1. INTRODUCTION
In this book, the French economist and philosopher Serge-Christophe Kolm discusses the problem of what he calls “macrojustice”, which “concerns the most general rules of society and their application to the distribution of the benefits from the main resources” (MJ 1)Footnote 1. He reminds his readers that the first and most important challenge for any policy analysis is the spelling out of its ethical foundations. On that basis, he argues in favour of what is nothing less than a paradigm shift in the theory of optimal income taxation, rejecting the whole traditional approach in public economics as being welfarist and too strongly focused on the problem of labour market distortions. As an alternative, he proposes an approach based on the ideal of equal freedom, and works out a specific tax proposal: equal-labour income equalization (or ELIE). He claims that ELIE is a general and operational solution for the organization of the distribution branch of taxation and public expenditure policy that does not create labour market distortions.
Serge-Christophe Kolm is a prolific writer, who has since the sixties of the last century published on an extremely wide range of topics, going from technical analyses of specific public policy issues (Kolm Reference Kolm1970) over an analysis of altruism and reciprocity (Kolm Reference Kolm1984) to Buddhism and psychoanalysis (Kolm Reference Kolm1982, Reference Kolm1987). He seems to write faster than many of us can read. His main focus of interest has always been the question of justice, however: how to find adequate normative foundations for what would be a “good economy” or a good public policy. In each of his previous books he has introduced original ideas, which sometimes have received less attention from the economics profession than they deserved. This is partly due to the fact that for a long period he preferred to publish in French for reasons of principle (Kolm Reference Kolm1988). However, it is also partly due to his style of argumentation. Reading his work is difficult. He writes thick books, in a somewhat repetitive style. In the course of the presentation, he often makes digressions that, although interesting, sometimes make it difficult for the reader to follow the main thread of the argument. He has in the past developed his own set of concepts and his own terminology, with which the reader has to familiarize himself. Reading Kolm requires an effort. His new book, Macrojustice, does not make an exception to that rule. Although the ELIE-proposal is rather simple, it is integrated in an elaborate philosophical discussion, covering more than 500 pages.
It is impossible to review all the ideas from this book in a short note. For economists, it is tempting then to focus on the specific ELIE-proposal. However, such a narrow approach would definitely not do justice to Kolm's book – since one of his main messages is precisely that one should not focus on narrow technical questions while neglecting the broader philosophical issues. I will therefore try in section 2 to give a brief summary of what I see as the main thread of the argument, emphasizing the points where Kolm's analysis deviates most from the traditional welfare economic approach. Although this summary will unavoidably be superficial and incomplete, I still start from it to suggest in section 3 some open questions. Rather than as a critique on the book, I present these questions as a kind of research programme. Section 4 concludes.
2. MACROJUSTICE AND ELIE (EQUAL LABOUR INCOME EQUALIZATION)
2.a. As noted before, Kolm has written in the past many books on a wide range of issues. For him, however, these different books are all related to each other. Taking the problem of macrojustice out of its broader setting, would therefore be highly misleading. Taking ELIE out of this broader setting would be even worse.
In Kolm's view, resorting to justice is only a third best (MJ 10–11). The first best would be a society in which the individuals are sufficiently able to control the birth of their desires (Kolm Reference Kolm1982). The second best would be “the good society” (“la bonne économie”) where people sufficiently like each other to remove all conflicts about sharing scarce resources (Kolm Reference Kolm1984). It is because problems of education and information make it impossible to realize these ideals in our modern large-scale societies that distributive justice as a third best becomes relevant. Moreover, macrojustice is only one of the dimensions of justice. Macrojustice is about the basic rights in society and about the principles for allocating the benefits from the main resources of society. In addition there are also many questions of so-called microjustice related to the social relations in smaller groups or with respect to more specific issues. And one can even consider a field of mesojustice, focusing on specific questions but that concern everyone: matters of education and health are the prime examples. As will become clear later on, there is no reason to suppose that the guiding principles of justice in these different spheres should be the same.
Moreover, justice is one issue, market failures is another. In Kolm's view the latter should be settled through liberal social contracts (Kolm Reference Kolm1985), through which the state implements the contracts that free agents would be willing to engage in if there were no problems of information and transaction costs. Some of these liberal social contracts have immediate distributive implications. Social insurance arrangements are an obvious example. And the fate of the poor as a collective good is another example (MJ 201–202). If everyone cares about the poor, the fate of the poor becomes a public good and free-rider behaviour is to be expected. Government intervention to help the poor can then be seen as the correction of a market failure.
2.b. How should we decide about the principles guiding macrojustice? Kolm's approach is made perfectly clear from the very beginning of the book: “Justice will be taken to be what society thinks it is. The reference will be to consensus or unanimity of the individuals. (. . .) No ‘moral intuition’ is proposed. The only exercise will be to observe and find out what society wants about the necessary questions, and to derive the logical conclusion. (. . .) In a strict sense, this study is not ethics (. . .). It is, rather, anthropology and logic (. . .). Finding the answer about what society should do in society itself and only there (without any input of an externally given rule) is ‘endogenous social choice’.” (MJ 11). At first sight, this reads like an empiricist approach, but of course Kolm's ideas are more subtle. It is indeed obvious that there is no unanimity in the real world about the desirable redistribution. According to Kolm, “for endogenous social choice the final criterion for a theory of justice is its consensual endorsement by sufficiently informed, aware and reflective individuals” (MJ 359, my italics). Kolm here comes close to the Rawlsian notion of an “overlapping consensus”. He rejects the reference to a kind of original position or veil of ignorance, however. For him, moral choices cannot in general be reduced to selfish decisions under uncertainty about one's future position in society.
2.c. What then does this method reveal about the principles that should guide the pursuit of macrojustice? According to Kolm, the basic macrojustice debate in all modern societies is about rights and freedoms. Social freedom means that individuals' acts should be free from forceful interference by others. Such social freedom is unanimously desired. The relevant basic principle of macrojustice then becomes the ideal of equal liberty. Moreover, given the importance of the freedom to act, all attempts to realize distributive justice should be based on so-called inelastic items, i.e. items which are not affected by individual actions, because this makes it possible not to interfere with individual freedom. This implies that the focus must be on the allocation of rights in personal capacities, which are given to the individuals and do not follow from their actions.
Kolm makes a distinction between “eudemonistic” and “productive” capacities. The former are the “capacities to derive satisfaction or happiness from a given situation or consumption” (MJ 94). According to Kolm, these eudemonistic capacities are completely discarded in the public debate about overall distributive justice. “Should someone with a joyful character pay a higher income tax for compensating or equalizing, or because she is less depressed than others by a lower disposable income? Or should gourmets pay lower income taxes because they can derive more happiness from the gain in income? These hardly seem serious questions” (MJ 98). This is not to say that considerations of subjective welfare could not play a role in other settings, e.g. for microjustice in smaller groups or concerning issues of suffering. It seems just that parents take into account the subjective feelings of happiness of their children when allocating a toy. But such considerations do not (and therefore should not) play a role in determining the rules of macrojustice.
It is worthwhile focusing on the reasons he puts forward to justify this neglect of eudemonistic capacities. He explicitly rejects the popular idea that people should be held responsible for their tastes (MJ 101–104). On the contrary, he claims that these eudemonistic capacities belong to the “core self”, because of their intimate relation to feeling, preferences, choice and action (MJ 101). In that sense, they are different from productive capacities: “Eudemonistic capacities are basically capacities to feel and experience emotions, using capacities to evaluate or discriminate, and resting in part on habit, experience, and culture. They can be seen as more intimate, more private, or more constitutive of the individual than the capacities to act, relate, or even to know (cognitive capacities), which constitute the bulk of productive activities” (MJ 104). The basic reason for not taking eudemonistic capacities into account in macrojustice is therefore respect for privacy and for personal integrity.
Not taking eudemonistic capacities into account leaves only productive capacities as the main focus of redistribution. Yet, within a liberal approach, redistributing the proceeds from productive capacities immediately raises the issue of self-ownership. Does a liberal approach not imply that individuals own themselves, hence own their productive capacities and the fruits of these productive capacities? And does this not exclude any redistribution, as redistribution implies forced labour? Here, Kolm suggests an interesting way out. Social freedom is having the right to use one's capacities in the way one desires. But such right to use one's capacities does not imply full self-ownership (for the whole of the productive capacities), since one can also be the tenant of oneself (at least partially). Having a tenancy-right means that one still has full freedom to act as one wants, but that one has to pay a rent for this use. Redistribution according to macrojustice is then about the allocation of these capacity rents.
Given this ethical background, it is not surprising that Kolm does not consider indirect taxation: people should be free to allocate their income as they want. It is more surprising that he largely neglects the taxation of the non-human natural resources: his main argument is that these only produce a small fraction of total output and that property rights in natural resources have been settled in the past by all kinds of contracts, mainly governed by microjustice. As for capital income: capital itself is produced by labour and capital income is labour income plus income from savings. There is a justice problem related to the taxation of bequests, but in “Macrojustice” this is left aside as being relatively unimportant.
2.d. How can the ideal of equal freedom be made operational? According to Kolm, the most natural (and classical) approach is free exchange from an initial equal allocation. Consider the simplest case of a society with n individuals, where the labour of each individual i can be represented by a scalar ℓi. Let us represent the productive capacities of individual i by wi. If total time is normalized to 1, wi is the amount she can produce when she works full-time. Define then “initial equal labour” as k. In a situation without redistribution, working k would provide individual i with an income wik. ELIE (equal-labour income equalization) now implies that these incomes from working k should be distributed equally over all individuals in society, so that everybody finally obtains the average kω, where ω = (1/n)Σwi. This implies that individual i receives a subsidy ti = k(ω – wi), which becomes a tax if wi > ω. Individuals can then freely choose their actual labour ℓi, which will result in disposable income yi = wiℓi + ti = kω + (ℓi–k)wi.
This approach reconciles two intuitions. Each individual keeps full freedom to act and to earn the corresponding (ℓi – k)wi in addition to the given kω, which can be seen as a minimum or basic income. The former component captures the idea of full self-ownership, the latter component the idea of equality. The relative weight given to both considerations will depend on k. If k = 0, we have full self-ownership and no redistribution at all, with possibly disastrous consequences for the unproductive individuals. If k = 1, we have full equality and we face the problem of the slavery of the talented. According to Kolm, the level of k reflects the degree of solidarity desired by society. A complete part of the book (MJ 277–360) is devoted to the question of how it can be determined through the method of endogenous social choice. Pragmatic approaches can start from observations of the minimum income that is generally acceptable in a given society – if we know this minimum income and ω, we can derive the corresponding k. One could also use observations of the level of k that is desired by the individuals with wi = ω; for these individuals with average productivity the choice of k has no influence on the tax they have to pay, and they can therefore be assumed not to be driven by self-interest. More generally, decisions about k have to result from an open dialogue between sufficiently informed and reflective individuals and Kolm discusses different methods which may be helpful in finding a consensus among the impartial views of the citizens.
ELIE is a transparent and attractive scheme. Kolm describes many different but complementary interpretations, among which: ELIE as reflecting a mixture of desert (for labour supply k, equal effort is rewarded equally) and merit (for ℓi > k, effort is rewarded according to productivity), as a financed universal basic income, as an example of generalized reciprocity (since each individual in a certain sense yields the proceeds (k/n)wi to everybody else), as a set of progressive redistributive transfers. He also describes and explains some necessary extensions. If labour is seen as a multi-dimensional concept (taking into account not only duration, but also speed, accuracy, effort, etc.), he proposes either to interpret k as a multidimensional “equal labour”, or to reinterpret duration and the corresponding wi so as to take into account these other labour dimensions. To tackle the problem of involuntary unemployment, he proposes to “truncate” the individual production function. This means that someone facing the labour market constraint ℓi ≤ ℓi 0 will be interpreted as having a productive capacity wi = 0 for ℓi ≥ℓi 0. The ELIE scheme can then be applied to these truncated production functions. Someone with ℓi = ℓi 0 < k will get the minimum income kω. This also holds for the fully unemployed with ℓi 0= 0. However, if a person has 0 ≤ ℓi < ℓi 0≤ k, i.e. if he freely decides to work less than ℓi 0, his income will be reduced by the corresponding loss in output.
The involuntarily unemployed probably are the main group of individuals with ℓi < k. In general, however, Kolm claims that these situations with ℓi < k are out of the scope of macrojustice. From an ethical point of view, the equalization labour k must in general be smaller than ℓi, since the consensual opinion in society is that people should not be taxed on the basis of their potential but unrealized earnings. In fact, in even the most redistributive countries, the current redistribution scheme amounts to redistributing the income earned in about two days per week: hence for normal full-time labour ℓi > k. There are of course individuals in society for which ℓi < k: these may be the handicapped, those who take part-time jobs, and perhaps a few productive individuals who choose to work very little. The latter in fact withdraw in a certain sense from the cooperative venture that is society. According to Kolm, these situations are exceptional or belong to the domain of microjustice.
2.e. Kolm emphasizes that ELIE is not only attractive from an ethical point of view, but that it is also a practical proposal. It can indeed be implemented as a tax on wages, if these are accepted as a good approximation of wi. Kolm points out that labour duration is most often observable – so, if one can observe labour income (as is necessary also for the present income tax schemes), one can in principle derive the necessary information about wi. For individuals such as the self-employed, whose wages are more difficult to observe, or for the more realistic case in which labour is seen as a multidimensional phenomenon, he suggests useful empirical approximations. Pay sheets are official documents and usually contain information about the most important labour characteristics such as productivity premiums or educational requirements. In any case, Kolm claims, collecting information about wi is not more difficult than collecting information about total labour income – and incomparably much easier than collecting the information about subjective utilities that would be necessary in a full-fledged application of the traditional welfarist approach.
I emphasized that Kolm advocates a tax on inelastic items (productive capacities) because he wants a scheme respecting the individual's freedom to act. But, of course, such a lump-sum tax does not introduce any distortions on the labour market. Pareto-efficiency is therefore an interesting by-product of ELIE. Moreover, Kolm suggests that his scheme, in which everybody is free to decide about her labour supply and each additional unit of labour earns an additional income of wi, is incentive-compatible, in that all economic agents, including the more productive ones, get the right incentives to reveal their own true productivity, at least for ℓi > k, for Kolm the only relevant situation for macrojustice.Footnote 2
3. A RESEARCH PROGRAMME?
In MJ, Kolm shows a tendency to present his work (including the ELIE proposal) as the final conclusive answer to the questions raised. Moreover, since all the ideas in his different books fit together, it is as if taking one brick out of the wall might let the whole building collapse. The reader therefore gets the impression that one can only take his message or leave it. But this is not a very fruitful attitude. Rather than as the definitive answer, Kolm's book on “Macrojustice” should be seen as part of a bigger philosophical and scientific debate. It contains many interesting and innovative ideas, but there are also open questions remaining. Some of these questions relate to the basic foundations of the approach, others are largely empirical. Taking the book seriously means that one attempts to make explicit these open questions, while largely remaining within the broader framework introduced by the author. I will now summarize some of these remaining questions, following the structure of the previous section.
3.a. Putting the issue of justice into a broader perspective and drawing attention to self-control, altruism and reciprocity is one of the most interesting points in Kolm's work. In the same way, his description (Kolm Reference Kolm1985) of a whole spectrum of liberal social contracts offers an interesting approach for thinking about non- (or less) redistributive aspects of government policy in a non-welfarist framework. However, at the same time, this broader approach also raises some difficult issues, which are perhaps not definitively settled by MJ.
One is the integration of the ELIE taxation scheme into the financing structure following from the other liberal contracts. For the financing of public goods, Kolm proposes a system of benefit taxation. Yet I mentioned already that some of these public goods have immediate distributional implications. One example is that of collective gift-giving (MJ 201–202), i.e. the fate of the poor as a public good. The distributional impact is even more obvious for other liberal social contracts in the sphere of social insurance: unemployment benefits are a prime example. While it definitely could make sense to distinguish these different domains in theory and perhaps even to keep their financing separate, there is no country in the world where this is done. There are always intricate interactions between social insurance and income taxation – and concern for the poor and the unemployed is expressed through multiple channels. In fact, the integration of these different domains is considered one of the main challenges of actual social policy. Kolm suggests that ELIE would come on top of unemployment insurance (MJ 213). But this raises difficult issues. The design of the unemployment insurance scheme (with the resulting problems of moral hazard) will have an effect on labour market incentives – and, in general, these incentives will result from the interaction between the different schemes, including benefit taxation and ELIE. Unemployment traps can follow from complex interactions, which are not obvious at first sight. There is a whole research domain here: how to give empirical content to the full set of liberal social contracts and to ELIE at the same time? What are the final distributional outcomes? And how do the different schemes interact in influencing behaviour?
Similar issues arise with respect to mesojustice (health and education). MJ basically takes the productive capacities as given (although Kolm discusses the issue of personal investment in education – I will come back to this). But productive capacities are heavily influenced by decisions in the spheres of education and health. Therefore government policy with respect to mesojustice has a direct effect on the distribution of productive capacities within society. This raises the philosophical issue that productive capacities are no longer given to individuals. At a more practical level, it suggests that mesojustice and macrojustice are necessarily linked to each other. Assume (realistically) that informational restrictions, leading, for example, to moral hazard in health insurance, make it impossible to fully realize the ideal of mesojustice. Would one then not expect some (re)interpretation of ELIE in the light of these imperfections in the sphere of education and health? Is it possible to work out the “freedom”-analogue of the welfarist public economic literature on the interaction between social insurance and income taxation (see, for example, Boadway et al., Reference Boadway, Leite-Monteiro, Marchand and Pestieau2006)?
The most challenging issue is the relationship between macrojustice, reciprocity (altruism) and self-control. Given the rich view of human beings that dominates Kolm's work, it cannot but strike the reader that the perspective on the labour market taken in MJ is a rather traditional economistic one. There is not much room for reciprocal relations in the process of wage formation itself and in the way the firms are run. Yet, there is by now a growing amount of empirical evidence that social factors do play a role in the process of wage determination. Should this not have some consequences for macrojustice? Moreover, and perhaps even more basically, there are also effects in the other direction. The degree of reciprocity in society is not given, nor is the willingness to “control” desires. Both are influenced by social institutions, including those of macrojustice. In fact, Kolm emphasizes with respect to taxation that one cannot think about second best without having defined clearly the first best. Should the same principle then not be applied to the problem of integrating macrojustice in the broader social setting? Should macrojustice not be defined with as main objective to move in the direction of a better economy with more altruism and more self-control? Kolm quotes Aristotle: “If you have friends, you do not need justice, but if you have justice, you want friends in addition” (MJ 11). But perhaps the number of friends you have is codetermined by the systems of taxation and social insurance? I am not claiming that Kolm would deny that this kind of interaction is important. I do claim, however, that a more structured study of these interactions could throw a clearer light on the specific features of macrojustice.Footnote 3 In fact, I think that such a structured study could strengthen the argumentation in favour of Kolm's ambitious overall project.
3.b. I mentioned already that Kolm's notion of “endogenous social choice” does not at all imply that the possibly uninformed and selfish majority in society should decide about the system and the degree of redistribution. The citizens concerned should be “sufficiently informed, aware and reflective”. But when are citizens deemed to be sufficiently informed and reflective? A clear answer to this question is necessary, if one does not want to end in a circular reasoning (defining someone as sufficiently informed and reflective if and only if she accepts Kolm's approach to social freedom). The welfare economists who advise the government on the basis of traditional social welfare functions may be a minority in society, but is there any empirical or theoretical foundation to state that they are not sufficiently informed and reflective (since apparently they do not share in the consensus about the neglect of eudemonistic capacities)? And if one gives up the idea of going for unanimity, what would then be a “sufficient” majority?
At an empirical level, taking seriously the idea of “endogenous” social choice opens up a whole area of research about the formation of distributional preferences. As noted before, Kolm rejects approaches in terms of a veil of ignorance and substitutes for them the idea of an informed dialogue. Interpersonal communication should lead to a better understanding of each other and therefore to the development of an impartial consensus view. Kolm is well aware that setting up and interpreting such a structure of interpersonal dialogue “requires a sufficient theory of the relevant psychological structures” (MJ 325). In fact, as also testified by much of his other work, he is definitely not a narrow-minded economist (or, for that matter, a narrow-minded philosopher) and is very open to contributions from other scientific disciplines. Yet, one sometimes gets the impression that he is using empirical findings in a somewhat casual way. To give an example: he is regularly claiming (or at least suggesting) that there is in most societies some kind of consensus about what a standard minimum disposable income should be. This statement is too easy. Although it is true that there is in most societies near-consensus about the desirability of a minimum income, it is much less obvious however that there is near-consensus about its level. And it is of course this level that would determine k. In any case exploring empirically the degree of consensus about the level of the minimum income – and interpreting the results in terms of ELIE – would be an interesting research topic.
There are other examples of useful research questions. Kolm claims that citizens with the average wage (for which the value of k has no direct financial implications) can be assumed not to be driven by self-interest. Is it possible to identify this group? What do they really think? Can one, by comparing their opinions to those of other groups, indeed identify self-interested biases in the views of the latter? And, more generally: how wide is the support for an ELIE-type mechanism? If the mechanism is explained to uninformed citizens, do they then change their view? Do they accept its basic underlying principles: non-welfarism, valuing freedom of choice, demarcating the limits of full self-ownership? How does Kolm's approach relate to the work of psychologists on the relative acceptance of principles such as needs, equality and desert? And can more survey work on opinions about distributive justice in society (see, e.g. the survey in Konow Reference Konow2003) help to draw the contours of a possible consensus?
3.c. In my view, Kolm's criticism of the traditional public economic approach to optimal income taxation is convincing. And the justification of ELIE in terms of freedom is an original and interesting contribution to the debate. Yet not everybody will be happy with Kolm's answer to the question of self-ownership of productive capacities, i.e. its proposal to reconcile social freedom with the duty to pay taxes by reinterpreting the latter as a rent to be paid for the use of one's productive capacities. The philosophical question is general and goes to the heart of liberalism, but it manifests itself most clearly with respect to the situation of the high skilled (with wi > ω), who prefer to work less than k. As Kolm admits, there is a remnant of the slavery of the talented here, in that these persons after taxes will have less than the “minimum income” kω – they have to pay wik into the scheme, while their labour income is only wili (with li < k by assumption). I will discuss later the question of incentive compatibility in this situation. But even disregarding the potential efficiency problem, from an ethical point of view this feature of ELIE implies that the talented are to some extent forced to work, certainly if kω is close to the minimum income needed to finance the necessities of life. As a matter of fact, Kolm speaks of “the minimal necessary labour of the more able” (MJ 126). He states that the situation li < k will be exceptional for productive individuals and is only relevant for a few eccentric persons who basically put themselves outside the cooperative venture of society. In his view, they do no longer belong to the domain of macrojustice. Yet, it is perhaps a too-easy way out to discard all such cases (even if rare) as being outside the realm of macrojustice. Principles of microjustice and macrojustice should remain coherent. Take the case of an extremely talented engineer (with a large wi) who prefers to live as a shepherd, i.e. to work for an hourly wage far below his productive capacity. Is it really the consensus opinion in society that this engineer should be forced to live below rather than at the minimum income? In my view, a coherent definition of macrojustice should be able to accommodate such cases of individuals whose life project implies that they do not want to realize the full market potential of their productive capacities. Or, formulated in another way: the idea of “minimal necessary labour” for the talented has to be defended as an integral part of macrojustice. Introducing some labour duty for the talented may perhaps be defensible, but how to integrate it into a theory of social freedom?
Specific questions arise about the relationship between Kolm's approach and some of the new perspectives opened in recent social choice theory. In fact, it seems fair to say that in these new approaches the rejection of “eudemonistic capacities” as a criterion for redistribution has more or less become the majority position (although usually the term is not used as such). This has led to a subtle analysis of different interpretations that may be given to “welfarism” (Fleurbaey Reference Fleurbaey2003). It also has resulted in alternative proposals of optimal income taxation, which deviate from traditional welfarism and introduce considerations of freedom and equal opportunities (Fleurbaey and Maniquet Reference Fleurbaey and Maniquet2006). In fact, it turns out that ELIE is fully analogous to the “equal rights solution” that has been axiomatized by Maniquet (Reference Maniquet1998) – itself a particular case of the reference-welfare-equivalent-budget solution proposed by Fleurbaey and Maniquet (Reference Fleurbaey and Maniquet1996).
Kolm mentions the existence of that literature but discards it because in his view Maniquet's axioms do not relate to social freedom and are therefore not “liberal”. In fact, he suggests that “for judging ELIE in this way, one should evaluate the interest (or necessity), meaning, significance, or importance of each of these basic formal properties” (MJ 457). This is undoubtedly true – none of the social choice researchers would object to this statement. But then Kolm continues saying that this discussion “would take us too far away from the present purpose” (MJ 457–458). This is surprising, as Maniquet's axioms turn around the limits of responsibility and focus on opportunity sets. They seem therefore very relevant for Kolm's approach. Kolm has a somewhat more extended discussion of Fleurbaey and Maniquet (Reference Fleurbaey and Maniquet2006) (MJ 496–498). However, it is fair to say that here also his discussion remains a bit superficial. From the point of view of social choice, this seems a missed opportunity. From a broader perspective, if one accepts the relevance of the axiomatic approach (and Kolm certainly does not reject it explicitly), then it would be very useful to analyse carefully how Kolm's approach relates to the rapidly growing formal literature on non-welfarist social choice (including, by the way, also the work on rights as game forms, see for example Gaertner et al., Reference Gaertner, Pattanaik and Suzumura1992). Some of this axiomatic work could gain from a stronger philosophical foundation – perhaps Kolm's approach could also gain from a more thorough axiomatic analysis. There is much room here for interesting work.
3.d. When considering ELIE on its own and disregarding to some extent the broader ethical framework, it can be seen as an extreme variant in the series of recent tax reform proposals aiming at reducing marginal tax rates. It is common practice to evaluate the distributional consequences of such proposals with microsimulation techniques. The available empirical work about flat tax schemes certainly suggests that the distributional effects of replacing the actual tax structure can be considerable. Why not try to do a similar exercise for ELIE? One could also focus on some of the specific policy prescriptions defended by Kolm, such as exempting overtime labour from the income tax. As most existing databases contain information on hours worked and on gross hourly wages, it is certainly feasible to perform these simulations.
Of course, if we find that introducing ELIE would lead to a reduction of the degree of fiscal redistribution in our society, this would in Kolm's view not be an argument against it – it would rather be an indication that the actual fiscal redistribution is not in line with the equal freedom-interpretation of macrojustice. Moreover, the results of the empirical exercises should be interpreted carefully. On the one hand, ELIE only aims at structuring the distribution branch of taxation and expenditure policy, while actual tax schemes cover much more and are also influenced by allocative considerations. On the other hand, in Kolm's view ELIE comes on top of the liberal social contracts, some of which have immediate distributional implications. Yet, the results derived from such simulations would undoubtedly give us a better insight into the economic effects of ELIE. And these effects are also crucial to evaluating the political feasibility of the proposal. If one wants to reform the tax system, one cannot neglect the status quo position. The political feasibility of any reform will be determined by the relative power position of the gainers and losers. Microsimulation exercises can identify these gainers and losers. This would, by the way, also be useful to understanding better the process of endogenous social choice.
A more detailed empirical analysis would also give some insight into the consequences of abolishing the tax on capital income. Kolm's argument for focusing on labour income is partly based on principles (capital is produced by labour, and the ownership of capital and of natural resources has been settled in the past by a set of contracts in the domain of microjustice), but partly also on empirical judgments (the bulk of income is labour income). The relevance of the latter argument can be tested. More specifically, even if the bulk of income is labour income, it is still true that non-labour income is distributed more unequally. Therefore, neglecting it in ELIE may undermine the redistributive potential of the tax scheme.
3.e. Kolm constantly stresses the fact that the focus of the traditional optimal taxation literature on incentives is a sign of misplaced emphasis. He claims that productive capacities are given and that ELIE is incentive compatible, at least for individuals with li > k. He also argues that there is a lot of information available on gross wages and that the number of hours worked is in principle observable. While it is probably true that the traditional economic literature has been one-sided in its consideration of informational issues, Kolm seems to exaggerate in the other direction. He regularly admits that there might be some incentive or informational problems with his scheme, but then tends to take the position that these problems are relatively unimportant. These are statements about facts, however, and not about ethical principles – and MJ does not contain much hard empirical information on these facts. Let us consider in turn the three points mentioned above.
First, Kolm admits that productive capacities are determined by education, and that educational choices may be influenced by the tax scheme. However, in his view, educational choices are made by the parents, not by the economic agents themselves. Moreover, “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 Reference Kolm2004: 140). This intriguing statement implies that even choices at the level of higher education are not influenced by financial incentives.Footnote 4 Is it really true that adolescents or their parents hardly take into account the income-earning potential of different educational choices? The least one can say is that this position is debatable, and that there is a growing amount of evidence that it is not valid empirically (Hanushek and Welch, Reference Hanushek and Welch2006). Better empirical evidence on these issues is crucial to evaluating the implementability of ELIE.
Second, as Kolm admits, ELIE is not incentive compatible for the agents with li < k. This problem is closely related to the labour duty of the talented, as described before. But it does not only hold for the “few eccentric” highly productive individuals who prefer to work less than k. It also holds for all the (in)voluntarily unemployed. Kolm's ingenious proposal to truncate their individual production function can only be implemented if wi, li and li 0 are observable. If this is not the case, these individuals have an incentive not to reveal their private information. Again, one can agree with Kolm that the popular emphasis on the incentive problems of the unemployed may be somewhat misplaced – but it seems strange to go to the other extreme and completely neglect the problem in economies with unemployment rates of more than 5%. At least, the topic deserves more careful scrutiny (some first calculations for Belgium are shown in Ooghe and Schokkaert Reference Ooghe and Schokkaert2007).
Third, is it really true that wage rates can be observed? Kolm argues that wage rates are prices, and that prices are publicly known. One should not neglect the specific characteristics of the labour market, however. There is an obvious problem with the self-employed (as mentioned by Kolm himself). But the problem is broader: as soon as one would really implement an ELIE-type scheme, in which the tax is based on wi, one may expect efforts to hide productive capacities. The indirect way of inferring wages from incomes and labour duration will probably be closed as soon as the economic agents are aware of the importance of wages for the calculation of the tax. It is true that there is now reasonably reliable information on wages available – but it is very optimistic to assume that this information will remain equally reliable if ELIE is introduced. Because of the special features of the labour market, it would be highly surprising if employers and employees together did not find creative solutions to profit from the zero marginal tax rate for li > k, while at the same time minimizing the tax kwi to be paid.
I think that Kolm's statement that ELIE would have no efficiency effects is unrealistic (or at least highly optimistic). This is not to deny, however, that the present literature is one-sided and that it would be very interesting to explore further the possibility of making more use in the income tax system of information about wages – or of other indirect information about productive capacities. Kolm makes a lot of interesting suggestions in his book about how one could implement ELIE in practice. It is a pity that by overstating his case, he makes it less convincing. ELIE is not a panacea that will cure all problems linked to the actual income tax, but it is an interesting proposal which deserves careful investigation.
4. CONCLUSION
It is impossible to do justice to Kolm's book on “Macrojustice” in a short review. This is a rich book, full of innovative and stimulating ideas. Moreover, to really understand it, one has to integrate it into Kolm's overall view, which is expressed in his other books. Together, these books convincingly convey an important message. When thinking about justice, the philosophical argumentation should come first – it is only useful to look at specific technical issues within a broader context. And justice is not the only value in society: self-control and altruism are equally important (and perhaps even more important) ideals. Justice is only third best. At the same time, Kolm defends a whole series of interesting specific policy proposals.
Macrojustice is a thick book, and not always easy to read. Kolm has developed his own terminology and his own style of argumentation, which makes it even more difficult to summarize his ideas in a few pages. Moreover, he has a tendency to present his work as the definitive answer to the questions he raises. This makes it a little difficult to look at it with some intellectual distance. Yet, such distance is badly needed. In fact, Kolm's work deserves being taking seriously – and taking a work seriously means that one points to the open questions. I suggested that Macrojustice in a certain sense opens up an interesting research program, not only for philosophers but also for empirical social scientists in general, and economists in particular.
When evaluating Kolm's contributions one should not forget what happened with some of his earlier work (Kolm, Reference Kolm, Guitton and Margolis1969, Reference Kolm1971, Reference Kolm1976a, Reference Kolm1976b): it is now almost generally recognized that all these contributions were far ahead of their time and contained a lot of results which have been “rediscovered” afterwards. In fact, some very famous economists have built up part of their reputation by working out ideas which were already present in Kolm's writings. Reading Kolm is not easy, but it is rewarding.
