allocation of benefits from society’s main resources). In other words, Kolm deals with the delicate combination of efficiency in the production of wealth and of justice in the distribution of income, leaving aside the numerous other contexts in which the reference to justice is also essential: first, issues of “microjustice” in the distribution of one particular scarce resource among few people (“local justice” in the sense of Elster (1992) applied, for example, to the selection of recipients for an organ transplant) and secondly, issues of “mesojustice” in the distribution of assets that remain specific (such as education or health), but with which everyone is concerned. For issues of fiscal and redistributive justice, which are often considered as two separate questions and are often the subject of controversy, Kolm proposes the generic integrated solution of the “Equal-Labour Income Equalisation” (ELIE). This solution has to be based on a dual consensus.

1.3 A Dual Consensus at the Core of Macrojustice: ELIE Transfers While the redistributive ELIE schemes result from a unanimous choice made by individuals, this choice is neither suggested nor imposed by any moral intuition or by any particular conception of justice. On the contrary, the only ethical dimension of the approach is the desire to help these individuals to become aware of what they really want when organising the society in which they live. The aim is to make explicit what Kolm calls “endogenous social choice” in this community6 , which, he argues, would lead to a dual consensus concerning the overall income distribution: first a general consensus on the scheme of income redistribution to be chosen – “Equal-Labour Income Equalisation” – (Sect. 1.3.1), and then a particular consensus on the extent of the equalisation to be achieved (Sect. 1.3.2). Once this dual consensus is obtained, the individual retains the freedom to choose the intensity 5

See in particular the synthesis presented in Modern Theories of Justice (Kolm 1996). At this stage, we focus solely on the content of this “endogenous social choice”; the conditions for its emergence will be discussed later (see below Sect. 1.5.2).

6

8

C. Gamel and M. Lubrano

with which to exploit his personal productive capacities by working. Without doubt, this freedom constitutes a third level for the theory of macrojustice that has to be carefully examined (Sect. 1.3.3).

1.3.1 The General Consensus on the Redistribution Scheme Regarding first the general scheme of ELIE transfers, consensus arises from the desire to reconcile “social freedom” defined as “individuals’ freedom from others’ forceful interference” and income redistribution which a priori introduces social interference. Conciliation is within the scope of a “process liberalism”, “the central social-ethical theory of the modern world” (see Kolm 2005, p. 20) which is distinct from classical liberalism in that it does not exclude redistribution, but requires a specific form for it.7 ELIE schemes contain a dual individual and collective dimension: – As regards the individual dimension, income redistribution should be based on inelastic variables that do not alter the behaviour of those involved (no price effect) and should leave invariant the redistributive capability of the economic system. These inelastic variables are individuals’ “natural resources”, i.e. their given personal capacities to earn income by working. It is therefore necessary to tax the market valuation of everyone’s productive capacities, regardless of the extent to which the individual decides to exploit them (working full-time, part-time or not at all). The estimated market value is based on the income he would earn from his productive capacities by working full-time. – The second dimension is more collective: individuals do not live in isolation but within a society of free men, for which they may express their degree of support and solidarity. In this context, they could accept that a part of the income they could earn from their productive capacities would be given to the community and shared equally among its members. For example, by giving to society the market value of two days of their working week, the individuals pay a lump-sum tax, which varies from one individual to another only in terms of the productive capacities of the individual concerned. In return, they receive from society the average value of all incomes derived from these two working days. The range of individual incomes is thus reduced by the equal sharing of the monetary value of one part of everyone’s productive capacities (two days in our example). 7

Of course, Kolm’s “process liberalism” also implies the adequate correction of the classical market failures (production of public goods, management of externalities, ...); on this point, see Kolm (1985).

1

Why Should We Debate the Theory of Macrojustice?

9

1.3.2 The Particular Consensus on the Extent of Equalisation to be Achieved This is the second level of consensus: not only would individuals accept such redistributive schemes, but they would be likely to agree, according to Kolm, on the choice of a specific redistributive scheme, i.e. on the percentage of the market value of their productive capacities that they would make available to society. This percentage is called the “parameter of equalisation”, k.8 Yet one might wonder how individuals (who know their personal productive capacities that will serve as a basis to evaluate their contribution) would be able to agree unanimously on the specific value of such a parameter. In this perspective, the two extreme cases may be found to be inadequate (Sect. 1.3.2.1), while intermediate ELIE schemes are more appealing (Sect. 1.3.2.2). 1.3.2.1 The Inadequacy of Extreme ELIE Schemes If the equalisation parameter k is zero, the contribution of each person is zero and “process liberalism” becomes “classical liberalism”, where no transfers are financed: the fraction k of the average income w N of the population that is redistributed is zero (k wN D 0). An immediate consequence easily appears in the classical framework of the individual leisure-income trade-off: for the same non-zero working time (` ¤ 0), the least productive individuals reach an income which is less than the income of the most productive individuals (see Appendix A and Fig. 1.1). In other words, the domain of choice of the most productive individuals includes that of the least productive individuals. The latter cannot fully adhere to such an unequal situation where, in the absence of any redistributive scheme (k D 0), they have less freedom. It is only when both types of individual do not work at all (maximum leisure time) that they can find themselves in a similar situation: neither of them has a positive income (y D k wN D 0). In contrast, if k is equal to one, each individual yields his whole income to society: in return, 100% of the population average income is redistributed (k w N D w). N Therefore, the most productive individuals pay a maximum negN  wmp / < 0), which provides a funding for ative net transfer (tmp D .w a maximal positive net transfer for the least productive ones (tlp D .w N  wlp / > 0). The intensity of the redistribution is so great that, with the same In algebraic terms, if wi is the earned income of individual i and w N the population average income, parameter k defines the level of transfer ti for individual i : ti D k.w  wi /. A simplified analysis of this transfer is presented in the two subsequent sections. For more details, see Appendixes A and B to this chapter. 8

10

C. Gamel and M. Lubrano

working time, the most productive have an income which is systematically lower than that of the least productive and therefore the domain of choice of the most productive is this time smaller (see Appendix A and Fig. 1.2). The most productive would not fully adhere to such an extreme redistributive scheme because they manage to reach the income of the least productive only when working at their full capacity (no leisure time). According to the now famous expression of Dworkin (1981, p. 312), this situation is called the “slavery of the talented”, and is also cited as such by Kolm (2005, p. 28). As a consequence, the case k D 1 is only a conceptual experiment and is not part of the “Kolmian” theory of macrojustice. 1.3.2.2 The Interest of Intermediate ELIE Schemes We now examine intermediate ELIE schemes, where k lies strictly between zero and one, which fits more closely with the consensual ethics of “process liberalism”. These schemes correspond to a more balanced situation, where the domain of choice of one category of individuals is no longer systematically included in the domain of choice of the other category. These situations are illustrated by Fig. 9.1 entitled “The geometry of ELIE” in Kolm (2005, p. 157). This Figure pretty well summarises the core of the theory and for the sake of clarity its highly synthetic construction is reconstituted in two successive steps (see Appendix B, Figs. 1.3 and 1.4). For the moment, these graphs illustrate the general framework of the theory, but do not allow one to make the difference between the case `  k and its opposite ` < k. Only the first of these is covered by the theory of macrojustice. The second represents another conceptual experiment which remains outside the scope of macrojustice, as with the earlier case of the “slavery of the talented”. As a starting point, let us consider the pivotal category of the population. For this category, the monetary value of its productive capacities wpi is equal to the average wage of the entire population. In this case, if for example k is equal to 0.4 – the equivalent of two working days in a five day week9 –, individuals in this pivotal category are required to pay 40% of the market value of their productive capacities, or k w, N and in return receive 40% of the average value of the productive capacity of the entire population, again k w. N The net transfer they receive t is neither negative nor positive: it is zero (tpi D k.wN  w/ N D 0), irrespective of their hours of work, i.e. whatever the intensity with which they decide to exploit their productive capacities. In these circumstances, the introduction of ELIE transfers does not change their budget constraint (see Appendix B, Fig. 1.3) and they receive an income 9

In this simple example, the five day week is assumed to be the maximum working time (` D 1).

1

Why Should We Debate the Theory of Macrojustice?

11

that remains insensitive to the introduction of these transfers: the income of individuals in the pivotal category depends solely on their working time `. It will be higher or lower than k w N (possibly zero), depending on whether ` is higher or lower than k (possibly zero). Compared to this pivotal category, the situations of the more productive and the least productive are characterised, for the same ELIE scheme (k D 0:4), by a non-zero net transfer, negative for the former and positive for the latter. Since their full-time wage is greater than the average wage, N Hence we have the most productive pay kwmp and will only receive k w. tmp D k.w N  wmp / < 0. Conversely the contribution of the least productive (k wlp ) will be lower than the allowance they receive (k w). N Hence we N  wlp / > 0. In contrast to those of the pivotal category, have tlp D k.w the most productive individuals together with the least productive individuals experience either the limitation or the extension of their domain of choice as the result of net non-zero transfers. Unlike the situation observed in the case of extreme ELIE schemes, none of the three categories is in a better situation than the others by having a larger domain of choice (see Appendix B, Fig. 1.4), since their respective budget lines intersect at point, K the coordinates of which are .1  k; k w/. N Moreover, if they all choose to work at the level ` D k, they all have the same leisure time .1k/, and also the same net income after transfers .k w/. N In this case, the “Equal-Labour Income Equalisation” leads to the quite remarkable situation of complete equality between all individuals.10 This situation of complete equality is a key point in the axiomatic analysis of the macrojustice theory proposed by Maniquet later in this volume. More precisely, it allows us to formulate one of the axioms specific to ELIE transfers, in the broader definition of the rules leading to first-best solutions to the problem of income tax. However, ELIE transfers do not necessarily lead to a situation of complete equality, since individuals are not forced to work at the level ` D k and remain a priori free to exploit as they wish their personal productive capacities. This point is very important for understanding the full scope of macrojustice theory as we will now show.

1.3.3 The Freedom to Use One’s Personal Productive Capacities When comparing, for an identical working time `, the situation implied by ELIE transfers for three individuals belonging to the three categories under inspection, we can make four important observations: 10

This situation of complete equality also exists in the case of extreme ELIE schemes, but is rejected, as already noted, at the corner of the individual domains of choice: complete inactivity .` D 0/ if k D 0, maximal working time .` D 1/ if k D 1.

12

C. Gamel and M. Lubrano

– If everyone has the same working time ` D k, the remaining time .1  k/ being dedicated to “leisure”, each person receives the same net income after transfers .y D k w/. N – If the working time ` of the three individuals is greater than k, they will be paid an income after transfers greater than k w; N the most productive of the three retains, however, an income higher than that of the other two and would thus have greater freedom, because his domain of choice would be larger in the zone ` > k.11 – The situation is reversed if the three people decide to work less than the retained value of the equalisation parameter. After transfers, the domain of choice of the most productive would be smaller than that of the other two and his liberty would therefore be more restricted in this zone ` < k. This case is not admissible in Kolm’s theory. – Finally, in the extreme case where the three individuals remain completely inactive, their individual situations are even more contrasted. The individual belonging to the pivotal category receives no income at all; the less productive individual receives a guaranteed income k.ww N lp /, depending on his productive capacity, low on account of his endowment; while the more productive individual must pay a lump-sum tax k.wN  wmp /, which is this time a function of his high capacity.12 This case is not admissible in Kolm’s theory if k > 0. In short, the theory of macrojustice is a three-storey construction, that any future implementation of the ELIE mechanism should take into account: a broad discussion concerning the general outline of redistribution (EqualLabour Income Equalisation), the selection of a particular ELIE transfer scheme (parameter k of income equalisation), the choice of the intensity at which productive capacities are to be used (working time `). While the first two levels result from a collective decision process which, according to Kolm, should be as consensual as possible, the third level is the outcome of an individual decision process, where personal freedom must remain as complete as possible. After this general presentation which has been made as simple as possible mainly for the benefit of non-economists,13 we now present the most original aspects of Macrojustice.

In Appendix B, Fig. 1.4, this situation is illustrated for ` D 0:9 and k D 0:4. In Appendix B, Fig. 1.4, when the individuals do not work at all (` D 0), positive ELIE transfers (guaranteed income in the case of under-endowment) and negative ELIE transfers (lum-sum tax in the case of over-endowment) are respectively represented by segments BD and BE. 13 In more comprehensive but obviously more complex presentations of the ELIE schemes, working time is not the only dimension which is considered for labour. Other more qualitative 11 12

1

Why Should We Debate the Theory of Macrojustice?

13

1.4 The Most Original Features of Macrojustice The originality of the theory of macrojustice can be understood both in terms of economic philosophy – concerning its foundations – as well as in terms of public economics – for its applicability. From the former point of view, the innovation concerns given personal capacities considered as a possible source of rent to be shared with others (Sect. 1.4.1). From the latter point of view, macrojustice theory introduces interesting incentive properties for taxation (Sect. 1.4.2). Therefore, like Schokkaert (2009), we regard this theory as promoting a “paradigm shift” that challenges the influence of the welfarist theory of optimal taxation in academic economics (Sect. 1.4.3).

1.4.1 Personal Capacities as a Source of Rent Income that can be Shared Kolm’s reflection goes beyond ordinary commutative justice (for market exchanges) and returns to the fundamental question of distributive justice (which applies to the allocation of initial endowments), a matter he treats of with a precision which is rare in the literature. The theory of macrojustice considers given personal capacities as a possible source of a rent, i.e. an income without an obvious cause, or at least as a source of an income the legitimacy of which is not self-evident. More precisely, this lack of legitimacy provides a justification for the idea that individuals may have to accept that a part k of their income wi is given to the community and equally shared among its members. Once the level of k has been chosen and the part k wi has been redistributed, individuals can peacefully enjoy the other part .1  k/wi which is now considered as self-owned. In this sense, Kolm seems to be in the tradition of Rawls:14 personal productive capacities can be regarded as the product of natural talents that the individual has inherited from his parents and which Rawls (1971, p. 104) refers to as the arbitrariness of the “natural lottery”, to emphasise that the individual has no moral merit in the selection of the genes that determine his

characteristics are taken into account such as intensity of work, initial training, existence of involuntary unemployment. As well, the equalisation parameter k no longer varies between 0 and 1, but varies in a smaller space Œ0; k e , where k e corresponds, according to Kolm (2005, p. 191), to a better specification of the current position of the supporters of income equality. For a presentation of all these refinements, see Kolm (2005, Chap. 8–13). 14 And maybe also of Mirrlees. See the developments concerning Rawls and Mirrlees in Kolm (this volume, pp. 52–55).

14

C. Gamel and M. Lubrano

talents or disabilities. We might also consider that the productive capacities of the individual are essentially the product of the socio-cultural environment in which he grew up and, in this case, Rawls puts into question the arbitrariness of birth and social environment to support the idea that the individual cannot be held responsible for the advantage or disadvantage generated by this environment. Therefore the distribution of rents arising from personal productive capacities, that everyone holds without morally deserving it, is a legitimate question that Kolm’s sharp presentation greatly clarifies: – Firstly, his presentation puts into perspective many theories of social justice such as the two criteria of Varian (1974), namely ex ante “wealth fairness” and ex post “full income fairness”. In the first case, only the non-human resources (natural or non-human assets inherited from past generations) are equalised. But there is no redistribution of human personal productive capacities. This corresponds to the first extreme situation of ELIE where k D 0. In the second case, personal individual capacities are considered as collective wealth, the income from which should be distributed equally, as in the second extreme situation of ELIE where k D 1. A very interesting contribution made by Kolm is to locate these extreme options in the full range of treatments that one can apply to the rent income arising from personal productive capacities .0  k  1/.15 – In doing so, the theory of macrojustice does not accept the fate of a dual genetic and social determinism that would deny free will to any individual. Neither does it accept the intervention of a fussy central planner who in the name of justice wishes to govern the lives of individuals in every detail. At the third level of ELIE schemes, an individual remains free to define a priori his own way of life by choosing the intensity with which he manages his personal productive capacities. Kolm’s framework is thus both egalitarian and liberal, proposing a “third way” between classical liberalism and naive egalitarianism. While such a prospect has been initiated by Rawls, the public economics point of view allows us to highlight another originality of Kolm’s work.

1.4.2 A Form of Taxation with Interesting Incentive Properties By considering given individual productive capacities as the sole basis for taxation, the theory of macrojustice leaves aside traditional fiscal policies, However, Kolm admits only 0  k  k e as already underlined in footnote 13. He does not include the individual non-human wealth in the field of macrojustice, whereas Varian takes it into account to define his two criteria. On this point, see also 1.5.3.

15

1

Why Should We Debate the Theory of Macrojustice?

15

which tax individual behaviour (work, saving or consumption) through the effects of this behaviour (labour income, savings, consumption expenditure). It relies instead on a variable which is a priori insensitive to individual behaviour (i.e. the level of their “natural resources” or “given capacities”), and thus avoids the introduction of incentive or disincentive mechanisms: – On the one hand, the “rich” are not forced to pay more (and work more) to supplement the income of the “poor” who would choose to remain idle: beyond a working time equal to k and a net lump-sum tax k.ww N mp / that they have to pay, their overtime work is now exempt from any taxation. – On the other hand, the “poor” are not encouraged to become inactive: whether they work or not, they are also required to provide their own contribution, based solely on the market value of their productive capacities wlp : in other words, their contribution reduces the net lump-sum grant k.w N  wlp /. But ELIE transfers do not create an “inactivity trap”, unlike the “degressive” mechanism of guaranteed income (such as the RMI in France), which reduces the amount paid to working poor when their labour income increases. In principle, the lump-sum taxation (up to k) of productive capacity income is therefore both simple and attractive for its useful incentive properties. If we put side by side the nice properties of ELIE transfers and the liberal egalitarianism in which the whole theory of macrojustice is inscribed, we can argue that Kolm lays the foundations for a non-welfarist theory, which initiates nothing less than a complete change of paradigm for both economic philosophy and public economics.

1.4.3 A Non-Welfarist Theory of Taxation: Toward a Paradigm Shift The change initiated by Kolm in the theory of taxation can easily be understood in term of the striking impact that the seminal theory of “optimal taxation” of Mirrlees (1971) had in the field of public economics. This theory is based on two fundamental arguments: – First, the maximisation of a “social welfare function” having individual utilities as arguments, most of the time chosen to be identical purely for convenience. This implicit reference to utilitarianism was natural in the early seventies, since normative economics was largely equivalent to welfare economics at that time. Even though the “new” ordinalist welfare economics has supplanted in some cases the “old” cardinalist one, reference to a social welfare function remains the rule. Most often research

16

C. Gamel and M. Lubrano

based on those functions is nothing more than an attempt to implement the utilitarian principle of “the greatest happiness of the greatest number”. – Second, the search for an adequate basis for taxation which could be regarded as an alternative to taxing individual productive capacities. The latter are by definition unobserved for the defenders of optimal taxation theory. These defenders agree to take the current income obtained from these productive capacities as a proxy for a taxation basis. Optimal taxation theory then seeks to maximise the efficiency of taxation on income by trying to identify the disincentive effects of taxation on labour supply. Kolm’s macrojustice theory questions these two fundamental arguments. – Regarding first the redistributive objective of taxation, the intention is indeed to abandon the welfarist perspective, and in doing this Kolm is again close to Rawls in his criticism of utility as a foundation for a theory of justice and redistribution. As a matter of fact, we notice that “eudemonistic capacities” – the personal ability to get satisfaction from a given situation – are rejected by Kolm as being outside the scope of macrojustice (Kolm 2005, pp. 26–27). And this rejection echoes the aim of the latter to “work out a theory of justice that represents an alternative to utilitarian thought generally and so to all of these different versions of it” (Rawls 1971, p. 22). – However the philosopher Rawls does not go as far as the economist Kolm in the definition of liberal egalitarianism as an alternative to utilitarianism. The former certainly defines and orders successively a principle of “equal liberties” and a principle of “fair equality of opportunity”; but at a third and lower level of the hierarchy, equality is no longer the rule. Rawls defines a “difference principle” that tolerates a certain degree of economic and social inequalities, regarded as essential incentives for the production of wealth, because the “legitimate expectations” of the most talented and most dynamic individuals must be rewarded. At this stage of our analysis, it is precisely on the issue of incentives that Kolm innovates at most, departing both from Rawls, whose theory is extended and completed, and from the theory of optimal taxation, which is questioned: – Rawls (1971, p. 278) certainly rejects the utilitarian approach but proposes as “part of the best tax scheme” a proportional expenditure tax, which can obviously generate disincentive effects on consumer behaviour. Moreover, the satisfaction of the legitimate expectations of the most productive can lead to the tolerance of strong economic and social inequalities, under the sole constraint that such inequalities could improve the situation of the poorest individuals. Thus the combination of production incentives

1

Why Should We Debate the Theory of Macrojustice?

17

and of consumption disincentives leads Rawls to depart significantly from his initial reference to equal liberties. Now, one of the major interests of the theory of macrojustice is to maintain such a reference throughout, since the equalisation parameter k is applied by definition to productive capacities that are insensitive to the strategic behaviour of individuals. Therefore Rawls’s maximin in primary goods (the difference principle) is perhaps no longer necessary if the choice of personal productive capacities as the fiscal basis for a distributive policy can lead to fewer incentive or disincentive effects than commonly admitted with other classical fiscal bases. – However, the paradigm shift implied by the theory of macrojustice perturbs the influence of the dominant theory of optimal taxation, against which Kolm develops two major criticisms. According to the first criticism, it is impossible to optimise income taxation starting from individuals utility functions that are necessarily different from one individual to another. This would imply a very fine degree of knowledge on the preferences and psychological profiles of everyone, knowledge that remains out of reach. Hence, the frequent use of identical utility functions that would be known by public authorities can only be an unjustifiable oversimplification. Thus, for Kolm (2005, pp. 197–198), the use of welfarism can only be relevant for issues of microjustice or mesojustice where information on personal utilities would be clearly accessible. The second criticism concerns less the optimisation of the tax system than the optimisation of the basis for taxation. If current income is taken as the correct proxy for productive capacities, no clear idea is proposed to legitimise the use of productive capacities as a first-best basis for taxation. But the theory of macrojustice shows that the important question is not knowledge of the stock of these productive capacities but to be able to tax the income that their full-time market exploitation could generate. As a consequence, the use of current income as a proxy variable is no longer necessary because wage rates are most of the time observable in the different segments of the labour market.16 Therefore, according to Kolm (2005, p. 167), information on the remuneration of the productive capacities is “on the whole more easily obtained than that needed for other taxes or aids”. In the present volume, Simula and Trannoy discuss this point of view, by underlining that the wage rate is known mostly indirectly by combining the observable gross income and the less easily observable working time; the authors attempt to find the interesting incentive properties of ELIE transfers in a classical welfarist framework.

16

For further developments on this point, see for instance Kolm (2005, pp. 175–179).

18

C. Gamel and M. Lubrano

As a whole, the applicability of ELIE redistribution schemes and the extent of the paradigm shift that it involves depend largely on the true availability of relevant information. Here we touch on one of the more sensitive points of theory of macrojustice, which explains and illustrates the richness of the debate.

1.5 Debating Macrojustice Without being exhaustive, we have identified at least five major issues raised by the theory of macrojustice and by ELIE transfers: the first two are mainly related to economic philosophy, the other three to public economics: – If the theory of macrojustice is a non-welfarist and possibly egalitarian theory of redistribution, it has nevertheless to be consistent with its claimed liberal inspiration. We can thus ask to what extent the integration of such an approach to redistribution respects the foundations of liberalism or at least involves an ad hoc reformulation of liberalism (Sect. 1.5.1). In addition, ELIE transfers require a dual and unanimous consensus of individuals within society, that Kolm calls “endogenous social choice”. However, what are the necessary conditions for the existence of such an “endogenous social choice” consensus (Sect. 1.5.2)? – As far as public economics is concerned, a first question deals with the scope of the theory of macrojustice, that Kolm explicitly limits to “human natural resources”. However, one can legitimately raise the question of why non-human individual wealth is neglected (Sect. 1.5.3). As we have already pointed out, it is essential to have the necessary information in order to compute the tax basis of ELIE transfers. The nature of the required information is not quite the same: in the case of the most productive individuals who are net payers of financial transfers, we have primarily to measure capacities (Sect. 1.5.4); in the case of the least productive individuals who are net beneficiaries of the system, we have mainly to evaluate their degree of responsibility (Sect. 1.5.5).

1.5.1 An ad hoc Reformulation of Liberalism? The “process liberalism” introduced by Kolm (see Sect. 1.3.1 above) raises the following preliminary question: does this constitute an ad hoc reformulation of liberal philosophy just for the sake of convenience (to legitimise redistribution in the framework of ELIE schemes)? In Kolm’s theory, we do

1

Why Should We Debate the Theory of Macrojustice?

19

not find, as in Marx, a critical reflection on non-human capital ownership (an aspect that we shall further discuss in Sect. 1.5.3), but we do find a more “revolutionary” vision of the ownership that the individual has over his own human capital. Indeed, the individual is no longer the full owner of himself, since Kolm introduced a dismemberment of ownership rights that concerns not the individual material wealth but his personal productive capacities. The legal ternary distinction between usus, abusus and fructus can shed light on this philosophical debate: – By the clause of abusus, the individual remains the bare-owner of himself: everyone is the sole holder of the right of “abuse” of his own human capital (to sell, to give or to destroy). Certainly the law strictly limits the possibilities for alienation of self-ownership (slavery is prohibited, organ donation is still possible but selling one’s organ is most of the time illegal). In fact, it is difficult to go against the individual decision not to undergo medical treatment or even to commit suicide. – The fructus attached to the individual productive capacities (the right to dispose of his human capital income) is now explicitly shared in the ELIE schemes between the individual himself and society up to the equalisation parameter k. It should be noted, however, that all tax systems based on labour income behave implicitly in the same way, but do not theorise as precisely as ELIE this questioning of the fructus clause. This is no doubt a major foundation for a liberal theory of taxation, which would aim at providing legitimacy for the principle of taxation. – The clause of usus (the right to exploit more or less intensely one’s human capital), is a priori respected by the theory of macrojustice, which strengthens its importance at the third level in its construction (see above Sect. 1.3.3). However, we have here not a principle but only a rule, which might thus tolerate certain exceptions: for instance, we have the case of the “very productive eccentric”, who would prefer to engage in poetry for which he has a great passion, rather than working and receiving a high salary. According to Kolm (2006a, p. 70), these particular cases “might be excluded when considering macrojustice”.17 In fact, excluding this special case from the field of macrojustice seems to hide a fundamental question: as shown in Fig. 1.4 of Appendix B, the budget constraint after negative net transfer requires that a productive eccentric

17

Consequently Kolm considers that the theory of macrojustice should be concerned only with people whose working time `i is higher than or equal to k; see for instance Kolm (this volume, p. 49). Moreover, the opposite cases (`i < k) are not so frequent because the upper bound of the domain of definition for the equalisation parameter k is not equal to one, but to k e < 1 in the more complex presentations of ELIE. See also footnote 13 above.

20

C. Gamel and M. Lubrano

should work at least part time to finance this transfer without having to reduce his non-human wealth. But in this case, his net income is zero and so we must assume that he receives capital annuities which are sufficient to cover his expenses. If his passion for poetry leads him not to work at all, the level of his annuities should be even higher to fund both his current expenses and the negative net ELIE transfer he has to pay. If this is not the case, the right to weakly use his high personal productive capacities or even not to exploit them at all is not respected (see the triangle BCE in Fig. 1.4 of Appendix B). Even if the question of “eccentric productive people” might seem to be marginal, it should not be understated. In the present book, Gamel underlines for instance that this question reveals one of the two major divergences between ELIE transfers and the basic income proposal inspired by Van Parijs’ “real libertarianism”. In other words, even if ELIE schemes exclude the extreme case where the equalisation parameter k is equal to one (see Sect. 1.3.2.1 and Fig. 1.2 of Appendix A), the “slavery of the talented” can remain effective, however in a weaker form, whatever the value of k ¤ 0 for “eccentric productive” people condemned to work against their will. As a consequence, the legal clause of usus is not respected if the Kolmian theory were extended to this particular case. In defence of this extended theory, we must add that this clause is also not respected in conventional tax systems, when the marginal tax rate becomes confiscatory, pushing people to work to pay their taxes. In conclusion, Kolm’s process liberalism may impose restrictions on selfownership rights that are far from being negligible and that could have incurred the wrath of more classical liberals, not just Hayek or Nozick, but maybe also Rawls.

1.5.2 A Consensual “Endogenous Social Choice”? From the point of view of economic philosophy, another aspect of the theory of macrojustice raises questions. Even when restricting the choice of a redistributive system to the sole ELIE schemes considered by Kolm .0  k  k e /, we can easily measure the ambition of Macrojustice’s objective: to bring about a dual consensus, not only on the general patterns of ELIE redistribution but also on the intensity of equalisation that will apply within that framework. In other words, the goal is to achieve convergence of individual preferences to a single parameter of equalisation k, despite the obvious divergence of individual interests.18 Kolm devotes the This divergence of interest would become stronger in the case of extreme ELIE schemes (k D 0 or k D 1), since in both situations one category of individuals dominates the others by a domain of choice which is systematically larger. Even though the second case (k D 1) is rejected by

18

1

Why Should We Debate the Theory of Macrojustice?

21

fourth part of his book (Kolm 2005, pp. 279–360) to a discussion of this issue when he examines “the degree of community, equality, reciprocity and solidarity”. In essence, the relevant method is “endogenous social choice”. It relies not on the interests of individuals, but on their social and ethical judgments, which cover a wide range of issues, ranging from their feeling of belonging to a community to their conception of solidarity and justice. We can note here a relative disagreement with the use of the Rawlsian “veil of ignorance” considered by Kolm (2005, p. 343) as being too “thick”. According to him, consensus and impartiality might rather arise from a better understanding of others, implying a form of dialogue in the sense of Habermas. If necessary, it can be supplemented by other more prosaic social choice methods (vote based on the method of the “moral surplus”, exchange agreement with mutually conditional concessions, ...). Also note that the pivotal category of individuals whose salary is equal to the average wage (see Appendix B and Fig. 1.3) has a special role to play: since by definition they do not pay contributions or to receive benefits, their opinions, which are not biased by personal interest, could serve as a reference for the emergence of the desired consensus. In this book, Livet is especially concerned with the case of less productive individuals. Their status of net ELIE transfer recipients could weaken the social consideration by other and more productive individuals. In this way, the emotional or affective dimension of the social bond must also be taken into account in order for the process of “endogenous social choice” to succeed. As a conclusion, finding a social agreement for the choice for a value for k might not be straightforward.

1.5.3 Why Neglect Non-Human Wealth? In addition to their personal capacities, individuals may possess another category of resources – their non-human wealth – composed of natural assets (land, mineral resources, ...) and other types of capital accumulated over time. Quite paradoxically, while Kolm deals with the most complex form of wealth (“human resources”) in an original way, he neglects the

Kolm (see footnote 13 above), the problem of strong divergence still remains for k D 0 (“classical liberalism”); therefore implementing a consensual endogenous social choice seems to be more difficult in that extreme case (k D 0) than in the other ELIE schemes considered by Kolm (k ¤ 0 and 0 < k  k e ).

22

C. Gamel and M. Lubrano

more traditional problem of non-human wealth and gives three reasons for doing so19 : – In a modern economy, 80% of national income, he says, consists of income from personal productive capacities (labour income), while the non-human wealth is at the origin of the remaining 20% (18% for capital and 2% for natural non-human resources). – Moreover, according to a long tradition dating back to Locke, capital can be regarded as labour accumulated in the past: assuming constant and uniform proportion of production factors, labour still accounts for more than 17.5% Œ18  80=.2 C 80/ in the 18% of national income generated by capital. “Hence, present and past labour accounts for more than 97.5% of national income”, he concludes in Kolm (2005, p. 84). – Finally, the distribution of the remaining 2.5% corresponding to nonhuman resources is essentially a microjustice question, because so many specific or local factors can determine their appropriation (unanimous agreement more or less explicitly formulated, right of the first occupant, succession of multiple local owners, ...). By dropping capital as an autonomous factor of production and by treating the distribution of natural assets only in terms of microjustice, the previous arguments tend to emphasise the marginal contribution of “non-human natural resources” in national income, a contribution that the Kolmian theory seems rightly to be able to ignore. However, one can also question whether the issue of macrojustice is addressed in too restrictive a way: does the distinction between the “human” and “non-human” characteristics of the natural resources really matter? In which case capital, because of its “produced” aspect, is not taken into account. Instead of that, the judicious question, from the point of view of justice, would be, for instance, to question the origin of all the available individual assets. The fact that it is “natural” (oil, physical or intellectual gifts) or “produced” (stock portfolio, professional qualifications) is ultimately secondary when compared either to the passive attitude of the individual (bequest) or to his entrepreneurial attitude (personal implication), when accumulating his wealth. In other words, the fact that human capital generates about 80% of incomes should not lead one to exclude non-human wealth (“natural” or “produced”) from the field of macrojustice. Through this simple example, we see that a greater consideration of wealth would be a delicate but legitimate task, that could provide fruitful extensions to the Kolmian theory. In this

19

See Kolm (2005, pp. 84–89); see also Kolm (1985, Chap. 10) and Kolm (1986).

1

Why Should We Debate the Theory of Macrojustice?

23

book, Lubrano shows that the accumulation of physical capital and the income it generates are paramount in determining the characteristics of the actual income distribution. The former depends crucially on initial conditions, which illustrates the importance of the question on the taxation of bequests.

1.5.4 How Should We Measure the Capacity of the Most Productive? Specifically, this question might be rephrased as: could an ELIE scheme effectively remove the disincentive effects that ordinary taxationredistribution schemes have on the labour supply of the most productive individuals (see Sect. 1.4.2)? Is the problem of incentives not simply shifted elsewhere? In ELIE schemes, the personal contribution of individuals to the financing of redistribution no longer depends on labour actually supplied and on income earned; no one is tempted to hide his earnings in order to escape the tax or to reduce it. However, the actual level of his productive capacities (or full-time earned wages wi ) becomes at the same time the strategic variable that the individual might have an interest in monitoring. In this book, de la Croix and Lubrano focus their interest on education decisions. They propose an overlapping generations model with heterogenous agents and endogenous growth based on human capital which exhibits the traditional trade-off between growth and redistribution. The distortionary effect of ELIE is removed when education is subsidised. Incidentally, even if the most productive individuals do not manipulate their education level, a powerful motive exists for them to hide their abilities, especially when they decide either not to work or to engage in a very small level of activity (` < k); but we know that this case is beyond the scope of Kolm’s theory. When one’s working time is greater than or equal to k, the temptation to hide the real level of one’s productive capacities can still exist, but this temptation is strongly reduced by the absence of any marginal taxation on additional earnings accumulated beyond k.20 In fact, the force of this criticism depends on the availability of objective information on the individual productive capacities. When people work full-time or even part-time, it is easy to identify their wage level wi in a flexible labour market. However, if the theory of macrojustice were extended to individuals who remain totally inactive, we should have to seek proxy

20

On this case, see also Kolm (this volume, pp. 117–118) for a slightly different point of view.

24

C. Gamel and M. Lubrano

variables: for instance the number of schooling years reported in a Mincer earning function. In this case, the nature and duration of the initial training could become the kind of strategic information that one may choose not to reveal. Moreover, for self-employed professionals like artists or writers, the level of productive capacities depends on personal factors such as know-how and talent which are highly variable from individual to individual and which can be reduced in case of partial or total inactivity: in this case, we are obliged to rely on a lump-sum valuation of the average income earned by the active members of the same profession. At this stage, the complexity of the questions raised cannot allow one to determine whether the information on the productive capacity necessary to implement ELIE schemes is actually more easily accessible than data required by other systems of redistribution.

1.5.5 How to Detect the Responsibility of the Least Productive? With regard to the useful incentive properties of the ELIE schemes, another question arises, concerning now the treatment of the least productive. We have already observed (see Sect. 1.3.2.2 and Fig. 1.4 of Appendix B) that the theory of macrojustice allocates them a minimum income when they do not work Œtlp D k.w N  wlp / > 0; whereas in the same situation, individuals in the pivotal category receive no benefits and the most productive must pay a N  wmp / < 0. negative transfer Œtmp D k.w Let us note here that the ELIE scheme differs markedly from all other systems of income support21 : it is not a basic income as a positive income is paid only to the least productive when no work is supplied. Neither is it a system of guaranteed minimum income or negative income tax, where the same level of income is guaranteed to all inactive people. In the ELIE schemes, minimum income varies from individual to individual and depends on the difference between the average wage rate and an estimate of his productive capacities wlp . In the present volume Leroux and Leroux explore the compatibility between the ELIE schemes and their proposed personal allocation system. According to these authors, a combination of these two systems is particularly efficient in reducing poverty. So, in the framework of ELIE transfers, the least productive can also have an incentive not to reveal the true level of their productive capacities. Indeed, 21

See on this point, Kolm (2005, pp. 240–242) and the graph entitled “Comparisons of distributive schemes” on page 241.

1

Why Should We Debate the Theory of Macrojustice?

25

when their liability is not engaged in their failure to work (intensive job search, for example), their productive capacities (and therefore their wage wlp ) are considered to be zero, because they can be offered no wage on the labour market. Accordingly, ELIE provides them with a minimum income “out of liability” k w. N The issue is then how to estimate in some cases the degree of liability of these individuals, especially when the concerned persons have an objective interest in pleading that their non-work status is not of their doing. The problem arises in particular when distinguishing situations of voluntary and involuntary unemployment: – In the first case, the individual productive capacities are by definition intact and the minimum income of the voluntary unemployed will be k.w N  wlp /, because their inactivity results from a deliberate choice. – In the second case, Kolm (2005, p. 128) treats the situation of involuntary unemployed on the same footing as that of the disabled whose access to employment is impossible. They receive the maximum transfer “out of liability” k w. N In the absence of a system of unemployment benefits, it is also notoriously difficult to make the difference in practice between individuals that are unable to work and those who lack the motivation to work (see for instance the phenomenon of “discouraged workers” who no longer seek a job). But in ELIE schemes, the phenomenon is potentially amplified by the financial incentive for an individual not to reveal his real productive capacity, so as to receive a maximum positive transfer. This incentive is symmetrical to that concerning the most productive, whose aim might be not to reveal their real productive capacities in some cases, so as to pay a minimum negative transfer (see Sect. 1.5.4). In doing so, ELIE faces the same difficulty as many systems of unemployment benefits, where the amount received depends on the classification into one of the above two categories of unemployment. The possibility, or even the incentive of not revealing the rent of his productive capacities, directly questions the ability of ELIE to act as a basis for targeted reforms of social transfers. In this book, Ooghe and Schokkaert explore precisely this field of research by studying the possible impact that the minimum income k w N stemming from ELIE transfers could have, if these transfers were widened to the unemployed and to persons of low productive capacities working less than k. According to these authors, the inefficiency entailed by that extension would be less than that of linear income transfers stemming from classical taxation theory.

26

C. Gamel and M. Lubrano

1.6 Conclusion In presenting this personal and simplified treatment of the theory of macrojustice, we hope that we have managed to capture the interest of many potential readers so as to encourage them to form their own opinion. This book, designed primarily as a multidisciplinary forum of exchange on macrojustice, should also help them in this intellectual endeavour. We cannot finish this presentation without acknowledging those people who, at one time or another, have assisted us during the past four years to complete this project. First of all, our colleagues Nicolas Gravel, Cecilia Garcia-Penalosa, Carine Nourry, Alain Wolfelsperger for their external refereeing contributions; Steve Bazen, Olivier Chanel, Jean Magnan de Bornier and Hubert Stahn for their friendly material and financial support; JeanSébastien Gharbi, whose PhD topic is on macrojustice, for his editorial help; Sébastien Mengin for his invaluable help with the LaTex format of Springer; Laximi Rodrigues and Marjorie Sweetko for their help in translation and, of course, Martina Bihn, editorial director at Springer for her constant and faithful support. We also acknowledge the logistical support of Greqam and Idep and a welcome grant from the University Paul Cézanne.

Appendix A The Inadequacy of Extreme ELIE Schemes In the graphical presentation of the microeconomic “income-leisure” tradeoff, the space of individual choices is bounded by the budget constraint y D w  w; its affine term w corresponds to the full-time wage rate of the individual. This wage rate is also the full time wage when his leisure time  is zero, since  is normalised between 0 and 1; the product w corresponds to the opportunity cost (loss of income) of a non-zero leisure time. This fulltime salary is obviously higher for those with a productive capacity that is paid at full-time wage wmp than for individuals that are regarded as less productive on the labour market (wage wlp ). Moreover, if we substitute leisure time  to working time ` (as ` D 1  ), the actual income y is equal to ` w. When the parameter of equalisation k is equal to zero (in the absence of any redistributive scheme), the domain of choice of the most productive is larger by inclusion than that of the less productive, and the latter cannot fully adhere to such a situation where their freedom is restricted. In Fig. 1.1, the two budget lines have a single common point K on the X axis. This

1

Why Should We Debate the Theory of Macrojustice?

27

w, y wmp

wlp

y = w − λw

ymp ylp

K

0 λ

0.5

1 − λ =  = 0.5

1

λ

Fig. 1.1 “Classical liberalism” (k D 0)

point corresponds to a zero income (y D k w N D 0) for a maximum leisure time ( D 1). In other words, both categories have no income if they do not work at all (` D 0). Apart from point K, for a same non-zero working time (` ¤ 0), the least productive earn an income that is consistently smaller than that of the most productive, as illustrated in Fig. 1.1 by the cases of fulltime activity (wlp < wmp , if ` D 1) and of half-time activity (ylp < ymp , if ` D 0:5). In contrast, if k is equal to one, everyone yields to society his entire income derived from his productive capacities: in return, 100% of average income is redistributed (k w N D w). N The domain of choice of the most productive is now included (as a result of a negative net transfer [tmp D .w N  wmp / < 0]) in that of the less productive (enlarged by a positive net N transfer [tlp D .ww lp / > 0]). In Fig. 1.2, when translating the budget lines, downwards for the former and upwards for the latter, the only common point K between these budget lines is located on the Y axis and corresponds to the average income coming from the redistribution of the five working days of all individuals: both categories have access to it by working full-time. However the most productive can hardly adhere to such a scheme, because apart from K, they have an income which is smaller than that of the less productive, as illustrated in Fig. 1.2 by the case of half-time work (ylp > ymp );

28

C. Gamel and M. Lubrano w, y wmp

ti = k(w¯ − wi) = (w¯ − wi)

K

k w¯ y lp

t mp < 0 t lp > 0

wlp yp

C

0 λ

0.5 1 − λ =  = 0.5

B 1

λ

Fig. 1.2 The “slavery of the talented” .k D 1/

They get an income equal to that of the least productive only when working a lot (` D 1) and when choosing point K on their budget line. This situation is called the “slavery of the talented”, a case which is not part of Kolm’s theory of macrojustice.22

Appendix B The Interest of Intermediate ELIE Schemes In intermediate ELIE schemes, k is strictly between zero and one, which better fits the ethics of Kolm’s “process liberalism” where 0  k  k e : the domain of choice of one category is no longer included in the domain

22

The most productive could even keep no positive income after transfer if their working time were lower than a minimal threshold (segment B  C ), a situation that would also exist, but to a lesser extent, in intermediate ELIE schemes (See Appendix B, Fig. 1.4).

1

Why Should We Debate the Theory of Macrojustice?

29

w, y ¯ wpi w,

ypi tpi = k(w¯ − w) ¯ =0

K

kw ¯

0

λ

1

0.2 1 − λ =  = 0.8 1−k

0.6

λ

k = 0.4

Fig. 1.3 The situation of the pivotal category (wpi D w) N in one ELIE scheme of “process liberalism” .k D 0:4/

of choice of the other category. We take here the example of a parameter of equalisation k equal to 0.4. In Fig. 1.3, we are interested only in the pivotal category of the population, those for which the monetary estimate of productive capacity wpi is equal to the average wage w N of the entire population. Individuals in this pivotal category are thus forced to pay 40% of the market value of their productive capacity, i.e. k w; N in return, they receive 40% of the average monetary value of the productive capacity of the entire population, or again k w. N The net N transfer t they receive is neither negative nor positive, but zero Œtpi D k.w w/ N D 0, irrespective of the true duration of their working time. Under these conditions, they reach an actual income [y D ` w] N defined by their budget constraint that remains insensitive to the introduction of ELIE transfers: Their income will be higher, equal to or lower than k wN (even zero), when ` is higher, equal to or lower than k (even zero). In Fig. 1.3 we illustrate as an example the case of an 80% activity (4 working days a week, ` D 0:8), which corresponds to the first case (` > k). Compared to this pivotal category, the situation of the most productive and that of the least productive will be characterised, for the same ELIE scheme (k D 0:4), by a non-zero net transfer, negative for the former

30

C. Gamel and M. Lubrano w, y wmp t mp < 0 y mp ti = k(w¯ − wi)

¯ wpi w, ypi wlp ylp

K

¯ kw tlp > 0 0

D C

λ = 0.1

1 − λ =  = 0.9 1−k

0.6

k = 0.4

B 1

λ

E

Fig. 1.4 The juxtaposition of the three categories (wpi , wmp , wlp ) in one ELIE scheme of “process liberalism” .k D 0:4/

[tmp D k.w N  wmp / < 0], positive for the latter [tlp D k.w N  wlp / > 0]. In contrast to the situation of the pivotal category, their budget constraint will be affected by the existence, in both cases, of net non-zero transfers. Figure 1.4 summarises and compares the situations of the three categories of individuals who have been successively identified: invariance of the budget constraint of the pivotal category, downward translation of the budget line of the most productive as a result of net negative transfers, upward translation of that of the least productive as a result of positive net transfers, all resulting by construction in the intersection of these lines at K with coordinates (1k, k w). N 23 Far from necessarily converging to point K in which equality would be perfect both in terms of income .y D k w/ N and in terms of leisure time . D k/, individuals can of course choose any other point in their respective domain of choice, since they are a priori free to exploit as they wish 23

Figure 1.4 is in fact a more complete version of Fig. 9.1 given in Kolm (2005, p. 157) and entitled “The geometry of ELIE” (see also Kolm, this volume, p. 104); our Figure allows to track down the downward translation of the budget constraint of the most productive.

1

Why Should We Debate the Theory of Macrojustice?

31

their personal productive capacities. In Fig. 1.4 the very simple example is represented where each individual belonging to any of the three groups chooses to work a lot .` D 1   D 0:9/: although compressed by ELIE transfers, a certain income hierarchy remains .ymp > ypi > ylp / and everyone’s income is higher than the amount k w. N

References Dworkin, R. (1981). What is equality? part 2: Equality of resources. Philosophy and Public Affairs, 10(4), 283–345. Elster, J. (1992). Local justice. Cambridge: Cambridge University Press. Fleurbaey, M. (2006). Book review: Serge-Christophe Kolm, Macrojustice, the political economy of fairness. Theory and Decision, 60(1), 113–118. Gamel, C. (2005). Compte rendu de l’ouvrage Macrojustice (2005). Revue de Philosophie Economique, 12(2), 181–191. Guitton, H., & Margolis, J. (1968). Economie publique. Paris: CNRS. Guitton, H., & Margolis, J. (1969). Public economics: An analysis of public production and consumption and their relations to the private sectors. London: Macmillan. Kolm, S.-C. (1963). Les fondements de l’économie publique : Introduction à la théorie du rôle économique de l’État. Paris: I.F.P. Kolm, S.-C. (1971). Justice et équité. Paris: Cepremap. Reprint, Paris: CNRS, 1972. English translation, 1997, Justice and equity. MIT Press: Cambridge, MA. Kolm, S.-C. (1982). Le bonheur-liberté (bouddhisme profond et modernité). Paris: P.U.F. Kolm, S.-C. (1984). La bonne économie (la réciprocité générale). Paris: P.U.F. Kolm, S.-C. (1985). Le contrat social libéral (théorie et pratique du libéralisme). Paris: P.U.F. Kolm, S.-C. (1986). L’allocation des ressources naturelles et le libéralisme. Revue Economique, 37, 207–241. Kolm, S.-C. (1996). Modern theories of justice. Cambridge (USA): MIT Press. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge (UK): Cambridge University Press. Kolm, S.-C. (2006a). Liberté, justice et efficacité: Distribution, impôts et transferts optimaux. Revue Économique, 57(1), 55–84.

32

C. Gamel and M. Lubrano

Kolm, S.-C. (2006b). Reply to J.E. Roemer’s review of Kolm S.C. Macrojustice, the political economy of fairness. Journal of Economics, 88(1), 87–91. Kolm, S.-C. (2009). A response to Erik Schokkaert on Macrojustice. Economics and Philosophy, 25(1), 85–98. Mirrlees, J. A. (1971). An exploration in the theory of optimum income taxation. Review of Economic Studies, 38(114), 175–208. Rawls, J. (1971). A theory of justice. Cambridge (USA): Harvard University Press. Roemer, J. (2005). Book reviews: Kolm S.: Macrojustice, the political economy of fairness. Journal of Economics, 86(3), 301–304. Schokkaert, E. (2009). Critical notice: Macrojustice as a research program. Economics and Philosophy, 25(1), 69–84. Sturn, R. (2008). Macrojustice, the political economy of fairness. By SergeChristophe Kolm. Economica, 75(297), 196–197. Varian, H. (1974). Equity, envy and efficiency. Journal of Economic Theory, 9(1), 63–91.

•

Part I

The Macrojustice of Serge-Christophe Kolm

Chapter 2

General Presentation Serge-Christophe Kolm

Abstract This chapter is a general summarised presentation of the problem of defining the best possible choice of the overall income distribution in macrojustice (as opposed to microjustice and mesojustice concerned with allocations directly of specific goods or in particular instances). The three classical polar principles advocate respectively self-ownership and transfers motivated by comparisons of individuals’ incomes and welfares. They are synthesised by people’s general ethical views in the society that has to implement the policy. Actual policies show the material possibilities (for instance the exemption of overtime labour earnings from the income tax that amounts to basing transfers on capacities). The result is a simple and richly meaningful distributive structure that means, jointly, equal real liberty; adding an egalitarian and a classical liberal parts of income; reciprocity by providing each other with the product of the same labour; and an equal basic income financed by an equal partial labour of each. This core principle is then applied taking all the actual economic and social phenomena into account. The questions this may raise are analysed and answered in the various chapters of this volume.

2.1 Foreword: Preliminary Remarks on this Book and Its Introduction 2.1.1 Problem, Achievements and Acknowledgments If “justice is the first virtue of societies” (Aristotle) or of their institutions (Rawls), macrojustice is the first virtue of justice. It concerns the overall

S.-C. Kolm EHESS, Paris and CREM e-mail: [email protected]

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_2, 

35

36

S.-C. Kolm

architecture of human interaction and organisation. Since a just patch on an unjust body is hardly a satisfactory solution – insofar as this can be avoided – macrojustice should be the first concern of the quest of men of good will for the Grail of defining what is just and what justice could mean. This volume, however, is restricted to economic macrojustice, that is, to the nature of economic interaction and to the distribution of the main resources – with an emphasis on the most important and delicate one, human capacities. Therefore, the main issue of “social justice”, to wit the class or caste structure of society, does not appear directly, although its economic basis is an essential aspect of it. The questions answered are not surprising, although the answers, often simple, are not (not yet?) commonplace. They include the following ones. What do we owe to each other (as Thomas Scanlon puts it)? How do we implement the fight on crippling poverty, wanted by everybody (a universal moral public good)? Which necessary concept and form of liberty and rights make “freedom from” and “freedom to”, freedom to act and possibility to act, or formal and real liberty, consistent with one another? What is the just and optimum tax on income or subsidy to it – if any –, as implied by unanimously desired properties of fairness and goodness? How can fairness, solidarity and public finance be respectful of the unanimistic Pareto efficiency required by social liberty, efficiency and stability (hence possibility and existence of the social state and policy)? Which structure of transfers permits disentangling, from these various properties, a single parameter describing each specific society’s choice of the distribution that implements its conception of itself as a solidaristic community, from no redistribution to substantial one according to the case? For instance, distributive transnational transfers hardly exist whereas national political communities often redistribute 30% of GNP with a large acceptance (these actual redistributions amount to an equal sharing of the income of 1–2 days per week, from the USA to Scandinavian homogeneous, social-democrat and welfare-state national communities). Moreover, social judgments and emotions aroused by yielding and receiving such transfers are essential for their possibility and desirability (Pierre Livet). Free education (and learning grants) financed by a tax on earnings has both essential effects of making this redistribution favourable to growth and of optimising educational choices and the production of acquired productive capacities (David de la Croix and Michel Lubrano). The obtained redistributive scheme relates to the classical semi-welfarist “optimum taxation” in various ways (“semi” because these studies consider almost always that individuals have the same utility function, thus discarding differences in tastes – as Mirrlees says – and in hedonic capacities). Bourguignon and Spadaro (2008) argue they show that this model does not describe, even implicitly, actual fiscal choices (as Rawls argued in 1971), and the present

2

General Presentation

37

French tax law shows that basing the tax on the wage rate without evasion is very easy. Nevertheless, the normative importance of this model and the fact that most income taxes are based on full earned income make the relations between both theories be of foremost importance (Laurent Simula and Alain Trannoy). Moreover, although the simple application of the obtained transfer scheme in a productive distributive society implies a notable minimum income, notably along with the various social insurances which have their own logic, focussing the receipts towards helping the poor is an important possibility, and this can be done in a humane and directly solidaristic social organisation of the distribution (Alain and Justin Leroux). Fiscal rules often also aim at particular incentives or have to remain in a transitional period, wage rates may still be uncertain and taxation has to be considered in an inter-temporal growth context: incorporating all such aspects is a requisite for application (Michel Lubrano). Beyond issues of inter-temporal disincentive and legitimacy, various reasons may require specific capital taxation which may logically extend the structure of the distribution of human resources, and the obtained distributive scheme should be compared with proposals of basic income (Claude Gamel). Although the basic obtained redistributive scheme does not apply to low labours (particular theories are proposed for these cases and for unemployment), and although very few people declare working less than 20 hours per week, a universal application of this scheme reveals an interesting way of financing a universal basic income with lower waste than that of standard taxation (Erwin Ooghe and Erik Schokkaert). Finally, since a main interest of the obtained distributive scheme is the variety of its justificatory properties (equal real freedom, balanced labour reciprocity, basic or minimum income financed by equal labour, etc.), more justifications can be added, they can be related to broader scopes of normative distributive properties, and this can be done most rigorously in an axiomatic framework (François Maniquet). Anyone interested in distributive or social justice should be extremely grateful to all these authors for their essential sophisticated additions to the initial, simple basic kernel, and I would like to claim to be the first of these interested admirers. I have no doubt that the reader will highly appreciate the symbiosis of scientific imagination, superb scholarship and sense of the human importance of the topic present in each of these contributions. Its having induced such great work is undoubtedly the main value of the initial logical framework. In addition to their noted crucial contributions to essential issues of the general theory of macrojustice, Claude Gamel and Michel Lubrano actually created the present volume which, from conception to completion, owes its very existence to their vision, initiative, skill, hard work, time and dedication. They thought of and organised the initial conference, demanded the chapters

38

S.-C. Kolm

and organised their reading, and, last but not least, prepared with great care the manuscript for publication. This deserves the intense gratitude of anyone interested in social justice or involved in its study. Almost all these tasks required scientific innovation. To take just an example, the association of economics and philosophy, which the editors chose as a principle of the enterprise, is not only a laudable (and probably too rare) aspect of scholarship: it is, in such branches of normative economics, simply unavoidable for making sense and progress (actually, I think the proper view to be just not to make a difference between these two fields). This book, as it stands, happens to have a simple logical structure. All the chapters from this one on are strictly complementary. After this presentation and a summary of the basic theory in Chap. 3, all following chapters have the same natural rationale. Each recognises that some important aspect has not been studied yet, and it analyses it. Often perfectly. This complements the rest of the volume. As far as I am concerned, I can only admire and applaud. I fully agree and add the author’s contribution to my understanding of the question.

2.1.2 Sometimes Distinct Possible Conceptions: Macrojustice and the Introduction The theory of macrojustice presented in the Introduction (Chap. 1) is on some points an original construction and is, therefore, much more interesting than a simple duplication of the content of the volume Macrojustice would be. It is sometimes a particularly perceptive analysis of the volume’s points, such as about the possible separation of rights in oneself in rights to act and rights to the rent (a priori economic value) of one’s productive capacities, or when providing an enlightening summary of the theory as a three-level structure: the basic redistributive scheme, the degree of social equalisation, and individuals’ free choice of labour (which, however, initiates as a basic right belonging to the first-level complex). On some other points, it chooses a larger or a smaller scope, a different reason or model of man, or, perhaps, it requires specification (including in a couple of ironical allusions). These differences may add richness, the sole inconvenience being – in an introduction – the risk that the reader credits the volume Macrojustice with ideas it fails to include. A basic simple structure for reducing inequalities in given earning capacities consists in redistributing the individual proceeds of the same labour k (which is also a degree of redistribution, equalisation, solidarity or community). This applies naturally only to individuals i who provide a total labour

2

General Presentation

39

`i not lower than k (Macrojustice, pages 115, 116, 117, 127, 133, 134, 155, among others). However, it is sometimes interesting to have a look at what an extension of this scope would give. The Introduction rightfully acknowledges this situation but the places and ways in which this is done risk to leave the reader with the impression that Macrojustice sometimes applies this policy to cases for which it does not. To be explicit about the consequences, and as the Introduction points out, the Macrojustice policy induces people to reveal their productive capacities (basically because, if wi is individual i’s wage rate, the part of disposable income .`i  k/wi is an increasing function of wi when `i > k), and it has no aspect of Dworkin’s “slavery of the talented” whatsoever. A most important property of the human mind is its capacity to take either a self-centered or an impartial point of view (Thomas Nagel’s “view from nowhere”). The Introduction and Macrojustice happen to make, about the relations between these two standpoints, two different hypotheses which both have grounds in reality. In Macrojustice, the impartial standpoint intervenes essentially in the collective political and moral choice of the degree of redistribution, along with the related senses of justice, community and solidarity, within a series of methods intending to reveal, induce, infer or use it, often relying on hypothetical information or agreement. This analysis is the topic of a separate part, Part 4, probably the most important of the volume (and certainly the deepest), noted in the Introduction. People can also have some sense of injustice in everyday life. However, a laudable didactic intention of the Introduction makes it present two people who just spontaneously agree on the distribution between them, each one solely “not fully adhering” to a situation in which she is less free than the other. Since this naive and irenic behavioural hypothesis is attributed to me (“according to Kolm”), I am sorry to have to say that I fully reject it. Is it a necessity of didactic simplification? Actually, both self-interested selves want to have more (no suffering is involved here), and both impartial selves can criticise any of the inequalities. The final result would require further considerations, and, indeed, that of a larger (macro) population. The next remark concerns just an omission, but a crucial one (if it were not an omission the whole presentation would be contradictory). The Introduction states that, in the theory, the individuals have no right in the value of their productive capacities, whereas the most distinctive feature of the Macrojustice theory is that this holds for the product of labour k only, the individuals being the full owners of their capacities for the rest. This restriction is also implied by the logic of the rest of the Introduction. The Introduction attributes this position to Rawls who, indeed, denies self-ownership in all productive capacities. This is also the view of “welfarism” and actually the concluding sentence of Mirrlees’s article of 1971 (hence it is also implied

40

S.-C. Kolm

by the Simula-Trannoy chapter in this volume). Rawls, however, justifies this position more in detail. It is also Dworkin’s theory of 1981, which he rejected. Actually, this theory is not even defined because the total time which, multiplied by the wage rate, gives this non-self-owned value is not a priori defined: is it 24 hours a day, or 16 only? In the theory of Macrojustice, this time is not total time but duration k only. We have noted that present-day national redistributions amount to a k from 1 to 2 working days a week. 2:5 days would make an extremely redistributive actual national society. Since k  0, the solution k D 0 is a priori a particular limiting case (Macrojustice, pages 147, 159, 160, 162, 285, 316, 317, 318, 321, 322, 565, among others). This is full self-ownership, an absence of redistribution, and what may be called “classical liberalism”. By contrast, the Introduction theory often bans this case from the scope it considers, which is, of course, a possible and a priori legitimate ethical choice. This case, however, may happen for a number of societies (insofar as such groups deserve this name). For instance, this happens for the “society of nations” since international distributive transfers are practically absent. For judging macrojustice theory, a classical liberal need only say: “I choose k D 0”. The basic point is that Macrojustice is analytical, not dogmatical. And the analysis refers to different spheres of reason. One of them gives the general distributive structure, without specifying the equalisation coefficient k, and hence with k D 0 as a particular borderline case. Then another, different type of reason gives the level k appropriate for a given society, “desired” by it in a certain sense. This latter stage also occurs for different theories, for instance for determining a social welfare function more directly applied. Then, if markets give very different earning capacities to people who constitute a sufficiently tight community, this society’s shared ethics will want the relevant distribution. The last difference between both presentations I will note here concerns the reason for the limited welfarism of macrojustice. The Introduction puts forward a question of information: the policy maker does not know the individuals’ utility functions whereas they can be better known in a choice of microjustice. This remark is obviously exact (consider for instance an allocative choice within a family), and it is noted in Macrojustice. However, the policy maker can make a choice in uncertainty (for instance, for determining a tax schedule, as in my books on public economics of 1970). Moreover, the basic reason seems, rather, that the public does not consider differences in tastes and capacities to enjoy to be relevant for choosing the macrojustice allocation – as tested about the income tax schedule, for instance – (Macrojustice, pages 96 and following). This holds even with known utilities. There exist, in addition, comparisons of variations of individuals’ “welfare” implicitly considered as the same concave function of individual income (a classical justification of transfers from rich to poor, for instance), but only

2

General Presentation

41

when at least one of the compared incomes is sufficiently low, which leads one to the realm of specific reasons of microjustice or mesojustice.

2.2 Overall Distribution: Structures and Possibilities 2.2.1 A Demanded Distributive Structure The next chapter, summarising the volume Macrojustice (Kolm 2005), shows that a few basic properties, desired by everybody, of the overall distribution of goods, imply a particular, simple and richly meaningful structure of the distribution of incomes and of the transfers it may need. This structure can apply with various degrees of redistribution (from none), depending on the particular ethos and sentiments of community of the society under consideration. This results, however, focusses on the few main variables. It has to be completed by the introduction of the other phenomena. This is the role of the other chapters of this volume. This method is universal. Science is cumulative cooperation. Initial simple structures are completed and applied, for the relevant important facts, by researchers who build the rooms of the common house of knowledge. The social ethics of the overall economic distribution exemplifies this collective endeavour. This chapter’s presentation outlines the necessary simple theory (which synthesises the classical alternatives), and it emphasises the necessary roles of the various essential parts of this building – not a palace yet, but no longer a hut, and rather, hopefully, a useful workshop for a product in demand. One first tries to present the resulting structure of the overall income distribution in one of the most natural ways. This consists in the association, in each individual income, of an egalitarian part and of a classical liberal part. The latter secures social liberty and efficiency (and some self-ownership) whereas the former provides solidaristic justice when it is required. The relative importance of these two elements depends unavoidably on the extent to which the particular society in which this is applied deeply (consciously or implicitly) feels that it is or should be a distributive community.1 The “egalitarian income” equally shares the products of the same labour (with different earning capacities): this is Equal-Labour Income Equalisation or ELIE. Moreover, individuals freely choose to work and keep the extra income earned: this is their classical liberal income or the classical liberal

1

Ways of making this sentence operational constitute the topic of Part 4 of the volume Macrojustice.

42

S.-C. Kolm

part of their income. As a particular a priori possible case, the egalitarian part may vanish and income is classical liberal only (as an example, distributive transfers are de facto presently negligible at the international level, across nations). The obtained distribution structure has a number of significant properties. The equal-labour equalisation achieves “from each according to her capacities, to each equally”. It provides a basic income (which is also a minimum income) financed by an equal labour of all. It also amounts to each yielding to each other her earnings from the same labour, in a general equal labour reciprocity. Moreover, the whole distribution turns out to amount to equal basic and real liberties. It also amounts to each receiving according to deserts (labour) for the equalisation labour and to merit (labour and given earning capacities) for the rest of labour. Practically, this distribution can be achieved by an income tax (or subsidy) with two bonuses: a uniform tax rebate and an exemption of overtime labour earnings above a rather low labour benchmark. These policy structures are actually applied and they induce no cheating for simple practical reasons shortly noted. This distribution provides the right incentives to use optimally one’s capacities and it secures social efficiency. This structure can apply all degrees of solidaristic distribution. As a particular case, a society may require no redistribution (classical liberalism). Other societies basically feel they ought to have substantial redistributions (e.g. present-day national political communities).2 Note that labour is by far the largest source of income, especially in an inter-temporal view when the value of capital, which is produced by definition, is attributed to its given or natural factors, human and non-human (the allocation of non-human natural resources is studied in Kolm 1985, Chap. 10, Kolm 1986, 2005 and in Sect. 12.3). This distribution results from or leads to four groups of studies none of which is less important than the other. (1) The philosophical underpinnings of distributive justice, such as the scope of people’s welfarist judgments, the types of liberty, the distinction between the free use and the value of capacities and the nature of communities. (2) The analysis of the properties of this simple redistribution scheme. (3) The analysis of the various methods for estimating the degree of redistribution or equalisation that corresponds to a given community – the same question is raised by all other distributive structures, and these methods apply to them also (for instance to finding the right social welfare function). (4) The introduction in the basic simple model of the various other aspects of reality to constitute the

2

This theory has been presented in a number of publications starting a long time ago (for a very short sample see Kolm 1993, 1996a,b, 2005).

2

General Presentation

43

general theory, such as: multidimensional labour including education; nonproportional earnings; imperfect markets; low labours and unemployment for all reasons; exceptionally high incomes; inter-temporal distribution; savings; bequest; the allocation of non-human resources; households; relations with other taxes, expenditures, aids and insurances; in particular specific aids; voluntary transfers; information and uncertainty; the implementation of tax and subsidy policies (modalities, information, checking, penalties); and so on.

2.2.2 Implementation The presentation of this theory may be helped by the comparison with the two best-known families of theories of distributive macrojustice, that of the “welfarist optimum income taxation” and that of John Rawls who presents his proposal in opposition to the former (which “is never applied”). Actually, as we will see, both can be said to have the same egalitarian ideal, equal income and equal labour (for the standard optimum tax model with identical utility functions as a result of discarding irrelevant differences in tastes – Mirrlees 1971 – and for Rawls after 1974). Yet, both hold that their ideal is strongly checked by disincentive effects due to the taxation of labour. As shortly noted, however, a present national experience proves that it is possible to base the tax on the wage rate rather than on total earned income, and thus not on labour, for wage labour (which is 9/10 of all labour). This is made by exempting overtime labour earnings over a rather low official labour benchmark from the income tax. There happens to be no cheating because this could practically not be done without the tax authorities being aware of it and informed about it.3 Other labour inputs than duration are taken off the tax base in various ways – for instance, taxes on extra earnings due to education finance free public education.4 This, and a number of other facts and reasons, shows that both welfarist tax studies and Rawls are undoubtedly much too pessimistic about the possibilities to base taxes much more on

3

The standard explanations of this fact by tax officials are now confirmed by economic analyses. Firms (with more than very few employees) cannot hide their actual pay accounting which is much used and known (some employee may report fraud). See for instance Kleven et al. (2009) and the literature they quote. This third-party reporting information explains why governments can extract high amounts of taxes. It should now be used not for this fiscal efficiency only but also for economic efficiency by choosing less elastic tax bases (and for moral “efficiency” by choosing the most “just” basis). Actually, not to declare overtime labour and pay is much easier than to declare false ones while falsifying the wage rate also. 4 The model of D. de la Croix and M. Lubrano in Chap. 11 studies subsidies to education.

44

S.-C. Kolm

given capacities.5 One may also add that the tax authorities’ information about total incomes is often very poor (classically, about 30% of the base evades the income tax on average, with much heterogeneity). This exemption of overtime labour earnings over a rather low benchmark is presented here only because it seems to be the best large-scale empirical proof of the possibilities for the tax-subsidy base to hit much closer to abilities than is usually assumed in scholarly circles. This tax structure exempting marginal labour has the classical economic virtue of being favourable to efficiency and freedom of exchange.6 As we have seen, the ELIE distribution can be realised in this way with two bonuses in the tax on earned income: the exemption of overtime labour earnings over a rather low benchmark and a uniform tax rebate or credit. The present French income tax has both features, in particular the exemption. This was presented in the volume Macrojustice, the Political Economy of Fairness. Discussions of these ideas – in which I put the tax exemption forward in order to be practical – happened to inspire the project and the realisation of this policy.7 It seems rather uncommon that a simple piece of theory has such an almost immediate effect on policy. And it is probably still less usual that this policy experience happens to answer a basic question in the field, and does it in a sense opposite to the “conventional wisdom” of the discipline. Just an effect of theory on policy is not usual in this specific field. The two best-known theories noted, the welfarist optimum income tax and Rawls’s are very generally found to be highly interesting, but have they been applied in their nearly four decades of existence? (Bourguignon and Spadaro 2008 find that the actual tax structure cannot be derived as the theory assumes it is).8 These theories’ egalitarian ideal could be achieved rather well, but – as noted below and in Chap. 3 – people would demand a classical liberal

5

See, for instance, Sects. 2.6.5, 2.6.7 and 3.4 below, and Chap. 10 of Macrojustice. In situations of normal employment, but the tax structure and law cannot be changed too often (and overtime labour is often demanded for tasks for which full hiring would not have been made). The exemption includes “social contributions” (financing all social security branches). The marginal wedge abolished is often 65%. 7 There were first discussions in economic and wide-audience journals, and conferences. A presidential candidate then made this exemption of overtime labour earnings the centerpiece of his program. Hence, this proposal from Macrojustice has been about the most discussed issue nationwide during the presidential campaign and after. This policy had been proposed in an optimum income tax study in Kolm (1974), but without publicity and actual effect. The present law has been effective since October 1st, 2007. 8 In contrast, a theory formally analogous to the optimum income tax, but applied to another field, the optimum non-linear tariff of public utilities (with, moreover, explicitly uncertain different individual utilities) was rapidly applied by the national electricity board (Kolm 1970a,b). An application to the schedule of the progressive income tax was proposed but was discarded for not taking account of all the ethical criteria used for this choice. 6

2

General Presentation

45

addition for unanimous improvement and freedom of exchange, and the distributive structure has to comply.

2.3 Relations with Basic Ideas Before pursuing the presentation of the distributive structure in question, it may be useful to continue to situate it with respect to the main ideas about its topic. Besides evaluating the macro distribution from the direct consideration of that of incomes, goods or leisure, two other values are commonly used in this evaluation. One is the concept of “welfare”. The other is the issue concerning free exchange, self-ownership and the classical liberal theory.9

2.3.1 Welfare “Welfare” is a most ambiguous term. People probably first take it to mean the set of consumption goods. Or, perhaps, income extensively understood and the amenities of life. Economists represent it by utility functions and therefore include in it a large psychological element. However, this “utility” is due to various psychological (and physiological) phenomena of different natures. If these phenomena need not be distinguished for many descriptive and even normative applications, they have to be for others, and in particular for important judgments concerning distribution. An elaborate example is provided by Bar-Hillel and Yaari’s (1984) study which shows that people’s moral evaluation of a distribution depends very much on whether the “utilities” it influences refer to individuals’ tastes or needs. Concerning the overall income distribution, for instance the justice of the income tax, questions of the type presented in Sect. 3.3.2 show that people hold two types of psychological differences to be irrelevant for evaluating this specific justice; (1) differences in individual tastes, and (2) differences in individual capacities to enjoy (hedonic capacities). One can discard these differences from individuals’ utility functions and obtain, as a result, an interpersonally comparable “residual” which can be considered as the personal “welfare” relevant for this issue, as shown in Sect. 3.3.1. This remark about tastes is classical. For instance, Mirrlees (1971) says that he considers identical individual utility functions because “differences in tastes raise different kinds of problems”.

9

As for equality, prima facie equal treatment of equals in the relevant characteristics is a requirement of rationality (e.g. Kolm 2010).

46

S.-C. Kolm

The irrelevance of tastes is classically shown by examples of “expensive tastes” (there are others of “cheap tastes”), for instance by Arrow, Rawls, Dworkin, Cohen and many others. The remaining interpersonally comparable welfare can then be used to define the egalitarian ideal. However, a classical liberal element of income may need to be added, as shortly noted. The noted psychological comparisons held to be irrelevant are so for this specific problem, overall distribution in “macrojustice” and for instance the income tax, and not for all distributions, of course. For distribution between people who know each other well and have developed a particular empathy for the others, for example within a family, differences in tastes and in capacities to enjoy are obviously generally of prime importance. Similarly, when higher utility means lower suffering, interpersonal comparisons of it are held to be relevant for the allocation of relief. However, these cases are not the general situation for a large society not in a state of disaster, in which they refer to issues of microjustice (including particular public or private aid when needed).

2.3.2 Classical Liberalism When the egalitarian part of income vanishes, the distributive structure presented becomes classical liberalism, a particular case. Classical liberalism is the standard theory based on rights and not what Murray Rothbard (1973) introduced as “libertarianism” in which the distribution results from force only. Using the same term for both constitutes an utmost confusion. The classical liberal theory is as necessary when there is an egalitarian part of income, since it describes what happens from the basis constituted by the given distribution that yields this egalitarian income. Hence, all aspects of this theory are useful. A place for classical liberalism is largely unavoidable in a large society. Replacement of pieces of it by processes that imply better social relations than market exchange (cooperatives, reciprocities, etc.) can be highly welcome, but this is not the present topic. Classical liberalism, possibly on the top of some egalitarian distribution, has two rationales for non-exchange transfers: the redress of past torts and violations of the relevant rights, and gifts. Gifts for helping people in need are joint contributions to a public good for the givers, the welfare or income of the beneficiaries. Hence they often have to be implemented by the public sector to avoid free riding. The government would even have to guess the givers’ desires because they are generally too numerous to agree directly by a contract. These actually voluntary transfers are then realised by public

2

General Presentation

47

constraint, as are fully forced distributive transfers when they exist. These two types of transfers are not distinguished in their realisation, and they are not in people’s mind either. There just is the awareness of the existence of different rationales and motives: benevolence, justice (the possible acceptance of which by payers requires them to be in another, more impartial, state of mind), but also, possibly, the benevolent desire that justice be realised, and the effect of knowing that others “provide their share”. The basically voluntary motive is far from negligible, as shown by the purely private transfers in some countries in spite of the public good aspect, and by various enquiries and political stances about public transfers. Yet the most conspicuous feature is the mixture of motives.10

2.3.3 Autonomy, Reciprocity and Justice This short paragraph may be irrelevant. However, since I am often questioned about the relations between different topics I have considered, a few words may clarify the place of justice. Justice, classically dubbed the “first virtue of society”, is better seen as its third-best quality. These last words of the previous book Modern Theories of Justice (Kolm 1996a) are also the first ones of Macrojustice. Two considerations render justice superfluous in so far as they can exist. These three levels correspond to three actual but problem-raising psychological possibilities of the human mind: the self-control of the birth of one’s desires, empathy and reciprocity, and impartiality. If people sufficiently controlled the birth of their own desires, as advanced Buddhist psychology shows the theoretical possibility of it, they would be satisfied without competing for scarce resources.11 However, the diffusion of this information and training cannot be sufficiently achieved, by far, in modern mass societies. If, then, people cared sufficiently about one another, the conflicts about scarce resources would also vanish, although the consensual issue of sharing them would still remain. This benevolence could be supported by the important

10

The classical liberal theory is the topic of two books of mine, one to analyse it (Kolm 1985) and the other to criticise it (Kolm 1984b). Collective giving is analysed in three works (Kolm 2008a,c, 2009b). 11 This analytical Buddhism is presented in the volume in French Happiness-Freedom, Deep Buddhism and Modernity (Kolm 1982). Thirty two of the 651 pages of this volume have been translated into English in The Multiple Self, edited by Jon Elster (1986, pp. 233–265) (there are also translations in other languages).

48

S.-C. Kolm

psychological phenomena of reciprocity.12 One may be content with reciprocity and non-profit activity sufficient for influencing the general ethos of society – a market economy, not a market society. Even this is not that easy in large societies, and, at any rate, benevolent and reciprocitarian relationships are more likely to thrive when they implement fair distributions. Justice and fairness, therefore, constitute the third level, with their own successive levels of macrojustice (basic rights and overall income distribution), mesojustice (education, health, etc.)13 and the multifarious issues of microjustice.

2.4 Liberty, Equality: Macrojustice from Endogenous Social Choice Free from forceful interference, as they want it, people act and exchange, given the distribution of the resources given to society. These resources include prominently the individuals’ capacities to earn and produce. Actual opinions and policies about the overall income distribution result from two opposite tendencies. One is the limitation of redistribution to gifts (possibly publicly implemented) and the redress of past torts, implied by the full selfownership advocated by classical liberalism – an actual possibility. Other views advocate reducing income inequalities. The present French income tax law shows that it is largely possible to base the redistribution on individuals’ capacities rather than on total earnings which also depend on the chosen actual duration of labour. It is, and will be more precisely, completed for the other dimensions of labour, notably education. Other experiences and considerations concerning taxation comfort this conclusion. Hence, this policy can have only quite limited overall wasteful disincentive effects. More or less egalitarian opinions object to the fact that some units of labour of different people are remunerated differently because of individuals’ unequal given earning capacities due to their own characteristics and to society’s demand for their talents. The classical liberal ethics does not share this view. Coexistence (synthesis or compromise) of these two ethics can be manifested by each determining one of two parts of each individual’s income. On the one hand, an egalitarian part of income equally shares the proceeds that the individuals earn during an equal partial “equalisation

12 See the two volumes The Good Economy: General Reciprocity (Kolm 1984) and Reciprocity, an Economics of Social Relations (Kolm 2008b), and the Introduction to the Handbook of the Economics of Giving, Altruism and Reciprocity (Kolm 2006). 13 See for instance Kolm (2001a).

2

General Presentation

49

labour” (with their different earning capacities). This part of income is the “equalisation or egalitarian income”. It is equal to the equalisation labour multiplied by the average wage rate (earning capacity per unit of labour). It vanishes in societies that choose so. On the other hand, the persons freely choose their total labour and keep the extra earnings they obtain with their different capacities: this is their “classical liberal income”. A little formalisation may help. If individual i’s income is denoted as yi , her chosen total labour as `i and her wage rate as wi (earning capacity Pper unit of labour), and if w denotes the average wage rate (w D .1=n/ wi where n is the number of individuals participating in this distribution) and k denotes the equalisation labour, the noted distribution consists in replacing kwi by kw for each participating individual i, and individual i’s resulting total income is therefore yi D k w C .`i  k/wi ; which shows these two egalitarian and classical liberal (respectively) parts of income. Classical liberalism is the particular possible case k D 0. This distribution applies for individuals i and a level k such that 0  k  `i for a necessary reason shortly noted (along with the cases of low or zero labour). This implies yi  k w, a minimum income. This distributive structure is the simple Equal-Labour Income Equalisation or simple ELIE. This core is further completed to constitute the general ELIE taking further phenomena into account, such as those noted in Sect. 2.2. Nevertheless, it shows simply the mentioned properties of this distributive philosophy for the participants. For instance, each participant i receives the net transfer ti D k  .w  wi / (a tax if ti < 0). This amounts to the noted tax structure since it is also a net tax of ti D .k=`o /wi `o  k w which, if `o denotes the benchmark labour duration, describes a tax exemption of the earnings of labour `i  `o plus a uniform tax rebate or credit of k w (with a tax rate of k=`o ). Philosophically, with ELIE each individual participant faces life with two assets : her given earning capacities per unit of labour wi , and a financial asset ti D k .wwi / which compensates, to an extent k to be chosen (which may be zero), the handicap of being endowed with a meager wi , thanks to the symmetrical partial solidaristic liability entailed by the godsend brute luck of being born with a high wi . The whole distribution can be seen as the classical social liberty (freedom from forceful interference), and in particular free exchange, from a given allocation of the given resources constituted by the basket of the productivity wi and of the lump-sum transfer ti . Individuals i with wi > w pay a tax ti > 0 if k > 0, which is the larger the larger wi , but their final disposable income yi D wi `i C ti D .`i  k/wi C k w is the

50

S.-C. Kolm

larger the larger wi , and hence the larger the tax they pay, for each `i > k. In particular, if individual i pays a tax ti > 0 rather than 0, for each `i > k her disposable income is higher by .`i  k/.wi  w/ > 0 (since wi > w has to hold). This also permits to have the same income or consumption by working less.14 The equalisation labour characterises the limit between the two ethics possibly present in the overall income distribution in macrojustice. In a free society, it results from a basic collective choice that manifests the extent to which this particular society considers itself a solidaristic community. One of the most meaningful summaries of any redistribution is its “equivalent equalisation duration”, that is, the duration such that the full equalisation of all incomes during it diminishes an index of inequality as much as does the actual redistribution. This may suggest orders of magnitude (labour is the main source of income, particularly if the value of capital is allocated to the human and non-human natural resources that produced it). For instance, present-day national redistributions have an equivalent equalisation duration about between 1 and 2 days a week (from the USA to Scandinavian social democracies). But there is practically no redistribution at the world level and this duration is zero (classical liberalism), whereas sharing is the very life of families. Part 4 of the book Macrojustice presents the various methods that permit to estimate the degree of redistribution desired by a society in any form (including for “social welfare functions”). It is at least as important as the rest of the study (and certainly deeper). It rests on the consideration of the extent and limits of the citizens’ classical impartial selves and capacities for empathy. Distribution is, then, a public good. Information about others and dialog can be favoured, notionally extrapolated thanks to the appropriate models, and these results can be shown to the participants. Other theories developed there are the “rational, recursive original position” and the similar “moral time-sharing”; the “distributive moral surpluses”; the notional “impartialisation of preferences”; and “fundamental insurance”. Individuals with an average wage rate are not affected by an ELIE redistribution and hence reveal their impartial opinion.

14

This implies that the full self-ownership of classical liberalism cannot be deduced from social liberty alone but has to resort to a concept of natural right – a central historical question of social ethical theory. Indeed, the derivation from social liberty is based on the reasoning that a tax forces the person to work more, which is a kind of “forced labour”, at least if she wants to maintain her consumption. In the present case, however, the tax implies the possibility to work less and consume more (unless it is banned because one assumes a priori that the productivity wi is self-owned, but then this can only be from a conception of an a priori natural right).

2

General Presentation

51

When all members of a society reject or require some aspect of a policy, since they include people who decide on this policy such as officials and voters, this opinion is a priori implemented. Contrary properties are just not implemented. Then, the useful analyst can only try to show society the best way to realise what it wants. This is “endogenous social choice”: Vox populi, vox dei (unanimity guarantees the protection of minorities). This applies, for instance, to the basic liberties, economic (Pareto) efficiency, and, in very different ways, the choice of the degree of redistribution. The risk of bypassing this issue is simply to propose unapplied and hence useless studies. Another aspect of the same question is that it consists in saying that useful proposals are constrained by a condition of democracy. Macrojustice concerns the most general rules of a society, and notably their application to the distribution of the value of the main resources in general income. Society also faces innumerable cases of microjustice concerned with allocations specific in terms of the nature of goods, reasons, people, occasions, circumstances or criteria. A field of mesojustice for particularly basic specific goods concerning everybody (e.g. education and health) is also fruitfully distinguished. All these “spheres of justice” of very different sizes are complementary. Since people hold that taxing leisure is shocking and providing a wage complement to hours that provide no wage is absurd, they neither vote for nor apply such policies. But ELIE consists in taxing the units of labour k by wi w when wi > w, and in providing a wage complement of wwi to these units when wi < w. If k > `i , the units of the leisure k  `i would pay this tax or receive this wage complement. Hence k  `i : the equalisation labour cannot exceed the labours freely chosen by the people who participate to an ELIE distribution. The few people who work less than the equalisation labour of an ELIE distribution cannot be part of this equalisation policy. Actually, extremely few people declare working less than 20 hours a week,15 and we have noted that such a k would correspond to a quite redistributive present society (about 25% more than present-day Scandinavian social-democrat national communities). People’s abnormally low labour has specific reasons which make them fall in the scope of microjustice with a variety of possible criteria. Moreover, most of these cases can be naturally brought back into the general case, as with total or partial involuntary unemployment (also the object of macroeconomic, insurance, formation and labour market policies), and part-time labour contracts and second wages in families (as in the present French tax law), whereas the income of people with low wage rates depends

15

Communication of Emmanuel Saez.

52

S.-C. Kolm

little on their labour.16 The equalisation income (k w) then is a minimum income, the choice of which is equivalent to the choice of the equalisation labour (k), and which adds to the various micropolicies, including insurances of all types.

2.5 The Possible Identity of the Supposedly Enemy Paradigms and the Additions of ELIE The two most-discussed theories of distributive macrojustice, both presented in 1971,17 are Rawls’s “Justice as Fairness” and the theory of the “welfarist” “optimum income taxation” (Jim Mirrlees). Rawls presents his theory as the alternative to the principle the other purports to apply, the maximisation of a Social Welfare Function function of individuals’ utilities, which, he says, just is never actually applied. In reality, however, the ideal of the “welfarist” theory of optimum taxation as it is practiced (and in Mirrlees 1971) can be seen to be the implementation of Rawls’s view after 1974. In brief, this latter view is that social justice consists of equal income and equal leisure (hence equal labour, the complement of leisure). And the objective function of this “welfarist” theory, in which individuals’ utility functions are replaced by the same function, is, notably, the minimisation of an index of a bidimensional inequality, in income or consumption and in labour (see Sect. 3.3.1). ELIE merely consists in the extension of this Rawls–Mirrlees theory in two respects demanded by society. First, people’s desire of unanimous improvement and freedom of exchange introduces a part of classical liberalism and self-ownership, which may be small or, as a limiting case, occupy the entire place, according to the self-conception of the society which implements the policy. Second, one optimises the tax base (and not the tax schedule only) and takes advantage of the observed possibility of a largely non-wasteful transfer policy. Even in 1971, Rawls and Mirrlees agree on the following three basic points. (1) Earning capacities. Rawls and Mirrlees both hold that individuals have no right in the economic value of their own given earning capacities, contrary

16

See the volume Macrojustice and Chap. 3, Appendix B. In 1971, also, both the analysis of the principle “equity-no-envy”, and the leximin in comparable “fundamental” preferences to be applied, notably, when the worst-off suffers and can be relieved at an acceptable cost, are introduced in the book Justice and Equity (Kolm 1971) (discussed by Rawls 1982). 17

2

General Presentation

53

to classical liberalism.18 However, societies often hold that people have some more or less extended rights in the economic value of their own capacities, and since society implements the policy, other views cannot be implemented. ELIE just acknowledges this fact and unavoidably presents the corresponding structure that respects freedom to act and exchange and secures social (Pareto) efficiency: the value of earning capacities is equally shared for the equalisation labour (which may be zero) and self-owned for the rest. (2) Tastes. Both Rawls and Mirrlees (1971) hold that individual differences in tastes are irrelevant for this issue. Mirrlees states that they “raise different kinds of problems” than those relevant for the income tax, and Rawls that they are irrelevant for “social justice” (“macro” and “not micro”, hence macrojustice). This is general opinion (see Sect. 3.3.2) and, therefore, ELIE also endorses this view just to be implementable. Mirrlees translates this view by using the same utility function for all individuals (this also suppresses the differences in individual capacities to enjoy, which also corresponds to general opinion as shortly noted). Since individuals’ actual utility functions are different, this is not their actual functions (except, possibly, for one of them). Nevertheless, this is a sound intuition which can be justified and implemented by defining a relevant function (Sect. 3.3.1.3) (but individuals’ actual reactions to taxation derive in fact from their actual utilities, which also are the relevant ones for the classical definition of Pareto efficiency). Then, the chosen Social Welfare Functions are symmetrical concave functions of these identical concave utility functions depending on disposable incomes and labours, and the choice of these variables that maximise them turns out to amount to minimising an index of inequality in these two variables.19 (3) Information. Mirrlees (1971) says that individuals’ wage rates can be observed by observing earned income and labour duration and that “we also have other means of estimating a man’s skill-level”. The present French income tax policy amounts to this (by taxing income earned during the benchmark duration only). However, Mirrlees bases his developed model on the different hypothesis that the tax base should be full earned income, which induces disincentives. This model can be an explanation of Rawls’s belief about the importance of incentives and disincentives (which leads him to propose the “difference principle”, a maximin in “primary goods” including income). Hence, both Rawls and Mirrlees’s developed model, in 1971, 18

Mirrlees’s text ends with: “the great desirability of finding some effective method of offsetting the unmerited favours that some of us receive from our genes and family advantages”. 19 This is a generalisation of the classical Pigou’s (1912) proof that utilitarianism (with identical concave utilities) is income-egalitarian, to several dimensions and more general Social Welfare Functions.

54

S.-C. Kolm

appear to be too pessimistic about the possibilities of redistribution with minimal wasteful disincentives. Rawls and Mirrlees pursued their work, however. In 1974, Rawls accepted Richard Musgrave’s proposal to replace his “difference principle” by the introduction of a second economic primary good, leisure. Rawls’s ideal, therefore, is equal income and equal leisure or labour. This can also be seen as an ideal, a very first-best, of Mirrlees’s (1971) social maximand (see Sect. 3.3.1.4). This maximisation might thus provide second-best Rawlsianism if the first-best has some defect. Three defects can indeed be detected. First, there generally are solutions preferred by all individuals to equal income and labour. However, this cannot be remedied by the maximisation of the social welfare function with identical – and hence non-actual – individual utility functions, which presents the same type of defect (Pareto inefficiency). Second, imposing income and labour violates individuals’ freedom to exchange, of which working and earning is an application. Third, common opinion favours some smaller or larger part of classical liberal self-ownership. The solution to these three defects consists in letting people freely work and earn from the equal income and labour basis. Since this free exchange is a basic liberty, this solution applies Rawls’s most basic and prioritarian principle. This solution also happens to be the ELIE distributive structure. The equal labour and income basis constitute the equalisation labour and income. This equalisation can be realised by choosing the right policy base, as noted. The only specification is that Rawls thinks that the citizens dialoguing for choosing a distribution choose the distribution they are actually submitted to (a “well-ordered society”) only if the equalisation labour is rather high. However, Mirrlees, in 1986, intends to consider individuals’ actual utility functions, hence with their actual tastes. Deriving the overall distribution from individuals’ utility functions (including tastes and propensities to happiness) is, indeed, the first-best theory in the sense that it would be applied in the “best” society in which the individuals would be sufficiently “good” persons altruistic towards one another. However, actual, imperfect people are like this solely in small groups such as the family or when higher utilities mean lower suffering (as with allocations in hospitals). This is not the case for large societies in normal situations. For the general opinion, indeed, interpersonal differences neither in tastes nor in individuals’ capacities to enjoy are relevant for determining the income tax (see Sect. 3.3.2). Hence, a theory with an objective function based on these items could not be implemented and the insurmountable difficulty of knowing these specific actual individual utilities is just not raised at this point. Finally, people’s discarding differences in individual hedonic capacities and tastes, and hence in the utility functions that represent them, can be seen

2

General Presentation

55

either as discarding these differences only for retaining some comparable “welfare” – Mirrlees’s (1971) approach which can be rationalised –, or as discarding individual utility functions altogether, which is Rawls’s view that leads him to an ideal equality of the arguments of these functions, income and leisure or labour. As we have noted, the former approach can be seen as desiring the second conclusion (minimise the bidimensional inequality), and the result should be extended by the addition of some free exchange. However, if utility functions are seen as representing not final satisfaction but the motives for choice by their maximisation, discarding them leaves the domains of free choice as the directly valued items. The freedom of choice permitted by these domains is a commonly demanded value in itself. Then, the moral principle becomes the equality of these liberties. The various possible ways of defining these equal liberties without referring to the discarded individual preferences turn out to yield the ELIE principle again (Sect. 3.6).

2.6 The General ELIE Research Program: The Present Volume 2.6.1 General ELIE Even though ELIE can be seen as just the necessary synthesis of the main paradigms of macrojustice (and complement to the distributive ones), Erik Schokkaert (2009) points out that the practical differences it implies amount to a full-fledged paradigm shift. A paradigm shift defines a new research program. The simple ELIE outlined above is a basic core theory that has to be completed to constitute the general ELIE taking account of other important relevant facts. For a number of them, this is done in Macrojustice and in previous works. Labour is multidimensional (duration, education, intensity, possibly of various types for the same person), and hence the equalisation labour is multidimensional. Pay may be a non-linear function of labour. Unemployment – the concern of other policies – can also be introduced in the general scheme. The present French tax law applies the solutions for part-time labour (by reference to a part-time labour contract) and for family incomes. It has been computed that replacing all aids to low incomes by ELIE transfers would benefit everybody (in the case of France). The financing of other expenditures can be integrated with this “distribution branch”. Possible principles of allocations of non-human natural resources are analysed in The Liberal Social Contract (Kolm 1985, Chap. 10). The ethics and fiscal treatment of bequest have also

56

S.-C. Kolm

been studied (Kolm 1985, 2006). Very high incomes are mostly de facto non-labour or collusive (with board members). Each chapter of the present volume provides a crucial part of this building, and it is at least as important as – and more elaborate than – the initial ELIE by itself. This covers most issues except, regrettably, developments and applications of the methods for determining the degree of equalities in income and self-ownership desired by a society (analysed in Part 4 of Macrojustice).

2.6.2 Axiomatic Foundation François Maniquet’s contribution is particular by being “upstream” and foundational. His terse axiomatic derivation of the simple ELIE structure constitutes an alternative to the one presented in Macrojustice. This elaborate, imaginative and rigourous construction was in fact initially due to be published as a chapter of Macrojustice (this was not done in order to keep homogeneity in style of presentation). Maniquet is a complete co“inventor” of the ELIE distributive structure, in a particularly appreciated logical form.20

2.6.3 Minimum Income, The Sociology and Psychology of Transfers, Community and Dignity If the average productivity of the society and the equalisation labour are sufficient, the minimum income (k w) also is. If this is not the case, specific policies of support to low incomes can be added. Of course, contributions are provided by the various types of social insurance,21 the noted cases of unemployment and part-time contracts, and donations (important in some societies) which may have to be actually implemented by the public sector because the poor’s income is a non-excludable public good for the givers. Note also that, at the other end of the income scale, very high incomes often include a rent of situation which is not labour (for instance a “reciprocity”

20

See Sect. 3.6.4, Maniquet (1998) and Kolm (1993, 1996a,b). Including transfers implementing hypothetical “fundamental insurances” against relative handicaps that occurred in the past when this is largely approved (such as with public health insurance with premia unrelated to health status, or possible similar policies for the “brute bad luck” of a poor education or of low earning capacities) (Kolm 1985).

21

2

General Presentation

57

with board members for top executives), and this part of income requires a specific treatment beyond the ELIE scheme (an issue of microjustice for macroincomes). Alain and Justin Leroux propose to eradicate poverty by completing an ordinary ELIE distribution with a specific support to the poor provided in a form particularly valuable and important from a social and relational point of view and for implementation efficiency. In an additional, specific, ELIE scheme (a “poverty ELIE”), the equalisation incomes would be voluntarily redistributed to the poor in the framework of voluntary associations of people having some common characteristic (each individual has to join one association). The method, tightly and convincingly analysed, replaces public assistance by humane or fraternal local solidarities, and transmutes a part of egalitarian fairness into collective altruistic help. This would improve not only the poor’s welfare but also the overall quality of social relations both with the poor and between people who help them – an essential aspect of the quality of a society. The elaborate analysis of social relations that may be induced by ELIE transfers is precisely the topic of Pierre Livet’s chapter. Yielding or receiving transfers induces opposite emotions due to social recognition, judgments and status. However, people with lower capacities may be credited for creating diversity favourable to society. The detailed analysis of these aspects reveals and enlightens an essential dimension of social transfers. It should help us make sure that transfers do not jeopardise dignity. This should be the case with the principle of ELIE which is the manifestation of a solidaristic community. Moreover, social emotions and judgments may induce people to pay their due, to help others when they have to – which happens when they have the luck of being endowed with high productive capacities –, and not to cheat by pretending to be less able than they are. People may be proud both to help others and to display their high abilities even though they are not responsible for them (both Voltaire and Mirrlees independently propose jokingly to tax intelligence, which people do not want to hide – I.Q. for the second).22 They may be particularly ashamed to exploit others and be social parasites living at their expense when they are able to earn their own living and – perhaps – to help less fortunate others. Therefore, Pierre Livet’s considerations may also be particularly important for the actual implementation of distributive principles.

22

See Voltaire, The Man with Forty Ecus (Philosophical Tales). The king to whom this advice is given answers the adviser that he would have to be exempted from this tax.

58

S.-C. Kolm

2.6.4 Education and Growth Education, formation and training that increase qualification, productivity and earnings constitute essential dimensions of labour, and hence of general ELIE. This includes education, formation and training both before and during one’s working life. The treatment of this issue is also particularly delicate. For instance, one should distinguish, in educational achievements, what is due to the person’s responsibility or effort and what is due to family help and influence, which should be considered along with family gifts and bequest (and includes motivation). Capacities to learn also intervene. Then, education can be treated strictly according to ELIE by redistributing equally individuals’ earnings due to education up to a certain level (the “equalisation education”). It can also be treated in other ways such as “spreading” education duration as an increase of the duration of the labours that use the increased productivity (Macrojustice), or by free public education (or other subsidy) compensating taxes on labour productivity boosted by the increased qualification. The issue is inter-temporal, and the analytical breakthrough is provided by the model of overlapping generations proposed by David de la Croix and Michel Lubrano. An individual chooses a level of education, works with a low productivity the rest of her youth (which shows the opportunity cost of education in foregone earnings), and then works as a mature worker with a productivity enhanced by education. This is a splendid precise example of tridimensional and inter-temporal labour, which permits enlightening studies of various effects and policies. The authors use it in particular for studying the trade-off between growth and redistribution when taxation is not based on given capacities only, and for proposing the noted education subsidy.

2.6.5 Low Labours Since general opinion refuses taxing or subsidising unused capacities, ELIE can apply only to individuals working no less than the equalisation labour, as we have noted. As we have seen, extremely few people declare presently to work less than 20 hours a week, which is about an upper bound for equalisation durations realistically possible in present-day national societies. On logical grounds, an easy theorem shows that, if an ELIE distributive scheme were applied to everybody and to these capacities actually used by people: (1) individuals who choose to work more than the equalisation labour have an interest to work with their most highly paid capacities; (2) on the contrary, individuals who choose to work less than the equalisation labour have an

2

General Presentation

59

interest to work at occupations that earn as little as possible (Sect. 3.8.2). The latter individuals, if the effect is complete, do not work at all and receive, as a subsidy, the equalisation income computed with a zero productivity (wage rate) for idle people. These are issues of microjustice, with many possible solutions and out of the present concern – yet one may nevertheless say something about them.23 Unemployed and handicapped people would be so subsidised, as people with very low earning capacities whatever their work (and part-time labour contracts and second jobs in family have been brought back to the equivalent full-time as indicated). For the very few other people who choose not to have a higher income, several points are in order. Rawls emphasises that people who choose to work very little are not members of the social process of cooperative production and therefore not party to its distributive scheme in “social justice” (i.e. macrojustice). These fugitives from cooperative production drop out from its distribution. They voluntarily withdraw from the general reciprocity of labour. Hence, neither would they be entitled to a subsidy from others’ labour, nor would the most able of them, who do not take advantage of the high pay they could have, incur a liability for this absent benefit. However, as shortly noted, the social waste is limited and these individuals may simply receive this universal social dividend or basic income. This may in particular permit them to do voluntary or associative work, or innovative art, and participate to valuable non-market production, activities and relationships. However, earning power may also be estimated from past activity, education and health. The better endowed may then simply be asked to take care of their own living – the solution of Rawls and Saint Paul (“he who does not work does not eat”). They may even be asked some contribution (certainly so when their work can save lives). Moreover, activity using one’s best capacities is a main dimension of a satisfactory life (Aristotle says of happiness), a source of pride, dignity and fulfilment, especially when it also serves others and provides social integration, relation and status. For these reasons, unemployed people are often despaired even when they benefit from a substantial unemployment compensation. Furthermore, able idle people may be ashamed to live on others’ labour, discontent with the attached blame, stigmata and lower status as Pierre Livet explains, reluctant to violate the very common folk-Kantian ethics (“what if nobody worked?”), and aware that ELIE schemes apply basic properties unanimously wanted to an extent (k) that corresponds to a deep collective desire in a democratic “well-ordered” society (some people do not cheat on taxes because of a civic culture and

23

Scholars happen to be more excited by weird issues of microjustice than by the simple fate of the toiling masses.

60

S.-C. Kolm

duty, and they are numerous in some countries – among which some with the highest rate of tax and transfers). Actually, very numerous empirical and experimental studies of the supply of labour in reaction to pay, in social psychology and economics, from those of “fairness” by Adam and others in the 1960s to those of reciprocity in labour by Fehr and others recently, show that people freely chose to work for an amount they consider “fair”, from a motive of “balance reciprocity”, in reaction to a pay they receive, rather than trying to maximise their “material self-interest” in this respect.24 Therefore, economists’ cherished anthropology of suspicion assuming that everybody is exclusively materially self-interested and tries to exploit others as much as she can is not necessarily warranted, to say the least. Assume it is, however, just to be on the safe side. Then, if an ELIE scheme is applied universally to everybody, and if some people choose to work less than the equalisation labour whereas their case is not brought back to the general case (very low wages, unemployment, part-time labour contracts) and their productivity has to be inferred solely from their statement or their actual activity, this implies some social waste. Nevertheless, a remarkable empirical analysis of Erwin Ooghe and Erik Schokkaert shows that this waste does not exceed that induced by ordinary tax schemes that take whole earned income as base. Since the working people also receive, as their equalisation income, the same equal amount as that received by idle ones (they also finance it by the same labour or according to their capacities), the result can be the implementation of a universal basic income advocated by many scholars such as Brian Barry, Yoland Besson, and Philippe van Parijs whose proposal is analysed in depth by Claude Gamel in this volume. This financing, based on actual workers’ given capacities, is less wasteful than that based on their full wage income proposed by van Parijs (although not in the situations of involuntary unemployment emphasised by this author). This Ooghe–Schokkaert financing of a universal income thus consists simply in extending an ELIE-type redistribution to everybody by taking as individual productivity that of the actual occupations chosen by the individuals. However, there remains some waste and the working people finance this equal universal allocation by the same equalisation labour or according to their capacities. One may therefore be interested in other solutions for financing a universal income, at least for non-working people. The only assets apart from human capacities are the non-human resources. These resources may be seen intertemporally, and then they reduce to non-human natural resources. They may also be seen at a given time, and then they also include all existing produced capital which, at this time, constitutes the bulk of non-human

24

An analysis of this motive is provided in Kolm (1984, 2008b).

2

General Presentation

61

resources and considerably extends this base. Both views have their rationale and tradition, as pointed out shortly, but the second may compensate the disincentive effect and induced inefficiency it entails in capital formation by being more practical and, at a given time, providing a much larger base. Claude Gamel advocates this second view, and, in an original analysis, proposes to tap these non-human resources in a way analogous to that in which ELIE schemes tap the human resource. Besides the financial result, this provides homogeneity to society by the universality of both the basic income and the uniform intent of its financing. Moreover, Sect. 12.3 shows that various rationales lead to add an ELIE distribution to an equal sharing of the distributable non-human resources whatever they are; this provides more compensation for poor earning power with more self-ownership.

2.6.6 Capital, Uncertainty, Family Size Claude Gamel’s discussion has the additional virtue of raising the important question of the status of capital and time. A traditional conception objects to taxing capital because the savings that created it were a free choice, the incomes from which they came have already been taxed and this would be double taxation, and this tax induces wasteful intertemporal disincentives. This intertemporal view would advocate restricting the meta-ELIE redistribution to the allocation of the non-human natural resources, according to any of the principles analysed in The Liberal Social Contract (Kolm 1985), for instance to help the most unlucky, miserable or poorest people, or those with the lowest earning capacities wi , or equally – the solution retained by the famous “left classical liberals” such as Hillel Steiner (1994) and Peter Vallentyne (Steiner and Vallentyne 2000a,b) – plus the redress of past torts and gifts possibly publicly implemented. However, first, this may not suffice. Second, many nations actually tax earnings from capital (and a few capital itself). Third, the problem may be to introduce a new distributive rule, and, therefore, the ownership accumulated under an ancient rule may not have to be respected. Therefore, taxing present-day ownership may be justified; it is actually performed and it is advocated in general by Ronald Dworkin (1981) and, with elaborate theories, by Maurice Allais (1989) and, very differently, by Claude Gamel in this volume. The taxation of capital along with ELIE is also studied by Michel Lubrano, with an explicit growth model, in an essential contribution which analyses the introduction of ELIE in an economy which has to keep a number of present-day policies and faces other problems. These other distributive policies may be necessary for several reasons: provisionally in a transition

62

S.-C. Kolm

period, because the chosen labour duration and society’s productivity do not suffice to satisfy all distributive needs, or because particular incentives or issues of fairness are sought. Beyond the tax on capital income, the basic question of adjusting ELIE for family size and the role of family allowances are analysed. The issue of the imperfect information about individuals’ earning capacities is also investigated. The concepts, proposals and results of this study constitute the threshold towards the actual application of ELIE policies, a concern without which all the rest is valueless.

2.6.7 Information and Second-Best Implementation of the ELIE Ethics by Taxes on Total Earnings: A Marriage of Paradigms An ELIE distribution induces individuals to work with their best capacities. The tax authorities may observe wage rates on pay sheets (official documents). They may deduce them from the observation of some corresponding earnings and labour. This may be, for instance, full earnings and labour or, as with the exemption of overtime labour earnings, the official duration of labour (e.g. in hours per week or in days per year for executives whose daily hours of work are unclear). They may estimate the person’s productivity by comparison with similar jobs. The authorities’ information about full earnings is sometimes good and sometimes very bad. However, this is the common tax base, often without other information (except family status). If this is followed, and in so far as the taxpayers are not motivated by civic, ethical or social considerations for implementing a distribution with clear properties of fairness and efficiency and collectively chosen, the problem faced is to implement a second-best ELIE with this tool and facing exclusively materially self-interested taxpayers. To solve this problem Laurent Simula’s and Alain Trannoy’s deep and elaborate study proposes two basic concepts and derive examples of solutions. In one case, they maximise a social welfare function which gives an ELIE policy as first best. In the other, they consider a second-best which realises ELIE transfers. The results depend on the specifications of the utility and social functions in the first case, and the result is unique in the second case. This constitutes a great example of the richness and flexibility of the classical optimum taxation studies, capable of providing very important analyses of social and ethical conditions different from those of their initial intuition. Can we conclude, from this synthesis of paradigms, that normative economics is more advanced than theoretical physics? On related issues, see also Simula and Trannoy (2011) and Fleurbaey and Maniquet (2011a,b) in the volume Social Ethics and Normative Economics, Fleurbaey et al. (2011).

2

General Presentation

63

2.6.8 Macrojustice in Normative Economics and Social Ethics Finally, Chap. 12 raises, broadens and deepens the analysis by briefly situating the obtained results in the full spectrum of the distributive principles of economics and social ethics. This is indispensable for four reasons. First, a principle is justified only if it is compared to all other possible alternative ones and found superior to each; and other principles have actually been proposed. Second, other principles may coexist with it, for instance in questions of microjustice or mesojustice, and the relations between them have to be considered. Third, this enlarged consideration may in particular be necessary for specific issues of application of the principle, for choosing properly justified solutions rather than ad hoc ones. Fourth, this broadened vision explains better, by comparison, the place and structure of the result obtained. In particular, one such deepening concerns the distinction and different treatment of various aspects of “utility”, a surgery also practiced by some other principles. And one enlargement is the integrated treatment of human and non-human resources. The principles of distributive justice are classified and appraised according to various criteria. The criteria of evaluation include the internal logic (some principles do not pass this test); the consistency with other properties (for instance a number of principles consist in equalities that happen to jeopardise Pareto efficiency); the possible meanings; and the relation with the general actual people’s judgment, which is crucial for the social possibility of implementation. The principles are characterised by their end-values, the items they directly evaluate. A class of them is basically concerned with psychological properties which can often be described as some structures of individuals’ utilities. This includes overall “welfarism”, “pure welfarism” discarding differences in individuals’ tastes and hedonic capacities, “ordinal welfarism” considering preferences orderings only and including principles in the families of “equity-no-envy” and of the “equivalence principle”, and co-ordinal welfarism that relates such preferences of various individuals. Other principles focus on freedom of some sort, such as “freedom from” and basic rights, domains of choice, and opportunities in various respects. Various goods provide both freedom as means and satisfaction from use, including income, the satisfaction of needs, human capital from education or health, or other “capabilities”. Indeed, an alternative to macrojustice is to reduce all justice to micro or macro issues. This is not actually possible in a modern society, however. Michael Walzer (1983) sees justice as equality in “spheres” for each main good or service. But one such sphere is much larger in volume than others, that related to income even if the market is not maximally extended. Other

64

S.-C. Kolm

theories also demand equality in various types of goods (e.g. opportunities, capabilities, etc.). However, these equalities violate Pareto efficiency because they interfere with people’s actions or because they are multidimensional, or both. In the former case, the solution is of the ELIE type (equal liberty provided by different domains of choice). In the latter case, the corresponding efficient second-best egalitarian solution leads to equality in income, hence to an issue of macrojustice. In the end, this comparison with a priori possible alternatives guarantees the necessity of the obtained distributive structure. * ** Derived from principles generally demanded, and hence socially implementable, this distributive structure is both philosophy and practice. Philosophically, it synthesises the three polar paradigms of macro social ethics. Classical liberalism, with its virtues of basic liberty and efficiency (possibly with public corrections) – and which may take the entire place in societies that want it – operates on the top of an equality in income and leisure (or labour) which is the ideal of Rawls and can be seen as that of the standard “welfarist” optimum income tax (Mirrlees 1971). Practically, the result can simply be an income tax with two universal bonuses: an exemption of overtime labour earnings over a low benchmark, and a uniform tax credit – the linearity is not always the case with general ELIE and it is not essential. The intensity of the redistribution, however, can be quite varied. The core simple ELIE is completed to become the general ELIE taking account of the above noted phenomena. The present volume provides such analyses. It shows essential aspects of this distributive philosophy. It investigates its relations with other distributive forms, whether they are complements, supportive methods, alternatives or interfering structures. In this process, this volume relates these issues to basic aspects of the economy such as efficiency, growth and employment, and of society such as formal and real freedom, social relations and consideration, and distributive and procedural justice. This result is finally situated in the field of the rather numerous economic and ethical theories that ambition to propose a solution for the overall allocation, with the logical and moral appraisal of these proposals.

References Allais, M. (1989). L’Impôt sur le capital. Paris: Hermann. Bar-Hillel, M., & Yaari, M. (1984). On dividing justly. Social Choice and Welfare, 1(1), 1–24.

2

General Presentation

65

Bourguignon, F., & Spadaro, A. (2008). Tax-benefit revealed social preferences. Working paper 2008-37, Paris School of Economics, Paris. Dworkin, R. (1981). What is equality? Part I: Equality of welfare; Part II: Equality of resources. Philosophy and Public Affairs, 10, 185–246, 283– 345. Elster, J. (Ed.). (1986). The multiple self. Cambridge: Cambridge University Press. Fleurbaey, M., & Maniquet, F. (2011a). ELIE and incentives. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer, (forthcoming). Fleurbaey, M., & Maniquet, F. (2011b). Kolm’s tax, tax credit and the flat tax. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer, (forthcoming). Fleurbaey, M., Salles, M., & Weymark, J. (Eds.). (2011). Social ethics and normative economics. Heidelberg: Springer. (forthcoming). Kleven, H., Kreiner, C., & Saez, E. (2009). Why can modern governments tax so much? an agency model of firms as fiscal intermediaries. Working paper 15218, NBER, Cambridge, MA. Kolm, S.-C. (1970a). L’Etat et le système des prix. Paris: CNRS-Dunod. Kolm, S.-C. (1970b). Prix publics optimaux. Paris: CNRS-Dunod. Kolm, S.-C. (1971). Justice et équité. Paris: Cepremap. Reprint, Paris: CNRS, 1972. English translation, 1997, Justice and equity. Cambridge MA: MIT Press. Kolm, S.-C. (1974). Sur les conséquences économiques des principes de justice et de justice pratique. Revue d’Economie Politique, 84(1), 80–107. Kolm, S.-C. (1982). Le Bonheur-liberté (Bouddhisme profond et modernité). Paris, Presses Universitaires de France. Kolm, S.-C. (1984a). La bonne économie: La réciprocité générale. Paris: Presses Universitaires de France. Kolm, S.-C. (1984b). Le libéralisme moderne. Paris: Presses Universitaires de France. Kolm, S.-C. (1985). Le contrat social libéral. Paris: Presses Universitaires de France. Kolm, S.-C. (1986). L’allocation des ressources naturelles et le libéralisme. Revue Economique, 37, 207–241. Kolm, S.-C. (1993). Efficient economic justice. Paris: CGPC. Kolm, S.-C. (1996a). Modern theories of justice. Cambridge MA: MIT Press. Kolm, S.-C. (1996b). The theory of justice. Social Choice and Welfare, 13, 151–182. Kolm, S.-C. (2001). On health and justice. In D. Wikler (Ed.), Global health: From goodness to fairness. Geneva: World Health Organisation.

66

S.-C. Kolm

Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge: Cambridge University Press. Kolm, S.-C. (2006). (a) Introduction. (b) Reciprocity: Its scope, rationales, and consequences. In S.-C. Kolm, & Jean Mercier Ythier (Eds.), Handbook of the economics of giving, altruism, and reciprocity (pp. 1–122 and 371–541). Amsterdam: North Holland. Kolm, S.-C. (2008a). The paradoxes of the war on poverty: Warm-glows and efficiency. Working Paper 2008-7, IDEP, Marseille. Kolm, S.-C. (2008b). Reciprocity, an economics of social relations. Cambridge: Cambridge University Press. Kolm, S.-C. (2008c). The theory of rules and the moral provision of public goods. Mimeo, Paris: EHESS. Kolm, S.-C. (2009a). Justice. In J. Peil, & I. van Staveren (Eds.), Economics and Ethics (pp. 291–300). London: Edward Elgar. Kolm, S.-C. (2009b). A response to E. Schokkaert on Macrojustice. Economics and Philosophy, 25(1), 85–98. Kolm, S.-C. (2009c). Social ethics and rationality, new directions for the opimum production of social justice: Meaningful welfare, equal liberties, social solidarity. In Conference on Inequality, New Directions, Ithaca, N.Y. Cornell University. Forthcoming in Journal of Economic Inequality (2011). Kolm, S.-C. (2010). Equality. In B. Badie (Ed.), International encyclopedia of political science. London: Sage Publications, Inc. Maniquet, F. (1998). An equal right solution to the compensationresponsibility dilemma. Mathematical Social Sciences, 35, 185–202. Mirrlees, J. (1971). An exploration in the theory of optimum income taxation. Review of Economic Studies, 38, 175–208. Mirrlees, J. (1986). The theory of optimal taxation. In K. Arrow, & M. Intriligator (Eds.), Handbook of mathematical economics (Vol.3). Amsterdam: North-Holland. Musgrave, R. (1974). Maximin, uncertainty, and the leisure trade-off. Quarterly Journal of Economics, 88(4), 625–632. Pigou, A. C. (1912). Wealth and Welfare. Macmillan, London. Rawls, J. (1974). Reply to Alexander and Musgrave. Quarterly Journal of Economics, 88, 633–655. Rawls, J. (1982). Social unity and primary goods. In A. Sen, & B. Williams (Eds.), Utilitarianism and beyond (pp. 159–185). Cambridge: Cambridge University Press. Rawls, J. (1999,1971). A theory of justice (revised edition) Cambridge, MA: Harvard University Press . Rothbard, M. (1973). For a new liberty. New York: Macmillan.

2

General Presentation

67

Schokkaert, E. (2009). Macrojustice as a research program. Economics and Philosophy, 25(1), 69–84. Simula, L., & Trannoy, A. (2011). When Kolm meets Mirrlees: ELIE. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer. (forthcoming). Steiner, H. (1994). An essay on rights. Oxford: Blackwell. Steiner, H., & Vallentyne, P. (Eds.). (2000a). Left-libertarianism and its critics: The contemporary debate. London: Palgrave Publishers. Steiner, H., & Vallentyne, P. (Eds.). (2000b). The origins of leftlibertarianism: An anthology of historical writings. London: Palgrave Publishers. Van Parijs, P. (1995). Real freedom for all. Oxford: Oxford University Press. Voltaire, F. A. (1768, 1938). L’homme aux quarante écus. In Contes philosophiques. Editions de Cluny, Paris. Walzer, M. (1983). Spheres of justice. Oxford: Blackwell.

•

Chapter 3

Economic Macrojustice: Fair Optimum Income Distribution, Taxation and Transfers Serge-Christophe Kolm

Abstract The basic, core theory of overall distributive justice in macrojustice is presented in this chapter. The basic facts are the following. (1) General opinion rejects differences in tastes and hedonic capacities as relevant for macrojustice in a society in a normal situation. (2) Experience shows the possibility of transfers based on given capacities with practically no disincentive effect (exemption of overtime labour earnings from the income tax). (3) Pareto efficiency is desired and a condition of stability. (4) Social liberty from given resources is desired and necessary. (5) Equal real liberty (for different domains of choice) is a priori desired and rational. The result is a simple distributive scheme rich of some twenty meaningful equivalent properties, including free exchange and labour from a given equal-labour income equalisation; general balanced labour reciprocity; basic income financed by an equal labour of each (or according to capacity); a “concentration” of total income; etc. The issues of the determination of the degree of redistribution and equalisation, and the relations with the rest of public finance are briefly recalled.

3.1 Justice, Liberty, Equality, Welfare and Information 3.1.1 Moral Principles of the Just Distribution Freedom, desert, need and luck determine opinions about the legitimacy and justice of income. My earnings are freely handed out to me by people who buy the services or the products of my labour performed with my capacities.

S.-C. Kolm EHESS, Paris and CREM e-mail: [email protected]

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_3, 

69

70

S.-C. Kolm

This voluntariness may make this income legitimately my own. Moreover, I may deserve the fruits of my effort. Underpaying labour is what Marx calls exploitation.1 However, I cannot be said to deserve – because I did not produce them – the other causes of these earnings: neither my capacities given by nature or my family, nor other people’s tastes, needs and capacities that determine their willingness to pay for my services or products. What do I deserve, therefore? People who provide the same effort or labour – or loss of leisure – (effort of formation included) may deserve the same income only, irrespective of their given talents and of the demand for them. However, people also commonly find some smaller or larger part of legitimacy in luck such as others’ demand, gifts of nature or family environment. This bases views about gifts, bequest, rights of families or “natural rights” (not to speak of god’s will) – fiscal laws about gifts, bequest or treasure discoveries reveal these choices. Needs, finally, become the main reason for augmenting very low incomes that are not sufficiently remedied by sharing for another reason. Note that possible reasons to diminish income inequality have to be different when the praised effect is to relieve poverty than when it is to make satisfactory incomes closer to one another only. Both people’s views and actual policies usually consider all these principles. Scholars are usually much less subtle and relish in dogmatism and ideology. Their theories may be morally right, but these theories are commonly blind to aspects that may be relevant and important, disrespectful of people’s thinking and experience, and, as a result, unavoidably sterile because unapplied by policies chosen by actual societies’ voters or officials.

3.1.2 Earning Capacities and Information For the overall income distribution in macrojustice, in particular, we have seen in Chap. 2 that both the “welfarist optimum income taxation” such as Mirrlees’s and Rawls’s “justice as fairness”, of 1971, hold that people have no a priori legitimate right in their productive and earning capacities. They are both fully egalitarian in this respect. However, they nevertheless advocate disposable incomes that increase strongly with these capacities. The reason is to induce people to work more or not to work less as a result of this tax. And this results from the fact that both consider an income tax based on earned income wi `i where `i is labour and wi the wage rate, for individual i. However, the present tax law of France shows that nothing is easier than to base the income tax on the wage rate wi , rather than also on labour `i , by

1

Marx does not distinguish in general various qualities of labour.

3

Economic Macrojustice

71

exempting the earnings of overtime labour over a low benchmark, for wage incomes which make up 9/10 of labour income, with actually no cheating (which could not escape the tax authorities’ notice and sanctions). Hence both noted theories are unfortunately mistaken: they optimise the tax schedule but forget to optimise the tax base and rest on an erroneous assumption in this respect. Moreover, a basic theorem shows that, with such a tax or subsidy, people are induced to work with, and hence to reveal, their most productive and remunerated capacities. This holds whatever these capacities, the income level and the fact that the transfer is a tax or a subsidy (Sect. 3.8). (Even if we wanted to base the tax on the actual wage rate because, if it is lower than the maximal, taxing the maximal taxes the discrepancy which represents a kind of leisure, in the end the chosen actual wage rate turns out to be the maximal one).

3.1.3 Cooperative Liberty and Moral Autonomy Let us consider, therefore, the third polar theory opposed to the noted two of solely maximising “social welfare” and of an ideal of equal incomes (the third summit of the “justice triangle” presented in Chap. 12).2 This is full self-ownership advocated by “classical liberalism”, which rejects all compulsory distributive transfers. Gifts, however, are possible. But desires that some other people have a higher income for a reason or a sentiment of justice, compassion or general altruism are generally shared by many people, whatever their intensity or weakness. If they give rise to gifts, this is joint giving. The receiver’s income is actually a public good (joint concern) for all people who care about it. Joint giving permits the beneficiary to receive something with a much smaller contribution from each giver. This does not happen if the gifts are non-cooperative, however. Then, indeed, there is at most one giver who crowds out all the others, and the result is wasteful (not Pareto efficient).3 Hence, this freedom should take the form of a free cooperative unanimous agreement between the givers about their gifts. But the agreement has a priori to be enforced by a public authority, as with the implementation of standard free contracts. In this case, specifically, this constraint avoids free riding (for this reason, Milton Friedman notices that these transfers should be enforced by the government). Moreover, the people concerned

2 3

See also Kolm (2009a). This holds whatever the type of non-cooperative behaviour (Kolm 1986a).

72

S.-C. Kolm

whatsoever about the low income are a priori much too numerous to be able to agree actually and directly between themselves. Hence this agreement has to be hypothetical, putative, and imagined by the public authority. That is, it is a “social contract” in the technical sense of the term. But, then, the corresponding freedom also is hypothetical and vicarious. However, people also freely follow a few ethical conducts of cooperation. They sometimes give given that the others contribute (not as a result of an exchange but as a conduct of “lateral reciprocity”).4 However, they may be sure that the others contribute if these others are forced to only. Then, there should again be (forced) taxes although this constraint is not actually binding. A still more moral cooperative conduct is the relatively common one that Kant theorises as the “categorical imperative” where “categorical” means unconditional: act as if others also acted correctly. This is a free act without external constraint. However, both with lateral reciprocity and with this imperative, Pareto efficiency requires specific relations between the contributions, and this may again require some kind of public intervention.5 Moreover, this reaches Kant’s (and Rousseau’s) definition of freedom, which is particular, although interesting: to be free is to dutifully obey the command of reason (here the imperative) against the desires of preferences describing both self-interest and altruistic sentiments (Kant’s “inclinations”). This puts forward the multiple rationales of a person in society: a self-interested individual, possibly augmented by altruistic sentiments (affection, liking, compassion, etc.), and an impartial self capable of objective judgments and of following the demands of rational duty.6 Then, answering people’s demand for helping reason to determine and point out the right actions and to control impulses, tends to need again some political dialogue and obligation, in the framework of democracy and in the end of unanimous consent. The ethics of liberty cannot avoid the subtleties of both the economics of cooperation and moral psychology.

3.1.4 Utility and Welfare The second type of human capacities considered in normative economic studies is a set of psychological or physiological characteristics which lead

4

See Kolm (2008b), Chap. 12. See Kolm (2008a, 2009c). 6 That is, implementing the impartial judgments described by Hume and Smith by the motivational autonomy put forward by Kant and Rousseau. 5

3

Economic Macrojustice

73

individuals to appraise their own situations or allocations, commonly summarised in the magnificently convenient device of the utility function. For distributive justice, however, people concur in holding that the various possible aspects or meanings of “utility” have very different scopes of relevance. According to the case, some are irrelevant and some are not. In particular, people’s propensities to suffer are very relevant for allocating means of relieving pain. Differences in individuals’ tastes and capacities to enjoy are commonly used with general consent when allocating between people knowing one another and feeling a strong empathy for the others (within families, for instance). But people do not think that someone should pay a higher income tax than someone else because the other enjoys more the euros taken away – and one is utilitarian – or less each euro left – and one is an egalitarian in utility (see Sect. 3.2.2). People do not think that I should finance the expensive tastes of someone else who can drink grandcru wines only, or her cheap tastes because she relishes cheap beer which permits to produce much utility at low cost. However, people also sometimes think that rich people should pay more because it reduces their welfare less. That is, they assume a concave “welfare function” comparable and identical for everybody. We shall see that it is indeed possible to define such a function (but it can describe neither individual behaviour nor Pareto efficiency). This identical concave individual welfare function favours equality in income and labour. However, such a solution is not Pareto efficient, almost everybody prefers to depart from it by working more and keeping the extra earnings, and this is an aspect of valued social liberty. This freedom from this equality solves these three problems (Sect. 3.3.1). This solution also happens to satisfy the possible definitions of equal total liberty that respect Pareto efficiency and social liberty (this is equal liberty provided by different domains of choice – Sect. 3.6).

3.1.5 Summary This chapter is a summarised but self-contained presentation of the theory of macrojustice presented in the volume with this title (Kolm 2005). For further developments of the hypotheses, analyses, consequences and relations with other distributive theories, and for the analysis of still other distributive principles and a general presentation of normative economics, the reader is referred to this volume. In particular, the methods of the determination of the degree of equalisation, redistribution or solidarity explicitly or implicitly desired by a society constitute Part 4 of Macrojustice. The core presented here is completed by specific analyses of related or interfering phenomena in the initial volume, in other works, and in all the other chapters of the present volume.

74

S.-C. Kolm

A summarised presentation of the result and its main meanings and a few basic questions make up Sect. 3.2. The next two sections provide the two bases of the optimum policy: the determination of the necessary principle of macrojustice (Sect. 3.3) and of the base of the transfers (Sect. 3.4). The conceptual and factual bases of the theory, concerned with liberties, resources and capacities, are laid down in Sect. 3.5. Section 3.6 shows how the result can be seen as deriving from a general principle of equality in liberties. The twenty or so meaningful equivalent properties of the simple result are gathered in Sect. 3.7. Section 3.8 presents the issues of incentive compatibility and of the exclusion of earnings that are only potential from the redistribution. Remarks about the determination of the degree of redistribution and the place of this distributive policy in overall public finance constitute Sect. 3.9. Section 3.10 concludes with remarks about implementation and convergence with other approaches. Finally, the theory with multidimensional labour and involuntary unemployment is presented in two appendices.

3.2 Overview and Basic Properties 3.2.1 Properties The distributive structure in question can be justified by any of its equivalent meaningful properties (Sect. 3.7). It is free exchange (of labour) from an equal allocation (of both income or goods and leisure or labour). This property in itself can refer to two possible basic social ethical principles. First, it is in itself a possible definition of equal liberty – free exchange is an application of the classical basic social liberty (civic or negative freedom). The other possible rational definitions of equal liberties – both social liberty and the real liberty added by the available means – happen to give the same result, notably the equal liberty provided by different (non-identical) domains of choice (Sect. 3.6). Second, this initial equality in income and labour (leisure) can be valued in itself or as a possible welfarist ideal (Sect. 3.3.1.3), and the extra free exchange is added for liberty, unanimity and efficiency. This result is yi D k w C .`i  k/wi D wi C ti ; where ti D k  .w  wi / is, if k > 0, a subsidy if wi < w and a tax of ti if wi > w. One has P ti D 0, the required balance for the “distribution branch”. The income

3

Economic Macrojustice

75

k w is that of an equal labour k paid at an equal wage rate (the average w). The income .`i k/wi is that of labour `i k paid at the individual wage rate wi . The former is according to deserts (labour, effort) and the latter according to merit (labour and capacities). The ELIE distribution itself – from each according to her capacities, or by an equal labour, to each equally – can also be seen in various ways. It amounts to each participant receiving the same basic income (k w) financed by an equal labour (k) of each. It also amounts to each yielding to each other the product of the same labour (k=n), in a general equal labour reciprocity. The amount k w is also a minimum income (since k  `i for the participants).

3.2.2 Pareto Efficiency The theory and the actual possibilities will turn out to imply that the outcome is close to Pareto efficient (with perfect markets, possibly thanks to the correction of imperfections which is another policy). Pareto efficiency is valuable for its classical property of non-waste. However, this absence of possible states preferred by all people (with the possible indifference of some of them) is also a condition of liberty, democracy and stability of the society. The liberty refers to the extra constraint that creates this situation. Moreover, when some possible states are unanimously preferred to the present one, a number of social processes induce leaving the latter. For instance, short of Pareto efficiency, a contending party can propose a program that wins with the unanimity of votes. Indeed, if, and insofar as, Coase (1960) is right to think that Pareto efficiency is always achieved if all actual constraints are taken into account (in particular transaction costs), then a model that does not have Pareto efficiency could not represent reality, and this may be due to its assumption of a policy inducing Pareto inefficiency. Pareto efficiency can a priori be achieved in two ways. One is from social interaction, in particular by the markets for which Pareto introduced this concept (public policy may have to correct their various possible defects in this respect). The other way is by a policy maximising a classical social welfare function, increasing function of individuals’ utilities. This second method can be the source of exciting mathematical exercises, but its actual application is highly problematic for four reasons. (1) The policymaker should sufficiently know the utility functions of all the individuals. (2) It should know the social welfare function that should aggregate them, whereas society never reveals its ethical distributive choices in this form (Rawls 1971, 1982, verified econometrically by Bourguignon and Spadaro 2008) and it

76

S.-C. Kolm

thinks that they cannot be expressed in this way for macrojustice (see below). (3) The policymaker should use all the available information about inelastic or less elastic variables to choose them as policy base (e.g. given productive capacities or wage rates). (4) At any rate, we will shortly see that this second method is not possible for macrojustice (and in particular the income tax) because general opinion rejects the relevance of some important features of utility functions for this question, notably the differences in tastes and in hedonic capacities (the remaining individual “welfare” function that can be defined is not individuals’ actual utility functions which correspond to their own evaluations, choices and actions – see Sect. 3.3.1.3).

3.2.3 Equality When some item proper to each individual is the direct focus of the ethical evaluation and no (other) differences in the individuals’ characteristics are deemed relevant for this judgment, then, as shortly explained, rationality demands prima facie equality (equal treatment of equals). This qualification “prima facie”, i.e in the absence of an overpowering reason, refers here, for instance, to impossibility for any reason or to the joint relevance of some other, interfering value. Rationality intervenes here in its weakest form and most common meaning of providing a reason, a justification, whatever it is, for this allocation, or even simply intending to. Indeed, a reason for justifying a person’s allocation refers to a number of facts relative to this person. If another person has an identical set of these relevant facts, the same reason indicates that she should have an equal allocation, prima facie. This equal treatment of equals in the relevant characteristics is a necessary condition for the possibility to provide a reason.7 However, the objective may also a priori be an aggregate of these individual items, for instance a sum. This is first of all a social viewpoint rather than one emphasising individuals as end-values. However, individual equality also intervenes in a sum by the equal weights (as forcefully emphasised by classical utilitarians), and hence for comparing variations of the items in question. The crucial distinction between the ethical views concerns the nature of the relevant items which can be, notably, for each individual i, bundles of goods xi , incomes yi and leisure i or labour `i , utilities ui or some other concepts of welfare, and freedoms for social or total liberties.

7

For a more complete discussion of the logical reason for equality, see Kolm (1971, 1996a, 2005, 2010a).

3

Economic Macrojustice

77

3.2.4 Who Owns the Economic Value of Given Earning Capacities? Scholars’ answers to this question are usually extreme, opposite, and rather dogmatic (they come about as a revealed moral truth). For classical liberals’ full self-ownership, the legitimate owner is the holder, from the “natural” ownership of her productive capacities plus free exchange of their product. For full “welfarism”, Rawls and income egalitarians, the holder has no particular a priori right in this value. Actual policies and most people’s thoughtful opinion stand somewhere in between. Individuals would have some more or less extended right in this value of their capacities. Similarly, in most countries someone who finds a treasure is entitled to keep some of it. Income taxes are not higher not only because of a possible disincentive effect but also because this may go too far in taking away the fruit not only of labour but also of the capacities it uses. The disincentive effect is also often more or less a pretext for implementing this view. However, it is also generally not considered unnatural that, in a community, people endowed with the “brute luck” of high earning capacities (due to their own characteristics and to society’s demand for their product) help people who happen to be much less fortunate in this respect. This is very generally accepted, to a substantial extent, in modern national communities. Individuals with higher earning capacities have a higher final disposable income as a consequence for two possible reasons. One is some right in the value of these capacities. The other is that taxing their labour income more has a disincentive effect costly for society. This latter reason occurs in both the “optimum income tax” studies and Rawls’s conception. The former reason occurs in particular when the base of transfers does not include full labour earnings. However, both the level and the structure of the final advantage of more productive people differ a priori in these two cases.

3.3 Society’s Principle of Macrojustice and Optimality 3.3.1 Theory 3.3.1.1 Endogenous Social Choice Optimisation begins with finding the relevant principle of optimality. The policy to be determined is intended for application. Hence it should be possible to apply the principle retained. This is not the case if everybody actually

78

S.-C. Kolm

rejects this principle, because this includes the persons who decide about its implementation, for instance voters and officials. Not to choose such principles is a condition for the social possibility of implementation. It can also be considered, on moral grounds, as a requirement of democracy. This is an aspect (the unanimity aspect) of the meta-principle of “endogenous social choice”, i.e. finding the principle of social choice in society itself. This section is divided in two parts. The present one shows the logic that leads to the choice of the principle. The empirical facts about people’s judgments are gathered in a second part. 3.3.1.2 Social Liberty and Given Capacities The main point is to discern the relevant distributive values. The bare bones of the question can be presented as follows. The basic value of our “free” societies is “social liberty”, the freedom from forceful interference, derived notably from non-violence, including the right to freely exchange without third-party interference, and entitling individuals to the result of their acts if they are legitimate owners of the means they use. More specifically, forceful interference can only either prevent forceful interference (protection) or result from a previous agreement. Social liberty plus the means used (which provide “real freedom”) constitute the person’s total liberty. Full selfownership implies social liberty (in line with the concept of ownership) but the converse depends on the specification of the concepts and on the policy (see Sect. 3.5.3.3). In particular, ELIE distributions are to be seen as allocating given resources rather than the product of labour exchange. We shall even see that, with an ELIE distribution, someone who pays a tax is de facto freer than someone who does not: she can work less and consume more; whereas someone who receives a subsidy is freer than if she did not. Hence the secure justification of self-ownership by “classical liberalism” is as a concept of “natural right” rather than of social liberty. With social liberty, what remains to be allocated are the resources that are given or “natural”. If, or insofar as, the distributive policy can be based on these resources only, it does not jeopardise the Pareto efficiency induced by efficient markets (a classical result of basic economics). These resources are either human given capacities or non-human assets. Capacities can be roughly but sufficiently and relevantly divided in two categories, productive capacities which, with free exchange, are also capacities to earn, and a family of capacities to enjoy, shortly discussed and represented by economists’ “utility functions”. The very largest part of economic value comes from labour using productive capacities, notably in an intertemporal view in which the value of capital, which is produced by definition, is allocated to

3

Economic Macrojustice

79

the human and non-human natural resources which, in the end, produce it (Sect. 3.5.2). Concepts of entitlement, legitimacy, and desert and merit are related to these values. The benefits from a capacity, or its shortcomings, can a priori be either attributed to its holder, or put in the common pool for redistribution by transfers of its product or earnings for a productive capacity or by compensatory transfers for the other capacities. This leads to three polar cases. Full self-ownership is the ethic of classical liberalism. On the contrary, everything goes into the common pool when the only end-values are individuals’ “utilities”. In an intermediate case, capacities to “enjoy” are self-owned but incomes are a priori fully available for redistribution. Actually, however, in a large distributive society two facts tend to produce intermediate cases with respect to both earning and “enjoying” capacities. One is the various possible divisions of the value (and output) of earning capacities between self-ownership and common distribution (this is the meaning of coefficient k with ELIE). The other is that a person’s “utility” results from several psychological phenomena and only some of them may be relevant. 3.3.1.3 Welfarism Between Necessity and Mistake A family of principles of optimality derives it from the consideration of individuals’ “welfare”. This was called “welfarism” by Hicks (1959), for criticising its excessive use by economists (probably for demanding more place for freedom). Welfare, however, is a problematic term. In common language, it seems, people understand it as meaning consumption goods, or income extensively reckoned with amenities. Economists, however, usually describe welfare by the level of “utility functions”. Now, when these “utilities” are given a tangible meaning – that is, when they are not solely the formal representation of orderings of a pure logic of choice – they are psychological phenomena, even intending to constitute an exhaustive coverage of general evaluative sentiments, emotions and feelings, such as satisfaction, pleasure and pain, happiness, enjoyment, well-being, or relief, while describing also tastes, needs, desires, urges, want, as well as aims, objectives, intentions, aspirations or ambitions, and intending to provide in the end an integrated explanation of choice and action. Let ui .xi / denote individual i’s such utility function, depending on her consumption goods represented by the bundle of their quantities xi 2 <m , which may also be her income that buys them, yi , and her leisure i . There are various ways of using these functions ui to evaluate society. A classical one is to do it by a classical “social welfare function” (SWF), an increasing

80

S.-C. Kolm

(or non-decreasing) function W .fui g/ the maximum of which is the social optimum. The issue concerns the scope of the validity of such principles.8 Various objections point out to choices that are not, cannot be or should not be made this way. They concern notably macrojustice. Some criticise the relevance of the functions ui or of aspects of them. Others emphasise directly other means, such as political processes or various types of political or private exchanges, including the classical liberal objection to any distributive tax in the name of self-ownership. John Rawls remarks in 1971 that “social justice” is never determined by thePmaximisation of a function W .fui g/ (he first refers to utilitarianism with ui but later makes it precise that he means any aggregate of the ui , including, in 1982, the limiting case of maximisation of the smallest).9 With philosophical precision, he criticises using functions ui which represent the individual concerns for “the good”, whereas he sees social justice as providing the individuals with the means of these choices only. Rawls then proposes an alternative theory because “this is the only way to criticise a theory”. His conclusion is an ideal equality of “primary goods” one of which is economic, income (or wealth), extended in 1974 to include a second economic primary good, leisure (“free time” might express better the value of a primary good). He also demands, with priority, “basic liberties”, the best concept of which is social liberty. In 1971 also, however, Jim Mirrlees applies to the determination of the optimum income tax the maximisation of a classical social welfare function which had been used with a similar logic for the microeconomic problem of finding the optimum non-linear tariff of public utilities.10 However, he replaces the individuals’ utility functions ui .yi ; `i / where yi and `i are individual i’s disposable income and labour by the same utility function u.yi ; `i / for all i. He gives for this the reason that “differences in tastes raise different kinds of problems”. Remarkably, this irrelevance of differences in tastes is precisely one of Rawls’s arguments for dropping functions ui altogether. Since individual utility functions ui actually differ from one another, this unique function u differs from them (except possibly for one) and, therefore, it can neither describe individuals’ choices nor define Pareto efficiency, 8

The two other main families of normative uses of utility functions are noted in Chap. 12 of this book which points out their meaning and their values. They are the family of “equity-no-envy”, an important meaning of which concerns equality and comparison of liberty, and the “equivalence principle”. 9 The maximin in sufficiently comparable utilities, and more generally the leximin, is “practical justice” in Kolm (1971). 10 Kolm (1970a,b) (see Mirrlees 1996). An application of this model to the determination of the optimum non-linear income tax schedule was withdrawn because this ethical principle was not accepted for this topic by fiscal authorities.

3

Economic Macrojustice

81

as assumed by the theory. Scholars working along this line almost always use this unique utility function, although, most of the time, by providing the different reason that they do not know the specific individual functions ui (the noted similar determination of optimum non-linear tariffs treated this problem by an a priori uncertainty about these different functions). Mirrlees reverted to the actual functions in 1986, but, as shortly noted, this probably does not fit with people’s general ethical opinion. This social rejection has a priori an advantage: it permits not to have to know the specific utility functions of all individuals.11 Therefore, there would be situations unanimously preferred to the maximisation of W .fui g/ both if the ui are the actual utilities (from people’s sense of justice) and if they are the same ui D u (for lack of Pareto efficiency). The crucial point is to know whether individual differences in a number of psychological (and physiological) possibilities to appreciate are relevant, or considered relevant, for macrojustice – and in particular the income tax. Section 3.3.2 proposes a number of tests to answer this question. The interpersonal differences are those in experiencing certain sentiments or feelings from some amount of income. The following notions can be so considered (roughly from more specifically personal to deeper, more objective and perhaps more relevant or less irrelevant ones): to fancy; to satisfy particular tastes; to derive pleasure, to like; to derive happiness (both deeper and stronger); to be satisfied, to derive satisfaction; to derive well-being. These feelings are due to hedonic capacities and tastes, also related to character; their relevance will turn out to be doubtful (contrary to the case of micro and local distributions in groups of people having much information about and empathy for one another). Then would come the satisfaction of basic needs and the relief of pain. Their interpersonal comparison is relevant for distribution, but this distribution concerns issues of microjustice in a society in a normal situation. Comparison of preferences (to prefer) does not in itself refer to a specific psychological reason and it can be caused by any of them. The concept of welfare has a special place, with probably some relevance in some domain, but when it is cleaned for the above irrelevant elements. The psychological differences between individuals can a priori intervene in different structural forms, depending on the shape of function W . The tests 11

The assumption that the tax authorities would know fully the utility functions of the individuals and not at all their wage rates is, of course, extraordinary. The 1971 article says more realistically that this information is “certainly not the case”. For the 1986 article, “The central element in the theory is information; public policies apply to individuals only on the basis of what can be publicly known about them”: but what is more private than individuals’ capacities to enjoy? The individuals would a priori distort their actions that would reveal their preferences used for taxing them, thus jeopardising economic efficiency and fairness. The full game-theoretic revelation model is not in this theory.

82

S.-C. Kolm

P are made easily in the two limiting cases, utilitarianism maximising ui on the one hand, and maximin in ui expressing an egalitarian ideal on the other hand, and if both give the same conclusion it certainly also holds for the other forms of function W . The conclusions of these tests – which will also consider entitlement from labour and effort – display the following general opinions. 1. Differences in hedonic capacities and tastes are irrelevant for macrojustice and in particular for optimum income taxation. 2. Differences in earning capacities have some partial relevance in this field. 3. Differences in hedonic capacities and in tastes are very relevant for proximity distribution in groups in which people have a strong empathy for and information about one another. 4. Propensities to suffer are very relevant for the relief of pain. However, cases 3 and 4 are issues of microjustice, except in exceptional cases of nation-wide disaster for case 4. Issues of suffering are the object of specific policies very important but micro (aid, welfare policies, social insurance, etc.). Consider view 1 by itself. If it results from an irrelevance of hedonic capacities and tastes and if these psychological (and physiological) elements suffice to determine utility functions fully, this view is that considered by Rawls, and it may lead to his result of an ideal equality of income y and leisure  (or labour ` D 1   if total time is measured as 1). 3.3.1.4 Strict Welfare However, there are also ethical reasons and sentiments referring to a notion of “welfare”. For instance, it is proposed that an extra dollar of tax should prima facie better be taken from a richer person than from a poorer one because it is less detrimental to her welfare. The same reason favours a progressive transfer of one dollar from the richer to the poorer. These views imply an individual welfare function which is a concave function of income and is the same for all individuals, and an aggregation of these functions for all individuals which is a sum (Pigou’s 1912 model) or more generally a symmetrical concave function of these functions.12 However, these reasons referring to individual welfare require that the lowest of the two incomes be sufficiently low. If both incomes are sufficiently high, these differential welfare effects are not distinguished (hence, if u.y/ is such a function of

12

Still more generally a Schur-concave function of them.

3

Economic Macrojustice

83

income y, we would have the derivatives u0 > 0, u00 < 0 and u000 > 0). However, at all levels there may also be a direct income inequality aversion due to the noted logical reason for equality – but this may require the joint comparison of more than two incomes (see Kolm 1999b). In relation, if result 1 is taken minimally and strictly, only differences in hedonic capacities and tastes are irrelevant. Then the rational solution is Mirrlees’s in 1971: replace individuals’ utility functions ui by the same function u. What should this function be, however (Mirrlees provides no cue)? We want to erase away interindividual differences in hedonic capacities and tastes (and possibly other elements) and have, as a residual, a utility function which would be the same for everybody and could represent individual “welfare”. This can be done thanks to the available SWF W Œfui .xi /g in the following way. Assume functions ui to be comparable (at least co-ordinal) and function W to be non-decreasing and increasing in at least one argument at each point, and symmetrical. Then define the function u.x/, in the relevant domain X with x 2 X  <m , by the value of u defined, for each x 2 X , by the equation W Œfui .x/g D W Œe u.x/ (3.1) where e is a vector of n ones. If all functions ui are increasing, or decreasing, or – if all ui and u had the properties of quantities – concave, or convex, (all strictly or weakly) in a domain, so is function u. If individual utility functions ui are uncertain for the policy-maker – which they actually are –, describe this uncertainty by their being stochastic variables uQ i , choose for W a specification of the risk-relevant von NeumanMorgenstern cardinal specification of this evaluation function, and define function u as, for each x, E W ŒfQui .x/g D W Œeu.x/: Function u is then an individual welfare function. Irrelevant differences in the ui have been averaged away with W as “averaging function”. The new social welfare maximand becomes13 W Œfu.xi /g:

(3.2)

As we have seen, the end-values of a conception of justice should be equal prima facie (ideally, in the absence of an overpowering reason), from P For example, if W is additively separable, theP generalised mean PW D gŒui .xi /, then u.x/ is 1 1 1 of ui .x/ with function g, u.x/ D g Œn ıui .xj /. In g ıui .x/ and W Œfu.xi /g D n i;j gP P P particular, if W D ui .xi /, u.x/ D u.x/ D .1=n/ ui .x/ and W .fu.xi /g/ D .1=n/ i;j ui .xj /. If W D min ui .xi /, u.x/ D min ui .x/ and W .fu.xi /g/ D mini;j ui .xj /. Chapter 12 of this book shows important properties of the “strictly welfarist” function W Œfu.xi /g. 13

84

S.-C. Kolm

a weak condition of rationality (the mere possibility of providing a reason for their value). In practice,14 function W is taken to be weakly concave and functions ui are taken to be strictly concave. The maximand should be inequality-averse in the end-values. This allows for two possibilities, however. One is “strictly welfarist” and takes the levels u.xi / as end-values, with aPstrictly concave function W (this excludes the corresponding utilitarianism u.xi //. Indeed, if the u.xi / are not all equal, X W Œfu.xi /g < W Œen1 (3.3) u.xi /: The other possibility takes the xi directly as end-values. Indeed, with a weakly concave W and strictly concave ui (and therefore u), if the xi are not all equal, X X W Œfu.xi /g  W Œen1 (3.4) u.xi / < W Œeu.n1 xi /; which shows that the maximand is strictly averse to the a priori multidimensional inequality in the xi . Moreover, function W is symmetrical both as a function of the levels u.xi / and as a function of the vector variables xi . The second inequality suggests a social peference for choosing, as end-values, the xi rather than the u.xi /. The “strictly welfarist” egalitarian case, with the possibility of basing the tax on wage rates (and the proposal to do it by exempting overtime labour income) is analysed in Kolm (1974b). Note that function u differs from the actual functions ui and can describe neither individuals’ reaction nor standard Pareto efficiency. The good-egalitarian interpretation will be used in a way that avoids these difficulties (and it is consistent with utilitarianism). Its main advantage, however, is that it leads to the structure considered here which has a number of other meaningful ethical properties. Now a basic epistemological principle of ethics is that a solution has to be evaluated by considering all its properties. Philosophically and ethically, the difference between taking as end-values the goods xi or the u.xi / amounts to discarding as irrelevant for macrojustice either individuals’ tastes and hedonic capacities or interpersonal differences in them only. The tests of comparative fairness presented in Sect. 3.3.2.2 do not seem able to discriminate between these two interpretations. Strictly speaking, their refer to differences, but this may result from a sentiment of irrelevance of the ui altogether. There is, moreover, a third possible conception. If utility functions ui are understood as meaning individuals’ choices For instance in the “welfarist” income tax studies, with xi D .yi ; `i /, the pair of income and labour (or leisure i ). 14

3

Economic Macrojustice

85

by their maximisation, then their irrelevance leaves the domains or possibilities of free choice as end-value, with a value which is the freedom of choice they offer. If this is concluded from an irrelevance of tastes, these discarded functions ui could be ordinal only, and, hence, the corresponding concepts of liberty and equal liberty should make no use of these functions and of the orderings they represent. The possible such concepts of equal liberty are analysed in Sect. 3.6. They all lead to the same result, which is also the one obtained by taking the allocations xi as end-value and correcting the result for questions of efficiency, liberty, and demanded partial self-ownership. Indeed, the egalitarian outcomes are, then, with the same xi D x for all i. Then, maximand (3.2) takes form (3.1), and the best x maximises u.x/ under the corresponding constraints. For macrojustice, xi can be taken as income yi and leisure i or labour D ` for all i (Rawls’s ideal `i D 1  i . With equality, yi D y and `i P solution). The distributive constraint is ny D wi ` or y D `w. Then the chosen ` is ` D k that maximises u.y; `/ under the constraint y D `w. This is described in Fig. 3.1 with labour ` and income y as coordinates as the maximisation of u.y; `/ under the constraint represented by the line with slope w from the origin. This is the behaviour of the “average individual” with the average utility function u.y; `/ and the average wage rate w. However, this solution has three defects. (1) Lack of freedom: individuals generally prefer to move from it, for instance to work more and keep the extra earnings. (2) Lack of Pareto efficiency, as shown by the same fact. (3) There is no element of specific self-ownership whereas this value is more slope wi

y

ui

slope w

K kw

0

u

k

Fig. 3.1 The two-part income: equality and liberty



86

S.-C. Kolm

or less present in general public opinion. The same solution solves these three questions. From the obtained distribution, let the individuals be free to work more and keep the extra earnings (this will be the relevant case). Then, if individual i chooses to work `i k, her extra gain is .li  k/wi and her income is yi D k w C .`i  k/wi . On Fig. 3.1, the individuals choose their preferred point on their budget lines which are the lines issued from point K with ` D k and y D k w, with slopes wi . This solution is, finally, a case of free exchange from an equal allocation, which is a principle of equal liberty. This aspect, and its equivalence with other principles of equal liberty, is developed in Sect. 3.6.

3.3.2 Is Society Welfarist? “Welfarism” and “welfarist” are now used in their present-day common sense in economics (irrespective of the interindividual “welfare function” previously defined). 3.3.2.1 The Scope of Full Welfarism: Proximity and Pain Welfarism is actually the first-best distributive principle. Indeed, it is applied both in the first-best society made of good, altruistic people who care much about others’ desires, tastes, satisfaction and happiness, and when goods are in painful scarcity and hence are the most needed. For instance, you may give the toy to your daughter rather than to your son because she enjoys it more or because she is a little sadder these days. These quasi-utilitarianism and maximin in “utility” are two limiting cases of welfarism. Welfarism is, indeed, a common principle of distribution within families. This extends somewhat to distribution between acquaintances. The people in question often have empathy towards the others based on a sufficient information, and they understand the reason for the allocating choice. Moreover, relieving deep pain is a foremost duty. Poverty is bad because it implies limited freedom but certainly also because of the sufferings it entails. Deeply depressed people are justly assisted. Surgeons transplant the rare organ to the individual who suffers the most or whom it relieves the most. Emergency care is allocated similarly. These are again instances of the two limiting cases of welfarism. Medical choices are indeed often welfarist in the sense that they refer to pain. Welfarism may be the best family of principles for managing a hospital, a health department, a welfare program, or some

3

Economic Macrojustice

87

hard cases of development. It becomes the general rule of society in case of a general disaster – war, famine, flood, etc. – (coupon rationing according to the basic needs of types of people or activities is a welfarist paragon). Courts estimate praetium doloris in order to compensate for pain. As we will see shortly, it makes a large difference, with respect to the choices of justice as people see them, whether a person’s utility is in the range in which it means lower pain or in the range in which it means higher pleasure. Distributive choices are strikingly more prone to be derived from comparisons of welfare when increased welfare means lower suffering than when it means extra pleasure.15 Proximity and suffering are, indeed, the two domains in which welfarist judgments are found to be common. Welfarism is commonly endorsed when it means “familism” or “dolorism”. However, these are instances of specific microjustice, not of macrojustice in a society in a normal situation in which specific welfare policies or insurance deal with cases of particular misfortune. It is remarkable that these two domains of common welfarism correspond to the two reasons for altruism: particular empathy related to proximity, and compassion for suffering. This leads to two remarks. First, some people hold that these cases are about altruism and not about justice. Second, the intellectual ancestors of welfarism, classical utilitarian philosophers such as J.S. Mill (1861) and H. Sidgwick (1874), associate utilitarianism and altruism. In the opposite type of social sentiments, people are indifferent to others’ welfare (or even envious) and try to take from them as much as they can. Intermediate social relations, that are neither so good as to be purely altruistic, nor so bad as to a priori rest on war or balance of force and threat only, notably within large-scale communities, have to rely on pure conceptions of justice for their principle of allocation. Now, the redistributive structure of the income tax allocates neither between suffering poor nor between related tax payers, but between the cooperative members of national communities. Is it and can it be derived from welfarism?

15

That is, allocation between suffering people tends to take the individuals’ different propensities to suffer into account, whereas allocation between normally satisfied people rather compares incomes directly and tends to discard the individuals’ different capacities to enjoy and to leave the individuals accountable for them. This may be formalised by attributing to each person an index which becomes her utility when it is low and her income when it is high. However, the usual distinction between income inequality and poverty, or the focus on macrojustice, enables one to take care of such differences. Bentham (1789) wrote: “to minimise pain, or, which comes to the same, to maximise pleasure”. This equivalence does not seem to be endorsed by common judgments for distribution in the noted cases.

88

S.-C. Kolm

3.3.2.2 Tests of Welfarism for Macrojustice16 Since a normative study can be applied only if people who actually influence its implementation sufficiently adhere to its normative criterion (they can be voters, people at large, politicians, tax officials, etc.), and the welfarist model of optimum income taxation is probably proposed for application, this model is based on the hypothesis that welfarism is an accepted principle for macrojustice. Does any test falsify this hypothesis, or not? Here are a few tests among many possible ones. The European Union Test If, as it is said, the people of Northern Europe are better at producing and those of Southern Europe more skilful at enjoying consumption, should the European Union set up a vast program of intra-European North–South income transfers? Should it tax the industrious Swedes for subsidising the Napolitans who make a feast from a meal? This would be the injunction of utilitarianism. Or perhaps, on the contrary, should this tax subsidise the Portuguese reputedly afflicted by a kind of mild sadness, in order to soothe their saudade? This would be required by a maximin in utility. Of course, everybody should help victims of deep poverty in the mezzogiorno or elsewhere, and such differences in tastes often influence the voluntary distribution within small groups with strong mutual sympathy such as the family; however these are cases of specific microjustice aiming at the relief of suffering or within such groups. The Earned Income and Legitimate Ownership Test17 “I take the 10 euros you just earned because I like them more than you do (or more than you dislike the labour of earning them)”. Is this a good reason? Or perhaps, on the contrary, “I take your earnings because you like your euros left more than I like mine”. Is this a better reason? Am I entitled to (or

16

All-purpose or universal welfarism (i.e., evaluating all social issues, including their distributive effects, by comparison of individuals’ satisfaction or happiness only) may be a scholar’s moral taste, and perhaps, indeed, the very first best principle, but there is no chance to see policy advice derived from it applied in domains for which the population finds this criterion not to be relevant. Benevolent dictators and philosopher kings alone can hold this view and apply it everywhere. They are myths. 17 Differences in satisfaction across individuals may be due to their allocations (of income and labour, notably), and not only to overall differences in tastes.

3

Economic Macrojustice

89

should I) take your money because it pleases me more than it pleases you? Or perhaps, on the contrary, because you enjoy your money left more than I am able to enjoy my own? These two opposite consequences of comparing our tastes for income are respectively utilitarianism and maximin in utility, the two polar cases of welfarism. If, however, your 10 euros enable me to buy the drug that saves my life, most people will excuse the theft; but this is a case of specific microjustice for the alleviation of suffering. The Tests of Cheap or Expensive Tastes, Preferences, Needs, Desires or Wants Should you finance somebody’s beverage because her special taste for cheap beer permits her drinking to produce utility at low cost (as utilitarianism requires)? Or because she only likes expensive wine (as egalitarian maximin or other welfarist principle may demand)? She may on the contrary have to compensate you for lacking this capacity for delicate pleasures. Tastes, and preferences that describe them, by themselves, are usually not seen as implying distributive norms. Nevertheless, they are a basic reason for distribution in small empathic groups (e.g. families), and you should probably give water to a thirsty passer-by, to relieve her pain cheaply. Bar-Hillel and Yaari’s (1984) experiments show the evidence of unanimous ethical judgments about distribution, that make a large difference depending on whether the issue means tastes or needs. And “to each according to her needs” is a classical principle. Indeed, vital and basic needs probably have to be satisfied for alleviating pain (or securing freedom). However, no such norm seems to attach to the fancy “needs” of the collector. More generally, desires and wants are considered self-accountable (no other person has a duty to satisfy them for the only reason that they are desires or wants – and not because their satisfaction would alleviate pain, for instance), except in particular social relations (and then their particular reason may induce such a duty); therefore, they are deemed self-accountable for macrojustice. This is Rawls’s “social justice”, for which he observes: “Desires and wants, however intense, are not by themselves reasons in matters of justice. The fact that we have a compelling desire does not argue for its satisfaction any more than the strength of a conviction argues for its truth”.18

18

The very large variety of psychological (and physiological) facts that economists’ concept of utility can describe have been pointed out. This plurality and versatility have a certain beneficial aspect in the analysis of behaviour – for instance, it permits the generality of choice theory. However, it raises problems for explaining specific behaviours, which depend on the type of motive, and particularly for normative analysis because different meanings often entail different judgments.

90

S.-C. Kolm

The Income Tax Test Should you pay a higher income tax than someone else because you like the euros taken away less than she does or, on the contrary, because you like the euros left more than she does – as utilitarianism and maximin in utility tend to require, respectively? Are, in fact, these considerations relevant for this issue? To begin with, do these comparisons of enjoyment make sense, are they possible? A joyful character or a happy mood may imply that happiness does not depend much on wealth (equanimity) or, on the contrary, that one is able to derive much pleasure from each euro (sybaritism). Hence, with utilitarianism or egalitarianism, there are four cases. Should such a fortunate person pay a higher tax because she stands the loss more easily, or because she is able to derive much pleasure from each euro left; or a lower tax because she would have enjoyed much the euros taken away, or because her satisfaction is both only little sensitive to wealth and not high? This latter case, however, is not satisfied if equanimity not only means low sensitivity to variation but also implies a sufficient level of satisfaction. Should propensity to spleen and melancholy be valid reasons to claim an income tax rebate? In fact, has the Internal Revenue Service ever thought about sending questionnaires to inquire about these relative propensities or capacities to enjoy? Or does it think that this would be irrelevant and, perhaps, abusively intrusive; that these psychological characteristics are private matters and not the concern of overall and general public policy and the income tax; that, for this question, people are accountable for their own tastes, entitled to their beneficial effects whereas they have to endure non-pathologically less favourable ones; and that such normal differences in tastes could not give rise to compensating claims on others’ incomes or liabilities towards them?19 The Implementability Test The welfarist theory of the optimum income tax is about a very important topic. It is very well known (and justly admired) by economists who want their work to be useful and seek application. Some eminent contributors to 19 Any more than, for instance, physical beauty. This self-accountability is a notion of selfownership. Responsibility is only one possible cause of accountability among various others. People can be held “responsible” for their tastes (Kolm 1966b, Dworkin 1981) insofar as they can influence them only. This a priori has limits, but, more basically, this question raises deep conceptual issues, such as the place of the “weakness of the will”, often relevant about tastes, which can be classified either as a constraint on choice or as a chosen property – somehow as with laziness, for instance (see Kolm 2005, pp. 101–104).

3

Economic Macrojustice

91

it have even had major economic responsibilities at world and national levels. Why, then, is this remarkable theory still waiting for the beginning of an application after nearly four decades? Can it be applied, at least in a democracy? To begin with, would officials and voters endorse its welfarist ethic? In fact, all the available information shows that, when it is explained to them, they discard its way of choosing the distribution by kinds of comparisons between individuals’ welfare or their variations, for this application beyond issues of suffering and small groups. Bourguignon and Spadaro (2008) show that the actual income tax schedules cannot be derived from the theory’s maximisation. The Distributive Opinion Test The opinions about overall distribution that exist in society have two polar positions; policies apply some mixture of them or compromise between them, and individuals also often endorse more or less some mixture. One polar position is income egalitarianism. It favours lower income inequality. Hence, it sees equality in incomes as the ideal. Since individuals have different utilities, this cannot result from this welfarism. The other polar position holds that earned income should belong to the earner (“classical liberalism”). It is not welfarist either. Hence, welfarism seems absent from actual moral positions about the overall distribution in macrojustice. The Rawls’s (and Many Other Scholars’) “Reflective Philosophy” Test John Rawls is the most famous of contemporary philosophers. His basic work, A Theory of Justice, is an indictment of welfarism for macrojustice (his “social justice” – he uses the term “macro” once and says “not micro”).20 He says he presents his own theory because a critique is fully convincing only if an alternative is proposed. Some economists hide this fact by calling “Rawlsian” a maximin in utility. However, Rawls’ maximin (his “difference principle”) is in “primary goods”, not in utility. This most basic point is unambiguous: “To interpret the difference principle as the principle of maximin utility (the principle to maximise the well-being of the least advantaged person) is a serious misunderstanding from a philosophical standpoint”

20

His view on this point is shared by a large number of scholars in the various disciplines (among others Dworkin 1981, but also “classical liberals”). Yet the rest of their conception, as that of Rawls, raises problems.

92

S.-C. Kolm

(Rawls 1982).21 Hence, his remarks that “Justice as fairness rejects the idea of comparing and maximising satisfaction” and “The question of attaining the greatest net balance of satisfaction never arises in justice; this maximum is not used at all” (1971), intend to point out a commonsense and moral inappropriateness of unsophisticated welfarism. The way in which policy choices are actually made corresponds to Rawls’s observation, and Bourguignon and Spadaro (2008) show econometrically that the choice of income tax schedules cannot be said to maximise implicitly a classical social welfare function. Therefore, Rawls naturally acknowledges: “A principle of equal liberty”. “A just social system defines the scope within which individuals must develop their aims, and it provides a framework of rights and opportunities and the means of satisfaction within and by the use of which these ends may be equitably pursued” (id.). These differences should not hide the fact that Rawls and welfarism (e.g. Mirrlees) share the same assumption about the most basic issue in the ethics of distribution: individuals are not entitled to the income-earning possibilities offered by their given productive capacities, in opposition to the view of “classical liberalism”.22 ;23

21

The leximin in interpersonally comparable utility is the eudemonistic “practical justice” in Kolm (1971), discussed by Rawls, and it was proposed as a limiting case and not for application to all problems of distribution. 22 Beyond these general conclusions, however, most of Rawls’ more specific proposals are logically problematic for specific reasons. (1) His maximin in “primary goods” (the “difference principle”) omits that the bases of transfers and taxation can be much less elastic (hence waste inducing) than they presently are – the issues of defining an index of these goods and of relating this to Pareto efficiency are much more secondary matters. (2) The theory of the “original position” and of the “veil of ignorance”, both in Rawls’s version and in Harsanyi’s (which gives a kind of utilitarianism or, at least, separable welfarism), are problematic because a selfish individual choice in uncertainty does not have the same structure (and objects) as a choice of justice (see Kolm 1996a, pp. 191–194, and Kolm 2005, pp. 358–360). (3) The classical theory of equal and maximal real basic liberties does not hold (see note 30 below). 23 In The Law of People (1999), Rawls rejects distributive transfers across “people” (say, nations). This certainly does not refer to help in case of emergency or disaster inducing much suffering, but to a view of pure distributive justice only. Also, Rawls’s theory of justice is explicitly for “social justice” or macrojustice and not for small local communities. This suggests the consideration of three levels with three rules for the allocation of the two types of capacities, productive capacities and hedonic capacities or capacities to enjoy. For small, tight, mutually altruistic groups (e.g. the family), all capacities are pooled and the rule is full welfarism. For large communities (e.g. nations), productive capacities are pooled but hedonic capacities and tastes are self-owned. Finally, there is no redistribution across nations which keep the benefits and shortcomings of their own stock of capacities of all kinds. The general motto of this system is: altruism within the family, social justice within nations, egoism beyond. Fully welfarist optimum taxation and classical liberalism extend, to the second level, the principles of the first and third ones, respectively.

3

Economic Macrojustice

93

The Constitution Test The basic principle of our societies, the transgression of which is unlawful and punished, is given by our constitutions and founding declarations. It consists in liberty and rights rather than welfare. Happiness is essential but private. “Men are free and equal in rights”. They should be secured the liberty and means to “pursue happiness” as they see fit, rather than some level of happiness.24 Property rights are basic, and the legitimacy of someone’s property of something is provided not by some beneficial consequence but by the condition of its acquisition, notably free actions and exchanges. Empirical Studies The non-welfarism of the relevant general opinion is confirmed by empirical studies such as those by Schokkaert and Lagrou (1983), Schokkaert and Overlaet (1989) and Schokkaert (1999). Bluntly welfarist optimum taxation thus seems to equate justice with charity or a nation with a family. This is very defensible on moral grounds, but it does not correspond to people’s view and hence could not be implemented. The demanded principle of macrojustice fully respects individuals’ sui generis tastes, hedonic capacities and preferences by leaving them in their private sphere.

3.4 The Optimum Tax Base 3.4.1 Practical Possibility The optimisation of taxes and transfers should be not only for their structure but also for their base, for what they depend on.25 The desired tax and subsidy base of macrojustice theory will happen to be individuals’ given earning capacities, hence excluding chosen labour, for several reasons. This is a condition for “social liberty” applied to free exchange (of labour). In relation, individuals are responsible for their chosen labour and are not for their given capacities. Moreover, basing the tax or transfer on a given item and not on a chosen one is a classical condition for Pareto efficiency (itself desired for

24 25

The 1789 Declaration of Rights and the American Declaration of Independence. The consequences of using various bases for the income tax are analysed in Kolm (1974b).

94

S.-C. Kolm

several reasons). The non-waste effect of Pareto efficiency is a reason for other theories to also a priori want the same tax base (see, for instance, the discussion in Mirrlees 1971). Therefore, the first thing that we would not want in the tax-subsidy base is the chosen duration of labour, `i . In this respect, the first piece of information to consider is the present tax law in France which exempts overtime labour earnings from the income tax. This is done without cheating for practical reasons – tax authorities would necessarily be informed about it. This exemption is over a rather low official labour duration,26 which could be reduced in order to apply to more people – the tax rates would be augmented as a compensation. This applies to wage labour which makes up 9/10 of all labour as in all developed countries. This nation-wide experience, beginning on October 1st, 2007 is the main fact.27 Other aspects of this tax law are relevant. Premia for previous formation and level of education, and productivity premia, over a given base, will also be exempted from the income tax in order to take account of these two other dimensions of labour, formation-education and intensity. The usual strong correlation between the type of job and intensity and required formation can also be used. At any rate, the tax on increased earnings due to formation finances the free public education system. For part-time labour contracts, the corresponding “complementary hours” are exempted. The taxation of households is the one usual in this tax system (“family quotient”). For non-wage labour (1/10 of total), lump-sum taxation is frequent for self-employed people, farmers and professionals. A given activity can be compared with similar ones for which the earnings per time unit are known such as wage labour, in a categorisation of activities. In all cases, the routine procedures of tax authorities use statements, matching those of buyers and sellers, checking accounts and durations (opening hours), random investigations and penalties, reference to standards for given types of jobs and activities, etc. When the tax base is total earnings, about 30% of the base evades the tax in all countries (see for instance Slemrod 2002 for the USA). This is a far cry

26

35 hours a week, or 1,607 hours per year, or 218 days per year for executives and others whose daily hours of work are unclear. The exemption is of both the income tax and the contribution to financing social security. The marginal “wedge” suppressed as a consequence is often 65%, a very high value. 27 The perfect availability of the necessary information in almost all cases, noted and explained in Chap. 2, implies a particular low cost of fiscal administration for this tax. The optimisation of fiscal implementation could thus leave penalty rates rather low (contrary to the case with costly controls – see Kolm 1973). The reason is not any morality of taxpayers but simply the properties and uses of firms’ accounting. There had also been, previously, a tax equal to earnings during a given length of time, also easily implemented.

3

Economic Macrojustice

95

from perfect information. Mirrlees (1971) takes earned income as tax base because “the natural, and one would suppose the most reliable, indicator of a man’s income-earning potential is his income”, but he ends up noticing that we also often know labour duration, and hence, with both, the wage rate, and that “we also have other means to know a man’s skill-level”. He suggests, however, that this would induce much hiding and evasion. The French tax experience, by selecting a given partial labour duration and the corresponding earning, shows the opposite. The fiscal administration needs to know neither earned income nor labour duration.

3.4.2 Incentive Compatibility These possibilities associate with the incentive compatibility of the obtained (ELIE) tax-subsidy policy (Sect. 3.8). Indeed, it induces individuals to work with their capacities that are the most highly remunerated, and hence the most useful to society (in this particular sense). Therefore, the actual, observed and taxed or subsidised wage rate is that which corresponds to these capacities.

3.5 Economic Liberties, Resources and Capacities 3.5.1 Liberties The obtained equally free exchange from an equal allocation of income and leisure or labour is a concept of equal liberty, and we will see that it is equivalent to other important concepts of equal liberty. Equal liberty can in fact be seen as a consequence of simply discarding utility functions for macrojustice, which discards tastes and hedonic capacities. Discarding utility functions can indeed lead to two types of variables as end-values. If “utility” describes pleasure derived from consumption, these variables are the bundles of consumption goods (perhaps income and leisure). If utility functions represent the orderings that lead to choice, the remaining variables are the domains of free choice and the moral value is the freedom they provide. These are the two natures of human beings seen by philosophical anthropology: sentient beings capable of pleasure and pains, and free agents capable of choice and action. In the second view, the moral ideal is equal freedom. However, the relevant economic liberty refers to two types of

96

S.-C. Kolm

freedom, defined by the nature of the constraint and the domain of choice, respectively. “Social liberty” is the basic, constitutional and legal rule of our “free” democratic societies. It means that individuals’ acts should prima facie be free from forceful interference by others individually, in groups, or in institutions. Individuals can only be forced not to force others or to do something about which they have agreed they could be forced – for instance in a previous agreement or contract.28 Free exchange without any kind of forceful interference by a third party is an important application. Social liberty implies the respect of the intended consequences of individuals’ respectful actions (including free agreements or exchanges) – such as rights they may create.29 Social liberty may have to be respected simply because it is the meaning of the constitutional basic rights and hence violating it should a priori be unlawful and punished. Moreover, it is wanted by practically everybody in societies where it prevails. It can also be intrinsically defended for its meaning of absence of direct violence (especially since – as we will see – it can be considered as compatible with a distribution banning poverty). Social liberty is non-rival. Indeed, each individual can have it at satiety, for all her actions that respect others. Hence, social liberty is equal for all in this sense. Incompatibilities and conflicts between individuals’ actions are due to issues about the allocation of other means (in particular of other rights), and this allocation results from the question of the allocation of resources (several actions of an individual can also compete for this individual’s means of various kinds).30 ;31 28

Constraints on some insufficiently informed or insane person imposing her to do what she would have chosen if she were fully informed or sane can be seen as extensions of this liberty. Another extension is that of public constraints that implement not only actual contracts but also implicit ones impaired by impediments of any kind (e.g. for financing public goods or internalising externalities). 29 Social liberty is the full theory of related notions presented under various names such as “civic or social liberty” (J.S. Mill), “negative freedom” (Kant, J.S. Mill, Berlin), “liberté civile” (Sismondi), “liberty of the Moderns” (Constant), “formal freedom” (Marx), or “process freedom”. The term liberty – rather than freedom – is sometimes restricted to social liberty (e.g. by some translators of Kant), but this has not gained general currency. 30 Another classical conception wants to associate to each basic right – which is social liberty for a broad kind of application – material means that make it “real”, and it wants the resulting freedom to be “equal for all and maximal” (Rousseau, Condorcet, the 1789 Declaration, J.S. Mill, Rawls). However, since there is no a priori limit to these associated means (to the size of the cathedral for the freedom of cult, of the various means of communication for the freedom of expression, of private planes and airports for the freedom to move, etc.), this would determine the totality of the allocation of goods, with no rule for choosing among the various goods. 31 Social liberty can also be supported by a logical requirement. Indeed, consistent individuals want not to be prevented from doing what they want to do, that is, they want social liberty for themselves. Yet, their opinion about justice in society has to be impartial, from the nature and definition of a

3

Economic Macrojustice

97

An individual can also have means, possibilities, other rights, assets and liabilities, which, given social liberty, provide “real freedom”. The conjunction of her social liberty and of these elements provides her total liberty.

3.5.2 Resources Social liberty and Pareto efficiency require distributive transfers to be based on inelastic variables (as far as possible), that is, on given resources. From the above, interpersonal differences in capacities to be satisfied are discarded for macrojustice. Intertemporally, capital is produced and the remaining given resources are the “natural” ones, human natural productive capacities used by labour (including through learning), and non-human natural resources. The latter account for only a very small fraction of the total value of the output.32 Hence, the problem of macrojustice is that of the allocation of the value of productive capacities.33 At a given time, capital income is labour income plus intertemporal exchange if the capital originates from savings from labour income. Hence, the remaining conceptual issue about capital income with social liberty is the ethical and tax treatment of bequest. Another intertemporal question raised by distributive reforms is the treatment of wealth accumulated in the past under different rules. These classical questions will not be touched in this short presentation.

concept of justice. Hence, this opinion has to want social liberty for everybody, if this is possible, and it is possible from non-rivalry. 32 As an order of magnitude and for example, the contributions of labour, capital and non-human natural resources to the value of yearly output are nowadays often about in proportion to 80, 18, and 2, respectively. However, capital is itself produced, and hence the recursive assignment of its share to the other resources gives an order of magnitude of 97.5% for labour and 2.5% for nonhuman natural resources. Moreover, labour uses productive capacities but not all of them, whereas “land” includes residential land. This order of magnitude is one of the most ancient and classical of economic ideas. Locke (1689) says that labour accounts for “9/10 and in fact, if everything is counted, 99/100” of the product (see also Ricardo and Marx, for instance). 33 Non-human natural resources are allocated in various ways including by criteria of microjustice (e.g. proximity, discovery, first occupancy, best use, needs, or various welfarist criteria); they are usually owned and have had several owners; they (notably new natural scarcities) or their value can be allocated in various ways (including equally shared, helping the neediest, used for specific services, or for provisioning the public budget). See Kolm (1985, Chap. 10), Kolm (1986b), Kolm (2005, pp.84–89) and Chap. 12 below.

98

S.-C. Kolm

3.5.3 Rights in Capacities 3.5.3.1 Use-Rights and Rent-Rights Finally, in the rights concerning an asset one classically distinguishes the right to use this asset, or use-right, and the value of the possibility to use it, or rent. This distinction is essential for human capacities because social liberty implies that the use-right belongs to the holder of the capacity (who can rent it out for a wage). The rent of a productive asset (notably a capacity) is equal to its productivity, the value of its possible production. However, the rent of someone’s productive capacities, for a certain time or labour, may a priori belong to some other person. Then, the former, who has the use-right if there is social liberty, pays this rent to the other. She is only the tenant of this part of her capacities (a necessary tenant, however, since social liberty implies that she has the use-right). If a person owns the rent of her own capacities for a certain time or labour, she has the corresponding ownership since ownership is use-right plus rent. In particular, there can be full selfownership. A person may both owe some rent of capacities of hers and own rents of others’ capacities (a reciprocity of this kind will happen to be the result of the theory of equal liberty). 3.5.3.2 Self-Ownership Self-ownership of given productive capacities (hence of earning capacities with free exchange) is the object of two very important, firm and opposite moral judgments. Income egalitarians (including Rawls’s ideal) and full welfarists reject it because individuals are not responsible for these capacities and hence do not deserve them.34 Classical liberalism, on the contrary, can be defined as the advocacy of full self-ownership. It has two types of motives for that. One is the sentiment that the person “naturally” owns herself (it used indeed to be called a “natural right”). There is an association between ownership and being part of: a person’s physical and mental capacities are hers because they “are her”. They belong to her (property) because they belong to her (being a part of). This is a concept of selfhood or integrity of self that focuses on the individualistic viewpoint.35

34

“They only took the pain to be born” (Beaumarchais). A difference is often seen, in this respect, between capacities to enjoy or choose (perhaps a utility function) and productive capacities because the latter are more instrumental and their product can be alienated (the former, being the person’s capacities to choose and to derive pleasure or pain, can be seen as belonging to a more intimate “core self”). Classical liberalism and welfarism amount 35

3

Economic Macrojustice

99

3.5.3.3 Social Liberty and Classical Liberalism The second classical liberal stance is its assimilation of self-ownership with social liberty. Full self-ownership certainly implies social liberty, because it implies holding the use-right in oneself and because the concept of ownership implies the absence of forceful interference. But is the converse true? Classically, the setting of social liberty is that an allocation of the given (natural) resources that exist in society is chosen, and, from and with this basis, there are free actions and exchanges in social liberty (this is for instance the form of the standard theories of markets). The standard justification of self-ownership by social liberty is that a forced transfer, say a tax, forces the individual to consume less or to work more, and therefore has the effect of a constraint on the individual’s actions. However, this demand may be part of a larger package such that, on the whole, the person has a given net asset rather than a liability. With an ELIE distribution, for instance, an individual i with a unit earning capacity wi > w is subject to the conditional liability of paying ti D k  .wi  w/ in the normal condition in which her free labour `i satisfies `i > k. Since she also has, as a positive asset, the high earning capacity manifested by her wage rate wi , on the whole she benefits of yi D wi `i Cti D .`i k/wi Ck w > 0. For given `i , this benefit is higher the higher wi , and therefore the higher the liability ti D k  .wi  w/. A higher tax ti (for a given k) permits individual i both to consume more (higher yi ) and to work less (lower `i ). It makes her de facto freer by inclusion of domains of choice. Moreover, the payment is transferred to people with a wi < w, which expands their domain of choice of consumption and leisure. Therefore, the classical liberal objection to transfers as violating social liberty because they induce forced extra labour or reduction in consumption, is – in a sense – not valid in this case, in which the initial allocation of the possibility wi and the transfers ti are considered jointly, as a package of the allocation of the given resources that exist in society. Note that the value wi depends both on the individual’s characteristics and on society’s demand for their services and hence valuation of these characteristics. This classical liberal objection would hold if it is considered a priori that the individuals are fully entitled to both their own productive capacities and the value that society attaches to them. But this can only result from the other type of

to allocating a priori to each person all of her capacities or none of them, respectively. Whereas egalitarians and in particular income egalitarians, such as Rawls, so allocate capacities to enjoy and choose to “socialise” all productive capacities (the unit value of which is the wage rate) – see Sect. 12.2.1.

100

S.-C. Kolm

classical liberal justification, for instance as some kind of natural right to full self-ownership (implying in particular free sale of one’s services). Hence, the problem of distributive justice in macrojustice is the allocation of the value or rent of individuals’ given productive capacities. There remains to see how the general principle of equal liberty solves this problem, and the resulting policy.

3.6 Equal Economic Liberty 3.6.1 Possibilities We want to consider the consequences of equality in all the economic freedom individuals have, given social liberty and Pareto efficiency. First of all, equal economic freedom should be defined. There is, equally for all, (full) social liberty of the acts to choose, exchange and earn from or with some given initial conditions or allocation. This has to be seen as equality in social liberty. The remaining equality concerns the initial given conditions. This initial equality can take four forms: 1. Equal initial allocation. The other forms describe properties of the given domains of choice. 2. Socially free individuals are susceptible to choose an equal allocation. 3. Identical domains of choice. 4. Equal overall freedom provided by different domains of choice. We will see that solutions 1, 2 and 4 give the same result, whereas solution 3 is impossible in the sense that it violates Pareto efficiency and social liberty if individuals’ preferences are not taken into account (from nonwelfarism or ignorance) to define the domain – and it may violate them even without this qualification. Note that the direct ethical relevance of possibilities and domains of choice and freedom occurs when preference orderings (and ordinal utilities), which determine individuals’ choices on the possibility sets, are deemed irrelevant for macrojustice (and not when intensities of preferences only are).36

36

There are other solutions that extend solution 3 into Pareto-efficient solutions, but they use individuals’ preferences even more and have other intrinsic handicaps. One considers individuals’ allocations that are equivalent, for each individual, to her best choice in the common possibility set (a case of “equivalence theory” – see Kolm 2005, Chap. 25). Another rests on the property that individuals can choose their allocations on identical domains of choice if and only if no individual prefers any other’s allocation to her own (Kolm 1971) and extends it to efficient maximin based on

3

Economic Macrojustice

101

3.6.2 The Simple Case, Notations To begin with, we consider the simple case of unidimensional labour and constant individual wage rates (linear wage functions), because it is an important case, it simplifies the presentation a little, the concepts and results extend straightforwardly to the general case of multidimensional labour (duration, intensity, formation, etc.) and non-linear production (see Appendix A), and the general case can often be reduced to the simple case by defining a duration of labour qualified for its other characteristics (id.). The case of involuntary unemployment is considered in Appendix B. There are n individuals, and each is indexed by i and has labour `i (seen as duration), and hence leisure i D 1  `i by normalisation to 1 of the total relevant time, a given wage rate wi , and an a priori unspecified tax or subsidy ti (ti > 0 for a subsidy and < 0 for a tax of ti ). The problem consists in determining the ti that yield equal total liberty. Individual i’s labour income is wi `i , her disposable income used to buy freely (non-leisure) consumption is yi D wi `i C ti ; (3.5) and her total income, which adds the value of leisure at its market price wi , is vi D yi C wi i D wi C ti :

(3.6)

We now consider a balanced P distributive budget (Musgrave’s 1959 “distribution branch”), and hence ti D 0.

3.6.3 Solution 1: Social Liberty from an Equal Allocation 3.6.3.1 A Solution This solution is the classical (equal) social liberty from an equal allocation.37 Social liberty implies free exchange. The allocation is that of the two goods, leisure (or labour), and income which can buy consumption (from free exchange). Free exchange is, first of all, of labour for earnings. If this initial equal labour is k (leisure 1  k), it provides each individual i with the income k wi , and, if this income is transformed into an equal piece of disposable income with balance of the distributive budget and no comparisons of potential freedom by inclusion of domains (Kolm 1999). These other solutions are discussed in Chap. 12. 37 See Kolm (1971).

102

S.-C. Kolm

P waste, each now receives the average k w, where w D .1=n/ wi is the average wage rate. Then, individual i is taken away k wi and provided with k w instead, that is, she receives the net subsidy-tax ti D k  .w  wi /:

P

(3.7)

We have ti D 0. The described operation is Equal-Labour Income Equalisation (the equal sharing of the incomes produced by a given labour equal for all) or ELIE. Labour k is the “equalisation labour”. Individual i freely chooses her (full) actual labour `i and the corresponding earnings wi `i . Equivalently, this can be described as her choosing labour `i  k above labour k, and hence earning the corresponding wi  .`i  k/ in addition to the given k w (we have noted that, for the problem of macrojustice, `i  k happens to hold – see also Sect. 3.8.1). At any rate, her disposable income and her total income are, respectively,

and

yi D wi `i C ti D k w C .`i  k/wi

(3.8)

vi D wi C ti D k w C .1  k/wi :

(3.9)

3.6.3.2 First Properties Formulas (3.7), (3.8) and (3.9) show remarkable properties in themselves. Form (3.8) shows that each individual income is made of two parts, an egalitarian part in which all individuals receive the same income k w for the same labour k, and a classical liberal part in which each individual i receives the full product of her extra labour (`i  k) at her wage rate wi , (`i  k/wi . The equalisation labour k is the cursor making the division between these two parts. Moreover, formula (3.8) shows that yi is close to k w if wi is small, whatever `i . At any rate yi  k w if `i  k, which happens to be the case relevant for macrojustice: there is a minimum income k w (hence a consensus about a minimum income implies a consensus about coefficient k, given that the properties that imply the structure ELIE are generally wanted). Formula (3.7) shows that this distributive scheme amounts to a universal basic income k w financed by an equal labour k of all individuals, or according to capacities (each individual i pays her earnings for this labour, k wi , which is also in proportion to her capacities wi ). The way in which the result has been obtained shows that it amounts to each individual i yielding to each other the sum k wi =n D .k=n/wi , that is, the proceeds of the same labour k=n. This is a general equal labour reciprocity. Formula (3.9) shows that an individual’s total income is the weighed average between her productivity wi and average productivity w, with k and 1  k as weights.

3

Economic Macrojustice

103

3.6.3.3 Rawls’s Final Solution In 1974, John Rawls, at the instigation of Richard Musgrave (1974), added leisure to his list of “primary goods”, thus bringing to two, income (related to wealth) and leisure, the economic primary goods.38 Rawls’s solution consists of basic liberties, the best description of which is social liberty which is full and hence equal for all and maximal, and an ideal of an equal initial allocation of primary goods insofar as this is not wasteful. The above solution consists in an initial allocation in which all individuals have the same quantity of each good, 1  k for leisure and k w for income, from which each individual freely trades labour for income in application of social liberty. No individual can have more of one good in her initial allocation without any other initial allocation of any good to any person being lower, and the final outcome is Pareto efficient. It seems, therefore, that this result may be said to be Rawls’s full solution (the solution he should have proposed for the distribution problem as he posed it after 1974 if the weak points are corrected).39 3.6.3.4 The Geometry of ELIE The result is shown in Fig. 3.2, with coordinates i and yi , `i D 1  i , budget lines with slopes wi , transfers ti and total incomes vi . The initial equal allocation is the point common to all budget lines K.`i D k; yi D k w/. When k varies from 0 to 1, point K describes the segment LM from point L.`i D yi D 0/ to point M (i D 0; yi D w) – however, only cases with k  `i turn out to be relevant for macrojustice. The particular case k D 0, and hence ti D 0 and yi D wi `i for all i, corresponds to the full selfownership of classical liberalism (this is for example the position of – among scholars – F. Hayek, M. Friedman, R. Nozick, and J. Locke). The choice of the coefficient or “equalisation labour” k is considered in Sect. 3.9.1.

38

Musgrave’s advice aims at avoiding maximising income without considering labour and leisure. However “free time” (time free from labour) could be considered a primary good in Rawls’s basic sense. 39 Coefficient k reflects the relative moral/social value attached to these two primary goods, and the choice of such a weight is a classical Rawlsian problem. However, Rawls’s thought is that, in addition, a relatively high level of k would be chosen in a “well-ordered society”. Such a society is constituted by persons having a shared political culture and choosing the rules of justice of their society by consensus in a dialogue, while being actually submitted to these rules (a condition of stability – see the theory in Kolm 2009c).

104

S.-C. Kolm

v2 w

M

v1 w1 y1 kw

K

t1

λ1

L

0

k 1

t2

Fig. 3.2 The geometry of ELIE

3.6.4 Solution 2: Socially Free Individuals are Susceptible to Choose an Equal Allocation Individuals who have social liberty and prefer higher income (consumption) and leisure choose an allocation on their budget line. If there is one individual allocation that they all are thus susceptible to choose, these lines pass through the same point representing this allocation.40 Equation (3.6) with some given ti represents this budget line for individual i, and if this common point is `i D k (i D 1  k) and yi D , it entails  C .1  k/wi D wi C ti

(3.10)

 D k w i C ti :

(3.11)

or P

For a balanced distribution ti D 0, and summing (3.11) for all i implies  D k w, hence form (3.7) for ti .

40

This form is a crucial axiom in Maniquet (1998).

3

Economic Macrojustice

105

3.6.5 Solution 3: Identical Domains of Choice 3.6.5.1 Properties If individuals’ choices include the choice of effort or labour and they have different capacities, and if the policy maker does not take individuals’ preferences into account, presenting identical domains of choice to all individuals violates both Pareto efficiency and social liberty (and hence it should be impossible in a democracy and it violates the basic rights).41 Consider, indeed, the five conditions: (1) Individuals freely choose in identical domains of choice. (2) They do not all have the same productivity. (3) Their preferences or utilities are irrelevant or unknown to determine the domain of choice. (4) Pareto efficiency. (5) Social liberty. Then, the two following results hold: (1) Properties (1), (2), (3), and (4) or/and (5), cannot hold jointly. (2) Properties (1), (2), and (4) or/and (5), may not hold jointly. 3.6.5.2 Proof of Result (1) The proof results from the conditions necessary for building such a common domain of choice. In the space of leisure or labour and disposable income (consumption), at an achieved state, (1) Pareto efficiency and social liberty imply that each individual’s marginal rate of substitution is equal to her marginal productivity (wi ); and (2) because this individual freely chooses in the domain offered to her, this state is on the domain’s border B and the marginal rate of substitution is equal to the border’s rate of transformation. Hence, at this state this latter rate is equal to the individual’s marginal productivity. If these productivities are identical and constant, this border can be a straight line with this slope. If not, this border should respect the following condition. Call Ei the “curve” (more generally, set of points) where individual i’s rate of substitution is equal to wi (an Engel curve). Then, border B should cut each Ei at a point where its slope should be wi (wi if the variable is leisure). This condition depends on the curves Ei , which are derived 41

This is for instance done by proposals of equality of opportunity understood as identity of possibility sets.

106

S.-C. Kolm yi

u2

E2

−w 2

B

E1 D u1 −w1 0

λi

1

Fig. 3.3 Identical domains

from the individuals’ preference orderings or utility functions. This border, and hence the common domain, cannot be built without these preferences or utilities. Figure 3.3 illustrates this condition.42 3.6.5.3 Proof of Result (2) A set of individual allocations can result from individual choices in identical domains if and only if no individual prefers another’s allocation to her

More precisely, in the space (i (or `i ), yi ), call D such a common possibility set, B its border limiting it towards larger i and yi , and t .i ; yi / the set of slopes of the tangents to B at point .i ; yi / 2 B (jt j D 1 if B is smooth). Call ui .i ; yi / individual i ’s utility function assumed to be increasing and differentiable, ui1 and ui2 its two first derivatives, and si .i ; yi / D  ui1 .i ; yi /=ui2 .i ; yi / the corresponding rate of substitution at point .i ; yi /. Denote ( i ; yi / for  all i the realised state. Pareto efficiency and social freedom imply si . ; y / D w . Individual i ’s i i i        free choice on D implies . i ; yi / 2 B and si .i ; yi / 2 t .i ; yi /. Hence, wi 2 t .i ; yi /. Call Ei D f.i ; yi / W si .i ; yi / D wi g individual i ’s relevant Engel curve. Therefore, B must satisfy the condition that, at its intersection with Ei , .i ; yi / 2 B \ Ei , one has wi 2 t .i ; yi /. If all wi were equal, any straight line with slope wi can be such a B, whatever the Ei . Yet, if not all wi are equal, the construction of B and D, to satisfy the condition, must take curves Ei into account, and, therefore, must take individuals’utility functions ui into account. Therefore, if B is built without consideration of the ui and the wi are not all equal, the result violates Pareto efficiency and social liberty, except fortuitously. Note that the various solutions correspond to various distributions. 42

3

Economic Macrojustice

107

own (Kolm 1971).43 Moreover, this latter property may be inconsistent with Pareto efficiency (Pazner and Schmeidler 1974, whose example is a case of the present simple model). Finally, social liberty with perfect markets implies Pareto efficiency.

3.6.6 Solution 4: Equal Liberty of Unequal Domains In order to define equal freedom of choice for different domains of choice, consider that domains can offer more or less freedom. Using these relations usually implicitly implies their transitivity, which we assume. Domains of choice are thus ranked by a (weak) ordering, the freedom ordering. This ordering will be assumed to be representable by an ordinal function, the “freedom function”, since this will suffice here. If D is a domain of choice (a set of possible choices), the freedom function F .D/ is such that, if D 0 is another domain, F .D/ D F .D 0 / if D and D 0 offer equal freedoms, and F .D 0 / > F .D/ if D 0 provides more freedom than D. (In particular, if the domains D and D 0 are identical, F .D/ D F .D 0 /). Let us apply this to the budget sets considered here. A generic individual can provide labour `  0, hence enjoy leisure  D 1  `  0, and consume consumption goods in amount y  0. Let us choose an arbitrary but given and fixed unit of account, for which the price of consumption goods is P > 0 (P D 1 if they are taken as this numéraire), and the generic individual’s wage rate and total income are W  0 and V  0, respectively. For a specific individual i, `, , y, W and V take the values `i , i , yi , Wi and Vi . An individual freely chooses her leisure  2 Œ0; 1 (and hence her labour ` D 1  ), and her consumption y  0, subject to her budget constraint Py C W   V

(3.12)

which defines her budget set, which is her possibility set or domain of choice in the space of y and . This set is classically characterised by the (total) income V and the prices P and W . The freedom function can be written, therefore, as F .V I P; W /: (3.13) If V , P and W are all multiplied by the same positive number, the budget set defined by condition (3.12) does not change. That is, function F is 43

Choices in identical domains clearly imply the absence of preferences for another person’s allocation (which the former individual could also have chosen); and when this property of preferences holds, each individual’s allocation is one of her preferred allocations in any set made of the individuals’ allocations and any allocations that no individual prefers to her own.

108

S.-C. Kolm

homogeneous of degree zero in its three variables V , P and W . Moreover, to describe market possibilities when incomes and prices can vary, the prices are usually summarised by a price index which is always taken as linear (as with the classical indexes of Paasche and Laspeyre and those derived from them). This index is the value of a specific bundle of goods in quantities which are the coefficients. Indexes derived from concepts of utility are irrelevant here since the intent is to describe a possibility of choice and utility functions are irrelevant. Write this index as  D ˛P C ˇW

(3.14)

where ˛ > 0 and ˇ  are constant numbers. One has F .V I P; W /  .V; / D .V; ˛P C ˇW /:

(3.15)

Function  is homogeneous of degree zero in its two variables V and  since multiplying V , P and W by the same positive number does not change the level F D  and multiplies the index  by this number. Hence, dividing both arguments of function  by  gives F D .V; / D .V =; 1/ D '.V =/

(3.16)

by definition of function '. Since functions F ,  and ' are ordinal and are increasing functions of V , V / is a specification of function ' (this is real (total) income, fittingly usually called purchasing power). Therefore, the V , P and W that provide equal freedom are such that V = D 

(3.17)

V D ˛P C ˇW:

(3.18)

for some given  , or Hence, individuals i with possibly different wage rates Wi have the same freedom if their total incomes Vi are Vi D ˛P C ˇWi ;

(3.19)

respectively. Hence, with real (i.e. in terms of consumption goods) wage rates Wi =P D wi and total incomes Vi =P D vi , vi D ˛ C ˇ wi

(3.20)

3

Economic Macrojustice

109

for all i. This implies that individual i receives the net real transfer

However,

P

ti D vi  wi D ˛ C .ˇ  1/wi :

(3.21)

ti D 0 entails .1  ˇ /w D ˛:

(3.22)

Then, denoting 1  ˇ D k, we have again (3.7) ti D k  .w  wi /: This is the same result as that of solutions 1 and 2. Moreover, individual i’s budget line in space (i ; yi ) is given in (3.6) wi i C yi D vi ; and it contains the point (`i D k; yi D k w) since .1  k/wi C k w D wi C ti D vi : This “equalisation point” K, independent of i, is common to all budget lines (which, therefore, constitute a “pencil” of lines). This result is a simple application of the general theory of the comparison and measure of economic liberty (Kolm 2005, 2009b).

3.7 Equivalent Properties and Normative Meanings Judging something can, and a priori should, be done according to its various properties. The obtained distributive scheme has in particular a number of characteristic (necessary and sufficient) properties or sets of properties, which have (more or less) different meanings (the key issue). Each can be taken as the scheme’s definition, and as its justification (or it can participate in it). Looking at the result from these different angles is necessary for fully “understanding” and finally evaluating it.44 There are more than twenty such different (although logically equivalent) meanings, which regroup into several types of issues.

44

The requirement that a principle should be evaluated from all its angles and possible meanings is a classical and basic meta-principle of social ethics, related, for instance, to Plato’s “dialectics” in Republic and to Rawls’s “reflective equilibrium”.

110

S.-C. Kolm

3.7.1 Equal Liberty The previous remarks have shown the following properties of the result. 1. 2. 3. 4.

Social liberty from an equal allocation. Susceptibility to choose some equal allocation with social liberty. Equal freedom of choice (for possibly non-identical domains). Rawls’s solution with leisure (post 1974 and completed).

3.7.2 ELIE A few other notable aspects are straightforward. 5. Equal-labour income equalisation: Redistribute equally the product of the same labour k of all individuals. k is the “equalisation labour”. 6. Equal pay for equal work, for labour k (the rate is the average wage rate w). This is one of the most widespread claims of justice. However, it refers here to differences in productivities. 7. From each according to her capacities, to each equally (where “according to” is taken to mean, as it most commonly does, in proportion to): take k wi proportional to wi and give the same k w. This associates two of the most widespread claims of justice. 8. Everyone works for everyone for the same labour (k) and for herself for the rest.

3.7.3 Responsibility and Brute Luck 9. Individuals are entitled to what they are responsible for, their labour and its product (`i and wi `i ). What is redistributed and taxed is what they are not responsible for, their given capacities (wi ), to some extent (k): yi D wi `i  k  .wi  w/:

3.7.4 The Two-Part Income: Egalitarian and Classical Liberal, Deserts and Merit, Work and Works Writing (3.8) as

yi D k w C wi  .`i  k/

3

Economic Macrojustice

111

has shown a decomposition of income in two parts induced by two different and opposed ethics, which can be seen in various ways. 10. Equality and classical liberalism. The two parts are an equal income k w and the market remuneration wi  .`i  k/ of labour `i  k. These are the two basic and opposed principles of overall distributive justice in our world. The level of coefficient k favours one or the other and delimitates their respective scopes. 11. Each earns according to deserts for labour k and to merit for the rest. Deserts is according to labour or effort, here k for the share kw. Merit means according to labour or effort and to capacities. This is the second part with individual labour `i  k and capacities wi . 12. To each according to her work (effort, input) and to her works (product, output). This classical distinction refers here respectively to k w in proportion to work k and to the individual’s product wi  .`i  k/.

3.7.5 Financed Universal Basic Income 13. Equal universal basic income financed by equal labour (equal sacrifice): The result ti D k w  wi k can be seen as providing the same basic income k w to each individual, and financing it by the same labour k from each (individual i pays the proceeds k wi ). 14. Equal universal basic income financed according to capacities (i.e. in proportion k wi of wi for individual i). A universal, unconditional and equal basic income has often been proposed by scholars and political figures. The Achilles’s heel of such schemes is the specification of their financing which should be sufficient and fair, and should not induce Pareto inefficiency. ELIE satisfies these conditions. The fairness cannot be an equality in money terms since this would cancel out the distributive effect. Hence, if capacities to enjoy are deemed irrelevant for this issue of macrojustice, the only remaining possibility is equality in labour provided.

3.7.6 Reciprocity A basic principle of fairness is reciprocity (in the framework of macrojustice, this is emphasised by Rawls).

112

S.-C. Kolm

15. General equal labour reciprocity: Each individual hands out to each other the proceeds of the same labour (r D k=n). Indeed, the ELIE operation amounts to equally sharing the proceeds k wi of each individual i’s labour k, hence to yield to each individual the proceeds .k=n/wi of the labour k=n of each individual i (and what an individual yields to herself can be discarded). That is, X X wj  n r w i D r wj  .n  1/r wi : (3.23) ti D k  .w  wi / D r j ¤i

This property has an aspect of fairness which is bound to be favourable to the acceptance of this scheme from sentiments of reciprocity.45 16. Each owns the rent of the same amount of each other’s capacities (r).

3.7.7 Progressive Transfers, Total Concentration ELIE belongs to the question of reducing inequalities, in a particularly meaningful and straightforward way (see also note 50). 17. Equal partial compensation of productivity differences: Each individual yields to each less productive individual the same fraction of the difference in their productivities, r  .wi  wj / from i to j if wi > wj . It suffices to consolidate the two transfers of the general equal reciprocity in each pair of individuals. Hence, ELIE amounts to a set of “progressive transfers” for total incomes. This set is, in fact, quite specific (Property 18). 18. Each individual’s total income is the weighed average between average productivity and this individual’s productivity, with weights k and 1k, as shown by (3.9) i D k w C .1  k/wi : 19. A concentration of total incomes: This formula also says that the set fvi g is a uniform linear concentration towards the mean of the set fwi g, with degree k. This structure of transformation of a distribution is that which can be said to be the most inequality-reducing.46

45 46

See Kolm (1984, 2006, 2008b). See Kolm (1966a, 1999b).

3

Economic Macrojustice

113

3.7.8 Tax Structure and Reform The fiscal structure and reform that realise ELIE are very simple, clear, natural, easy to implement, and made of a few elements each of which is classical. 20. An exemption of overtime labour over some given labour, and an equal tax credit or rebate, from a flat tax. Indeed, the transfer can be written as the net tax  ti D .k=`o /wi `o  k w

(3.24)

for some given labour `o chosen such that `o  `i for the chosen labours `i relevant for macrojustice. The first, positive, term is the flat tax with rate k=`o on the earnings wi `o of labour `o , hence with a tax exemption of the corresponding overtime earnings of labour `i  `o . The second term is the tax credit or rebate kw equal for all. This tax structure is simple, clear, with two bonuses – an exemption and a rebate. For example, the tax exemption of overtime labour over a low duration is the new general law in France, which has also the equivalent of a universal equal rebate (resulting from an income tax credit). 21. Tax reform. The ELIE distributive structure can be obtained from actual income taxation by a series of a few simple and rather classical tax reforms: – A negative income tax or income tax credit for low incomes, which exists in many countries. – Replace actual labour by a given labour in the tax schedule, which is obtainable by exempting earnings over a given labour not exceeding actual (full-time) labours. – Flatten the tax schedule, which is often advocated for a reason of simplicity (and incentive)47 – an ELIE scheme can a priori be made as redistributive as one wants by choosing a sufficiently high coefficient k. – If the scheme concerns the “distribution branch” in “functional finance”, balance the budget. Formally, from the income tax on labour income f .wi `i /, one thus successively obtains, with constants a > 0, b > 0, c, and `o >P0 : f .wi `i / < 0 if wi `i < a; f .wi `o / or bwi `i C c; bwi `o C c; and, if f .wi `i / D 0, o o bw` C c D 0 and hence, noting b` D k, k  .wi  w/ D ti . 47

A flat tax is for instance implemented in all Eastern European countries including the nine fastest growing countries of the European Union.

114

S.-C. Kolm

3.7.9 Other Meanings 22. Bi-numéraire equal sharing of the value of productive capacities. An amount of a productive capacity (with a given productivity) can be measured by the labour that can use it (or time of use), or by the output it can produce. In an equal sharing, the choice of this measure makes a difference because individual productivities differ. If an amount of an individual’s productive capacities is measured by the labour input that can use it, each individual has initially 1 and the given allocation without any transfer is equal. If this amount is measured by the output it can produce, however, the total initial endowment of individual i is wi . Both goods – income-consumption and leisure-labour-lifetime – can be taken as numéraire. Amounts of both are classically compared across individuals. The general solution consists in measuring a fraction of the capacities, say k, in income-value, and the rest, 1  k, in labour-value. For individual i, the equalisation of the first share transforms income k wi into k w, and the second share is already equal for all in labourvalue, 1  k. The result is the net income transfer ti D k  .w  wi /. One can also directly write the total income of individual i from the two parts, vi D k w C .1  k/wi .48;49

3.8 Real Gains, Incentive Compatibility 3.8.1 Irrelevance of Non-Realised Advantages As we have noted, a concentration transformation of a distribution is, in a sense, the most inequality-reducing transfer structure. Hence, the inequalityreducing effect of a redistribution is meaningfully measured by the coefficient

48 With ELIE as the solution of Rawls’s full problem, k thus measures the relative importance attached to the two economic primary goods: income relative to leisure-labour. With the measure in labour value only, equality is satisfied by full self-ownership which is classical liberalism, but is also Marx’s view (he defines “exploitation” as theft of this property by low wages). 49 ELIE has other interesting and meaningful properties. For instance, Maniquet (1998) derives, from a number of basic axioms, a state which is about the one chosen by the individuals submitted to such a distributive scheme. Moreover, it is securing that ELIE can be derived from the most famous general presentation of principles of justice, that of Plato (Laws) and Aristotle (Nicomachean Ethics), with each person receiving the fruit of her labour wi `i in “commutative justice”, and an equal share (with the appropriate measure) of what is given to society in “distributive justice”, achieved by compensatory transfers since their capacities are attached to the individuals (“diorthic justice” – see Kolm 2005, pp. 248–249).

3

Economic Macrojustice

115

of the concentration which produces the same effect on some measure of inequality. For a redistribution and an inequality index, the “equivalent ELIE” produces the same “decrease” in inequality in total income: its k is the degree of inequality reduction or equalisation of this redistribution.50 Consider now the three following facts and judgments. (1) Present redistributions in nations amount to equally redistributing the incomes of 1–2 days per week (from the USA to Scandinavia). Hence, de facto – even for the most redistributive policy a country could actually achieve –, for normal full-time labour one has `i > k (the cases of total or partial unemployment are the object of Appendix B and the equivalent treatments of part-time labour and second wages in families have been noted). Actually, we have seen that in fact, extremely few people declare working less than 20 hours a week. (2) Moreover, people commonly understand that individuals who benefit from a high wage rate be taxed to help people who are not as lucky, but not when this advantage remains a mere unused potential possibility of income. Precisely, people do not agree with a tax on earning capacities that entail no earning because they are not used, that is, with a tax on leisure measuring its value by the earnings this time could provide were it used at labour (taxing to induce work is something else and has to be justified). ELIE with k > `i would so imply, when demanding the amount k wi , demanding the value of leisure (k  `i ), (k  `i /wi , in addition to the value of the whole product wi `i (for equally redistributing the proceeds). If the redistribution of k w is jointly taken into account, this would imply demanding .k  `i /.wi  w/ on leisure (k  `i ) for wi > w, in addition to (wi  w/`i . If wi is quite low, the tax kwi is negligible and ti and yi are both about equal to kw, whatever `i . If wi < w remains substantial, and `i < k, people would again not agree with taxing leisure (k  `i ) at unit value wi for the share .k  `i /wi of the tax k wi (then equally redistributed). If the subsidy k w is taken into account, people would correspondingly find absurd to subsidise the unused and inactive productive capacities in leisure (k  `i ) because they have a 50

This degree of inequality reduction of a redistribution is equal to the relative decrease in the absolute form of any synthetic index of inequality (Kolm 1966b). Indeed, P for any distribution of incomes (or other quantity) xi whose set is x and average x D .1=n/ xi , one can, for an index of inequality, distinguish the absolute form I a .x/ and the relative form I r .x/ D I a .x/=x. A synthetic inequality index is by definition such that I a .x/ is equal-invariant (invariant under any equal variation of all the xi ) and I r .x/ is intensive (invariant under any multiplication of all the xi by the same number). Then, the absolute form is also extensive (linearly homogeneous). A concentration of coefficient k of the distribution amounts to an equiproportional decrease of all xi in proportion k, which similarly decreases the absolute index, and an equal increase that restores the total sum or the mean, which property. Examples P Pdoes not affect this index. Hence the notedP of such indexes are .1=n2 / i j max.xi  xj ; 0/ (absolute Gini), .1=n/ jxi  xj, and the standard deviation.

116

S.-C. Kolm

relatively low potential productivity wi < w, by the part (k  `i /.w  wi ) of the subsidy k.wwi /. In brief, people do not find it normal to tax leisure and find it absurd to provide a wage complement for hours that produce no wage. Hence, this opinion implies that people who pay an actual distributive tax k wi and receive k w as counterpart are people who choose to work `i  k. This common view has to be obeyed in a democracy. (3) The very few productive individuals who choose to work very little mostly choose not to benefit from society’s supply of a favourable wage, and hence arguably do not have to be taxed for this advantage. They choose to drop out of the cooperative venture of collective production (and division of labour), from its advantages, and, hence, from its liabilities. People who choose not to contribute to this joint venture whereas they could may not be entitled to a reciprocal share of the product. These fugitives from production are not, as Rawls (1982) puts it, “fully cooperating members of the society engaged in social cooperation over a complete lifetime for mutual advantage”, and hence are not party in the sharing of benefits. These last two points mean that what would be at stake concerns actual advantages that people actually derive from their productive capacities and society’s demand for them, rather than these capacities and demand per se – hence as potential earnings. The cases in which the chosen `i is lower than k are particular cases: partial or full unemployment, the few eccentric productive people who drop out of cooperative social production, victims of particular handicaps, part-time jobs which are often second wages in families, etc. These particular cases deserve particular criteria and treatments. They are, therefore, out of the scope of overall distributive justice in macrojustice. However, some can also be more or less brought back into the general case, as with involuntary unemployment (Appendix B), the case of people with capacities without market value (wi D 0), and the tax treatment of part-time labour contracts and of households (as in the present French tax law). The case of the tiny fraction of people – if any – who could earn high wages for a moderate effort but decide to live “on welfare” if they can is not a concern for macrojustice for three sets of reasons: the noted ethical reasons and opinions entailing k  `i ; this is a particular situation (out of the definition of macrojustice); and its rarity (not an issue for overall justice). These work evaders are the object of classical other proposals and discussions.51

51 These are, for example, people who can earn ten times the average income for some standard labour but would prefer to stop working and live on – for instance – 1/5 to 1/3 of average income. For the very few able people who choose to work very little, there are three classical proposals. (1) They should earn their sandwich, “he who does not work does not eat” (Saint Paul), the solution endorsed by Rawls. (2) They should have a “right to laziness” (Paul Laffargue) and perhaps

3

Economic Macrojustice

117

Finally, for all these related reasons, distributive macrojustice is concerned with normal full-time labour, or at any rate with `i  k, only (including the cases which are assimilated to such a situation by the appropriate theory or device). One consequence is that, for macrojustice, yi D wi `i C k  .w  wi / D wi  .`i  k/ C k w  kw:

(3.25)

That is, there is a minimum income of k w.52 As noted, the case k D 0 is full self-ownership. A case of k D 2:5 days a week for a nation would correspond to a very high redistribution (there can, in addition, be various policies of more specific microjustice).

3.8.2 Incentive Compatibility and Information If wi denotes the highest wage rate individual i can obtain, this individual can also generally earn various rates w0i < wi by not using her best (most highly paid) skills at work.53 She may make such a choice if she thinks that

receive a basic income (utilitarianism may support this position, which is eloquently defended by Van Parijs 1995). (3) We may try to persuade them that they should make other people somewhat benefit from the talents endowed to them by nature, providence or their parents by working a little (at a high wage rate). If their productive capacities are due to subsidised public education which they accepted, they might be asked to refund this cost to the rest of society. If they had to pay for their possible advantage in earning capacity, they would pay ti D k  .wi  w), for which they should work k  Œ1  .w=wi / < k; however, if they still choose `i < k, we will see that they may have an interest in hiding their skills and their value wi (yet, diplomas, previous jobs, etc. often make some estimate possible and E. Ooghe and E. Schokkaert show that, at any rate, the resulting waste would be very small). Finally, sheer coercion might be restricted to the limited (and possibly highly remunerated) draft of exceptional talents indispensable to society or to other people’s life. Note that freedom of choice should a priori refer to the full domain of possible choice in the space of income and leisure rather than to a subset of it only – such as the case `i D 0 put forward by solution (2), which makes the corresponding basic income be “freedom from labour” rather than general “real freedom”. Moreover, there are other distributive units than nations; for instance, transfers are intense in a family, but they are gifts rather than taxes (each likes the others’ enjoyment and consumption). 52 One consequence is that, in a society, since w is given, choosing a minimum income and choosing a level of equalisation labour k amounts to the same – given that the structural properties that lead to ELIE happen to be largely wanted (social liberty, Pareto efficiency, non-welfarist macrojustice). The frequent rough consensus about a minimum income implies the same convergence of views about coefficient k. This relation is more valid the more the minimum income refers to a norm of income (and consumption and lifestyle) rather than to the alleviation of physical suffering (which may elicit relief provided by microjustice policies). The differences in the opinions about a minimum income and the evolutions of these views are the object of other studies (e.g. Kolm 1974a), and they can be enlightened by the theories of the determination of coefficient k in Part 4 of Macrojustice. 53 See Dasgupta and Hammond (1980).

118

S.-C. Kolm

the fiscal authority bases taxes and subsidies on this actual and observed w0i , in order to diminish the tax or transform it into a subsidy if wi > w, or to augment the subsidy if wi < w (hence she would benefit whatever w if k > 0, and therefore she need not know w to behave this way). The individual may think that the government would take the observed w0i as base either because it deems the actual wage rate to be the appropriate basis for the reasons presented in the previous section (not taxing or subsidising unused capacities of value (wi  w0i )), or because it mistakes it for the value of capacities wi , or for any mixture of these reasons. Individual i thus chooses both labour `i and skills that earn w0i wi , that maximise some increasing ordinal utility function P

ui Œ1  `i ; .`i  k/w0i C k w0 ;

(3.26)

where w0 D .1=n/ w0j .54 Variables `i and w0i are independent. The derivative @ui =@w0i has the sign of `i  k C k=n if individual i takes the w0j for j ¤ i as given (no collusion), but whatever they are. Therefore, individual i chooses w0i D wi if `i > k  Œ1  .1=n/. This is the case for macrojustice in which `i  k (see the previous section). Hence, the individuals choose to work with their best skills and thus to “reveal” their capacities and to exhibit their economic value. The government can understand this (it need not know individuals’ utilities, but only that individuals prefer higher disposable incomes for given labour). Hence, it need not raise questions about basing its taxes and subsidies on the actual values of capacities wi or on the observed wage rates w0i since using the latter as base makes them be the wi . And the individuals can in the end know this conclusion.55

3.9 The Degree of Redistribution and Public Finance 3.9.1 The Degree of Redistribution Coefficient k, technically the equalisation labour, is a degree of redistribution, equalisation, and solidarity with regard to the unequal endowments of 54

Choosing a more remunerated but more painful or disagreeable activity, or the contrary, is considered as working more or less, and a corresponding full analysis has to consider, in a framework of multidimensional labour (see Appendix A), the relevant dimension(s) that affect both the productivity and the painfulness or intrinsic attractiveness of labour. 55 If the government used the wi if it could know them, with ti D k .wwi /, and each individual i could choose her skills used and w0i wi , her income would be `i w0i C k  .w  wi /, and she would also choose w0i D wi if she chooses to work at all (`i > 0) and hence when `i > k.

3

Economic Macrojustice

119

productive capacities. The value k D 0 corresponds to full self-ownership and an absence of redistribution from it, and redistribution increases with k. Specifically, k is a degree of common ownership of the value or rent of given productive capacities (and 1  k is a corresponding degree of selfownership) – and this commonly owned part is equally shared for lack of relevant other differences among individuals. Coefficient k also has the various important meanings derived from the various meanings of an ELIE distribution (Sect. 3.7). The level of coefficient k has been derived, in Sect. 3.3.1.4, from strict welfarism, that is, welfarism cleaned for differences in individuals’ tastes and hedonic capacities which people find irrelevant for macrojustice. The result rests on the choice of a social welfare function, however. Yet the rich tangible meanings of coefficient k make it particularly susceptible to be the direct object of social ethical evaluations. The structure of ELIE has been derived from properties which are essentially wanted by all for macrojustice. Could this also hold for the level of coefficient k, given that it has opposite effects on the interest of individuals depending on whether their wi is above or below the average w (since ti D k  .w  wi /)? In any instituted society, it is largely held that people with insufficient means and earning capacities should be helped by some redistribution. More precisely, in a given society, there often is some kind of rough consensus about what a standard minimum disposable income should be. As we have noted, since this level is k w with ELIE and w is given, such a common view determines a coefficient k (the poor can also benefit from more specific measures of microjustice).56 Moreover, in a number of peaceful societies the overall level of income redistribution is generally directly more or less accepted or approved of (with particularities due to the fact that it has the nature of a public good), or the various standard opinions in this respect vary in a relatively limited range. Then, the coefficient k of an ELIE equivalent to the actual redistribution (the degree of this redistribution, see Sect. 3.8.1) provides an answer. Reforms towards this ELIE structure can de facto benefit everybody, as we will see. However, this level of redistribution also often evolves, and this is done more efficiently and in accordance with common views if the distributive structure also evolves towards an ELIE scheme. For more direct inquiries, however, although the opinion of an individual “small in a large number” has in itself no actual influence – and hence 56

This is why, as noted, the standard minimum income particularly plays this role of revealing a consensual k of an ELIE distribution when it refers to a norm of income or consumption rather than to physical sufferings of poverty which provides classical reasons for various insurance schemes and specific aids in income or in various goods or services.

120

S.-C. Kolm

no influence on this individual’s self-interest –, people’s expressed views are often influenced by their interest, even though people also have a socialmoral judgment (the view of the “impartial spectator in their breast” as Adam Smith 1759 puts it). However, ELIE provides a neat possibility of obtaining people’s social-moral views cleaned from their self-interest. It suffices to consider the opinion of individuals with an average wage rate wi D w. Indeed, for them ti D 0 whatever k: their interest is not affected by the level of k. Their opinion about this level thus a priori only expresses their impartial social-ethical view. This would a priori provide an unbiased sample of these views in society. Individuals’ social ethical views are a priori globally closer to one another than their interests in questions of distribution (less polarised for an ELIE), because they are altruistic and because they are impartial (by nature and definition of a conception of justice).57 Nevertheless, they may differ. However, these views depend on the various influences the individuals have been submitted to, their life experiences, their reasoning – and, possibly, some given sensitivity. Hence, they a priori become more alike when people are informed about the others’ arguments and know vividly about their experiences. The means are essentially information and social dialogue. This has practical limits, but divergences can be reduced by showing the results of a number of analyses. Some are complete theories of impartial views such as the “recursive” “original position” or “moral time sharing” (each individual assumes she is each individual with the same probability or successively in time, and these views are submitted to a similar process recursively). A theory of fair dialogue in which each individual has equal power to influence others (as with the Athenian isegoria) shows the convergence to what Habermas calls the “ideal speech”. In an equilibrium surplus theory of transfers, the money values of self-interests cancel out and only moral values determine the result. Individuals’ own impartial views can be derived from observed conducts and preferences and extrapolated. And so on. These are theories to determine the just distribution from society’s conception.58 The distributive coefficient k depends on the society in which this distributive policy takes place. It expresses the extent to which this society considers itself a community of resources and solidarity. We have noted the levels of k of the ELIE equivalent to the present-day national distributions. These actual distributive policies are often not based on the less elastic possible items and also generally induce other waste. Simply reforming them – notably

57

See Kolm (2005, Parts 4 and 5). All these analyses, others for the same purpose, and their results, are presented in Part 4 of the volume Kolm (2005).

58

3

Economic Macrojustice

121

the income tax and the main aids to low incomes – with everybody gaining at each step can be done towards an ELIE with a similar coefficient k.59 However, the social and political dialogue about the degree of community, solidarity and redistribution will go on. Moreover, there can be, and often are, various communities of redistribution for the same person – for instance at levels of a region, a nation, or supranational (e.g. the European Community). Then, there can be an ELIE and a k for each community, with a net addition of the transfers, and possibly some evolution and shift in time of the responsibility for distribution.

3.9.2 Place in Public Finance If distributive justice is achieved by such a policy, the financing of other public expenditures should a priori be by the method that is neutral in this respect, benefit taxation.60 This is the classical budget optimisation by “functional finance” (e.g. Musgrave 1959). A number of services can then be associated with their financing, and they can be given financial and hence managerial autonomy, which is often favourable to efficiency. The users’ benefits are more or less estimated by the usual benefit-cost analyses of public expenditures, but this is sometimes difficult. Other principles of financing are also classically proposed. One of them is taxation “according to capacities” which, for earned income, should be capacities to earn, i.e., the tax is in proportion to the wi . Another principle is “equal sacrifice”, which, if it does not simply mean equally in income, and if this distribution of tax liabilities aims at macrojustice which discards capacities to be pleased, should be equal sacrifice in labour. These two classical principles are in fact equivalent: each individual i pays wi L in which L is both the coefficient of proportionality and the equal labour. This is in fact how ELIE finances the basic income k w. Each taxpayer i then pays the product .k C L/wi of her labour k C L, the same for all, and she receives the amount k w plus the benefit of other

59

This is a factual result suggested by numerical examples rather than a theoretical necessity since ELIE solutions are only a subset of the Pareto-efficient states. It is in particular shown that ELIE schemes can supersede all present-day supports to low incomes with everybody benefiting (Kolm 2005, pp. 118–122, and see Chap. 12 of this book). 60 With some rule for allocating the surplus for public goods (possibly the outcome of a fictive and implicit exchange or agreement for respecting the spirit of social liberty – a “liberal social contract”, Kolm 1985, 1987, 2005, pp. 67–69).

122

S.-C. Kolm

public expenditures.61 Of course, all these financing principles can be jointly present, for various types of public expenditures.

3.10 Conclusion Facts and necessary or commonly held judgments – Pareto efficiency, social liberty, private accountability of tastes and hedonic capacities for macrojustice – have been shown to imply a macrojustice policy which is simple, clear, understandable, richly meaningful, made up of fiscal properties actually used, more easily implementable than present taxes and aids, and which can be installed progressively or rapidly by largely supported reforms. Its structure amounts to several distributive principles and policies which are logically equivalent but have different and very important social meanings: equal social and real liberty; an exemption of overtime labour earnings and a given tax credit; a universal basic income financed by an equal labour of all; an equal sharing of the proceeds of the same labour of all; each yielding to each other the product of the same labour; and a number of other meanings presented in Sect. 3.7. This is complemented, when needed, for possibly remaining specific issues of microjustice. Implementation can rest on both the obtained theoretical properties and the actual experiences of application of aspects of this scheme. These experiences include tax exemption of overtime labour (over a rather low benchmark), minimum incomes realised in various ways, a tax equal to the earnings during a given period, exemption of productivity and formation premia (for the intensity and formation dimensions of labour), and, less important, uniform tax rates. The exemption of overtime labour over a rather low benchmark and the experience of it is particularly important. For the rest, the various routine procedures of estimation of fiscal administrations are used. From the economic point of view, the tax base suppresses the elasticity due to labour supply and demand for most dimensions of labour, hence a priori it improves efficiency, and this can practically be translated into a performance more favourable for everyone than other actual or proposed distributive schemes with the same degree of equalisation. This favours political implementability. The ongoing social debate can then focus on the degree of solidarity appropriate for the society in question and its evolution, by considering its various related practical aspects (minimum or basic income, 61 If b is the per capita non-ELIE budget, L D b=w and each individual i pays bwi =w of it. In Fig. 3.3, point K moves to the left by L (if people also pay some equal amount, this moves point K down by this amount), and the slopes of budget lines remain the same.

3

Economic Macrojustice

123

comparing tax liabilities measured in labour, levels of tax burdens, relevant measures of inequality, etc). The obtained ELIE distributive structure relates to a number of existing or proposed ones. We have seen the realisations of exempting overtime labour, paying the earnings of a period, and minimum incomes. The basic income is discussed in many scholarly and political circles, with the problem of finding an efficient, sufficient and just financing (this can be the proceeds of an equal labour of all). One such financing proposed is a flat tax; this amounts to Mirrlees’s (1986) final proposal of a flat income tax with a negative part, and is studied by Atkinson (1995) – ELIE only adds exemption of overtime income above some given labour.62 All reforms that tend to base taxes or aids on less elastic items a priori go in the proper direction, and much is possible in this respect. Moreover, if, as Kenneth Arrow (1963) proposes, “The fundamental function of any theory of social welfare is to supply criteria for income distribution”, the ELIE tax-subsidy structure constitutes a solution to this general problem too. The issue is that if “social choice” is derived from “individual values” – as Arrow’s title suggests – and individual values all discard tastes and hedonic capacities for this problem of macrojustice, this social choice does the same. In fact, a large “overlapping consensus” (Rawls’s term) of values points to the solution described here. These values are varied, but they include prominently efficient associations of equality, liberty and equal liberties. Other schools of economic thought acknowledge the basic importance of freedom. Classical liberals (such as F. Hayek and M. Friedman) advocate full self-ownership (k D 0), but they justify it by social (defensive) liberty which they think forbids transfers, whereas ELIE transfers can be seen as the given basis from which this freedom develops, these transfers happen to give more de facto real freedom to people who pay higher taxes, and they can ultimately simply be the necessarily public implementation of the free joint giving or aid entailed by benevolence and duty. Another school of economic thought (James Buchanan, Public Choice) rightly emphasises that public policy results from individuals’ preferences, but considers it to be the truce resulting from freedom to fight. However, people also want fairness and justice, they always defend their interest by appealing to some value – which suggests that such claims have

62

Hence, when students of uniquely “welfare”-determined income taxation face the problem that their refined and well worked-out second-best proposal is complex, not understood by the public and politicians who, at any rate, disagree with its ethics for this application, with a regressive tax for high incomes (Phelps 1973a,b, Sadka 1976 and Seade 1977 – this had been previously shown for non-linear tariffs of public utilities), informational and conceptual difficulties (utilities), and high administrative costs, they come to consider an intuitive pragmatic third best in the direction of the first best implied by standard moral judgements.

124

S.-C. Kolm

some effect –, their political opinions consist in principles, and the “social contract” between their interests for which they can settle has to be in terms they deem appropriate to define this fairness in their community.

Appendix A Multidimensional Labour, Nonlinear Production Labour has a priori various dimensions, such as duration, individual effort and costs in previous education and training, intensity (strength, concentration), speed, etc. Moreover, the output may not be a linear function of labour. Let `i denote a multidimensional labour of individual i, and pi .`i / the corresponding earnings.63 All the reasonings, results and meanings presented for the simple case can be repeated for this general case practically identically. The equalisation labour k is now multidimensional. The tax-subsidy is

where p.`/ D .1=n/

P

ti D p.k/  pi .k/

(3.27)

pi .`/, and individual i’s disposable income is

yi D pi .`i /  pi .k/ C p.k/:

(3.28)

This multidimensional case can often practically be reduced to a onedimensional case with labour duration adjusted for the other characteristics of labour. Indeed, labour can generally be considered as a flow, and as steady in some given period (which can be taken as short as one wants). Then, if `0i denotes the duration of labour `i and `00i the set of its other parameters, function pi can be written as pi .`i / D li0 qi .li00 /. If individuals’ particular productivities are of the classical “output augmenting” type qi .li00 / D ai f .`00i /, then pi .`i / D wi Li where Li D `0i f .`00i / is individual i’s “labour duration augmented for the other characteristics of labour”, and wi D ai is the corresponding competitive wage rate.64 In the expression of earnings from labour `i , pi .`i /, labour `i represents items chosen by individual i, and the function pi ./ the other items, that is, individual i’s productivity and the labour market. Formation, education and training (as health care) increase later productivity. They depend on the 63

For macrojustice, the effects of other persons’ labour on an individual’s earnings pass through the prices. 64 The educational input can also be taken into account by “spreading” the formation time on later labour (that uses its benefits) (see details in Kolm 2005, Chap. 8).

3

Economic Macrojustice

125

persons’ given capacities for learning. They also involve acts of the individual and possibly various costs for her (time, effort, direct costs, foregone earnings, etc.). However, the bulk of the formation and education received in the first period of life is provided by the family, or determined by it through choice, support, information, and induced motivation. Globally, at a macro level and apart from exceptions, individuals’ level of education is essentially a given sociological phenomenon. Hence, for macrojustice and as a first approximation, its effects on earnings can be incorporated in the productivity pi ./ or the wage rate wi under consideration. Individuals are generally more directly responsible for training and formation undertaken later in life.65 Note that the effects of different pi ./ or wi are equalised only for labour k and not for the rest of labour. This effect of the family should also be considered with the issue of bequest – its cost can be seen as a part of it.66 Both individual and family-induced educational choices could be sensitive to future taxation, but this is much attenuated by the fact that taxes decades later are very uncertain and by the non-pecuniary values of education as providing larger occupational opportunities and freedom of choice, jobs that are less painful and more interesting and gratifying, the status of educational level and occupations, culture, and the continuation of family traditions. Moreover, the cost of providing education is more or less publicly financed in many countries. D. de la Croix and M. Lubrano (this volume) propose to match the increased income tax due to more education by a subsidy to education, in a model which also permits the direct application of the ELIE principle to this input.

Appendix B Unemployment Situations of unemployment raise particular specific issues, but, given their importance, they should be related to the general results for macrojustice. If wi D 0, individual i’s labour is neither supplied for income nor demanded, and the formula ti D k  .w  wi / gives yi D ti D k w, the minimum or basic income. If wi is low, ti and yi are close to k w, whatever `i . These 65

A refinement of the analysis can also find ways of taking account of some individually chosen effort at the end of the educational period. 66 There is even a ground for compensating sociological differences more than those due to intrinsic individual capacities which belong to the person’s self, but this issue is not pursued in this simple presentation.

126

S.-C. Kolm

people’s actual labour level makes little financial difference.67 Hence, the general principle can be applied to these cases (apart from the other policies of formation, education, taking care of handicaps, etc.).68 In involuntary unemployment, the individual faces a constraint `i  `oi . It can be partial or total (duration zero). It can be for duration or for other dimensions (for instance as under-qualification for formation). Reasons for discarding cases `i < k from macrojustice may not hold any longer for this case: these people do not voluntarily abstain from participation in social production, and their number may not be small (or at least for `oi D 0). Of course, good macroeconomic policy in the first place, unemployment insurance, and specific policies about the labour market and formation are in order. However, the obtained distributive policy can have three important positive effects on employment. By basing taxes and subsidies on items less elastic than actual labour, it generally induces higher labour. The other two effects concern involuntary unemployment in the strict sense. First, the income support to people with low wage rates provided by the obtained scheme can supersede, to everybody’s benefit, a number of wage rigidities of public or private nature which are important causes of unemployment (minimum wages, “collusion” in any form, etc.).69 Second, the general results for macrojustice can also apply to the case of involuntary unemployment, by using the logical device of considering someone who cannot work more as someone who cannot earn more by working more (and works in order to earn). What the market presents to the individual is then described solely in terms of the remuneration of each labour (however, for partial unemployment it cannot be a linear function of labour). Considering one-dimensional labour for simplicity in presentation, the outcome is that someone involuntarily unemployed at `oi  k (in particular totally unemployed) has income p.k/ Q which derives from the average P p.k/ D .1=n/ pi .k/ by replacing the pi .k/ of such individuals by pi .`oi / (0 for full unemployment). This results from the application of the noted device by replacing the function pi .`i / by its truncation at `oi 70 : Pi .`i / D pi .`i / if `i  `oi and Pi .`i / D pi .`oi / if `i  `oi , with pi .0/ D 0 for full unemployment. Then, applying the ELIE scheme to functions Pi gives ti D P .k/  Pi .k/ and yi D Pi .`i / C ti D Pi .`i /  Pi .k/ C P .k/. If `i D `oi and `oi  k, For other levels of wi , the case of individuals who choose to work very little (`i < k) is treated as indicated in Sect. 3.8.1. 68 Low wi at a given time only is normally the object of an insurance compensation (health, unemployment – see also below –, etc.). 69 Computations of the effects are provided in Kolm (2005, Chap. 7). 70 A particular case can be pi .`i / D wi `i . 67

3

Economic Macrojustice

127

Pi .k/ D pi .`oi / D Pi .`oi / D Pi .`i /, and therefore yi D P .k/ D p.k/. Q This is in particular the case for full unemployment, `oi D 0. Moreover, if, when `oi > 0, person i chooses to work less than `oi , her income is reduced by the corresponding loss in output.

References and Bibliography Arrow, K. (1963). Social choice and individual values. New York and New Haven: Wiley and Yale University Press. Atkinson, A. (1995). Public economics in action: The basic income/flat tax proposal. Oxford: Oxford University Press. Bar-Hillel, M., & Yaari, M. (1984). On dividing justly. Social Choice and Welfare, 1(1), 1–24. Bentham, J. (1789). An introduction to the principles of morals and legislation. New York: Delphin Books. Reprint 1961, in The Utilitarians. Bourguignon, F., & Spadaro, A. (2008). Tax-benefit revealed social preferences. Working paper 2008-37, Paris School of Economics, Paris. Coase, R. H. (1960). The problem of social cost. Journal of Law and Economics, 3(1), 1–44. Dasgupta, P., & Hammond, P. (1980). Fully progressive taxation. Journal of Public Economics, 13, 141–154. Dworkin, R. (1981). What is equality? Part I: Equality of welfare; Part II: Equality of resources. Philosophy and Public Affairs, 10, 185–246, 283–345. Hicks, J. (1959). Essays in world economy. Oxford: Basil Blackwell. Kolm, S.-C. (1966a). Les choix financiers et monétaires (théories et techniques modernes). Paris: Dunod. Kolm, S.-C. (1966b). The optimal production of social justice. In H. Guitton, & J. Margolis (Eds.), Proceedings, international economic association conference on public economics, Biarritz, 1966. Economie publique, Paris, CNRS, 1968, pp. 109–177. Public economics, MacMillan, London, 1969, pp. 145–200. Reprinted in The foundations of 20th century economics, landmark papers in general equilibrium theory, social choice and welfare, selected by K.J. Arrow, & G. Debreu, 2001, Cheltenham, Edward Elgar, pp. 606–644. Partial reprint in Journal of Economic Inequality, 2007, 5, pp. 213–234. Kolm, S.-C. (1970a). L’Etat et le système des prix. Paris: CNRS-Dunod. Kolm, S.-C. (1970b). Prix publics optimaux. Paris: CNRS-Dunod.

128

S.-C. Kolm

Kolm, S.-C. (1971). Justice et équité. Paris: Cepremap. Reprint, Paris: CNRS, 1972. English translation, 1997, Justice and equity. MIT Press: Cambridge, MA. Kolm, S.-C. (1973). A note on optimum tax evasion. Journal of Public Economics, 2(3), 265–270. Kolm, S.-C. (1974a). La transition socialiste. Paris: Ed. du Cerf. Kolm, S.-C. (1974b). Sur les conséquences économiques des principes de justice et de justice pratique. Revue d’Economie Politique, 84(1), 80–107. Kolm, S.-C. (1984). La bonne économie: La réciprocité générale. Paris: Presses Universitaires de France. Kolm, S.-C. (1985). Le contrat social libéral. Paris: Presses Universitaires de France. Kolm, S.-C. (1986a). Freedom, cores and public goods. Working paper 50, CERAS, Paris. Kolm, S.-C. (1986b). L’allocation des ressources naturelles et le libéralisme. Revue Economique, 37, 207–241. Kolm, S.-C. (1987). Public economics. In J. Eatwell, M. Milgate, & P. Newman (Eds.), The new Palgrave: A dictionary in economics (pp. 1047–1055). London: Macmillan. Kolm, S.-C. (1996a). Modern Theories of Justice. Cambridge, MA: MIT Press. Kolm, S.-C. (1996b). The theory of justice. Social Choice and Welfare, 13, 151–182. Kolm, S.-C. (1999a). Freedom justice. Working paper 99-5, CREM, Université de Caen, France. Kolm, S.-C. (1999b). Rational foundations of income inequality measurement. In J. Silber (Ed.), Handbook of income inequality measurement (pp. 19–94). Kluwer, Dordrecht. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge: Cambridge University Press. Kolm, S.-C. (2006). (a) Introduction. (b) Reciprocity: Its scope, rationales, and consequences. In S.-C. Kolm, & Jean Mercier Ythier (Eds.), Handbook of the economics of giving, altruism, and reciprocity (pp. 1–122 and 371–541). Amsterdam: North Holland. Kolm, S.-C. (2008a). The paradoxes of the war on poverty: Warm-glows and efficiency. Working Paper 2008-7, IDEP, Marseille. Kolm, S.-C. (2008b). Reciprocity, an economics of social relations. Cambridge: Cambridge University Press. Kolm, S.-C. (2008c). The theory of rules and the moral provision of public goods. Mimeo. Paris: EHESS. Kolm, S.-C. (2009a). Justice. In J. Peil, & I. van Staveren (Eds.), Economics and ethics (pp. 291–300). London: Edward Elgar.

3

Economic Macrojustice

129

Kolm, S.-C. (2009b). The rational, recursive original position. A fully determined impartial endogenous social welfare function. In Conference in honour of Maurice Salles, University of Caen. Kolm, S.-C. (2009c). Social ethics and rationality, new directions for the opimum production of social justice: Meaningful welfare, equal liberties, social solidarity. In Conference on inequality, new directions, Ithaca, N.Y. Cornell University. Forthcoming in Journal of Economic Inequality (2011). Kolm, S.-C. (2010a). Equality. In B. Badie (Ed.), International encyclopedia of political science. London: Sage Publications, Inc. Kolm, S.-C. (2010b). On real economic freedom. Social Choice and Welfare, 35(3), 351–375. Locke, J. (1960,1689). Second treatise of government. Cambridge: Cambridge University Press. Maniquet, F. (1998). An equal right solution to the compensationresponsibility dilemma. Mathematical Social Sciences, 35, 185–202. Mill, J. (1957,1861). Utilitarianism. New York: Bobbs-Merrill. Mirrlees, J. (1971). An exploration in the theory of optimum income taxation. Review of Economic Studies, 38, 175–208. Mirrlees, J. (1986). The theory of optimal taxation. In K. Arrow, & M. Intriligator (Eds.), Handbook of mathematical economics (Vol. 3). Amsterdam: North-Holland. Mirrlees, J. (1990). Taxing uncertain incomes. Oxford Economic Papers, 42, 34–45. Mirrlees, J. (1996). Information and incentives: the economics of carrots and sticks. Nobel lecture, Royal Swedish Academy of Sciences. Musgrave, R. (1959). The theory of public finance. New York: McGraw-Hill. Musgrave, R. (1974). Maximin, uncertainty, and the leisure trade-off. Quarterly Journal of Economics, 88(4), 625–632. Nozick, R. (1974). Anarchy, state and utopia. New York: Basic Books. Pazner, E., & Schmeidler, D. (1974). A difficulty in the concept of fairness. The Review of Economic Studies, 41(3), 441–443. Phelps, E. (1973a). Economic justice. Hammondsworth: Penguin Education. Phelps, E. (1973b). Taxation of wage income for economic justice. Quarterly Journal of Economics, 87(3), 331–354. Pigou, A. C. (1912). Wealth and welfare. London: Macmillan. Rawls, J. (1974). Reply to Alexander and Musgrave. Quarterly Journal of Economics, 88, 633–655. Rawls, J. (1982). Social unity and primary goods. In A. Sen, & B. Williams (Eds.), Utilitarianism and beyond (pp. 159–185). Cambridge: Cambridge University Press.

130

S.-C. Kolm

Rawls, J. (1999). The law of people. With the idea of public reason revisited. Cambridge, MA: Harvard University Press. Rawls, J. (1999,1971). A theory of justice (revised edition). Cambridge, MA: Harvard University Press. Sadka, E. (1976). On income distribution, incentive effects and optimal income taxation. Review of Economic Studies, 42, 261–268. Schokkaert, E. (1999). Monsieur tout-le-monde est post-welfariste. Revue Economique, 50(4), 811–31. Schokkaert, E., & Lagrou, L. (1983). An empirical approach to distributive justice. Journal of Public Economics, 21, 33–52. Schokkaert, E., & Overlaet, B. (1989). Moral intuitions and economic models of distributive justice. Social Choice and Welfare, 6, 19–31. Seade, J. (1977). On the shape of optimal tax schedules. Journal of Public Economics, 7, 203–236. Sidgwick, H. (1874). The method of ethics. London: Macmillan. Slemrod, J. (2002). Tax systems. NBER Reporter, Summer 2002:8–13. Slemrod, J., & Yitzhaki, S. (1987). Tax avoidance, evasion and administration. In A. Auerbach, & M. Feldstein (Eds.), Handbook of public economics (pp. 1423–1470). Amsterdam: North-Holland. Smith, A. (1976,1759). The theory of moral sentiments. Oxford: Oxford University Press. Van Parijs, P. (1995). Real freedom for all. Oxford: Oxford University Press.

•

Part II

Philosophical Aspects of Macrojustice

Chapter 4

ELIE and the Emotions Related to Social Recognition Pierre Livet

Abstract Does the ELIE system of redistribution always lead to positive social emotions and improve social recognition? Receiving a transfer in ELIE could be a sign that your social status is that of the less talented: even if you make the sacrifice of working more in order to contribute to the transfer, your productive capacities may have no market value. This could decrease the strength of the social bond. In order to avoid this, we need to be able to attach social recognition to the passive sacrifice implied in being less talented. One way is to relate this situation to the value for collective adaptability of a level of randomness in the distribution of talents. The redistribution procedure – the sacrifices of the more talented people – leaves the less talented people free to contribute or not by their work to their community, and their passive sacrifice is thereby changed into an active one. Social recognition can thus become mutual.

4.1 Introduction ELIE was formulated from the perspective of distributive justice. As we have several concepts of justice at our disposal, a scheme of redistribution should be evaluated with reference to the concept of justice that we choose. We suggest an evaluation combining the perspective of a reflection on the concept of justice with the perspective of the affective experience of social situations. In the latter, people are not only affected by the satisfaction or non-satisfaction of their basic interests, but also by the level of social recognition and social status that other people assign to them. The main emotions related to justice,

P. Livet CEPERC, Université de Provence. e-mail: [email protected]

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_4, 

133

134

P. Livet

that is, resentment and the sentiment of having received one’s due, are linked to social recognition of one’s due by other people, and this requires other people to recognise the social status and value of one’s actions and qualities. Social recognition is also related to the strength of the social bond. In Macrojustice, Serge Christophe Kolm considers, for the sake of discussion, two extreme positions: liberalism and complete redistribution. Extreme liberalism implies that you keep all your income for yourself. Complete redistribution implies that the parameter of transfer (k in Kolm’s formalisation) is equal to 1, in which case everybody receives the same income, which is equal to the average. If k D 0, everybody does whatever he likes for himself with his own income. This extreme liberalism corresponds to the weakest social bond. The fact that people belong to the same community is of no importance. Complete redistribution, conversely, implies the strongest social bond, as everybody is more eager to ensure a common level of income than to satisfy his personal additional desires. Of course Kolm suggests that k has to be less than one and more than zero. But then we have to check what affective impact such moderate redistribution would have on the emotions and sentiments arising from social recognition. Social recognition is strongly related to the strength of social links, but opens the way to competition for better social recognition, so that people expect a social community and at the same time try not to have homogenous status. Apparently, Kolm takes as given the strength of the social bond in a society, and parameter k has to be determined in accordance with this particular strength. As a consequence, redistribution in accordance with k is not supposed to increase this strength or to decrease it. Or does Kolm implicitly assume that this redistribution necessarily increases the strength of the social bond? If this is the case, I first propose to show that this is not necessarily true. There are plausible scenarios in which the mechanisms of emotions related to social recognition have the effect of decreasing this strength. But my investigation does not end with this result, and I go on to attempt to specify conditions in which social recognition can increase the strength of the social bond. The third section attempts an interpretation of Macrojustice and ELIE that integrates the impact of the emotions from social recognition, and satisfies conditions for the strength of the social bond to increase.

4.2 Situations where ELIE Transfers can Decrease the Strength of the Social Bond People who are endowed with better capacities than the average obtain social recognition rather easily, particularly if their capacities are detected and rewarded by higher wages, as is assumed by Kolm. However, people with

4

ELIE and the Emotions

135

limited capacities do not obtain such social recognition. On the contrary, their low wages give them low social status and this situation produces negative emotions in them. Is ELIE redistribution1 able to compensate for these negative emotions? At first sight yes: the income level of these disadvantaged people is increased by the income transfer from the better-endowed people. This change must surely induce some positive emotions. But these emotions are not emotions linked to social status and social recognition. They are only emotions linked to income and to the interests and desires that are better satisfied by a higher income. What about emotions specific to social recognition? This depends on which income distribution group the people belong to. Remember that in the ELIE system you receive a fraction k of the average wage w, N and give in turn the same fraction k of the market value of your production capacities wi . If you belong to a group for which k.w N  wi / > 0, you are receivers and not contributors.2 To receive additional income by ELIE transfer and not to contribute to ELIE transfer is proof that your social status is that of the less talented people, at least as concerns the talents that are socially recognised. The positive emotions induced by your increasing income are thus always affected by the negative emotions related to your inferior status. There is another source of difficulty. What means do we have at our disposal to manifest our attachment to a socially recognised value? Our actions in favour of better respect for this value might be our answer. But are these actions really motivated by this specific value, or are they motivated by other personal interests? Our fellow-citizens might ask. As long as they can explain our action both by our commitment to the value and by our other personal interests, they cannot be sure. Therefore, in order to show other people that we are motivated above all by the social value, we need to choose actions some of which could be taken as contrary to these other personal interests. This is the condition for other people to be convinced that we are acting for the sake of the value, to socially recognise our close association with this value and to recognise us as a supporter of the value.3 For the sake of brevity, let us call “sacrifice” this choice of actions contrary to some of our personal interests4 . Social recognition requires such sacrifices.

1

And other redistribution schemes. As C. Gamel has pointed out, in ELIE most of the people benefit and contribute at the same time. However, I focus first on the two extremes, in as much the perverse situations that I examine are more prominent in other redistributive systems in which the people who benefit and the people who contribute are never the same, than in ELIE. 3 Notice that, as a consequence, socially recognisable values require this sacrificial behaviour. 4 Interests and values are not definable in themselves in this framework. Values are recognisable only by contrasting them to interests, but the same motivation can be treated in one case as a 2

136

P. Livet

ELIE ensures some degree of social recognition: more talented people (or better educated, or born to a higher social status) make sacrifices by working longer than they would do if they were not required to transfer part of their wages or incomes to other people. As working no longer than want to is assumed by everyone to be a basic personal interest, they sacrifice this individual but socially identified interest, so that their sacrifice is socially recognised: they make it clear that they are supporters of the value of social justice. But the less talented or less educated people (those who benefit and do not contribute) do not seem to be in a position to make sacrifices. They just work as long as they have to work in order to obtain the income that satisfies their personal interests. Were they to work longer, this longer work could not be clearly assigned to a motivation arising from a commitment to a value, but rather to a desire to obtain a higher income, satisfy more personal interests, obtain income exempt from taxes. “If individual i with N chooses not to work, she nevertheless receives disposable income wi < w ti D k.w N  wi / > 0” (Kolm 2005, p. 154). So in order to get k w N and to contribute to the transfer, she has to work. This can be considered as a sacrifice, if this work has a market value. But “individuals who can earn nothing in the labour market, hence with wi D 0, have an income equal to the average equalisation income k w” N (Kolm 2005, p. 154). They work, they exert themselves, but their productive capacities in this market have no market value. Therefore if your working time is close to the average working time, given N (Kolm 2005, p. 154), the closer that “the lower wi , the closer yi is to k w” your income yi is to the average transfer, the more obvious it becomes to other people that your capacities are of no market value. In a society in which market value is the reference, your social value is very low. You choose to work, you consider it as a sacrifice, but it can be considered by other people as a sign that your capacities are useless. The very means by which you try to gain social recognition are counterproductive in this respect. The result could be that the social divide between the most talented and/or educated people and the least talented/educated could be reinforced. The former already have higher social status, and to this social status, the sacrifice that they are making by following ELIE adds greater social recognition. They are doing something for the value of justice. The latter people have lower status, and cannot render any sacrifices they might make socially recognisable by their fellow-citizens. Even if they make sacrifices with the hope of value and in another as an interest. The distinction is a relative one, as we restrict ourselves to the condition of manifestation of values. Interests are supposed in a society to be the most universal self-oriented motivations. Different members of a society are supposed to have roughly the same interests, but could be committed to different values.

4

ELIE and the Emotions

137

behaving like ordinary citizens and alleviating the burden that they are supposed to be placing on the shoulders of the community, these very sacrifices would be tantamount to recognising their inferiority once again. They have to recognise the social sacrifice of the more talented (not only of the most talented, but also of the more talented, those who at the same time contribute and receive) but at the same time cannot be recognised as sacrificing their personal interests for the sake of social justice. They are mainly passive, not active, subjects or objects of justice. They may believe that other people think that they are living on government handouts. Their plausible affective reaction could be resentment against these more flavoured people and a sentiment of injustice. What about the majority of the people of mixed status (receiving and contributing)? Remember that social recognition is insensitive to differences smaller than a particular threshold. As a consequence, the difference interpreted as sacrifice has to be sufficiently high in order for this sacrificing behaviour to gain social value. In order for the participation of these people to be socially recognised, conferring on them a feeling of worth, they have to contribute much more than they receive: if they contribute and receive the same amount, they can be suspected of taking with one hand what they sacrifice with the other, so that the sacrifice is nullified. If this is true, then anyone receiving more than they contribute can be afraid of being poorly socially valued. If we focus on these potential situations, ELIE, whose aim was justice, appears to be inducing a sentiment of injustice in the minds of the least flavoured, and fear or anxiety about their social value for many of the people of middle status. Although ELIE is better in this respect than other redistributive systems in which the people who benefit are never the people who contribute, the problem still has to be solved. Moreover, more talented people believe that those less talented should acknowledge their effort for justice instead of taking the situation as unjust; consequently, the subjective distance between these two classes of people increases, and the strength of the social bond decreases. In this scenario (only potential, as there are other compensating mechanisms, as we will see), ELIE would start with a given strength of social bond, determining a given value for parameter k, but once the social emotions come into play this value would be decreased. The possibility thus arises that by successive steps we could move towards an absence of social bond, ending up with the extreme situation of pure liberalism (everyone is only motivated by his personal interests). Note that real social situations show a different pattern. Among those most likely to share their income with others and to work harder so as to be able to transfer some money to people who need it, are the poor; the rich are much less inclined to work longer in order to transfer money to poorer

138

P. Livet

fellow-citizens (as shown in a survey by Leroux and Swaton).5 This could simply be a result of the mechanism of social recognition which we have just briefly described. Rich people are already socially recognised as having better social status, provided their wealth is not just inherited but is partly a result of their work. They are supposed to have sacrificed time and pleasure for their work, so that their wealth is not only the result of having worked for their personal interests, but also the result of having sacrificed other values (like good times, family life, and so on) for the sake of economic activity. And even if they do not work at all, but enjoy their leisure as they like, they can be supposed then to be sacrificing interests related to profit in order to seek hedonistic values! In their situation, they do not have to make any more sacrifices than those that they are supposed to have already made. On the contrary, poor people have to make sacrifices in order to get a bit of social recognition, and these sacrifices can be added to an already unfavourable situation!

4.3 Conditions for Strengthening Social Bonds The reader may object that such a social mechanism of recognition, at least if reduced to the previous scenario, necessarily weakens the social bond in every case, not only with ELIE. But our experience shows that this is not always true. Social recognition, common sense says, can also strengthen the social bond. What are the conditions for this positive effect? Our commitment to socially recognised values has to be shown by actions that partly sacrifice some of our supposed self-interest. If we can believe that our fellow-citizens acknowledge this sacrifice, we feel socially recognised, and this gives rise to positive emotions. The implication is that they recognise the kind of personal interests that we have sacrificed, and the kind of value for the sake of which we are supposed to be making these sacrifices.6 Now if, in addition to this simple social recognition, we want the strength of the social bond not to decrease but rather perhaps even to increase, we have to add at least one extra condition. The recognised sacrifice must have favourable consequences for other elements of the society and what is more, these consequences must give other people the opportunity not only to better satisfy their personal interests, but also to make sacrifices in turn that increase their own social recognition. In addition, these latter people have to be able to acknowledge that this increase in their own social recognition 5

See Leroux (2004). This does not imply that they support this same value. They can recognise it as a social value and not personally support it.

6

4

ELIE and the Emotions

139

is partly due to the sacrifices of the former group. If these conditions are satisfied, we can say that the increase in social recognition is mutual. But is this sufficient? For example, increasing the income of poorer people by ELIE transfers makes these transfers socially recognisable. If the acknowledgment conditions are satisfied, the strength of the social bond is maintained or increased by the mechanism of social recognition. Are there means by which distributive justice could satisfy these additional conditions? One way is to shift from the point of view of individuals to that of the community, but in a way that remains compatible with methodological individualism. Let us consider the distribution of talents and status in a society. It is at least partly random. You do not choose your genes, nor the social status of your parents. From the viewpoint of the community and its evolution, such randomness should not necessarily be considered as a defect. Randomness is one face of a coin whose other face consists of variations, mutations and other combinations of genes, social mutations and explorations of diverse forms of human ability to deal with different environments. In order to be able to cope with random variations of environments, some randomness in the determination of capacities of individuals is useful. From this perspective, people who are born with less favourable genetic and social endowments pay the collective price of the adaptability of the whole society to various environments. They make the sacrifices required for the collective adaptability. More talented people cannot despise them because of this, without despising by the same token the proportion of their own advantages that is due to chance. In order to give a social value to their advantages, talented people have either to acknowledge the value of passive sacrifices from less talented people, or to discard both passive sacrifices and passive advantages.7 But it still remains difficult for the less talented person to believe that her passive situation can count as a sacrifice showing her commitment to a value. What could offer the possibility of an active choice to less talented people? There is no answer if we remain stuck in a situation deprived of social distributive justice (the extreme liberal situation). But the existence of a procedure of distributive justice makes an answer possible. If there is some redistribution, poorer people have the opportunity to choose either to work less (as the transfer ensures them a better income without needing to work so long), or to work as they would do without the transfer and obtain a higher income. The first advantage of such a distributive transfer is that it allows the

7

Discarding both passive drawbacks and passive advantages is the way chosen by Rawls, but it is doubtful that real people can really discard them. On the contrary, we suggest here a way to transmute these passive endowments into actively socially recognised ones.

140

P. Livet

more talented people to use their passive advantages by transmuting them into recognisable sacrifices, which is the only way to demonstrate that such passive advantages have a social value.8 Moreover, the additional income that is given to the less talented by the transfer offers them the opportunity to make an active sacrifice, if they choose to work despite their now socially ensured income. What about our previous objections? Is this active sacrifice more likely to be socially recognised as a behaviour of social value? Is it not still a sign of the less useful capacities of the people who have no other sacrifice at their disposal? It can be argued that the interpretative situation has changed. Taken together, a sign of less powerful adaptive capacities and a sign of great adaptive capacities are now, from the collective point of view, the random aspect of the necessary conditions for the long term collective adaptability of the community. The passive sacrifice of the less talented people, taken together with the passive advantage of the more talented people, has a collective social value. This inclusion of passive sacrifice in the collective adaptability imparts a different meaning to their situation: it is not only the sign that their capacities are useless or less valuable. As their passive sacrifice has new meaning, when the less talented people make additional sacrifices – perhaps not only in labour, but also in cooperative activities –, these active sacrifices are not reduced to signs of their inferior status. Now they can be positively socially recognised, due to the combination of the existence of distributive transfer and the collective interpretation of random assignments of natural talents and initial social positions as related to a collective capacity for adaptive evolution. The adaptive interpretation partly blocks the pejorative interpretation, and the existence of the transfer makes it possible for the work of the less talented people to be considered as an active sacrifice. Thus, the sacrifice of the more talented people affords the less talented people the opportunity to make sacrifices showing their support for the collective good. Such sacrifices are, of course, recognisable by the more talented people. Conversely, the less talented people perceive the sacrifices of the more talented as socially recognisable – they are also sacrifices for the collective good. In addition, each group has the opportunity to recognise the contribution of the other group to the social value of its own contribution. Our conditions for mutual social recognition are satisfied. It could be objected that, as the absolute amount of income sacrificed by the talented people is much greater than the amount of income sacrificed

8

We do not assume here that more talented people have a social responsibility with respect to poor people and the community, but only that they are in quest of social recognition, as all citizens are – anti-conformists included, even if they choose means opposite to those of conformists.

4

ELIE and the Emotions

141

by less talented people, the talented people can still despise the sacrifices of the poorer ones. But a more careful examination needs to be made. Suppose that before an agent has made any transfer, the level of his capacities (measured by his wages) is in the upper part of the random distribution of talents: a given degree of sacrifice by work leads to a great amount of transfer of income. If the level is located in the lower part of the random distribution, this very same degree of sacrifice leads to smaller transfer. The higher you are in the income distribution, the more efficient the relation between work and income obtained by your sacrifice to be transferred. How should this relation be interpreted? It is only an implication of the collective evolutionary adaptation perspective: taking into account the randomness of being more talented or less talented implies taking into account this variation in efficiency through the random distribution. Here again, despising the lower amount of the transfer obtained by the sacrifice of the less talented would mean also despising the larger amount of transfer obtained by the sacrifice of the more talented. Our previous reasoning is not undermined by this objection. In this scheme, on the one hand the less talented people are useful to the society through making passive sacrifices that could be considered necessary to ensure the adaptability of the whole society, including the more talented people; and on the other hand, distributive justice makes use of the other side of the coin, that is the more talented people, in order to give the less talented people the opportunity to make active sacrifices and receive active social recognition. The asymmetry of the distinction between the less and the more talented people has been balanced by acknowledging that in a random distribution, there cannot be a group of more talented people without another group of less talented ones.9 In addition, due to the distributive scheme, each of the two groups now has the opportunity to make active sacrifices showing their commitment to social values (the simplest sacrifice is to work longer). And these sacrifices, as they are of the same type, can be recognised from the perspective of both groups. Mutual social recognition follows, and the strength of the social bond is maintained or increased.

4.4 Conditions for Increasing Social Bonds and ELIE Transfers Is this mutual recognition possible with ELIE transfers? Kolm’s framework seems interpretable in accordance with our mutual sacrifice framework. More talented individuals have to earn the equivalent of transfer t in order 9

Not just a tautological claim, stating that more implies less, but the claim that whatever social progress we make, inequalities in talents are an effect of ensuring adaptability by recourse to the introduction of a random mechanism.

142

P. Livet

for the sacrifices of less talented people to be socially recognisable. Once they have done so, and this counts as their own sacrifice, they are free to use their time as they like (leisure or labour). Once this transfer is ensured, less talented people have freedom of choice and can choose to sacrifice or not a part of the transferred income. Parameter k is indeed the key to the freedom of choice needed for a sacrifice to be an active one and to be recognised by other people in a way that transmutes the negative emotions of sacrifice into the positive ones of social recognition. This interpretation of the role of parameter k can be supported by Kolm’s remarks. He says that “an individual more productive than another one can choose more disposable income and consumption for each labour ` > k and less for each ` < k,10 and more leisure for each consumption y > kw N and less for each y < k w N 11 ” (Kolm 2005, p. 159). The freedom of choice – reduced by Kolm to the power of consumption – of the talented individual is higher when he works more than k, and lower when he works less than k. “With a lower w, freedom can be maintained by a proportionately higher transfer t (possibly a lower tax).12 The coefficient of proportionality, independent of levels and individuals,13 is the equalisation labour k” (Kolm 2005, p. 160) if the freedom function is linear in t and w. As the social identification of an active sacrifice implies a significant freedom of choice (higher than the threshold of social sensitivity), working more than k in Kolm’s system could also be a sign of active, and incidentally socially valuable, sacrifice. Parameter k has to be sufficiently high to give by transfer the less talented people the possibility of having some freedom of choice, but not so high as to prevent both more and less talented people from having sufficient freedom of choice to work more then k. If k is too high, this freedom of choice is reduced, and with it the possibility of making recognisable active sacrifices. If k is too low, more talented people have much more freedom, but as transfers are lower, less talented people have less freedom of choice. The only difference between our interpretation and Kolm’s formalism is that we cannot ensure that the freedom function is linear in transfer t and productivity rate w. The positive consequence interpretation would also be valid if the function were not a linear one, but it would also have negative 10

As in the first case, if he chooses to work longer it is only in order to increase his income and he is free to do so, and in the second case, he has to work to finance the transfer of income determined by k w. N 11 Where w N is the average unit of productivity, as in the first case, he can freely renounce a better income in exchange for longer leisure, and in the second case, he has to renounce more leisure in order to work to earn the amount of transfer k. 12 Freedom is maintained either by giving the less talented a higher transfer t or imposing lower tax on him. 13 As it is not dependent on any individual index.

4

ELIE and the Emotions

143

consequences. In this case, we would in addition need a weighting device, giving not only different weights to the sacrifice of the same income by more talented people and by less talented people, as Kolm already does, but also varying the weights in a non-linear way as the productivity rate (that is, a measure of the capacity or capability, the talent) and the amount of transfer vary. But then a unicity property could be lost. Remember the figure “The geometry of ELIE” (Kolm 2005, p. 157). Kolm represents different budget lines for people with different capacities. If you are more talented, your income is higher, but your transfer is negative (a tax) and if you are less talented, your income is smaller, but your transfer is positive. The different lines cross each other at a single point K, the coordinates of which are N If the relation between the difference in labour `i D k, and income yi D k w. capacities and the freedom of choice – and of making socially manifest active sacrifices – were non-linear, this could imply that there would no longer be a single K point for every budget line, for any relation whatsoever between the amount of transfer (from the less talented side) and the productivity rate (from the more talented side), as the non-linear variations could be different ones for different relations between different talented people and different less talented ones. Whether there is such a single point or not, the main idea and principle remains the same. While we have to find some equivalence between the sacrifices of the talented people and the opportunities for sacrifice that their transfers allow the less talented people, we must remember that fairness and justice are not a simple matter of equivalence. Our aim is not to equalise two independent values, for example a deficit of income for the poor people and a transfer of income from the rich people, or to maximise an aggregation of the different incomes and the transfer plus the incomes earned by the people’s work. The process of evaluation is rather a recursive one. From the outset, it needs to assume a supposed level of transfer, which produces a given level of freedom of choice for the more talented as well as for the less talented people. This provides the opportunity to make socially visible sacrifices. These sacrifices enable social recognition and social values. The possibility of obtaining such social recognition motivates each individual to determine his amount of labour. The distribution of these amounts of labour modifies the amount of transfers possible, which in turn modifies the freedom of choice, and so on. In addition, these modifications have to improve mutual recognition of the sacrifices of the more talented and the sacrifices of the less talented. We have to collectively choose the value of parameter k that leads by this recursive process to a satisfactory state of social justice. It is only by combining this fixed point scheme with the acceptance of the random character of our endowments and social status at birth, that we can

144

P. Livet

integrate the affective importance of social recognition into the concept of distributive justice.

References and Bibliography De Sousa, R. (1987). The rationality of emotion. Cambridge, MA: MIT Press. Goldie, P. (2000). Emotion. Oxford: Oxford University Press. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge, MA: Cambridge University Press. Leroux, A. (2004). Eliminer la pauvreté en France. Paris: Economica. Livet, P. (2002). Emotions et rationalité morale. Paris: PUF. Rawls, J. (1999). A theory of justice. Cambridge, MA: Harvard University Press.

Chapter 5

Basic Income and ELIE Transfers: Argument for Compatibility Despite Divergence Claude Gamel

Abstract Van Parijs (1995) with basic income and Kolm (2005) with ELIE transfers have both revisited the ethical foundations and the redistributive patterns of the tax system. Despite being formally close, both propositions diverge because the financing of basic income is not really guaranteed and the treatment by ELIE transfers of “eccentric productive people” who choose not to work is not obvious. Both projects remain nevertheless compatible: from a philosophical point of view, Van Parijs tries to equalise individuals’ “external endowments”, while Kolm exploits only their “internal endowments”; from an economic point of view, TECIE transfers which would be based on “external endowments” could thus complete ELIE transfers stemming from “internal endowments”. The first examination of this “hybridisation” provides the framework of our conclusion.

5.1 Introduction: The Formal Closeness of Two Projects In a ten years’ interval between their publication, Van Parijs’ book – Real Freedom for All (1995) – and Kolm’s book – Macrojustice (2005) – have contributed, in their own right, to revisiting the ethical and philosophical foundations of the tax and income redistribution system. Van Parijs’ aim is to give a suitable financing to the project of “basic income”, whereas Kolm’s

C. Gamel GREQAM-IDEP, Université Paul Cézanne e-mail: [email protected]. The author thanks Serge-Christophe Kolm, Michel Lubrano and Alain Wolfelsperger for their comments on earlier versions of the paper.

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_5, 

145

146

C. Gamel

priority is to discuss the question of “macrojustice” through “equal labour income equalisation”. The two perspectives, each in its own way, include deep reflections on freedom, “real freedom” for one, and “social freedom” for the other: – Van Parijs’ basic income project consists in giving to freedom a real content, which goes well beyond its formal aspects: it means offering the individuals the income they need to lead their life as they please (marketable, non-marketable or even domestic activities), with the amount of resources the society is able to provide for their financing. Therefore the project is not a rewriting of existing minimum income guarantees, the latter being nothing but social benefits receivable by definition only below ceiling resources, whereas the purpose of basic income is to provide each individual (resident or national) with a lifelong minimum income irrespective of his or her professional situation, civil status or income from other sources.1 – Kolm’s reflection on macrojustice aims at reconciling in an explicit way “social freedom”, defined as “individuals’ freedom from others’ forceful interference”, and income redistribution which seems to imply such interference. The objective is to build a double consensus about macroeconomic redistribution: the first agreement being about a general scheme of income redistribution – “equal labour income equalisation” – and the second agreement being about the intensity of the equalisation. Once this double consensus has emerged, everyone remains free to choose the intensity for exploiting his or her income-yielding personal productive capacities. These capacities provide also the redistribution system with an inelastic tax base. The purpose of our paper is to define to what extent these two projects can appear to be compatible despite the fact that they are not convergent. At first sight several arguments may justify our research: – First of all, the two authors are prompted by the closely related philosophical foundations of their reflections: Van Parijs’ “real libertarianism” (1995, pp. 25–29) and Kolm’s “process liberalism” (2005, pp. 20–23) include common critical views on classical liberalism and try to legitimise in a liberal society the setting of a redistributive system which would be both fair and efficient.

1

Basic income would thus replace (totally or partially) many forms of existing minimum social benefits targeted at life’s various up and downs (dependents, unemployment, low income, single parenting, handicap, old age).

5

Basic Income and ELIE Transfers

147

– Moreover, it is true that Kolm has underlined several times in his book Macrojustice that ELIE transfers and basic income are quite different2 ; but in his more recent writings, he is inclined to minimise that difference, so that he regards ELIE transfers as an “equal universal basic income financed by equal labour” (this volume, p. 111)3 and according to personal capacities of everyone. – Lastly, it would seem that the real divergence between basic income and ELIE transfers relies on the fact that Van Parijs’ basic income is a universal grant, whereas Kolm’s ELIE transfer can vary from individual to individual, according to personal productive capacities. Now it would be wrong to forget that universal basic income is not equal for everyone since the situation of handicapped people might be compensated by a higher amount of basic income; this compensation, which according to Kolm is because of lower productive capacities, is also justified by Van Parijs because “real freedom” would be curbed. In answer to the question we ask here, our own analysis begins by confirming and even by stressing Kolm’s initial position on the difference between the two projects, which seems to be a real divergence (Sect. 5.2). The latter results not only from more or less credible modalities of financing (Sect. 5.2.1), but also and more fundamentally, from the very dissimilar treatment which both authors reserve to non-working people (Sect. 5.2.2). Even though the starting positions are so sharply different, the fact that both systems are definitively incompatible is not yet granted. We will then try to show that ELIE transfers and basic income can be seen as two projects which are more complementary than opposite (Sect. 5.3). From a philosophical point of view, Van Parijs, who is inspired by Dworkin on this point, tries to finance basic income by using people’s “external endowments”, while they are fully neglected by Kolm: in the design of ELIE transfers, only “human natural resources” are considered, that is to say “internal endowments” (Sect. 5.3.1). From an economic point of view, it is then possible to consider a more complete system of redistribution: as a complement to ELIE transfers which are based only on “internal endowments”, we can make use of “external endowments” to define TECIE redistributive transfers – “Totally Exploited Capital Income Equalisation” (Sect. 5.3.2). The first examination of this “hybridisation” provides the framework of our conclusion (Sect. 5.4).

2 3

See especially Kolm (2005, pp. 224, 238, 240–242). See also Kolm (2006, p. 64) and Kolm (2007, p. 76).

148

C. Gamel

5.2 The Reasons of a Theoretical Divergence Till now the time gap between the publications of the two books has not yet given Van Parijs the opportunity to voice his opinion about ELIE schemes, but on the other hand Kolm repeatedly formulated his opinion about basic income projects, the financing of which seems to him especially open to criticism.4 Even though Van Parijs has devoted one whole chapter to this question,5 the solution he promotes (the use of “employment rents”) is a complex one and above all it really lacks credibility (Sect. 5.2.1). As a matter of fact, from our point of view, the divergence cannot be reduced to the single problem of financing, because both projects also differ by their conception of labour. Both authors obviously disagree on the role of labour in social integration and this disagreement is made fully apparent in the treatment of “eccentric productive people” by ELIE (Sect. 5.2.2).

5.2.1 The Financing of Basic Income Through “Employment Rents” Sustainable financing of basic income directly follows from Van Parijs’ “real” version of libertarianism that we shall detail later on (see Sect. 5.3.1). For the while, let us concentrate on the principle which stems out of his analysis: the available wealth at a given time, whether it consists of natural resources or assets inherited from past generations, should be equally shared between all the individuals. 5.2.1.1 Wealth as an Inadequate Funding of Basic Income Van Parijs (1995, p. 101) suggests that the amount of assets composing this wealth, estimated at the competitive market price, should be taxed to be redistributed in an egalitarian way between all the individuals: “An equal distribution of their value therefore amounts to taxing the value of all gifts and bequests at 100% and distributing the proceeds in the form of a uniform basic income”.

4 Particularly the financing of basic income is neither balanced (Kolm 2005, p. 224, Sect. 2.4 “Financial balance”) nor compatible with “process liberalism” he advocates (Kolm 2005, p. 238 Sect. 5 “Freedom and dignity”). 5 See Chap. IV “Jobs as assets” (Van Parijs 1995, pp. 89–132).

5

Basic Income and ELIE Transfers

149

Once the tax base has been defined, we should aim at the highest possible level of basic income compatible with economic efficiency. At this stage, Van Parijs (1995, p. 101) does not underestimate the negative effect of an excessive tax rate: “As the total amount that gets saved, invested, or preserved may well be adversely affected by high rates of taxation, 100% is not likely to be the choice that maximises the tax yield, nor the level of the grant”. The French situation, for instance, gives us an idea of such a divergence: the total value of bequeath and donations represents less than 3% of GNP every year, while the total tax yield from donation and inheritance is about 0.25% of GNP. This represents a gap from 1 to 12. From these figures, Van Parijs (1995, p. 102) concludes that the corresponding estimates for the level of basic income go “from the pathetically low to the frankly negligible”. In this context, taxation efficiency does not seem to be compatible with a decent level of basic income. How would it be possible to enlarge the taxation base in a coherent way, if we do not want to tax wealth too heavily at the time of inheritance? Having investigated several lines of research, Van Parijs supports the idea that people’s wealth cannot be reduced to nonhuman wealth (financial assets, buildings, estates) and to human skills (human capital). Indeed, since the labour market cannot be cleared according to the usual “walrasian” definition, the impossibility of lowering the current wage rate prevents all those who would like to work to get a job; as a consequence, Van Parijs (1995, p. 108) considers that “the holding of a job constitutes a third type of asset”. In societies with high involuntary unemployment, it is fair to deal with “employment rents” in the same way as we do with nonhuman wealth: both must be equally redistributed, because in both cases these assets are not reachable to everyone. More precisely, the employment rent is made by the difference between the income the employed earn from their job and the (lower) income that they would earn if the labour market were to clear according to the “walrasian” assumption. Once this gap is evaluated, its amount is added to the initial tax base designed for financing basic income. 5.2.1.2 The Limits of Financing Through “Employment Rents” Let us now summarise the many theoretical and practical obstacles which come against the levy of Van Parijs’ “employment rents”.6

6

A more complete statement of these obstacles can be found for the French-reading people in one of our previous works (see Gamel 2004).

150

C. Gamel

– Unlike other assets the valuation of which is made at the competitive market price, the rent associated with the holding of a job corresponds to the difference between the effectively earned wage and the wage which would have been received if the “walrasian” equilibrium of the labour market had been reached. How can this hypothetical equilibrium wage, which is indispensable for the estimation of the rent, be determined? As a matter of fact, it would be necessary to organise an independent set of auctions, each matched to a particular type of job, but Van Parijs himself concedes that this idea does not make much sense. – As for the other types of assets, a 100% taxation of this rent would have more or less intense counterproductive effects; in this particular case, these effects create a lower incentive to work, a lower profitability for firms and a lower creation of wealth, all these reducing indirectly the “sustainable” level of basic income. – The practical difficulty of taxing these rents at an optimal rate therefore forces Van Parijs to propose an approximate method consisting in taxing, not the rent itself, but the entire wage linked to the job. “We can keep going, he writes (1995, p. 115), protected as we are against the risk of taking away more than is legitimate by the absence of involuntary employment: no one is stuck with a job with a negative rent”. – However this argument seems quite weak, because it can induce a legitimisation of a maximal tax pressure on labour income. The limit to tax pressure would be only the fact that an excessive taxation would push the too heavily taxed employee to give up his job and to be satisfied with basic income. In fact, the true limit to such a drift lies in a general principle that is also valid for the other assets; according to this principle, “wages should be taxed up to the point at which the tax yield, and hence the basic income financed by it, is maximised” Van Parijs (1995, p. 116). – We can notice that the same principle can be easily applied not only to the income of wage earners but also to the income of the self-employed workers and to that of capital owners (1995, pp. 118–119); therefore we become aware that Van Parijs manages to justify a particularly elaborated fiscal system which suspiciously looks like those found in all developed countries: the only difference is that his system, by assumption, has been optimised. Finally, the financing of basic income which, originally, had to rest only on wealth levy at the time of inheritance, eventually mobilises, from concessions to concessions, a big part of the pre-existent income tax system. The solution recommended by Van Parijs is complex and is not quite suitable

5

Basic Income and ELIE Transfers

151

for “optimising capitalism”7 as he would aim at: for basic income, as for many other ideas that stand out as landmarks in the history of economic thought, the question of financing appears to be a tough test, insofar as even the most attractive theoretical projects could lose in practice some or all of their credibility. 5.2.1.3 The Financing Test a Priori Easier for ELIE Transfers Kolm’s project seems to pass this financing test much more easily, provided however that the double consensus on the macroeconomic income distribution is accepted, as the theory of macrojustice suggests it8 : – According to the general consensus in ELIE, redistribution is based on the personal capacities of individuals to earn an income while working; these capacities are inelastic “given (or natural) resources”, because they don’t vary when people modify their behaviours as producers, consumers or savers. It is thus advisable to tax a market price of the productive capacities of each individual, whatever the working time he or she chooses to exploit them (while working full-time, part-time or not at all). Then the market price is set up according to the wage he or she could draw from his or her productive capacities, while working full-time. Furthermore everyone does not live alone, cut off from the rest of the world, but lives inside a society of free men towards which he or she is able to express a degree of membership and solidarity. As a result, everybody agrees that some of the income he can draw from his own productive capacities is given to the community to be equally shared. For example, if the individual gives to society the market value of two days in his weekly work, he discharges a lump sum tax, variable from individual to individual, according only to his or her personal productive capacities. In return, everyone receives from the community the average value of all the incomes resulting from these two working days. The gap between the individual incomes is thus reduced by the leveling division of the money value of the same part (in the example 2 every 5 days that is 40%) of everyone’s productive capacities. – From that, the choice of a precise scheme of ELIE redistribution depends on the market value of productive capacities. More precisely, this choice depends on the percentage of this market value that people give to society;

7

See Chap. VI «capitalism justified» (Van Parijs 1995, pp. 186–233). For a more detailed statement of this double consensus, see Gamel and Lubrano (this volume, Sects. 1.3.1 and 1.3.2).

8

152

C. Gamel

it might be supposed that individuals will choose its amount in a more consensual way by avoiding two extreme situations: in the first case the parameter of equalisation would be zero (no redistribution at all) and, in the latter case, equal to one (egalitarian redistribution of the whole available capacities). Thus, if this “parameter of equalisation” k is in between 0 and 1,9 the situation of the more productive people and that of the less productive ones are to be characterised by a net transfer which is not equal to zero, negative for the first ones, positive for the second ones. Because their full-time remuneration on the labour market is higher than the average wage, the most productive will have to pay kwmp and will receive only k w; N from which: tp D k.w N  wmp / < 0. Conversely the contribution of the second ones kwlp will be lower than the allowance k wN they would receive; from which: tlp D k.w N  wlp / > 0. By construction, negative and positive net transfers are thus balanced and the financing of ELIE transfers, which is present from the very beginning of the story, is set up without any difficulty. We can thus justify Kolm’s scepticism about basic income: “This balance makes a difference, he writes (2005, p. 224), with distributive schemes that consider solely the structure of taxes – notably the income tax – or solely subsidies without considering their financing – notably for aids to low-income earners (or proposals to grant the same ‘universal basic income’ to everyone).” Moreover, as the fiscal basis of financing is inelastic and the tax system is a lump sum one, the most productive people are incited to work, because if their working time ` is higher than k, the income they get from their labour activity beyond k is completely tax-exempt. Symmetrically, the least productive people are not led to step out from labour activity: firstly, the positive transfer they receive will not increase if they reduce their working time deliberately (no “inactivity trap”), and, secondly, the level of this positive transfer only depends on their productive capacity (wlp ), whatever its low value, so as to guarantee the financial balance of the ELIE scheme. Our previous remarks highlight, beyond the question of financing, the major difference between basic income and ELIE transfers: on the one hand, the propensity to tax “employment rents” so as to complete the financing of the former, on the other hand, the concern of the latter to eliminate the

More precisely, according to Kolm (2005, pp. 187–195), the k parameter of equalisation ought to be chosen in a more restricted space, between 0 and k e (< 1); the k e “income-egalitarian equalisation labour” seems to be “the best specification of the set of moral stands or ‘intuitions’ of present-day income egalitarians” and is to be found “in the neighbourhood of the lowest normal full-time labours” (2005, p. 190). As a consequence, the extra income which is earned with productive capacities used above k e remains out of the field of redistribution. Even if in some respects this specification has to be considered, however it does not change our further analysis.

9

5

Basic Income and ELIE Transfers

153

counterproductive incentives towards work. Indeed, it is especially the conception of labour and the role it has to play in the social insertion that make both systems of redistribution to diverge. For that purpose, the treatment of “eccentric productive people” in the ELIE schemes reveals, in our opinion, a critical point of Kolm’s thesis that Van Parijs’ basic income can elude, by definition.

5.2.2 The Treatment of “Eccentric Productive People” in the ELIE Schemes According to Kolm, the situation of “eccentric productive people” who choose not to work is one case among many: in all these cases, the individuals are unable to provide their contribution kwi to the financing of ELIE transfers using only their labour income `i wi , because their working time is lower than the value of the parameter of equalisation k. To illustrate his comment, Kolm (this volume, p. 116) mentions as special situations “partial or full employment shortly considered, the few eccentric productive people who drop out of cooperative social production, victims of particular handicaps, part-time jobs which are often second wages in families, etc. These particular cases deserve particular criteria and treatments. They are, therefore, out of the scope of overall distributive justice in macrojustice.” 5.2.2.1 The “Eccentric Productive People”, as a “Particular Case” Among Many Others. . . In spite of their being “out of the scope of overall distributive justice”, Kolm admits however that some of these particular situations can be reintroduced in his analysis of macrojustice, to begin with the case of the involuntary unemployed persons. The latter are individuals whose working time is equal to zero (`i D 0), while their productive capacities are either valued on the labour market (wi ¤ 0) or not (wi D 0). In all cases, since society is incapable of exploiting the skills they offer on the labour market, involuntary unemployed people are entitled to the maximal transfer t D k w N 10 : Kolm (2005, p. 213) writes “there is no difference, in the end, whether the Whatever his productive capacity wi may be, an individual i earns an income yi which corresponds to his personal income free of his contribution to the ELIE scheme [.`i  k/wi ]; as the income received from the community (k w) N is added to this net income, we have the following N A full-time involuntary unemployed person only perceives ELIE result: yi D .`i  k/wi C k w. transfer k w, N since the other constituent of his income is equal to zero: either the productive capacity 10

154

C. Gamel

impossibility to obtain an income is due to the productive capacities (as valued by the market) or to the constraint of the market; someone who cannot sell her labour or its product is equally aided, regardless of the reason for this impossibility”. A similar treatment can be suggested for other particular cases evoked above: – Part-time involuntary unemployed persons or part-time employees happen to be in the same situation (`i < k), but the transfer to which they will be entitled has to take into account the partial exploitation of their productive capacities. This transfer will be maximal only if their working time `i comes up against the much too low working time (`0i ) that is available on the labour market (`i > `0i ); if this is not the case (`i  `0i ), the transfer will be reduced to take into account the free decision the individual makes, in so far he is not forced by the lack of work.11 – The handicapped persons whose productive capacity is not equal to zero should be treated in a like manner, if they do not succeed, in spite of their willingness, to work at least at the level of k. No doubt that the proposal would be to consider the maximal part-time they are capable of as the equivalent of the constraint of working time (`0i ) for the part-time unemployed persons. As for the disabled persons who are completely incapable of working, they should receive the same maximal transfer (t D k w) N as the full-time unemployed persons. However, Kolm carefully emphasises the limited scope of these situations for macrojustice; in particular he thinks that it is easy to deal with the most sensitive cases of total or partial unemployment: first of all, the root causes of involuntary unemployment must be treated, to start with labour market rigidities which ELIE transfers might also contribute to reduce.12

wi is considered as being equal to zero or the available working time (`0i D 0) is substituted to the working times k and `i that could not have been reached (`i D k D `0i D 0). 11 In the first case (k > `i > `0i ), the quantitative constraint on the labour market imposes: `i D k D `0i ¤ 0; even if the available working time `0i is not equal to zero, part-time and full-time unemployed persons receive the same income (k w), N since the available working time (`0i ¤ 0) is substituted to the working times k and `i that could not have been reached (see previous footnote). In the second case (k > `0i > `i ), the constraint on the available working time (`0i ) has no effect on chosen working time `i (`i  `0i ), but only on k .k > `0i /, that cannot be reached. As a consequence, if we replace only k by `0i in the “personal” constituent of the individual’s income, this constituent is no longer equal to zero but becomes negative [(`i  `0i /wi < 0] and the total N is so much reduced. income of the concerned person [yi D .`i  `0i /wi C k w] 12 “Of course, good macroeconomic policy in the first place, unemployment insurance and specific policies about the labour market and formation are in order. [. . . ] the income support to people with low wage rates provided by the obtained scheme can supersede, to everybody’s benefit, a

5

Basic Income and ELIE Transfers

155

Finally, Kolm (this volume, p. 117) writes that “distributive macrojustice is concerned with normal full-time labour, or at any rate with `i  k, only”, for all these related reasons. Now, by introducing this supplementary constraint, Kolm neglects the particular situations where this constraint might not be respected, whereas they can occur because of a deliberated choice of the concerned individuals; besides, he significantly reduces the scope of macrojustice theory compared with its initial basic ambitions: the freedom of exploiting one’s personal productive capacities is no longer complete, whereas this freedom is still the third stage of macrojustice theory (see Sect. 5.1). More precisely, two different points may be criticised: – Involuntary unemployment and constrained part-time employment are not the only cases to be considered, since voluntary unemployment and chosen part-time employment can also occur. In such cases, the ELIE scheme is far from introducing nice incentive properties; on the contrary, it can maintain confusion and even increase it. For example, in the case of unemployment, and even if no unemployment benefit is paid, it is sometimes very difficult to distinguish between the unemployed who cannot really work and those who lack motivation to be employed. Now ELIE transfers increase that difficulty, because “involuntary” unemployed are well treated and receive the highest benefit Œt D k w, N whereas “voluntary” unemployed are considered responsible for their situation and endowed with a reduced positive transfer Œtlp D k.w N  wlp / if they have low productive capacities. ELIE schemes thus induce a counterproductive monetary incentive, which leads people to reveal neither their preferences (see the behaviour of “faux chômeurs” in the “unemployment trap”), nor their real productive capacities wlp (the transfer t to be received would be reduced). As that first point was already raised in the general introduction of this book,13 we have now to tackle the second point of our critical analysis. – Unlike the previous case which affects individuals receiving a positive transfer, this second point concerns individuals who are more productive than the average and are net contributors to redistribution (t < 0); these persons could nevertheless choose not to work or, at least, not to work enough to let them pay the tax (kwmp ) they besides have to fulfil. That situation was also evoked in the general introduction14 : A “very

number of wage rigidities of public or private nature which are important causes of unemployment (minimum wages, “collusion” in any form, etc.).” (Kolm, this volume, p. 126). 13 See Gamel and Lubrano (this volume, Sect. 1.5.5). 14 See Gamel and Lubrano (this volume, Sect. 1.5.1).

156

C. Gamel

productive and eccentric person” would prefer to devote all his time to poetry rather than to work at the measure of his skills, despite being highly remunerated on the labour market. He would be thus compelled either to work at least part-time or to be paid enough capital earnings, so as to pay his contribution to the ELIE scheme and to face his current expenditures.15 That case is certainly much less frequent than the case of unemployment: as Kolm notes (this volume, Sect. 3.8.1), in developed countries the redistributive systems amount to equally redistributing the incomes of one or two days of work per week (k D 20% or k D 40%) and it is uncommon (except for cases of involuntary unemployment) that individuals, who are able to work, do not exceed these thresholds, even if they choose to work only part-time. For all these reasons, the situation of “eccentric productive people” seems to be symbolically of a great importance, because, in those cases, Kolm’s “social freedom” is no longer relevant: concerning their way of life, these persons experience a really forceful interference of society and their situation is comparable to a kind of “slavery of the talented”, according to Dworkin’s meaning of this expression. In other words, except in the extreme case (k D 1), that situation could also exist in a softer and more casual way. 5.2.2.2 . . . But an Emblematic Situation of Work-Based “Social Cooperation” At first, Kolm was inclined to neglect the emblematic scope of these very rare situations: “The productive individuals who have not any other private means but choose to work a very short time are really very few; for that reason, they may not be included in the main rule of macrojustice”, he asserted Kolm (2007, p. 79)16 ; but more recently he took great care to strengthen his position with two important arguments we have now to analyze:

15

More precisely, the “eccentric productive” person is completely unable to provide his contribution kwmp to the financing of ELIE, if his net personnel income ymp D .`mp  k/wmp C k w N is not positive. In other words, ymp  0, if `mp  k.1  w=w N mp /; this situation is illustrated by segment BC in Fig. 1.4 in Gamel and Lubrano (this volume, Appendix B). As a consequence, the individual can pay kwmp and face his current expenditures only with other (capital) earnings that he must have in hands. If his working time is between k and k.1 w=w N mp /, his net personal income remains positive but is lower than the benefit k w N he receives, because he has to use a part of it to pay his contribution kwmp . 16 Kolm’s shortcut is also illustrated in Figure “The geometry of ELIE” (Kolm 2005, p. 157): although this figure provides an attractive synthesis of macrojustice theory, it does not include the negative transfers that the most productive individuals have to fulfil; consequently it is impossible to track down, among them, the “eccentric productive people” whose situation is very delicate.

5

Basic Income and ELIE Transfers

157

– First of all, public opinion would not agree with a tax on unexploited productive capacities, which therefore would not ensure earnings, because it would mean taxing the opportunity cost of leisure time (at the wage level it could have been paid, had the person used it to work). “This common view has to be obeyed in a democracy”, he concludes (this volume, p. 116). In our opinion this assertion is at least inconvenient, because it induces to weaken the general consistency of ELIE schemes and to remove one of the most original characteristics of macrojustice theory, namely the distinction between the rent of productive capacities, that the individual may share with others, and the freedom of exploitation of these capacities he must keep under his full control. Now, if we follow Kolm, this freedom is no longer applied in the entire field of possible working times (0 < `i < 1) but in a more restricted domain (k < `i < 1), since the individual may no longer choose a lower working time than k. Rather than to bow to public opinion, it would be better to try to convince it to maintain the social freedom of the individuals in its entirety, as far as this extended macrojustice theory including “eccentric productive people” can be considered. On this point, it would be at least worthwhile to experiment a kind of dialogue that would be consistent to the “kolmian” process of “endogenous social choice”.17 – This rather factual argument about public opinion is completed by a second one, which is more theoretical (Kolm, this volume, p. 116): “The very few productive individuals who choose to work very little mostly choose not to benefit from society’s supply of a favourable wage, and hence arguably do not have to be taxed for this advantage. They choose to drop out of the cooperative venture of collective production (and division of labour), from its advantages and, hence, from its liabilities”; and Kolm adds: “These fugitives from production are not, as Rawls (1982) puts it, ‘fully cooperating members of the society engaged in social cooperation over a complete life time for mutual advantage’18 and hence are not party in the sharing of benefits”. In this last sentence, the reference to Rawls is not unintentional since, a few pages before, Kolm (this volume, p. 103) had presented ELIE schemes as “Rawls’s final solution” (the solution he should have proposed for the distribution problem as he posed it after 1974): more precisely, in answer that year to Musgrave’s critical

17

See Kolm (2005, pp. 279–360) ; for a brief presentation, see Gamel and Lubrano (this volume, Sect. 1.5.2). 18 Although Kolm does not specify the page where to find this quotation, the latter seems to refer to a passage (Rawls 1982, p. 164), in which Rawls defines the profile of the persons involved in his conception of the “justice as equity”.

158

C. Gamel

observation about the difference principle (or maximin criterion),19 Rawls had introduced leisure time as a new “primary good”, so as to complement income.20 Therefore, under the “rawlsian” veil of ignorance, the individuals are supposed to agree on an initial allocation where everyone has the same leisure time and the same income: according to Kolm’s theory, this initial allocation concerns the parameter of equalisation k defining, in the income-leisure trade-off, the intersection point of all the budget lines, the co-ordinates of which are .1  k; k w/. N 21 Finally, setting the two references to Rawls side by side perfectly illustrates the conception of labour Kolm agrees with and the thorough difference between Kolm’s and Van Parijs’ conceptions.22 Indeed, at the end of the eighties, Van Parijs has sought to base the principle of a universal basic income on the “rawlsian” theory of justice and on its principle of difference, but Rawls (1988, p. 257) set against him a flat refusal in these terms: “Twenty four hours less a standard working day might be included in the index [of ‘primary goods’] as leisure. Those who are unwilling to work would have a standard working day of extra leisure, and this extra leisure itself would be stipulated as equivalent to the index of primary goods of the least advantaged. So those who surf all day off Malibu must find a way to support themselves and would not be entitled to public funds”.23 Consequently, one single minor difference between Kolm and Rawls seems to prevail: “eccentric productive people” evoked by the former are not to be taxed on productive capacities that don’t yield any income, whereas, according to the latter, people “who are unwilling to work” are not entitled to public benefits. Both positions remain still quite compatible in so far as both exclude people in question from rights (Rawls) as well as duties (Kolm) allocated to the other “fully cooperating members of the society”. 19

“Implementation of maximin thus leads to a redistributive system that, among individuals with equal earnings ability, favours those with a high preference for leisure. It is to the advantage of recluses, saints, and (non-consulting) scholars who earn but little and hence will not have to contribute greatly to redistribution” (Musgrave 1974, p. 632). 20 See Rawls (1974, p. 654). However, his position will be more explicitly expressed in 1988, when he answered to Van Parijs (see the next indented line). 21 See point K on Fig. 1.4 in Gamel and Lubrano (this volume, Appendix B). Incidentally, the level of the k parameter of equalisation results from the already mentioned “kolmian” process of endogenous social choice; in all events this process remains quite different from the “rawlsian” veil of ignorance. 22 Rawls’, Van Parijs’ and Kolm’s conceptions are too briefly compared by Kolm (this volume, Chap. 3, footnote 51). 23 The treatment to be reserved to the “surfer of Malibu” has been so much discussed between Rawls and Van Parijs that it can explain the title of Van Parijs (1991): “Why surfers should be fed: the liberal case for an unconditional basic income”; as a more anecdotal consequence, the cover of Van Parijs (1995) is also illustrated by a splendid image of a surfer in the roller of a wave.

5

Basic Income and ELIE Transfers

159

“Eccentric productive people” are confronted in fine with the following dilemma: either maintain their preference for not working and live outside society, or give it up and adopt a way of life corresponding to the social standards but not to their first aspirations. Now, such a dilemma might not occur in the case of a universal basic income, because of its unconditional feature. 5.2.2.3 Basic Income: A Seemingly Better Solution for “Eccentric Productive People” Concerning the multiple “particular situations” listed by the theory of macrojustice, Van Parijs’ position indeed seems more comfortable than Kolm’s: – Everyone is entitled to the universal basic income which is unconditional, while ELIE transfers grant a positive income only to the least productive individuals; ELIE transfers cannot be considered as a kind of universal basic income, apart from reducing it wrongly, as Kolm does (this volume, Sect. 3.7.5), to the constituent part k w N of ELIE transfers, while the part kwi would proceed from everybody’s participating to its financing. – In addition, as basic income considers every unemployed person in the same way, any reference to the involuntary or voluntary nature of unemployment is of no issue for discussion. Therefore, unlike ELIE transfers (see Sect. 5.2.2.1), basic income is not affected by the phenomenon of the “unemployment trap” and also remains perfectly neutral to work, at least at this preliminary stage of reflection.24 – Can these attractive properties of basic income be maintained through appropriate financing of the transfer? Concerning basic income, that question is really the core of the problem whereas, concerning ELIE transfers, the question of a well-balanced financing is resolved from the very beginning, as we have already observed (see Sect. 5.2.1.3). – In fact, such a balanced financing of basic income is conceivable but, instead of being based on taxation of “employment rents” (see Sect. 5.2.1.2), it ought to be achieved with a proportional (flat rate) income tax.

24

From the point of view of microeconomic theory, this basic neutrality can be analysed as the absence of any substitution effect in the income-leisure trade-off, when basic income is received (upward translation of the income line, without any change of its slope). As the opportunity cost of leisure time remains unaffected, basic income introduces neither incentive nor disincentive to work. For the moment, let’s put aside the question of the tax levy which may finance transfers (see next footnote); the final effect of basic income on the propensity to work will then be defined by the direction of the income effect, depending on whether the beneficiaries consider leisure time either as a “normal good” they try to increase or as an “inferior good”, in which case a higher income has still priority; this second behaviour is not excluded for the poorest people, when social integration through labour is prevailing in their mind (for more details, see Gamel et al. 2006).

160

C. Gamel

Such a solution at least reveals the very sense of the exploratory attempt led by Atkinson (1995), the details of which cannot be set out here; nevertheless this attempt leads to an unexpected likeness: basic income would have in fact the same macroeconomic impact as the systems of negative income tax.25 – With such a financing, basic income certainly loses its basic neutrality concerning labour and turns out to adopt the goal of any negative tax, that is to say to avoid the “unemployment trap” by making activity more attractive. However, the main point is preserved for all those who still choose to remain inactive and especially for “eccentric productive people”: the latter would be granted the same basic income as the other individuals and could therefore dodge the dilemma into which ELIE transfers would lock them (see Sect. 5.2.2.2). – Last but not least, the difference between universal basic income and negative income tax is not reduced to a purely accounting difference (ex ante entitlement to basic income with perhaps a later income tax levy or ex post possible allocation of negative income tax once the declarations of resources have been first verified). Indeed, basic income has some singular advantages over negative income tax if we consider, from a more sociological outlook, the perception of each redistributive mechanism by their possible beneficiaries.26 Finally, as far as social integration through labour is concerned, the particular case of eccentric productive people emphasises the major theoretical divergence between universal basic income and ELIE transfers. This

25

To be more precise, if we combine basic income with a proportional tax on all individual’s other incomes, the resulting macroeconomic income distribution could be also obtained by a combination of negative and positive income tax (with the same flat rate). One single difference then remains: basic income and positive income tax are merely superimposed on the whole income distribution system (everybody receives basic income but has also to pay the proportional tax on all their personal incomes except basic income), whereas negative income tax and positive income tax are in fact juxtaposed and cover two distinct zones (in the first zone, the poorest individuals receive the transfer that is the negative tax, while in the second one, the other individuals finance it by the positive tax on their higher incomes). For fuller developments, see Gamel (2004, pp. 307– 309). Van Parijs (1995, p. 57) already evoked the possibility of such a convergence between basic income and negative income tax (through an explicit graph in a brief appendix of his Chap. 2). 26 Because of the ex ante unconditional allocation, basic income, that would be more fitted to treat poverty in case of emergency situations, should reduce the risk aversion of beneficiaries owing to the minimum level of guaranteed income and would induce no “stigmatisation effect” on the poorest people; all those advantages would be achieved by a reduction of management procedures and costs, because of the individualised allocation of basic income (irrespective of the family situation) and because of its neutrality between activity and inactivity. For further developments, see Gamel (2004, pp. 309–310).

5

Basic Income and ELIE Transfers

161

divergence prevents us from concluding that ELIE transfers are a kind of basic income which would be correctly financed. As regards financing, Van Parijs’ position is indeed very weak, either from the very beginning with the problematic taxation of the employment rents or later with the main result of Atkinson’s proposition. In the latter case, basic income turns into an advanced variant of negative income tax and loses one of its main features, namely its basic neutrality between working and non working; but this important feature can no longer be secured with ELIE transfers owing to their original financing. That is why the theory of macrojustice is still worth being investigated further, in order to find how to combine this basic neutrality with an adequate financing. More generally speaking, the question is, could Kolm’s and Van Parijs’ schemes, despite their divergence, be compatible? Our answer to this question is detailed in the following part.

5.3 A Route Towards a Possible Compatibility Although we have found two important points of difference between basic income and ELIE transfers, there is a serious reason for believing in their compatibility: according to Van Parijs, basic income is legitimated by its organising the egalitarian sharing of “external endowments” among individuals, whereas the aim of Kolm’s ELIE transfers is to equalise incomes stemming from their personal productive capacities which could be also called “internal endowments”: the two notions complement one another so easily that they are in fact dependent, from our point of view, on the same philosophical research (Sect. 5.3.1). As the uncertain financing of basic income is also to be considered, a budgetary solution of substitution must be found; hence it seems conceivable that the fiscal treatment of Van Parijs’ “external endowments” should be inspired by the one Kolm has reserved for “internal endowments”. Consequently “TECIE” distributive transfers (“Totally Exploited Capital Income Equalisation”) would thus be added to ELIE transfers to complete income redistribution (Sect. 5.3.2).

5.3.1 “External” and “Internal” Endowments: Two Complementary Philosophical Notions As regards modalities of sustainable financing suggested by Van Parijs for basic income (see Sect. 5.2.1), we have voluntarily left aside the philosophical foundations that originally explain the choice of those modalities, as well

162

C. Gamel

as the notion of “external endowments” that plays a key part for transforming foundations into modalities. It is time for us to tackle the whole question. 5.3.1.1 External Endowments Equalisation, as Foundation of Basic Income In order to legitimise basic income, Van Parijs (1995) uses the “libertarian” thought, according to which private property gives to freedom its best protection: The system of property rights protects the ownership of everyone on himself (self-ownership) and on the possessions rightfully acquired from people who voluntarily gave or sold them (security). However, we are directly interested in an important point of debate and even of conflicts, which still lingers on in the libertarian thought, namely the conditions of appropriation of assets (lands, natural resources), which are originally unowned. Indeed, the libertarians are so very heterodox that their diverse representatives would occupy a very wide space, from the extreme right-wing to the extreme left-wing in the usual range of the philosophical and political values: on the one hand, we find the pure and simple assertion of the property right of the occupant (Rothbard 1989) and, on the other hand, the assertion of an equal right of everyone on possessions which are initially the property of nobody (Steiner 1994), and, between these two positions, we still have at least a third one, namely the necessity of respecting a suspensive clause (“lockean proviso”), which forbids only to appropriate the asset if the resulting situation of other people is lowered (Nozick 1974). As for the way of considering the rights of the current descendants of the first “owners”, we can easily measure the extent of the possible differences between the conclusions to be drawn from these positions. In this debate on the original appropriation, Van Parijs’ analysis can be located well beyond Nozick’s position and closer to Steiner’s: according to the latter, if everyone has indeed the property of assets he acquired by rightful means, any material good contains natural resources on which every human being has an equal moral right, because initially land and its resources do not belong to anybody. From Steiner’s point of view, it is thus quite consistent with libertarian principles to tax and to redistribute among all the part of the total income which corresponds to the value of natural resources included in it. In fact, Van Parijs’ analysis appears to be even more radical than Steiner’s: he starts with the notion of “external resources” defined by Dworkin (1981, p. 307) to point out all the resources an individual can have besides his personal talents and capacities. According to him and unlike Steiner, it is relevant that the whole set of these external resources ought to be redistributed

5

Basic Income and ELIE Transfers

163

equally through the highest possible basic income which would be compatible with economic efficiency. Of course these external resources include natural resources, such as Steiner defines them to be, but many other assets too; in fact they include any useful external asset (including factories and technologies) people are endowed with to pursue their ideal in life. Whether these “external endowments” are natural or have been produced is not really the matter. In spite of all his efforts, Van Parijs does not succeed in combining the equalisation of those widened external resources with any sustainable financing of basic income; which leads him to propose, as already mentioned (see Sect. 5.2.1), the levy of “employment rents”, anyway a problematical issue. Whatever the outcome, a key point is now to be introduced: beyond “external endowments” and “employment rents”, Van Parijs (1995, pp. 60–68) does not at all intend to use other resources individuals have at their disposal, namely personal talents and productive capacities. Both are part of their “internal endowments” and he explicitly analyses them within a “dworkinian” thinking process: he supposes at first that people’s “internal endowments” are identical, before investigating the consequences of the more realistic situation in which these internal endowments differ markedly. But far from studying these differences from the fiscal point of view as tax bases that would be variable among individuals,27 he considers them, as Dworkin does: on the one hand, these differences imply handicaps for people with the weakest internal endowments and, on the other hand, they justify the social benefits that the same people should receive as compensations. At a further stage of the process, Van Parijs (1995, pp. 68–84) deviates from Dworkin’s analysis: to define the amount of compensation, he does not use the method of fictitious insurance markets, especially because of its delicate distinction between “handicap” and “lack of natural talents”. Instead, Van Parijs suggests that the basic income of people considered as handicapped would be increased according to the criterion of “undominated diversity”.28 27

In Sect. 4.7 “An inconsistent proposal? Employment rents with unequal talents”, Van Parijs (1995, pp. 119–125) focuses only on compatibility between “employment rents”, which are for him an essential notion, and existing “internal endowments”, which are quite different among individuals. When he gives up the initial hypothesis of identical internal endowments, the difficulty of measuring employment rents rises by far; Van Parijs (1995, p. 123) ends by concluding that persons with good qualified jobs through high internal endowments ought to be bound by the general rule consisting in taxing not the rent itself, but the entire income linked to the job. As in the general case (see supra Sect. 5.2.1.2), the level of taxation is only restricted by the absence of involuntary employment (“no one is stuck with a job with a negative rent”). 28 See Van Parijs (1995, Chap. III); the criterion of “undominated diversity” is derived from an idea of Ackerman (1980) and from the methodology of the “envy-freeness” theory, of which Kolm (1972) is one of the precursors. According to this criterion, when the situations of two individuals are considered, the endowment of the first one must not always be preferred to that of the second

164

C. Gamel

Now, inside a given amount of financial resources, the effect of such an increase is to reduce the already weak level of basic income to be paid to the “normal” (non-handicapped) people in the population and, under some bad circumstances, this level might even dwindle to zero. 5.3.1.2 From Kolm’s “Natural Resources” to Van Parijs’ “Internal Endowments” Unlike Van Parijs’, the main interest of Kolm’s thinking process is to consider what he calls individuals’ “given (or natural)” resources as the object of macrojustice and ELIE transfers. However we still have to verify that these given resources, which yield everybody’s personal productive capacities, fit well with Van Parijs’s internal endowments. According to the latter (Van Parijs 1995, p. 60), people’s endowments appear to be “their talents, their abilities, their capacities in all areas of life”, which seems to be selfevident: insofar as his analysis is focused on handicaps, the lack of internal endowments can be felt in professional life (production) as well as in personal life (consumption). In Kolm’s analysis the field of people’s given resources is at first much broader but becomes gradually reduced, so as to fit the scope of macrojustice: – First of all, ELIE transfers are essentially based on “inelastic” items which should be “the given resources, classically called the “natural resources”. These resources are nonhuman or human (capacities)” (Kolm 2005, p. 45). – In the same passage, he adds immediately that “many nonhuman natural resources are specific and local [. . . ]. And, for many others, their initial allocation has been made to past owners different from the present ones, and this cannot practically be reversed”. As a consequence, the nonhuman natural resources happen to be out of the field of macrojustice. – The same is true for human resources that are not designed for production, but according to Kolm (2005, p. 46) the reason is different: “Consensus endorses full self-ownership of one’s capacities to be satisfied or happy – or eudemonistic capacities –”. – “Hence the rent of productive capacities”, he concludes (Kolm 2005, p. 46), “remains the only distribuand available for global distributive justice in macrojustice”.

one by any other third individual who would compare them. In other words, consider the endowment of a person whose “handicap” is compensated with an increased basic income; if some third individual thinks that increase is enough, it might happen that this endowment would be preferred to that of a “normal” person. In this case, the criterion of undominated diversity is verified.

5

Basic Income and ELIE Transfers

165

In others words, what Kolm calls natural resources is in a first step larger than Van Parijs’ internal endowments but is finally reduced to a subset of these endowments: natural resources are only talents, abilities and capacities designed to produce and to earn an income. If eudemonist capacities make a persistent difference between Van Parijs’s internal endowments (which include them) and Kolm’s natural resources (which do not), the gap can easily be explained: with basic income, the former tries to compensate handicaps “in all areas of life”, while the latter considers that this problem ought to be dealt with outside the field of macroeconomic income redistribution (inside the family unit or by public arrangement of local justice); moreover eudemonistic capacities would be unable to provide a reasonable funding basis of basic income in any case. From this fiscal point of view, it is much more surprising that Kolm excludes nonhuman resources from the field of the relevant “natural resources”, that is to say, excludes from the field of macrojustice external endowments that are on the contrary carefully considered by Van Parijs. So as to justify his position, Kolm (2005, pp. 84–89)29 produces three series of arguments: – In a modern economy, the shares of incomes from labour, capital and nonhuman resources can be seen roughly as 80, 18 and 2%. Since capital is produced from nonhuman and human resources, the value of capital contains a value due to past uses of nonhuman resources equal to 0, 5% [ 18  2=.80 C 2/] and the remaining value (17, 5%) proceeds from labour accumulated in the past, namely from past uses of human natural resources. Then Kolm refers to Locke (1689) in order to emphasise that his own conclusion resides in an old tradition: as human productive capacities are by far the main source (97; 5% D 80% C 17; 5%) of the total value, to which non human resources makes a marginal contribution (2; 5% D 2% C 0; 5%), the analysis of macrojustice must concentrate on human natural resources only. – In other respects the distribution of the nonhuman natural resources essentially follows from considerations of microjustice, because of the so many specific or local factors which can determine their appropriation, i.e. unanimous agreement which are more or less explicitly formulated, right of first occupancy, attention either to egalitarian considerations or to aims of general interest, and so on. – Finally, much of the natural resources were possessed by many successive owners (many present owners of land or stocks in mining companies bought them from previous owners). “For them, there was no relevant 29

See also Kolm (1985) and Kolm (1986, pp. 207–241); in Kolm (1985) the whole Chap. 10 (pp. 127–168) is dedicated to “the distribution of natural [nonhuman] resources”.

166

C. Gamel

difference between buying these assets or property in produced goods” (Kolm 2005, p. 88). It would therefore be difficult to organise an appropriate distribution of nonhuman natural resources, by only taxing the present owners of these resources. Besides, even if the taxation of these nonhuman resources is conceivable, Kolm (2005, p. 89) admits that “this can be done for produced goods as well”. It seems to us that the first two series of remarks stem from a rather non convincing argumentation: even if the relative weight of human resources at the origin of value is very large, considering nonhuman resources within the framework of macrojustice theory cannot basically be disqualified only because their importance would be marginal. Besides, in the history of human societies, the complexity of the institutional arrangements concerning the appropriation of these resources do not necessarily imply that they must be taken out of the field of macrojustice. On the contrary, it is first dubious that these arrangements at the micro-social level should be the result of a process always considered as justified or fair, and secondly, the allocation of the property rights of the nonhuman natural resources is of course a determining factor for the good process of classical liberal societies30 ; but this allocation is equally decisive for the complete exercise of “social freedom” in Kolm’s acceptance of the expression (“freedom from others’ forceful interference”): the full ownership of a natural resource, whether human or nonhuman, protects the owner but excludes others from any enjoyment and enforces a major social “interference” on them. 5.3.1.3 Whether External or Internal, “Given” Endowments Subjected to “Moral Arbitrariness” In other words, the main question is: Why is it necessary to keep the original solution of macrojustice only for human natural resources (dismembering of self-ownership and social levy – at level k – of the rent which stems from it)? While the same principle would certainly be considered to be far less audacious if it were applied to the nonhuman part of natural resources (dismembering of property of these resources and social levy – at level k – of the rent which stems from them). Here is the third line of Kolm’s argumentation (see Sect. 5.3.1.2) that seems to us a decisive one, because it could

30

See for instance Nozick’s famous principles of “justice in acquisition”, “justice in transfer” and even “rectification”; according to this author they are essential for entitling the property of nonhuman assets (Nozick 1974, pp. 150–153).

5

Basic Income and ELIE Transfers

167

accept philosophical complementarities between macrojustice theory and Van Parijs’ approach. Kolm (2005, p. 89) notices in a few words that nonhuman natural resources cannot be easily processed in a different way from other nonhuman assets (either financial or manufactured) which constitute the nonhuman wealth of the individuals. From the conventional economic and fiscal point of view, we do not indeed process differently patrimonial assets whether they are natural or not. So we are far from concluding with him that, in spite of this closeness, any nonhuman resource remains disqualified from the point of view of macrojustice theory. We must emphasise on the contrary that this economic and fiscal proximity between natural and non natural resources is coupled with an even deeper ethical proximity: both resources were “given” to the individual (by legacy or inheritance), so as to make what Van Parijs just calls “external endowments”. Whether external resources coming from inheritance or gift are natural (landed property) or are artificial (productive or financial assets), their difference fades with the increase of wealth which results from this inheritance or gift, while the recipient cannot claim the full property of assets as producer or buyer: he is in possession of both types of assets only because he had “brute good luck” – in the “dworkinian” meaning of that term.31 At this point, it seems conceivable to look for an extension of macrojustice theory that would goes well beyond the only internal resources. In this prospect, a comprehensive study still has to be done to really know whether internal and external endowments are complementary from the philosophical point of view of macrojustice theory. Within the framework of that “kolmian” theory, if Kolm himself had wanted to include non human natural resources in it, it seems to us that no essential difficulty would have emerged from these complementarities: As macrojustice theory is based on the “revolutionary” dismembering of self-ownership,32 it could admit as well a more “classical” dismembering of people’s property rights over the nonhuman part of their wealth. However, as philosophical complementarities of internal and external endowments are first conceived within the framework of a theory of justice, it is useful, in our opinion, that not only “dworkinian” but also “rawlsian”

31

“I shall distinguish, at least for the moment, between two kinds of luck. Option luck is a matter of how deliberate and calculated gambles turn out – whether someone gains or loses through accepting an isolated risk he or she should have anticipated and might have declined. Brute luck is a matter of how risks fall out that are not in that sense deliberate gambles. If I buy a stock on the exchange that rises, then my option luck is good. If I am hit by a falling meteorite whose course could not have been predicted, then my bad luck is brute (even though I could have moved just before it struck if I had any reason to know where it would strike)”. (Dworkin 1981, p. 293). 32 On this point, see Gamel and Lubrano (this volume, Sect. 1.5.1).

168

C. Gamel

arguments should be put forward. Indeed, since the moral arbitrariness of the “natural lottery” of personal talents can also justify the fact that individuals should share with others their own productive capacities (internal endowments inherited from parents and valued by education), the moral arbitrariness of the social birth sphere should also induce them to agree to share their nonhuman productive capacities (external endowments extracted from nature or inherited from the past generations). Here, it doesn’t matter very much that these productive resources (“internal” or “external”) are “natural” (great intellect or oilfield), in so far they have been “given” and “received” through inheritance (stocks portfolio) or education (school fees financing). In both cases we cannot consider that the individual morally deserves the lucky or unlucky fate he gets according to his initial resources. To better emphasise the importance of “given” or “inherited” assets (whether they are “internal” or “external”), we will use, as Van Parijs does,33 the term “endowments” rather than any other term which could be considered at first sight as equivalent (resources, capacities, opportunities,...); as a consequence, we shall also avoid, with few exceptions, the use of those terms later on in the present text. Finally, it results from the previous developments that Van Parijs’s and Kolm’s points of view were initially very different: the sharing of external endowments on which the first one theoretically justifies basic income has at first sight nothing to do with the personal productive capacities that the second one turns into the fiscal basis of ELIE transfers. Upon further examination, we note a first relative convergence of both authors: Kolm’s human natural resources are included in Van Parijs’s internal endowments, even if the aims that justify why they exploit these notions remain different (fiscal resource for the first one, source of handicaps to be compensated for the second one). But the main result is actually to be found elsewhere, as the two authors complement each other when they study the problem of distributive justice concerning the allocation of the endowments that are “given” to the individuals: whether they are external according to Van Parijs (“inherited” nonhuman resources) or internal according to Kolm (“received” human productive capacities), the intellectual processes of both authors have thereby a common denominator which does matter at the philosophical level and can make them perfectly compatible.

33

“I prefer to use the term ‘endowments’ where others might have used, for example, ‘resources’, ‘opportunities’, or ‘capacities’ to stress the restriction to what people have been ‘given’, whether at the start or later on, whether intentionally or unwittingly, whether in the form of external goods and purchasing power or as bodily and mental features” (Van Parijs 1995, p. 251, footnote 1).

5

Basic Income and ELIE Transfers

169

In that case, a thorough study of this compatibility at the economic level needs to be done: the challenge is to conceive an income redistribution which would remain coherent but would be more complete: in addition to ELIE transfers, it would contain TECIE transfers – “Totally Exploited Capital Income Equalisation” – that we have now to define.

5.3.2 “TECIE” Transfers in Addition to “ELIE” Transfers The starting point of our attempt is indeed to give Van Parijs’ external endowments the fiscal treatment which Kolm keeps for the internal endowments.34 In this prospect, we have to consider all the consequences of the complementarities which have been tracked down at the philosophical level between both types of endowments: In so far as Kolm succeeds in legitimising, by dismembering of self-ownership, that everybody has to share with all other people the rent of “received” internal endowments, it must be a priori possible to justify, by a similar dismembering of wealth ownership, that everybody has to share with others the rent of “inherited” external endowments. 5.3.2.1 The Rent Sharing of External Endowments: “TECIE” Transfers More precisely, let us go back now to the ternary legal distinction dividing the property rights into three attributes, which has been already presented in the first chapter of this book (Sect. 1.5.1): By the clause of “abusus”, everybody partly remains (as “bare-owner”) the owner of himself as well as the owner of the assets he has, since in both cases he can alienate them within legal limits; by the clause of “usus” he remains also the sole holder of the right to exploit more or less intensively his internal endowments (human capital) as well as his external endowments (nonhuman assets). On the contrary, the “fructus” linked to these two categories of resources, which entitles everybody to the rent incomes both of them can produce, must be shared between the person concerned and the community in proportion of a k “parameter of equalisation”. So the individual has to pay a fraction k of rent incomes which could be gained from the productive capacities of his external endowments; in return

34

Henceforth internal endowments will be reduced to “human natural resources” in Kolm’s sense of this expression: productive capacities of the individuals are only considered and eudemonistic capacities are excluded (see Sect. 5.3.1.2).

170

C. Gamel

he must receive a fraction k of similar rent incomes that are linked to the productive capacities of all other people’s external endowments. Let Ci be the estimate of the external endowments of individual i, that is the money value of his nonhuman wealth (“cut off” from nature or “inherited” from past generations) and let ri be its reference rate of return, would this nonhuman wealth be completely exploited and valued; the product ri Ci is thus the ci “full” rent of the individual’s external endowments.P Moreover let CN be the mean value of all the external endowments C D Ci of n N individuals, so that we have: C D C =n ; finally let rN be the mean rate of return of C , if C were fully exploited; the product rN CN is thereby the “full” average rent income cN of the community’s members. So individual i has to discharge his contribution k ci D k ri Ci and to receive the amount k cN D k rN CN , which implies a new transfer ti0 coming from external endowments; nevertheless these transfers ti0 are mimicked from the model of ELIE transfers ti , which are based on internal endowments. Finally we have the following formula: ti0 D k.cN  ci /: This transfer ti0 means equalising at the level k rent incomes which could have been produced by the individual’s external endowments, if they were completely exploited; hence we call this transfer “Totally Exploited Capital Income Equalisation”, in abbreviated form TECIE. When the individual holds external endowments that yield a ci full rent income which is above the cN average full rent income, he becomes a net payer to the TECIE transfers scheme; in the opposite case, he becomes a net payee. At this stage of our presentation, we can observe the first difference between ELIE transfers and TECIE transfers. Concerning internal endowments or personal productive capacities, Kolm uses, as a basis of the equalisation to be operated, incomes stemming from everybody’s full-time work, so that, in our opinion, ELIE transfers could as well have been called FTLIE transfers (“Full-Time Labour Income Equalisation”). However, as maximal working-time durations are quite comparable among individuals, it is easy to define what can be called full-time work; as a consequence Kolm rightly chose to stress on the fact that, concerning this quantitative dimension, the margins of use of personal productive capacities are relatively homogeneous and even equal, so that only “Equal Labour Income” has to be equalised. Moreover, in the basic model of macrojustice theory, he considered that the maximum full-time working time is equal to 1, which leads to remove from the equation of ELIE transfers the quantity of labour (` D 1) on which the equalisation has to operate: only the wage rate wi remains Œti D k.w N  wi /, as a signal of the unit quality of internal endowments that distinguishes everybody.

5

Basic Income and ELIE Transfers

171

In other words, the maximal intensity of use of received internal endowments comes up against physiological limits concerning the volume of work that can be produced; as a consequence, those limits are nearly common to all individuals. As regards the maximum intensity of use of external endowments, there is on the contrary no maximal volume of inherited wealth which would be common to all the people. By the way, we realise that internal endowments are different among individuals only in a single dimension (their quality which determines the wage rate wi ), while external endowments differ not only in quality described by the rate of return ri (which depends on the nature of inherited assets), but also in quantity (the volume of nonhuman capital Ci ). This last remark provides also a new ethical reason for integrating external endowments into macrojustice theory in spite of the lower incomes they provide, which was the reason why Kolm put them aside (see Sect. 5.3.1.2). When the heterogeneity of external endowments is compared with that of internal endowments, the question is not reduced to the previous features. Here appears a second difference between ELIE and TECIE transfers. Concerning the quality of endowments, the level of heterogeneity is all the more increased because the composition of external endowments is much more diversified than that of internal endowments. Except in special cases, the internal endowments of the individual, that is to say the skills transferred to him, are mostly homogeneous, which induces him to value them only on a single labour market, at the current wage rate wi on this market. The same is not to be observed for external endowments: the inherited wealth can often consist of several kinds of assets, from the most liquid ones (cash) to the most non-liquid ones (real or land estate) and, between these two types, various intermediate kinds of assets can also be found (financial assets, works of art, and so on). Therefore, for each category j of wealth which is part of individual i’s external endowments, we should estimate its amount Cij and especially the rij rate of return which can be obtained if the asset is totally exploited. In some cases, this estimation is easy (interest rate or rate of return on all kinds of transferable securities), in other cases the estimation is more difficult but quite possible (rate of return of land or real estate if they were rented, or the rate of return that the cash part of wealth could yield if it were converted into liquid investments). Finally, if there were no organised renting market, some hoarding values (works of art, jewels, bullions,. . . ) would yield only non monetary returns (enjoyment of holding them or pleasure of admiring their beauty, for example), even though the owners look forward to possible capital gains in case of sale or resale; in others words, their monetary return rij would be equal to zero, even if their market value Cij is not.

172

C. Gamel

In any case, it remains possible to estimate for each individual i the total value ofP his external endowments through the value of his various assets Cij .Ci D Cij /; at the same time, we can define the rate of return ri as the average of the rates of return rij , as far as this average is weighted according to the relative value of every asset Cij in the total amount of external endowments Ci . 5.3.2.2 “ELIE Transfers” and “TECIE Transfers”: Two Formally Similar Processes Once these differences between internal endowments and external endowments are noted, the development of TECIE transfers can follow the same three stages process as for ELIE transfers35 : searching for an agreement on TECIE transfers based on individuals’ external endowments, searching for an agreement on the intensity of the equalisation to be operated on these endowments (choice of a parameter k of equalisation) and finally everyone’s freedom of exploiting more or less intensively the income-yielding capacities of his external endowments. If we suppose that a more consensual way of choosing the parameter k of equalisation is to be found between both extreme positions (0 < k < 1), Fig. 5.1 below transposes into TECIE transfers the synthetic presentation of “the geometry of ELIE” which illustrates the three stages of this process.36 The principle of this transposition is as follows: while the exploitation of the concerned individual’s internal endowments was monitored by his “incomeleisure” trade-off, the exploitation of his external endowments hereafter depends on an “income-fallow” trade-off: – In the first case (“geometry of ELIE”), the net income yi generated by internal endowments is lower than the full-time wage wi in an inverse ratio to the opportunity cost of leisure time . If leisure time  is calibrated between 0 and 1, the equation of his budget line can be written as: yi D wi  wi i D `i wi ; the actual income of individual i depends on the level of working time (`i D 1  i ), during which he exploits the personal productive capacities he has received. – In the second case (“geometry of TECIE”), we are interested in the net income yi ensured by the external endowments of individual i; yi is lower than the maximal rent income ci in an inverse ratio to the relative part of endowments he has left unexploited. Let us note this relative part the “rate

35 36

See Gamel and Lubrano (this volume, Sect. 1.3). See Fig. 1.4 in Gamel and Lubrano (this volume, Appendix B).

5

Basic Income and ELIE Transfers

173

of fallow” ˇ 37 ; on Fig. 5.1 below, the equation of the budget constraint is now: yi D ci  ci ˇi D ei ci . The actual income of individual i depends on the intensity of exploitation (ei D 1  ˇi ) of the nonhuman assets he has inherited. If we go on with the analysis of Fig. 5.1, the central category of the population is the part of the population for which the full rent income (cce D rce Cce ) of external endowments would be equal to the average full rent income of the whole population (cN D rN CN ), would external endowments be totally exploited (e D 1). As a consequence, if the parameter k of equalisation is equal to 0.4, an individual belonging to this central category has to pay 40% of the potential rent income of his external endowments (k cce ) and has to receive in return 40% of the potential average rent income of the whole population (kN c); N the net transfer from which he does profit is therefore sys0 D k.cN  c / D 0. The individual is neither net tematically equal to 0 Œtce ce payer nor net payee of TECIE transfers, whatever the rate of exploitation ei he chooses; in other words, the intensity with which he decides to extract an income from his external endowments does not change anything. Individuals not belonging to this central category either finance a negative net transfer or receive a positive net transfer. As the external endowments of the first ones can be more productive than the average (cmp > c), N they have to 0 N Hence we have: tmp D k.cN  cmp / < 0; pay k cmp and they receive only k c. conversely, because of potentially less productive external endowments, the tax k clp of the second ones will be lower than the benefit they will receive 0 D k.cN  clp / > 0. On Fig. 5.1 the situations of (k c), N which induces: tlp each of the three successively listed categories of individuals are compared and nobody can fail to have noticed the parallel between “the geometry” of TECIE and “the geometry” of ELIE: in both cases, the budget constraint of the central category is invariant, the budget line of the individuals with potentially most productive external endowments is translated downward (negative net transfer) and that of individuals with potentially least productive endowments is translated upward (positive net transfer); by construction, the whole result of these similar observations consists in the intersection of these lines at point K, the coordinates of which are (1  k; k w) N for ELIE transfers and (1  k; k c) N for TECIE transfers. Our analysis of Fig. 5.1 can also be refined by comparing the individuals according to another criterion, the intensity ei with which they exploit their

37

We use the word “fallow” only in a symbolic and illustrative way, which emphasises that any kind of asset could be kept unexploited, wholly or partly. As a consequence the agronomic point of view concerning the rate of fallow of the farmlands is not considered.

174

C. Gamel

c;y c mp

 <0 t mp

y mp t  = k (¯ c − c) c¯;c ce yce ylp c lp

K

k c¯ t lp > 0

0

D C

¯ = 0.1

1−¯ = e = 0. 9 1−k

B 1

0.6

k = 0. 4

¯

E

Fig. 5.1 The geometry of TECIE

external endowments, instead of the level of their external endowments Ci : Let us consider three individuals belonging to a different category cmp , cce or clp . If they all choose the same rate of exploitation equal to k (k D 0:4), by construction (see point K) they all get the same net income k cN out of external endowments which are still quite different. If each of them chooses a same rate of exploitation above k, they all earn an income which is higher than k c, N but this income is different among them, because it is conditional upon the potential rent income of their own external endowments. As Fig. 5.1 shows with the hypothesis of a high rate of exploitation (e D 0:9), we have then: ymp > yce > ylp ; after transfers t 0 , individuals who are best endowed with external assets still have the highest incomes, because the incomes they

5

Basic Income and ELIE Transfers

175

earn through the “overexploitation” of their external endowments well above k (e > k) remain completely tax-exempt.38 We would obtain exactly the opposite result if our three fellow men would choose a rate of exploitation of external endowments lower than k: on the one hand everyone would get an income lower than the allocation k cN he is entitled to, on the other hand the position of the three incomes would be inverted (ymp < yce < ylp ). All of them are penalised by the “underexploitation” (below k) of their external endowments, but, as they have to discharge a lump sum tax k ci which is proportional to the potential rent income they can get from these endowments, in this case the best endowed people earn the weakest net incomes (after TECIE transfers). We get an extreme situation when our three fellow men do not exploit at all their external endowments that lie thus in full fallow: in that case, after TECIE transfers the individual of the central category has a net income equal to 0 (point B on Fig. 5.1); the individuals with the lowest external endowments have a positive income which is reduced to the positive TECIE transfer Œk.cN  clp / > 0 and varies according to the “under-endowment” of which they are victims (segment BD). The individuals with the highest external endowments have to discharge a lump sum tax Œk.cN  cmp / < 0 which varies according to the “overendowment” they enjoy (segment BE). As these endowments are not at all exploited, they have to finance the tax in another way than the rent income they could have drawn from them. Such behaviours are those of “fallow-land owners” which amount in case of external endowments to the behaviours of “eccentric productive people” who do not work while they do have high internal endowments (see Sect. 5.2.2). 5.3.2.3 “Eccentric Productive People” and “Fallow-Land Owners”: An Incomplete Symmetry The question is then to know whether both situations are completely comparable: are “fallow-land owners” compelled to work in order to finance their More exactly, the income yi an individual i can get out of his external endowments at the same time depends on the gross income ei ci their exploitation yields and on the TECIE transfer, which includes the received allocation k cNi and the paid contribution k ci ; in short we have: yi D ei ci C k cN  k ci or: yi D .ei  k/ci C k c. N The first term is the individualised part of his income coming from the “overexploitation” (above k) of his external endowments; the second part defines the socialised part of income – the same equal allocation k cN everybody is entitled to in return of the personal lump-sum contribution k ci paid in full settlement. As a consequence, for any rate of exploitation ei chosen above k, the individual’s additional incomes coming from external endowments are tax-exempt. For TECIE transfers, we thus find the same property of inelasticity of the contribution in relation to people’s behaviour that Kolm already emphasised for ELIE transfers. 38

176

C. Gamel

negative TECIE transfers, just in the same way as the “eccentric productive people” who had to be provided with capital incomes in order to finance their negative ELIE transfers? In that case, the blow to “social freedom” for ELIE payers would be of the same order as the blow to “real freedom” for TECIE payers. In other words, “fallow-land owners” could not really have fallow-lands while “eccentric productive people” could not really stand “idle”, because the society would enjoin on everyone a minimal exploitation of his endowments, whether they are external or internal. Before answering this question, we can note, as external endowments are logically very different from internal endowments, that both situations are not completely symmetric.39 As already emphasised (see Sect. 5.3.2.1), external endowments are quite more heterogeneous than internal endowments; therefore the individual can pay TECIE transfers by exploiting only a part of his assets (the most profitable ones) in an intensive way, which allows him to keep another part unexploited: For example, if the “fallow-land owner” is a feudal lord unwilling to share the magnificence of his real estate preventing paying visitors from seeing or hiring the place, he ought to exploit in a more intensive way (eij > k) other elements j which are part of his fortune (cultivation of a greater part of the surrounding lands that lie “fallow”, or even investments of the sale proceeds of hoarded assets such as works of art, gems and so on). Now we have also to consider two facts: on the one hand, the rent income ci fixing the level of TECIE transfers Œk.cN  ci / takes only into account the initial structure of the inherited endowments (and not the further level and structure of the fortune); on the other hand, parameter k of equalisation does not usually enforce a rate of exploitation eij which is very high. As a consequence of these two facts, the degree of freedom secured by restructuring inherited external endowments is all the more high because the part of hoarded or less profitable assets was also originally high. Of course such a degree of freedom does not exist for “eccentric productive people” whose internal endowments are mostly very homogeneous. Another difference exists for the respective rates of accumulation of human wealth and nonhuman wealth, which include internal endowments in the first case and external endowments in the second case: in the framework that we develop, redistributing a part of the rent incomes (at the level k) is justified by the fact that the individual has been “given” these endowments either by education or by inheritance (see Sect. 5.3.1.1) and the rent income 39

For paying for negative TECIE transfers, we leave out the false solution which would consist in alienating the external endowments by putting them up for sale; such possibilities cannot be considered for paying ELIE transfers with internal endowments, since of course slavery is excluded.

5

Basic Income and ELIE Transfers

177

is then worked out according to the volume and the structure of “received” endowments. In other words, if the individual himself manages to get richer later on by accumulating new (human or nonhuman) assets, the resulting increase of his wealth remains outside the framework of our extended theory of macrojustice, since the various incomes stemming from this increase are not to be considered. Hence, wage rises which are connected later on to work experience or additional training efforts must not be bound by ELIE transfers, as well as TECIE transfers must not take into account the capital incomes yielded by the individual’s personal further investments. Now, human capital and nonhuman capital do not follow the same rate of accumulation throughout life: the first type is especially concentrated in the earliest ages of life (school education, on the job training at the beginning of the professional career), before physical fatigue or intellectual tiredness reduce the investments made (in the form of in-service training or retraining); so the part of internal endowments in an individual’s human capital is a deciding factor throughout his life. On the contrary external endowments might be spread more easily over the life cycle. On the one hand, the personal gathering of nonhuman assets is progressive and is not as concentrated on a particular segment of life, on the other hand, the relative weight of external endowments depends whether the individual is still young or already old when he receives them; the second possibility tends to be rather more frequent owing to the increase of life expectation, so that a generation may be skipped (when grandchildren directly inherit external endowments from grandparents). In other words, it seems easier to pay taxes for financing TECIE transfers, since the taxation is worked out according to external endowments which are more often inherited at very variable ages of life. On the other hand, taxation for financing ELIE transfers depends on internal endowments which are mostly received at the earliest ages of life, so that their relative weight in the individuals’ human capital does not decrease rapidly. Owing to this second difference also, the social freedom of “fallow-land owners”’ seems better secured than that of “eccentric productive people”. However, after these last remarks, our evaluation which separates ELIE transfers from TECIE transfers is now exhausted: far from replacing the first ones by the second ones, the aim of this article is indeed to superimpose TECIE transfers on ELIE transfers. As both appear to be complementary from the point of view of distributive justice, our comment is now to investigate the basis of a double scheme of income redistribution, which would be more complete than Van Parijs’s and Kolm’s single systems. Through the combined effects of this double scheme, ELIE transfers and TECIE transfers could perhaps diminish their respective failings and find an additional legitimacy. More generally speaking, we have to draw the first conclusions from

178

C. Gamel

this “hybridisation” of Van Parijs’ and Kolm’s systems so as to estimate its possible fruitfulness for further investigations.

5.4 Conclusion: Could “Hybridisation” be Fruitful? The fact that our conclusion can concern the degree of fruitfulness to be expected from the established link between such elaborated intellectual processes as Van Parijs’ and Kolm’s is in itself one first result: it was far from being reached when we started this rather bold intellectual adventure. Except for the visible nearness between Van Parijs’ “real freedom” and Kolm’s “social freedom”, the only strong point we could originally rely on was Kolm’s staunch conviction, that ELIE transfers could be interpreted as an accurately financed basic income. Far from confirming such an assertion, the first part of our research somewhat emphasised a basic difference which does not amount to an “ordinary” financing problem: the two authors also disagree about the contribution of labour as a factor of social integration, as exemplified by the case of “eccentric productive people” which is very revealing even though it is rather special. The pursuit of our reflection on the feasible compatibility of two such different projects could have come to a deadlock. However the careful reading of Van Parijs’ and Kolm’s theses reveals that, beyond a common culture on the modern theories of social justice, they really have a common objective, that is the equalisation of the incidence of wealth (in the broad sense of this term) by means of transfers which would at the same time be fair and efficient; but, in this prospect, the two authors complement each other, since they are situated on both sides of the “dworkinian” split between external resources and internal resources. From then on, our task consisted at first in supporting this guess by emphasising how much Kolm’s natural resources fitted to Van Parijs’ internal endowments, although each author does not give them the same status: natural resources provide the tax base for ELIE transfers according to Kolm, whereas, according to Van Parijs, internal endowments are at the origin of handicaps for which basic income should provide compensation. How much Van Parijs and Kolm complement each other can then be verified through the external endowments for which Van Parijs does not really succeed in finding financing for basic income, while Kolm neglects them completely without advancing enough convincing reasons. As a consequence the route was open to apply to Van Parijs’ external endowments the method used by Kolm only for internal endowments, with the aim of drawing from this a more adapted way for the financing of basic

5

Basic Income and ELIE Transfers

179

income. As TECIE transfers are the result of this investigation, the question is now whether these transfers can take up such a challenge, without distorting the property of basic income, that is to say its basic neutrality between work and non-work. It was indeed the end we expected, but after all it does not seem to be completely achieved (Sect. 5.4.1). Nevertheless the superposition of ELIE and TECIE transfers results in a new prospect initially underestimated: it concerns a system of transfers the hybridisation of which, far from being an inconvenience, appears to be a factor of strong consistency (Sect. 5.4.2). Let us recapitulate successively both points of this provisional evaluation.

5.4.1 Basic Income Financing Through TECIE Transfers: A Cautious Evaluation Table 5.1 below lists nine possible situations which result from the combination of ELIE transfers and TECIE transfers: as the case may be, the individual can use internal endowments yielding a full-time wage which is higher (wmp ), equal (wce ) or lower (wlp ) than the full-time average wage (w); N he can also use external endowments yielding a full rent income which N is higher (cmp ), equal (cce ) or lower (clp ) than the full average rent income c. Let us leave aside, for the while, the intermediate situations of Table 5.1 N and 2nd line: cce D c: N at best, in these central cat(2nd column: wce D w egories of the population, individuals perceive, as net payees or net payers, the existence of only one out of the two transfers (cells 2, 4, 6 and 8); as for individuals of cell 5, which is a central category in two ways, they perceive neither ELIE transfers nor TECIE transfers (t D t 0 D 0). Therefore, in all these cases the superposition of ELIE transfers and TECIE transfers is not so easy to identify and their respective evaluation seems at first sight separate. Therefore the most characteristic situations are to be found in the four “corners” of Table 5.1: superposition of ELIE and TECIE benefits (cell 1) by “systematically under-endowed” individuals and, in contrast to the first diagonal, superposition of ELIE and TECIE taxes (cell 9) by “systematically over-endowed” individuals; in the other diagonal, compensation of ELIE transfers by TECIE transfers (and vice versa), both transfers being quoted with algebraic opposite signs (cells 3 and 7). From the framework of Table 5.1, let us superpose the cases where “eccentric productive people” could appear, cases where those people would be forced to work because of N This constraint is obvinegative ELIE transfers (3rd column: wmp > w). ously relaxed in cell 3, because receiving a TECIE benefit could provide the concerned individuals with the means necessary for paying the ELIE tax,

180

C. Gamel

Table 5.1 The joint effect of ELIE transfers and TECIE transfers [According to the levels of full-time wage (wmp , wce , wlp ) and full rent income (cmp , cce , clp )] wlp < w N -1ELIE: positive, TECIE: positive

wce D w N -2ELIE: zero, TECIE: positive

wmp > w N -3ELIE: negative, TECIE: positive

cce D cN

-4ELIE: positive, TECIE, zero

-5ELIE: zero, TECIE: zero

-6ELIE: negative, TECIE: zero

cmp > cN

-7ELIE: positive, TECIE: negative

-8ELIE: zero, TECIE: negative

-9ELIE: negative, TECIE: negative

clp < cN

wholly or partly. As a consequence “eccentric productive people” without any family inheritance might also choose more easily to remain unoccupied; in this case the basic neutrality between work and non-work, which is a typical feature of basic income, is reinforced by the hybridisation of ELIE and TECIE transfers. This first result a priori does not seem to be confirmed by the situation of cell 9, which is another characteristic cell of the 3rd column: far from being able to compensate for the financing of ELIE transfer with the benefit of a TECIE transfer, the individual has to be considered as a net payer at the same time in both systems; in this case “eccentric productive people” should work not only against their will, but they should make sure as well that the exploitation of their external endowments is high enough to finance a TECIE transfer. Nevertheless, if necessary, a degree of freedom remains and should not be neglected: if the value of k allows it, the individual, while “over-exploiting” (emp > k) large external endowments (cmp > c), N might finance simultaneously both taxes without having to work for all that. The case of eccentric productive people “over-exploiting” external endowments N might also appear in cell 6: despite medium-sized endowments (cce D c), the TECIE transfer to be financed is just equal to zero.40 Finally, the hybridisation of ELIE and TECIE transfers does not secure in an absolute and equivalent way the basic neutrality between work and non-work specific to basic income; but this hybridisation provides with significant margin all “eccentric productive people” who would not like to work, whatever their situation (cells 3, 6 and 9): this solution protects their “real 40

Such an overexploitation is even possible in cell 3, wouldn’t positive TECIE transfer be enough; however in this case, low external endowments that would explain a positive TECIE transfer (clp < c) N do not mean endowments that would be equal to zero (clp D 0).

5

Basic Income and ELIE Transfers

181

freedom” in a better way than Kolm’s ELIE transfers alone, while seeming to provide them with better guarantees of financing than the exploitation of Van Parijs’ “employment rents”.

5.4.2 Hybridisation of Transfers: A Prospect to be Thoroughly Studied More generally speaking, the hybridisation of ELIE and TECIE transfers seems to lay the foundations for a redistributive system that would be both different and more complete than those from which it is inspired. For that purpose let us carry on the analysis of the other characteristic cells of Table 5.1. N cells 7, 8 and 9 are those where a On 3rd line of the figure (cmp > c), TECIE transfer to be paid can restrain the freedom of “fallow-land owners” and prevent them from keeping external endowments unexploited. However the hybridisation gives them prospects which are symmetric of those of “eccentric productive people”; in cell 7, the constraint is relaxed, wholly or partly, owing to the receiving of a positive ELIE transfer: the young man with high private means but limited personal productive capacities (wlp < w) N might keep “fallow-lands” in the nonhuman wealth he inherited. In cell 8, a degree of freedom is also to be found: as the young man has N the ELIE transfer in this case middle-sized internal endowments (wce D w), he is entitled to is indeed equal to zero but, if the k parameter of equalisation is not too high, he can possibly “over-exploit” his personal productive capacities by working a lot (working time `i > k), without giving up his behaviour of “fallow-land owners”. It is the same in cell 9 where personal N the same strategy productive capacities are over the average (wmp > w); of overexploitation could allow him to finance simultaneously both types of transfers. Finally, the fallow-land owners, whatever their situations are (cells 7, 8 and 9), are not inevitably compelled to abandon their favourite lifestyle: their “social freedom” is better protected by adding Kolm’s ELIE transfers, because they can diversify the financing of the redistribution beyond Van Parijs’ taxation of the only external endowments. We shall note however that the hybridisation of ELIE and TECIE transfers reaches its limits in cell 9, where the concerned young man with high private means might have to choose either the attitude of “eccentric productive people” or the behaviour of “fallow-land owners”: If the level of the k parameter of equalisation is not too high, he can indeed choose either not to work while overexploiting his external endowments or to let his inheritance lie fallow while working a lot; but having to finance both ELIE and TECIE transfers

182

C. Gamel

at the same time prevents him from adopting both attitudes simultaneously. Is this dilemma really unbearable for him? Indeed we can think yes, if we consider that both taxes to be paid are already a sufficient counterpart of his double “brute good luck”. His double luck being his having drawn two winning numbers in the “natural lottery” and in the “social lottery”; in other words, his having been given internal endowments and external endowments which are both over the average. We can as well think no and consider that his being the elected representative of the Providence finally imposes a small restriction of freedom on him and on very few other people, in the very unlikely event of their stubbornly refusing to choose between “idleness” and “fallow”. On the opposite end of the diagonal line, cell 1 corresponds with the situation where, on the contrary, the individuals are badly treated by fate and have internal endowments and external endowments that are lower than the average; they thus receive in return two positive ELIE and TECIE transfers. As far as the accumulation of these transfers would be enough for them, they also have a unique “privilege”, which is keeping their low endowments simultaneously unexploited (by idleness and fallow); but we can doubt whether such an opportunity could be in all cases considered in this way, since it would amount to neglecting the strong incentive power of such a system of redistribution: on the one hand this system would logically exempt from any charge incomes that the individual would draw from his low initial endowments if nevertheless he would decide to exploit them, on the other hand, it would also exempt incomes he could draw by his own efforts (savings or training) from increased internal or external resources above their initial amount. In short, the hybridisation of ELIE and TECIE transfers would be worth studying thoroughly beyond the main outlines sketched out above. Whether such a system would be applicable obviously raises many so far unanswered questions, concerning the fiscal data to be collected to reconstitute everyone’s external endowments and TECIE transfers; but, since sophisticated tax systems on the inheritance are already applied in developed countries, these data seem a priori less difficult to gather than statistics on internal endowments and Kolm’s ELIE transfers. Nevertheless hybridisation raises new and not obvious questions and would imply a detailed investigation: for example, is it really required to have the same parameter of equalisation for both types of transfers or is it useful to have a different k parameter for transfers of each type (ELIE and TECIE)? From another point of view, are everyone’s full rent incomes wi and ci to be fixed once for all, according to the initial level of his internal and external endowments, or is it conceivable that the re-appraisal of these data

5

Basic Income and ELIE Transfers

183

would be processed in exceptional cases, according to the lucky or unlucky chance events that happen throughout his life? To answer to those questions, the main point is to keep in mind the ethical consistency of this hybrid system, which aims for both fair and efficient redistribution of everyone’s initial endowments, whether they are internal or external. This redistribution thus concerns neither all the incomes the individuals later receive nor all the assets they accumulate throughout their life. However it is fundamental that the object of redistribution should be only the rent incomes of the initial endowments, because these endowments are resources that individuals receive but do not merit: They cannot be held responsible for the initial endowments that they inherit with good luck or bad luck, while fundamentally differences in these initial endowments induce a lack of “equality of opportunity” (in the deep sense of the expression) among individuals throughout their life. That is why the best summary of the ethics of ELIE and TECIE hybridised transfers is perhaps to be found in a reformulation of the notion of “democratic equality” that Rawls introduced more than 35 years ago in A Theory of Justice” (1971, Sect. 12). According to him, “democratic equality” results from combining the “equality of fair opportunity” which fights against disparities induced by differences in social positions and the “difference principle” which compensates for the disparities from natural origin: from Rawls’ moral point of view both disparities are quite equally “arbitrary”. As far as we are concerned, we will not try to distinguish between “nature” and “society” or even between “innate” and “acquired” characteristics. In each case both elements are highly interwoven when they are transmitted to the children within the family unit, whether the transmitted “values” are material, immaterial or even moral. On the other hand this complex blend provides a decisive argument for considering that internal endowments and external endowments must be treated in the same way: therefore not only internal endowments through ELIE transfers but also external endowments through TECIE transfers ought to be equalised. After all, if Kolm and Van Parijs agree to appear as inseparable heirs of this revised “rawlsian” philosophy, our own research on the compatibility between basic income and ELIE transfers would benefit from a happy presumption. With the basic support of both of them, our investigation of the hybridised transfers could be continued under the most favourable auspices: this indeed is the only wish we finally dare to express. . .

184

C. Gamel

References Ackerman, B. (1980). Social justice in the liberal state. New Haven: Yale University Press. Atkinson, A. (1995). Public economics in action: The basic income/flat tax proposal. Oxford: Oxford University Press. Dworkin, R. (1981). What is equality? Part 1: Equality of resources. Philosophy and Public Affairs, 10(4), 283–345. Gamel, C. (2004). Comment financer l’allocation universelle ? La stratégie de Van Parijs en question. Recherches Economiques de Louvain, 70(3), 287–315. Gamel, C., Balsan, D., & Vero, J. (2006). The impact of basic income on the propensity to work: Theoretical issues and microeconometric results. Journal of Socio-Economics, 35(3), 476–497. Kolm, S.-C. (1972). Justice et équité. Paris: CNRS. English translation Justice and equity, (1997), Cambridge: MIT Press. Kolm, S.-C. (1985). Le contrat social libéral. Paris: Presses Universitaires de France. Kolm, S.-C. (1986). L’allocation des ressources naturelles et le libéralisme. Revue Economique, 37, 207–241. Kolm, S.-C. (2005). Macrojustice, The political economy of fairness. Cambridge: Cambridge University Press. Kolm, S.-C. (2006). Liberté, justice et efficacité: distribution, impôts et transferts optimaux. Revue Economique, 57(1), 55–84. Kolm, S.-C. (2007). Macrojustice: distribution, impôts et transferts optimaux. Revue d’Economie Politique, 117(1), 61–89. Musgrave, R. (1974). Maximin, uncertainty, and the leisure trade-off. Quarterly Journal of Economics, 88, 625–32. Nozick, R. (1974). Anarchy, state and utopia. Oxford: Basil Blackwell. Rawls, J. (1971). A theory of justice. Cambridge: Harvard University Press. Rawls, J. (1974). Reply to Alexander and Musgrave. Quarterly Journal of Economics, 88, 633–55. Rawls, J. (1982). Social unity and primary goods. In Sen, A., & Williams, B. (Eds.), Utilitarianism and beyond, (pp. 159–186). Cambridge, UK: Cambridge University Press. Rawls, J. (1988). The priority of right and ideas of the good. Philosophy and Public Affairs, 17(4), 251–276. Rothbard, M. (1989). The Ethics of Liberty. Atlantic Highlands (New Jersey): Humanities. Steiner, H. (1994). An essay on rights. Oxford: Basil Blackwell.

5

Basic Income and ELIE Transfers

185

Van Parijs, P. (1991). Why surfers should be fed: The liberal case for an unconditional basic income. Philosophy and Public Affairs, 20, 101–131. Van Parijs, P. (1995). Real freedom for all: What (if anything) can justify capitalism? Oxford: Oxford University Press.

•

Part III

Economic Analysis of Macrojustice

•

Chapter 6

An Axiomatic Study of the ELIE Allocation Rule François Maniquet

Abstract This chapter presents an axiomatic analysis of the allocation rules that assigns each economy with its set of ELIE (Equal Labour Income Equalisation) allocations. Two fairness properties, directly inspired by Serge Kolm’s theory of macrojustice are defined. Then, minimal lists of axioms are identified that, when combined to those two fairness properties, characterise the ELIE allocation rules. Finally, other fairness properties are defined that are not satisfied by the ELIE allocation rules.

6.1 Introduction Serge Kolm’s contribution to welfare economics is immense. He is not only among the very first ones having defined and named public economics, but he has contributed to the development of all branches of the field. In the last decade, he has devoted a large fraction of his work to the definition and justification of a particular system of first best income taxation, which he calls Equal Labour Income Equalisation, in short ELIE. Briefly, this system consists of taxing agents according to their labour productivity, and in such a way that the hypothetical income each of them would earn, should she choose a labour time of k, would be equal among all agents. We thus have a family of tax systems, depending on the value of k. As we will see below, the larger k, the more redistributive the system, where redistribution means lump sum transfers from high skill to low skill agents. In his recent contributions, Kolm has developed a long justification of the ELIE tax system, and has deeply analysed the social choice of the value of k. The justification is derived from a combination of heuristic, economic F. Maniquet CORE, Université catholique de Louvain e-mail: [email protected]

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_6, 

189

190

F. Maniquet

and philosophical arguments. In this chapter, I discuss the axiomatic foundations of the ELIE proposal. That does not mean that Professor Kolm has not discussed the axioms that could justify ELIE (see, in particular, his contribution to this book), but he does not seem to have tried to fully axiomatise his solution. The discussion we present here is inspired by a series of papers that Marc Fleurbaey and I have written on the same question. As a matter of fact, we reached the conclusion that the ELIE system was a prominent solution to the first best income taxation problem independently of Kolm, and from a different and purely axiomatic perspective. From a more technical point of view, the ELIE allocation rules are characterised in Maniquet (1998), and are members of a larger family of rules, called Reference Welfare Equivalent Budget rules, characterised in Fleurbaey and Maniquet (1996). I think the axiomatic discussion developed here can be useful in several respects. First, it sheds some light on Kolm’s own justification to ELIE. Indeed, some of the axioms that are part of the results here are very close to the non-axiomatic arguments that Kolm develops, and I try below to insist on the relationship between the axiomatic and the non-axiomatic arguments. Second, it allows us to shed light on the limit of the ELIE proposal, by identifying a list of axioms that ELIE DOES NOT satisfy. Out of the axioms I use in the discussion below, only two are specific to the ELIE rule, and they both are related to the idea that a fraction k of one’s earning ability should be used to the benefit of society. One axiom requires that every agent should be at least as well off as if everybody were requested to work a fraction k of her time and the resulting output were divided equally. The other, weaker, axiom requires that if it is Pareto efficient to request from all agents to work a fraction k of her time and to share the output equally, that is precisely what society should do. The main axiomatic justification of ELIE combines either axiom with efficiency, fairness and cross-economy robustness properties. The chapter is organised as follows. In Sect. 6.2, I introduce the model. In Sect. 6.3, I define the axioms and give the main axiomatic characterisations of the ELIE rules. In Sect. 6.4, I discuss the limits of the ELIE rules. In Sect. 6.5, I give some concluding comments.

6.2 The Model A bundle of goods is composed of a quantity of labour, denoted by `, and a quantity of a consumption good, denoted by c. The consumption set of an agent is defined as X D Œ0; 1  RC . There is an infinite set N of possible

6

An Axiomatic Study of the ELIE Allocation Rule

191

agents. An economy is a list e D .wN ; RN / 2 E where N  N is the finite set of agents, wi 2 RC is the earning ability of agent i 2 N , that is, with a labour contribution of `i ; agent i produces an amount wi `i of consumption good, and Ri are the preferences over X of agent i 2 N , that is, Ri is a complete ordering on X (strict preference and indifference will be noted Pi and Ii ; respectively). Preferences are assumed to be continuous, decreasing in labour contribution, increasing in consumption, and convex. An allocation for an economy e D .wN ; RN / 2 E is a list .`N ; cN / 2 N X where .`i ; ci / represents agent i’s bundle of goods, for all i 2 N . An allocation .`N ; cN / 2 X N is feasible for an economy e D .wN ; RN / 2 E if the total scheduled consumption can be produced given the scheduled labour contributions, that is, if X X ci  wi `i : i2N

i2N

Let F .e/ denote the set of feasible allocations for e D .wN ; RN / 2 E . An allocation rule is a correspondence ' that associates to each economy e D .wN ; RN / 2 E a nonempty set '.e/ of feasible allocations for e. A feasible allocation .`N ; cN / 2 X N for an economy e D .wN ; RN / 2 E is Pareto efficient if no other feasible allocation for that economy guarantees as much satisfaction of their preferences to all agents and a strictly larger satisfaction to at least one agent. Let P .e/ denote the set of Pareto efficient allocations for e D .wN ; RN / 2 E . In this model, we have all convexity properties needed for the second fundamental theorem to hold. Also, the value of the equilibrium price of the consumption good is given by the linear technology and, therefore, is independent of the agents’ preferences. In a competitive world, each agent i 2 N will, therefore, face a wage rate of wi . Consequently, if e D .wN ; RN / 2 E and .`N ; cN / 2 P .e/, then ci  wi `i represents the lump sum transfer which agent i 2 N should receive in the competitive decentralisation of .`N ; cN /. The allocation rules we are interested in are parameterised by a coefficient k 2 Œ0; 1 that must be interpreted as a fraction of the individual labour time. For any given k 2 Œ0; 1, the corresponding allocation rule selects the Pareto efficient allocations having the property that the incomes which agents would have earned, given their individual wage rate and lump sum transfer, by contributing exactly k, are identical among agents. Formally, Allocation rule The k-ELIE rule ' k : Let k 2 Œ0; 1. For all e D .wN ; RN / 2 E : .`N ; cN / 2 ' k .e/ if and only if .`N ; cN / 2 P .e/ and ci  wi .`i  k/ D cj  wj .`j  k/, for all i; j 2 N . Figure 6.1 exemplifies a k-ELIE rule for the case k D 0:5 in a four-agenteconomy with only two possible skills and two possible preferences. Agents

192

F. Maniquet c

R z4

z3

R

w

w

z1 z2

0

k = 0. 5

1



Fig. 6.1 An ELIE allocation in a four-agent-economy

1 and 3 are the low skill agents (their skill is equal to w < w0 ), and agents 1 and 2 have a preference for leisure (their preferences are R, more leisure oriented than R0 ). The intercept of agents 1 and 3’s budget lines measures the transfer to low skill agents in that economy. One can easily realise that this transfer would be zero if k D 0. On the other hand, this transfer would be maximised for k D 1. In this sense, the choice of k amounts to choose the degree of redistribution, where redistribution takes place from higher skill to lower skill agents, and not necessarily from high income to low income agents. When k D 0, the k-ELIE rule coincides with the wealth-fair rule studied by Varian (1974). When k D 1, the k-ELIE rule coincides with the fullincome fair rule defined by Pazner and Schmeidler (1978a). This chapter is dedicated to an axiomatic analysis of the k-ELIE rules. I present the axioms in the next section. The model described in this section is extremely simple, but can be easily generalised in different ways. First, the number of consumption goods need not be restricted to be one. This restriction allows us to speak indifferently of income or consumption. With many different goods, income would be the price value of the production (pre-tax income) or of the

6

An Axiomatic Study of the ELIE Allocation Rule

193

consumption (after-tax income), but all the results presented below would easily generalise. On the other hand, it is crucial to have only one input, labor contribution. More precisely, it is crucial to have a basis of labour contribution comparison, in order to be able to define a common labour contribution that can be used as a reference. The underlying technology is linear in this model. This can be generalised, provided convexity assumptions are still satisfied. With a general production function, equilibrium profits would typically be strictly positive, and property rights on the production sector need to be specified. If these rights are endogenous, then the k-ELIE rules are easily generalised by requiring that they be equally distributed. Fleurbaey and Maniquet (1996) and Maniquet (1998), in fact, axiomatise this generalisation of the k-ELIE rules.

6.3 Axiomatic Characterisations In this section, I list a series of desirable properties that k- ELIE rules satisfy. I also show that these rules are the only ones to satisfy them. It means that this set of properties give an axiomatic justification to the k-ELIE rules. I also point out that the axiomatic justification I propose is well in line with the less formal justifications on which Kolm has insisted in his works. The first axioms that are imposed follow from the Pareto efficiency requirement. They are justified by Kolm on the basis of social freedom and the resulting unanimity principle. They are also common throughout the whole literature, and I do not discuss them here. The first axiom prevents a rule from selecting inefficient allocations. Definition 6.1. An allocation rule ' satisfies Pareto efficiency if and only if for all e D .wN ; RN / 2 E and .`N ; cN / 2 '.e/, we have .`N ; cN / 2 P .e/. The second axiom requires that all allocations Pareto equivalent to a selected allocation (that is, all agents are indifferent between the bundles they are assigned in the two allocations) be also selected. I take it as requiring that the satisfaction level which an agent reaches at her assigned bundle is all what matters, compared to the material composition of that bundle. Consequently, society should be indifferent between two allocations that leave all agents equally well off, even if the bundles change so that, for instance, they are less similar to each other. Definition 6.2. An allocation rule ' satisfies Pareto indifference if and only 0 / 2 F .e/, if for all if for all e D .wN ; RN / 2 E , .`N ; cN / 2 '.e/, .`0N ; cN 0 0 0 0 i 2 N; .`i ; ci / Ii .`i ; ci /, then .`N ; cN / 2 '.e/.

194

F. Maniquet

Before we move on to the next axioms, it may be interesting to add a remark about the relationship between this axiom and welfarism. Welfarism is the ethical theory which considers that justice depends on how individual welfare is distributed in a society. Formally, it means that the ethical observer should choose a specific way of aggregating welfare levels (summing them up leads to utilitarianism, for instance) and that the aggregation rule should not depend on the profile of utility functions that are used to measure individual welfare levels. Kolm and many others have criticised welfarism (see, for instance, Kolm (2005), Chap. 10, Sect. 4), arguing that justice is rather a question of how incomes, resources or freedom, are distributed. In the debates on welfarism, it has proven useful to introduce another, weaker, notion of welfarism, called economy by economy welfarism by Blackorby et al. (1990), which consists of dropping the (crucial) requirement that the aggregation rule does not depend on how individual welfare levels are computed. Pareto indifference turns out to be equivalent to economy by economy welfarism (see Blackorby et al. 1990), but it should be clear that this does not mean that Kolm’s criticisms against welfarism also apply to Pareto indifference. Still, I do not consider this axiom as important as the previous one. As a matter of fact, however, it does not conflict with any of the following axioms, and, therefore, does not force us to drop any of the requirements we would like to impose. As we will see in the results below, it plays a crucial role in singling out the ELIE rules. We now come to the two core axioms of our discussion. They capture the heart of Kolm’s justification of the ELIE rules. They both follow the idea that agents do not necessarily have the right to fully benefit from the a priori value of their productive capacities, that is, their earning abilities (or wage rates). Kolm gives long justifications for that ethical premise. As a consequence, let us assume that society considers that the benefit associated to a fraction k of each agent’s labour time should be a common asset of all the agents. Let us observe that the simple idea of forcing agents to choose a labour contribution of k and sharing the production equally would typically conflict with Pareto efficiency, and, in Kolm’s terms, with process-freedom, as this requires that each agent have the right to use one’s capacities. Here we propose two ways of circumventing this difficulty by still capturing the ethical ideal. The first one consists of giving a welfare lower bound, or safety net, to each individual. Interestingly enough, this safety net does not depend on one’s productive skill. It is defined with respect to the allocation where all agents have the same labour time and output is divided equally. More precisely, the k-lower bound requires that every agent be at least as well off at any selected allocation as at the allocation composed of an equal labour contribution of k and an equal sharing of the production.

6

An Axiomatic Study of the ELIE Allocation Rule

195

Definition 6.3. An allocation rule ' satisfies k-lower bound if and only if for all e D .wN ; RN / 2 E , .`N ; cN / 2 '.e/: for all i 2 N , 0 1 X 1 .`i ; ci / Ri @k; wj k A : jN j j 2N

This axiom was introduced in Maniquet (1998), whereas the second one was not. It is logically weaker than the previous one. It consists in focussing on economies where there is no conflict between imposing a labour time of k and Pareto efficiency, that is, economies where the allocation composed of an equal labour contribution of k and an equal sharing of the production is efficient (and, therefore, is compatible with a free use of one’s capacities). In such economies, the following axiom requires that this allocation, or any Pareto equivalent allocation, be the selection. Definition 6.4. An allocation rule ' satisfies k-efficient selection if and / 2 '.e/, if there exists only if for all e D .wN ; RN / 2 E and .`N ; cNP 0 / 2 P .e/ such that `0 D k and c 0 D 1 .`0N ; cN j 2N wj k for all i 2 N , i i jN j 0 0 then .`i ; ci / Ii .`i ; ci / for all i 2 N . The next two axioms are consistent with the second major ethical value defended by Kolm. That is the idea that choices are private matters, which should not be taken into account in the distribution of society’s assets. Consequently, agents are free to choose their consumption bundle. That is, they should not be given resources as a function of their preferences. Of course, it is clear that society needs to take preferences into account and respect them, what is captured by the Pareto efficiency axiom, but preferences should only matter for efficiency purposes. In the terms of Fleurbaey (1995a), agents should be held responsible for their preferences. The first axiom provides the clearest interpretation of this ethical requirement. It uses the celebrated notion of no-envy, which was introduced in the formal analysis of fairness by Kolm (see Kolm 1998), after Tinbergen had defined it (see Tinbergen 1946). Formally, a no-envy allocation is one where no agent strictly prefers the bundle assigned to any other agent to her own. An equivalent statement is the following one. A no-envy allocation is one that can be obtained by letting all agents choose in identical opportunity sets. As is well known since Pazner and Schmeidler (1974), preventing envy among agents having different productive skills is incompatible with Pareto efficiency. This is why the no-envy requirement is restricted to agents having the same productive skill. The resulting axiom was studied at length in Fleurbaey and Maniquet (1996), and is reminiscent of a similar axiom introduced in Fleurbaey (1995b)’s analysis of a different model.

196

F. Maniquet

Definition 6.5. An allocation rule ' satisfies no-envy among equally skilled if and only if for all e D .wN ; RN / 2 E , all .`N ; cN / 2 '.e/, if wj D wk , then .`j ; cj / Rj .`k ; ck /. The second axiom, called Maskin monotonicity, is, under specific assumptions, logically stronger than no-envy among equally skilled (for analyses of that logical relationship, see Moulin 1993 and Fleurbaey and Maniquet 1997). It requires the following: assume an allocation has been selected for some economy, and assume that the preferences of one agent change in such a way that her assigned bundle in the original allocation moves upwards in her preferences (the upper contour set shrinks at that bundle); then, the allocation should remain in the selection. That definition calls for several remarks. First, contrary to what the statement may lead to think, preferences are not considered as a variable of the economies. The statement “her preferences change” should be understood as meaning that we are comparing the selection for two different economies that could both be economies to which the ELIE proposal should apply. Maskin monotonicity is therefore a property requiring some consistency of the ethical judgement across economies. That kind of properties are key in social choice since Arrow’s contribution (on the logical relationship between Arrow’s independence of irrelevant alternatives property and Maskin monotonicity, see Maskin 1985). Second, Maskin monotonicity is ethically appealing only if one agrees to let agents responsible for their preferences, that is, to prevent society from allocating resources in a way that, except for efficiency reasons, depends on preferences. If an allocation is efficient in some economy and the preferences of some agent change in the way described above, then this allocation is clearly efficient again in the new economy. This change in preferences, therefore, does not justify that this agent receives more, or less, resources (a similar justification is proposed by Gevers 1998). Hence the requirement. Third, Maskin monotonicity is a necessary and sufficient condition for Nash implementation in the model we study here (see Maskin 1999). Nash implementation requires that there exists a mechanism giving the right incentives to the agents to reveal their private information, that is, their preferences, when the Nash equilibrium concept describes agents’ behaviour adequately, which is hardly the case in large societies (because agents are not fully informed about one another, or because they don’t have the opportunity to learn about each other’s preferences). Consequently, we cannot hope to build on implementation theory to design the appropriate institutions leading to ELIE allocations in large societies. On the other hand, the ELIE rules could be applied to small population as well, like workers of self-managed firms, for instance, and the fact that those rules satisfy Maskin monotonicity

6

An Axiomatic Study of the ELIE Allocation Rule

197

and, therefore, are Nash implementable may help one design the appropriate institutions.1 Definition 6.6. An allocation rule ' satisfies Maskin monotonicity if and 0 / 2 E , .` ; c / 2 '.e/, if only if for all e D .wN ; RN /; e 0 D .w0N ; RN N N 0 0 0 0 0 for all i 2 N , wi D wi and f.`i ; ci / 2 X j.`i ; ci / Ri0 .`i ; ci /g  f.`0i ; ci0 / 2 X j.`0i ; ci0 / Ri .`i ; ci /g, then .`N ; cN / 2 '.e 0 /. The last axiom is not related to Kolm’s discussion of the ELIE proposal, though it follows an idea introduced in Kolm (1998). Consistency properties have become common in bargaining theory, cooperative game theory and have been introduced in the fair allocation literature by Thomson (1988) (see the survey in Thomson 1990). Consistency refers to the idea that a solution should be re-negotiation-proof after some agents have left the economy. More precisely, if an allocation has been selected for an economy, and a subset of agents leave the economy with their assigned resources, then the sub-allocation concerning the remaining agents should be in the selection for the sub-economy. Defining sub-economies requires to define what resources are left when some agents are no longer in the economy. This is particularly difficult when production is involved, as the resulting production set need not be in the original domain (for a discussion and a solution to this point, see Moulin and Shenker 1994). Here I propose a weak version of the axiom. Assume that the economy, at a selected allocation, can be decomposed into subgroups of agents such that individual transfers within each group sum up to zero. The consistency axiom below requires that, within each group, the sub-allocation coincides with what the rule would have recommended if that subgroup were an economy. Definition 6.7. An allocation rule ' satisfies consistency if and only if for all P e D .wN ; RN / 2 E , .`N ; cN / 2 '.e/, M  N , if i2N nM .ci wi `i / D 0, then .`M ; cM / 2 '.wM ; RM /. This requirement should be viewed as a cross-economy robustness test for allocation rules. If the rule requires that there be no transfer of resources between two groups, then the global recommendation should be consistent with what the rule would have recommended within each group. I find interesting to note that the k-ELIE rules satisfy this consistency property, but I certainly do not consider it as a major requirement. In particular, this requirement should never be imposed at the price of a fairness axiom. 1

Of course, that requires to use an adequate mechanism, the definition of which is still under debate.

198

F. Maniquet

The first part of our axiomatic analysis consists of listing the major axioms which k-ELIE rules satisfy. That leads to the following theorem. Theorem 6.1. Let k 2 Œ0; 1. The k-ELIE rule satisfies Pareto efficiency, Pareto indifference, the k-efficient selection, the k-lower bound, no-envy among equally skilled, Maskin monotonicity and consistency. (The simple proof of this result is omitted.) The k-ELIE rules are Walrasian rules with specific lump sum taxes that only depends on the productive skills. It is therefore not a surprise that they satisfy Pareto efficiency, Pareto indifference, no-envy among equally skilled and Maskin monotonicity. The interest of the result above lies in the fairness properties of the k-efficient selection or the k-lower bound, which embody the key ethical aspects of these rules. The next results show that the k-ELIE rules are the only ones satisfying the above axioms. In fact, this can be done with two subsets of the axioms. Theorem 6.2. Let k 2 Œ0; 1. The k-ELIE rule is the smallest rule, with respect to inclusion, which satisfies Pareto indifference, the k-efficient selection and Maskin monotonicity. Proof. Let ' satisfy the axioms. Let e D .wN ; RN / 2 E and .`N ; cN / 2 ' k .e/, which implies that ci wi .`i k/ D cj wj .`j k/, for all i; j 2 N . 0 We need to prove that .`N ; cN / 2 '.e/. Let e 0 D .w0N ; RN / 2 E be defined 0 0 0 0 by: for all i 2 N , wi D wi and Ri is such that for all .`i ; ci /; .`00i ; ci00 / 2 X , .`0i ; ci0 / Ri0 .`00i ; ci00 / , ci0  wi `0i  ci00  wi `00i : Let c k 2 RC be defined by ck D

1 X wi k: jN j i2N

By construction,

..k; c k /; : : : ; .k; c k // 2 P .e 0 /:

By k-efficient selection, ..k; c k /; : : : ; .k; c k // 2 '.e 0 /: Note that for all i 2 N, .k; c k / Ii0 .`i ; ci /: By Pareto indifference, .`N ; cN / 2 '.e 0 /. By Maskin monotonicity, .`N ; u t cN / 2 '.e/.

6

An Axiomatic Study of the ELIE Allocation Rule

199

The proof above closely follows Gevers’ characterisation of the Walrasian correspondence (see Gevers 1986; for a similar proof, see also Moulin 1990). The lesson to draw from this result lies in the use of the k-efficient selection axiom. It is sufficient, when combined with Pareto indifference and Maskin monotonicity, to uniquely identify the lump sum tax system which needs to be implemented in each and every economy. Any kind of Walrasian rule based on different lump sum taxes would also satisfy Pareto indifference and Maskin monotonicity. Pareto efficiency is not required for the result to hold: it is implied by the other axioms. The allocation rule selecting allocations ..k; c k /; : : : ; .k; c k // in each and every economy satisfies k-efficient selection and Maskin monotonicity, but fails to satisfy Pareto indifference; it is also inefficient. Kolm has criticised welfarism from different points of view. Most importantly, macrojustice is about the allocation of resources among agents having equal rights, and it is not a question of allocation utilities, welfare or happiness. When combined with Pareto efficiency, that justifies ordinalism and non-comparability, the strongest utility invariance properties compatible with efficiency. It is interesting to study where ordinalism and non-comparability come from in the axiomatic characterisations of the ELIE rules. In the characterisation above, ordinalism and non-comparability directly follow from Maskin monotonicity (see Maskin 1985). We now turn to the second characterisation. Theorem 6.3. Let k 2 Œ0; 1. The k-ELIE rule is the smallest rule, with respect to inclusion, which satisfies Pareto indifference, the k-lower bound, no-envy among equally skilled and consistency. Proof. Let ' satisfy the axioms. Let e D .wN ; RN / 2 E and .`N ; cN / 2 ' k .e/, which implies that ci wi .`i k/ D cj wj .`j k/, for all i; j 2 N . We need to prove that .`N ; cN / 2 '.e/. Let c k 2 RC be defined by ck D

1 X wi k: jN j i2N

Let n 2 N be sufficiently large that for all i 2 N there exists .`0i ; ci0 / 2 X such that .`i ; ci / C .n  1/.`0i ; ci0 / D n.k; c k /: Let us assume, without loss of generality and for the sake of clarity, that 0 / 2 E be defined by: M D f1; : : : ; N D f1; : : : ; jN jg. Let e 0 D .w0M ; RM jM jg; jM j D njN j; and for all i 2 N , m 2 f2; : : : ; ng: w0i D wi , Ri0 D 0 is such that for all .`0i ; ci0 /; Ri , w0iC.m1/jN j D wi and RiC.m1/jN j .`00i ; ci00 / 2 X ,

200

F. Maniquet 0 00 00 0 0 00 00 .`0i ; ci0 / RiC.m1/jN j .`i ; ci / , ci  wi `i  ci  wi `i :

For instance, for e D ..w1 ; w2 /; .R1 ; R2 // and n D 3, we would get e 0 D ..w1 ; w2 ; w1 ; w2 ; w1 ; w2 /; .R1 ; R2 ; R10 ; R20 ; R10 ; R20 //: 0 / 2 '.e 0 /. By feasibility, we know that Let .`0M ; cM X ci0  wi `0i  0:

(6.1)

i2M 0 Assume .`0M ; cM / … ' k .e 0 /. Then, there exists j 2 N such that X cj0 C.m1/jN j  wk `0j C.m1/jN j < nc k  wj k;

(6.2)

m2f1;:::;ng

otherwise, inequality (6.1) would not hold. From inequality (6.2), we will derive a contradiction. By the k-lower bound, for all m 2 f2; : : : ; ng, cj0 C.m1/jN j  wj `0j C.m1/jN j  c k  wj k: Consequently, cj0  wj `0j < c k  wj k D cj  wj `j , with the corollary that .`j ; cj / Pj .`0j ; cj0 /: Let cj00CjN j 2 RC be defined by cj00CjN j  wj `0j D cj0 CjN j  wj `0j CjN j : This construction guarantees that .`0j ; cj00CjN j / Ij CjN j .`0j CjN j ; cj0 CjN j /. 000 0 Consequently, it is possible to define .`000 M ; cM / 2 F .e / that is Pareto 0 0 000 000 0 indifferent to .`M ; cM /, and such that .`j ; cj / D .`j ; cj0 / and .`000 ; j CjN j 000 0 00 000 000 0 cj CjN j / D .`j ; cj CjN j /. By Pareto indifference, .`M ; cM / 2 '.e /. By monotonicity of the preferences, .`0j ; cj00CjN j / Pj .`0j ; cj0 /; violating noenvy among equally skilled, the desired contradiction. That proves that 0 .`0M ; cM / 2 ' k .e 0 /, with the consequence that X ci0  wi `0i D 0: i2N 0 / 2 '.e/. But we also have .`0 ; c 0 / 2 ' k .e/, By consistency, .`0N ; cN N N which implies that .`i ; ci / Ii .`0i ; ci0 / for all i 2 N . By Pareto indifference, t u .`N ; cN / 2 '.e/.

6

An Axiomatic Study of the ELIE Allocation Rule

201

6.4 Discussion The main contribution of the preceding section consisted, first, in the definition of two fairness properties, the k-efficient selection and the k-lower bound axioms, directly inspired by Kolm’s theory of macrojustice, and second, in the identification of the minimal lists of axioms that, when combined with those fairness axioms, characterise the k-ELIE allocation rules. In this section, I would like to list the main appealing fairness properties that are not satisfied by the ELIE rule. Of course we know that it is typically impossible to find solutions that meet all the desirable properties one can think of, but in this case, I would like to argue that the axioms I list below can also find justifications in Kolm’s theory of macrojustice. A clarification remark is needed here. I do not mean that there is some inconsistency in Kolm’s theory. Kolm is clear in his hierarchy of values and in his deducing a precise conclusion (the superiority of the ELIE allocations) from those values. But in the axiomatic theory of resource allocation, theorists typically do not look for a one-to-one relationship between philosophical arguments and axioms. On the contrary, a weak or vague link to philosophical principles is in general sufficient to motivate an axiom, provided its statement is sufficiently clear, simple and meaningful. Consequently, this section is dedicated to showing how Kolm’s Macrojustice theory can be used to motivate the definition of further axioms in the model where ELIE rules can be defined. The first such axiom states that two agents having identical preferences should not envy each other. When two agents have the same preferences, they are not necessarily equals, as their earning abilities may differ. Requiring that they do not envy each other means that they should be given equal opportunities, or equal resources, which implicitly requires that agents with low earning abilities be compensated for their low ability by some larger transfer from high ability agents. That follows from the idea that agents should not be considered as the full owners of the value of their productive skills, a core idea of Kolm’s theory. This axiom is formally defined as follows. Definition 6.8. An allocation rule ' satisfies no-envy among equal preferences agents if and only if for all e D .wN ; RN / 2 E , all .`N ; cN / 2 '.e/, all j; k 2 N , if Rj D Rk , then .`j ; cj / Rj .`k ; ck /. If two agents having identical preferences do not envy each other, they consume bundles they deem equivalent. We could even say that they reach an equal level of satisfaction, although that may misleadingly let one think that we introduce interpersonal welfare comparison, whereas it is not the case: only bundles are compared. That kind of equality should be linked

202

F. Maniquet

to the property of “identical treatment of individuals identical in relevant characteristics” (Kolm 2005, p. 398) when productive assets are viewed as irrelevant in the distribution of the value of these assets. This axiom does not conflict with Pareto efficiency. Moreover, it is compatible with both k-efficient selection and k-lower bound (see below). On the other hand, when Pareto efficiency is imposed, it conflicts with no-envy among equally skilled (as proven in Fleurbaey and Maniquet 1996, following the seminal contribution of Fleurbaey 1995b; for similar proofs, see Bossert 1995, Iturbe and Nieto 1996 and Valletta 2009; for a proof that this incompatibility does not disappear when productive skills are endogenous, see Athanasiou 2007; for an alternative study of the combination of these two principles, see Roemer 1993; see the survey in Fleurbaey and Maniquet 1999a). That is exactly where the ethical choice needs to be made: either one insists on responsibility, that is, non-discrimination on the basis of preferences, and no-envy among equally skilled is given priority, or one insists on compensation, that is, the fact that agents should not bear the consequences of a poor endowment in productive skills, and no-envy among equal preferences agents is given priority. Kolm has chosen the former route. But it is important to observe that his idea of sharing the value of everyone’s earning abilities is compatible with both the compensation objective and the responsibility objective. A much weaker version of the no-envy among equal preferences agents axiom is, actually, satisfied by each k-ELIE rule. Indeed, if we restrict the application of this property to agents having such preferences that they always choose a labour contribution of k, independently of their wage rate, then two such agents would be assigned the same bundle of goods (a bundle of the type .k; c k /) and the axiom is satisfied.2 The idea of restricting the application of the axiom to agents having some specific preferences comes from the fact that it is impossible to eliminate envy among all groups of agents with identical preferences. So it does not mean that we favour certain agents, but simply that we would like to save as much as possible of the initial requirement, and this can only be done by restricting its application to some specific set of agents. Fleurbaey and Maniquet (1996) propose a characterisation of a family of rules that include the k-ELIE rules by imposing

2

Given the continuity assumption, these preferences do not belong to the domain we consider here, as the ranking of bundles composed with a labour time equal to or smaller than k follow from consumption maximisation, whereas ranking of bundles composed of a larger labour time follow from labour time minimisation. The result we refer to, therefore, requires to enrich the domain by adding these preferences.

6

An Axiomatic Study of the ELIE Allocation Rule

203

an axiom of no-envy among reference preferences agents: if two agents have pre-specified reference preferences, they should not envy each other. The second important axiom that is not satisfied by any k-ELIE rule is the following skill solidarity axiom (introduced in Fleurbaey and Maniquet 1999b; that way of formalising solidarity properties comes from Roemer 1986 and Moulin and Thomson 1988). It requires that when comparing two economies composed of the same agents with the same preferences but possibly different productive skills, it must be the case that agents unanimously prefer the selected allocation in one economy to the selected allocation in the other. To phrase it differently, if productive skills change, even if some skills decrease whereas others increase, all agents are affected in the same direction: either they all gain, or they all loose. Note that a change in the earning profile of an economy is the typical consequence of a technological shock. That solidarity requirement is another possible interpretation of the idea that earning abilities partly escape the private sphere. Solidarity properties in the sense of a socialisation of the effects of changes in the external world are closely related to Kolm’s theory, as they are part of the fundamental result that justice should be equality of something (see Kolm 2005). If a solidarity axiom is satisfied, the direction in which welfare levels are affected by a change in the external world is equalised among agents. Definition 6.9. An allocation rule ' satisfies skill solidarity if and only if for 0 0 / 2 E , all .`N ; cN / 2 '.e/; .`0N ; cN /2 all e D .wN ; RN /; e 0 D .w0N ; RN 0 0 0 0 '.e /, if for all i 2 N Ri D Ri , then either .`i ; ci / Ri .`i ; ci / for all i 2 N , or .`0i ; ci0 / Ri .`i ; ci / for all i 2 N . A special case of changes in skill profiles is when skills are permuted between agents. Fleurbaey and Maniquet (1999b) prove that the axiom of no-envy among equal preferences agents is a consequence of some general anonymity axiom and skill solidarity, when changes in the skill profile are restricted to permutation in the skills of equal preferences agents. Their proof is a replication of Kolm’s argument that rationality implies equality (of individuals identical in relevant characteristics; see Kolm 2005, p. 398; for a longer treatment of that point, see Kolm 1996, Chap. 2). The discussion above can be partly summarised by the definition of the following allocation rule, which satisfies the relevant properties. It belongs to the large family of egalitarian equivalent allocation rules, introduced by Pazner and Schmeidler (1978b). An allocation belongs to the selection recommended by this rule if and only if it is Pareto efficient and each agent is indifferent between the bundles she is assigned and a common reference bundle with a labour contribution of k (assigning that reference bundle to all agents simultaneously is typically unfeasible). To phrase it differently,

204

F. Maniquet

this rule selects allocations that are Pareto equivalent to some hypothetical egalitarian allocation where all agents contribute exactly k. Allocation rule The k-egalitarian equivalent rule ' k : Let k 2 Œ0; 1. For all e D .wN ; RN / 2 E : .`N ; cN / 2 ' k .e/ if and only if .`N ; cN / 2 P .e/ and there exists c  2 RC such that .`N ; cN / Ii .k; c  / for all i 2 N . Theorem 6.4. Let k 2 Œ0; 1. The k-egalitarian equivalent rule satisfies Pareto efficiency, Pareto indifference, the k-efficient selection, the k-lower bound, no-envy among equal preferences agents, skill solidarity and consistency. (The simple proof is also omitted.) Kolm has criticised the notion of egalitarian-equivalence (see, for instance Kolm 2005, Chap. 26), but his criticism implies that egalitarian equivalence should not be imposed as a prima facie desirable property. Note that this criticism, therefore, does not apply here, as egalitarian equivalence is not imposed as an axiom, but it is derived from the combination of other, more fundamental, axioms.

6.5 Conclusion In this chapter, I have formally discussed the axiomatic properties that can be used to justify the ELIE allocation rules. These properties include Pareto efficiency, fairness and cross-economies robustness properties. I have also pointed out two sets of properties that are sufficient to axiomatically characterise the ELIE rules. In each case, the key fairness axiom is related to the idea that the benefits of a fraction k of one’s labour time should be shared equally among all the members of the economy. I have also discussed the main drawbacks of the ELIE solutions, that is, the desirable fairness properties that these solutions do not satisfy. It is of course not a surprise that a solution fails to satisfy all the desirable properties one can think of. The axiomatic approach is precisely helpful in identifying the necessary choices social decision makers face. A central question has not been discussed here: the incentive question. How can we decentralise an ELIE rule (or any other solution to the problem at hand)? Kolm has argued that the information necessary to decentralise ELIE were, in general, immediately available. When this information is not available, several (second best) taxation mechanisms have been proposed that are consistent with the properties discussed in this chapter (see Fleurbaey and Maniquet 2011 and Simula and Trannoy 2011). Discussing these taxation mechanisms is beyond the scope of this contribution, but the important point to make is that the key axioms to discuss remain those that have been analysed here.

6

An Axiomatic Study of the ELIE Allocation Rule

205

References Athanasiou, E. (2007). Educational outcomes and equality of opportunity. Technical report, mimeo. Blackorby, C., Donaldson, D., & Weymark, J. (1990). A welfarist proof of Arrow’s theorem. Recherches Economiques de Louvain, 56, 259–286. Bossert, W. (1995). Redistribution mechanisms based on individual characteristics. Mathematical Social Sciences, 29, 1–17. Fleurbaey, M. (1995a). Equality and responsibility. European Economic Review, 39, 683–689. Fleurbaey, M. (1995b). Three solutions to the compensation problem. Journal of Economic Theory, 65, 505–521. Fleurbaey, M., & Maniquet, F. (1996). Fair allocation with unequal production skills : the no-envy approach to compensation. Mathematical Social Sciences, 32, 71–93. Fleurbaey, M., & Maniquet, F. (1997). Implementability and horizontal equity imply no-envy. Econometrica, 65, 1215–1219. Fleurbaey, M., & Maniquet, F. (1999a). Compensation and responsibility. In K. Arrow, A. K. Sen, & K. Suzumura (Eds.), Handbook of social choice and welfare (Vol. 2). Amsterdam: Elsevier. Fleurbaey, M., & Maniquet, F. (1999b). Fair allocation with unequal production skills : the solidarity approach to compensation. Social Choice and Welfare, 16, 569–583. Fleurbaey, M., & Maniquet, F. (2011). ELIE and incentives. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer. (forthcoming). Gevers, L. (1986). Walrasian social choice : some simple axiomatic approaches. In W. Heller, R. Starr, & D. Starett (Eds.), Social choice and public decision making (Vol. 1, pp. 97–114). Cambridge: Cambridge University Press. Gevers, L. (1998). Equality versus what? In J.-F. Laslier, M. Fleurbaey, N. Gravel, & A. Trannoy (Eds.), Freedom in economics (pp. 235–240). London, New York: Routledge. Iturbe, I., & Nieto, J. (1996). On fair allocations and monetary compensations. Economic Theory, 7, 125–138. Kolm, S. C. (1971, 1995, 1998). Justice et équité, (Justice and equity). Paris, (Cambridge): CEPREMAP, (MIT). English translation by H. See. Kolm, S. C. (1996). Modern theories of justice. Cambridge, London: MIT Press. Kolm, S. C. (2005). Macrojustice, The political economy of fairness. Cambridge: Cambridge University Press.

206

F. Maniquet

Maniquet, F. (1998). An equal-right solution to the compensationresponsibility dilemma. Mathematical Social Sciences, 35, 185–202. Maskin, E. (1985). The theory of implementation in Nash equilibrium: a survey. In L. Hurwicz, D. Schmeidler, & H. Sonnenschein (Eds.), Social goals and social organization: Essays in memory of Elisha Pazner (pp. 173–204). Cambridge: Cambridge University Press. Maskin, E. (1999). Nash equilibrium and welfare optimality. Review of Economic Studies, 66, 23–38. Moulin, H. (1990). Joint ownership of a convex technology : Comparison of three solutions. Review of Economic Studies, 439–452. Moulin, H. (1993). On the fair and coalition-strategy-proof allocation of private goods. In K. Binmore et al. (Eds.), Frontiers in game theory (pp. 151–163). Cambridge: MIT Press. Moulin, H., & Shenker, S. (1994). Average cost pricing vs serial cost sharing: An axiomatic comparison. Journal of Economic Theory, 64, 178–201. Moulin, H., & Thomson, W. (1988). Can everyone benefit from growth? Two difficulties. Journal of Mathematical Economics, 17, 339–345. Pazner, E., & Schmeidler, D. (1974). A difficulty in the concept of fairness. Review of Economic Studies, 41, 441–443. Pazner, E., & Schmeidler, D. (1978a). Decentralization and income distribution in socialist economies. Economic Inquiry, 257–264. Pazner, E., & Schmeidler, D. (1978b). Egalitarian equivalent allocations: A new concept of economic equity. Quarterly Journal of Economics, 92, 671–687. Roemer, J. E. (1986). Equality of resources implies equality of welfare. Quarterly Journal of Economics, 101, 751–784. Roemer, J. E. (1993). A pragmatic theory of responsibility for an egalitarian planner. Philosophy and Public Affairs, 22, 146–166. Simula, L., & Trannoy, A. (2011). When Kolm meets Mirrlees: ELIE. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer. (forthcoming). Thomson, W. (1988). A study of choice correspondences in economies with a variable number of agents. Journal of Economic Theory, 46, 237–254. Thomson, W. (1990). The consistency principle. In T. Ichiishi, A. Neyman, & Y. Tauman (Eds.), Game theory and applications (pp. 187–215) San Diego. Proceedings of the 1987 International Conference, Academic. Tinbergen, J. (1946). Redelijke inkomensverdeling. Haarlem: De Gulden Pers. Valletta, G. (2009). A fair solution to the compensation problem. Social Choice and Welfare, 32, 455–478. Varian, H. (1974). Equity, envy, and efficiency. Journal of Economic Theory, 9, 63–91.

Chapter 7

An Exploration of Incentive-Compatible ELIE Laurent Simula and Alain Trannoy

Abstract Simula and Trannoy (2011) have shown that ELIE is confronted with implementation issues when the policymaker cannot observe the time worked by every individual. This paper tries to fix this problem. To this aim, we characterise the second-best allocations which are the closest to ELIE first in terms of welfare and then in terms of transfers. In the former perspective, we consider a welfarist setting in which the social weights are those required by ELIE to be generated as a first-best allocation. These weights are defined by the tangent hyperplane to the first-best Pareto set at the ELIE allocation. We show that, in the absence of income effect on labour supply, the closest solution to ELIE is the laissez-faire. In addition, simulations for a Cobb–Douglas economy show that the second-best transfers may then be substantially different from ELIE. This is why, in the latter perspective, we construct second-best allocations which are both incentive-compatible and generate net transfers coinciding with the first-best ELIE transfers. We show that the unique solution is Pareto-efficient in the constraint set.

7.1 Introduction In Mirrlees (1971) seminal article, the policy-maker aims at redistributing income from the high to the low productive individuals. However, it is confronted with a basic separation between public and private information: “the L. Simula (B) Uppsala University and Uppsala Center for Fiscal Studies e-mail: [email protected] A. Trannoy EHESS, Marseille and IDEP-GREQAM e-mail: [email protected] Very helpful comments from François Maniquet are gratefully acknowledged. The usual caveat applies. C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_7, 

207

208

L. Simula and A. Trannoy

government can observe the total product of each individual, that is the product of the wage rate and the amount worked, but is unable to observe either of these alone” (Mirrlees 1997). As a result, the tax scheme must be based on gross income instead of skill levels. However, as emphasised by Mirrlees himself in the conclusion of his paper, “it would be good to devise taxes complementary to the income-tax, designed to avoid the difficulties that tax is faced with”. The tax scheme recently proposed by Serge-Christophe Kolm belongs to this family (Kolm 2005). It is derived from fundamental principles of justice and corresponds in essence to a linear tax based on productivity. It has a remarkable feature when every individual to whom it applies provides a quantity of labour in excess to what is needed for him to pay his net tax. In this case, everyone gives the fruit of the same labour duration to the society from which he receives the average donation. This structure is referred to as “Equal-Labour Income Equalisation” or ELIE by Kolm and constitutes a simple way of equally apportioning the common contribution between the citizens, according to their means: “from each, to each other, the product of the same labour” and “from each, to each other, according to her capacities”. It is argued by Kolm (2007, unpublished manuscript) that ELIE is incentive-compatible in the sense that “it induces people to work with their full abilities”. In this respect, ELIE would resolve the fundamental trade-off between equity and efficiency and justify leaving the second-best analysis for the first-best one. In fact, the incentive-compatibility of ELIE depends on which variables are observable and can thus be included in the contract between the taxpayers and the policymaker. If it seems natural to consider that gross income is observable, the verifiability of time worked appears to be more problematic. In practice, when both gross income and time duration can be observed in a second-best world, ELIE actually provides the right incentives for every agent whose labour can be time-clocked or measured by similar mechanisms. However, this is not always the case for all individuals in brainwork occupations for which time and attendance solutions for employee labour tracking are irrelevant. In Simula and Trannoy (2011), it is established – in a continuous population framework – that these individuals have an incentive to misreport their productivity through the gross-income/labour combination they choose as soon as time worked is not observable and verifiable. This is an unfortunate state of affairs because the latter individuals are likely to be the most productive in the population, i.e., those who are supposed to pay the higher taxes under ELIE. Consequently, incentive-compatibility of ELIE is not a trivial matter. This paper considers a discrete population version of the framework developed by Mirrlees (1971), initiated by Stiglitz (1982) and Guesnerie and Seade (1982), and focuses on the cases in which the first-best ELIE

7

An Exploration of Incentive-Compatible ELIE

209

allocation is not envy-free. Consequently, ELIE is not implementable when productivity is not observable by the policymaker. ELIE must thus be replaced by an income tax scheme. Our primary objective is to construct income tax schemes which are “as close as possible” to ELIE. Proximity is defined is two ways. First, we keep the social weights which generate ELIE in the first-best setting to define the second-best social objective. In this sense, this approach yields the closest solution in terms of welfare. We then maximise this social objective function to find the best income tax scheme in the possibility set constrained both by the incentive-compatibility constraints and the tax revenue constraint. We show that the income tax can be substantially different from ELIE. In particular, in the absence of income effect on labour supply, i.e., for quasilinear-in-consumption preferences, the closest solution to ELIE is the laissez-faire. Second, because Kolm analysis puts the stress on the specific shape of the ELIE transfers, we endeavour to stay as close as possible to ELIE transfers. To this aim, we depart from the welfarist framework to adopt a non-welfarist viewpoint as favoured by Kolm in Macrojustice (Kolm 2005). We look for the income tax scheme for which everyone pays the same tax or receives the same transfer as in the firstbest while the second-best incentive-compatibility constraints are satisfied. We show that there is a unique solution which is Pareto-efficient and corresponds to a simple monotonic chain to the left, i.e., to an allocation of gross-income/consumption bundles in which every individual is indifferent between his own bundle and that of his nearest less productive neighbour. The general methodology to get the closest solution in terms of transfers is then illustrated through an example. A related analysis is made by Fleurbaey and Maniquet (2011) who also address the incentive-compatibility of ELIE in both above-mentioned informational settings. A major difference comes from the diversity of preferences over leisure they introduce. Kolm regards this diversity as a private matter and not as a legitimate ground for redistribution, contrary to productivity differences. Therefore, from an ethical viewpoint, it seems justified – as a first pass – to focus on the latter and to consider the restrictive framework where all individuals have the same preferences. To cope with this additional heterogeneity, Fleurbaey and Maniquet employ a different social welfare function from utilitarianism, which is less demanding in terms of interpersonal utility comparisons. An axiomatic foundation of ELIE is provided in this context. The paper is organised as follows. Section 7.2 addresses the incentivecompatibility of ELIE. Section 7.3 examines the closest solution in terms of welfare. Section 7.4 is devoted to the closest solution in terms of transfers. Section 7.5 provides concluding comments.

210

L. Simula and A. Trannoy

7.2 ELIE and Incentive-Compatibility 7.2.1 Setting We consider a competitive economy in which the technology exhibits constant returns to scale. The population consists of I  2 individuals, indexed by i 2 I WD f1; : : : ; I g : For convenience, only one person has a given productivity level. This simplification is not particularly restrictive because the distance between two productivity levels is free to vary. Without loss of generality, the vector of productivities w WD .w1 ; : : : ; wI / is taken to be monotonically increasing, belonging to the set o n (7.1) ˝ WD w 2 RICC jw1 <    < wI : Person i’s wage rate is fixed, equal to his productivity wi : The time endowment of each individual has been normalised to equal 1: Consumption good is chosen as the numéraire. When person i works `i units of time, his gross income amounts to i 2 I: (7.2) zi WD wi `i ; All individuals have the same preferences over consumption and leisure. These preferences are represented by a cardinal utility function U W RC  Œ0; 1 ! R: The function U is assumed to be strictly concave, twicecontinuously differentiable, strictly increasing in consumption and strictly decreasing in labour. Moreover, leisure is assumed to be a normal good. Note that person i has a personalised utility function in the gross-income/consumption space, defined by u W RC  Œ0; wI  ! R with u .x; zI wi / WD U .x; z=wi / :

(7.3)

Person  i’s marginal rate of substitution of gross income for consumption at the xj ; zj -bundle is defined as     u0z xj ; zj I wi : (7.4) s xj ; zj I wi WD  0  ux xj ; zj I wi The Spence–Mirrlees condition is assumed to be satisfied. It ensures that, for any given gross-income/consumption bundle, more productive individuals have flatter indifference curves. Formally: Assumption 7.1 Given any .xk ; zk / 2 RC  Œ0; wI  ;   s .xk ; zk I wi / > s xk ; zk I wj , i < j:

(7.5)

7

An Exploration of Incentive-Compatible ELIE

211

7.2.2 ELIE The tax policy can now be introduced formally. With a Kolm formula tax scheme of degree k; person i is required to transfer kwi to society in exchange for which he receives kE .w/ ; where E .w/ is the average productivity level in the population. Hence, the tax function is Ti .k/ D k .wi  E .w// :

(7.6)

The ELIE tax scheme of degree k combines (a) the Kolm formula tax scheme of degree k with (b) a condition on the endogenous individual labour supply. In the absence of involuntary unemployment, this condition states that all productive individuals must provide labour in excess of k to take part in the overall redistributive mechanism (Kolm 2007, pp. 26–27, unpublished manuscript). Because it can be argued that a redistributive tax schedule should be universal with the same schedule applied to everyone, special attention is paid below to the case in which all citizens face the ELIE tax scheme of degree k:1

7.2.3 ELIE, Envy and Implementability In the first-best setting, the policymaker observes the productivity of every individual. He implements the tax scheme Ti .k/ based on individual skill levels. We say that person i envies person j if and only if, given his own productivity, the former would prefer the labour/consumption bundle of the latter, i.e., if and only if     zj zi U xj ; > U xi ; : (7.7) wi wi Figure 7.1 shows the budget lines and the choices of two individuals, with w2 > w1 : person 2 may envy person 1 at the ELIE allocation because of the progressivity of the tax schedule Ti .k/ ; but the converse is impossible. More generally, when there are more than two individuals, person i may envy person j at the ELIE allocation only if wi > wj ; but the contrary is impossible. Figure 7.2 shows a situation where person 2 does not envy person 1 at the

1

These desiderata are incorporated in the actual tax schedules in many developed countries. In France, for instance, the 13th article of the Declaration of the Rights of Man and of the Citizen, which has a constitutional status, states that the common contribution should be equitably distributed among all the citizens in proportion to their means.

212

L. Simula and A. Trannoy xi

xi

1

1

2

2

li

0

zi

0

Fig. 7.1 ELIE and envy xi 2

1

0

zi

Fig. 7.2 ELIE without envy

ELIE allocation. In the absence of envy at the ELIE allocation, ELIE remains implementable even when individual productivity is private knowledge. If there is at least an individual who envies another one in the population at the ELIE allocation, ELIE is no longer implementable when productivity is private knowledge. The problem is then to find a second-best solution which is “as close as possible” to ELIE. But what is the “closest” solution? At least two different approaches can be considered: the closest in terms of welfare and the closest in terms of transfers. The next two sections examine each of them respectively. Both of them introduce incentive-compatibility constraints in the programme of the policymaker. It is thus worthwhile to examine the implications of these constraints before going further. By the taxation principle, an income tax schedule corresponds to a mapping  w ! RC  R (7.8) wi .xi ; zi / ; which satisfies the incentive-compatibility constraints   (7.9) 8 .i; j / 2 I 2 ; u .xi ; zi I wi /  u xj ; zj I wi ;

7

An Exploration of Incentive-Compatible ELIE

213

and the tax revenue constraint I X iD1

zi 

I X

xi :

(7.10)

iD1

The incentive-compatibility constraints ensure that everyone prefers his own consumption/leisure bundle to that of anyone else. They place structure on feasible income tax schedules. First, the gross income and net income vectors of an incentive-compatible allocation must be non-decreasing in productivity, i.e., such that .x1 ; z1 /      .xI ; zI / ;

(7.11)

with .xi1 ; zi1 / << .xi ; zi / if .xi1 ; zi1 / ¤ .xi ; zi /, i D 2; : : : ; I: This condition corresponds to the second-order condition for incentive compatibility in the continuous population framework. Second, given an incentivefeasible consumption vector satisfying (7.11), it proves sufficient to check the downward and upward adjacent incentive compatibility constraints to get an incentive-compatible allocation, providing the Spence–Mirrlees condition holds. This can be formally stated as follows. Lemma 7.1. Given x1 ; : : : ; xI and z1 ; : : : ; zI satisfying (7.11), u .xi ; zi I wi /  u .xi1 ; zi1 I wi / ; u .xi ; zi I wi /  u .xiC1 ; ziC1 I wi / ;

i D 2; : : : ; I i D 1; : : : ; I  1;

(7.12) (7.13)

imply (7.9). The proof is standard and thus omitted (see Cooper 1984). Many patterns of links between gross-income/consumption bundles satisfy the conditions in Lemma 7.1. This is notably the case if all adjacent downward incentive-compatibility constraints are binding, i.e., if .xi ; zi / and .xiC1 , ziC1 / are both on person i C 1’s highest (feasible) indifference curve: u .xi ; zi I wiC1 / D u .xiC1 ; ziC1 I wiC1 / ;

i D 1; : : : ; I  1:

(7.14)

Allocations satisfying (7.14) are called simple monotonic chains to the left. The following result is then easy to establish. Proposition 7.1. Let an allocation be a simple monotonic chain to the left satisfying (7.11). Then it satisfies all the incentive compatibility constraints (7.9).

214

L. Simula and A. Trannoy

7.3 The Closest Solution in Terms of Welfare 7.3.1 Methodology When at least one individual envies another one in the first-best, ELIE is not implementable in the second-best. Designing an incentive-compatible solution then implies a loss in welfare. A first solution is to consider the social weights which generate ELIE in the first-best and then look for the allocation that maximises the corresponding social welfare function subject to the tax revenue constraint (7.10) and the incentive-compatibility constraint (7.9). The optimum allocation is the closest one to ELIE in terms of social welfare in the sense that, given the social weights which generate ELIE in the first-best, it minimises the deadweight loss resulting from asymmetric information. This welfarist approach is certainly clearly distinct from the Kolmian analysis, but it allows us to use the tools and results of the Mirrleesian optimal income tax literature. Obviously, there is no unicity of the social welfare function which generates ELIE in the first-best information framework. Which function should we pick up? We know that the allocation obtained for the ELIE tax scheme belongs to the frontier of the utility possibility set because it is Paretoefficient. There is an hyperplane tangent to the Pareto set at the ELIE allocation. The individual social weights  WD .1 ; : : : ; I / 2 RICC are defined by this hyperplane and a social welfare function W which is additively separable with respect to these social weights is adopted. Formally, W W .RC  Œ0; wI /I ! R; with W ..x1 ; z1 / ; : : : ; .xI ; zI // D

I X

i u .xi ; zi I wi / :

(7.15)

iD1

Because W ..x1 ; z1 / ; : : : ; .xI ; zI // is homogeneous of degree 1 in ; the social weights can be normalised without any loss in generality. We choose to set E ./ D 1: Figure 7.3 shows the utility possibility sets in the first-best  F and the second-best when I D 2: The first-best solution is uF D uF 1 ; u2 : The tangent to the first-best Pareto frontier at uF has slope 1 =2 : We consider the family of linear social indifference curves with slope   1 =2 and look for the second-best solution. It is obtained at uS D uS1 ; uS2 where the second-best Pareto frontier is tangent to the highest (feasible) social indifference curve. Though this approach seems rather natural, it is not free from drawbacks: in particular and as in most of the optimal income tax models, it implies that individual preferences have a cardinal meaning. Yet, it

7

An Exploration of Incentive-Compatible ELIE

215

u2 uF uS

u1

0

Fig. 7.3 Utility possibility set

is compatible with any finite level of inequality aversion.2 The problem of the policymaker can thus be stated as follows: Problem 7.1 (Welfare Approach). Choose an allocation a 2 RIC Œ0; wI I which maximises the social welfare function W .a/ ; where the social weights  generate ELIE in the first-best and E ./ D 1, subject to the incentive compatibility constraints (7.9) and the tax revenue constraint (7.10). Given the utility maximisation programme max

.xi ;`i /2R2C

U .xi ; `i / s.t. xi  wi `i  k .wi  E .w// ;

(7.16)

person i’s indirect utility is

  F  k .w  E .w// ; ` : Vi .k/ WD U w`F i i i

(7.17)

Let ˇi denote the Lagrange multiplier of person i’s budget constraint. The first-order condition with respect to consumption is Ux0 .xi ; `i / D ˇi :

(7.18)

By the envelope theorem, @Vi .k/ : (7.19) @ Œk .wi  E .w// In the first-best, the social policymaker maximises W ..x1 ; z1 / ; : : : ; .xI ; zI // subject to the tax revenue constraint (7.10). If  stands for the Lagrange multiplier of this constraint (which must be binding at the social optimum), the first-order condition with respect to consumption is ˇi D 

2

Note that it is not compatible with the maximin since all social weights are strictly positive.

216

L. Simula and A. Trannoy

Ux0 .xi ; `i / D

 : i

Combining (7.18)–(7.20), one obtains   @V .w; k/ 1 i D    @T .w; k/  ; D F 0 Ux .wi .`i  k/ C kE.w/; `F i /

(7.20)

(7.21) i 2 I:

(7.22)

In Simula and Trannoy (2011), we have established that, when the marginal utility of consumption is strictly decreasing, these social weights  are increasing in productivity provided there is ALEP substitutability between consumption and leisure and the substitution effect on labour supply is larger than the income effect.

7.3.2 Features of the Solution It is possible to derive qualitative features of the solution to Problem 7.1 for different specifications of individual preferences. We first consider separable preferences, linear with respect to consumption, i.e., U .xi ; `i / D xi  v .`i / with v0 > 0 and v00 > 0:

(7.23)

By (7.22), the social weights which generate ELIE in the first-best are i D  for every i 2 I . Because preferences are quasilinear in consumption and the social weights i are all equal, the solution to Problem 7.1 is the laissez-faire. Proposition 7.2 casts light on a striking feature of ELIE. Proposition 7.2. Consider that individual preferences are quasi-linear in consumption. When ELIE is not envy-free, the closest incentive-compatible allocation in terms of welfare is the laissez-faire. We now turn to Cobb–Douglas preferences. Guesnerie and Seade (1982) have examined the solution to the optimal income tax model in a discrete population framework. They consider well-behaved individual preferences, which are assumed to be concave, increasing in consumption and decreasing in leisure. The social objective function is a weighted sum of individual utilities, which is supposed to be defined on and increasing in individual utilities. In this context, designing a nonlinear tax scheme is equivalent for the policymaker to setting a step function in the gross income/consumption space.

7

An Exploration of Incentive-Compatible ELIE

217

A “corner” is thus obtained at every observed gross income/consumption bundle. A nonlinear tax schedule corresponds to a set of corners, each of which being chosen by the individuals who maximise their utility at this bundle. Without loss of generality, these corners can be unambiguously arranged in a South–West/North–East direction. When an additional assumption regarding the shape of the welfare improving transfers is made, the optimal income tax which maximises the objective function subject to the incentive-compatibility constraints and the tax revenue constraint is a simple monotonic chain to the left, provided the Spence–Mirrlees condition is satisfied. In our framework, this assumption may be formulated as follows (see Röell 1985).   i < j Assumption 7.2 (VWR) For each pair of corners Ci ; Cj with  i j (and thus Ci << Cj /; given K > 0 small enough, there exists ı ; ı 2   it is desirable to distribute K  ı i ; ı i to agent i and Œ0; K2 such that   K  ı j ; ı j away from agent j; provided incentive effects are ignored. Proposition 7.3. Under Assumption VWR, any set of corners .C1 ; : : : ; CI / satisfying (7.11) is welfare-dominated by a simple monotonic chain to the left. t u

Proof. See Röell (1985).

K might be thought of as the quantity of money indirectly transferred from a more to a less productive agent. Under Assumption VWR and the Spence–Mirrlees condition, the optimal income tax schedule must be a simple monotonic chain to the left. We now examine whether the specific shape of the social weights which generate ELIE in the first-best implies Assumption VWR. For Cobb–Douglas preferences U .xi ; `i / D ˛ ln xi C .1  ˛/ ln .1  `i / ; in which case `F i

 D max 0; ˛ C .1  ˛/

k .wi  E .w// wi

(7.24)

:

(7.25)

By (7.22), when `F i > 0 for every individual, i D  Œwi .1  k/ C kE .w/ :

(7.26)

Given our normalisation, E ./ D  E .w/ D 1 ,  D

1 ; E .w/

(7.27)

218

L. Simula and A. Trannoy xi 1

1

C2

2

2

C1

zi

Fig. 7.4 Transfer along the direction given by the gross-income axis

from which

wi C k: (7.28) E .w/ Let I D 2 and consider a transfer .0; ı/ with ı > 0 from person 2 to person 1 (see Fig. 7.4). This transfer – along the direction given by the gross-income axis – satisfies VWR. Let  



z1  ı w1 C k ˛ ln x1 C .1  ˛/ ln 1  W .ı/ D .1  k/ E .w/ w1  



z2 C ı w2 C k ˛ ln x2 C .1  ˛/ ln 1  : C .1  k/ E .w/ w2 i D .1  k/

(7.29) Hence,



1˛ w1 Ck W .ı/ D .1  k/ E .w/ w1  z1 C ı 

1˛ w2 Ck ;  .1  k/ E .w/ w2  z2  ı 0

(7.30)

from which W 0 .0/ D ( .1  ˛/

w1 Ck .1  k/ E.w/

w1

w2 .1  k/ E.w/ Ck 1 1  1  `1 w2 1  `2

) : (7.31)

If W 0 .0/ > 0, then an increase in ı increases social welfare. This is the case when 1 .1  k/ E.w/ C wk2 1  `2 > W 0 .0/ > 0 , 1 1  `1 .1  k/ E.w/ C wk1 1 .1  k/ E.w/ C L2 , > 1 L1 .1  k/ E.w/ C

k w2 k w1

;

(7.32)

7

An Exploration of Incentive-Compatible ELIE

219

where Li denotes person i’s leisure. Equation (7.32) provides us with a sufficient condition for VWR to be satisfied. Using Proposition 7.3, we can therefore proceed as follows to find a solution to Problem 7.1: (a) Assume the solution to Problem 7.1 is a simple monotonic chain to the left. (b) Compute the candidate allocation and check that (7.32) is satisfied. If so, the candidate simple monotonic chain to the left solves Problem 7.1.

7.3.3 An Example For illustrative purpose, we consider that the population consists of two individuals, who both have the same preferences, represented by the utility function (7.24). The policymaker chooses x1 ; z1 ; x2 ; z2 to maximise

  z1 W .x1 ; x2 ; z1 ; z2 / D 1 .k/ ˛ log x1 C .1  ˛/ log 1  w1

  z2 ; (7.33) C 2 .k/ ˛ log x2 C .1  ˛/ log 1  w2 where 1 .k/ and 2 .k/ are given by (7.26), subject to the incentive compatibility constraints     z1 z2 ˛ log x1 C .1  ˛/ log 1   ˛ log x2 C .1  ˛/ log 1  ; w1 w1 (7.34)     z2 z1  ˛ log x1 C .1  ˛/ log 1  ; ˛ log x2 C .1  ˛/ log 1  w2 w2 (7.35) and the tax revenue constraint z1  x1 C z2  x2 D 0:

(7.36)

There are three possibilities as regards the incentive compatibility constraints: (a) 1 D 2 D 0; (b) 1 D 0 and 2 > 0; (c) 1 > 0 and 2 D 0: We apply the strategy described above (for the computational procedure, see the Appendix). For example, assume w1 D 1, w2 D 1:5 and ˛ D 1=2: Figure 7.5 shows person 1’s and person 2’s first-best indirect utilities (denoted U1 and U2 respectively) as well as person 2’s utility if he were able to choose person 1’s leisure/consumption bundle (denoted U21 ). For k  0:166; person 2 does not envy person 1 and ELIE remains implementable in the second-best. In Simula and Trannoy (2011), we have shown in a continuous population framework that, when gross income is verifiable

220

L. Simula and A. Trannoy

-0.5

U21 U2

Utility

-0.55

-0.6

U1

-0.65

-0.7

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

k

Fig. 7.5 First-best utility levels

and time worked non-verifiable, every individual facing the ELIE tax scheme of degree k has an incentive to understate his true productivity level. Therefore, this tax scheme is not implementable as a direct truthful mechanism in weakly dominant strategies for every value of k ¤ 0. This shows a basic difference between the continuous and discrete population versions of the optimal income tax model. In the discrete framework and for k  0:166 person 2 would have an incentive to mimic the behaviour of an individual with productivity between w1 and w2 : However, this productivity level does not exist in the population. Hence, this is because he would loose too much in terms of income that person 2 chooses not to mimic person 1: For k  0:444; person 1 chooses to work less than k units of time in the first-best. For 0:166 < k < 0:444; the second-best optimum allocation is a simple monotonic chain to the left (we can check that (7.32) is verified), i.e., person 2 is indifferent between his consumption/gross-income bundle and that of person 1 while the constraint preventing person 1 from mimicking person 2 is inactive. Figure 7.6 compares the transfers obtained in the firstbest (which remain implementable in the second-best for k  0:166) with the second-best transfers solution to Problem 7.1. For every 0:166  k  0:444; the second-best solution is such that person 2 transfers money to person 1: Hence, person 1 is a net recipient of and person 2 a net contributor to the tax policy, as under ELIE. However, the magnitude of the transfers is significantly altered. In the second-best, the transfer to the low-skilled agent is increased compared to the first-best. This is because person 1 now faces a distortive marginal tax rate in order to prevent person 2 from mimicking him (Fig. 7.7).

7

An Exploration of Incentive-Compatible ELIE

221

0.18 0.16

Net Transferts

0.14 0.12 0.1

Second-Best

0.08 0.06 0.04

First-Best

0.02 0

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

k

Fig. 7.6 Net transfers to person one -0.6 -0.65 -0.7

First-Best

Utility

-0.75 -0.8 -0.85 -0.9

Second-Best

-0.95 -1

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

00.4

k

Fig. 7.7 Person one’s utility level

7.4 The Closest Solution in Terms of Transfers As already noted, the welfare approach adopts the viewpoint of the welfarist optimal income tax literature to construct incentive-compatible transfers close to ELIE in the sense specified above. This welfarist perspective seems to significantly differ from Kolm’s analytical framework. Moreover, the numerical simulations have shown that the gap between ELIE and the second-best transfers is then rather substantial. In this section, we construct

222

L. Simula and A. Trannoy

second-best allocations which are as close as possible to ELIE in terms of transfers. Because the focus is on the transfers themselves, we do not require individuals to work at least k units of time. Hence, we concentrate on Kolm’s tax schemes of degree k (the ELIE tax schemes of degree k being a particular case of Kolm’s tax schemes).

7.4.1 Methodology The basic idea is to construct a second-best allocation for which all individuals receive the same transfers and pay the same taxes as under ELIE. If this construction is possible, we say that the second-best allocation replicates the first-best ELIE transfers. Because we consider cases where at least one individual envies another one in the first-best, the second-best allocation we construct will not coincide with the first-best one; only the taxes and transfers will be the same. Let us first consider that the population consists of two individuals, i.e., I D 2: We are interested in cases where at least one person envies another one since otherwise the first-best allocation is still incentive-compatible in the second-best. According to Fig. 7.1, we thus assume that person 2 envies person 1: To construct a second-best allocation which replicates the ELIE transfers, person 1’s and person 2’s bundles must stay along their respective first-best budget lines. Given this requirement, how shall we modify the first-best bundles so as to make them incentive-compatible?  there  A priori, are three possibilities: (a) choosing a bundle .x2 ; z2 / >> x2F ; zF 2 for per F F son 2; where xi ; zi is person i’s first-best bundle, (b) choosing a bundle   of (a) and .x1 ; z1 / << x1F ; zF 1 for person 1 or (c) choosing a combination   (b). Let us consider person 2’s indifference curve through x2F ; zF 2 : Because   F F s x2 ; z2 ; w2 D 1; it is impossible to find higher indifference curves which intersect person 2’s first-best budget line. The only possibility is to find lower indifference curves. But that does not induce incentive-compatibility. In fact, the only possibility is thus to decrease person 1’s utility. So, person 2 gets his  first-best utility level x2F ; zF . The minimum decrease in person 1’s utility  2  is at the junction of person 2’s indifference curve is obtained when x1F ; zF 1  F F through x2 ; z2 and person 1’s budget line. Hence, the solution must be (b). The implication is that person 2 receives his first-best bundle while person 1’s choices are distorted (to the minimum). This is illustrated on the lefthand side of Fig. 7.8. This construction is not always possible as shown by the right-hand side. Indeed, there might be not intersection between person 2’s indifference curve through his first-best bundle and person 1’s first-best

7

An Exploration of Incentive-Compatible ELIE D F1

x F2 1

x F1

223

D F2 2

D F2

D F1 1

x F1 x F2

2

x1

0

z1

z F1

z F2

0

z F1

z F2

Fig. 7.8 Closest solution in transfers: methodology

budget line. In that case, ELIE transfers cannot be replicated using a secondbest income tax scheme. Now, assume I D 3: Person 3 receives his first-best bundle. The minimum incentive-compatible decrease in person 2’s utility is obtained when he receives the bundle at the junction of his budget line and person 3’s highest (feasible) indifference curve. Now, let us consider person 2 and person 1 and keep person 2’s bundle fixed. Using the same argument as in the twoperson case, it is clear that incentive-compatibility now requires a decrease in person 1’s utility. This reduction is minimised when person 1’s bundle is at the junction of person 2’s indifference curve through his second-best bundle and person 1’s first-best budget line. By induction, the same procedure applies recursively when I > 3; the loss in utility being minimised at each step. Recall an allocation for which person i is indifferent between his own bundle and that designed for the closest less productive person, i.e., for which u .xi ; zi I wi / D u .xi1; zi1 I wi / ;

i D 2; : : : ; I;

(7.37)

is called a simple monotonic chain to the left after Guesnerie and Seade (1982). The problem addressed in this section can thus be summarised as follows. Problem 7.2. Construct a simple monotonic chain to the left for which zi  xi D k .wi  E .w// ;

i D 1; : : : ; I:

(7.38)

In fact, because the Spence–Mirrlees condition is met, the satisfaction of all local incentive-compatibility constraints induces the satisfaction of all global incentive-compatibility constraints (Cooper 1984). Hence, pairwise comparisons of adjacent bundles are sufficient to get incentive-compatibility.

224

L. Simula and A. Trannoy

In the above construction, the second-best bundles must also be on the first-best budget lines. By construction, every individual i < I incurs the minimum loss in utility so as to replicate the ELIE transfers by the means of an income tax. Consequently: Proposition 7.4. If there exists a solution to Problem 7.2, then it is unique and Pareto-efficient in the constraint set.

7.4.2 The Geometric Construction To construct a solution to Problem 7.2, it is useful to define Di .k/ WD f.x; z/ 2 RC  Œ0; wI  W x D z  k .wi  E .w//g ;

(7.39)

as person i’s budget line when ELIE transfers of degree k are implemented in a first-best environment. Since there is no distortionat the top,  the most productive individual must receive his first-best bundle xIF ; zF I solution to (7.40) max u .x; zI wI / s.t. .x; z/ 2 DI .k/ :  So, .xI ; zI / D xIF ; zF I : It is then possible to determine person I  1’s bundle .xI 1 ; zI 1 /. On the one hand, person I must be indifferent between .xI ; zI / and .xI 1 ; zI 1 / because we are looking for a simple monotonic chain to the left. Therefore, .xI 1 ; zI 1 / must belong to the set 

U .xI ; zI I wI / WD f.x; z/ 2 Œ0; xI   Œ0; zI  W u .x; zI wI / D u .xI ; zI I wi /g : (7.41) On the other hand, the constraint zI 1  xI 1 D k .wI 1  E .w// must be satisfied, which means that the bundle .xI 1 ; zI 1 / must belong to person I  1’s first-best budget line DI 1 .k/ : As a consequence, .xI 1 ; zI 1 / D U .xI ; zI I wI / \ DI 1 .k/ :

(7.42)

The construction then proceeds recursively until .x1 ; z1 / is obtained. Figure 7.9 illustrates it for a two-person population (in this figure, k is suppressed from the notation; x F , zF and T F correspond to first-best levels; x; z and T to second-best levels; indifference curves are labeled by productivity levels). The two budget lines for the low-skill individual cross at the new equilibrium for this individual.3 A difficulty that might arise is that the intersection may be obtained for negative values of zi : In that case, individual preferences just need to be reinterpreted: a negative quantity of labour would then correspond to a transfer in labour to person i:

3

7

An Exploration of Incentive-Compatible ELIE

225

x 1

x2 = x F2

1

2

x F1

x1 –T1

1–t

–T F1

z1

z F1

z2 = z F2

Z

–T2 = –T F2

Fig. 7.9 Closest solution in transfers: geometric construction

7.4.3 An Example Complete solutions for the approach in terms of transfers are now provided. To this aim, it is assumed that the population consists of two individuals (I D 2/ who both have Cobb–Douglas preferences, U .x; `/ D ˛ ln x C .1  ˛/ ln .1  `/ ; ˛ 2 .0; 1/ :

(7.43)

Person 2’s choices are not distorted. He thus gets his first-best bundle   D .˛ .w2 .1  k/ C kE .w// ; ˛2 C .1  ˛/ k .w2  E .w/// x2F ; zF 2 (7.44) So, his indirect utility amounts to   u x2F ; zF I w WD 2 2 # "   E .w/ 1˛ 1˛ ˛ ˛ .w2 .1  k/ C kE .w// 1  k C k ln ˛ .1  ˛/ w2 D u2 : (7.45) Person 1’s second-best bundle must belong to the left-hand part of person 2’s highest indifference curve,

226

L. Simula and A. Trannoy

  U x2F ; zF I w 2 WD 2 (   1 ˛1 ) i h i h u z 2 1 : .x; z/ 2 0; x F .w2 /  0; zF .w2 / W x D exp ˛ w2 (7.46) h i As a consequence, it is the unique solution in .x; z/ in 0; x f .w2 /  i h 0; zf .w2 / (if any) to the following system: 8 <

1

1 1 ˛

z D ˛ .1  ˛/ ˛ 1 w2



: x D z  t .w I w; k/ : 1

2 w2 C k w1 w 2

 ˛1  1

z w2

1 ˛1

2 C k w1 w 2 ;

(7.47) For illustrative purposes, we consider the same economy as in the closestsolution-in-welfare approach: ˛ D 1=2; w1 D 1 and w2 D 3=2: In this case, the first equation in (7.47) is zD

1 .3  k=2/2 k  : 8 3  2z 4

Hence, the income tax scheme is characterised by 8      p p < .x1 ; z1 / D k C 1 6  2 6k  k ; 1 6  2 6k  k ; 8 8 4  3 k 3 k : .x2 ; z2 / D 4  8; 4 C 8 :

(7.48)

(7.49)

It is therefore possible to make the ELIE transfers incentive-compatible for every k 2 .0; 1/ : Gross income and consumption levels are illustrated in Fig. 7.10, panels (1) and (2); person 1’s marginal tax rates and second-best lump-sum transfers on panels (3) and (4). The implicit marginal tax rate faced by person 1 is given by p  4 6k  1 1  ˛ x1 D T 0 .x1 ; z1 I w1 / D 1  p : (7.50) ˛ w  z1 2 C 2 6k C k Hence, the marginal tax rate faced by person 1 increases with k: Its sign is given by T 0 .x1 ; z1 I w1 /  0 , k  1=6 ' 0:1666; 0

0

T .x1 ; z1 I w1 / D 0 for k  1=6:

(7.51) (7.52)

For k  1=6; ELIE is indeed incentive-compatible as previously noted; hence both agents face a zero marginal tax rate.

7

An Exploration of Incentive-Compatible ELIE

227

(1) Consumption

(2) Gross Income

0.7

2

0.8

0.6

1:FB

0.7 0.6

0.5 1:SB

0.5

0.4

0.4

0.3

0.3 0.2

2 1:FB 1:SB

0.2

0.1

0.1

0

0 0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

k

k

(3) Marginal Tax Rate

(4) Second-Best Lump-Sum Transfer

0.9

1

0.9

1

70.00 0.25 60.00 50.00

0.20

40.00 0.15

30.00 20.00

0.10

10.00 0.00

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.05

0.2

0.3

0.4

k

0.5

0.6

0.7

0.8

k

Fig. 7.10 Closest solution in transfers: example economy

7.5 Concluding Comments When they exist, the closest solution in terms of welfare and the closed solution in terms of transfers are both Pareto-efficient. Yet, they do not belong to the same Pareto set. Indeed, the constraint set of Problem 7.2 is included in that of Problem 7.1. The income tax corresponding to the closest solution in terms of welfare may be substantially different from ELIE. This is notably the case for quasilinear preferences. For every value of k for which there is envy in the first-best, the closest solution is the laissez-faire, which is only obtained for k D 0 in the first-best. However, Kolm analysis puts the stress on the very specific shape of the ELIE transfers and does not adopt a welfarist perspective. This justifies the second approach we have proposed, in which the second-best transfers exactly correspond to the first-best ones. Our procedure yields a unique Pareto-efficient outcome in the constraint set.

228

L. Simula and A. Trannoy

Appendix The Lagrangian of this optimisation problem is 

 z1 L D 1 ˛ log x1 C .1  ˛/ log 1  w1 

 z2 C 2 ˛ log x2 C .1  ˛/ log 1  w2  

 z1  ˛ log x2  .1  ˛/ log 1  C 1 ˛ log x1 C .1  ˛/ log 1  w1  

 z2  ˛ log x1  .1  ˛/ log 1  C 2 ˛ log x2 C .1  ˛/ log 1  w2 C  Œz1  x1 C z2  x2  :

 z2 w1  z1 w2 (A.1)

The first-order conditions for a maximum are @L @x1 @L @z1 @L @x2 @L @z2 1 2

˛ .1 C 1  2 / ;  1 C  1 2  D0,  D ; z1  w1 z1  w2 1˛ ˛ D 0 , x2 D .2 C 2  1 / ;  2 C  2 1  ; D0,  D z2  w2 z2  w1 1˛  0 .D 0 if (7.34) is not binding/ ;  0 .D 0 if (7.35) is not binding/ : D 0 , x1 D

(A.2) (A.3) (A.4) (A.5) (A.6) (A.7)

Substituting (A.2) and (A.4) in the tax revenue constraint (7.36), z1 C z2 D x1 C x2 D

˛ ˛ 2˛ .1 C 2 / D , z1 D  z2 ;   

(A.8)

because 1 C 2 D 2: Moreover, note that x 1 C x2 D

2˛ 2˛ , x1 D  x2 :  

(A.9)

Case (i). We first examine whether the first-best solution remains incentive-compatible when individual productivity becomes private knowledge. In the first-best, person i’s labour and consumption choices are respectively given by

7

An Exploration of Incentive-Compatible ELIE

( `F 1 D x1F `F 2





˛ C .1  ˛/k 1  0

E.w/ w1



229

if w1 > w0 ;

(A.10)

if w1  w0 ;

˛Œw1 .1  k/ C kE.w/ k ŒE .w/  w1    E .w/ D ˛ C .1  ˛/ k 1  w2 D

if w1 > w0 ; if w1  w0 ;

(A.11) (A.12)

x2F D ˛ Œw2 .1  k/ C kE .w/ ; where w0 WD

(A.13)

k ˛ E Œw ; k C 1˛

(A.14)

is a productivity threshold under which individuals are idle, provided that w0   : We check whether there is at least one individual which envies another one. Case (ii). The incentive-compatibility constraint preventing the highly skilled from mimicking the low skilled is binding. By (A.5), z2 D w2 

1˛ .2 C 2 / : 

(A.15)

Moreover, equalising (A.3) and (A.5), 2  2 C 2 1  D ; z1  w1 z1  w2 z 2  w2

(A.16)

in which (A.8) is substituted to get rid of z1 : ˛ 

1   z2  w 1

˛ 

2  2 C 2 D :  z2  w2 z2  w2

(A.17)

Now, use (A.15) to get rid of z2 ; ˛ 

C

1˛ 

2 .2 C 2 /  22



˛ 

C

1˛ 

1 .2 C 2 /  .w1 C w2 / D

2 C 2

1˛ 

.2 C 2 /

; (A.18)

which is an equation in 2 : Solving (A.2), one gets 2 as an implicit function of ; given ˛; w1 , w2 and k; 2 D 2 . / :

(A.19)

230

L. Simula and A. Trannoy

Plugging 2 . / in (A.15), z2 is obtained as a function of ; z2 D z2 . / ; from which z1 D z1 . / because of (A.8). Substitution of z1 . / and z2 . / in the binding incentive-compatibility constraint (7.35) yields: ˛ log x2 C .1  ˛/ log 1 

w2 

1˛  .2



C 2  //

!

w2

 ˛ D ˛ log  x2 C  ! ˛ 1˛ C . C  . //  w 2 2 2   : (A.20) .1  ˛/ log 1  w2 Solving it, x2 is obtained as a function of ; x2 D x2 .w/ : Using Then, x1 is determined as x1 . / D ˛  x2 : Finally, we choose the value of  which maximises W .x1 . / ; x2 . / ; z1 . / ; z2 . // and check that (7.34) is satisfied. We check whether condition (7.32) is satisfied. If not, we would also consider the last case and then compare (i)–(iii). Case (iii). The incentive compatibility constraint preventing the low type from mimicking the high type is active. The procedure follows the same lines as in case (i), but (7.34) now plays the part of (7.35).

References Cooper, R. (1984). On allocative distortions in problems of self-selection. Rand Journal of Economics, 15, 568–577. Fleurbaey, M., & Maniquet, F. (2011). Kolm’s tax, tax credit, and the flat tax. In M. Fleurbaey, M. Salles, & J. Weymark, (Eds.), Social ethics and normative economics. Heidelberg: Springer. (forthcoming). Guesnerie, R., & Seade, J. (1982). Nonlinear pricing in a finite economy. Journal of Public Economics, 57, 157–179. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge: Cambridge University Press. Kolm, S.-C. (2007). Economic macrojustice: fair optimum income distribution, taxation and transfers. Paris: EHESS. Mirrlees, J. A. (1971). An exploration in the theory of optimum income taxation. Review of Economic Studies, 38, 175–208. Mirrlees, J. A. (1997). Information and incentives: the economics of carrots and sticks. The Economic Journal, 107, 1311–1329.

7

An Exploration of Incentive-Compatible ELIE

231

Röell, A. (1985). A note on the marginal tax rate in a finite economy. Journal of Public Economics, 28, 267–272. Simula, L., & Trannoy, A. (2011). When Kolm meets Mirrlees: ELIE. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer. (forthcoming). Stiglitz, J. (1982). Self-selection and Pareto efficient taxation. Journal of Public Economics, 17(2), 213–240.

•

Part IV

Combinations of ELIE with Other Targeted Transfers

•

Chapter 8

Is ELIE a Wasteful Minimum Income Scheme? Erwin Ooghe and Erik Schokkaert

Abstract ELIE can be interpreted as a minimum income scheme, financed by lump-sum taxes. While Kolm is careful in stating that his theory of macrojustice does not apply to individuals voluntarily working less than the “initial equal labour” k, we consider an extended scheme in which equal-labour income equalisation is also applied to these individuals. This extended ELIE may induce social waste as individuals with a low taste for working may opt for voluntary unemployment. We simulate the magnitude of this social waste with microdata for Belgium and compare extended ELIE with a first-best scheme and a second-best scheme (based on a linear income tax), implementing the same minimum income. The social waste induced by extended ELIE is intermediate between the social waste induced by the first- and second-best schemes, and remains relatively small for realistic levels of redistribution. Assumptions about the preferences of the voluntarily unemployed play a crucial role.

8.1 Introduction Serge-Christophe Kolm’s (2005) book Macrojustice is concerned with “the most general rules of society and their application to the distribution of the benefits from the main resources” (Kolm 2005, p. 1). Kolm rejects the traditional welfarist approach to income taxation and substitutes for it an ideal E. Ooghe (B) Department of Economics, UCBrussels and KULeuven e-mail: [email protected] E. Schokkaert Department of Economics, KULeuven and CORE, UCLouvain e-mail: [email protected] Erwin Ooghe would like to thank GREQAM, Marseille for their hospitality as well as the Fund for Scientific Research – Flanders for their generous support.

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_8, 

235

236

E. Ooghe and E. Schokkaert

of equal freedom. Given the crucial importance of the freedom to act, taxes and transfers must be based on so-called inelastic items, i.e., items which are not affected by individual actions. In Kolm’s view these inelastic items must be the productive capacities of the individuals. He then proposes an operational tax-benefit scheme, called ELIE (equal-labour income equalisation). The basic idea of ELIE is simple. Society fixes an amount of “initial equal labour” and distributes all the proceeds from working this amount of labour equally over all individuals. This equally distributed amount can be interpreted as a kind of minimum income. Individuals with a productivity smaller than the average in society receive a transfer, individuals with a larger than average productivity have to pay a tax. Each individual keeps full freedom to choose her actual amount of labour time and may keep for herself the income resulting from working more than the “initial equal labour”. The basic motivation for ELIE is an ethical one, i.e., the importance of respecting the freedom to act. However, in Kolm’s view, ELIE also has attractive incentive properties. It is basically a tax on wages, not on incomes, and it is incentive-compatible for all those who are working more than the “initial equal labour”: indeed, for them, additional units of labour remain untaxed and earn an additional income equal to the individual’s productivity. However, there are two potential problems with this argumentation. First, it assumes that wages are perfectly observable for all those who are working. While it is true that there is now reasonably reliable information on wages available, or that wages can be calculated from observations on income and on labour time, one may fear that this information could become much less reliable when ELIE were to be introduced. Second (as acknowledged by Kolm himself), the incentive-compatibility argumentation does not go through for those who work less than the “initial equal labour”, including the (in)voluntarily unemployed. These can get a minimum income if they do not work at all, and therefore may have an incentive to hide their true productivity. While Kolm proposes to treat the involuntarily unemployed as individuals with a zero market productivity, he explicitly excludes the voluntarily unemployed from the domain of application of macrojustice, arguing that redistribution is only acceptable with respect to resources that are actually used. One could formulate an alternative ethical principle, however, in which redistribution pertains to all resources. Moreover, when implementing ELIE in the real world, the distinction between voluntary and involuntary unemployment will be very difficult to make. In this paper we therefore analyse a redistribution scheme in which the basic principle of ELIE is extended to all individuals working less than the “initial equal labour”. To make clear that this is not exactly Kolm’s scheme, we call it “extended ELIE”.

8

Is ELIE a Wasteful Minimum Income Scheme?

237

In this paper we do not focus on the basic ethical foundations of (extended) ELIE, but on the incentive issues.1 Moreover, we will take it for granted that wages are observable for those who are working more than the initial equal labour and we will only focus on the issue of voluntary unemployment, linked to the minimum income feature of extended ELIE. In Sect. 8.2, we propose a simple model of the labour market, which is very similar to Kolm’s model. We introduce three tax-benefit schemes: (1) a first best ELIE scheme, in which productive capacities are perfectly known, (2) the extended ELIE scheme, in which productive capacities are perfectly known for the working population, but not for the (in)voluntarily unemployed, and (3) a traditional second-best (linear) income tax scheme, in which incomes are observed and taxed but productive capacities are not known. We show how extended ELIE may induce social waste. This immediately raises the (empirical) question of the amount of waste induced. In Sect. 8.3, we simulate the results for the three tax schemes with Belgian microdata. It turns out that the social waste in the extended ELIE scheme is intermediate between the other ones (as expected) and its relative magnitude is highly sensitive to some of the assumptions, especially to the taste of the voluntarily unemployed. Section 8.4 concludes.

8.2 Does Extended ELIE Induce Waste? Three Minimum Income Schemes To explain the effects of the extended ELIE tax-benefit scheme, we propose a stylised model of the labour market –which is similar to Kolm’s model (see, e.g., Kolm 2005, Chap. 9–10, pp. 144–184). We assume that individuals differ both in their productive capacities (with constant marginal productivities) and in their preferences for leisure. Tastes and productivities are assumed to be independently distributed. The continuous density function of the productive capacities w  0 is given by f (with f > 0 on RC ). Each individual belongs to one of a discrete number of taste types i 2 N D f1; 2; : : : ; ng. We use pi > 0 to denote the proportion of individuals with taste type i 2 N . To simplify the analysis we assume that preferences are quasi-linear in consumption. Summarising, we have: Assumption A1: Gross income y equals w`, i.e., a multiplication of individual productivity w  0 and (adjusted) labour `  0, which is itself

1

A related analysis can be found in Simula and Trannoy (2011).

238

E. Ooghe and E. Schokkaert

a function of labour duration, intensity, speed, and so on (Kolm 2005, Chap. 9, p. 145).2 Assumption A2: Individuals have preferences over consumption and labour. They like consumption (net income) c D y C t, with t a transfer, they dislike labour `, and preferences are strictly convex (Kolm 2005, Chap. 9, Fig. 9-1, p. 157). More specifically, the preferences of each taste type i 2 N D f1; 2; : : : ; ng can be represented by a utility function Ui W .c; `/ 7! c 

1C" 1 " ` " ; ˛i 1 C "

with " > 0 the labour supply elasticity and ˛i > 0 a taste (or ambition) parameter. We assume that, if individuals are indifferent between several consumption-labour bundles, they choose the bundle with the lowest labour time. Assumption A3: Tastes and productivities are independently distributed. Note that the preference specification in A2 excludes income effects, i.e., giving an amount of money to an individual does not change her labour supply decision (see, e.g., Atkinson 1990, Diamond 1998). A further consequence of Assumption A2 is that individuals end up with either y > 0; ` > 0 (and thus w > 0 must hold for these individuals) or y D ` D 0 (with w > 0 for the voluntarily unemployed and w D 0 for the involuntarily unemployed). Other combinations of y and ` are not possible. As a matter of definition, individuals with a ‘real’ productivity w D 0 are said to be involuntarily unemployed (irrespective of `); the voluntarily unemployed are defined as individuals with a productivity w > 0 who choose ` D 0. Traditional tax theory assumes that gross income y.D w`/ is observable, while productive capacity w is not. A linear income transfer system with a uniform lump sum transfer B and a constant marginal tax rate  can then be defined as follows: R1 LINEAR INCOME TAX : T W w` 7! B w`, with B D  0 .w`/f .w/d w On the other hand, ELIE assumes that the productivities w are observable, either directly, e.g. from a paysheet, or indirectly, by observing y D w` and `, and dividing the former by the latter (Kolm 2005, Chap. 10, p. 172).3 We Note that –in contrast with (Kolm 2005, Chap. 9, p. 145)– we do not restrict ` to be bounded in some interval, say Œ0; 1. 3 Two other possibilities arise. First, in a dynamic setting one could obtain w from previous tax files. In that case our model still applies with ` measuring life-time labour. Second, one could estimate a wage equation and “measure” w on the basis of sex, age, years of schooling, highest educational degree and so on. Although this is a promising approach, two additional problems arise: (1) how do we deal with the error term? It is indeed possible that individuals with a non-zero estimated wage 2

8

Is ELIE a Wasteful Minimum Income Scheme?

239

assume that, if the paysheet reveals w, then it also reveals both y and `, with w D y=`. As a consequence, we can write the transfer as t D T .y=`/, with T a function of ‘real’ productivity. If (all the) wages are perfectly observed, we define the first best extended ELIE transfer scheme for a given parameter k  0 as follows: FIRST BEST EXTENDED

TkFB

ELIE:

Z

W y=` 7! k .w  y=`/ ; with w D

1 0

wf .w/ d w;

where a budget constraint has been imposed such that the average extended ELIE-transfer is equal to zero. Note that the parameter k determines the degree of redistribution and plays therefore an analogous role as  in the linear income tax, while the minimum income kw is analogous to the uniform lump sum transfer B: If one wants to implement the first-best extended ELIE scheme, one immediately faces the problem that y=` is not well-defined when y D ` D 0. This raises the issue of what productivity should be ascribed to the (voluntarily and involuntarily) unemployed, and therefore the issue of what transfer they should receive. Kolm proposes that for the application of ELIE, an individual i facing the labour market constraint `i  `0i should be interpreted as having a productive capacity wi D 0 for `i > `0i ; the involuntarily unemployed (with `0i D 0) thus have wi D 0. But since we cannot distinguish between voluntarily and involuntarily unemployed, we define the ‘revealed’ productivity b w equal to 0, whenever y D ` D 0 Kolm (2005, Chap. 13, p. 215). Of course, this redefinition implies that the budget requirement also has to be adapted by replacing the mean ‘true’ Z productivity w by the mean X wD pi wf .w/d w  w, with ‘revealed’ productivity, i.e., b i2N

wj`i .w/>0

`i .w/ the labour choice of an individual with taste type i and productivity w. The extended ELIE transfer scheme then becomes EXTENDED ELIE WITH SOCIAL WASTE :   w  y=` ; TkS W W y=` 7! k b Z X def wD pi wf .w/ d w. (8.1) with y=` D 0 if ` D 0 & b i2N

wj`i .w/>0

are in reality unproductive; (2) how to deal with the incentives to invest in education? Kolm argues that educational choices are made by the parents, not by the economic agents themselves, and that “education obtained long in advance (and which provides not only income but also status and other things) is bound not to be very sensitive to tax computations performed much later” (Kolm 2004, p. 140). This intriguing statement implies that even choices at the level of higher education are not influenced by financial incentives. In the light of the empirical literature, this position is at least debatable.

240

E. Ooghe and E. Schokkaert c w¯ wO k w¯

k



Fig. 8.1 It might be beneficial to hide your true productivity

Equation (8.1) immediately shows the reason for the terminology “extended ELIE with social waste”. Indeed, it is obvious that individuals can hide their ‘true’ productivity w by not working and thus revealing a productivity b w D 0. Figure 8.1 (with labour, not leisure, on the horizontal axis) illustrates this point. The reader familiar with ELIE will recognise the family of budget curves which cross at .k; kw/. Suppose for simplicity that all individuals have the same preferences, but differ in productivity. Given wellbehaved preferences (Assumption A2), we can construct a unique cut-off wage wı such that individuals with this productivity level are indifferent between shirking and working, i.e., the indifference curve tangent to the wı -budget line goes through the point .`; c/ D .0; kw/. Furthermore, all individuals with a lower wage w < wı will shirk, while individuals with a higher wage w > wı will work. Of course, if there are individuals shirking, the budget constraint will then lead to a downwards shift of the whole transfer scheme for everybody. Since individuals can only choose to reveal their true productivity (choose b w D w) or to hide it (choose b w D 0  w), our model is a peculiar case of Dasgupta and Hammond (1980), in which individuals can choose to work w  w. Given the specific structure of at (and reveal) any productivity level b the extended ELIE scheme as displayed in Fig. 8.1, it is always the case that, if it is beneficial for someone with a productivity w > 0 to hide it and to work at a lower rate b w, with 0 < b w < w, then it is optimal for this individual w D 0. Therefore, the current model with two discrete to choose to reveal b choices b w D w and b w D 0 is sufficient to capture the essential features of Dasgupta and Hammond (1980).

8

Is ELIE a Wasteful Minimum Income Scheme?

241

We will now further analyse the labour market equilibrium if extended ELIE is implemented. Individual decisions are linked to each other through the budget requirement. Letting I ./ denote an indicator function which equals 1 in case the statement between brackets is true, and zero otherwise, we define: L ABOUR SUPPLY EQUILIBRIUM: A labour supply equilibrium for the extended ELIE scheme in (8.1) is a list `1 ; `2 ; : : : ; `n of maps `i W RC ! RC W w 7! `i .w/, one for each taste type i 2 N , such that, for each i 2 N , for each w 2 RC and for each ` 2 RC we must have      Ui w`i .w/ C k b w  wI `i .w/ > 0 ; `i .w/      Ui w` C k b w  wI .` > 0/ ; ` ; (8.2) wD with b

X i2N

Z pi

wj` i .w/>0

wf .w/ d w.

Note that, since individuals with taste type i and productivity w are assumed to be ‘atomic’, i.e., their proportion is negligible with respect to the total population, they cannot influence b w by choosing ` different from w also appears at the right-hand side of (8.2). `i .w/, and thus b Proposition 8.1 shows the existence of a unique labour supply equilibrium. It is completely defined by cut-off productivity levels wıi for i 2 N , such that (1) individuals with taste type i and a productivity w  wıi choose to hide their true productivity and thus to remain voluntarily (if w > 0) or involuntarily (if w D 0) unemployed, while (2) individuals with taste type i and a productivity w > wıi choose to reveal their true productivity and thus to work. Since there are no income effects, labour supply does not depend on the minimum income. The proof of the proposition is given in the appendix. Proposition 8.1. If Assumptions A1–A3 hold, then there exists a unique  labour supply equilibrium `1 ; : : : ; `n for the extended ELIE transfer scheme defined in (8.1) such that each individual with taste type i 2 N and ‘real’ productivity w chooses `i .w/ D 0 if w  wıi D k 1=" .1 C "/1=" =˛i and `i .w/ D w" ˛i" , otherwise. The social waste, induced by extended ELIE, is caused by the fact that some individuals prefer to shirk, i.e., to get voluntarily unemployed, if they are entitled to a minimum income when they do not work at all. The amount of social waste will then be determined by the (shirking) cut-off levels wıi : As expected, these increase with k, and hence with the level of the minimum income. They decrease with the taste for working ˛i . Indeed, as was already illustrated by Fig. 8.1, the crucial condition for shirking is that the utility of

242

E. Ooghe and E. Schokkaert

shirking with income k b w, is larger than the utility of working `i .w/ .> 0/  with income w`i .w/ C k.b w  w/: It is therefore clear that in a realistic setting, the minimum income scheme implied by extended ELIE induces some social waste. The crucial question now becomes how much social waste is induced. The next section provides an estimate, based on a (somewhat rough) calibration and simulation exercise for Belgium.

8.3 How Much Waste is Induced? An Empirical Analysis We base our simulation study on the individual observations available in the last wave of the Panel Study of Belgian Households (2001). We calibrate the parameters of our model so as to replicate as well as possible these individual data (Sect. 8.3.1). We then show in Sect. 8.3.2 the results for the benchmark values of the parameters. The most interesting insights are obtained from the sensitivity analysis in Sect. 8.3.3.

8.3.1 Data and Calibration We restrict ourselves to the individuals in the potential workforce. The latter is defined as consisting of those individuals who either (1) have work (possibly temporarily suspended), or (2) who do not have work, but are neither retired, nor sick, nor handicapped, and so on. The total sample size of the potential workforce is equal to 3 789 individuals. Unemployment among the potential workforce equals 22.8%. Figure 8.2 presents a kernel density estimate of labour supply (in actual hours worked per week) for the working population only. The ‘typical’ peaks around half-time and fulltime, together with the large group of unemployed (not taken up in Fig. 8.2) suggest to use three taste types, ˛L (low) ˛M (medium) and ˛H (high), for those voluntarily not working (` D 0), those working in between 0 and 30 hours a week (0 < `  30) and those working more than 30 hours a week (` > 30), respectively; the taste values used in the simulation will be calibrated later on. The data do not allow us to distinguish the involuntarily employed (with w D 0) from the voluntarily unemployed (with 0 < w  wıi ). We therefore define a parameter ˇ, indicating the fraction of the voluntarily unemployed in the total group of the unemployed (22.8% of the sample). In our benchmark simulation we (arbitrarily) put ˇ D 0:5. The proportions off individuals in the different groups are then equal to

8

Is ELIE a Wasteful Minimum Income Scheme?

243

Fig. 8.2 Kernel estimate of the ‘actual’ labour supply density of the working population

pL D 0:114, pM D 0:123 and pH D 0:649; the proportion of involuntarily unemployed is denoted p0 D .1  ˇ/ 0:228 D 0:114. Figure 8.3 shows a density estimate of the logarithm of ‘revealed’ gross w for the employed. If we assume that no one has an incenhourly wages b w in .0; w/, the revealed productivities of the tive to reveal a productivity b employed must correspond to their true productivities. The distribution is approximately normal with a sample mean and standard deviation equal to 2.65 and 0.36, respectively; this corresponds with a mean gross hourly wage equal to e 15: Note that lognormality cannot be statistically rejected (Shapiro–Wilk test). In Fig. 8.4 we look at the logarithmic wage density again, but now separately for the medium and the high taste types. Equality of both distributions cannot be statistically rejected (Kolmogorov–Smirnov test), which gives some support for our independence hypothesis between productivity and taste levels (Assumption A3). Finally, the current Belgian net income scheme –which maps gross into net incomes (exclusive of social benefits)4 – can be very well approximated by a loglinear scheme. Figure 8.5 plots the logarithm of net income versus the logarithm of gross income and

4

Notice that, given quasi-linear preferences in income, there are no income effects and thus, labour supply decisions are not influenced by lump-sum social transfers. However, some social transfers are not lump-sum –e.g., unemployment insurance benefits–and therefore the calibration exercise will be distorted to some extent.

244

E. Ooghe and E. Schokkaert

Fig. 8.3 Kernel estimate of the gross hourly log wage density

Fig. 8.4 Kernel estimates of the gross hourly log wage density for the medium and high taste types

8

Is ELIE a Wasteful Minimum Income Scheme?

245

Fig. 8.5 The Belgian tax system is approximately loglinear

shows a tight log-linear fit given by c 1:42y 0:76 (the explained variance equals 89%). Let us now combine all the previous information in order to calibrate the three taste values ˛i , for i D L; M; H . The taste types ˛M and ˛H are calibrated to ensure that the theoretically predicted mean labour hours per week, given the ‘estimated’ actual Belgian net income scheme c 1:42y 0:76 , is equal to the observed mean labour hours per week for both types (23.9 and 41.4 hours/week respectively). Recall that all individuals are endowed with a preference technology as in Assumption A2. We assume that the elasticity of labour supply " is the same for everyone and equal to 0.25, which lies within the range of plausible empirical estimates (see Blundell and MaCurdy 1999). Since the ‘true’ productivities are independently distributed from the taste types, the productivity distribution is the same for each taste type and given by F W w ! f .w/ with F .0/ D p0 D 0:114 (the proportion of involuntarily unemployed), and F .w/ D 0:114 C .1  0:114/ G.w/ for w > 0, with G a lognormal distribution function with mean 2.65 and standard deviation equal to 0.36 (to mimick Fig. 8.3). We then still have to fix a value for ˛L ; i.e. the preference parameter for those who are voluntarily unemployed. By definition these are individuals with w > 0: To calibrate ˛L , we assume that they would just ‘survive’ in a laissez-faire economy, where survival means working 10 hours a week (which would provide them on average with a (net

246

E. Ooghe and E. Schokkaert

and gross) income equal to e 603.68 per month). This assumption results in ˛L .10/  178 150. We will now first show the results for the benchmark case. Remember that we made three crucial assumptions about the benchmark parameters: we have assumed that ˇ (the fraction of the voluntarily unemployed in the total unemployed) D 0:5; that these voluntarily unemployed would be able to just survive in a laissez-faire economy (˛L .10/  178 150), and that the elasticity of labour supply " D 0:25: We will show in Sect. 3.3 the results of a sensitivity analysis with respect to each of these assumptions.

8.3.2 Results for the Benchmark Case Figure 8.6 shows the average gross income per capita y (in e per month) as w (in e per month). a function of the implementable minimum income k b For each value of k, proposition 1 allows to calculate the cut-off values w and k b w. On the basis of the latter value wı which we use to calculate b for the minimal income, we calculate the second-best linear tax rate which w. Figure 8.7 shows the relative implements the same minimal income k b efficiency cost of each scheme, i.e., the difference between the average

avg. gross income (Euro/month)

3000

2950

2900

2850

2800

First best Elie Second best Poly Elie

2750

2700

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.6 Average gross income as a function of minimum income, given " D 0:25 & ˇ D 0:5 & ˛L .10/

8

Is ELIE a Wasteful Minimum Income Scheme?

247

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.7 Efficiency cost as a function of minimum income, given " D 0:25 & ˇ D 0:5 & ˛L .10/

gross income of the first-best extended ELIE scheme and the scheme under consideration, expressed as a percentage of the average gross income of the first-best extended ELIE scheme. In 2002–2003 the actual minimum income in Belgium for a single individual was about e 600 per month. This corresponds to redistributing the value of one day work per week and will be the focal point of our analysis. If extended ELIE could be implemented in a first-best way, there would be no efficiency cost to society.5 This first best gross income per capita is given by the horizontal solid line in Fig. 8.6. The second-best (linear income tax) scheme shows gross income per capita for a linear income tax scheme which implements the same minimum income, i.e., with B D k b w. In this case, there is of course an efficiency cost related to the imposition of a marginal tax rate . Figure 8.7 shows that this cost increases almost linearly, and becomes about 6% of gross income for a minimum income of e 600 per month. We are most interested in the effects of the realistic extended ELIE scheme (8.1). w there is no efficiency cost at all, because all those For very low values of k b with w > 0 prefer to work.

5

Of course, this holds as long as participation constraints –the constraint that the utility of working should be higher than the utility of the bundle .c; `/ D .0; 0/– are not binding.

248

E. Ooghe and E. Schokkaert

However, once the minimum income approaches e 350 per month, the group with the lowest taste type is moving from working to not-working. From about e 500 onwards, all individuals with a low taste for work have become voluntarily unemployed, and thus, gross income remains stable again. Admittedly, this is somewhat artificial, as it follows from our assumption of a discrete number of taste types. Yet, assuming more taste types would only smooth Figs. 8.6 and 8.7. To give an idea how such a smooth scheme would look like, we added in Fig. 8.7 a (polynomial) trend line through the extended ELIE-values. The basic message remains the same: extended ELIE induces social waste, but for the benchmark values of our parameters it is still a considerable improvement compared to the current practice of second-best. At a minimum income of e 600 per month, the efficiency cost of extended ELIE is about 2%, i.e., one third of the efficiency cost of the linear income tax.

8.3.3 Sensitivity Results We now analyse to what extent the previous results are sensitive to the main choices: the labour supply elasticity ", the proportion of voluntarily unemployed ˇ, and the low taste type ˛L . In each of these cases we only show the welfare cost as a % of gross income, for different levels of the minimum income. We start with the elasticity of labour supply ". Figs. 8.8 and 8.9 present the same simulation as in Fig. 8.7, but with " equal to 0.125 (half the benchmark) or 0.5 (double the benchmark); all figures of the sensitivity analysis can be found in the appendix. Changing " has the expected influence on the incentive cost of the linear income tax, which varies between 4% (for " D 0:125) and 7% (for " D 0:5) for a minimum income of e 600 per month. The effect on the social waste induced by extended ELIE goes in the other direction: the larger ", the (relatively) closer extended ELIE comes to the first-best. Notice first that the gross income share of the (potentially) voluntarily unemployed is equal to pL .˛L /" =.pL .˛L /" C pM .˛M /" C pH .˛H /" / in a laissez-faire economy. The larger ", the lower their production share. But while in the linear income tax case all taste types continuously reduce their labour supply when the minimum income increases, this is not the case for extended ELIE, in which only the low taste types decide to become unemployed when the minimum income becomes sufficiently high. Therefore, increasing " brings the waste induced by extended ELIE (relatively) closer to the incentive cost of the first-best extended ELIE scheme. In fact, for " D 0:125, the costs of the linear income tax and of extended ELIE are similar for reasonable values

8

Is ELIE a Wasteful Minimum Income Scheme?

249

of the minimum income. For large ", extended ELIE is much more efficient than the linear income tax. The next two Figs. 8.10 and 8.11 show the sensitivity with respect to ˇ (the proportion of voluntarily unemployed) by changing it to respectively 0.25 and 0.75. If ˇ were equal to 0, there would be only involuntary unemployment, i.e. all unemployed would have w D 0, and extended ELIE would induce no waste. In fact, the efficiency cost for extended ELIE remains small for ˇ D 0:25. For increasing values of ˇ, the waste induced by extended ELIE increases, while the efficiency cost of the linear income tax slightly decreases. The latter phenomenon is due to the fact that the increase in ˇ is linked to an increase in the number of individuals with w > 0. Therefore, the same level of minimum income can be financed with a lower tax rate . For ˇ D 0:75, the welfare cost of extended ELIE approaches that of the linear income tax at a minimum income of e 500 per month, but the differences grow larger again for higher values of the minimum income. Finally, recall that the benchmark value for the low taste type ˛L is chosen such that these individuals –who are voluntarily employed in the data– would ‘survive’ in a laissez-faire economy, where survival means working 10 hours a week. This assumption provides them a (net and gross) income equal to e 603.68 per month. We change our assumptions about ˛L and assume that in a laissez-faire economy the low taste types would work either 5 hours a week (earning on average e 301:84 per month) or 15 hours a week (earning e 905:52 per month), respectively. Figures 8.12 and 8.13 present the results. The assumption on the taste type of the voluntarily unemployed is crucial. In case of ˛L .5/, the low taste types choose to be unemployed already from a minimum income of e 200 onwards, but the welfare cost of extended ELIE then stays constant at about 1.2% of gross income. In the case of ˛L .15/, the choice to shirk is postponed until e 550, but then the welfare cost of extended ELIE is increasing faster, because the voluntarily unemployed work harder in the counterfactual situation without shirking. Still, at a minimum income of e 750 per month, the welfare cost of extended ELIE is only half that of the linear income tax.

8.4 Conclusion Kolm (2005) largely focuses on the situation of individuals working more than the “initial equal labour time” k, because he considers this to be the true realm of macrojustice. In economies with a large unemployment rate, however, there is a considerable fraction of the population that works less than k. Some of these will be voluntarily unemployed, and, in fact, it is

250

E. Ooghe and E. Schokkaert

impossible for the government to distinguish the voluntarily and involuntarily unemployed. Extending the basic principle of ELIE to all individuals working less than k may induce some voluntary unemployment because it grants all unemployed individuals at least a minimum income. In this paper we analysed the resulting incentive problems and we used Belgian microdata to get a better idea about the empirical significance of the phenomenon. It turns out that it is not negligible. Yet, for reasonable parameter values, the welfare cost of this extended ELIE scheme is (much) smaller than the welfare loss induced by a linear income tax which would grant the same minimum income. As expected, a crucial role is played by the assumptions made about the preferences of the unemployed. The relevancy of our work for “Macrojustice” should be put into perspective. First, incentive issues are not the main argument in favour of ELIE. To some extent, they are only a byproduct. Our analysis does not add anything to or detract from Kolm’s ethical argumentation in terms of freedom. Second, in Kolm’s overall view, ELIE is only part of a set of coherent policy proposals, which also include traditional unemployment insurance and a change in labour market policies. Our analysis captures only one aspect of this broader program, and the results might be different in a broader setting. In fact, Kolm explicitly states that the principles of macrojustice do not apply to the voluntarily unemployed. That is the reason why we used the terminology “extended ELIE” for the scheme that we have analysed. Third, our finding that preferences are important raises interesting issues related to Kolm’s view on (macro)justice as a third best, where the first best would be a society in which individuals are sufficiently able to control the birth of their desires and the second best would be a society in which people sufficiently like each other to remove all conflicts about sharing scarce resources. In such a broader view on society and on human beings, preferences can certainly no longer be seen as exogenous. Our results suggest that more research going beyond a narrow view on individual preferences, could also throw a clearer light on the strengths and limitations of ELIE.

Appendix Proof of Proposition 1 Recall that there are no income effects, so in defining the labour supply equilibrium for extended ELIE, we can treat b w as a given scalar. The proof proceeds as follows. First, we show that, irrespective of the taste type i 2 N , (1) `i .0/ D 0 and (2) if `i .w/ > 0 for some w > 0 , then `i .w0 / > 0

8

Is ELIE a Wasteful Minimum Income Scheme?

251

for all w0  w. Both (1) and (2) the existence of a unique ˇ together imply  cut-off level wıi D sup w  0 ˇ`i .w/ D 0 for which indeed `i .w/ D 0 for w  wıi and `i .w/ > 0 for w > wıi . Afterwards, we calculate the cut-off levels for each taste type as well as the uniquely defined labour supply `i .w/ > 0 provided by individuals with w > wıi , which is simply their utility maximizing labour supply given the extended ELIE budget constraint. First, (1) is trivial from (8.2), since for w D 0 and choosing ` D 0 it reduces to     Ui k b w; `i .0/  Ui k b w; 0 ; which is, given that utility is strictly decreasing in `, only possible if `i .0/ D 0. Second, (2) is trivial for w0 D w (since `i is, by definition, a map). We show it for w0 > w by contradiction. Suppose not, thus, suppose w < w0 and `i .w/ > 0, but `i .w0 / D 0. Since type w chooses `i .w/ > 0 in equilibrium we must have that       w ; `i .w/  Ui k b w; 0 (8.3) Ui w`i .w/  k w  b which is (8.2) for ` D 0. Similarly, since type w0 chooses `i .w0 / D 0 in equilibrium, we need satisfaction of       w; 0  Ui w0 `i .w/  k w0  b w ; `i .w/ ; Ui k b which is (8.2) for ` D `i .w/. The last two equations lead to         w ; `i .w/ Ui w0 `i .w/  k w0  b w ; `i .w/ ; Ui w`i .w/  k w  b or, (given that there are no income effects),         Ui w `i .w/  k ; `i .w/  Ui w0 `i .w/  k ; `i .w/ . Given w < w0 , and given the preference technology, the last inequality is possible only if `i .w/ D k > 0. Plugging in `i .w/ D k > 0 in (8.3) yields     w; k  Ui k b w; 0 Ui k b which contradicts that utility is strictly decreasing in `, as assumed in A2. Third, call wıi the cut-off level for taste type i 2 N . Individuals with a lower w will not work, thus, `i .w/ D 0 for w  wıi . Individuals with a higher w will choose to work, which, given the functional form of preferences in A2 and the labour supply equilibrium for extended ELIE in (2), implies that `i .w/ must correspond with labour supply that maximises individual utility, thus, `i .w/ D w" ˛i" for w > wıi . The cut-off level can now

252

E. Ooghe and E. Schokkaert

be derived from the fact  productivitytype w which isindifferent  thatit is the w; 0 and w1C" ˛i"  k w  b w ; w" ˛i" , or, given between the bundles k b A2, wıi solves 

kb w D wıi

1C"

 1C"   1 "   ı " " " ı b wi .˛i / .˛i /  k wi  w  ˛i 1 C " "

which gives us wıi D .k 1=" .1 C "/1=" /=˛i for each i 2 N . For given k, this expression fixes the (unique) cut-off levels for the different taste types. We also know the labour supply for individuals with w > wıi . w and of the minimum Using (2), the corresponding (unique) value of b income k b w immediately follow. This completes the proof since the minimum income does not influence labour supply, because of the absence of income effects.

Changing the Cost of Taxation "

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.8 Efficiency cost as a function of the minimum income, given " D 0:125 & ˇ D 0:5 & ˛L .10/

8

Is ELIE a Wasteful Minimum Income Scheme?

253

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.9 Efficiency cost as a function of the minimum income, given " D 0:5 & ˇ D 0:5 & ˛L .10/

Changing the Proportion of Voluntarily Unemployed ˇ

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.10 Efficiency cost as a function of the minimum income, given " D 0:25 & ˇ D 0:25 & ˛L .10/

254

E. Ooghe and E. Schokkaert

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.11 Efficiency cost as a function of the minimum income, given " D 0:25 & ˇ D 0:75 & ˛L .10/

Changing the Low Taste Parameter ˛L

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.12 Efficiency cost as a function of the minimum income, given " D 0:25 & ˇ D 0:5 & ˛L .5/

8

Is ELIE a Wasteful Minimum Income Scheme?

255

Cost (% of avg. gross income)

10 Elie Second best Poly Elie

8

6

4

2

0

0

100

200

300 400 500 600 Minimum income (Euro/month)

700

800

Fig. 8.13 Efficiency cost as a function of the minimum income, given " D 0:25 & ˇ D 0:5 & ˛L .15/

References Atkinson, A. (1990). Public economics and the economic public. European Economic Review, 34(2–3), 225–248. Blundell, R., & MaCurdy, T. (1999). Labor supply: a review of alternative approaches. In O. Ashenfelter, & D. Card (Eds.), Handbook of labor economics (Vol. 3, Chapter 27, pp. 1559–1695). Amsterdam: Elsevier. Dasgupta, P., & Hammond, P. (1980). Fully progressive taxation. Journal of Public Economics, 13(2), 141–154. Diamond, P. (1998). Optimal income taxation: an example with a U-Shaped pattern of optimal marginal tax rates. American Economic Review, 88(1), 83–95. Kolm, S.-C. (2004). Liberty and distribution: macrojustice from social freedom. Social Choice and Welfare, 22, 113–145. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge: Cambridge University Press. Simula, L., & Trannoy, A. (2011). When Kolm meets Mirrlees: ELIE. In M. Fleurbaey, M. Salles, & J. Weymark (Eds.), Social ethics and normative economics. Heidelberg: Springer.

•

Chapter 9

ELIE-Minating Poverty? Limits of the Mechanism and Potential Improvements Alain Leroux and Justin Leroux

Abstract Using French data, we show that ELIE performs rather weakly when it comes to addressing the issue of poverty. Yet, eliminating poverty is also a valid normative property of any redistribution mechanism. We suggest combining ELIE with another redistributive solution aimed specifically at alleviating poverty: the personal allowance (PERAL) mechanism (Leroux 2004 and 2007). We argue that ELIE and the PERAL mechanism, more than being compatible, are in fact complementary.

9.1 Introduction In his presentation of the general structure of ELIE, (Kolm 2005) convincingly argues that the procedure is at the same time just, efficient and incentive compatible. These appealing properties are the three main concerns of the literature on mechanism design, but have rarely been achieved in combination, a frequent incompatibility of which the well-known Gibbard–Satterthwaite theorem is an emblematic example in the voting context. Thus, Kolm’s discovery can only be commended. Yet, because Kolm’s concern is the redistribution of income among the members of a vastly heterogeneous society, some rich, some poor, one cannot help but wonder whether ELIE is able to satisfactorily address the issue of poverty in developed countries. As it turns out, ELIE performs relatively poorly in this sense. Using French data, we show that the amount of redistribution–represented by the A. Leroux (B) GREQAM, Université Paul Cézanne e-mail: [email protected] J. Leroux HEC Montréal, CIRANO and CIRPÉE e-mail: [email protected]

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_9, 

257

258

A. Leroux and J. Leroux

parameter k of ELIE–necessary to eliminate poverty in France would be unreasonably large. Hence, despite the major appeal of the “macro” properties of ELIE, such as process freedom and Pareto efficiency, we deem this “micro” drawback to be an important limit of ELIE. We propose to augment ELIE with a mechanism aimed specifically at alleviating poverty: the personal allowance (PERAL) mechanism. We argue that the PERAL mechanism is very much in line with the driving principles motivating ELIE. Moreover, from a practical standpoint, we build upon previous studies (Leroux 2004, 2007) to suggest that redirecting a small fraction of the income collected with ELIE (the equivalent of a bit more than 1 hour per week) toward the PERAL mechanism would be enough to eliminate poverty altogether. Our claim regarding the limit of ELIE we identified here and the potential gains achievable by incorporating the personal allowance should not be interpreted as a criticism of ELIE or, if so, merely as a constructive one. In fact, Kolm himself suggests in this volume (p. 51) that ELIE must be amended with other “micro” mechanisms: “Macrojustice concerns the most general rules of a society, and notably their application to the distribution of the value of the main resources in general income. Society also faces innumerable cases of microjustice concerned with allocations specific in terms of the nature of goods, reasons, people, occasions, circumstances or criteria. [...] [T]hese “spheres of justice” of very different sizes are complementary.”

9.2 Limits of ELIE This section aims at illustrating the fact that ELIE is unable to satisfactorily solve the poverty issue by examining its limits using French data. Before doing so, we first consider a very simplified example, using a linear income distribution, in order to describe the inner workings of ELIE with respect to the issue of alleviating poverty.

9.2.1 A Simplified Example: A Linear Income Distribution We consider here a simplified situation in which the income distribution is linear in the population-income space. In other words, the top income of the d th decile is equal to d times the top income of the first decile of population. Such an income distribution is far from being realistic but exhibits neat graphical properties which we shall exploit to illustrate the consequences of ELIE with respect to poverty relief. See Fig. 9.1.

9

ELIE-Minating Poverty?

259

Income

pre -ELIE distribution

post -ELIE distribution

ymean = ymed p poverty line ( = 0 5ymed) k ∗ ymean

1st

nd

2

rd

3

th

4

th

5

th

6

th

7

th

8

th

9

th

10

Population (deciles)

Fig. 9.1 ELIE on a linear income distribution

We denote by ymean and ymed the mean and median incomes of the population, respectively. While ymed is typically smaller than ymean in practice, the two values coincide in this simplified example. The value of ymed plays no role in the determination of the income transfers associated with ELIE but is instrumental in defining the poverty line in the case of France, which will be the country of interest in the next sections. Specifically, in France, any person whose income is less than half the median income is officially considered poor.1 Thus, the aim of this exercise is to describe the evolution of the proportion of the population falling below the poverty line as the redistribution parameter, k, varies. Recall that “[c]oefficient k [...] is a degree of redistribution, equalisation, and solidarity [...]” (Kolm, this volume, p. 118). Strictly speaking, this redistribution parameter, k, is interpreted in Kolm (2005) as a duration of labour, which can be adjusted to take into account other dimensions of the labour task considered such as effort, risk involvement, tediousness, etc. However, there is also a sense in which k can be interpreted as a proportional income tax.2 Indeed, if we assume in a rough approximation that the large majority of the population works full time, so that income is a good proxy for wage, collecting the fruit of a given duration of labour (as in ELIE) amounts to collecting the corresponding percentage of

1

Note that in other countries, such as in the US, the poverty line is not determined relative to the income distribution within the nation but relative to the income needed to purchase a minimal bundle of goods deemed necessary for a decent standard of living. The choice of a poverty line is beyond the scope of this text and we shall simply take as given the criterion used in France, given that it is our country of study. 2 Note that in other contributions to this volume, the parameter k also includes leisure time.

260

A. Leroux and J. Leroux

income (as a proportional income tax would). Note, however, that this rough “equivalence” only holds quantitatively but not, for instance, in terms of the disincentives usually associated with a proportional income tax. It follows from the above approximation that the net income transfer awarded to individual i amounts to k.ymean  yi /. Clearly, this transfer is positive for individuals whose income is below the mean income level and negative for those whose income is above it. Graphically, it follows that the post-transfer distribution is also a line, which is obtained by rotating the original line clockwise around ymean (and, de facto, around ymed ). Moreover, because ymed is not affected by the transfers stipulated by ELIE, neither is the poverty line. Hence, ELIE necessarily reduces the size of the poor population, p. In fact, poverty is eliminated for any value of k greater than 0.5, which amounts to the equivalent of 2.5 days of labour per week. This value is quite large compared to current income tax rates, even for France. But, the pre-transfer income distribution being arbitrary, this result is not grounded in any reality and there is no need for alarm (yet).

9.2.2 ELIE in France We are now ready to consider real data. By applying the same reasoning to the income distribution of France, we are able to answer the following question: Can ELIE eliminate poverty? Clearly, it can, as a value of k D 1 results in the egalitarian allocation of income. Nevertheless, as Kolm argues, such an extreme value of k is untenable on grounds of social stability (unanimity would massively be violated) and incentive compatibility (why should I work at all if my final income is independent of my effort?). Hence, a more interesting question is: Can ELIE eliminate poverty for reasonable values of k? We argue that the value of k necessary to relieve France of poverty is unreasonably large. In this section, we first determine this value of the redistribution parameter and then discuss the practical consequences of implementing such a value. The current income distribution of France is depicted in Table 9.1, which contains information on primary incomes (i.e. before tax and prior to any redistributive transfers) by decile. The contents of Table 9.1 were inferred from INSEE data on total income and on financial aid (see Table 9.2 and Table 9.3 in the Appendix). This income distribution is almost linear for the lower half of the distribution, with a slight concavity for very low incomes, and a sharper convexity at high incomes (see Fig. 9.2). However, one should note that the current

9

ELIE-Minating Poverty?

261

Table 9.1 Primary income in euros by decile (Year 1999) Decile 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th

Top income in decile 4,660 8,828 12,210 15,738 19,476 23,667 28,568 35,149 46,082 –

Mean income in decile 2,453 7,417 10,912 14,295 17,807 21,676 26,178 31,774 40,061 69,510

Mean income within decile 70000 60000 50000 40000 30000 20000 10000

1st

2nd

3rd

4th

5th

6th

7th 8th

9th

10th

Fig. 9.2 Primary income distribution for France (in deciles, Year 1999)

income distribution of France is necessarily a consequence of the various policies in effect in France, including minimum wage policies. Because ELIE itself takes care of such issues as minimum wage, it is not meant to be implemented on top of current minimum wage policies, but instead of them. Therefore, before applying ELIE, we must first make an effort to recover what the income distribution would have been were these policies not in effect. Because such counterfactual information is–of course–unavailable, we opted for correcting the current distribution in the following simple but conservative fashion. First of all, we should consider that current minimum wage policies only affect the lowest portion of the income distribution. In other words, we take the view that the convex shape obtained for high incomes in the actual data accurately reflects the shape of the counterfactual distribution. Moreover, one

262

A. Leroux and J. Leroux

70000

Mean income within decile Linearly adjusted portion

60000 50000 40000 30000 20000 10000

1st 2nd

3rd

4th

5th

6th

7th

8th

9th 10th

Fig. 9.3 Couterfactual primary income distribution for France (Year 1999)

may expect this counterfactual distribution to be convex for low incomes as well, instead of the concave shape we actually observe. Thus, for simplicity, we shall correct for the effect of minimum wage policies by adopting a linear distribution in the low-income range (i.e., up to the median earner, see Fig. 9.3). This rectification–literally–is a conservative one and will actually underestimate the value of k necessary to eliminate poverty, as we shall discuss later. We can now evaluate the performance of ELIE concerning the alleviation of poverty. Values of interest after redistribution, such as the minimum and median income level, the poverty line and the proportion of the population falling below it can all be expressed in terms of k and the (pre-transfer) values of ymean and ymed . Specifically, the post-transfer minimum income level equals k  ymean , which is an immediate consequence of the lumpsum feature of the redistribution of ELIE. Similarly, the post-ELIE median income level: 0 ymed D ymed C k.ymean  ymed /: Also, recall that in France the poverty level is tied to the median income 0 . Finally, our linand its post-transfer value is therefore equal to 0.5 ymed ear approximation of the lower half of the income distribution allows us to compute the proportion of the population falling below the poverty level: Â Ã kymean 1 0 p D 1 4 .1  k/ ymed

9

ELIE-Minating Poverty?

263

Numerically, with values of ymed D e19; 476 and ymean D e24; 208 from the data, we obtain that eliminating poverty (so that the minimum income level becomes equal to the poverty level) would require a value for k of at least 0.446. It is interesting to compare this value (of 44.6%) with the 43.6% at which French citizens are currently taxed on average (including direct and indirect taxes, as well as social contributions). Moreover, as the current socio-economic context suggests, it may be difficult to convince French citizens to pay more taxes than they already do. Hence, one can expect the implementation of ELIE to face strong resistance from the citizens of France. Not only that, but devoting 44.6% of the income of the citizens solely to redistribution would imply that an additional 10 or 12% would be needed to maintain some form of government (national defense, police, judicial system, etc.). Also, considering that 43.6% of income currently finances almost entirely education and health care, we argue that eliminating poverty in France via ELIE would be a difficult pill to swallow for the French population. How much poverty can ELIE reasonably eliminate in France? Considering that roughly 10% of GDP is necessary to ensure the maintenance of a minimal government structure, and taking current average tax rate as given at 43.6% (thus corresponding to a value for k D 0:33, after financing of a minimal government), ELIE would reduce poverty to roughly p 0 D 9:7% of the population (down from p D 19% if no redistribution were carried out, Fig. 9.4). Once again, it is worth comparing this value to the current size of the poor population, which is of 7%. In other words, ELIE would be less effective than the current redistributive system (even without the burden of financing education and health care, thus making the poor effectively even poorer). Our conclusion is that ELIE cannot satisfactorily eliminate poverty under its current form, but must be modified in order to do so. We devote the remainder of the text to providing specific suggestions towards improving the performance of ELIE.

9.3 The Personnal Allowance (PERAL) Mechanism The sole purpose of the PERAL mechanism is to redistribute wealth so as to alleviate poverty. As such, its goal is to channel wealth from the well-todo to the needy, which appears to be at odds with a procedure like ELIE which awards every citizen the same lump sum. And, while ELIE’s collection process does differentiate between individuals, thus leading to the needy to become “net-receivers” and the well-to-do to become “net-givers”,

264

A. Leroux and J. Leroux Mean income with in decile

70000

Post-transfer distribution

60000 50000 40000 30000 20000 10000

ymed ymed p p 1st

2nd

3rd

4th

5th

6th

7th

8th

9th

10th

Fig. 9.4 Post-ELIE income distribution for France

this fixed lump sum is not sufficient to pull all the needy above the poverty line, as we illustrate in the previous section for the case of France. Moreover, as Figs. 9.1 and 9.3 indicate, those which are pulled out of poverty are the ones closer to the poverty line. While this feature makes good sense on efficiency grounds, it leaves an entire population of individuals which are hopelessly doomed to remain poor. In the absence of a better mechanism, the performance of ELIE regarding the issue of poverty could be deemed acceptable (if poverty were ever an acceptable circumstance). However, it turns out that such a mechanism does exist: the PERAL mechanism, which we describe in this section (see Leroux 2004, 2007 for a more detailed exposition of the mechanism). Preliminary studies show that the PERAL mechanism could alleviate poverty entirely at the much cheaper fare of 3% of GDP (in the case of France3 ) compared to the 44.6% of GDP (corresponding to the value k D 0:446 previously calculated). This section is devoted to describing the PERAL mechanism as well as making explicit its relationship with ELIE. Far from being at odds with one another, we find that both mechanisms are likely to be complementary, with ELIE governing the general redistribution

3

See Leroux (2004).

9

ELIE-Minating Poverty?

265

of wealth while the PERAL mechanism fine-tunes the new distribution so as to alleviate poverty.

9.3.1 The Mechanism The mechanism takes place in two steps. First, the government collects a portion of the citizens’ income and redistributes the generated revenue equally among all citizens (just as ELIE would) in the form of vouchers. We call this lump sum their base allowance. So far, the PERAL mechanism is entirely compatible with ELIE, if not identical to it. Indeed, the PERAL mechanism does not specify the collection method, which could very well be conducted as in ELIE, with its own value of the redistribution parameter, l, taken to be the analog of Kolm’s k, though much smaller in size.4 The second step concerns the final redistribution of the citizens’ base allowance, which is done via a new type of entity, which we call redistributive mutual companies. These mutual companies are groups which citizens voluntarily create and join, and to which they deposit their base allowances (vouchers). They then reallocate the corresponding revenue between their members, with the objective of targeting the needy. Before discussing the relationship between ELIE and the PERAL mechanism, and given the novelty of redistributive mutual companies, we devote the next section to describing their inner workings in some detail.

9.3.2 Redistributive Mutual Companies As the name suggests, the purpose of redistributive mutual companies is to redistribute wealth. Also, as described in the previous section, the total amount of wealth to be redistributed consists of the sum total of its members’ base allowances. What is less clear, at this point, is the composition of these mutual companies, which is what we shall now specify, along with precise rules as to their function. Redistributive mutual companies are created by citizens. In other words, anyone can create a redistributive mutual company so long as certain rules are met. First of all, a specific criterion governs the eligibility of individuals

4

Thus, the tax revenue collected would be based on the sum of the parameters, kCl: For illustration purposes, a value of l D 0; 03 corresponds to 3% of GDP, which is all that would be necessary to alleviate poverty in the case of France.

266

A. Leroux and J. Leroux

to adhere to a given mutual company: all members of a given mutual company must have at least one observable characteristic in common. In fact, this common characteristic de facto defines the mutual company, and becomes its label. This characteristic could be tied to geographical status (e.g., people born in a given region) or to a hobby (e.g., people who take part in a given sport), but many others possibilities can arise spontaneously as more and more people relate to given characteristics. Indeed, because redistributive mutual companies would result from the personal initiative of citizens, and are not imposed by any governing authority, their defining labels should reflect the member’s sense of identity. The reason for placing this condition on eligibility is to foster sympathy and, in turn, voluntary redistribution between members. By joining a mutual company which is built around a characteristic with which they identify, people are likely to feel sympathetic towards their fellow members, simply because they can relate to them via their common characteristic. As a result, helping out a fellow member in distress becomes easier. The phenomenon of increased generosity is likely to arise even if many members of the same mutual company may never meet.5 Indeed, redistributive mutual companies may become quite large. In addition, they should be required by law to be of a minimum size, say of a few thousand members, so as to avoid clannish behaviour. Without such a lower bound on size, the common characteristic which defines a given mutual company may be so specific as to intentionally exclude the vast majority of the population. On the contrary, the purpose of mutual companies is to include every citizen in the PERAL mechanism, and to ensure that everyone will be able to choose one from several mutual companies. Moreover, certain characteristics may not be allowed to be the defining characteristic of mutual companies. It would be the governing institution’s role to forbid mutual companies based around criminal activity or indecent behaviour. Likewise, mutual companies defined by political, religious or ethnic backgrounds could be deemed as inappropriate. In addition to the above-mentioned instances, the governing institution may also intervene to ensure that each redistributive mutual company abides by the regulations in place. In particular, their accounting would be subject to 5

Increased generosity can be seen as the result of two reinforcing effects. A direct effect, due to proximity to the net receiver, via the common characteristic, makes givers derive greater satisfaction from knowing that their contribution will be targeted towards someone with whom they feel a certain affinity. But an indirect effect, albeit a second-order one, accentuates this increase in generosity. Indeed, because net receivers are benefiting from contributions from their fellow members, they are likely to be more parsimonious in their demands. As a result, and knowing that the net receivers will likely be more considerate than if they were supported by an anonymous government, givers will have an even greater desire to help their fellow members.

9

ELIE-Minating Poverty?

267

scrutiny as well as their procedure to reallocate the funds among its members. The PERAL mechanism being one of assistance to the needy, with the elimination of poverty as its ultimate goal, any subversive reallocation of funds (e.g., nepotism, lotteries, etc.) would be forbidden. However, aside from guaranteeing that funds be used towards alleviating poverty, the governing institution is not allowed to interfere in the actual method of redistribution within mutual companies. Redistribution is decided upon collectively via democratic election. In addition, each mutual company is free to implement as little or as much redistribution as decided upon by its (democratically elected) board. Naturally, people in need would choose to avoid companies which decide to not operate any redistribution whatsoever–thus “refunding” each member their base allowance, minus a possible tax incentive–in favour of those which redistribute more. Such behaviour is understandable, as people in need are likely to value financial aid more than group identity, and vote with their feet accordingly. Freedom to choose one’s mutual company, and to switch away from it as its mode redistribution ceases to meet one’s expectations, is essential to achieving an equilibrium state where the needy receive sufficient financial support. Now that the composition and the internal structure of redistributive mutual companies have been clarified, we turn to perhaps the most important interrogations. Will it be enough to alleviate poverty? At what price?

9.3.3 Redistribution It is quite natural to be skeptical regarding the performance of the PERAL mechanism. Unlike existing financial assistance programs, which allocate resources to the needy according to a predetermined schedule on observable life circumstances (number of children, employment status, income level, etc.), the PERAL mechanism relies entirely on the–optional–sympathy that citizens may feel towards one another. Therefore, concerns as to whether enough redistribution will be carried out so as to eradicate poverty are legitimate. Indeed, should every mutual company choose to refund all members their base allowance, there would be little hope.6 In preliminary studies, Leroux (2004, 2007) shows evidence that more than sufficient redistribution would take place to eliminate poverty under the PERAL mechanism in the case of France. The study relies on interview data 6

Nonetheless, the presence of a fiscal incentive to redistribute should generate some revenue, which can then be devoted to aiding the needy.

268

A. Leroux and J. Leroux

which, despite the care taken in designing the survey and the selection, is not exempt of shortcomings. Nonetheless, even pessimistic corrections of the data lead to the conclusion that people would be generous enough to channel more than enough wealth towards the needy. In comparison to the existing aid programs in France, Leroux (2004, 2007) finds that the PERAL mechanism is likely to perform better (i.e., to eradicate poverty entirely, whereas the current programs still leave 4 million people below the poverty line, including 1 million children) with less funds (amounting to a base allowance of less than 73 Euros per month). The study assumes a fiscal incentive of 50%, meaning that 50 cents of every euro of base allowance not devoted to redistribution would be collected by the governing authority as tax revenue, which would then be reinvested in the mechanism.7 Knowing that the extent to which sympathy towards people in need translates into actual wealth being transferred is extremely encouraging. However, one must also make sure that this burst of generosity is efficiently allocated. Two major concerns arise. The first one relates to the PERAL mechanism itself, and to the possible imbalance in the formation of the redistributive mutual companies. For instance, one may be concerned that mutual companies form around characteristics which are highly correlated with income. In other words, there may be too many “rich companies” and “poor companies”, both reluctant to redistribute wealth, either by choice (the former) or by necessity (the latter). As a solution against such an undesirable structure of mutual companies, the governing authority would be able to establish public mutual companies, whose financing is augmented via the tax revenue from the fiscal incentive to redistribute (at a rate of 50% in the study described above). The sole purpose of the public mutual companies is to redistribute this tax revenue among the poor who were unable to find a mutual company capable of rescuing them satisfactorily. Finally, another major concern relates to the redistribution of wealth within mutual companies, which we now describe. Unlike current government programs, which PERAL does not award financial assistance based on objective characteristics, so how can one make sure that aid will adequately target the needy as intended by the program? Actually, recipients of financial aid in the PERAL mechanism would be determined via face-to-face dialogue between potential recipients and representatives of the

7

It should be noted that such second order taxation would be in addition to–and independent of– the lump-sum collection carried out initially (and governed by the parameters k and l), which may appear to be in contradiction with the principles of ELIE. Nonetheless, a study in Leroux (2004) evaluates at 0.4% of GDP the size of this tax revenue necessary to induce sufficient participation to eradicate poverty in France, which is of the order of k/100. We thank Claude Gamel for this observation.

9

ELIE-Minating Poverty?

269

mutual companies in charge of evaluating the needs of the claimants. These representatives should also be members of the mutual company who choose to carry out their duty voluntarily, truthfully and impartially. At first glance, such a glowing description seems to apply only to a select few, rendering the mechanism seemingly inoperable and in danger of being crippled by corruption. Yet, many individuals exist who fit the description accurately. They are among us, and most of them are not high-profile individuals but simply people who wish take part in the well-being of their community and who are neither after glory, money nor any tangible advantages. Indeed, the qualities we envision for representatives of mutual companies are similar to those possessed by the municipal counselors of our local governments today. Finally, talk of the private management of public funds often raises eyebrows. This legitimate concern stems from the very strength of the mechanism: if mutual companies are able to decide who will be the beneficiaries of the mechanism, how can abuse be prevented? It is expected that unscrupulous individuals would attempt to exaggerate their needs in order to take advantage of the fact that wealth is not be distributed according to objective and observable characteristics, but through dialogue. In addition to government audits from the Internal Revenue Service which will discourage blatantly unreasonable reallocations, it would be up to the representatives to distinguish between legitimate claims and exaggerated ones. Because representatives of mutual companies would only be net givers, they would have little interest in manipulating the redistribution process. Finally, rotation schedules whereby no representative would be assigned to the same claimant for a long period of time should discourage collusion with the recipients of aid. To sum up, the fact that resources are redistributed according to criteria decided upon within the mutual companies is actually the main strength of the PERAL mechanism. Because members of a mutual company share at least one common characteristic of their choice and, in all likelihood, a characteristic which is most meaningful to them, they are likely to feel closer to each other than to the rest of the population. Hence, they will be more prone to redistributing wealth among themselves (as opposed to the anonymous government or, worse, to some presumed unscrupulous individual who decides to take advantage of the system). Moreover, members who end up being net receivers will be less likely to exaggerate their needs, knowing that the burden will fall on their fellow members (once again, as opposed to the anonymous government or to a mass of strangers towards whom they feel no affinity whatsoever). By emphasising closeness (of tastes, of geographical location, or otherwise), the personal allowance mechanism reinforces the individuals’ proclivity towards solidarity or, more specifically, towards

270

A. Leroux and J. Leroux

mutual aid.8 Moreover, as is intended, this “mutual aid via closeness” argument (increased generosity in giving combined with more restraint in asking) would still hold even when the size of these mutual companies reaches thousands of members.

9.4 ELIE and the PERAL Mechanism The PERAL mechanism described in the previous section has no other goal than to eliminate poverty in rich countries. In other words, it is not aimed at solving issues of redistribution beyond that of poverty. Hence, the PERAL mechanism is a “micro” mechanism, grounded on fairness considerations at the microeconomic level, whereas ELIE is a “macro” mechanism, aimed at respecting macrojustice (Kolm 2005). Therefore, these two mechanisms are not competing with one another. In fact, we show here that they are complementary, both in the normative sense and in terms of implementation. Naturally, the first step of the PERAL mechanism in our description (i.e., its financing) can be seen as a direct offshoot of the ELIE mechanism. Hence, it will suffice to show that the second step of the PERAL mechanism (redistribution) is in line with ELIE’s intrinsic properties of process freedom and Pareto efficiency. The PERAL mechanism clearly respects process freedom, as all citizens are free to take their base allowance to the mutual company of her choice (and switch at any time), be it by intrinsic affinity or because the redistributive policy suits them best. Moreover, within a mutual company, they are free to vote for whatever redistributive policy they choose to support and define the company’s guidelines for redistribution. Lastly, we check that Pareto efficiency is satisfied by establishing that the redistributive structure of the PERAL mechanism does not give rise to perverse disincentives. In other words, one must ensure that no individual (or only a small minority) will decide to work less in anticipation of receiving a larger personal allowance (recall that the “collection” portion of the mechanism is already taken care of, thanks to Kolm’s profound analysis of ELIE). Because transfers are made within mutual companies, and because the members of a given company are somewhat close, our “mutual aid via closeness” argument above applies, and we can expect such manipulative behaviour to be kept to a minimum.

8

See Leroux and Leroux (2009) for a formal distinction between the concepts of solidarity and mutual aid.

9

ELIE-Minating Poverty?

271

Hence, the PERAL mechanism is entirely in line with the driving principles of ELIE. Moreover, one can expect this mechanism to solve the issue of poverty elimination using only a fraction of the income collected with ELIE. Indeed, collecting 3% of GDP corresponds to an equalising labour, `, of barely more than 1 hour per week.

9.5 Conclusion We illustrated, using the example of France, that ELIE would not perform satisfactorily regarding the issue of poverty elimination. Despite its major appeals as far as macrojustice is concerned, we showed that the redistributive parameter, k, would have to be unreasonably large in order to eliminate poverty. Moreover, for values comparable to the current average tax rate, we saw that ELIE performed much worse than existing aid programs. To correct for this “micro” (but important) drawback, we suggested amending ELIE with another mechanism which is aimed specifically at eliminating poverty while retaining the driving principles of freedom and Pareto efficiency: the PERAL mechanism. Backed by previous studies, we argued that the PERAL mechanism could eliminate poverty using only a small fraction of the income collected by ELIE (the equivalent of a bit more than 1 hour per week). We argue that ELIE and the PERAL mechanism, more than being compatible, are in fact complementary.

Appendix Table 9.2 Mean yearly French household income (excluding capital income), in deciles, in euros, and portion of total income 1st decile 2nd decile 3rd decile 4th decile 5th decile 6th decile 7th decile 8th decile 9th decile 10th decile

Top income in decile

Mean income in decile

% of total income

7,304 11,091 14,099 17,219 20,631 24,653 29,361 35,757 46,642

3,845 9,318 12,601 15,640 18,863 22,579 26,904 32,324 40,548 69,930

1 4 5 6 7 9 11 13 16 28

(Source: «INSEE, revenus fiscaux 1999, hors revenus du patrimoine»)

272

A. Leroux and J. Leroux

Table 9.3 Proportion of financial aid in gross income by decile, in % Decile « Alloc, familiales » « Autres prest, familiales » « Alloc, logement » « Minima sociaux »

1st 6.6 5.3 12.9 11.4

2nd 4.4 4.1 7.6 4.3

3rd 3.1 3.2 4.4 2.7

4th 2.3 2.3 2.4 1.6

5th 1.8 1.7 1.2 0.9

6th 1.5 1.3 0.6 0.6

7th 1.1 0.9 0.3 0.4

8th 0.9 0.5 0.1 0.2

9th 0.7 0.3 0.1 0.1

10th 0.4 0.1 0.0 0.1

(Source: «Enquête revenus fiscaux 2000, INSEE-DGI»)

References Kolm, S.-C. (2005). Macrojustice. Cambridge: Cambridge University Press. Leroux, A. (2004). Eliminer la pauvreté en France. Paris: Economica. Leroux, A. (2007). Peut-on éliminer la pauvreté en France? Paris: Economica. Leroux, A., & Leroux, J. (2009). L’allocation personnelle: une nouvelle approche de l’assistance, entre solidarité et entraide. In A. Leroux, & P. Livet (Eds.), La pauvreté dans les pays riches (pp. 363–378). Paris: Economica.

•

Part V

Econometric Evaluations of ELIE

Chapter 10

The Redistributive Aspects of ELIE: A Simulation Approach Michel Lubrano

Abstract This paper analyses the problems linked to the implementation of the Equal Labour Income Equalisation (ELIE) scheme proposed by Kolm (2005). It successively studies the influence of uncertainty in the information about individual incomes, the impact of equivalence scales and finally the consequences of capital accumulation. If uncertainty does not modify fundamentally the equity properties of ELIE, equivalence scales can have non trivial consequences depending on the relation between income and fertility. Finally, capital accumulation introduces strong inequalities in the income distribution which are not removed by taxation. The paper relies on simulations of the income distribution, calibrated on French data and on the use of taxation indices.

10.1 Introduction Macrojustice, as discussed in Kolm (2005) concerns the rules for distributing the benefits of the main social resources. Because initial dotations are unequally distributed including human capacities, macrojustice implies some kind of redistribution which depends on the socially accepted degree of solidarity or altruism. Kolm (2005) considers that the main resources are labour capacities so that redistribution should imply an equal labour income

M. Lubrano GREQAM and CNRS, Marseille e-mail: [email protected] This paper is a revised version of the paper given in Marseilles for the workshop on Macrojustice organised by IDEP around Serge-Christophe Kolm in April 2006. The author thanks all the participants for their comments and especially Serge-Christophe Kolm. Later comments by David de la Croix, Kaddour Hadri, Claude Gamel and Nicolas Gravel greatly improved the paper. Usual disclaimers apply.

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_10, 

275

276

M. Lubrano

equalisation or ELIE. “Practically, each individual receives or yields in proportion to the difference between her wage rate and the average, the proportion being the equalisation labour. Hence, a difference in productivity is compensated by a proportional difference in income transfer.” As capital was initially produced by labour, the real means of production at the steady state are labour and non-human natural resources. The contribution of labour to total resources is accounted for 97%. Consequently, Kolm (2005) assumes that the question of taxation can be centred on labour only: Labour is the main source of income in society. This being said, let us turn to usual public finance theory in order to situate the possible impact of ELIE. Public finance has different aims and functions that can be summarised as follows. – A redistributive function which is based essentially on progressive income tax and various policies of assistance such as the RMI (minimum income for insertion) in France or allowances for old people. Obviously an implementation of ELIE should replace most of these existing taxes and transfers.1 – An incentive function. For instance, a carbon tax aiming at reducing CO2 emissions, or family allowances that aim at promoting a greater number of children per family. ELIE cannot be a full substitute for these taxes and transfers. – The production of public goods. This aspect of public finance aims at organising and improving economic growth and capital accumulation. Here again, we are outside the role and attributions of ELIE. These functions cannot of course be totaly independent one from another. But we would like to be able to separate as much as possible these different effects. We would not like for instance that a redistributive policy might generate disincentives or that an incentive policy could have regressive distributional features. This is one of the reasons why ELIE is based on capacities and not on effective labour supply. As its basis for taxation is inelastic, there should be no inefficiency disincentives for labour supply with ELIE. The simple presentation model of Kolm (2005) assumes that the distribution of talents is perfectly known. We shall detail a number of reasons why this might not be the case. It is thus reasonable to investigate the sensitivity of the ELIE scheme to an uncertainty in the information about the distribution of talents. Are the basic properties greatly modified or is uncertainty

1

ELIE can be completed by taxes and transfers for solving questions of meso and micro justice if the value of w N and the chosen k are such that the allowance k w N is still too low.

10 The Redistributive Aspects of ELIE

277

only a minor problem? This question is essential for judging the redistributive performances of ELIE. In other words, from a purely redistributive point of view, is it worth changing the existing system for ELIE? Modern societies and particular France are concerned with the behaviour of households as regards to fertility decisions. Family allowances have been implemented as an incentive to fertility. They have been quite successful in France for maintaining a decent rate of birth, compared to other European countries. But family allowances, on the other side, might have some unwanted redistributive effects. So what is the exact trade-off between family incentives and redistribution? Because of their incentive motives, ELIE cannot be a substitute to them, but has to be a complement. How can the two systems cohabitate? How will family allowances modify the redistributive properties of ELIE? This is also an important question. Kolm (2005)’s introductory model does not consider capital income by arguing that physical capital is itself produced by labour, so that for macrojustice (not for microjustice) capital can be neglected. His introductory model ignores the question of accumulation and economic growth and so does not give any hint on how to investigate the trade-off between growth and redistribution. On the one hand, individual saving rates rise with the level of income. Any redistributive scheme, by transferring resources from rich to poor tends to lower the aggregate rate of saving and by consequence the rate of capital accumulation. On the other hand, capital income modifies the total income distribution and has a strong influence on inequality. In most countries and certainly in France, wealth inequality is much stronger than income inequality. To answer the above questions, we shall have to formalise the ELIE scheme in various contexts. The basic ELIE formula is simple and linear. But this attractive simplicity hides in fact many non trivial properties. For instance, the degree of redistribution depends on the characteristics and shape of the initial gross income distribution. When mixed with other public finance mechanisms, the initial model is also much complexified. For these two reasons (shape of the distribution, and more complex models), we have to resort to simulation techniques. We have clearly two ways of proceeding. We could start from a microeconomic sample of French wages and simulate from this empirical distribution. The alternative solution is to simulate a given density and calibrate its parameters so as to match some inequality or redistributive indices computed from the empirical distribution of gross or net incomes in France de la Croix and Lubrano (this volume) adopt the same type of calibrated simulation as well as many authors do in the literature. This approach is much simpler to implement because it is easier to get the figures for some indexes than the figures of individual incomes collected from household surveys. Moreover, calibrated simulation is more adapted

278

M. Lubrano

when one want to study variants of the original model, because in this case, the income distribution can be simulated in different way across the variants of the initial model. A fiscal system has been characterised in the econometric literature by a number of indices that we recall in the appendix. They are, apart from the well-known Gini index, the progressivity index of Kakwani (1977), the vertical equity index of Reynolds and Smolensky (1977) and the horizontal inequity index of Atkinson (1980) and Plotnick (1981). We detail these indices in an appendix, because it is rather difficult to find a comprehensive presentation of the related literature which might not be well known to all the readers of this book. The paper is organised as follows. In Sect. 10.2, we give a mathematical presentation of the ELIE scheme. In Sect. 10.3, we simulate this model and verify its redistributive properties. We show that a small value of the fundamental redistributive parameter is needed in order to obtain the same degree of redistribution that exists in France. With Sect. 10.4, we enter the core of the debate, examining the consequences of uncertainty in the information about the distribution of talents. In Sect. 10.5, we introduce family allowances while Sect. 10.6 details a growth model à la Solow (1956) to investigate the relation between capital accumulation and income distribution. Section 10.7 concludes.

10.2 The Initial Formulation of the Model The economy we consider is composed of n individuals with a labour supply `i and a potential wage rate wi corresponding to individual capacities or productivities. Taxation bears on wi , independently of `i , the effective quantity of supplied labour. That, with the assumption that there is a total freedom for the amount of labour supplied, means that taxation has no distortive effect on labour supply. Consequently, the question of optimal taxation can be treated in a first best framework provided that wage rates or productivities are perfectly known. Kolm imagines a self-financing Pdistributive system where taxes ti D 0. As labour is the main and subsidies, both noted ti , balance with resource, taxes and subsidies are evaluated in terms of a quantity of labour. An equal quantity of labour k is taken from each individual and measured at his productivity so the tax is equal to kwi . Meantime, an equal amount k w, Q to be determined, is redistributed to each individual so that the net transfer for individual i is equal to Q  wi /: ti D k.w

(10.1)

10 The Redistributive Aspects of ELIE

279

The value of k has to be chosen by society while w Q is determined by the assumption that transfers are totaly financed by taxes. As a matter of fact, the system must verify X X Q wi / D 0: (10.2) ti D 0 ) k.nw P N This equation has only two solutions. Either k D 0 or wQ D wi =n D w. Thus w Q is equal to the mean wage w. N It plays a central role in the system and we shall call it here the pivot because it corresponds to the limit between individuals that pay and individuals that receive. Remark 10.1. w N determines the position of richer and poorer individuals. In the EU and in the UK as well, any individual with an income below a fraction of the mean income in society is considered as poor. In France, the same definition uses instead the median. When the income distribution is not symmetric, this makes a difference. So the choice of w N as a pivot is not neutral. When accounting for taxes and transfers, disposable income or net total income is given by yi D wi `i C k.w N  wi / D k w N C wi .`i  k/:

(10.3)

In this framework, taxation is based on capacities and not on actual income. So that an individual determines his labour supply independently of k. But he is implicitly constrained to work at least for k days. His income is positive N i /. Everybody receives the same basic income k w N only if `i > k.1  w=w plus a fraction `i  k of his labour productivity (wi .`i  k/ can be read as the income produced by a working time equal to `i  k). This is the Equal Labour Income Equalisation or ELIE. ELIE is thus a kind of universal basic income.2 Everybody receives k w. N Contrary to a usual basic income, the financing mechanism is already contained in the definition of ELIE as everybody also pays kwi for financing the system.

10.3 Characterising the Initial Model We start from a hypothetical population of n D 10;000 individuals. For the time being, we restrict our attention to individuals, leaving the question of household composition to Sect. 10.5. We suppose also that wages are for the 2

The interpretation of ELIE as a basic income mechanism is uncontroversial when everybody is working at least at level k .li > k/. It can lead to controversies in the opposite case .li < k/.

280

M. Lubrano

time being the sole income. These two hypotheses characterise the initial simple model of Kolm. We are here interested in exploring the consequences of wage heterogeneity on redistribution, or in other words to show that ELIE properties heavily depends on the shape of the gross income distribution. We shall thus simulate a calibrated distribution of gross income and analyse the resulting distribution of net income (after taxes and transfers) when k varies. In the original model of Kolm, the individual is free to choose the amount of labour he decides to supply and that amount can be influenced by the value of k. For the ease of the simulation, we shall assume that labour supply is exogenous (does not depend on k in any way, and we shall moreover suppose that `i D 1, which means that every individual is working full time. Consequently, the net total disposable income (10.3) is transformed into N  wi /: yi D wi C k.w

(10.4)

This assumption is a non trivial one. In the original model of Kolm, wi is the hourly wage rate which has to be multiplied by `i or order to get the gross income `i wi , and the individual is free to chose his quantity of supplied labour. Here, in our simulation model, we assume that `i exogenous and equal to 1. This means that wi now represents gross labour income. In a way the first best framework is lost. Our simulation model also supposes that wi is perfectly known. We shall relax this last assumption in Sect. 10.3. We assume that gross wages are distributed according to a Gamma distribution with  degrees of freedom and scale parameter s: f .wj; s/ D

1  1 w s : s w e  ./

We recall that the mean of this distribution is equal to =s and its variance to =s 2 . We then impose the restriction s D  in order to get a normalised mean of 1.3 We have then to calibrate this distribution by adjusting its remaining parameter . For a normalised mean, the empirical distribution of gross wages can be characterised in a simple way by its Gini coefficient. This coefficient was equal to 0.327 in France in 1998.4 To obtain the same Gini coefficient for our simulated sample, we had to choose  D 2:75. Note 3

The annual mean disposable net income in France was 28,935 euros in 2004 and the median 24,599 euros. As we have chosen a normalised mean, the horizontal axis can be read as x times the mean. It is then easy to interpret the right tail of the distribution. Actual income figures can be obtained just by multiplying the figures of the horizontal axis by the empirical mean given in this footnote. 4 Some of these data can be found for instance on the Web site of the World Institute for Development and Economic research, UNU-WIDER World Income Inequality Database, Version 2.0a, June 2005. http://www.wider.unu.edu/wiid/wiid.htm.

10 The Redistributive Aspects of ELIE

281

1.40 1.20

k k k k

1.00

= = = =

0.0 0.2 0.3 0.4

0.80 0.60 0.40 0.20 0.00

0

0.5

1

1.5

2 w

2.5

3

3.5

4

Fig. 10.1 Income distribution before and after ELIE transfers. Multiply by 28,935 euros (the annual mean disposable net income in France in 2004) to have an idea of the range of net incomes in France

however that these figures are in a way not exactly comparable because here we have individuals and earnings for the time being when real data are concerned with households and total incomes. Note also that with the Gamma distribution, very high incomes are not so well represented. In our sample, the maximum income is 6.27 times the mean. We could have chosen another usual distribution such as the logNormal or the Pareto without changing much the results. We choose the Gamma for the sake of algebraic simplicity. All moments exist, contrary to the Pareto.5 Let us now apply the ELIE scheme to this sample, taking three different values for k: 0.2, 0.3, 0.4, as suggested in Kolm (2005) so as to obtain the distribution of net incomes (after taxes and transfers) as given by (10.4). A coefficient k D 0:2 corresponds to a tax rate of one working day in a five working day week. And k D 0:4 corresponds to two working days. Figure 10.1 displays a non-parametric estimate of the different income distributions obtained when varying k. We immediately notice three facts: firstly, ELIE introduces a concentration of net incomes around the mean; secondly it raises the minimum disposable income as k grows; thirdly, if poverty is

5

A real alternative would have been to choose a richer distribution with much more parameters in order to get a better tail behaviour. We can quote the Generalised Beta II distribution which has four parameters. However, the calibration of the parameters is far from trivial. We should have used at least the deciles of the income distribution for this purpose and not just a simple Gini coefficient. The gain for the subsequent analysis is not clear.

282

M. Lubrano

Table 10.1 The original ELIE model k 0 0.2 0.3 0.4

Mean wages 1.013 1.013 1.013 1.013

Gini 0.327 0.262 0.229 0.196

Fiscal pressure 0.00 4.70 7.05 9.39

Partial fisc. pres. 0:00 7:23 10:84 14:45

Kakwani

Reynolds

Atkinson

0.000 0.454 0.454 0.454

0.000 0.065 0.098 0.131

0.00 0.00 0.00 0.00

Redistribution 0.000 0.065 0.098 0.131

defined an income lower than half the mean, a k greater than 0.4 is needed to totaly eliminate it (see Leroux and Leroux in this volume as an illustration). The decrease of the Gini coefficients given in Table 10.1 shows that ELIE strongly reduces inequality. But the chosen values for k entail Gini coefficients which are much lower than that computed on French data for net incomes and which is equal to 0.27 in 2000 and 2001. We should have chosen k D 0:17 to recover such a value for the Gini coefficient, which would corresponds to slightly less of one working day. This is a relatively low value when compared to the importance of social contributions and payroll taxes in France. However, we are studying here only redistribution; and redistribution, when properly measured, is rather weak in France. For instance, income taxes are only a small part of GDP in France. In this simple model, a rather low value for k is finally realistic. When examining various variants of the basic model, we shall see that the French Gini coefficient for net incomes is reproduced for increasing values of k as we depart from the original model. It is rather easy to verify that ELIE complies with the Pigou–Dalton requirement, which means that it does not change the mean of the distribution after taxes and transfers and that it entails Lorenz curves that get closer to the main diagonal without crossing, see Fig. 10.2. As a consequence, all concentration curves, computed on the sole basis of positive taxes are identical.6 Let us now detail various indices which characterise taxation, leaving aside transfers. Taxes, noted t ti , concern only individuals who are above the mean:  0 if wi  w N t ti D N otherwise: k.wi  w/ Those below the mean receive only transfers. The marginal tax rate is constant and equal to k for those paying taxes. However, despite this linear

For instance, as wi  G.; /, then by properties of the Gamma distribution, yi D .1k/wi Ck w N is also Gamma with yi  G.; =.1k//Ck. Its mean is equal to 1 and its variance to (1k/2 =. So ELIE does not change the mean but reduces the variance.

6

10 The Redistributive Aspects of ELIE

283

1.00

0.80

k = 0.0 k = 0.2 k = 0.3 k = 0.4 Concentration

0.60

0.40

0.20

0.00 0.00

0.20

0.40

0.60

0.80

1.00

Fig. 10.2 Lorenz and concentration curves associated with ELIE 35.00 30.00 25.00

k = 0.2 k = 0.3 k = 0.4

20.00 15.00 10.00 5.00 0.00 0.00

1.00

2.00

3.00

w

4.00

5.00

6.00

7.00

Fig. 10.3 Taxation rates implied by ELIE (in percentage)

structure, ELIE is a progressive taxation scheme. The mean taxation rate, obtained by dividing taxes by gross wages, is txi D t ti =wi D k.1 

w N /: wi

(10.5)

This rate grows with wi , but less than proportionally as illustrated on Fig. 10.3. The shape of this graph, which illustrates the bearing of taxation on our particular population, depends heavily on the shape of the distribution of income. On the contrary, the marginal rate of taxation is constant and

284

M. Lubrano

equal to k for individuals above the mean, and of course equal to zero for those below the mean. The usual index of progressivity of Kakwani (1977) is strong with 0.45 compared to its value computed on actual French data which is only 0.31. It is not a function of k because of the linear structure of ELIE (constant marginal rate). It is only a function of the shape of the distribution and of the chosen pivot for ELIE.7 A smaller index could be obtained with a lower  entailing stronger asymmetry and stronger initial inequality. The average fiscal pressure, computed as the ratio between the sum of total taxes and the sum of gross incomes, is very low, because only incomes which are above the mean are taxed. But even when the fiscal pressure is computed for only those actual paying taxes, it is not very high. The amount of inequality removed as indicated by the Reynolds index as well as the redistribution index grows with k. Horizontal equity, as defined in the appendix is perfect, taxes and transfers imply no rank permutation. The Atkinson–Plotnick horizontal inequity index is always zero. This is due to the fact that there is no effect due to family composition or to the combination of different taxes. We can conclude that the pure ELIE transfer scheme is a rather powerful mechanism of redistribution. Its implementation in France would require a rather low value for k to reproduce the same Gini coefficient for net income. However, ELIE is much more progressive. That would mean putting the burden of redistribution more on richer people than what it is now.

10.4 Uncertainty in the Information About Wages Wages and income are not in general perfectly known by the tax authority. The basic wage rate can be public, but there are bonuses and extra hours for instance that can be more difficult to know for the tax authority. These can be important for some categories. We have then tax evasion and tax avoidance. The former is illegal but some professions are famous for practicing it. The later is perfectly legal and results from an optimisation behaviour as described for instance in Stiglitz (1985). Of course in this game of tax evasion and tax avoidance, richer individuals have more opportunities than poorer ones.

7

The pivot is chosen here equal to the mean because the system is balanced in this case. Choosing for instance the median would have led to an unbalanced system. The pivot can be chosen lower than the mean if extra resources are to be collected for paying for instance civil servants and collective equipments. But in this case the original ELIE is distorted.

10 The Redistributive Aspects of ELIE

285

We are now going to modify the initial model so as to take into account this aspect which is important for practical implementation. We have to know how much ELIE looses of its nice redistributive properties when it becomes difficult to have a precise information about the individual talents. We now modify the taxation scheme so that wi would be lowered by a random effect ei : N  wi  ei /: (10.6) ti D k.w We want to implement the idea that uncertainty in the information about wages increases with its level and that the importance of dissimulation is random. For that purpose, ei is a strictly positive process and once again we chose the Gamma distribution for convenience. Its mean will depend on the level of wi and on its rank in the distribution. The maximum rate of tax avoidance (expressed as a fraction of gross income) is supposed to be 30%. More precisely, if the wi are sorted in increasing order, we shall have i ei  0:3 wi G.; /; n

(10.7)

where G.; / means the Gamma distribution. This formula implies that ei is positive with a mean equal to 0:3  wi  i=n and a variance equal to .0:3  wi  i=n/2 =. The richer people will on average manage to hide 30% of their income while the poorer people will manage to hide a negligible part of it. For the system to be balanced, the pivot has to be taken equal to the mean of the declared wages wi  ei and no longer equal the mean of the exact wages wi . Introducing a random dissimulation (Table 10.2) entails little changes when compared to the initial Table 10.1. Redistribution is of course weaker as well as the fiscal pressure on richer people. A weak horizontal inequity appears due to the random character of the modeled tax avoidance. In order to get a Gini of 0.27, we must chose a k D 0:24, a value slightly higher than in the original case where a value of k D 0:17 was sufficient. Despite the fact that we have introduced a rather strong possibility of tax avoidance and evasion, the final influence is weak. We just have to increase k in order to get the same amount of inequality. Redistribution is simply slightly less important. Taxing capacities instead of incomes seems a crucial Table 10.2 ELIE and uncertainty in the information about wages k 0.0 0.2 0.3 0.4

Mean wages 1.013 1.013 1.013 1.013

Gini 0.327 0.282 0.261 0.239

Fiscal pressure 0.00 3.16 4.73 6.31

Kakwani

Reynolds

Atkinson

Redistribution

0.000 0.437 0.437 0.437

0.000 0.044 0.066 0.089

0.0000 0.0000 0.0002 0.0003

0.000 0.044 0.066 0.088

286

M. Lubrano

hypothesis because incomes are rather easy to measure whereas capacities are not. The results of this section show that finally this hypothesis is not so crucial because when relaxing it, the main properties of the model are not so much altered. We must however keep in mind that we have supposed that labour supply is exogenous and thus does not depend on the value of k. We could also justify this conclusion on the ground of some figures concerning tax evasion. The Government of Quebec evaluates that tax evasion comes at 74% from firms in 2005 and that this percentage has steadily increased up to 79% in 2008. The main question is thus not whether to tax wi or wi `i when analysing labour supply.

10.5 Family Allowances and Family Composition ELIE is a fair and powerful redistributive scheme with nice labour incentives properties. However, incentives are not limited to the labour market, as underlined in the introduction. In this section, we want to investigate how the pure redistributive ELIE mechanism can interact with the incentive mechanism of family allowances. Family allowances exists in many European countries and are very important in France. They have been successfully designed to increase natality. With family allowances, we have a mechanism which mixes redistribution and incentives. Clearly, ELIE cannot be a substitute for family allowances, simply because it is not a targeted mechanism. If we want to experiment the feasibility of implementing ELIE in France, we have to study the properties of a mix system combining ELIE with a complementary mechanism of redistribution based on family composition and not solely on the position of gross income versus mean income.8

10.5.1 ELIE and Household Composition The basic unit i is no longer an individual but a household. We suppose that a household is composed of two adults earning the same amount each and a random number of children noted chi that can be zero.9 The relationship

8 In France, family allowances are independent of income, except for the complement familial which is proportional to wages. 9 We exclude for simplicity mono-parental households. This is certainly a limitation of our analysis. We are concerned with the whole French income distribution, of which households with a

10 The Redistributive Aspects of ELIE

287

between income and fertility is rather controversial. Becker (1981) argues for a positive relationship on the ground that children are consumer goods competing with alternative goods. But he also put forward a trade-off between quality of education and quantity, so that the relation can be negative (see also de la Croix and Doepke (2003) for an alternative point of view). How could we decide between these alternative views? We had access to survey data collected from interviews in Marseilles in 2006 at the occasion of an exercise in experimental economics concerning willingness to pay for a better air quality. The sample is made of 549 individuals, coming from distinct households for which we have the number of children, the total net income of the household and the marital status of the responder. We can thus run a regression which links the number of children to family income and marital status. Cel is a dummy variable indicating if the responder is single or not. Expressed in logarithms, this regression has a low R2 D 0:205, but clearly indicates a slightly positive relation between fertility and family income as we have: ln.1 C chi / D 0:034 C 0:061 ln.1 C wi /  0:40 cel: Œ0:16

Œ2:43

Œ9:01

Following this regression10 , we shall suppose that the number of children follows a Poisson distribution with parameter  (which represents both the mean and the variance) and we note chi  P . /. The average number of children in France is two per family. We suppose that part of the fertility decision is independent of income, say for the first child, but that the decision for an extra child depends on income. We have thus chosen  D 1 C wi =w. N The realised number of children in a family for the 10,000 households of the sample lies in the interval [0,9]. In Fig. 10.4, we compare the distribution of the number of children in a household when fertility depends on gross income to a distribution that would suppose that fertility is totaly independent of income, while keeping the same mean. Clearly, the assumption that fertility depends on income has a strong influence on the decision to have a second child, but lowers the probability to have more than three children for the given gross income distribution. As we now define the income distribution of a household and no longer the income distribution of an individual, comparisons between households become more difficult. Clearly a couple with many children has not the same needs as a couple without any children. We have to introduce an equivalence scale in order make household incomes comparable. We have chosen a rather single adult are a significant fraction. We also do not introduce the possibility of income disparities between the two parents and their consequences on taxation. 10 A non-parametric regression, not reported here, indicates also a positive and quasi linear relation.

288

M. Lubrano Uniform fertility 4500

4000

4000

3500

3500

3000

3000

Total number

Total number

Fertility increasing with w 4500

2500 2000 1500 1000

2000 1500 1000

500 0

2500

500 0

1

2 3 4 5 Children per family

6

7

0

0

1

2

3 4 5 Children per family

6

7

Fig. 10.4 Distribution of the number of children for the two scenarios

mild equivalence scale which gives a weight of 1 to the first adult, 0.7 to the second adult and 0.3 to each child. This is a mix between the Oxford equivalence scale and the OECD equivalence scale. Let us call sni the number thus obtained for household i. If we divide wi by this number, we have an income expressed in terms of family needs. The basic ELIE formula is changed into:   wi wi yi (10.8) D2 C 2k w Ns  sni sni sni P where w N s D n1 i wi =sni .

10.5.2 The Influence of Equivalence Scales Might ELIE gain some incentive properties for fertility decisions just by using an equivalence scale without introducing a specific family allowance mechanism? The answer might depend on the assumption made for fertility which determines in which type of family children are located. It is thus wise to consider two variants: one where fertility is independent of income and one where it is income dependent. Let us first suppose that the number of children per household is drawn from a Poisson distribution with  D 2. We have simulated an income distribution for 10,000 households composed of two adults earning each the same wi and having a random number of children, independent of their income. Results are displayed in Table 10.3. In column U , we have divided the total gross household income by 2 (the number of adults in the household) while in column S, total income is divided by an equivalence scale depending on the number of children. We then apply the usual ELIE scheme on the resulting scaled income, using (10.8). Considering an equivalence scale slightly increases inequality, fiscal pressure and decreases redistribution. The order of magnitude of these changes is small, but significant. It would be stronger with a different

10 The Redistributive Aspects of ELIE

289

Table 10.3 Equivalence scales and ELIE with a random fertility k D 0:0 k D 0:2 k D 0:3 k D 0:4 U S U S U S U S Gini 0.327 0.342 0.262 0.273 0.229 0.239 0.196 0.205 Fisc. pres. 0.000 0.000 4.697 4.911 7.045 7.366 9.393 9.821 Kakwani 0.000 0.000 0.454 0.448 0.454 0.448 0.454 0.448 Reynolds 0.000 0.000 0.065 0.068 0.098 0.103 0.131 0.137 In columns U, household income is divided by 2, while in columns S it is divided by an equivalence scale.

Table 10.4 Total net income per family with a uniform fertility Children 0 1 2 3 >3

k D 0:0 U S 2.062 2.062 2.025 2.025 2.025 2.025 2.018 2.018 2.003 2.003

k D 0:2 U S 2.054 1.959 2.025 1.984 2.025 2.039 2.020 2.088 2.008 2.160

k D 0:3 U S 2.051 1.907 2.025 1.963 2.025 2.046 2.020 2.123 2.010 2.239

k D 0:4 U S 2.047 1.856 2.025 1.943 2.025 2.053 2.021 2.158 2.012 2.317

equivalence scale. Column U is identical to the initial case of Sect. 10.3. With an equivalence scale, the French Gini coefficient is recovered for k D 0:19 instead of k D 0:17 when there is no equivalence scale. Let us now consider the income distribution conditional on the number of children. What is the more favourable system for large families? In Table 10.4, we give the mean income of the 10,000 households having zero, one, two, three and more children. We give the total income after ELIE taxes and transfers, the exact income that the household will have in its pocket, once transfers are operated. This figure is obtained by re-multiplying the net income obtained in (10.8) by sni . In column k D 0 of Table 10.4, there is of course no difference between U and S. For k > 0, the U columns are simply an estimate of the conditional mean of the income distribution. The computed marginal mean income is equal to 2.025. Let us concentrate on the columns marked S. When k is positive, ELIE seems to redistribute money as a function of the number of children. Households with less than two children receive less that the mean income. Household with more than two children receive more and transfers increase with the number of children. As a conclusion, an equivalence scale seems to introduce a trade-off between fertility incentives and fairness.

10.5.3 Equivalence Scales and Income Dependent Fertility If we believe in the incentive effect of family allowances, fertility should be assumed to depend on income. This assumption is going to change

290

M. Lubrano

Table 10.5 Equivalence scales and ELIE with income dependent fertility

Gini Fisc. pres. Kakwani Reynolds

k D 0:0 U S 0.327 0.307 0.000 0.000 0.000 0.000 0.000 0.000

k D 0:2 U S 0.262 0.245 4.697 4.393 0.454 0.465 0.065 0.061

k D 0:3 U S 0.229 0.215 7.045 6.589 0.454 0.465 0.098 0.092

k D 0:4 U S 0.196 0.184 9.393 8.753 0.454 0.465 0.131 0.123

Table 10.6 Equivalence scales and ELIE with income dependent fertility (continued) Children 0 1 2 3 >3

k D 0:0 U S 1.502 1.502 1.712 1.712 1.997 1.997 2.271 2.271 2.892 2.892

k D 0:2 U S 1.607 1.500 1.774 1.720 2.002 2.001 2.222 2.273 2.719 2.860

k D 0:3 U S 1.659 1.499 1.806 1.724 2.005 2.003 2.197 2.273 2.632 2.843

k D 0:4 U S 1.711 1.498 1.837 1.729 2.008 2.005 2.173 2.275 2.546 2.827

dramatically the previous results. We have redone the same simulation exercise as in Sect. 10.5.2, but this time with chi  P . / where  D 1 C wi =w N instead of  D 2. Table 10.5 shows that the use of an equivalence scale now decreases inequality as shown by the Gini indices computed on net income after ELIE transfers, instead of increasing it as in the case of random fertility. The fiscal pressure is lowered while progressivity is increased. Thus the dependence between fertility and income has a drastic influence on the redistribution properties of ELIE when an equivalence scale is used. This influence is now positive in term of fairness. A value of k D 0:13 is now sufficient to get the French Gini coefficient on net income instead of k D 0:17 in the original model without equivalence scales. Let us now examine the money which actually enter the household pocket. Table 10.6 shows first that before any redistribution (k D 0), the fertility model implies that households with children have more money than households without children, while the computed average income remains the same at 2.025. This is just a consequence of our assumption made on fertility. When k > 0, the mean income per household type is not significantly changed. Families with more than two children get more money with the equivalence scales and family with less than two children get less money as in the previous configuration. So the incentive mechanism introduced by the equivalence scales is unchanged, while the fairness properties of the model were increased.

10 The Redistributive Aspects of ELIE

291

10.5.4 ELIE and Family Allowances ELIE is a mechanism designed for macrojustice and not for incentives, because it is not a targeted mechanism. It can be viewed as a mechanism aiming at reducing the distance between average income and actual income, independently of any other factors, provided of course that all the `i are the same. Here, we have introduced a new dimension, which is family composition and its corollary equivalence scales. Depending on the modeling of fertility, the use of an equivalence scale can confer to ELIE some incentive properties, but at the expense of a lower degree of fairness measured here as progressivity and lower inequality. When fertility increases with income, the incentive effect remains and fairness is restored. Family allowances is a sensitive topic in France and also in many other European countries. A government who would like to implement the redistributive ELIE mechanism cannot politically suppress family allowances, even if we can truly suspect this mechanism to distort the nice properties of ELIE. Let us try to conceive a complementary redistribution scheme based solely on family composition and independent of income. We need extra resources in order to finance explicit family allowances, which means distributing an amount of money only on the basis of family composition, independently of income. The redistributive ELIE system balances because the pivot is equal to the mean w N s . If we choose a pivot lower than the mean, say w Nq < w N s , we get the needed extra amount to redistribute: Tr D 2

n X iD1

k.w Nq 

wi / > 0: sni

So that net income is now defined as   wi wi yi D2 C 2k w Nq  C Tr chi =nch; sni sni sni

(10.9)

(10.10)

where nch is the total number of children in the complete sample. How can we calibrate w N q ? In the ELIE scheme, the Kakwani progressivity index is independent of k, but varies with the size of the pivot. In the previous sections, the Kakwani index was much higher than its French value of 0.31. We can decide to chose w N q so as to match the value of 0.31 for the Kakwani index. Compared to Table 10.5, inequality has slightly increased (Table 10.7), but mainly progressivity was calibrated to be much lower. This is translated in a huge increase in the fiscal pressure because now much more households are taxed. The Reynolds index is smaller indicating that taxes are less successful in reducing inequality.

292

M. Lubrano

Table 10.7 Family allowances and ELIE with income dependent fertility Gini FiscPres Kakwani Reynolds

k D 0:0 0.306 0.000 – 0.000

k D 0:2 0.249 8.154 0.310 0.059

k D 0:3 0.222 12.231 0.310 0.089

k D 0:4 0.197 16.308 0.310 0.119

Table 10.8 Equivalence scales and ELIE with income dependent fertility (continued) Children 0 1 2 3 >3

k D 0:0

k D 0:2

k D 0:3

k D 0:4

1.488 1.731 1.955 2.344 2.857

1.387 1.676 1.967 2.409 3.090

1.337 1.649 1.974 2.441 3.206

1.286 1.621 1.980 2.474 3.322

Table 10.8 show that large families get much more money. But we have seen that this extra amount of money is not obtained in a fair way. If ELIE had to be applied, explicit family allowances should be avoided. Equivalence scales are enough.

10.6 Introducing Capital Stock Kolm (2005) claims that since capital is by definition produced, the other primary resources finally account for about 97.5% for labour and 2.5% for the non-human natural resources. So that the problem of overall distribution in macrojustice is the allocation of the rights in the value of productive capacities, i.e. labour capacities.The idea is attractive from a normative point of view, but it is not adapted if the objective is to confront the ELIE scheme to an empirical reality. Kolm’s point of view implicitly assumes that capital and credit markets are efficient so that there is no problem of capital reallocation in the long term. However, the above citation alluded to the simple model of Kolm and does not precludes further refinements of this simple model.

10 The Redistributive Aspects of ELIE

293

40.00 38.00 36.00

U.S.A France Germany

34.00 32.00 30.00 28.00

1980

1982

1984 1986 1988 1990 Source: Piketty (2004)

1992

1994

Fig. 10.5 Capital share in some OECD countries

10.6.1 Stylised Facts We observe a fundamental inequality in the sharing of the fruits of capital and of labour which sets forward the question of the efficiency of any redistributive policy. At the macro level, the share of capital in total value added can remain relatively constant despite of large wage rises as in the US, or experience a marked increase as in Europe and especially in France as shown in Fig. 10.5 and as noted in Piketty (2004). The modification of the capital share in France implied a reallocation of more than 10% of total GNP on a rather short period of 10 years. This is more than what any fiscal redistributive policy could have operated. At the micro level, we can notice that capital allocation remains extraordinary stable in the long term between households. This is due to several reasons. There is first the question of initial conditions in the dynamic process of wealth transmission: death duty does not manage to reallocate efficiently ownership because rates of taxation are too low for that purpose. Secondly, the rate of saving is different between poor and rich individuals. A high rate of saving is necessary for accumulation. Moreover, as documented in Direr and Weitzenblum (2006), mostly rich individuals have access to really profitable investments. Thirdly, credit market is rationed, banks avoid financing risky projects initiated by lower income holders as exemplified for instance in Piketty (1994). Finally, wages and capital gains are not taxed in the same way. In France, for the same level of income, wage earners are more taxed than holders of capital gains. All these features put together contribute to inequality in the income and wealth

294

M. Lubrano

distributions. As a matter of fact, in every country, there is a stronger inequality in wealth distribution, than in income distribution. For instance, the Gini index for net income distribution was reported to be 0.27 in France. It rises up to 0.65 for wealth inequality in 2004 (see for instance Cordier et al. 2006). Regulating the distribution of income between capital and labour can become thus one of the major goals of income distribution policies. This justifies introducing capital accumulation in the ELIE framework and see what is the performance of ELIE for regulating the inequality in income distribution generated by capital accumulation.

10.6.2 General Features The model we shall use for capital accumulation11 is inspired from Solow (1956) where the rate of saving determines the rate of accumulation. In the original growth model of Solow, population grows at a fixed rate while there is no depreciation for capital. Here we want to keep the labour force constant with the same wage distribution for ease of comparison with our previous results. We shall thus suppose that the rate of growth of the labour supply is zero while capital depreciates at a fixed rate. The fundamental properties of the model should not be affected. We consider that there is one active generation with an economic activity lasting for 30 periods. In this framework, we can leave aside the question of inheritance taxes (death duty). We just have to precise initial conditions, saving behaviour and transfers. The first key assumption concerns the rate of saving. As we have an heterogenous population (contrary to the model of Solow), we shall suppose that the rate of saving varies across the population. Our key assumption is that rich people save more than less rich people. More precisely, the rate of saving will be a positive function of income. This assumption summarises the stylised facts detailed above and is central to obtaining a distribution of wealth more concentrated than the distribution of income. The second key assumption follows Direr and Weitzenblum (2006) where only rich people have access to really profitable investments. Capital profitability will also be dependent on income. With this model, we want to compare two types of taxation-redistribution in the presence of capital accumulation. In a first scenario, we introduce ELIE only for labour income. There is a flat tax on capital income, very much like what happens in France. The fruit of these taxes is redistributed 11

A complete model would have to consider overlapping generations such as in Direr and Weitzenblum (2006). We would like, at least in a first attempt, to leave this question aside.

10 The Redistributive Aspects of ELIE

295

uniformly among the individuals as would be a basic income. This scenario aims at reproducing more or less the French situation or a minimal adaptation of ELIE to France. The second scenario is a kind of full application of ELIE when there is capital accumulation. Capital income is added to labour income and serves a basic to the ELIE formula. Capital and labour incomes are equally treated.

10.6.3 A Dynamic Model of Capital Accumulation Initial conditions are crucial for any non-stationary processes, because they have a direct influence on their trajectories. We have calibrated the wage distribution on the French Gini coefficient and normalised it with a mean equal to one. We have to do the same for the initial wealth distribution. We suppose that the initial capital stock, s0;i , is a function of labour income and that it is not equally distributed in such a way so as to reproduce the French Gini coefficient computed for wealth. Our proposed formula is s0;i D ˛0

wi wi 1I.wi > w/ N C .1  ˛0 / 1I.wi < w/; N r r

(10.11)

where r is the mean rate of return on capital and 1I is the indicator function. With ˛0 D 0:75, we get a Gini coefficient of 0.651 for the initial capital stock. These initial conditions are interpreted as a bequeath of the previous generation. With an initial mean rate of return of 3%, the income generated by the initial capital stock will represent roughly 39% of total income at t D 0. The rate of saving determines the rate of accumulation. Capital stock available at time t, si;t comes from capital at time t  1, si;t 1 , diminished of a fixed depreciation rate ı, but augmented of savings. It is commonly admitted that the saving rate increases with income, but less than proportionally (see e.g. Dynan et al. 2004). We propose modeling the saving rate as p (10.12)  i D  w i ; where  is a scaling parameter. The mean saving rate is equal to 15% in 2005 in France. This rate is reproduced for  D 0:156. For individuals with a labour income lower than w, N this rate is on average equal to 12%; for individuals with a labour income greater than w, N this rate is equal to 19% on average, given our wage distribution. Following Direr and Weitzenblum (2006), profitability of capital must be a function of income to illustrate the fact that only rich individuals have access to really profitable investments. We shall suppose the rate of return

296

M. Lubrano

on capital, ri;t , is a random variable with a mean which is a linear function of the square root of wi and a fixed variance: p ri;t  N.r wi ; r2 / (10.13) where N.:; :/ is a Gaussian random variable. At each period, we will have a new draw for the returns. We have previously chosen r D 0:03 and now we chose r D 0:01. This formulation does not exclude temporary negative returns. Capital income at time t, sri;t , is given by: sri;t D si;t ri;t ;

(10.14)

while capital stock is defined by si;t D .1  ı/si;t 1 C i yi;t :

(10.15)

The growth rate of the capital stock depends on the balance between the proportion of depreciation ı and the share of net income devoted to savings. As savings depends on net income, the growth rate of capital is a function of k, the redistribution parameter. There is a clear trade-off between growth and redistribution. Net income is given by the combination of labour income, capital income and the implementation of a given redistribution scheme. In our first scenario, ELIE is applied only to labour income, while capital income is taxed at a fixed rate  of 20% (a realistic value in the French context). The product of capital taxation is used to finance a basic income. We have: N C .1  /sri;t C alt =n; yi;t D wi  k.wi  w/

(10.16)

where alt is the fruit of taxation on capital income and is given by X sri;t : (10.17) alt D  i

We can now calibrate ı. As there is a trade-off between capital growth and redistribution, we chose ı so as to maintain over the 30 periods a constant ratio between capital income and total income for a given value of k. For k D 0:3, the corresponding value of ı is found to be 0.0175. In our second scenario, capital and labour incomes are treated in the same way (which they are not in reality). We add capital and labour income and then apply the ELIE formula to total income; yi;t D wi C sri;t  k.w N cap  wi  sri;t /; (10.18) P with w N cap D n1 .wi Csri;t /. Simulating these two variants for our 10,000 individuals will give us clues on the properties of ELIE in the presence of capital accumulation.

10 The Redistributive Aspects of ELIE

297

10.6.4 A Trade-Off Between Growth and Inequality We first analyse the trade-off between growth and inequality in our simple calibrated growth model. Table 10.9 clearly illustrates the trade-off resulting from ELIE. Increasing k deceases the net total income but also inequality both for net income, capital income and capital stock. There is still more inequality on capital income than on labour income. The two systems of taxation-redistribution, that is to say, on one side ELIE applied only to labour income but coupled with a flat tax on capital income, and a basic income and on the other side the full ELIE applied to the sum of labour and capital income, are strictly equivalent for k D 0:2. But for greater values of k, the full ELIE is much more powerful to reduce inequality. For k D 0:2, the share of capital income increases slightly in both cases, compared to the initial situation. For k D 0:3, it diminished with the full ELIE while it remains slightly constant in the first scenario. The redistributive power of ELIE is made even more apparent if we examine Fig. 10.6 which displays for both scenarios the income distribution. With the full ELIE, the right tail of the income distribution is much flatter. The group of very rich individuals is strongly and negatively affected by ELIE taxes. With a progressive taxation, the extreme right tail disappears. Moreover, the importance of this group collapses for earners beyond five times the mean, while individual around twice the mean beneficiate largely from the full ELIE scheme. The change of the initial income distribution has its roots in the modification of capital distribution as illustrated in Fig. 10.7. Initial conditions are largely modified after thirty periods, because everybody can save and Table 10.9 The trade-off between growth and inequality k

Mean net total inc.

Mean capital share

0.000

1.659

0.389

Mean capital Gini net stock income Initial situation 16.540 0.417

0.00 0.20 0.30 0.40

1.670 1.660 1.655 1.650

0.393 0.390 0.388 0.386

0.000 0.200 0.300 0.400

1.684 1.660 1.648 1.636

0.399 0.390 0.385 0.381

Gini capital income

Gini capital stock

0.706

0.651

ELIE on labour income only 17.004 0.410 16.842 0.367 16.761 0.346 16.680 0.324

0.680 0.673 0.670 0.666

0.601 0.590 0.584 0.578

ELIE on total income 17.240 0.470 16.842 0.367 16.650 0.318 16.462 0.269

0.692 0.673 0.664 0.654

0.619 0.590 0.575 0.561

298

M. Lubrano

1.00

1.00

Initial k = 0.0 k = 0.2 k = 0.3 k = 0.4

0.80

Initial k = 0.0 k = 0.2 k = 0.3 k = 0.4

0.90 0.80 0.70

0.60

0.60 0.50

0.40

0.40 0.30

0.20

0.20 0.10

0.00

0

1

2

3

4

5

6

7

8

0.00

0

1

2

3

Partial ELIE

4

5

6

7

8

Full ELIE

Fig. 10.6 Income distribution with capital accumulation 0.07

0.07

Initial k = 0.0 k = 0.2 k = 0.3 k = 0.4

0.06 0.05

0.05

0.04

0.04

0.03

0.03

0.02

0.02

0.01

0.01

0.00

0

20

40

60

80

Partial ELIE

Initial k = 0.0 k = 0.2 k = 0.3 k = 0.4

0.06

100

0.00

0

20

40

60

80

100

Full ELIE

Fig. 10.7 Capital distribution

accumulate in our model, even if all are not equal in their opportunities and saving rates. But the capital distribution is still bimodal. The full ELIE has a larger impact on holders of capital stock greater than 60 times the mean wage. The full ELIE is able to greatly diminish their importance. This corroborates the remarks made in Piketty (2001) on the disappearance of large fortunes at the beginning of the twentieth century with the introduction of a progressive taxation on income. However, the redistributive scheme of ELIE should be combined with a careful study for tuning death duties. Initial conditions are of a considerable importance, even after 30 periods, to determine the shape of the capital distribution.

10.6.5 Analysing Taxation The previous picture can be completed by analysing some taxation indices as reported in Table 10.10. For k D 0:20, we have seen that the macroeconomic

10 The Redistributive Aspects of ELIE

299

Table 10.10 The trade-off between growth and inequality k

Mean net total inc.

0.00 0.20 0.30 0.40

1.670 1.660 1.655 1.650

0.000 0.200 0.300 0.400

1.684 1.660 1.648 1.636

Gini Fiscal Kakwani total inc. pressure ELIE on labour income only 0.410 7.868 0.214 0.367 10.663 0.238 0.346 12.072 0.247 0.324 13.490 0.254 ELIE on total income 0.470 0.000 – 0.367 6.945 0.350 0.318 10.298 0.352 0.269 13.574 0.355

Reynolds

0.053 0.092 0.111 0.131 0.000 0.092 0.136 0.179

situation is the same for both scenarios. However, we see now that the full ELIE system implies a lower fiscal pressure and a stronger progressivity of taxation in order to achieve at the same redistribution. Fiscal pressure starts to become greater with the full ELIE only for k D 0:40. Finally, we note that with capital accumulation, our calibrated model needs a k D 0:40 together with the full ELIE in order to reach the French value of 0.270 for the Gini coefficient on net income. We can thus conclude that introducing capital accumulation leads to a very strong modification of the income distribution and of the ELIE principle of equity based on the sole labour factor. Clearly, ELIE has to be completed by a taxation-redistribution of bequeath.

10.7 Conclusion The ELIE scheme is a principle of taxation and transfers that has many attractive properties in terms of macrojustice and equity. In this paper, we have tried to address the question of its implementation. We examined that nice principle of taxation-redistribution and tried to see how its properties would be transformed when it has to be applied in a specific context with its own characteristics where it would be of course difficult to totaly change a complete system. Taxation systems cannot be changed completely in 1 year. Reformation is always discussed at the parliament which means that some categories will pay or receive more or less than what the original idea would imply. In France family allowances are historically very important and certainly been successful in maintaining a decent rate of natality, even if we have seen that their impact in term of equity is dubious. Since capital is much more mobile than labour, the impact of capital taxation is always discussed in terms of incentives and international fiscal competition more than in terms of equity.

300

M. Lubrano

We have tried to propose options which make the basic ELIE model more realistic for the French case. We have found that uncertainty on wages does not modify fundamentally the properties of the model and this is a nice result. However it might lower greatly its efficiency. Family allowances represents a greater challenge. The incentive properties of family allowances can be recovered simply by introducing equivalence scales. Their impact is very sensitive to the relation between income and fertility. If fertility is increasing with income, an equivalence scale can increase the redistributive efficiency of the system in the sense that the same degree of inequality can be obtained with a lower k. But we can have exactly the reverse property with a different model for fertility. The incentive effect can be greatly increased if we introduce a specific family allowance mechanism, but at cost of a huge increase in inequality. Finally, capital accumulation has the strongest distortion impact and changes many of the ELIE properties. In the presence of capital accumulation, ELIE should be completed by a mechanism of taxation-redistribution on bequeath which is outside the scope of this paper. Our model was calibrated in such a way so as to reproduce most of the inequality and taxation indexes obtained on French data. Various values for k were considered. They are directly linked to the desired degree of aversion for inequality in society. In the pure ELIE model, k had to be rather low to fit French data. But in this basic model, the income distribution does not reproduce correctly the importance of high incomes. A model with capital accumulation is necessary to obtain a distribution representing correctly high incomes. In this case k has to be rather high in order to reproduce the value of the various taxation indexes. The determination of k thus heavily depends on the model considered. The more we depart from the original model, the greater k has to be. Further investigation is needed so as to estimate the desired degree of aversion for inequality in society. Many papers have been written on this topic. Building on the seminal work of Kolm (1969) and Atkinson (1970), we can quote Gevers et al. (1979) followed by Van Praag et al. (1980) or Amiel et al. (1999), all using experiments and questionnaires. This is planned for future work.

Appendix About the Instruments Most of the indices used in the literature devoted to taxation and transfer analysis are based on the geometry of the Gini index. Let us start from a series of gross income X with density f .x/ and distribution F .q/. The

10 The Redistributive Aspects of ELIE

301

Lorenz curve is defined as 1 p D F .q/ ) LX .p/ D xN

Z

q 0

xf .x/ dx:

Let us now suppose that the observations have been ordered in ascending order. The Lorenz curve can be estimated using LX .p D i=n/ D

i 1 X xj : n xN j D1

The Lorenz curve indicates the proportion LX .p/ of total income that is hold by the proportion p of the total population. Perfect equity corresponds to the main diagonal LX .p/ D p. The Gini index G is defined as the surface between that diagonal and the Lorenz curve: Z 1 Z 1 Œp  LX .p/ dp D 1  2 LX .p/ dp GD2 0

0

which can be approximated by, using ascending order GD1C

N 1 X 1 C .N  i C 1/xi : N xN N 2 iD1

Let us now define a tax function t.x/ and the amount of taxes ti D t.xi /. The mean tax rate is obtained as the ratio of total taxes by total gross income: Z tN 1 1 t.x/f .x/ dx D : gD xN 0 xN It will be used extensively in the next definitions. The concentration curve CT .p/ indicates the proportion of total taxes that the population holding the proportion p of the total income pays: Z q 1 p D F .q/ ) CT .p/ D t.x/f .x/ dx: g xN 0 Let us now order the observations ti according to the ascending order of xi . The concentration curve can be estimated by i i 1 X 1 X tj D tj : CT .p D i=n/ D n g xN n tN j D0

j D0

A concentration index CI is defined in the same way as the Gini index as a measure of the surface between the concentration curve and the diagonal

302

M. Lubrano

Z CI D 1  2

1 0

CT .p/ dp:

Let us go from gross income xi to net income yi by subtracting taxes ti . The concentration curve of net incomes CY .p/ gives the proportion of total net income that is held by the population holding the proportion p of the total gross income: Z q 1 p D F .q/ ) CY .p/ D Œx  t.x/ f .x/ dx; .1  g/xN 0 which can be estimated by CY .p D i=n/ D

i i X 1 1 X .xj  tj / D yj : n .1  g/ xN n yN j D0

j D0

There exists a deterministic relationship between the three curves: LX .p/ D gCT .p/ C .1  g/CY .p/: The various indexes that characterise a taxation and redistribution scheme are defined in the same way as the Gini index. They correspond to surfaces between two Lorenz and/or concentration curves. The progressivity index of Kakwani (1977) measures the surface between the Lorenz curve associated to gross incomes and the concentration curve of taxes. It characterises progressivity as a departure from proportionality: Z 0 ŒLX .p/  CT .p/ dp D CIT  GX : KD2 1

This index is positive if a tax is progressive, equal to zero if the tax is proportional and negative if the tax is regressive. The negative bound is .1 C GX / and the positive bound .1  GX / where GX is the Gini index for gross income xi . The index of vertical equity of Reynolds and Smolensky (1977) measures the reduction in inequality brought in by taxes as the surface between the concentration curve of net income and the Lorenz curve of gross income Z 0 RS D 2 ŒCY .p/  LX .p/ dp D GX  CIY : 1

It equivalently measures the reduction operated by taxation on the Gini index of gross income. Because of the deterministic relationship existing between Lorenz and concentrations curves, these two indices are related by

10 The Redistributive Aspects of ELIE

RS D

303

g K: 1g

The index of horizontal inequity of Atkinson (1980) and Plotnick (1981) is concerned by the unequal treatment of equals or re-ranking operated by taxation. It measures the surface between the concentration curve and the Lorenz curve of net income: Z 0 AP D 2 ŒCY .p/  LY .p/ dp D GY  CIY : 1

If taxes do not induce a rank permutation, then the concentration and Lorenz curves of net income are identical and this index is zero. Or the Gini and concentration index of net income are identical. The redistributive index measures the surface between the Lorenz curve of net and gross incomes: Z 0 ŒLY .p/  LX .p/ dp D GX  GY : IR D 2 1

This index is at value between 0 and 1. It increases with the degree of redistribution. It is simple to discover that this index is equal to the difference between the index of vertical equity and the index of horizontal inequity: IR D RS  AP. For more details on the topic, see Lambert (2001), Duclos and Tabi (1999) or Essama-Nssah (2000).

References Amiel, Y., Creedy, J., & Hurn, S. (1999). Measuring attitudes toward inequality. Scandinavian Journal of Economics, 101(1), 83–96. Atkinson, A. B. (1970). On the measurement of inequality. Journal of Economic Theory, 3, 244–263. Atkinson, A. B. (1980). Horizontal equity and the distribution of the tax burden. In H. Aaron, & M. Boskin (Eds.), The economics of taxation (Vol. 318, Chapter 1, pp. 3–18). Washington DC: Brookings Institution. Becker, G. (1981). A treatise on the family. Cambridge: Harvard University Press. Cordier, M., Houdré, C., & Rougerie, C. (2006). Les inégalités de patrimoine des ménages entre 1992 et 2004. Données sociales, INSEE, pp. 455–464. de la Croix, D., & Doepke, M. (2003). Inequality and growth: why differential fertility matters. American Economic Review, 93(4), 1091–1113.

304

M. Lubrano

Direr, A., & Weitzenblum, T. (2006). Modéliser la distribution des richesses en France. Annales d’Economie et de Statistique, 82, 151–186. Duclos, J.-Y., & Tabi, M. (1999). Inégalité et redistribution du revenu avec une application au Canada. L’Actualité Économique, 75(1-2-3), 95–122. Dynan, K. E., Skinner, J., & Zeldes, S. P. (2004). Do the rich save more? Journal of Political Economy, 112(2), 397–444. Essama-Nssah, B. (2000). Inégalité, pauvreté et bien-être social. Louvain: De Boeck Université. Gevers, L., Glejser, H., & Rouyer, J. (1979). Professed inequality aversion and its error component. The Scandinavian Journal of Economics, 81(2), 238–243. Kakwani, N. C. (1977). Measurement of tax progressivity: an international comparison. Economic Journal, 87, 71–80. Kolm, S.-C. (1969). The optimal production of social justice. In J. Margolis, & H. Guitton (Eds.), Public economics: An analysis of public production and consumption and their relations to the private sectors (pp. 145–200). London: Macmillan. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge (UK): Cambridge University Press. Lambert, P. J. (2001). The distribution and redistribution of income (3rd ed.). Manchester and New York: Manchester University Press. Piketty, T. (1994). Inégalités et redistribution. Revue d’Economie Politique, 104, 769–800. Piketty, T. (2001). Les hauts revenus en France au 20e siècle: Inégalités et redistribution, 1901–1998. Paris: Grasset. Piketty, T. (2004). L’Economie des inégalités. Paris: La Découverte. Plotnick, R. (1981). A measure of horizontal inequity. The Review of Economics and Statistics, 62(2), 283–288. Reynolds, M., & Smolensky, E. (1977). Public expenditure, taxes and the distribution of income. New York: Academic Press. Solow, R. M. (1956). A contribution to the theory of economic growth. The Quarterly Journal of Economics, 70(1), 65–94. Stiglitz, J. E. (1985). The general theory of tax avoidance. National Tax Journal, 38(3), 325–337. Van Praag, B., Goedhart, T., & Kapteyn, A. (1980). The poverty line–a pilot survey in Europe. The Review of Economics and Statistics, 62(3), 461–465.

Chapter 11

The Trade-off Between Growth and Redistribution: ELIE in an Overlapping Generations Model David de la Croix and Michel Lubrano

Abstract The ELIE scheme of Kolm taxes labour capacities instead of labour income in order to circumvent the distortive effect of taxation on labour supply. Still, Kolm does not study the impact of ELIE on human capital formation and investment. In this paper, we build an overlapping generations (OLG) model with heterogenous agents and endogenous growth driven by investment in human capital. We study the effect of ELIE on education investment and other aggregate economic variables. Calibrating the model to French data, we highlight a trade-off between growth and redistribution. With a perfect credit market, ELIE is successful in reducing inequalities and poverty, but it is at the expense of lower investment in education and slower growth. In an economy with an imperfect credit market where individuals cannot borrow to educate, the trade-off between growth and redistribution is not overturned but is less severe. However, it is possible to overturn completely that trade-off simply by changing the base of taxation for the young generation which is equivalent to subsidising education.

D. de la Croix (B) IRES and CORE, Université catholique de Louvain e-mail: [email protected] M. Lubrano GREQAM and CNRS, Marseille e-mail: [email protected] David de la Croix acknowledges the financial support of the Belgian French speaking community (Grant ARC 09/14-018 “Sustainability”) and the Belgian Federal Government (Grant PAI P6/07 “Economic Policy and Finance in the Global Economy: Equilibrium Analysis and Social Evaluation”). Michel Lubrano acknowledges the financial support of the ANR research project NT05-3-41515-STAHN-Hubert: Economie de la Connaissance. We thank Claude Gamel, Cecilia García-Peñalosa, Serge Christophe Kolm and Carine Nourry for useful comments. Remaining errors are solely ours.

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_11, 

305

306

D. de la Croix and M. Lubrano

11.1 Introduction Equal Labour Income Equalisation or ELIE is a form of taxation and transfer imagined by Kolm (2005) in his theory of macrojustice. In this scheme, labour is the only source of income. ELIE proposes to tax labour capacities, and not labour supply in order to circumvent the distortive effect of taxation on labour supply.1 Kolm does not consider capital income by arguing that physical capital is itself produced by labour, so that for macrojustice (not for microjustice) capital can be neglected. Lubrano (this volume) builds a simple neo-classical growth model à la Solow (1956) and analyses how the ELIE scheme reduces disposable savings and slows down physical capital accumulation. In the most simple model considered in Kolm (2005), capacities are considered as given. So that the individual has no action on them and labour is reduced to labour time. In more complex versions of ELIE, see e.g. Kolm (2005, Chap. 8), labour is considered to be multidimensional, which means that individuals can devote a part of their labour time to improve their capacities, initial education being an outstanding example. In this case, education is simply added to actual working time, Kolm (2005, p. 142). However, no formal model of education decisions is given, because “in most cultures educational choices are little affected by taxes on earnings to be paid decades later in unknown situations” (Kolm 2005, p. 180). In this paper, we consider on the contrary that when individuals can modify their capacities and their gross wages by investing in human capital, taxing capacities is likely to have an effect on their incentives to educate. The aim of this paper is thus to evaluate what are the effects of taxation and redistribution on human capital accumulation. We introduce an ELIE-like scheme in an overlapping generations (OLG) model where heterogenous agents choose how much to invest in education when young. The initial model comes from Azariadis and de la Croix (2006), itself based on an extension of Azariadis and Drazen (1990) to a world with heterogeneous agents. This model has two important characteristics. First, both growth and the income distribution are endogenous. We can therefore study how these two variables co-move facing changes in the environment. Second, individuals differ by their abilities, but not by their inherited wealth. Taxing labour income will affect incentives to educate and will redistribute resources from the rich to the poor, and not the opposite as it is the case when agents differ in their initial endowments of physical capital, as in García Peñalosa and Turnovsky (2007). 1

See Cardia et al. (2003) for an estimation of this distortion based on a general equilibrium model.

11 The Trade-off Between Growth and Redistribution

307

We will focus on long-run income growth and inequality without paying attention to possible losers and gainers along the transition path. For that purpose, we will characterise the balanced growth path of our overlapping generations model and analyse the effect of introducing an ELIE-like scheme on growth and inequality. We will first consider the effect of an ELIE-like scheme on inequality, education and growth in a world where the credit market is perfect, i.e. where all individuals can freely borrow for their educational investment. We shall then consider a situation where human capital cannot be collateralised and where individuals cannot borrow. In this case, we expect a taxation-redistribution scheme to be less harmful for growth by redistributing resources towards those who are constrained in their education decision. Indeed, as stressed by Bénabou (2005), the trade-off between growth and redistribution generated by a taxation scheme can depend on the availability of a credit market. A scenario of “growth-enhancing redistribution” may seem relevant when the capital market is less well-functioning or even unavailable. This scenario might also be relevant if the young and old generations are taxed differently in order to subsidise education. The paper is organised as follows. The model is presented in Sect. 11.2. The extension to the case of an imperfect credit market is proposed in Sect. 11.3. Section 11.4 is devoted to the calibration and simulation of the long-run equilibrium. The trade-off between growth and redistribution is analysed in Sect. 11.5. The trade-off between growth and redistribution is overturned by a different implementation in Sect. 11.6. The last section concludes.

11.2 An Overlapping Generations Model The model is set up in discrete time, with time going from 0 to 1. A unit of time measures the length of a generation. At each point in time, two generations of workers are alive. Junior workers (aged 18–39 to fix ideas), and senior workers (aged 40–62). Assuming that individuals are born at age 18 and die at age 62, we abstract from childhood and old-age.2 The number of individuals born at time t is Nt . At one date t, total population includes Nt young workers and Nt 1 old workers. Young workers chose either to work 2

The complete model should include four generations (childhood, junior worker, senior workers and old age). But as our main concern is the trade-off between growth and intra-generational redistribution, we neglect for simplicity the childhood and old-age generations. Moreover, a four generation model would not be feasible because generations are constrained to be of the same length.

308

D. de la Croix and M. Lubrano

directly or to devote one part of their time to specialise with an advanced program of education. Each young individual i born at t is endowed with one unit of time. ıit is the proportion of this time devoted to further education while 1ıit the proportion of time devoted to work and earn money. The trade-off they face is therefore between studying to improve their human capital for getting more money when old and working for getting more money right now. Workers benefit from their further education during their second period of life, when reaching seniority (i.e. age 40). This modeling choice reflects the idea that the skill premium becomes much more important after 40. For the old generation, there is a retirement age which is determined by a parameter  that can be supposed to be the same for everybody and to be constant over time. This exogenous and policy parameter is a device to modulate the length of the activity of the old generation. For instance if the length of a generation is 22 years, the period of activity of the old generation will be   22. If people retire at 59, the value of  can be determined as follows: 1   D .62  60/=22 !  D 10=11. The use of this parameter is simply to allow for early retirement, knowing that in France the official age of retirement is 60, while the mean retirement age is most of the time around 59 and that workers retire at 65 in some professions. Heterogeneity is introduced by supposing that each agent i born at time t has a different ability. His ability vector denoted it D .itY ; itO / is drawn from a distribution defined over R2C (a bivariate lognormal for instance) with mean .1; 1/0 and a variance-covariance matrix ˙ . itY is related to physical strength and is attached to the working ability when young. For the same individual, itO incorporates elements related to his intellectual capacities (IQ) and thus to his ability to learn and to make education profitable when he will be old in t C 1. We have two generations living at the same time. The old generation, born at t  1 is characterised by a vector t 1 drawn at t  1 while the younger generation is characterised by a second bivariate vector t , drawn at t.

11.2.1 Human Capital and Growth At each date t the old generation has an average human capital stock hNt . Along a balanced growth path, it is growing over time at rate g. Average human capital determines a cultural environment from which everyone draws benefits. The stock of human capital of a member of the young generation (say at age 18) results from the combination of his environment hN t and of his personal characteristics itY hYit D itY hNt :

(11.1)

11 The Trade-off Between Growth and Redistribution

309

There are stronger and weaker individuals and that makes differences in the wage they can earn. The wage rate per efficient unit of human capital is denoted w and it is time independent along the balanced growth path. The difference in earnings across young individuals results from differences in abilities itY and differences in the number of hours worked 1  ıit . However, because the mean of itY was supposed to be unity, the average wage rate in the young generation is equal to w. A young individual has to decide which proportion 1  ıit of his time he will work in order to earn .1  ıit /witY hNt

(11.2)

and which proportion ıit of his time he will devote for advanced studies in order to increase his future human capital stock at time t C 1 when he will be old: O hO .ıit /hN t : (11.3) itC1 D it Coupled with itO , .ıit / tells how much education can be transformed into true future capacities. It will determine the expected earning in t C 1: witO .ıit /hN t :

(11.4)

When old, the individual will earn money by working the first fraction  of his second period of life. He will rely on his savings for the last 1   part of his life.3 The function is assumed to be increasing, concave and satisfies boundary conditions lim

ı!0

0

.ı/ D C1; lim

ı!1

0

.ı/ D 0;

(11.5)

implying that it is always optimal to spend a strictly positive time span for building human capital. hN t is the average human capital of the old generation at time t (hence of individuals born in t  1), while the average capital stock for the next generation is Nt 1 X N hO (11.6) ht C1 D itC1 : Nt iD1

Along a balanced growth path, the growth factor, denoted by G, is constant. Using (11.3), we can now characterise G as Nt 1 X hN t C1 D itO .ıit /: GD Nt hN t

(11.7)

iD1

The parameter  is here only to analyse the sensitivity of the model to the retirement age. The actual period of retirement beyond 62 is not included in the model. 3

310

D. de la Croix and M. Lubrano

In this model human capital is the sole engine of growth. Growth depends on two terms. The first term involves the heterogenous abilities of old workers itO . Thus growth depends in a way on the results of a genetic lottery. The second term is the decision to invest in human capital when young ıit . This second factor results from the profitability of education. It can be influenced by economic policies, taxation and redistribution.

11.2.2 Income and Education Decisions Individuals have to take decisions concerning investment in education (ıit ), consumption citY and saving sit . Young individuals at present time t consume O citY and save sit . At future time t C 1, they will consume citC1 and will not save any longer; they are not altruistic and it is therefore optimal for them to consume all their wealth when old. Their preferences are represented by a utility function which depends on present and future consumption only. It has the form of: O : (11.8) ln citY C ˇ ln citC1 The old generation does not take into account the fact that the young generation benefits from its human capital. The intergenerational transmission channel operates with (11.1) and is totaly involuntary. The utility function is simple and short-sighted. Earnings when young are devoted to consumption and saving w.1  ıit /hYit D citY C sit :

(11.9)

O that will be When getting old, the young generation will consume citC1 financed partially by future labour income (in a proportion equal to  < 1 because of early retirement age), and partially by past savings, O citC1 D whO itC1 C Rsit

(11.10)

where w is the wage per unit of human capital and R is the return on capital between the two periods; both are constant along a balanced growth path. We now determine the life cycle total income ˝it for the young generation: ˝it D Œ.1  ıit /witY C 

w O  .ıit /hN t : R it

(11.11)

Since preferences do not depend on leisure, and as long as the capital market is perfect, the individual decision problem is separable. We first maximise life-cycle income to determine optimal education. Second, we maximise utility given income to determine optimal saving and optimal consumption.

11 The Trade-off Between Growth and Redistribution

311

The optimal ıit maximising the life cycle income is given by: 0

.ıit / D

itY itO

R:

(11.12)

This implicit equation gives the optimal value of ıit and thus represents the trade-off between studying and working. The opportunity cost of an additional year of education is itY w while its discounted benefit is itO 0 .ıit /w=R. Up to now, in the absence of any redistribution mechanism, the decision of educating depends solely on the ratio between physical and intellectual abilities and on the retirement age.

11.2.3 Firms Firms produce the final good with the following CES production function:4 1= Yt D A.˛Kt C .1  ˛/L ; t /

(11.13)

where Kt is the stock of physical capital, Lt is the labour input in efficiency units, A is a parameter measuring total factor productivity, ˛ 2 .0; 1/ is related to the capital share5 and 1=.1 C  / is the elasticity of substitution between the two factors. It is convenient for the rest of the paper to note the production function in its intensive form, which means explaining y D Y =L as a function of  D K=L: y D f ./ D A.˛  C 1  ˛/1= :

(11.14)

If w is the wage rate, the labour share is defined as ˛ w D 1  : y  .1  ˛/ C ˛

(11.15)

For 1=.1 C  / D 1, i.e.  D 0, we have the Cobb–Douglas case with its constant labour share. For 1=.1 C  / < 1, i.e.  > 0, the wage share depends positively on the evolution of the wage rate compared to the total factor productivity. For 1=.1 C  / > 1 and  < 0, this is just the reverse. A non-constant labour share might be justified for developed countries as made apparent in Duffy and Papageorgiou (2000) with an elasticity of substitution slightly greater than 1. 4

For more details on the algebra of the CES production function, see e.g. de la Croix and Michel (2002). 5 Arrow et al. (1961) call ˛ the distribution parameter.

312

D. de la Croix and M. Lubrano

We assume that capital depreciate fully after one period. The competitive behaviour of firms leads to the equalisation of marginal productivity to prices so that the rate of return of capital and the wage rate are given by: R D A ˛.y=/.C1/ w D A .1  ˛/.y/.C1/ : The capital stock is formed by the aggregation of all savings Kt C1 D

Nt X

sit :

(11.16)

iD1

Labour market clears so that labour input in efficiency units is given by: Lt D

Nt N t1 X X O .1  ıit /itY hN t C it1 .ıit1 /hN t : iD1

(11.17)

iD1

In this model, labour supply in efficiency units is for one part a function of personal abilities, and is thus partly exogenous, but for the other part it results from an endogenous decision depending on the profitability of education. Any taxation-redistribution scheme is going to modify labour supply because it will alter the profitability of education for the old generation. Let us investigate now how an ELIE-like scheme can be introduced in an overlapping generations model and under which conditions it can keep its neutrality on labour supply in a dynamic setting.

11.2.4 Implementing ELIE in an OLG Model In order to have a self contained presentation, let us briefly summarise the main characteristics of a simplified ELIE-like scheme. In a given society, the ELIE scheme is a self financed distributive system where taxes and subsidies, P both denoted ti in Kolm (2005) and in this volume, balance with ti D 0 where i is an index covering all the individuals at a given period of time. An equal amount of labour is taken from each individual i measured in terms of his productivity while an equal monetary amount, k w, Q is redistributed to all, Q The variable wQ is determined so that the net transfer is ti D k.wi  w/. so as to balance the system. k is a parameter which measures the taxationredistribution rate. One of the main characteristics of ELIE, is that income is not taxed, but capacities are taxed in order to have the most inelastic taxation base as possible. In our OLG model, the modeler directly observes the working capacities of the individuals as the vectors t and t 1 are exogenous

11 The Trade-off Between Growth and Redistribution

313

and given a priori. So they are a perfect inelastic base for taxation. Finally, as ELIE concerns only active age individuals, we are not going to apply this scheme to the fraction of time 1   during which people are pre-retired. Note that  is a fixed policy parameter on which the individual has no action. The introduction of an ELIE scheme is going however to modify indirectly and in the long term the incentives faced by individuals in this economy. First, taxation of individual i is based on his labour capacities which are partially exogenous with the observed abilities itY and itO and partially endogenous because there is a decision to be taken for ıit , the degree of higher education. Taxation will reduce the future return on education. It should be noted however, that when considering the generations separately, ELIE has no distortive effect because education decisions are taken in the first period while the resulting enhanced human capital is taxed in the second period. So in the short term, the taxation base is perfectly inelastic as can be seen from (11.18). The second aspect of ELIE is that there is an equal redistribution to everybody. In the initial model, young people might have to borrow in order to get supplementary education. Here, the ELIE redistribution increases the incentives to educate by providing grants to young individuals who have a weak endowment itY of physical strength. For the young generation, we can either decide to tax their physical strength capacity whatever their decision of getting educated during a fraction ıit of the period. Or, we can decide to tax only the fraction of their time devoted to work and leave aside from the taxation base the fraction of their time during which they decide to educate. This last option is a subsidy to education. In this section, we shall present the benchmark model where the whole capacities of the young generation are taxed. This is in a way a pure implementation of ELIE which taxes capacities independently of labour supply decisions,6 here represented by ı. Refinements and by the way more realistic cases are considered in Sect. 11.6, where in particular we examine the consequences of using as a taxation base only the fraction of the strength capacities that is devoted to actual work and not to education. While being a priori contrary to the philosophy of ELIE, we shall see in Sect. 11.6 that this option is particularly attractive. For the simple benchmark case, when the whole capacities are taxed, the young age budget constraint is hN t Œ.1  ıit /itY w  k.itY w  y/ Q D citY C sit : 6

(11.18)

In his Chap. 8, Kolm (2005, p. 142) considers multidimensional labour where education is part of labour supply and where education time is added to labour, at least in a first approximation: “Another natural way of introducing training, formation, or education in the considered structure consists in adding the corresponding time to the duration of labour stricto sensu.”

314

D. de la Croix and M. Lubrano

The net transfer to a young individual i is given by: Q hN t : titY D k.itY w  y/

(11.19)

The old generation born at time t  1 also receive an identical transfer Q hNt 1 . This transfer must be multiplied k yQ hN t , which can be written as k yG by  because only active workers are concerned by ELIE transfers. Symmetrically, only the fraction  of their capacities is taxed. Once they have left the labour market, old workers only consume their previous savings. There is no taxation nor redistribution. ELIE is kept independent of any type of early retirement system here. The budget constraint for old people born at time t  1 writes: O O .ıit1 /w  k.it1 .ıit1 /w  G y/ Q D citO  Rsit1 :  hN t 1 Œit1 (11.20)

The net transfer to an old individual i is: O titO D k.it1 .ıit1 /w  G y/ Q hNt 1 :

(11.21)

The taxes he has to pay are a direct function of his human capital and of the decision to educate he took in the previous period. But what he receives depends on the level of both yQ and the growth rate g which are a function of past collective decisions to educate. We have to determine the constant yQ which will balance the budget of the system jointly for the young born at time t and for the old born at time t  1, because both live at time t. This means 2 !3 Nt N t1 O X X it1 .ıit1 / w  yQ 5 D 0 (11.22) Q C k hN t 4 .itY w  y/ G iD1

iD1

because hN t 1 =hNt D 1=G. As the mean of itY is one, we get a simplified expression for y: Q 3 2 Nt1  X w O 4Nt C it1 .ıit1 /5 : (11.23) yQ D Nt C Nt 1 G iD1

We retrieve the usual result of Kolm that yQ is equal to the mean wage w N in the case of a stationary economy (G D 1) and a degenerate function with .:/ D 1. We are here is a formal dynamic model which of course has properties which are different from the original ELIE. This model is nevertheless compatible with ELIE when it is reduced to the static case (G D 1 and .:/ D 1).

11 The Trade-off Between Growth and Redistribution

315

11.2.5 Optimal Education and Savings with ELIE To determine optimal education in the presence of the ELIE redistribution scheme, we maximise income over the life cycle as a function of ı: # " O .ı /w C kG yQ .1  k/ it it : Q C max hN t .1  ıit /itY w  k.itY w  y/ R ıit Proposition 11.1 (Optimal education with perfect credit market). Life cycle income of individual i is maximised for ıit satisfying 0

.ıit / D

itY itO

R : 1k

(11.24)

Proof. The first order condition (11.24) corresponds to a maximum because the function .:/ is concave. t u The individual choice for ı depends on his capacities, on the taxation rate k and on the endogenous rate of return on capital. Since is a concave function, education is increasing in the ratio of IQ to strength and in the age of retirement , and decreasing in the tax rate k and in the rate of return on capital r. There is thus a clear distortive effect of ELIE on the decision of educating. The previous case (11.12) can be recovered of course with k D 0. But it can also be recovered if , the retirement age, is made a function of k with for instance  D =.1 Q  k/. Other solutions are also possible and will be studied in Sect. 11.6. Saving is determined by young people, taking into account the old people that they will become. It is convenient to rewrite the income of the young for the first period as ! Y D Œ.1  ıit / Y w  k. Y w  y/ Q hN t (11.25) it

it

it

and the income of the young when they will become old in the second period Q hN t : (11.26) ! O D Œ O .ıit /w  k. O .ıit /w  G y/ itC1

it

it

Optimal saving sit is determined by a utility maximisation under two constraints: O max log citY C ˇ log citC1 subject to

!itY D citY C sit O O !itC1 D citC1  R sit :

316

D. de la Croix and M. Lubrano

The solution is given by sit D

ˇ 1 !itY  !O : .1 C ˇ/ R.1 C ˇ/ itC1

(11.27)

As usual in OLG models where individuals work during two periods, savings depends positively on income when young !itY and negatively on the disO (see de la Croix and Michel 2002). In this counted income when old !itC1 first version of the model, we do not impose a positivity constraint on sit . This means that individuals can freely borrow in the first period. The sole constraint is that their savings are zero at the end of the second period. This implies that the credit market is perfect. In particular, individuals can decide to borrow for educating.

11.2.6 Equilibrium We shall now collect the different parts of the solution in order to provide a formal definition of the equilibrium. For that purpose, it is convenient to define the following intensive variables, following the notations adopted for the CES production function: !O iY D !itY =hNt

!O iO D !itO =hNt

KO D Kt =hNt

sOi D sit =hNt

LO D Lt =hNt :

For defining a stationary equilibrium, we suppose that the size of the young generation Nt is kept proportional to N Y while the size of the old generation Nt 1 is kept proportional to N O so that their relative size is constant. Definition 11.1. Given the policy parameter k, an equilibrium with a perfect credit market is  A vector of individual variables fıi ; !O iY ; !O iO ; sOi g satisfying for i D

1:::NY:

0

Q !O iY D.1  ıi /iY w  k.iY w  y/;

(11.28)

Q !O iO DiO .ıi /w  k.iO .ıi /w  G y/;

(11.29)

.ıi / D sOi D

iY

iO

R ; 1k

ˇ 1 !O iY  !O O : 1Cˇ .1 C ˇ/R i

(11.30) (11.31)

11 The Trade-off Between Growth and Redistribution

317

O L; O g satisfying Q K;  A vector of aggregate variables fG; y; Y

N 1 X O GD Y i .ıi /; N iD1 3 2 NO X  w 4N Y C jO .ıj /5 ; yQ D Y O G N C N

(11.32)

(11.33)

j D1

Y

N X sOi O KD ; G

(11.34)

iD1 Y

O D L

N X

O

.1 

ıi /iY

C

iD1

N X

jO .ıj /;

(11.35)

j D1

O L: O  DK=

(11.36)

 And a vector of prices fR; wg satisfying

R DA ˛.f ./=/C1 w DA



C1

.1  ˛/.f .//

(11.37) :

(11.38)

11.3 Imperfect Credit Market We define an imperfect credit market as an environment in which young households cannot credibly commit their future labour income as a collateral against current loans. As in Kehoe and Levine (1993), we assume that individuals are allowed to borrow up to the point where they are indifferent between repaying loans and suffering market exclusion. Since everyone dies at the end of the second period, default involves no penalty and is individually optimal. As in this context it is optimal for them never to pay back their credits, banks will always refuse to lend them money. The borrowing constraint then takes the very simple form: sit  0. Let us first identify the individuals who are going to be affected by this constraint. Proposition 11.2 (Earnings profile and borrowing constraint). There exist a function  . Y ;  O /, such that individual i is credit constrained if and only if  .itY ; itO / < 0. The function  .:/ is implicitly defined by:

318

 .itY ; itO /

D. de la Croix and M. Lubrano

D

ˇ.1  k  ıit /itY

   G k yQ O  .1  k/it .ıit /  ˇ ; R R w (11.39)

with ıit given by (11.24). The function  .:/ is increasing in itY and decreasing in itO . Proof. The function  .itY ; itO / is derived from the condition sit  0 using the saving function (11.27) and the definitions of incomes (11.25) and (11.26). Since ıit is an increasing function of itY and a decreasing function of itY , the sign of the partial derivatives of  are not ambiguous. t u As in De Gregorio and Kim (2000) and in de la Croix and Michel (2007), households with a steep potential earning profile would like to borrow in order to study longer, but credit rationing prevents them from doing so. All others have positive saving and study as long as they wish. Hence constrained individuals are those with a relative low strength and high IQ. Note that the threshold function  depends on prices through (11.39). For example, when yields r are high, there will be fewer constrained households, other things being equal. Hence, although our borrowing constraint is very simple, the proportion of rationed people depends on equilibrium prices. For the constrained households, (11.24) no longer determines their education choice. Instead, these households maximise an autarkic utility, i.e. the utility they could reach without being able to use the credit market to smooth consumption. More explicitly, they choose education in order to maximise the utility function (11.8) where the consumption arguments were replaced O O D !itC1 by actual wages, with no possibility of saving: citY D !itY and citC1 so that   Q max ln .1  ıit /itY w  k.itY w  y/ ıit   C ˇ ln .1  k/itO .ıit /w C kG yQ C .1 C ˇ/ ln hN t C ˇ ln : Proposition 11.3 (Optimal education with imperfect credit market). The autarkic utility of an individual i with  .itY ; itO / < 0 is maximised for the unique value of ıit satisfying !! yQ k G yQ ˇ 0 .ıit / D .ıit / C : (11.40) 1  ıit  k 1  Y 1  k itO w it w Proof. The left hand side of (11.40) is decreasing in ıit , going from C1 to 0 as ıit goes from 0 to 1. The right hand side of (11.40) is increasing in ıit . Hence there exists a unique ıit equalising these two terms. t u

11 The Trade-off Between Growth and Redistribution

319

We can now define the equilibrium with an imperfect credit market. Definition 11.2. Given the policy parameter k, an equilibrium with an imperfect credit market is  A vector of individual variables f!O iY ; !O iO ; ıi ; sOi g satisfying for i D

1 : : : N Y (11.28)–(11.29) and   k G yQ .ıi / C 1k O i w 0   .ıi / D   yQ ˇ 1  ıi  k 1  Y

if  .iY ; iO / < 0

i w

D

iY

iO

R 1k

sOi D0 D

1 !O iO ˇ !O iY  1Cˇ 1Cˇ R

if  .iY ; iO /  0; if  .iY ; iO / < 0 if  .iY ; iO /  0:

O L; O g satisfying (11.32)–(11.36).  A vector of aggregate variables fG; y; Q K;  and a vector of prices fR; wg satisfying (11.37)–(11.38) When the credit market is perfect, ELIE acts as a obstacle to the decision of educating. When the credit market is imperfect, the ELIE scheme can help the constrained individuals in their decision to educating. The ability of ELIE to promote education will then depend on the proportion of constrained individuals.

11.4 Numerical Simulation of the Equilibrium The objective of this section is to calibrate and simulate the benchmark version of the model. Doing so will allow us to assess the size of the tradeoff between growth and redistribution in the perfect market case and to determine whether it is modified by the presence of borrowing constraints (imperfect credit market). Assumed that one period of the model is 22 years. It is then useful to define the annual growth rate of income and the annual interest rate as: g D G 1=22  1;

r D R1=22  1

320

D. de la Croix and M. Lubrano 2 1.9 1.8

y (d)

1.7 1.6

g g g g g

= = = = =

0.55 0.50 0.45 0.40 0.35

1.5 1.4 1.3 1.2 1.1 1 0.02

0.03

0.04

0.05

0.06

0.07

d

Fig. 11.1 Production function of human capital

11.4.1 A Priori Information We first choose a functional form for the production function of human capital and the distribution of abilities. The production of human capital has to satisfy the two limit conditions (11.5) to guarantee an interior solution for all agents. We use:   1  ı ı ; .ı/ D b  where  2 .0; 1/ and b is a scale parameter used as a degree of freedom for calibrating the model. In Fig. 11.1, we have graphed this function for a range of values for  that are within the domain compatible with our calibration exercise. The scale parameter b was adjusted accordingly to obtain a nice graph. The psychological discount factor of individuals is set to 3% per year. As we have assumed that one period of the model is 22 years, we have: ˇ D 0:9722 D 0:512. The growth rate of population n D N Y =N O 1 can be directly computed from official data which yields 1Cn D 1:177.7 Finally, we have taken N Y D 10;000, which implies that N O D NY =1:177 D 8;496. The abilities bivariate index . Y ;  O / is assumed to be distributed over a generation according to a bivariate lognormal distribution. The usual way 7

The total population in France is available from the Web site of INSEE http://www.insee.fr/fr/ffc/ pop_age2.htm, Population totale par sexe et âge au 1er janvier 2007, France métropolitaine. From these annual data, we computed the ratio between the population born between 1960 and 1981 and the population born between 1938 and 1959. The value of this ratio is 1.177.

11 The Trade-off Between Growth and Redistribution

321

of obtaining a lognormal distributed random variable is to take the exponential of a normal random variable. Let us thus consider a bivariate normal distribution with mean D Π1 ; 2 0 , and variance-covariance matrix p p     2 1  O2 = Y2  Y2 O2 2 p O p D Y : : (11.41)  Y2 O2

O2  O2 = Y2 O2 = Y2 This matrix has three parameters: the correlation  and the two variances Y2 and O2 . The resulting lognormal distribution has marginal means equal to exp. i C i2 /, marginal variances equal to &i2 D .exp. i2 /  1/ exp.2 i C

i2 /. The correlation coefficient % is independent of the means and equal to exp. Y O /  1 %D p : .exp. Y2 /  1/.exp. O2 /  1/

(11.42)

When  D 0, % D 0, but when  ¤ 0, then j%j < jj. Even if  is kept fixed, % varies with i . It is convenient, for elicitation purposes, to reparametrise this matrix in Y2 ,  and the relative variances O2 = Y2 . We do not have much information to calibrate this variance-covariance matrix. The parameter Y2 can be adjusted to match a measure of inequality for the observed income distribution in France. The Gini coefficient obtained on French gross income data and equal to 0.327 in 1998.8 This will be matched with the Gini coefficient implied by the model (computed in Appendix A). But we have no precise procedure to calibrate the two other parameters. It seems reasonable to assume that the ability to work when young is equally dispersed as the ability to work when old. However, ability in youth only reflects different endowments in efficient labour, while ability in old age also embodies the ability to accumulate human capital. We select O2 = Y2 D 1 in a first step and will carry some sensitivity analysis for O2 = Y2 D 1:5. The parameter  directly influences to proportion of types in society. With  D 0, the four possible types detailed below are in equal proportion. We will take  D 0 as a benchmark and we will carry sensitivity analysis for  D 0:9 which maximises the proportion of the type for which education makes an important difference. We assume that people retire at the age of 59 as reported by the OECD in 2002. This imply that  ' 10=11. The productivity parameter b governs the long-term growth rate of output per capita. We shall adjust it on the observed growth rate of GDP per capita

8

This figure comes from the Human Development Report of the United Nations http://hdrstats. undp.org/indicators/147.html. It can also be found elsewhere, such as in the World Fact Book of the CIA.

322

D. de la Croix and M. Lubrano

that we collected from Maddison (2007) data over the period 1981–2003. We have G D 1:44, which gives an annual growth rate g of 1.67%. The parameter  determines the time spent on education in the first period of life. We shall adjust it so as to match the observed share of time devoted to education. We assume that the first period of the model covers ages 18–39. Doing so supposes that higher education is an alternative to working, but elementary and secondary education is not. The percentage of time devoted to schooling is computed using Education at a Glance from OECD (2006) (Indicator A3, page 53). We use Tertiary type A and B graduation rates and obtain ı D 0:075. As far as technology is concerned, we borrow from Duffy and Papageorgiou (2000) the conclusion that, in developed countries, the elasticity of substitution between capital and labour adjusted human capital is of the order of 1.1 and we set  D 0:1. From Askenasy (2003) we take that the share of capital in value added, =y, is 0.35 (it fluctuates between 0.32 and 0.38 in the last 30 years). The physical capital share parameter ˛ will we set to match this value. We also learn from this study that the rate of return on capital in the manufacturing sector fluctuates between 9% and 14%. We do not use this information directly but it will serve as a benchmark to check whether our equilibrium r is in line with the data. Finally, the scale parameter A is normalised to 1. Varying A leaves everything else unchanged provided that we adjust ˛ to keep the same capital share. We now summarise the available a prior information in Table 11.1.

11.4.2 Calibration of the Model In order to impose the a priori information on growth, education and inequality, we have four parameters of adjustment, the two parameters of the production function of human capital .ı/, a scale parameter for the variance covariance matrix of the lognormal distribution Y2 , and the capital share parameter ˛. Given starting values for the rate of return on capital r and the wage rate w, the model is solved iteratively using the fixed point algorithm Table 11.1 A priori information used for calibration

O2 = Y2 1.0 1.5 1.0

 0.0 0.0 0.9

N Y =N O 1.177 1.177 1.177

 0.91 0.91 0.91

=y 0.35 0.35 0.35

ˇ 0.512 0.512 0.512

 0.1 0.1 0.1

g 1.67% 1.67% 1.67%

ı 0.075 0.075 0.075

Gini 0.327 0.327 0.327

The first line is used for calibrating the benchmark model. The last two lines are used for sensitivity analysis.

11 The Trade-off Between Growth and Redistribution

323

described in Appendix B, conditionally on initial given values for  , b, Y2 and ˛. As a by-product, the model produces a vector ı, a growth rate g, an income distribution for which a Gini coefficient is computed, and a capital share in value added =y. The four adjustment parameters are then updated using the following scheme N  D  C .0:075  ı/ b D b C .:0167  g/

Y2 D Y2 .0:327  Gini/ ˛ D ˛ C .0:35  =y/ and the process is iterated until convergence is reached. We found the following solution displayed in Table 11.2. The obtained rate of return on capital is 10.14% on an annual basis which is within the range provided by Askenasy (2003). Let us now calibrate the model with an imperfect credit market. When we keep the same parameters, we see from the second line of Table 11.2 that credit rationing entails a drop in education and growth, and incidently in inequality too. The capital share increases. In order to make comparisons between the two cases, we have to recalibrate some of the parameters. We recalibrate b in order to match the same growth rate as before, which requires an increase in this coefficient to compensate for the loss of growth due to the imperfection of the credit market. We also recalibrate Y2 to match the required level of inequality. Matching the same growth-inequality pair in the two versions of the model allows to compare the trade-off between growth and redistribution across them. We do not alter the parameters  and ˛ to make the two models similar in this respect. The results are reported in the line labeled (1) of Table 11.2. We also report a calibration where we only adjust parameter b and leads to a similar result. In the last line of Table 11.2, the one labeled (3), we report a calibration of the imperfect credit market model where we compute the four parameters  , Table 11.2 Calibration and solutions of the initial model Perfect credit Imperf. credit (0) Imperf. credit (1) Imperf. credit (2) Imperf. credit (3)

˛ 0.475 0.475 0.475 0.475 0.473

 0.456 0.456 0.456 0.456 0.503

b 2.387 2.387 2.421 2.431 3.027

Y2 0.204 0.204 0.223 0.204 0.211

r (%) 10.14 9.94 10.02 10.04 10.02

g (%) 1.67 1.59 1.67 1.67 1.67

ı 0.075 0.070 0.070 0.070 0.075

Gini 0.327 0.316 0.327 0.316 0.327

=y 0.350 0.352 0.351 0.351 0.350

In (0), no parameter is adjusted. In (1), b and Y2 are adjusted, in (2), only b is adjusted, in (3) the four parameters are adjusted.

324

D. de la Croix and M. Lubrano

b, Y2 and ˛ in order to match the four targets g, ı, Gini and the capital share. We see that doing so requires an important rise in the elasticity of human capital to education  .

11.4.3 Heterogenous Behaviour Without Redistribution To better grasp the logic of the model, we distinguish four groups of individuals, depending on their abilities  Y when young and  O when old. Given the median of each marginal of the joint distribution of . Y ;  O /, we classify each individual in a two by two entry table. Type 00 has a physical strength  Y lower than the median and an intellect  O lower than the median. For convenience, we call this type white collars. Type 10 has a physical strength  Y greater than the median and an intellect  O lower than the median. We call this type blue collars. Type 01 a physical strength lower than the median and an intellect greater than the median. We call this type academics. Finally, type 11 has a higher physical strength and a higher intellect. We call this type managers. Table 11.3 presents some characteristics of these different groups. As  D 0, each type represents 25% of our sample. All types decide to educate, but according to different degrees. Types 00 and 11 choose to educate around the mean, type 01 (academics) chooses to educate twice the mean, while type 10 (blue collars) has the lowest decision of education. These decisions have marked consequences. Type 10 (blue collars) are major savers because they will earn well above the mean when young, but below the mean when old. Type 01 (academics) net saving is roughly zero. 41% of this group borrow to finance longer education and have the prospect of earning a very high wage when old. Notice that the Table 11.3 Education and saving decisions Y

Education O 0 1

0 1

(0.075) 0.062 0.021

0 1

(0.070) 0.064 0.021

Net savings 0 1

0.154 0.063

(0.017) 0.011 0.032

Borrow. prop. Income young 0 1 0 1 Perfect credit market (0.116) (0.073) 0.001 0.026 0.410 0.048 0.043 0.023 0.000 0.027 0.102 0.098

0.129 0.066

(0.018) 0.011 0.032

Imperfect credit market (0.000) (0.074) 0.003 0.000 0.000 0.047 0.023 0.000 0.000 0.105

0.044 0.100

Life cycle income 0 1 (0.086) 0.056 0.107

0.067 0.115

(0.087) 0.055 0.110

0.065 0.117

Physical strength when young is indicated in column and intelligence when old is indicated in line. The mean value for each small two-two table is indicated between brackets. The total life cycle income for a young individual is given by !Q it D !itY C Œ.1  k/itO .ıit /w C kG y=R. Q

11 The Trade-off Between Growth and Redistribution

325

group of academics will earn the minimum when young (roughly the same as white collars), but will receive the maximum when old so that over the life cycle their earn more than white collars. Managers and blue collars receive a similar average income when young, but a quite different one when old. Considering now the equilibrium with an imperfect credit market, we observe that the academics 01 are strongly hit by the impossibility to borrow. Their education is reduced, and their life-cycle income as well. We have analysed the sensitivity of these results to the choice of . Increasing correlation up to  D 0:50 and recalibrating the model to fit observed growth, education and inequality, we observe that this higher correlation between abilities greatly diminishes the proportion of borrowers. The type proportion is changed to 0.333 (white collars and managers) and 0.167 (blue collars and academics). In this world, 33.1% of the academics borrow money for educating which relates to an average of 7.8% in the whole economy (instead of 11.6% in the benchmark calibration). We do a similar exercise for O2 = Y2 D 1:5 while keeping  D 0 to investigate the robustness of the results to different relative variances. In this case the type proportions are only very slightly modified compared to the benchmark. 36.6% of the academics borrow money for educating which relates to an average of 10.4% in the whole economy. Results are close to those of the benchmark.

11.5 The Trade-off Between Growth and Redistribution We now introduce the ELIE transfer system. We do so by letting k vary between 0 and 0.40. Let us recall that k D 0:40 means that for a working week of five days, the product of two days is taken for redistribution. Remember that in our model both inequality and growth are endogenous. We have seen in Proposition 11.1 that individual investment in education is negatively affected by taxation k, but this was only a partial equilibrium effect, for a given rate of return on capital. We will now investigate whether this partial equilibrium effect carries over to the general equilibrium framework; the numerical simulation will also allow us to quantify this effect. In Sect. 11.6, we investigate how this trade-off can be overturned by a different implementation of ELIE. As the system balances, money is taken from some individuals and distributed to others. If the focus of the analysis was on the life cycle of one generation in the previous section, it has now to be on the two generations together. This means that at time t, we have to study the interaction between young and old and detail the possible intergenerational transfers. The ELIE

326

D. de la Croix and M. Lubrano

transfer system has the particularity of reducing inequality in the income distribution. We will investigate by how much does the ELIE scheme affect the Gini coefficient of young and old incomes.

11.5.1 Simulations Results Let us first analyse the impact of ELIE on macroeconomic variables, before analysing its impact in term of inequality and poverty. First consider the case of a perfect credit market. The impact of ELIE on the growth rate is negative as shown in Table 11.4. The young generation decides to educate less and growth in our model is affected solely by the growth of human capital. However, ELIE also decreases the proportion of young individuals that are obliged to borrow to finance their supplementary education. The lower investment in education allows an increasing capital labour ratio which in turn implies an increase in the wage rate per efficient unit of human capital and a lower rate of return on capital. The decrease in the rate of return on capital dampens the negative effect of k on the decision of education but does not overturn it. Considering now the imperfect credit market case, the question here is whether the negative effect of redistribution on growth via the distortive highlighted above can be overturned by a positive effect of redistribution on growth through the easing of borrowing limits bearing on poor people. The answer is no. Table 11.4 shows that growth is still decreasing in k, indicating that the predominant effect is still the distortion one. But the drop Table 11.4 Macroeconomic impact of ELIE k

g r (% annual)

ı

0.0 0.1 0.2 0.3 0.4

1.67 1.52 1.35 1.15 0.91

10:14 9:95 9:73 9:48 9:19

0.075 0.069 0.063 0.056 0.050

Saving Percent. % credit rate borrowers constr. Perfect credit market 22.88 11:63 0:00 22.86 9:28 0:00 22.85 7:04 0:00 22.83 4:99 0:00 22.82 3:39 0:00

0.0 0.1 0.2 0.3 0.4

1.67 1.54 1.39 1.21 0.99

10:02 9:89 9:72 9:51 9:25

0.070 0.066 0.061 0.055 0.049

Imperfect credit market 23.61 0:00 13:77 23.42 0:00 11:05 23.25 0:00 8:27 23.11 0:00 5:85 23.00 0:00 3:87

Gini

Headcount poverty

0.327 0.295 0.264 0.232 0.200

0.281 0.242 0.195 0.143 0.082

0.327 0.298 0.268 0.237 0.205

0.287 0.251 0.204 0.153 0.090

11 The Trade-off Between Growth and Redistribution

327

in growth is slightly less severe than in the previous case, indicating that the effect of ELIE on borrowing limits helps to limit the cost of taxation in terms of growth. If the proportion of constrained individuals on the credit market were larger, the compensating effect of ELIE would have been larger. The right panel of Table 11.4 provides inequality measures for different k in the two economies. Not surprisingly, ELIE manages to reduce inequality in the population (this is also true for within group inequality, but inequality remains greater in the older generation.) We define a poverty level as 60% of the mean income.9 We compute a head count measure of poverty as the proportion of individuals below the poverty level. Increasing k from 0:00 to 0:40 allows to decrease poverty from 28:1% of the population to 8:2%. Comparing the model with perfect credit market to the one with imperfect credit market, we observe that ELIE diminishes the Gini coefficient in the same way in both cases, but is slightly less efficient at reducing poverty when credit market is imperfect. Notice that, if one wishes to totally remove poverty, one needs to push k as high as 0:60. The elimination of poverty by the ELIE scheme is further illustrated in Fig. 11.2. We observe that the income distribution is fairly regular and corresponds to the shape of a log-normal distribution when there is no redistribution. The ELIE scheme shifts the whole distribution to the right (poverty 18.00

k k k k k

16.00 14.00

= = = = =

0.0 0.1 0.2 0.3 0.4

12.00 10.00 8.00 6.00 4.00 2.00 0.00 0.00

0.05

0.10

0.15 0.20 Net income

0.25

0.30

Fig. 11.2 Income distribution with perfect credit market

9

There are various ways of defining a relative poverty line. EUROSTAT defines the poverty level as 60% of the median income. France and INSEE use 50% of the median income. The European Commission once used 50% of the mean in its reports. See Atkinson (1998) for a discussion.

328

D. de la Croix and M. Lubrano

reduction), except for the extreme right tail which is dampened (inequality reduction). We only report the graph for the perfect market case. The imperfect market case produces a very similar graph.

11.5.2 Assessing the Size of the Trade-off We have seen in the previous section that increasing the value of k reduces growth but promotes redistribution and reduces poverty. We measure inequality using a Gini index computed on the total income distribution and poverty using the P0 index of Foster et al. (1984). The importance of the trade-off between growth and redistribution is illustrated in Fig. 11.3 where the graph of 1Gini and 1  P0 against the annual growth rate is displayed. We give this graph both when credit market is perfect and when it is imperfect. The difference between these two cases is not negligible, but not large either. We can measure the difference in the trade-off by comparing the slope of the two curves. The slope with perfect credit market is equal to 5.90, which implies that reducing the Gini by 1 point costs 0.059 in term of annual growth rate. In the imperfect market case, the slope is lower in absolute value and equal to 5.58. The ratio of two slopes is 0.95. Hence the trade-off between growth and redistribution is slightly less severe with an imperfect credit market, but is far from being overturned. As far as poverty is concerned, reducing poverty by 1 point costs 0.0376 in terms of annual growth rate. This cost drops to 0.0347 when the credit 1.70

1 – Gini 1 – Gini rationing 1 – P0 1 – P0 rationing

Annual growth rate

1.60 1.50 1.40 1.30 1.20 1.10 1.00 0.90 0.65

0.70

0.75

0.80

0.85

0.90

0.95

Fig. 11.3 Trade-off between redistribution and growth for k D 0 to k D 0:4

11 The Trade-off Between Growth and Redistribution

329

market is imperfect. Here the ratio of the two slopes is 0.92. These numbers indicate that ELIE is quite good at reducing poverty at a relatively low cost in terms of growth, and that this is even more true if individuals face borrowing constraints. Still again, we are far from a case of “growth-enhancing redistribution”, where redistribution lifts so much the credit constraints that the disincentive effect on growth is overturned, as suggested by Bénabou (2005) for instance.

11.5.3 The Trade-Off with a Larger Proportion of 01 Type When the credit market is imperfect, the ELIE scheme helps some poor individuals to invest in education. These are the 01 types (the academics) who have a strong potential in terms of future income growth but little resources when young. This is why the trade-off between growth and redistribution is less severe when the credit market is imperfect. But ELIE also redistributes to the 00 types (white-collars), which is fine as far as equality is concerned, but is of no help as far as growth is concerned. This explains why the tradeoff is not modified much when the credit market is imperfect. Unless ELIE is targeted towards the 01 households, its effect on borrowing constraints is not strong enough to suppress its negative effect on growth. To illustrate this point, we consider a calibration of the model with a strong negative correlation between the two ability shocks:  D 0:9. In that case the economy is mostly composed of academics and blue collars (43% of population each). Among young individuals, ELIE will redistribute in favour of academics, without “wasting” too much resources on white collars, who form a small 7% fraction of the population. Hence, ELIE is much more targeted towards persons with a strong growth potential. Assuming such a strong negative correlation is of course unrealistic, but this simulation is meant to illustrate the properties of the ELIE scheme as a function of the type of ability distribution in the population. We recalibrate the model with a perfect market using  D 0:9. This gives  D 0:413, b D 1:890, Y2 D 0:159 and ˛ D 0:475. The empirical correlation between the two shocks is % D 0:74. We also recalibrate the model with an imperfect credit market to obtain the same growth and inequality without ELIE (b D 1:931 and Y2 D 0:188). Then we simulate various levels of redistribution by letting k vary between 0.00 and 0.40. The distance between the curves with perfect and imperfect credit market is now more important. This is because ELIE is now more targeted towards 01 people and plays therefore a greater role in alleviating the credit constraints for the individuals with a strong growth profile. The ratio between

330

D. de la Croix and M. Lubrano

the two slopes is now 0.87 (against 0.95) for the 1Gini slopes and 0.81 (against 0.92) for the 1  P0 slopes. Notice finally that the case with  D 0:9 is the most favourable situation to generate a positive influence of ELIE on growth. Although we know little on the distribution of abilities in the population, and hence the parameters of this distribution are subject to a large uncertainty, it seems pretty clear now that no parameter configuration would be able to reverse the trade-off between growth and redistribution. We have to find something else.

11.6 How to Overturn the Trade-Off The original ELIE has no distortive impact because the tax base is chosen independently of labour supply decisions. Once we introduce a decision for educating in a two generation model, the distortive effect reappears. We have studied up to now the least favourable case. We have given indications on how to reduce the distortive effect of ELIE by an alternative implementation. We now explore two possibilities which are equivalent to either subsidising education in the first period or making it more profitable in the second period.

11.6.1 Education Subsidies The crucial decision of educating has to be taken in the first period. ELIE had a disincentive effect on that decision, because labour capacities when senior are taxed at a proportional rate while the opportunity cost when young is not tax deductible (the whole physical capacity itY was taken as a basis for taxation). In doing so, we had a dogmatic vision of ELIE where the taxation basic must be independent of labour supply decisions and thus of ı. What happens if we now decide to apply ELIE only to the sole fraction of itY that is devoted to actual work and to exclude the fraction which is devoted to education? The taxation base is no longer itY , but .1  ıit /itY . The young age budget constraint becomes hN t Œ.1  ıit /itY w  k..1  ıit /itY w  y/ Q D citY C sit ;

(11.43)

instead of (11.18). The net transfer to a young individual i is now: Q hN t ; titY D k..1  ıit /itY w  y/

(11.44)

11 The Trade-off Between Growth and Redistribution

331

replacing (11.19) of the benchmark. As in the benchmark, they receive k yQ hNt (while equilibrium yQ will be different). But in the benchmark case, they had to pay kitY whN t , while here, they have only to contribute to the system for k.1  ıit /itY whN t . The more they educate, the less they contribute to the system in the first period. There is thus a subsidy to educating and implicitly a transfer from the old generation to the young generation. Let us now determine the constant yQ which will balance the budget of the system. Balanced budget implies 2 !3 Nt N t1 O X X  .ı / it1 it1 w  yQ 5 D 0 Q C k hN t 4 ..1  ıit /itY w  y/ G iD1

iD1

(11.45) because hN t 1 =hNt D 1=G. This implies 3 2 Nt N t1 X X w O 4 .1  ıit /itY C  it1 .ıit1 /5 : yQ D Nt C Nt 1 G iD1

(11.46)

iD1

We can no longer simplify the expression using the assumption that the mean of the itY is one. To determine optimal education in this new scheme, we maximise income over the life cycle as a function of ı: h Q max hNt .1  ıit /itY w  k..1  ıit /itY w  y/ ıit # .1  k/itO .ıit /w C kG yQ : C s R The first-order condition for a maximum is given by 0

.ıit / D

itY itO

R;

(11.47)

which is the same expression as in the case when there is no ELIE scheme (11.12). Hence, when education time is deductible from taxes, the distortive effect of ELIE should disappear. The implicit subsidy implied by the deductibility exactly offset the effect of the tax bearing on future income. Compared to the benchmark, we will no longer have the distortive on education choices; but we will have a lower transfer yQ since the tax basis has been shrunk.

332

D. de la Croix and M. Lubrano

11.6.2 Linking Early Retirement to Redistribution As an alternative to subsidising education in the first period, we can make it more profitable in the second period, simply by increasing the early retirement age . Equation (11.24) suggests to link this age to the intensity of redistribution by implementing Q ; 1k where Q is the retirement age in the case without ELIE scheme. Then, the optimal rule for education (11.24) becomes (11.47) again. Increasing the length of active live raises the return on education investment. Here, by letting the retirement age increase with redistribution, we compensate the negative effect of the ELIE tax on education by increasing length of active life and, hence, the return on education. Again here, the distortive effect of ELIE disappears. We are left with a rising labour supply as redistribution increases. D

11.6.3 Numerical Assessment The two alternative efficient implementations of ELIE have clearly different macroeconomic properties, despite the fact that they both imply the same decision function for educating. They are not equally feasible. If subsidising education is possible whatever the value of k, postponing retirement as a function of k can be implemented only for a small range of values of k. Here as  is already close to 1, this solution can work only for k  0:1. Table 11.5 illustrate the macroeconomic properties of these two implementations using the same calibration as before with a perfect credit market. Table 11.5 Macroeconomic impact of ELIE with subsidies to education k

g

r (% annual)

ı

0.0 0.1 0.2 0.3 0.4

1.67 1.68 1.68 1.68 1.69

10.14 10.13 10.12 10.11 10.11

0.075 0.075 0.075 0.076 0.076

0.0 0.1

1.67 1.73

10.14 10.53

0.075 0.078

Saving Percent. rate borrowers Subsidising education 22.88 11:63 22.91 9:20 22.94 6:64 22.98 4:41 23.01 2:76 Decreasing pre-retirement 22.88 11:63 22.39 10:18

Gini total

Headcount poverty

0.327 0.294 0.262 0.229 0.196

0.281 0.241 0.193 0.140 0.078

0.327 0.294

0.281 0.242

11 The Trade-off Between Growth and Redistribution

333

We give between brackets results for the option consisting in postponing retirement. When education is subsidised, the rate of growth of the economy is no longer decreasing with k, but is even slightly increasing with it (more increasing). The rate of return of capital remains more or less constant (increases) as well as the capital share (decreases). The wage rate increases (decreases). Education slightly increases (strongly increases). The percentage of borrowers decreases and is lower than in the benchmark model (decreases slowly and less than in the benchmark model). The share of the young generation in total income slightly increases (decreases). If we now look at inequality and poverty, they are both slightly more reduced compared to the benchmark model. But postponing retirement decreases inequality less in the young generation but more in the old generation, compared to the solution of subsidising education. It is thus fairly possible to find an implementation of ELIE that has no distortive effect. On the contrary, that new implementation can even be growth enhancing, even if the credit market is perfect. The solution of subsidising education is much easier to implement and certainly more politically feasible than postponing retirement. In the benchmark model, the equilibrium wage rate and the pivot for redistribution yQ are roughly equal. When education is subsidised, w is 3.5% higher than y. Q So there is slightly less to redistribute, but in both cases w and yQ increase with k at exactly the same pace. Moreover, inequality in the young generation is unaffected by subsidising education. Inequality in the old generation is significantly reduced when education is subsidised in the first period so that overall inequality and poverty are more reduced in that case than in the benchmark model.

11.7 Conclusion The ELIE scheme of Kolm (2005) proposes to tax labour capacities instead of labour income in order to promote social freedom. Once the individual has paid kwi to society, he is free to dispose of his supplementary earnings as he pleases. As a secondary effect, the taxation redistributive scheme ELIE has no distortive effect on labour supply. The question of human capital formation and investment is addressed by Kolm (2005) only marginally as a dimension of labour (its quality and efficiency, Chap. 8), or as a piece of information (Chap. 10) for practically implementing ELIE. Otherwise, ELIE is confined to a static world with no consideration for dynamics, growth and inter-temporal optimisation. In this paper, we have built an overlapping generations model with heterogenous agents and endogenous growth driven by investment in human

334

D. de la Croix and M. Lubrano

capital. We have studied the effect of the ELIE scheme on education investment decisions and other aggregate economic variables. The fundamental question is to decide how to implement the ELIE scheme in this growth model and which part of the capacities to use as an inelastic basis for taxation. Clearly, the whole capacities can be taxed for the old generation. For the young generation, theory shows that, ceteris paribus, the implemented ELIE has a negative effect on investment decisions in education if the whole capacities are taxed. This effect arises because the implemented ELIE taxes future labour income, which reduces the return to investment in human capital in an inter-temporal optimisation. The distortive effect of this implementation of ELIE is completely overturned if the part of the capacities that are used for financing education are taken out of the tax base. This is a form of subsidy to education. This result can be seen paradoxical, because the implemented ELIE looses its distortive effect in a dynamic setting just when its basis is made elastic for the young generation. Calibrating the model on French data, we illustrate the traditional tradeoff between growth and inequality when whole capacities are taxed. In its crude implementation, ELIE is successful at reducing inequality and poverty, but at the expense of a lower investment in education and a slower growth rate. In a world with an imperfect market where individuals cannot borrow to educate in the first period, the trade-off between growth and redistribution is modified. Indeed, in such a world, ELIE helps poor students to finance their education which counteracts partly its negative effect on the future return to education. But since ELIE redistributes to all poor people, and not only to those with a strong growth potential, the beneficial effect of ELIE obtained by releasing borrowing constraints is quantitatively small. Using an alternative implementation of ELIE, growth can remain constant while inequality is reduced. This variant of the model, calibrated on French data, shows that education has to be subsidised if we want to escape from the traditional trade-off between growth and redistribution. Moreover, the usual argument according to which students should pay high fees at the university because those fees are partly compensated by their discounted future earnings is wrong. Our model shows that when there is redistribution, high fees have a disincentive effect on education decisions. And it also shows that subsidising education when there is redistribution enhance growth and reduces inequality in a better way.

11 The Trade-off Between Growth and Redistribution

335

Appendix A Model’s Income Distribution We give here the formula to derive the net income distribution of the population living at time t. It is formed by the concatenation of the vector of income of the young generation and of the vector of income of the old generation living at the same time. For the young generation the budget constraint gives: Q !iY D .1  ıi /"Yi w  k.iY w  y/

for i D 1:::N Y :

For the old generation, the net income is, still up to the multiplicative factor hN t , established for the working part of that generation given by the budget constraint (11.20) !jO D ŒjO .ıj /w  k.jO .ıj /w  G y/=G Q

for j D 1:::N O :

Taking into account that ht 1 D ht =G. The net income distribution is thus given by:  0  0  0 Y NY O NO : ! D Œ!i iD1 ; Œ!j j D1 The income distribution is computed for the age group 18–62. Accordingly, the relevant income of the old generation is here !itO , and  does not enter this formula, contrary to (11.20). We compute the Gini coefficient for !.

Appendix B Numerical Methods The model is solved using a traditional fixed point algorithm. We give below the procedure to compute the equilibrium with credit constraints for given parameters. The case without constraints is just a simplification of this more complicated case. We first have to fix starting values for the aggregate variables r, w and y. Q Then we apply the following algorithm. – Step 1 identify constrained agents running (11.39) for the first generation using t , store the results in id1 . Do the same for the second generation using t 1 and store the result in id 2 .

336

D. de la Croix and M. Lubrano

– Step 2 – compute the optimal ı using (11.24) – compute the constrained optimal ıc solving (11.40) using a fixed point algorithm – ıj D id j ı C .1  id j /ıc for j D 1; 2. – Step 3 compute the growth rate g, the different income and transfers vectors, and the vector of savings. Deduce K, L, r, w and y. Q Q If the sum of the absolute changes – Step 4 Check the change in ı1 and y. is greater than 106 , go to step 1. Otherwise deliver the needed vectors and equilibrium values. In the unconstrained case, step 1 does not exist and step 2 does not involve computing ıc .

References Arrow, K. J., Chenery, H. B., Minhas, B. S., & Solow, R. M. (1961). Capitallabor substitution and economic efficiency. Review of Economics and Statistics, 43, 225–250. Askenasy, P. (2003). Partage de la valeur ajoutée et rentabilité du capital en France et aux Etats-Unis: une reévaluation. Economie et Statistique, 363–365:167–189. Atkinson, A. B. (1998). Poverty in Europe. Yrjö Jahnsson Lectures. Oxford: Blackwell. Azariadis, C., & de la Croix, D. (2006). Financial institutional reform, growth, and equality. In T. S. Eicher & C. García-Peñalosa (Eds.), Institutions, development, and economic growth, CESifo Seminar Series, (pp. 33–64). Cambridge (USA): MIT Press. Azariadis, C., & Drazen, A. (1990). Threshold externalities in economic development. Quarterly Journal of Economics, 105(2), 501–526. Bénabou, R. (2005). Inequality, technology and the social contract. In P. Aghion & S. Durlauf (Eds.), Handbook of economic growth (Vol. 1, Chapter 25, pp. 1595–1638). Elsevier. Cardia, E., Kozhaya, N., & Ruge-Murcia, F. J. (2003). Distortionary taxation and labor supply. Journal of Money, Credit and Banking, 35(3), 350–73. De Gregorio, J., & Kim, S.-J. (2000). Credit markets with differences in abilities: education, distribution and growth. International Economic Review, 41(3), 579–607.

11 The Trade-off Between Growth and Redistribution

337

de la Croix, D., & Michel, P. (2002). A theory of economic growth: Dynamics and policy in overlapping generations. Cambridge: Cambridge University Press. de la Croix, D., & Michel, P. (2007). Education and growth with endogenous debt constraints. Economic Theory, 33, 509–530. Duffy, J., & Papageorgiou, C. (2000). A cross-country empirical investigation of the aggregate production function specification. Journal of Economic Growth, 5, 87–120. Foster, J., Greer, J., & Thorbecke, E. (1984). A class of decomposable poverty measures. Econometrica, 52, 761–765. García Peñalosa, C., & Turnovsky, S. (2007). Growth, income inequality, and fiscal policy: what are the relevant trade-offs? Journal of Money, Credit and Banking, 39(2–3), 369–394. Kehoe, T., & Levine, D. (1993). Debt-constrained asset markets. Review of Economic Studies, 60(4), 865–888. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge (UK): Cambridge University Press. Maddison, A. (2007). Contours of the world economy, 1-2030AD; Essays in macroeconomic history. Oxford: Oxford University Press. http://www. ggdc.net/maddison. OECD (2006). Education at a glance. Paris: OECD Publications. Solow, R. M. (1956). A contribution to the theory of economic growth. The Quarterly Journal of Economics, 70(1), 65–94.

•

Part VI

Selective Comments by Serge-Christophe Kolm

•

Chapter 12

Macrojustice in Normative Economics and Social Ethics Serge-Christophe Kolm

Abstract These concluding comments and considerations briefly summarise the achievement of this volume’s set of complementary contributions, answer the main questions posed and pending, and provide the necessary basic analysis and evaluation of the other distributive principles which are alternatives or complements to the one obtained here. The basis is the synthesis between the three polar possible ethical principles for macrojustice: income justice, self-ownership and the proper welfarism. Its central piece is a distributive coefficient k which can provide solutions from the pure self-ownership of classical liberalism for k D 0 to freedom-respecting income-egalitarian ideals. This choice results from the impartial moral judgment of the distributive society in question, representable in particular by comparisons between the “pure welfare” of members and which can be revealed by a number of methods. Non-human resources can be allocated according to various principles, but their equal sharing results from various types of association with the solidaristic equal-freedom allocation of the value of the human resources, and it permits more self-ownership for the same degree of distribution. The relevant introduction of the obtained distributive policy makes everybody better-off. Distributive principles alternative to the one obtained include those based on ordinal welfarism (equity-no-envy and the equivalence principle) and reductions to mesojustice or microjustice. Finally, the moral public goods of justice or caring about others’ needs elicit various types of motives which make the nature of the corresponding transfers be quite more subtle and rich than simple coercion or voluntariness.

S.-C. Kolm EHESS, Paris and CREM e-mail: [email protected]

C. Gamel and M. Lubrano (eds.), On Kolm’s Theory of Macrojustice, c Springer-Verlag Berlin Heidelberg 2011 DOI 10.1007/978-3-540-78377-0_12, 

341

342

S.-C. Kolm

12.1 The Results Obtained: A Synthesis 12.1.1 Basic Logic The contributions of the various chapters of this volume make it a unique compendium of the issues raised by the question of overall distributive justice in macrojustice (as opposed to allocations specific in nature, circumstances or people involved, in questions of microjustice or mesojustice). They cover both the basic pure logic of the question and the analysis of and solutions to problems met by its application. The result, first of all, avoids naïveté, dogmatism and inconsistency about both the relevant and demanded ethical principles and actual informational possibilities. People’s satisfaction, happiness, non-suffering and similar formulations is what people want and this translates in economic terms as Pareto efficiency (this condition of non-waste in these individual values is also the impossibility of further unanimous improvements, hence a condition of democracy and – insofar as all actual constraints of all types are taken into account – something favourable to the stability and hence to the possible existence of the social state described). There remains, however, to determine the distribution of welfare or incomes, and people do not find comparisons between people’s tastes, desires and capacities to enjoy to be relevant for macrojustice such as that realised by the income tax or main transfers (they are found relevant in small empathic groups or for the relief of suffering). This simple observation of people’s opinions and of the ways in which these policy choices are made is confirmed by a large variety of tests and systematic enquiries (for example, the econometric analysis of Bourguignon and Spadaro (2008) proves that governments do not maximise a classical social welfare function, even implicitly, when they choose the income-tax schedules). However, judgments sometimes refer to comparisons of variations or levels of individuals’ “welfare” as if individual welfare were the same function for all individuals (for instance of income) – this is for instance one of the reasons presented for making richer pay more than poorer or for transferring income from richer to poorer when the lower income is sufficiently low. A key theoretical point is the possibility to define normative concepts of individual welfare cleaned of sui generis differences in tastes and hedonic capacities, and to compare it (or its variations) across individuals. However, the result is not Pareto efficient and other allocations are preferred by everybody (moreover, this “utility function” identical for all individuals does not describe their choices as required by some theories). Pareto efficiency can nevertheless be obtained by letting people free to choose, exchange and work

12 Macrojustice in Normative Economics and Social Ethics

343

from a given allocation – a most classical economic framework (for instance that of Pareto himself).

12.1.2 Solution One such solution is simply the full self-ownership of classical liberalism. This implies no transfers except gifts (and perhaps compensation of past torts), however, whereas present-day national redistributions commonly exceed 30% of GNP (in Europe) – up to 40% in Scandinavian countries – and this level of solidarity is roughly commonly accepted. These national distributions are seen and obtained as a political or moral synthesis of, or compromise between, two opposite (and themselves composite) ideals or tendencies: on the one hand a lowering of income inequalities – criticised in the name of justice in consumption and freedom of choice and of some kind of compared “welfare” or needs – and, on the other hand, some respect of the fruit of one’s labour with one’s productive capacities. The simplest way to implement the resulting structure of distributive justice is to add an “egalitarian income” equally sharing the product of the same partial labour (with different productivities) and a “classical liberal income” earned by each person choosing her own labour. On theoretical grounds, this structure happens to be the one resulting from all concepts of equal total liberty that respects Pareto efficiency (with the relevant social liberty). Practically, this implies a distributive tax based on productivity (wage rate). The actual realisation of an exemption of overtime earnings over a low benchmark shows a way of implementation and the possibility of such a tax (France since October 1st 2007, with de facto no cheating).1 Associated with a standard tax credit, this can realise simply the distribution in question.

12.1.3 Aspects of Implementation This shows how this fiscal system, with a clear distributive ethics in terms actually posed by society, and respecting social efficiency, can replace most actual ones, which induce waste and are not rationally build up from the social values, by reforms which are clear, simple to implement, and use fiscal

1

Wage income makes up 9/10 of all labour income. Wage rates are in pay sheets which are available legal contract documents. Other people can be taxed according to their categories, with all the usual methods of information of the tax authorities.

344

S.-C. Kolm

forms which actually exist. The properties of this distributive structure and further important complements answer the main questions that applications may face. People who participate in such a distributive scheme happen to be induced to work with their best possible skills (incentive compatibility). They have a minimum income related to the “equalisation labour” or degree of redistribution or social solidarity. Social insurances of all kinds have their own rationale and accompany this overall basic distribution. The same holds for policies of microjustice or mesojustice applied to particular situations. A number of these cases, however, – part-time incomes, unemployment compensations, etc. – can be brought back to the general case by the appropriate device, theory and policy. If average productivity and the degree of redistribution are too low for an equal sharing of the proceeds of the distribution labour to sufficiently alleviate poverty, this receipt can be redistributed more to the poorest – possibly by the particularly interesting social system proposed by Alain and Justin Leroux. Education is a dimension of labour and can be introduced in the general scheme in various ways, including public subsidies financed by the income tax (which taxes the enhanced productivity) as proposed and studied by David de la Croix and Michel Lubrano.2 The widespread free public education just does this, and, at any rate, the bequest aspect of education supported or induced by the family and the various social externalities of education make this field a main one for policies of mesojustice. Lubrano also provides the important analysis of the interferences of the general distributive scheme with specific policies. Capital taxation and the allocation of non-human natural resources are studied elsewhere (Kolm 1985), but Claude Gamel proposes to relate their logic to that of the distribution of the human resource, and an overall treatment of human and non-human resources is proposed in Sect. 12.3 below. The extremely rare very productive individuals who choose to work very little escape cooperative production and its distribution and raise issues of microjustice: they may be included in part-time labour contracts, demanded that they finance their own consumption (which needs little labour) and either persuaded to work or drafted when their particular skills can save lives. Even if they were provided the minimum or basic income and choose not to work, by exploiting their working fellow citizens, the resulting social waste would be limited as shown by Erik Schokkaert and Erwin Ooghe. In all cases, however, providing or receiving transfers can arouse social judgments and sentiments analysed in depth by Pierre Livet. This can, in particular, induce people who pay a distributive tax to try less to avoid it or to hide its basis (their capacities) because

2

Moreover, their overlapping generation model provides an example of a three-dimensional labour which shows, in particular, how the most direct form of multidimensional ELIE can be applied.

12 Macrojustice in Normative Economics and Social Ethics

345

of sentiments of “warm-glow”, solidarity, fairness or compassion, or a sense of duty. In so far as such sentiments and behaviour are absent and people can hide their wage rate, if the government – for some reason – does not use exemption of overtime labour earnings, takes the full earned income as tax base and maximises a social welfare function, then the problem is that of the “welfarist income tax”. Laurent Simula and Alain Trannoy show how such a case can nevertheless importantly relate to the distributive scheme obtained.

12.1.4 Method Knowledge always begins with distinguishing a very small number of variables and basic relations between them, and then it enriches this simple model with the various other relevant facts of reality. The initial framework is not criticised but completed, just as flesh and muscles do not criticise the skeleton but complete it. Scientific epistemology follows this path, just it does it systematically. In economics this gives, for instance, as initial basic simple (or simplistic) stylised structure, the Harrod–Domar model for growth theory, Hicks’s IS-LM, the simple Keynesian model, or even basic supply and demand, and the other sciences follow exactly the same pattern. The field of macrojustice is no exception, with the simple ELIE scheme as candidate to this structure with about the same very small number of variables and equations. Completion then came for issues such as: multidimensional labour; unemployment; checking the sufficiency or specifying the theory for very low or very high incomes (as with microphysics and cosmology for physical magnitudes); the allocation of the non-human natural resources; the methods of determination of the distribution coefficient (k) for each society; showing the possibility of basing the tax-subsidy on the required base; and the topics of the contributions of the various chapters of this volume with the solutions they propose which can only be approved and endorsed.

12.1.5 Structure: Self-Ownership and Solidarity; Freedom, Justice and Efficiency The basic structure should be on the right track, however. For distributive justice, this should be the right ethical and logical conception. The bulk of distributive justice occurs in macrojustice (Rawls’s “social justice”), and, there, the bulk of resources is constituted by the human resources – capital

346

S.-C. Kolm

is itself produced, and non-human natural resources are also studied (see Sect. 12.3). Human resources are roughly but sufficiently divided into capacities to produce or earn valued by a wage rate wi for individual i and capacities to evaluate often represented by utility functions ui .x/. The benefit from an element or aspect of a capacity can be allocated to its holder or to “society” and notably equally shared if there is no other relevant reason. All opinions about this can a priori exist. For earning capacities, the product of a given amount of labour can be equally shared, whereas that of the freely chosen rest of labour is self-owned – as is the use of these capacities when they are not used by labour. For utility functions, general opinion holds that differences in tastes, desires and capacities to enjoy are not valid reasons to affect income distribution in macrojustice. However, it also holds that comparisons of something called welfare can provide such a reason when the lowest of the compared incomes is sufficiently low. A corresponding comparable individual welfare function – the same for all individuals – can be obtained by averaging away individual differences with some chosen social welfare function (Sects. 3.3.1 and 12.2). The resulting individual welfare function can be the one commonly used (without explaining this choice) in classical “welfarist” taxation studies, for instance in Simula and Trannoy’s chapter in this volume. However, since this function differs from individuals’ utility functions, the result cannot be Pareto efficient (and this welfare function cannot describe individuals’ choices). Yet Pareto efficiency is achieved differently with the noted division of the value of earning capacities. Indeed, if earnings are self-owned for a freely chosen labour above a given amount (of labour or of earnings), there is no marginal distortion and free exchange classically secures Pareto efficiency. Then, the ideal welfare equalisation can intervene in the determination of the given (redistributed) part of disposable income only. This policy, however, requires distributing the value of earning capacities corresponding to given labours. And this is possible thanks to the noted large possibility to base taxes and subsidies on earning capacities. The result is the simple ELIE policy. It contains an a priori undetermined parameter, the equalisation labour k. This coefficient of distribution or equalization may depend on the equal welfare just noted but also on other considerations. Its level that can be said to be desired by the society which applies this policy can be determined by a series of methods (Part 4 of Macrojustice). The lowest level, k D 0, is full self-ownership, advocated by classical liberalism which is, therefore, a particular case of the moral philosophies described here. Then, earning capacities are fully self-owned. There is no relevant morally compared individual welfare because there is no relevant social welfare function comparing the levels of individuals’ utilities. The opposite case is that of the highest k – say k D k e – that gives an ELIE distribution consistent with free labour supply and democracy (Chap. 11 of

12 Macrojustice in Normative Economics and Social Ethics

347

Macrojustice). This is the solution desired by a non-dictatorial realistic egalitarian. This level is always lower than the lowest labour considered to be full-time. It depends on the society in question, regarding its conception and sense of various related notions: community, solidarity, equality, formal and real freedom, entitlement with respect to given assets, and duty to contribute to the collective production. This k D k e corresponds to Rawls’s position and – had he seen the possibility of distributive policies much less distortive than those he envisioned – he would think this level to be rather high (for nations) and reached in a “well-ordered society” with unanimous discursive approval.3

12.1.6 Concluding Closure: The Ethical Synthesis, Alternatives, Non-Human Resources, Unanimous Improvement, Desired Distribution, Types of Moral Motives The questions posed about macrojustice theory are now briefly answered. They divide into three groups. A basic one concerns its ethical foundation and, hence, the comparison with alternative theories based on different ethical principles. The synthesis between income justice, self-ownership and welfarism provided by ELIE is presented with greater precision (Sect. 12.2). The differences between the alternative theories depend much on distinguishing as relevant or irrelevant for justice various psychological properties represented by structures of utility functions. This goes from discarding individual differences in tastes, desires and hedonic capacities to obtain the same individual welfare function for everybody, to the ordinal welfarism of criteria in the family of equity-no-envy and of the “equivalence principle” (Sect. 12.6), and to co-ordinal comparable “fundamental utilities” (these are topics of Part 5 of Macrojustice). Preferences vanish from public endvalues in full income justice and self-ownership, and, at a macro level, in solutions that consider mesojustice (or microjustice) only – e.g. equalities in “spheres of justice” or of “opportunities” or “capabilities” (Sect. 12.5). The second issue concerns the allocation of given non-human resources. General solutions are analysed in Kolm (1985, 1986), noted in Macrojustice, but the principles that lead to ELIE (notably equal liberties) lead in various ways to equal sharing complementing ELIE and permitting a larger share of selfownership (Sect. 12.3). The third class of issues concerns the application, notably for low incomes and with the possibility to improve everybody’s

3

A theory of this concept is proposed in Kolm (2009b).

348

S.-C. Kolm

situation (Sect. 12.4); the ways of determining the impartially best degree of solidarity (Sect. 12.7); the provision of social justice as that of a nonexcludable universal moral public good and the motives that could permit it, which shows that the opposition of coerced and voluntary transfers is often a mistaken conception in many respects; and the important intrinsic value of solidaristic motivations (Sect. 12.8).

12.2 Situation in Distributive Principles 12.2.1 The General Structure of Distributive Principles: The “Justice Triangle” Economics divides human resources in capacities to produce and earn (the market price of their output is the wage rate) and capacities to “enjoy” or hedonic capacities (represented by utility functions ui .//. Labour produces by far the largest part of economic value, notably in an inter-temporal view in which capital is produced. Hence the allocation of human resources provides the bulk of the economic allocation in a society. Such a resource, or its value or output, can be allocated to its holder, or not. This makes, de facto, three polar cases represented in Fig. 12.1. In one case, each individual owns all of herself. This is (full) self-ownership, advocated by classical liberalism. In a second polar case, the individuals own their hedonic capacities but not, a priori and for a moral reason, the product of their productive capacities. This case can be called income justice. In the third polar case, no holder of a resource is a priori entitled to its value. The policy tools of distribution are transfers, notably of incomes, which, for hedonic capacities, provide the required compensations. The end-values are the values on which the ethical judgment about the distribution intends to apply. They are disposable incomes for income justice. However, each individual appraises her disposable income or the consumption it provides with her hedonic capacities. When all resources are available for distribution, the end-values are individuals’ levels of satisfaction, “happiness” and the like represented by their utility level depending on both their income (consumption) and their utility function. This case can be labeled welfarism (after John Hicks 1959). Moreover, individual incomes for income justice and “utilities” for welfarism have to be aggregated for defining the respective optimum distribution. This can be in the form of maximising some function of incomes or of utilities, respectively. Two polar cases are equality with the highest possible values and the highest sum. Equality can be described by a maximin or a leximin, which can be retained even when the constraints are such that it does

12 Macrojustice in Normative Economics and Social Ethics

349

Self-ownership

ELIE

yi

ui equality, max min

Income justice

Welfarism highest sum

Fig. 12.1 The justice triangle

not give equality (which implies that possible states with equality are dominated by other possible states in which some individuals have more and none has less – Pareto inefficiency for the case of utilities). Equality in income is a standard principle, and the maximin in income would be Rawls’s (1971) proposal restricted to this economic primary good. The leximin in comparable individual utilities which can be ordinal only (co-ordinal – a comparability justifiable by the theory of “fundamental preferences”) is eudemonistic “practical justice”.4 The second polar case is the highest sum. This is utilitarianism for utilities. The highest social income (sum of individual incomes) is implicit in proposals of having the highest growth rate, and the economist and judge Richard Posner (1977, 1981) proposes that it be the principle for all judicial decisions. There tends to be a correlation between the material of the principle of distributive justice and the size of the society for which this distribution is considered. Welfarism is a frequent case in small tight societies in which individuals know each other and have empathy towards her. An absence of distributive transfers and hence self-ownership is the standard case internationally, at world level. Finally, income justice is commonly demanded at

4

Kolm (1971). For the issue of comparability in some relevant domain, see also Kolm (1994, 1996a, 2005). Maximin requires only comparability between the worst-off and each other.

350

S.-C. Kolm

intermediate levels, for instance for nations (in which, however, the two other principles also intervene for various reasons).

12.2.2 Pure Welfarism and Beyond The solution obtained for macrojustice synthesises the noted three families of principles. It does it in a particular way, however. In the justice triangle (Fig. 12.1), this solution (ELIE) is represented by a point on the segment between income justice (egalitarian one) and self-ownership. It is a mix of both these principles, to an extent represented by coefficient k. However, one way of determining this separation refers to welfarism. This is explained in Chaps. 2 and 3 to which a few important precisions are added here. Section 3.3 shows that society holds differences in hedonic capacities and tastes to be irrelevant for the overall distribution in macrojustice. This can be interpreted in two ways (see Fig. 12.2). In one way, utility functions are discarded altogether (as Rawls does, for instance). This solution gives itself rise to two possibilities. Either the end-values become the arguments of the utility function, notably consumption goods (bought or valued by income) and leisure, or utility is taken to represent the principle of choice and the endvalues become the domains of choice valued for the freedom they provide. The second way discards only differences in individual utility functions, and the remaining identical individual utility function is taken to mean individual welfare, since comparing individuals’ welfare as if they were provided by the same function is a view which is actually considered (it is, for instance, one reason for preferring to give to a poorer than to a richer, or for favouring a transfer from a richer to a poorer when the lower of these two incomes is sufficiently low). This erasing of differences in tastes and hedonic capacities is achieved thanks to some Social Welfare Function W .fui g/, symmetrical function of comparable utilities ui of individuals i, non-decreasing and increasing in at least one argument at each state. Individual i’s allocation is xi 2 X and ui D ui .xi /. Denoting, for any number ˛, w.˛/ D W .e˛/ where e is a vector of n ones (n is the number of individuals), function w is increasing. Then, the individual welfare function is function u.x/ defined as, for each x 2 X , (12.1) u.x/ D w1 ı W Œfui .x/g: Individual i’s strict (or pure) welfare is u.xi / and the corresponding welfarist maximand is

12 Macrojustice in Normative Economics and Social Ethics

351

u = ”objective midfare”, or ”strict welfare” Strict welfarism W [ {u(xi)}]

super-equality etc.

difference averaged away

macro-irrelevances Full welfarism ui(xi), W [ {ui }]

discard ui e.g. Rawls

Equal income yi Rawls (1971) (yi , λi ) Rawls 1974 (yi , i )

liberty Pareto efficiency some self-ownership

ELIE Equal liberty

Fig. 12.2 The graph of solutions

W Œfu.xi /g D W Œfw1 ı W Œfuj .xi /gj g:

(12.2)

This is a function of all individuals’ evaluations of all individuals’ allocations uj .xi /. One can show that, in the two polar cases of “equity” defined as ui .xi /  ui .xj / for all i; j (sometimes problematically called “no envy”)5 and of “adequacy” defined as ui .xi /  uj .xi / for all i; j – the ui being comparable – (Kolm 1971), W Œfu.xi /g  W Œfui .xi /g: The difference between both values means that some part of the individuals’ satisfaction – say that which is due to their sui generis tastes or hedonic 5 The description of the sentiment of (strong) envy requires the consideration of other utility functions, of the form Ui .xi ; xj /, with Ui .xi ; xj / < Ui .xi ; xi /  UQi .xi / expressing the painfulness of this sentiment and where UQi .xi / so defined represents individual i ’s “envy-free preferences” (see Kolm 1995).

352

S.-C. Kolm

capacities rather than to their individual welfare as defined here – is left to the accountability of the individuals rather than counted in the purely welfarist social evaluation. With adequacy, the replacement of all functions ui by the “average” u for evaluating xi lowers the value left for the social evaluation. And, with equity-no-envy, replacing in form (12.2) the uj .xi / by uj .xj / to obtain W Œfui .xi /g augments (does not reduce) the social value. This function u of individual welfare can then be used for standard welfarist studies, for instance in the field of taxation such as income taxation with xi D .yi ; `i / where yi and `i are individual i’s disposable income and labour respectively, or for describing Pigou’s (or Bentham’s) utilitarianism as the highest ˙ u.yi / – with an application in the classical theory of inequalities.6 Figure 12.2 summarises the rest of the discussion proposed in Chap. 2 and 3. We have seen that an ideal of equal xi , rather than equal u.xi /, is socially preferred (gives a higher value of W ). Then, for xi D x for all i, the same solution obtains by maximising u.x/ or the uncorrected maximand W .ui /, from relation (12.1). For instance, with x D .y; `/ with income y and labour `, this solution maximises u.y; `/ under the distributional constraint y D w` where w is the average wage rate. Equality in income and labour can also be an egalitarian ideal by itself. It is, for instance, that of Rawls after 1974. However, people prefer, from this state, to work more and keep their extra earnings, hence this state is not Pareto efficient, and, in addition, society usually values some amount of respect of classical liberalism. This leads to this free choice from the obtained egalitarian distribution, with various possible divisions of incomes in these two parts, depending on the initial choice of function W .

12.3 Macrojustice and Given Non-Human Resources: Self-Ownership-Enhancing Affirmative Action 12.3.1 Non-Human Ressources The simplest, core presentation of the theory of macrojustice focuses on human resources and omits non-human ones for a reason of relative overall importance in economic value (and of practical possibility). Moreover, the various general possible principles for allocating non-human natural resources are analysed in depth in the volume The Liberal Social Contract (Le Contrat Social Libéral), Kolm 1985, Chap. 10; and in Kolm 1986.

6

Kolm (1966).

12 Macrojustice in Normative Economics and Social Ethics

353

However, the distributions of human and non-human resources have interfering effects. This may be used to solve some problems faced by these distributions. For instance, one may be insufficient by itself in possible amount and allocation. Or the availability of non-human resources may permit to support people with low earning capacities without transfers of the value of the capacities of able people, that is in the respect of full selfownership. The association of ELIE macrojustice with the allocation of non-human resources is also the topic of the chapter of Claude Gamel in this volume. Including capital, as he does, or keeping to natural resources is a question discussed in Sect. 2.5.6. When a more or less extensive part of capital should be included in the given resources for the reasons noted there, this can be added to the natural resources in the analyses.

12.3.2 The Solutions In the various solutions studied for the allocation of given non-human resources there are, in particular, maximin or leximin in comparable utilities or welfare (“practical justice” in “fundamental preferences”), equal sharing, and unanimous agreements (notably putative ones, that is, “liberal social contracts”). The school that calls itself “left-libertarian” (“classical liberal” should replace “libertarian”)7 defends full self-ownership and an equal sharing of non-human natural resources (see for instance the references to Peter Vallentyne’s and Hillel Steiner’s works in Chap. 2). The term “left” is here to distinguish equal sharing from the appropriation of non-human natural resources by first occupancy. This equal sharing can be justified by the noted rationality of equality. However, self-ownership is itself an equality in the sense that one person owns one person, but the persons so owned are very different notably in capacities to produce or to enjoy. In particular, the emphasis on social liberty for justifying self-ownership suggests the consideration of total freedom. With such an equal sharing of given nonhuman resources, and self-ownership, more productive people have more total freedom by inclusion of domains: they can have more income and consumption for the same labour, or the same income and consumption with lower labour. However, with given non-human resources available for distribution, it is possible to remedy more or less this inequality in total freedom while keeping full self-ownership. It suffices to give more of the value of these resources to people with lower productive capacities. This policy is 7

“Libertarian” was initially used, in scholarly English, by theories in which the relation between people is force or a balance of force rather than rights, which is the opposite (e.g. Rothbard 1973).

354

S.-C. Kolm

more “leftist” than classical equal-sharing “left-libertarianism”. It uses the non-human resources for the “affirmative action” of giving more of them to people less endowed with productive capacities by full self-ownership. The inequality of freedom of choice by inclusion of domains provided by different capacities or wage rates wi is maintained by an equal sharing of the given non-human resources but can be remedied by giving more to the individuals with lower wi . If equal liberty provided by these domains is defined by the properties described in Sect. 3.5, it will turn out that the share of given non-human resources received by each individual is an equal share plus the net transfer of an ELIE scheme (a positive or negative transfer according as wi 7 w). Another approach would be to consider lump-sum transfers of any sign with a net sum equal to the total value of the given non-human resources. The same theory of equal liberty yields the same result with the difference that some people may now be net payers. These individuals no longer enjoy full self-ownership. Full self-ownership for all is respected if the coefficient k of the obtained ELIE is sufficiently limited for the ELIE tax paid by the most productive individuals to fall short of the equal share of the given non-human resources. These results are now shown.

12.3.3 Compensatory Sharing Denote as R the value of the given non-human resources, r D R=n the per capita value and ri the part allocated to person i, with ri  0 and ˙ ri D R. The allocation with compensation for lower wi is with ri D f .wi / where f is a decreasing function. As in Sect. 3.5, and for the same reasons, the various concepts of equal total freedom imply that the individual budget lines have a common point. Individual i’s budget line is yi D ri C `i wi . Then, there should be k D `i and  D yi such that  D ri C wi k for all i. Adding and dividing by n give  D r C kw. Hence ri D r C k  .w  wi / for all i. This amounts to the superposition of an equal sharing of R and of an ELIE distribution with coefficient k. The case k D 0 is that of the “leftw D libertarians” (for natural resources). ri  0 for all i implies, noting b max wi , r C k  .w  b w/  0, hence k  r=.b w  w/ (if the wi are not all equal). Such solutions are those of “equal-freedom left-libertarianism”. The w  w/ and therefore most “egalitarian” in a sense is with k D k o D r=.b w  w/ D r  .b w  wi /=.b w  w/ ri D r  Œ1 C .w  wi /=.b for all i, with ri D 0 for wi D b w.

12 Macrojustice in Normative Economics and Social Ethics

355

12.3.4 Independent Superimposition More generally, macrojustice with given non-human resources should be considered a priori. The principle of distribution may be some equal sharing, from a requirement of rationality (because no relevant characteristic differentiates people for this allocation), but the basis that has to be shared has to be chosen. This basis can be the value of the non-human resources. For standard ELIE, it is the product of everybody’s labour for the “equalisation labour” (k). Both can be associated. The result respects full self-ownership if the per capita value of the non-human resources suffices to compensate the ELIE largest payment, that from the most able person. But society may also choose otherwise.

12.3.5 Integrated Distribution Moreover, this dual equal sharing may be a consequence of the single principle of equal liberty (with its several definitions that give the same structure). Let Ti denote the total net lump-sum allocation received by individual i, from the allocation of R and from interpersonal transfers (Ti < 0 denotes a tax of Ti ). Since interpersonal transfers cancel out, ˙ Ti D R. If this allocation intends to compensate more or less the inequalities in total freedom due to the differences in earning capacities, one should have Ti D g.wi / where g./ is a decreasing function. Then, the various concepts of equal total freedom presented in Sect. 3.5 imply, for the same reasons, that the budget lines of individuals i, the equations of which are yi D Ti Cwi `i , have a common point `i D k; yi D  for all i. Then  D Ti C wi k for all i and adding and dividing by n gives  D r Ckw, and therefore Ti D k .wwi /Cr. That is, each individual i receives her net allocation ti D k  .w  wi / in an ELIE distribution with coefficient k plus an equal share of the value of the given non-human resources, r. This is a superposition of these two distributions.8 Pure ELIE is the case r D R D 0, and the sole equal distribution of the nonhuman resources is the case k D 0 (advocated, notably, by “left-libertarians” for natural resources).

8 All these cases of addition of the same given income r for everybody to the incomes of an ELIE scheme are represented in Figs. 3.1 and 3.2 by an upward translation of the whole figure by r.

356

S.-C. Kolm

12.3.6 ELIE and Full Self-Ownership However, with k > 0, k  .w  wi / < 0 if wi > w, and one may a priori have Ti < 0. This is a possibility, with a tax of Ti . But an interesting property occurs if Ti  0 for all i. Then, indeed, nobody pays a tax. This implies full self-ownership (in spite of the ELIE part of the distributive program). w/ C r  0 for b w D max wi and therefore k  This occurs when k  .w  b k o D r=.b w  w/ if the wi are not equal. Then, the highest ELIE distribution (highest k) that respects full self-ownership is with net transfers w  w/ D r  .b w  wi /=.b w  w/: Tio D r  Œ1 C .w  wi /=.b Since Tio  0 for all i, ˙ Tio D R and Tio D 0 for wi D b w, this can be seen as a distribution of the given non-human resources that gives more to less productive individuals and zero to the most productive one (as in a previous case that considers the allocation of given non-human resources only). Particular cases are w; Tio D 0: wi D b wi D w; Tio D r: w/: wi D 0; Tio D r=Œ1  .w=b For this wi D 0, if b w is much larger than w, Tio ' r. The Tio for low wi may provide a sufficient minimum income, thanks to r > 0. If this is not the case, either a somewhat higher k is chosen and individuals i with wi > w C .r=k/ pay a tax, or another policy measure is introduced, aiming specifically at avoiding poverty, and it has to be financed. For the general case allowing some tax, one has Ti  0 and hence full self-ownership for all individuals i with wi  w C .r=k/.

12.4 Application The modalities, possibilities and effects of applications of the analysis of macrojustice are largely investigated in the volume Macrojustice (2005). Some practical applications resulted immediately, such as switching the base of the income tax to individual productivity instead of labour income in one country (France) by an exemption of overtime labour earnings over a rather low benchmark, with the corresponding de-elasticisation with regard to labour. The case of low incomes has been particularly studied. Simulations have shown that “An appropriate ELIE scheme can replace at once

12 Macrojustice in Normative Economics and Social Ethics

357

low-income and unemployment subsidies and minimum wage laws, by making everyone better-off”.9 More generally, for each actual transfer system there is a coefficient of equalisation such that the replacement of the relevant transfers by the corresponding ELIE scheme makes everybody better-off. Full ELIE schemes do not induce Pareto inefficiency, contrary to present transfer systems, and each of the latter is Pareto dominated by some of them. However, a reform may, in addition, desire to modify the intensity of the equalisation. An issue, of course, is which transfer measures should be replaced. A number of policies and arrangements have their own logic and hence remain. This is notably the case of insurances of all kinds, including social insurances (this should also include cases of “fundamental insurance”, that is, putative insurance against a given handicap, when they are generally approved, for instance that included in public health insurance with premia independent of health status in various countries). Another example is provided by family allowances, important in some countries, the interaction of which with an ELIE distribution is presented by Michel Lubrano in his chapter. Another possible implementation policy studied in the volume Macrojustice consists in a progressive introduction of an ELIE scheme, beginning with a low coefficient k and augmenting it progressively, and correspondingly reducing and suppressing other general supports of low incomes in so far as they become superseded by the new distribution. When social productivity and the distributive coefficient k are high enough, the minimum income kw may be an acceptable level. If this level and the measures palliating specific causes of poverty such as the various social insurances are not sufficient, other particular aid, in the realm of microjustice, has to be added. It belongs to the category of assistance, however, a rationale very different from that of the rights to the allocation of the benefits from given resources, notably concerning their effect on human dignity.

12.5 Mesojustice, Alternative or Complement to Macrojustice Given the proper income distribution, social liberty implies that people are free to spend their income as they wish by exchange (or to give it). Forceful interferences with these processes are justified by failures of some kind, and their proper rationales are often the implementation of some hypothetical free action or exchange hampered by the cause of this failure.10 These causes 9

Pages 120–122. This reproduces conclusions from a study of the General Planning Commission about the application of an ELIE transfer system, see Kolm (2002). 10 A “liberal social contract” (see Kolm 1985, 1987, also Kolm 1996a and Kolm 2005).

358

S.-C. Kolm

for interference are classical: externalities, non-excludable public goods, macro-economic problems, lack of information, transaction costs, etc. One particular “good” cumulates reasons to be the object of particular policies, education. It is, in fact, jointly an input of labour, an element of bequest by parental inducement and support, a source of externalities from parents to children and from the educated persons to the rest of society which interacts with them and to the social culture and civilisation in itself, and an issue of formation and information which may require vicarious interference. The other dimension of human “capital”, health, is a question of both insurance and allocation of a given human natural resource or liability, the propensity to be sick or basic health. Insurance with uniform premia (not related to health status, hence not a purely private insurance) amounts to an equal sharing of the part of this natural liability in the cost of care for the corresponding coverage. Education and health are the two main domains of issues of “mesojustice”. With the proper income distribution, concerns about specific consumption not justified by some relevant failure of free choice violate social liberty. A particular bundle of goods is appropriate for determining whether the minimum income is sufficient (with the overall distribution, the various insurances, etc.) or should be completed by the appropriate policy, in relation with concepts of basic needs (or “capabilities” as used by M. Nussbaum).11 But demanding a priori an equality in each of various consumption goods, as sometimes advocated, is problematic. It is, indeed, de facto not Pareto efficient – people have different tastes, preferences and needs –, which implies waste and a possible instability with regard to social and political processes. One may, then, define the least unequal of the Pareto-efficient allocations of these goods (using the theory of multidimensional inequality).12 The result consists of the overall allocations such that no individual prefers any individual allocation in the convex hull of the actual individual allocations to her own – the property of “super-equity”, now sometimes called “convex no-envy”.13 Then, however, if all individuals consume some amount of each of the goods in question, this result implies an equal income (with the “efficiency prices”).14 The conclusion is that such multidimensional equality (as proposed, for instance, by A. Sen for “capabilities”) could hardly be a proper “intermediate” between welfarism and income justice (as Rawls’s).

11

Nussbaum (1992). In Kolm (1977). 13 Kolm (1973). 14 Kolm (1996b). 12

12 Macrojustice in Normative Economics and Social Ethics

359

However, people sometimes value morally equality in some goods. Following Max Weber, Michael Walzer (1983) even argues that justice is equal allocation in each of specific “spheres of justice” (surprisingly, the most extreme in this respect is an economist, James Tobin, who in 1970 advocated equality for each good!). Yet Walzer admitted adjustments giving somewhat more of a good to people who value it particularly. Nevertheless, such specific egalitarian ideals exist for some goods. These “sphered goods” can include notably health, educational opportunities and some basic education, and the availability of various services. These are issues of microjustice and mesojustice.15 An aspect of specific equality is found in the common demand for equality of opportunity, which can apply to various goods. This principle can actually mean many things and, as a result, all politicians favour it and all philosophers severely criticise it.16 A basic issue is that, when the acting people have different capacities, identity of possibilities a priori implies a violation of Pareto efficiency, because it is de facto performed by an interference with individuals’ freedom of choice distorting their marginal choices (for example, for labour earnings, this would consist in applying the same wage rate to people with different productive capacities). This refers to an ideal of deserts (to each according to her effort). This diriment shortcoming of equality of opportunity constitutes a most important example of opposition between justice and efficiency (and hence unanimity). The solution rests on the alternative definition of equality of opportunity not as identity of domains of choice but as equality of liberty offered by adequately different domains of choice. Then, it is an ideal of partial merit, i.e. merit from some benchmark – merit is to each according to her effort and capacities. This is precisely the logic used in macrojustice for the opportunity to work and earn by defining equal liberty for domains of choice different for people with different productive capacities (Sect. 3.6).

12.6 Ordinal Welfarisms and Their Relations to Macrojustice Macrojustice analysis begun with the observation that some psychological aspects of what individual “utility” intends to mean are actually considered irrelevant for this topic, with the conclusion that some structural properties of utility functions have to be discarded for this application. However, the most

15 16

See, for instance, Kolm (2001a) for the case of health. See Kolm (2001b, 2010).

360

S.-C. Kolm

classical view about discarding aspects of these functions holds irrespective of and beyond this specific application: this is the conception that individual preference ordering and ordinal utility alone have meaning. The standard view holds moreover these utilities not to be interpersonally comparable. The concept then corresponds to what can be used to “explain” choices. This leads to two families of allocative principles. One uses individuals’ preferences between the allocations of various individuals (including themselves) or individual allocations derived from them (equity-no-envy is the most common of these structures). The other finds social states between which all individuals are indifferent to be equivalent (the equivalence principle).17 Each individual preferring her allocation to that of any other is equityno-envy: no individual prefers any other’s allocation to her own.18 The overall allocation obtained for macrojustice does not satisfy it because of the classical liberal part of the obtained incomes. However, it satisfies the meaningful and more interesting variant of realistic equity: no individual prefers any other’s allocation that she can have to her own (Kolm 1971). This is simply due to the fact that the agents choose their allocation in their possibility set, it is just a property of freedom of choice. Moreover, individuals’ evaluations of others’ allocations, and in particular equity-no-envy, appeared as a result of erasing individuals’ differences in tastes and hedonic capacities (Sect. 12.2.2). And a variant of this principle, super-equity (no individual prefers any allocation in the convex hull of individual allocations) had a role as the least unequal efficient multidimensional allocation (Sect. 12.5). The most remarkable property of equity-no-envy is that it is a principle of equal liberty in the sense that it holds if and only if there exists a domain of free choice, the same for all, such that each individual’s allocation could be her choice in this domain (these domains are the set of actual individual allocations plus any possible individual allocations that no individual prefers to her own). This property, or, directly, individuals’ comparisons of others’

17

Theories of optimum income distribution and taxation based on such concepts are found notably in Bös and Tillman (1985), Kolm (1991) and Fleurbaey and Maniquet (1996) for equity-no-envy, and Fleurbaey and Maniquet (1999), Luttens and Ooghe (2008) and Chap. 25 of Macrojustice for the principle of equivalence. 18 In 1921, the Dutch physicist Paul Ehrenfest – a co-author of Albert Einstein – was asked by an 18-years old student of his named Jan Tinbergen what a fair wage is. After some reflexion, he answered: “If I prefer to do my job with my wage to doing your job with your wage and you similarly prefer your own situation to having mine, this is a fair wage” (private communication in Paris in 1962). Tinbergen notes this criterion in his book (in Dutch) of 1946. Duncan Foley (1967) proposed the same notion for consumption goods. The analysis of this principle was carried out in Kolm (1971), Varian (1974, knowing the previous work) and an abundant further literature. A number of related concepts, properties, variants and applications have been proposed (see a survey in Thomson 2008). The theory of actual envy is proposed in Kolm (1995) with the concept of envy-free preferences and the relation with the above “equity” principle.

12 Macrojustice in Normative Economics and Social Ethics

361

allocations to their own, extends to defining more or less free (by inclusion of putative domains of choice) and, with efficiency, to the definition of a maximin in this freedom (Kolm 1999). This efficient second-best equity-no-envy and the restriction to “realistic” equity are useful notably because a main shortcoming of equity-no-envy is that it may not be consistent with Paretoefficiency (Pazner and Schmeidler 1974, for instance). This opposition is at the heart of the problems of optimum taxation and equality of opportunity. The same tax schedule for everybody’s full earned income, or imposing the same wage rate to everybody, as well as identity of opportunities, give identical domains of choice, hence equity-no-envy of the outcome, but they entail Pareto inefficiency. This criterion is also a basis of the issues of incentive compatibility and implementation – for instance in the chapter by Simula and Trannoy in this volume. The equivalence principle was introduced to reduce a multidimensional inequality to a one-dimensional one – an inequality in consumption goods and labour to an inequality in equivalent incomes for a given labour – (Kolm 1966) and to find possible (and efficient) states equivalent to an equality in all goods (“egalitarian equivalence”, Pazner and Schmeidler 1978). It is used to provide a variety of solutions of the problem of overall allocation which are discussed in Chap. 25 of Macrojustice. However, denote as x a social state, say that two social states are equivalent if all individuals are indifferent between them, denote as P the set of possible – or Pareto-efficient – social states x, as ˘ the set of social states each equivalent to any of the states of P (we have ˘  P ), and, for any set X of social states, as C.X / a social state selected in X by some criterion C . The equivalence principle demands one to choose a state x 2 P equivalent to a state x 0 D C.˘ /. However, this x is in general different from C.P /. This appears a priori as a serious conceptual problem for the equivalence principle. Nevertheless, this principle is usually applied for criteria C such that C.P / does not exist. For instance, this criterion is an equality in several goods and this is impossible or not Paretoefficient. Another solution would be to extend the criterion C to a criterion C 0 which has a solution on P (C 0 .P / exists), and if C 0 is an extension of C for which C.˘ / was found good or the best in ˘ , then C 0 .˘ / D C.˘ /. Then, however, C 0 .P / is a priori not equivalent to C 0 .˘ / D C.˘ /, and this normally whatever the extension C 0 of the properties of C found valuable. For instance, if C is an equality, C 0 could define an inequality ordering or measure and C 0 .X / would be the (or a) x 2 X with the lowest inequality. Then, the least unequal x D C 0 .P / is not equivalent to a state of ˘ with equality, whatever the chosen ordering or measure of inequality. Saying that some person is happier than some other is sometimes not meaningless (for instance one is happy and the other is unhappy). This leads to the comparison of individuals’ utilities by smaller or larger in the relevant

362

S.-C. Kolm

domain. This is the minimal structure19 that permits to define a symmetrical social welfare function W .fui g/ (as in Sects. 3.3.1 and 12.2.2), and a leximin or maximin in utility.20 This maximin is relevant when there is a suffering minority – it is then not difficult to point out its members – and when the policy can sufficiently improve its situation.21 ;22

12.7 The Degree of Distribution, Community, Reciprocity and Protection Welfarist income taxation and other studies have to make precise the social welfare function they use. The same holds for the much simpler coefficient k of an ELIE policy. In fact, this coefficient can be derived from a social welfare function – the function that averages differences in tastes and hedonic capacities away as society demands for this application – as presented in Sect. 3.3.1. Note that the frequent idea of using an arbitrary social welfare function just to obtain properties specific to Pareto-efficient states is itself somewhat problematic since such states can also be obtained by not interfering with the agents’ choices of marginal amounts. Coefficient k – or its multidimensional equivalent – is, we have seen, the “equalisation labour or duration”, a degree of community of resources, of solidarity with respect to “natural” endowments, of general labour reciprocity (“what we owe to each other” in T. Scanlon’s terms) and of concentration of total incomes (leisure included), and minimum income as a share of average productive capacity. This coefficient, and its value for the “equivalent ELIE” for other distributive structures (cf. Sect 3.8.1), manifest an essential property of the distributive society under consideration: the extent to which it is a community rather than a meeting of individuals only (a bag of potatoes, as Marx puts it),23 with respect to economic resources. Actually, present-day nations redistribute a large part of their social income (30–35% in most European countries and up to 40% in Scandinavian 19

“Fundamental” preference or utility in Kolm (1971). “Practical justice” in the same reference. Maximin solely requires comparisons between the worst-off and others. 21 In Kolm (1974), the maximin in comparable utility is applied to optimum taxation with the analysis of the various tax bases (incomes, given wage rates – with the suggestion to exempt overtime labour earnings – , etc.) and of the various reasons that lead to inequality (disincentives, savings and thus investment and growth, inflation, risk-taking, education, social organisation, etc.). 22 Remember that this is not Rawls’ “difference principle” which is a maximin in “primary goods”, yet Rawls (1982) discusses interestingly the concept of a fundamental preference. 23 In The 18-Brumaire of Louis-Napoleon Bonaparte. 20

12 Macrojustice in Normative Economics and Social Ethics

363

“social democrat” national communities). These orders of magnitude happen to be largely accepted. This is the result of long-term national histories, with jumps in levels of transfers occurring as a consequence of the major national traumas (wars and economic crises).24 The determination of what the degree of fiscal redistribution should be cannot avoid considering citizens’ opinions in this respect. Indeed, society realises this policy, considering citizens’ view is a condition of democracy, and these opinions may a priori have moral value. However, people are of two minds – at least – in this respect. On the one hand, their self-interest (and that of the persons they like) are affected by the paying or receiving. On the other hand, social individuals have the capacity to take an impartial viewpoint, abstracting from their own particular situation, which provides their objective moral social opinion (the “view from nowhere” as Tom Nagel puts it – but is there such a place as nowhere?).25 Knowing the latter view, however, requires bypassing expressions influenced by interests. Beyond the classical means found in political expressions or inquiries (in which the opinion of a single small agent has no actual influence), the particular ELIE distribution provides two technical ways of erasing self-interests from people’s expression. People’s opinion about the distribution is assumed to derive from their self-interest and from their social moral judgment. One method consists in demanding the opinion of people with average wage rate: since individual i’s net transfer is ti D k  .w  wi /, for individuals i with wi D w this is ti D 0 whatever k, and therefore the choice of k does not affect these people’s self-interest. Hence these persons’ opinion about this level reflects their social moral judgment only (and this may be taken as a sample for the whole population). The second method, the “moral or distributive surplus”, applies to any set of distributive transfers (the only property used is ˙ ti D 0).26 The surplus of an item is the algebraic sum of its money values for the agents or of their willingness to pay for it (these two values are a priori different).27

24

Typically, after the world wars the expenditures of the swollen war budgets are switched in large part to aid, notably to the victims of wars (in particular to veterans and their families). The collective national defence itself pushed up the sense of national community. In the US, public aid exploded during the Great Depression and this is the origin of its present level (relatively moderate compared to other countries). 25 T. Nagel (1986, 1991). 26 See Kolm (1966, 2005). The former reference fully considers the possibility of satisfaction derived from other people’s satisfaction and general equilibrium. The second also considers other types of surpluses such as the egalitarian surplus. 27 They are money equivalents evaluated in two different states, the present state and the projected state, respectively. This may raise classical Scitovsky’s inversion problems – related to the questions of the “principle of equivalence” noted earlier – , but this is not an obstacle when the solution chosen is the “surplus optimum” shortly noted.

364

S.-C. Kolm

The part of this value that measures self-interest is the amount of the transfer itself (taken with its sign). These values globally cancel out in the sum. Hence this surplus is an aggregate of individuals’ social moral evaluations. Moreover, this is applied to determine the corresponding optimum as a distribution such that no surplus from another distribution to it is negative and no surplus from it to another distribution is positive (money value and willingness to pay are inverted in these two operations). This aggregation is a particular one, yet also a classical one (Jeremy Bentham, for instance, says both that utility should be measured in money “for lack of a better measure or we should bid adieu to moral” and that “the pleasure that people derive from other people’s pleasure” should be included). This method works even if the self-interested values are much larger than the moral ones (including “lexical egoism”). Voting rules can approximate the surplus with usual distributions of individuals’ evaluations. A number of other methods of determining people’s impartial views, aggregating them when they differ and determining the resulting degree of distribution are presented in the most important Part 4 of the volume Macrojustice (2005).28 However, these moral political opinions are not really individual phenomena but social ones. People generally adhere to an existing position, for instance by supporting a political party or voting for a political program. These social options result from long and permanent political and moral debates. Moreover, an individual’s choice depends on the particular influences she happens to be submitted to, her various life experiences and her particular sensibility (and, possibly, rationalisations of her self-interest, from unconscious ones to Max Weber’s pharisianism). These causes of moral opinions are not themselves moral (the influence of people’s moral reasoning only is). Moreover, these opinions also deal, by nature, with what other people should have or do: they are externalities, in a reciprocal way. Therefore, they differ from preferences about consumption which may a priori have to be “respected” as they are. These problems subside or vanish if people sufficiently know others’ information, reasonings and life experiences. They are thus induced to adopt the logical parts of these reasons, and to take others’ experiences into account if they are sufficiently vividly described to them (they should not only know about them but have 28

Theories of the “original position” and of “moral time sharing” are people’s hypothetical consideration that they could or will become any individual in person and situation as alternative possibilities or successively in time (with equal chances or durations, or emphasising the particularly bad cases). These theories are handicapped both by the problem of assimilating self-interested choices in such conditions to choices of justice, and by the fact that different people in these hypothetical situations have different preferences about being in the various cases and about either risk or the aggregation of experiences at different times (see Kolm 1996a, 2005, 2009a for the analysis of these questions and possible answers to them).

12 Macrojustice in Normative Economics and Social Ethics

365

at least some notion of how it could feel to experience them). Criticism from others can induce them to correct mistaken reasonings. Even others’ different sensibilities tend to be taken into account by the universal sentiment of empathy when they are sufficiently known. This leads to the consideration of dialogue. Ideally, the notional opinion taking account of all relevant experiences and reasons provides the solution. Dialogue normally fosters the convergence of opinions, possibly towards consensus. It is, however, more or less hampered by the costs, time and capacities required to transmit and use information. The consequences of these limits can be more or less remedied by the social organisation of the dialogue and by its modeling which can show its likely outcome and thus hasten the process. An essential point is that the proper dialogue is not just influence but transmission of basic information which induces and justifies opinion (if it were just influence the obtained consensus would a priori depend on the initial prejudices based on partial information). The virtue of dialogue is enhanced by a number of qualities that public policy can promote, such as truthfulness (including sincerity), responsiveness to others’ arguments, and, notably in so far as interests can also influence expression, a balance between views possibly implemented by an equal right to public expression (the Athenian isegoria, deemed to be a main condition of democracy, and applied as an equal time of speech in the agora). Consensus permits to realise what each person wants and democracy, in peaceful relations. From the epistemic point of view, however, it is achieved by both the best solution and truth, with perfect information, but it does not guarantee them short of this condition.29

12.8 The Motives of Solidarity 12.8.1 Wanting to be Forced to be Free Actual judgments, motives and private and public actions about the distribution constitute an aspect of societies which is both massive and rich in psychological and social subtleties (and paradoxes). Two facts are basic. One is the psychological duality of individual citizens who are both selfish and capable of impartial and moral judgments – as Pascal puts it, “man is neither 29

Other aspects of the ethic of dialogue are provided by Ackerman, Perelman’s “rhetoric” and Apel’s and Habermas’ “discourse” or “communicative ethics”, and dialogue between people with “a shared political culture” secures the “stability of justice” in Rawls’s description of the “wellordered society” (Rawls 1971, Chap. 7).

366

S.-C. Kolm

angel nor beast”. The other is that their concern about social justice or others’ needs is a collective concern, a public good for them. More precisely, it is a universal (everybody shares more or less such sentiments), non-excludable and moral public good. As a consequence, some forced transfers can make everybody better off – and hence this obligation may result from a more basic, logically prior and possibly hypothetical agreement. This may result from the non-excludability property of the public good of caring about others or justice, for checking the corresponding free riding. Moreover, individuals may also accept or value this obligation in so far as it permits them to do their moral duty in spite of the resistance of their “selfish self”, to be actually as good people as they think they should be (given that other people are also similarly forced to contribute) – as Ulysses tied by taxes to the mast of moral goodness to resist the call of the sirens of self-interest. Actually, the simple lack of information limits individuals’ possibility to relieve deprivation or injustice at best by direct donations (charitable private organisations have some views of needs, but often a partial and biased one, and they lack the general politically built relevant conception of distributive justice; collective local entraide can have this latter dimension and the factual information at its limited level). The lack of information also extends to knowing the other numerous co-givers and, along with “transaction costs”, prevents direct agreement about contributing. Hence both the information about the object and the unanimous free agreement are vicarious. Such a putative agreement is by definition a social contract. However, even actual contracts are implemented thanks to public constraint. All these reasons make freedom take the form of public constraint.

12.8.2 A Hierarchy of Moral Motives for Joint Transfers Nevertheless, the ethical aspect of both cooperation in itself and the distributive objective can influence the realisation which can be more or less cooperative, moral and voluntary in various possible ways.30 The provision of a public good may be the realisation of a putative agreement between the people concerned, and the most classical social ethical theory is that of the social contract which consists in hypostasising such an agreement into the social ethical obligation of its would-be outcome. Moreover, cooperation

30

The large number of relatively small contributors prevents voluntariness from resulting from the repeated character of contributions only, since an individual free riding or contributing, as well as another individual’s abstention to punish a free rider, are not even perceived – the latter also “punishes” everybody.

12 Macrojustice in Normative Economics and Social Ethics

367

often rests on an ethic of “lateral reciprocity”, that is, each individual voluntarily does her share, in particular gives, given that the others also do – this is a conditional moral motive, not a “selfish-like” conditional exchange with the others (this reciprocity is “lateral” because standard reciprocity is that of direct gifts between the actors). However, in order to be sure that the others contribute, it may be necessary that they are forced to. But this constraint on all participants is not actually binding (it forces each person to do what she wants, but she wants it because the others do their share). This kind of motive accounts for the high level of voluntary tax compliance in some countries (actually those that pay the highest distributive taxes). This obligation can be avoided by sequential contributions, with some people who contribute unconditionally at the beginning. However, this does not fit for a steady regime of distribution. Another solution is for people to go one step more in ethical conduct and to contribute as if other people similarly contributed too. This is a classical motive, revealed by expressions such as “what if nobody did it” which is, for instance, the main “reason” people give for voting in large elections in which a single vote actually makes no difference, and this is the motive philosophically hypostasised by Kant into his more refined “categorical imperative”. However, in both lateral reciprocity and the categorical imperative, Pareto-efficiency for individuals’ preferences including their moral values requires particular relations between individual contributions.31 A minimal conclusion is that the issue of coercion or voluntariness in transfers, and particularly in collective transfers, is by no means clear-cut. There is a number of cases, motives and subtleties on psychological, social and logical grounds.

12.8.3 Beyond Interest and Justice: The Virtue Ethics of Intrinsic Value Moreover, motives are not valuable as causes and means of social realisations only. They constitute essential values in themselves, by determining the intrinsic quality of actions directed by them and of people moved by them. This goes beyond public actions just discussed. In particular, a main manifestation of social liberty is standard exchange between self-interested agents, the constituent of markets. Although this peaceful interaction has this particular advantage over the use of force, fights and threat, it is not, in itself,

31

See Kolm (2009b).

368

S.-C. Kolm

the best type of human relations guided by the most intrinsically valuable motives. Social liberty is also respected by benevolent giving, by relations of genuine reciprocity that extend the possible scope of transfers realised by gifts, and by domains of gratuity which allow people to find some respite from the debasing grasp of economic motivations and the permanent clash of opposed interests.

References Bös, D., & Tillman, G. (1985). An envy tax: Theoretical principles and applications to the German surcharge of the rich. Public Finance, 40, 35–63. Bourguignon, F., & Spadaro, A. (2008). Tax-benefit revealed social preferences. Working paper 2008-37, Paris School of Economics, Paris. Fleurbaey, M., & Maniquet, F. (1996). Fair allocation with unequal production skills: The no-envy approach to compensation. Mathematical Social Sciences, 32, 71–93. Fleurbaey, M., & Maniquet, F. (1999). Cooperative production with unequal skills: The solidarity approach to compensation. Social Choice and Welfare, 16, 569–583. Foley, D. (1967). Resource allocation and the public sector. Yale Economic Essays, 7, 45–98. Hicks, J. (1959). Essays in world economy. Oxford: Basil Blackwell. Kolm, S.-C. (1966). The optimal production of social justice. In H. Guitton, & J. Margolis (Eds.), Proceedings, international economic association conference on public economics, Biarritz, 1966. Economie publique, Paris, CNRS, 1968, pp. 109–177. Public economics, MacMillan, London, 1969, pp. 145–200. Reprinted in The foundations of 20th century economics, landmark papers in general equilibrium theory, social choice and welfare, selected by K.J. Arrow, & G. Debreu, 2001, Cheltenham, Edward Elgar, pp. 606–644. Partial reprint in Journal of Economic Inequality, 2007, 5, pp. 213–234. Kolm, S.-C. (1971). Justice et équité. Paris: Cepremap. Reprint, Paris: CNRS, 1972. English translation, 1997, Justice and equity. Cambridge MA: MIT Press. Kolm, S.-C. (1973). Super-équité. Kyklos, XXVI(fasc.4):841–843. Kolm, S.-C. (1974). Sur les conséquences économiques des principes de justice et de justice pratique. Revue d’Economie Politique, 84(1), 80–107. Kolm, S.-C. (1977). Multidimensional egalitarianism. Quarterly Journal of Economics, 91, 1–13.

12 Macrojustice in Normative Economics and Social Ethics

369

Kolm, S.-C. (1985). Le contrat social libéral. Paris: Presses Universitaires de France. Kolm, S.-C. (1986). L’allocation des ressources naturelles et le libéralisme. Revue Economique, 37, 207–241. Kolm, S.-C. (1987). Public economics. In Eatwell, J., Milgate, M., & Newman, P. (Eds.), The new Palgrave: A dictionary in economics (pp. 1047–1055). London: Macmillan. Kolm, S.-C. (1991). The normative economics of unanimity and equality: equity, adequacy and fundamental dominance. In Arrow, K. (Ed.), Markets and welfare (pp. 243–286). London: Macmillan. Kolm, S.-C. (1994). The meaning of fundamental preferences. Social Choice and Welfare, 11, 193–198. Kolm, S.-C. (1995). The economics of social sentiments: the case of envy. Japanese Economic Review, 46(1), 63–87. Kolm, S.-C. (1996a). Modern theories of justice. Cambridge MA: MIT Press. Kolm, S.-C. (1996b). The theory of justice. Social Choice and Welfare, 13, 151–182. Kolm, S.-C. (1999). Freedom justice. Working paper 99-5, CREM, Université de Caen, France. Kolm, S.-C. (2001a). On health and justice. In D. Wikler (Ed.), Global health: From goodness to fairness. Geneva: World Health Organisation. Kolm, S.-C. (2001b). To each according to her work? Just entitlement from action: Desert, merit, responsibility and equal opportunities. Working paper 01-07, IDEP, Marseille. Kolm, S.-C. (2002). La théorie des transferts sociaux et son application. Rapport, Commissariat Général du Plan / Maison des Sciences de Paris: l’Homme (MSH), 412 pages. Kolm, S.-C. (2005). Macrojustice, the political economy of fairness. Cambridge: Cambridge University Press. Kolm, S.-C. (2009a). The rational, recursive original position. A fully determined impartial endogenous social welfare function. In Conference in honour of Maurice Salles. University of Caen. Kolm, S.-C. (2009b). Social ethics and rationality, new directions for the opimum production of social justice: meaningful welfare, equal liberties, social solidarity. In Conference on inequality, new directions, Ithaca, N.Y. Cornell University. Forthcoming in Journal of Economic Inequality (2011). Kolm, S.-C. (2010). Equality. In Badie, B. (Ed.), International encyclopedia of political science. London: Sage Publications, Inc. Luttens, R., & Ooghe, E. (2008). Is it fair to make work pay? Economica, 74(296), 599–626.

370

S.-C. Kolm

Marx, K. (1852). The 18th brumaire of Louis Bonaparte. New-York: Die Revolution. Nagel, T. (1986). The view from nowhere. Clarendon: Oxford. Nagel, T. (1991). Equality and partiality. Oxford: Oxford University Press. Nussbaum, M. (1992). Human functioning and social justice: In defence of Aristotelian essentialism. Political Theory, 20(2), 202–246. Pazner, E., & Schmeidler, D. (1974). A difficulty in the concept of fairness. The Review of Economic Studies, 41(3), 441–443. Pazner, E., & Schmeidler, D. (1978). Egalitarian-equivalent allocations: A new concept of economic equity. Quarterly Journal of Economics, 92, 671–687. Posner, R. (1977). The economic analysis of law, (2nd ed.). Boston: Little Brown. Posner, R. (1981). The economics of justice. Cambridge, MA: Harvard University Press. Rawls, J. (1982). Social unity and primary goods. In A. Sen, & B. Williams (Eds.), Utilitarianism and beyond (pp. 159–185). Cambridge: Cambridge University Press. Rawls, J. (1999,1971). A theory of justice, revised edition. Cambridge MA: Harvard University Press. Rothbard, M. (1973). For a new liberty. New York: Macmillan. Thomson, W. (2008). Fair allocation rules. In K. Arrow, A. Sen, & K. Suzumura (Eds.), Handbook of social choice and welfare. Amsterdam: North-Holland. Tinbergen, J. (1946). Redelijke inkomensverdeling (2nd ed. 1953). Haarlem: De Gulden Pers. Tobin, J. (1970). On limiting the domain of inequality. Journal of Law and Economics, 13, 363–378. Varian, H. (1974). Equity, envy, and efficiency. Journal of Economic Theory, 19, 63–91. Walzer, M. (1983). Spheres of justice. Oxford: Blackwell.