ESSENTIAL EDUCATIONAL ACHIEVEMENTS AS THE CURRENCY OF EDUCATIONAL JUSTICE
Fábio D. Waltenberg*
* PhD. in Economics, Assistant professor in the Department of Economics at Universidade Federal Fluminense (UFF) (Niterói, Rio de Janeiro, Brazil) and Researcher, Centro de Estudos sobre Desigualdade e Desenvolvimento (CEDE), (Niterói, Rio de Janeiro, Brazil). Address: Rua Tiradentes, 17, Niterói, Rio de Janeiro, Brazil, CEP 24210-510. E-mail: email@example.com.
This article has been developed out of a chapter of the author´s doctoral dissertation, defended at the Economics Department of the Université Catholique de Louvain, Belgium. For helpful comments and suggestions, the author thanks the dissertation jury, and in particular his supervisor, Prof. Vincent Vandenberghe. The author is also grateful to two anonymous referees, and the editor of this journal, for equally useful comments and suggestions. The author is fully responsible for errors or misinterpretations that might have remained in this final version. For financial support, the author is grateful to the Belgian French Community´s programs ARC 02/07-274 and ARC 03/08-235, and to the Hoover Chair associate-member allowance.
This article was received on February 2, 2009 and his publication approved on November 23, 2009.
¿Cómo puede definirse la justicia en la educación? Hacerlo de una manera rigurosa requiere la elaboración de algunas alternativas metodológicas importantes. Este artículo, se concentra en la más básica de ellas: los logros educativos esenciales (currency of educational justice). Dado que el nivel en el cual la justicia debe ser valorada es una cuestión sin resolver, se exponen brevemente algunos problemas bien conocidos con puntos de vista normativos que adoptan el ámbito de "macrojusticia", y explican por qué la perspectiva de "mesojusticia" de Amartya Sen es preferible. Después de mencionar algunas objeciones, se defenderá una versión de "educacionismo" (educationism), en el cual la manifestación relevante de la justicia resulta estar asociada a los logros educativos esenciales, un atributo que es simultáneamente un funcionamiento relevante y un determinante posible de capacidades.
Palabras clave: educación, teorías de justicia, imparcialidad de la educación, economía normativa. JEL: D63, I21, I20.
How can we define justice in education? To do so in a rigorous fashion requires making some important methodological choices. In this article, we focus on the most basic of them, namely the choice of the "currency of educational justice". Since the level at which justice is to be assessed is an open question, we briefly expose some problems with well-known normative viewpoints that adopt a "macrojustice" scope, and explain why Amartya Sen´s "mesojustice" approach is preferable. After addressing some objections, we end up defending a version of "educationism", in which the relevant currency of justice turns out to be essential educational achievements, an attribute which is simultaneously a relevant functioning and a plausible determinant of capabilities.
Key words: education, theories of justice, educational fairness, normative economics, functionings. JEL: D63, I21, I20.
¿Comment peut-on définir la justice à l´éducation? Le faire d´une façon rigoureuse requiert l´élaboration de quelques alternatives méthodologiques importantes. Cet article, se concentre sur la plus élémentaire : les réussites éducatives essentielles (currency of educational justice). Étant donné que le niveau dans lequel la justice doit être mesuré c‘est une question encore sans solution, on expose brièvement quelques problèmes bien connus des points de vue normatives basés sur la "macro-justice", qui expliquent pourquoi la perspective de "meso-justice" d´Amartya Sen lui est préférable. Après avoir mentionné quelques objections qu´on peut lui adresser, on défendra une version de l´"educationisme" (educationism), dans laquel la manifestation plus saillant de la justice apparaît associée aux réussites éducatifs essentiels, un attribut qui est associé simultanément à un fonctionnement remarquable et qui peut déterminer les capacités elles-mêmes.
Mot clés: éducation, théories de la justice, justice dans l´éducation, économie normative. JEL: D63, I21, I20.
A person´s education is a valuable asset, because of the benefits in terms of future income stream it entails, because it paves the way for many other life opportunities, and because it has an intrinsic importance. Being a valuable asset, it is normatively relevant to investigate the parameters of its distribution among a given population. Many observers –inside or outside academia– could set as a desirable social goal to increase an education system´s equity or fairness. However, are we sure we know what an equitable or fair education system means? Is it possible to define educational justice in a rigorous way? Answering that would provide guidance for both theoretical and empirical studies regarding inequalities of education, and possibly for policies aimed at remedying unfair educational inequalities.
The issue is inescapably complex, suggesting the path of amulti-disciplinary approach. A promising strategy consists of investigating theories of distributive justice –developed by political philosophers and normative economists– and trying to understand how they can help us shed light on the specific sphere of education. Such a strategy is put forward by a prominent normative economist working in the related field of justice in health; he recommends combining recent "advances in the systematization of the analysis of justice on the one hand", and "in the rather technical sub-field of the measures of unjust inequalities on the other hand" (Kolm, 2002).
In this article, we try to take a modest step towards defining educational justice, taking the methodological procedure mentioned above as a guideline. We start in the next section, by presenting a more detailed motivation justifying the choice of this article´s object. In Section 3, we state how the research question at stake can be framed following Kolm´s recommendation, by means of a partition of the issue in three parts, the first of which is addressed here - namely, defining the relevant attribute of educational justice. Since the level at which justice is to be assessed is an open question, before moving to the substantive argumentation of the article, we briefly recall some problems with well-known theories of justice (welfarism and Rawls´s justice as fairness), which adopt a "macrojustice" perspective. Relying on Sen´s "mesojustice" approach in the particular sphere of education constitutes what we label "educationism", whose advantages and disadvantages are discussed in Section 5. Section 6 is devoted to addressing some important objections that can be made against educationism, leading us to our final claim: that the currency of educational justice should be essential educational achievements. Section 7 concludes.
EDUCATION: A VALUABLE ASSET WHOSE NORMATIVE INVESTIGATION STILL NEEDS TO BE DEVELOPED
Education as a valuable asset
At least since the seminal works of the 1960s (Schultz, 1963; Becker, 1964), economists explicitly and formally recognize that education has an important economic value. Educational outcomes are important means for achieving a wide array of personal goals, not only because of the strictly monetary returns it provides. Indeed, while educational achievements can be good predictors of an individuals´ future earnings capacity, they also arguably impact her likelihood of going to college, the social position she holds, and so on.
Furthermore, education is likely to be positively correlated to "advantages" valued by various theories of justice, and not just within the specific normative framework usually adopted by economists (cf. next sections). Being more educated might, under certain bounds, enhance the odds that an individual will be "happier" (Hartog & Oosterbeek, 1998; Corbi & Menezes-Filho, 2006). In addition to that, more educated individuals might also score higher in the distribution of "primary goods" defined by Rawls (1999), of functionings and capabilities defined by Sen (1985), and also of other "mesojustice" attributes (cf. Section 4), such as health status and civic engagement (Hartog & Oosterbeek, 1998, Dee, 2004, Grossman, 2005.)
Education is also an end or an achievement, since it is –per se– an element of a "good life". Being educated arguably has an intrinsic value, regardless of the effect education might have on other, contemporaneous or future, goals. The United Nations Human Development Index (HDI) ranks countries not only according to their income per capita, but also by (rough) measures of health status and education.1 Along the same line, education is defined by Fleurbaey (1995) as one of the "core individuals´ achievements" about which society should care.
While correlated, income and educational outcomes are not perfect substitutes. When the "currency of justice" –that is, the relevant normative attribute– is income, for a given society to achieve social justice as defined by John Roemer (1998), that society could opt either to massively change the allocation of school resources (cf. Betts & Roemer, 2004) or to redistribute income substantially (cf. Roemer et al., 2003). However, while extremely important for many reasons, massive income redistribution does not, so to speak, "solve" the problem of social justice. It will always be necessary to educate people as well as possible such that, as Betts & Roemer (2004) suggest, people can earn a substantial fraction of their income on their own –a preventive or empowering perspective. Rawls (1999: xiv) defends a property-owning democracy, which for a number of reasons, while relying on competitive markets, tries to disperse the ownership of wealth and capital. He specifies this should be done: "(. . . ) not by redistributing income to those with less at the end of each period (. . . but rather by ensuring the widespread ownership of productive assets and human capital (. . . ) at the beginning of each period." Commenting on Rawls´ ideal society and relating it to educational justice, Van Parijs (2004) says: "From this perspective, an educational system that equips all citizens with all they need not only to find a decently paid job, but also to get along in the other aspects of their lives, such as choosing a doctor, renting a flat or selecting an internet provider, is of paramount importance." According to such views, which we share, income redistribution should be employed to complement people´s income and to insure them, but should be accompanied by efforts aiming at allowing individuals to improve their own personal skills.2
Summing up, the educational level of a person is a valuable asset, not only because of the benefits in terms of future income stream it entails, but also because it paves the way for other life opportunities, and finally, because it may be viewed as having an intrinsic importance. For those reasons, we believe that a better understanding of different aspects related to the distribution of educational achievements is a relevant research topic, certainly related to the study of income distribution, but which is important for its own sake.
A research program in need of development
There are a few contributions in the economic literature whose focus is on justice in education taking theories of distributive justice as a starting point, but normative economics of education actually seems to lag behind related fields such as normative health economics: "a great amount of work has been done by economists in what concerns equitable allocations of health care and relatively a small amount in what concerns education" (le Clainche, 1999: 78).
The treatment given to justice in education and to justice in health in a wellknown handbook series reflects the differences in the importance attributed to research on justice in these two fields. While the handbook of health economics (Cuyler & Newhouse, 2000) contains three chapters explicitly concerned with normative health economics, and partly based upon, or at least referring to, theories of distributive justice, the analogous handbook of economics of education, published more recently than its health economics counterpart (Hanushek &Welch, 2006) does not contain any chapter or part devoted exclusively to "equity and education".3
The shortage of contributions in normative economics of education was also noticed in the public-economics literature: "given the apparent importance of the subject, the economic theory of education in its relation with distributive justice and income redistribution is still under-developed" (Fleurbaey et al., 2002). In that article, the authors are concerned with the relationship between education and general justice, and not with justice in education. While it provides insight on important trade-offs, it does not answer questions such as the one we raise here.
An educationist (Brighouse, 2002) tries to shed some light on the relationship between a particular theory of justice (egalitarian liberalism), and justice in education. Before trying to accomplish this endeavour, he says he is puzzled by the fact that, although highly influential in other normative disciplines over the past thirty years, Rawlsian theories of distributive justice had been ignored by education scholars –possibly explained by the neglect of educational issues by political philosophers themselves. He adds that by ignorance, or misunderstanding, of egalitarian liberalism, egalitarian scholars "have missed out on a rich tool for critique of current practices and policies, and for the guidance of future reforms". Thus, it seems to be the case that normative analyses of education have been rare even in disciplines that place education at the very centre of their research agenda.
There is a flourishing literature regarding normative issues in education in studies on equity in education financing. Empirically oriented papers belonging to this stream of the literature (e.g., Murray et al., 1998; Iatarola & Stiefel, 2003) employ simple and intuitive normative concepts such as horizontal and vertical equity, and now and then refer to more complex ones such as equality of opportunity (although usually a rigorous definition of the concept is absent). Another stream of this literature (e.g., Hoxby, 1999), provides more formalized analysis through standard economic models. A feature of almost all the contributions in this line of research, and particularly the empirical ones, is that equity is usually assessed at the input level. The main variables at stake –typically, the dependent variables in econometric studies– are the monetary resources allocated to pupils, such as per-pupil expenditures, teacher´s wages or school resources. A serious problem with that approach, however, is that monetary resources are only part of the inputs required to produce education, as an abundant literature in the economics of education has taught us.4 Even if monetary inputs (school resources) were uniformly distributed to all students of a given population, the outcomes of these students would certainly still be quite unequal –and partly unfair, whenever those outcome gaps were partly explained by differential circumstances, which are beyond the control of pupils and their families (Roemer, 1998). To appropriately measure equity in education, it is necessary to go beyond this input approach and to turn, instead, to outcome equity.5 We will return to this topic when talking about Amartya Sen´s framework.
Levin (1994) is an older, but extremely interesting, contribution, since it is exclusively devoted to equity in education. The author defines an equitable educational situation as one in which "representatives of different racial, gender, and socio-economic origins have about the same probabilities of reaching different educational outcomes". He first criticizes approaches of equity as equality of access or equality of funding (resources), claiming that they are not sufficient conditions for achieving educational equity - a view which we share for reasons explained in the previous paragraph. Achieving educational equity would require compensatory education funding efforts in the form of "accelerated schools", aimed at increasing the schooling performance of at-risk students. While working with the intuitive notion of outcome equity stated above, this contribution comes very close to the more systematic notion of equity conceptualized by John Roemer a few years later. The article has the usual drawback of lacking a more detailed and systematized definition of justice in education, possibly due to the fact it is not based on the normative literature.
Meuret (2001) provides an overview of the way various theories of distributive justice can shed light on educational inequalities. Firstly, the author presents the theories criticized by Rawls, which are, according to him, utilitarianism, meritocracy, and the principle of redress. Secondly, he describes Rawls´ theory in its relation with education. Finally, those who criticized or extended Rawls´s work are discussed: communitarianism (Walzer), libertarianism and egalitarianism. It is a quite rich descriptive piece of work, but the analysis does not try to address the lack of careful attention of those sets of theories discussed with regards to the educational sector.
An extensive and systematic survey of the literature is out of the scope of this article. What we wanted to point out with what is reported in this subsection is that there are still a limited number of contributions in normative economics of education which are: (i) based on contemporary theories of distributive justice, (ii) explicitly spell out and try to address challenging trade-offs, and (iii) provide a solid basis for empirical applications. Whereas fragmented contributions do exist, to our knowledge no systematic treatment of educational justice, combined with an empirical approach, has been done.
EVALUATING AND COMPARING DISTRIBUTIONS: WHICH STEPS AND AT WHAT LEVEL?
Having exposed the reasons justifying the choice of our research question, we now start adopting Kolm´s methodological suggestion, mentioned in the introduction, which consists of trying to combine two different strands of literature: one regarding theories of distributive justice (or normative issues); the other concerning how to measure unjust inequalities.
Modern normative economics is mainly about describing fair social states, or ranking different social states according to their degree of fairness (Fleurbaey, 1996). But if we want to compare social states and rank them, what set of criteria should guide us? The challenge at stake could be defined as one of ´evaluating and comparing distributions of relevant attributes´. According to Sen (1992), Cowell (1995), or Lambert (2001), doing so requires making three main choices, regarding:
Although there are of course overlaps between these three domains, such a sub-division of the larger issue is useful to organize the reasoning. Because a detailed analysis of the three points would not be treatable in one single article, we limit our focus here to the first choice, that of defining an appropriate currency of educational justice.
Before moving ahead, it is necessary to distinguish at which "level of justice" a normative analysis is to be conducted. Kolm (2002) distinguishes hree possible levels: (i) macrojustice, which concerns "basic rights of a society and the resulting global distributive justice", (ii) mesojustice, concerning "issues which are specific but widespread, important both intrinsically and in volume, and which elicit policies that can affect almost everybody" –justice in health and in education are examples of this, and (iii) microjustice, which concerns particular, very local, situations where issues of justice are raised (e.g., resource allocation inside a hospital or a school).
Many observers would tend to consider education simply as one element that can contribute more or less to social justice, via its impacts on individual "advantage" functions. In other words, all of them would adopt –implicitly or explicitly– an integrated, or general-justice, perspective, that is, "macrojustice" in Kolm´s terms. However, defining the appropriate level at which justice is to be assessed is not a theoretically settled issue. While according to Fleurbaey (1996), an essential element of any theory of economic justice is its "extent or field of application", he admits that in this respect there is not yet an appropriate method allowing us to determine the adequate extent or field of application of a given economic theory of justice.
THREE MAJOR NORMATIVE VIEWS IN THEIR RELATIONWITH EDUCATION
Since the appropriate level at which justice is to be assessed is an open question, before reaching the core of this article, we need to briefly recall some problems with two well-known macrojustice approaches: welfarist and Rawlsian. Then, we move on to our preferred framework –Sen´s functionings and capabilities– which is instead a mesojustice approach.
The welfarist benchmark: a macrojustice approach based on subjective attributes
Suppose we assume: focus on end-states, welfarism, and self-centeredness.6 Suppose individuals´ utility levels are measurable and interpersonally comparable. Finally, suppose that the conditions which are required for the second welfare theorem to be valid break down. Since it is not possible to reallocate initial endowments in such a way that any Pareto-efficient allocation can be reached, it is not possible to partition "equity" and "efficiency" into two separate and distinct problems. Under such conditions, defining a social optimum necessarily requires taking simultaneous decisions concerning equity and efficiency, as stated by Atkinson & Stiglitz (1980), who claim that "both [are] subsumed on the objective of maximizing social welfare", according to the inequality aversion parameter of some objective function. In such a framework, one could make use of a broad class of social welfare functions (SWFs), within the boundaries of the welfarist perspective, which is the standard normative framework in economics.
Focusing on end-states and assuming self-centeredness in educational issues does not differ much from doing so in the general case: an endstate (or an educational outcome in the particular case) is valued according to the benefit it provides to its holder. Assuming welfarism, in turn, is a more delicate issue, among other reasons because the relationship between educational variables and individuals´ welfare is intrinsically multi-fold. When we restrict our analysis to a welfarist framework, important choices have to be made regarding an appropriate description of the relationship between education, income, other relevant variables, and welfare. Particularly, the role of education in generating current and future well-being, and the complementarities between education and other goods in that process, would need to be addressed, a task that could be accomplished in different ways, such as: (i) relying on empirical evidence, (ii) making specific choices depending on the nature of the problem under scrutiny, (iii) turning to a reasoned argumentation or a consistent formal model. However, although useful in other contexts, we do not believe the welfarist framework is the most suitable one for dealing either with social evaluation in general or with educational justice in particular.
Welfarism requires evaluations based on individuals´utility, that is, according to individuals´own preferences, aspirations, beliefs, and desires, and the extent to which individuals are happy or satisfied. Individuals´material holdings, as well as their degree of freedom, entitlement of rights, emancipation, selfrespect and autonomy can also be valued, but only through their effects on welfare –and never per se. While utility may express individual welfare to a certain extent –with wellknown conceptual and fundamental caveats7 – it is difficult to defend it as a good basis for judging social quality and for designing social policies, unless we are willing to base important (objective) choices a society has to make –such as those governing the distribution of educational inputs which will impact on the distribution of educational outcome, a valuable asset, for reasons explained above– on a contestable (subjective) psychological8 metrics, with a great risk of tolerating, among others, the serious problems of cheap and expensive preferences. As an example of cheap preferences in education, we could imagine a situation in whichmany individuals are extremely lowskilled (say, illiterate) –as long as they are happy, welfarism would not consider it as a problem.
Primary goods: a macrojustice approach based on objective attributes
Amajor non-welfarist approach has been developed by the political philosopher John Rawls (1971), whose theory of "justice as fairness" is a response to the welfarist conception of the good society, which prevailed as a dominant view among political philosophers and economists until the 1960s. Trying to reconcile concerns for equality, liberty, and efficiency, Rawls derives two principles of justice which lead him to claim that a just society is one that assigns the highest possible level of "primary goods" to the worstoff individuals (i.e., the "maximin" rule), provided that a certain level of liberty for all is guaranteed. Primary goods are those which every rational individual would want, under appropriate conditions, whatever his or her conception of justice, and whatever his or her life plan.9 Rawls firmly rejects interpreting his theory in the spaces of utilities, well-being, education or health status. The space of primary goods is the appropriate one according to him.
Clearly, when we turn from welfarism to Rawls, the currency of justice shifts from a subjective measure (utility) to an objective one (a set of primary goods). The problems related to the subjectiveness of utility vanish away when primary goods replace utility - for example, information not related to individuals´ welfare can now be taken into account in the evaluation of society. The drawback is that the virtues of a subjective attribute also vanish away. Individual preferences are taken into account only to the extent that they do not clash with the two principles of justice, but in fact Rawls imposes, so to speak, not only that primary goods are the attributes to be valued, but also that they must be valued equally by all individuals. As argued by Sen (1992), Rawls does not take sufficiently into account the diversity of human beings and assumes that all individuals can benefit equally from a given set of primary goods. However, "equality of primary goods can go hand in hand with strong inequalities in actual freedoms enjoyed by different people".
Furthermore, even if primary goods were suitable attributes, those interested in defining educational justice would be faced with important difficulties when searching for guidance in Rawls´s theory, since education is simply not defined by Rawls as one of those primary goods.
Functionings and capabilities: a mesojustice approach based on partly subjective and partly objective attributes
While Rawls built his work as an alternative to welfarism, Amartya Sen built his own theory trying to go beyond welfarism, Rawls´s "justice as fairness", and libertarianism.10 Sen´s "central idea is to see the basal space [i.e. the currency of justice] in terms of what people are able to be, or do, rather than in terms of the means they possess", in a clear contrast with both Rawls´ primary goods and welfarists´ utility. While admitting that freedom is very important, Sen tries to do better than the libertarians, who neglected to a great extent the social consequences of the requirements and constraints their theory imposes. Sen (1985) proposes using the ingredients of quality of life as the currency of justice, by means of two core concepts: (i) "functionings": doings and beings that, taken together, constitute the quality of life; (ii) "capabilities": sets of functionings from which a person can choose.
In the debate over the appropriate currency of justice, one important contribution of Sen consists of trying to reconcile the advantages of both objective and subjective approaches, placing his theory in a somewhat intermediate position between welfarism and Rawlsianism. Sen puts forward an objective currency of justice, and in this respect, he joins Rawls, and opposes welfarism. Sen also takes some distance from Rawls, since functionings are valuable objective achievements of individuals –and not means to achievements such as Rawls´ primary goods. Moreover, while primary goods are defined by Rawls himself, and are to be evaluated according to a scale which is fixed across societies and individuals, Sen´s functionings are to be defined by each society. More importantly, Sen leaves room for individuals´ diversity, since each individual may convert means into ends in a different fashion –for example, converting primary goods (or educational inputs) into quality of life (as expressed, for example, by his or her educational level). Roughly speaking, as compared to functionings, utilities are "too subjective" and primary goods are "too objective".11
Relying on inputs would mean relying on means for achieving or being something, which is a misplaced strategy, because of the different ability individuals have when it comes to converting means into ends. Here, we can see that our claim against focusing on inputs in the specific sphere of education (cf. Section 2) echoes the much more general argumentation made by Sen. Based on Sen´s viewpoint, relevant attributes should be endstate variables, that is: outcomes or achievements. In the educational sector, such end-state variables could take the form of an individual´s total years of schooling, or the certificates she has obtained, or the score she gets on a particular test.
THE EDUCATIONIST APPROACH
We now examine more carefully what is at stake when a mesojustice approach, or segmented justice, is adopted in the educational sphere -what we label "educationism".
Conceptual pros and cons
Educational outcomes can be a reasonably objective currency of justice. They can be more accurately measured (e.g., in terms of schooling years or points on a test) than utility or than advantages such as "freedom", "extended resources", or "opportunity for welfare". They can be compared across individuals. Taking such currency of justice saves us from some difficulties faced by cardinal welfarism. Indeed in a welfarist framework, weighing losses and gains of a given policy is unfeasible, since in practice it is impossible to compare subjective attributes such as utilities across individuals.
In an educationist approach, in contrast, it is possible to assess and compare the scores of students on achievement exams, to know more or less precisely the years of schooling or the highest schooling level attained, or to obtain at least rough measures of other educational outcomes. Even what we could interpret as the "ultimate educational outcome", at least from a strictly economic perspective, namely, individuals´ earnings, is also measurable and comparable.12 The difficulties related to measuring attributes, and to comparing them across individuals, do not pose great obstacles for educationism.
Admittedly, it is possible to, in a sense, "measure utility", by adopting the classical definition of utility and working with happiness as the attribute. As a metrics, happiness and some kinds of educational outcomes share similar limitations: it is not reasonable to claim pupil A "knows twice as much" as pupil B just because the former scores 600 in the PISA exam and the latter scores 300, just as one cannot say individual C is twice as happy as individual D (i.e., restricted comparability). Besides, in both cases, the scale on which variables are measured is artificially constructed, as opposed, for example, to a monetary scale.13 Having said that, while happiness has the drawback of being a strictly subjective variable, scores reflect an objective knowledge, or set of skills. For example, knowing how to solve an algebraic operation reflects an objective knowledge a person has, and not a subjective mental state. "Small bits" of knowledge can be aggregated in order to express the skills a person possesses.
There is a potential flexibility in the definition of what educational outcomes are to be chosen as attributes of justice. At least in principle, many kinds of skills can be evaluated through an educational certification system, and not only standard ones, such as literacy or numeracy skills. The assessment can be extended to what is important in today´s life, such as the knowledge of English as a foreign language, or computer literacy (Van Parijs, 2004). An educationist approach does not restrict itself a priori to attributing value only to strictly academic achievements. Rather, there is room for a wide array of skills to which a given society decides to grant priority, in contrast, for example, with pre-determined primary goods.
Finally, educational outcomes can presumably accommodate non-welfarist and non-consequentialist concerns. It is acknowledged that traditional welfarist approaches have difficulties in taking into account non-welfarist and/-or non-consequentialist objectives such as freedom, emancipation (e.g., of women or ethnic groups), self-respect and autonomy - except to the extent that they impact a person´s utility function. Educational outcomes qualify not very controversially as a relevant functioning, as already claimed. But would more educated people have a larger capability set than less educated people? While we acknowledge this is a controversial issue, we tend to believe more educated people would not only be capable of "being" more and "doing" more (i.e., achieving more functionings), but also of making more and better choices. While a minimal level of education is required for a person to be minimally free, emancipated, autonomous, self-aware, and to respect herself, a higher level of education would very likely enlarge the extent to which these objectives are achievable.
Is educationism in conflict with macrojustice?
One major drawback of segmented justice, according to Kolm (2002), would be the risk of incurring Pareto inefficient situations: "an equal allocation of each of several goods is generally Pareto inefficient because people have different tastes." For example, individual A might prefer to have more education and less health, whereas individual B might want exactly the opposite, and the achievement of equality in health and education separately may be sub-optimal in the Pareto sense. Both Kolm (2002) and Tobin (1970) believe this problem can be attenuated by partitioning the bundle of consumption in the particular fields at stake between those which are substitutable with other goods and those which are not. For example, in health issues, the former could be represented by drugs used for comfort, while the latter would be basic health care needs that are used to keep one alive: "the case where health can uncontroversially be considered in isolation is when it matters with priority" (Kolm, 2002). In the case of basic health needs, treating justice in health as an autonomous separate entity is unlikely to produce (morally relevant) Pareto inefficient allocations. The drugs used for comfort, in turn, could arguably be dealt with within the framework of income distribution, that is, as a macrojustice problem. Obvious analogues in education are basic (compulsory) schooling which would not be substitutable with other goods, and post-compulsory education, which would be substitutable.
But then, how would a segmented justice approach inform us on how to choose, for instance, between how much of basic education and how much of basic health? What to do when faced to such trade-offs? While recognizing the status of the theoretical challenge embodied in such a point, we question its relevance for most practical matters. In situations of extreme misery, priority should be defined as ensuring that the most basic necessities are provided –shelter, security, basic health, etc.– such that the tradeoff is elusive. In sufficiently developed countries, basic necessities of life are ensured, so that resources are abundant enough for social planners to simultaneously care about justice in basic health and justice in basic education, with relative budgets defined according to some practical (democratic) procedure. In middle-income countries (where situations of both extreme misery and abundance coexist), the trade-off would be more pressing, but in this case –again– supplying more basic functionings would be granted priority.
More generally speaking, we could fear that some ranking of functionings would need to be pre-established somehow, such that macrojustice would return –"through the back door", so to speak– to the very core of the analysis. There are at least two lines of answer to that challenge. Firstly, it is indeed possible to rank a certain number of social states, recurring only to ordinal rankings. For example, when calculating the HDI, if two countries perform equally well in terms of health and income, and one of them performs better than the other in terms of education, the former will rank better than the latter, even without an explicit ordering of the three functionings which compose the HDI (that is, without explicitly solving the macrojustice problem).
Secondly, incompleteness can be seen as a normal feature of normative evaluations: "Both well-being and inequality are broad and partly opaque concepts. Trying to reflect them in the form of totally complete and clearcut orderings can do less than justice to the nature of these concepts. There is a real danger of overprecision here" (Sen, 1992, 48). In a reasonable range of contexts and situations, practically some share of the budget is assigned to the minister of education, another to that of health, and so forth. Segmented justice would not allow us to say much about the fairness of a particular partition of the overall budget, but would guide us on how to fairly distribute specific budget shares, such as the educational one. We also quote Sen here, who defends a plurality of approaches when it comes to evaluating inequalities: "Indeed, pluralist proposals make up much of practical ethics, even though descriptive homogeneity evidently appeals to many moral philosophers (utilitarians among them)" (Sen, 1992, 132).
Indeed, people ordinarily think in a ´segmented way´, that is, they intuitively want justice to be made in different sectors or aspects of life, and such concerns are not necessarily irrational or wrong. Tobin (1970), for example, diagnoses that the widespread tolerance of Americans towards general inequality is "tempered by a persistent and durable strain of (. . . ) specific egalitarianism" (his emphasis). While economists may be ´instinctively´ tempted to want to provide people with cash income through taxand-transfer schemes, Tobin claims that an economic rationale can also be employed to acknowledge and justify that some particular kinds of inequalities are more severe than others. He reminds us, for instance, of economic models that have been warning us about the limits of income redistribution, a result which paves the way for focusing on the distribution of specific goods.
Other arguments supporting a particular attention to specific inequality can be found in the strand of the public finance literature devoted to the study of the pros and cons of in-kind versus cash transfers. Gasparini & Pinto (2005) review three classic arguments in favour of in-kind transfers - merit goods (individual consumption of some valuable specific goods would be too low under cash transfers); more efficient redistribution, under specific conditions (already mentioned in the previous paragraph); and market failure (sub-optimal consumption, especially in the presence of positive externalities). A fourth argument, closer to "real-world debate" according to them, is added: equality of opportunity (public intervention in markets such as education and health care, for example, in order to reduce the dispersion in the consumption of ed and health care, which determine future life opportunities of individuals).14
Another reason for not opposing segmented justice a priori is related to feasibility, or tractability. Fleurbaey (1996) says that "feasibility is a practical value that is worth to be taken into account in the assessment and in the definition of justice". According to Kolm (2002), attaining macrojustice might be too difficult a task, and so pursuing justice in particular aspects or sectors may be a useful intermediary strategy for ultimately enhancing society´s macrojustice. Although less ambitious than welfarists´ and most non-welfarists´ conceptions of (macro)justice, an educationist approach may be easier to handle in empirical studies and for policy-related issues. And the same might be true for other specific goods such as health care, housing, and others.
Summing up, we can say that segmenting justice is not, in itself, necessarily an irrational instinct or a misplaced reasoning. On the contrary, it may turn out to be a useful way of downgrading complex problems (e.g., achieving macrojustice) into simpler ones (e.g., achieving justice in different spheres). Mesojustice objectives might make sense as independent, autonomous objectives. Alternatively, provided some appropriate partitions are undertaken (i.e., between substitutable and non-substitutable goods), segmented justice objectives can even be compatible with more standard normative criteria (e.g., Pareto-efficiency for subsets of educational outcomes).
THE CASE FOR FOCUSING ONLYON ESSENTIAL OUTCOMES
Important objections can be raised against educationism. First of all, objective attributes have their own problems, as mentioned before, such as overlooking the diversity of human beings, by assuming that all individuals can benefit equally from a given set of the attribute (e.g., different people might make different use of educational outcomes), which is not necessarily accurate. Related to that objection is the one saying educationism would mean paternalism, given that some third party (the state, the scholar, the policymaker) decides, in the place of children or their parents, that education is valuable for them.
In addition to the arguments supporting a particular attention to specific goods (cf. previous section), we believe it is possible to address these objections by means of two partitions of the set of educational outcomes. Firstly, one can partition the schooling process in such a way that the respect for human diversity is fully preserved for a subset of an individuals´ schooling life. An educationist policy-maker could state, for example, that society must make sure that all citizens will attain a minimal level of educational outcomes (e.g., acquiring basic literacy and numeracy skills), whereas going beyond such level is a personal choice, dictated by each person´s aspirations or preferences (Fleurbaey, 1996; Trannoy, 1999).
Secondly, another partition can be designed which is possibly orthogonal to the first one: subdividing the potentially infinite set of subjects / skills into a subset of essential ones (e.g., mathematics, reading, basic scientific knowledge, English as a foreign language, and basic computer skills) and a subset of complementary ones (e.g., other academic skills, artistic skills, athletic skills, rhetoric skills, etc.). According to each society´s priorities, the educationist policymaker would confine his attention to the essential subjects - those which every person needs in order to live well in contemporary world –and ignore the other ones, which are not of concern for educational justice and belong to the private sphere of individuals.
With those two partitions of the broad set of educational outcomes, the critiques of paternalism would be confined exclusively to those educational outcomes which are obtained in the initial levels of the schooling process, and would be related exclusively to essential subjects. Initial years of schooling are basically attended by children, who cannot be considered as fully autonomous citizens; it would be wrong to assume they are capable of taking fully-informed and fully-rational decisions. Children´s outcomes depend strongly on the behaviour adopted, and the decisions made, by their peers, parents, teachers and other school staff, and the educational authorities (e.g., decisions concerning education resource allocation). For these reasons, "consumer sovereignty" of children should not be taken for granted, and because of that some degree of paternalism might be tolerated, and even recommended.
The partition of the educational process into essential and non-essential subjects (or skills) neutralizes the critiques on the non-essential subjects. The latter are taken to be ordinary goods, meaning that each individual is absolutely free to decide which amount he prefers to consume of them. As for the essential subjects, it is difficult to claim they are not of paramount importance (i.e., a merit good), firstly per se, and secondly as the ingredients for any other social goal people might set for their lives. If this is true, it should not be a problem to impose that they cannot be traded off against other goods, and that some degree of paternalism would, again, be justified.15
Designing a conceptually rigorous and empirically useful definition of educational justice is a great research challenge. In this article we have discussed some issues related mainly to the first of the three main choices that have to be made when it comes to evaluating distributions. As a result, we have set the currency of justice to be essential educational outcomes, an attribute which is simultaneously a relevant functioning (an achievement) and a plausible determinant of capabilities (freedom to achieve different functionings). The argument could be summed up as follows:
What is done here –proposing a definition of the attribute of educational justice– constitutes only one of three important choices. It is still necessary to specify an appropriate aggregation (and/or evaluation) procedure, and to define reference groups or units of analysis. Both discussions would be natural complements to this article, and both still represent considerable research challenges.
Only by expanding our understanding on those three subsets of the larger problem, would we be closer to specifying a fully-fledged and rigorous definition of educational justice. Such a definition would provide a more informed direction to theoretical and empirical studies regarding inequalities of education, and possibly to policies aimed at remedying unfair educational inequalities. If we are successful in accomplishing that task, we would avoid "[missing out on] a rich tool for critique of current practices and policies, and for the guidance of future reforms" (Brighouse, 2002).
1 Although typically there is a correlation between a country´s GDP per capita (PPP) and its position in the HDI ranking, the two orderings do not necessarily coincide. For example, in 2003, Sweden ranked 22nd in GDP per capita, while it was 6th in the HDI ranking; South Africa, in turn, was 52nd in terms of GDP per capita, but only 120th in the HDI ranking.
2 Among the reasons for adopting such views, we could mention the warnings made by Roemer (1998), that redistribution mechanisms are not a panacea, for at least three reasons: (i) tax-and-transfer mechanisms create their own inefficiencies, (ii) individuals may derive self-esteem from occupying skilled jobs, (iii) income taxation distributes money not only from advantaged to disadvantaged types of individuals, but also from high-effort to low-effort individuals (of similar types), which is ethically contentious.
3 There are only three chapters devoted to specific aspects of the black-white performance gap.
4 Including, among many others: Coleman et al. (1966), Hanushek (1986, 1997, 2002), Card & Krueger (1992), Figlio (1999), Woessman (2001, 2003) and various authors in Hanushek and Welch (2006).
5 An exception is Iatarola & Stiefel (2003), who estimate both input equity measures and output equity measures, trying to find out what factors influence pupil performances in standardized tests, taking as the dependent variable the educational outcome, and not inputs anymore. They claim that their paper "is one of the first to measure output equity, using levels and changes in test scores to do so". Previously, some chapters in Berne and Picus (1994) had also tried to go beyond the input equity approach, exploring output equity issues. More recently, Waltenberg & Vandenberghe (2007) set output equity (conditional on pupils´ circumstances) as the framework for an empirical investigation using Brazilian test score data.
6 "Self-centeredness" means here that well-being is, so to speak, filtered by each person´s individual utility function. We do not call it "egoism", because a person could be positively affected by the well-being of another person. While egoism is incompatible with altruism, self-centeredness is not.
7 The strongest critiques on welfarism essentially contest two interlinked implications: (a) that nonutility information (freedom, emancipation, autonomy, collective rights . . . ) is ignored, and (b) that every variable which impacts on utility is assumed to be relevant, allowing situations of "cheap preferences" or "expensive tastes". See, for example, Rawls (1971), Nozick (1974), Sen (1992), Van Parijs (1995), and Laslier et al., (1998).
8 Following Rawls, a series of alternatives have been suggested in a literature, which can be unified under the label of "nonwelfarist" (Maguain, 2002), or "postwelfarist" (le Clainche, 1999), and that include, among many others: Rawls (1971), Dworking (1981a,b), Sen (1985), Cohen (1989), Arneson (1989), Fleurbaey (1995), Van Parijs (1995), and Roemer (1998).
9 Primary goods are the following ones: (a) basic liberties, (b) freedom of movement and choice of occupation, (c) powers and prerogatives of offices and positions of responsibility, (d) income and wealth, (e) self-respect.
10 Libertarianism responds to welfarism challenging the latter´s focus on end-states, and suggesting we care instead about procedural variables such as liberties and rights. We do not discuss it here, because it does not seem to be insightful with regards to education. Probably their advocates would only require something like equal distributions of inputs across individuals, a position we reject for reasons outlined throughout this article. For different accounts of libertarianism, see Atkinson & Stiglitz (1980), Fleubaey (1996), Sen (2000), Williams & Cookson (2000), Arnsperger & Van Parijs (2000).
11 Sen does not limit the analysis to actual achievements (functionings), but takes into account the real opportunity to accomplish what people value (capabilities). He makes clear that both are expressed in the same space (that of functionings), in which functionings are points, and capabilities are sets. Although focusing on capabilities would be the first-best, he admits that measuring a capability set is not an easy task, and he is enthusiastic about empirical applications focusing on functionings, not only because these are feasible (second-best) applications, but also because in any case, functionings constitute "a much finer basis of evaluation of the quality of life and economic progress than various alternatives more commonly recommended, such as individual utilities or commodity holdings" (Sen, 1992, 53).
12 The difference between measuring income in an educationist approach and in a welfarist approach is that in the former we do not assume that income is a proxy for individuals´ (subjective) welfare, but rather that it is merely one of the various (objective) educational outcomes.
13 It makes sense, though, to say that a person has completed twice as many years of schooling as another. Moreover, 2earsïs not an artificial scale.
14 Walzer (1983) goes further. Not only does he claim that people are not necessarily wrong to think in a segmented way, but he also defends it as indeed the correct approach to justice. His seminal book published in 1983 contains a radical defence of segmentation of justice. Each sphere of justice (education, health, labour market etc.) produces a different kind of good, for different users, and the principles that should regulate the distribution of goods must be specific to each sphere. He rejects a unified conception of justice, and believes that the only general principle that has to be respected at the societal level is that of non-interference of one sphere into the other.
15 Schooling systems of many countries have compulsory attendance laws (up to a given age), and thus are already paternalistic. Yet they are not vigorously contested, for many reasons, including the prospects of generating positive externalities.
 Arneson, R.J. (1989). Equality and equal opportunity for welfare. Philosophical studies, 56, 77-93.
 Arnsperger, C. and Van Parijs, Ph. (2000). Éthique économique et sociale. Paris: Éditions la Découverte.
 Arrow, K. (1973). Higher education as a filter. Journal of Public Economics, 2, 103-216.
 Atkinson, A.B. and Stiglitz, J.E. (1980). Lectures on Public Economics. New York: McGraw Hill.
 Becker, G. (1964). Human capital: a theoretical and empirical analysis with special reference to education. New York: Columbia Univ. Press.
 Berne, R. and Picus, L. (eds.) (1994). Outcome equity in education. Thousand Oaks, Ca.: Corwin Press.
 Betts, J.R. and Roemer, J. (2004). Equalizing Opportunity through Educational Finance Reform. Unpublished manuscript. Paper presented at the CESIfo / PEPG Conference on Schooling and Human Capital Formation in the Global Economy, Munich, September 3-4.
 Blanchflower, D.G. and Oswald, A.J. (2004). Well-being over time in Britain and the USA. Journal of Public Economics, 88,(7-8), 1359-1386.
 Brighouse, H. (2002). Egalitarian liberalism and justice in education. Londres: Institute of Education, University of London.
 Brighouse, H. and Swift, A. (2006). Equality, priority and positional goods. Ethics, 116, 471-97.
 Card, D. and A. B. Krueger (1992). Does School Quality Matter? Return to Education and the Characteristics of Public Schools in the United States. Journal of Political Economy, 1(100), 1-40.
 Cohen, G. (1989). On the currency of egalitarian justice. Ethics, 99(4), 906- 944.
 Coleman, J.S.; Campbell, E.Q.; Hobson, C.J.; McPartland, J.; Mood, A.M.; Weinfeld, F.D. and York, R.L. (1966). Equality of Educational Opportunity. Washington, DC: US Government Printing Office.
 Corbi, R.B. and Menezes-Filho, N.A. (2006). Os determinantes empíricos da felicidade no Brasil. Revista de Economia Política, 26(4),120-140.
 Cowell, F.A. (1995). Measuring inequality, 2nd edition Harvester. Wheastsheaf: Prentice Hall.
 Cuyler, A.J. and Newhouse, J.P. (eds.) (2000). Handbook of Health Economics. North-Holland: Elsevier.
 Dee, T.S. (2004). Are there civic returns to education?. Journal of Public Economics, 88(9-10), 1697-1720.
 Dworking, R. (1981a). What is equality? Part I: equality of welfare. Philosophy and Public Affairs, 10(3): 185-246.
 Dworking, R. (1981b). What is equality? Part I: equality of resources. Philosophy and Public Affairs, 10(4): 283-354.
 Figlio, D.N. (1999). Functional form and the estimated effect of school resources. Economics of Education Review, 18(2), 241-252.
 Fleurbaey, M. (1995). Equal opportunity or equal social outcome? Economics and Philosophy, 11, 25-55.
 Fleurbaey, M. (1996). Théories économiques de la justice. Paris: Economica.
 Fleurbaey, M.; Gary-Bobo, R. and Maguain, D. (2002). Education, distributive justice, and adverse selection. Journal of Public Economics, 84, 113-150.
 Frey, B. and Stutzer, A. (2005). Happiness research: state and prospects. Review of Social Economy, 63(2): 207-228.
 Gasparini, L. and Pinto, S. (2006). Equality of opportunities and optimal cash and in-kind policies. Journal of Public Economics, 90(1-2), 143-169,
 Grossman, M. (2005). Education and Nonmarket Outcomes. NBER Working Paper, 11582. New York: National Bureau of Economic Research - University of New York Graduate Center.
 Hanushek, E. (1986). The Economics of Schooling: Production and Efficiency in Public schools. Journal of Economic Literature, 24(3), 1141-1177.
 Hanushek, E. (1997). Assessing the Effects of School Resources on Student Performance: an Update. Educational Evaluation and Policy Analysis, 19(2), 141-164.
 Hanushek, E. (2002). The Failure of Input-based Schooling Policies. NBER Working Paper, 9040. New York: National Bureau of Economic Research - University of New York Graduate Center.
 Hanushek, E. and F. Welch (eds.) (2006). Handbook of the Economics of Education. North-Holland: Elsevier.
 Hartog, J. and Oosterbeek, H. (1998). Health, wealth, and happiness: why pursue a higher education. Economics of Education Review, 17(3), 245-256.
 Hoxby, C. (1999). The productivity of schools and other local public goods producers. Journal of Public Economics, 74(1), 1-30.
 Iatarola, P. and Stiefel, L. (2003), Intradistrict equity of public education resources and performance. Economics of Education Review, 22, 29-78.
 Kolm, S.C. (2002). On health and justice. Paris: Institute for Advanced Studies in the Social Sciences.
 Lambert, P. (2001). The distribution and redistribution of income, 3rd edition. Manchester: Manchester University Press.
 Le Clainche, C. (1999). L´allocation équitable des biens d´éducation et des soins de santé dans une problématique post-welfariste. In: D. Meuret. La justice du système éducatif. Paris: De Boeck Université.
 Levin, H.M. (1994). The necessary and sufficient conditions for achieving educational equity. In: Berne, R. and Picus, L. (eds.) (1994). Outcome equity in education. Thousand Oaks, Ca.: Corwin Press.
 Maguain, D. (2002). Les théories de la justice distributive post-rawlsiennes: une revue de la littérature. Revue Economique, 53(2): 165-199.
 Meuret, D. (1999). La justice du système éducatif. Paris: De Boeck Université.
 Meuret, D. (2001). School equity as a matter of justice. In: W. Hutmatcher, D. Cochrane and N. Bottani. In pursuit of equity in education: using international indicators to compare equity policies. Dordrecth: Kluwer Academic Publishers.
 Murray, S.E.; W.N. Evans; R.M. Schwab (1998). Education-finance reform and the distribution of education resources. American Economic Review, 88(4), 789-812.
 Nozick, R. (1974). Anarchy, state and utopia. New York: Basic Books.
 Rawls, J. (1971). A theory of justice. Cambridge, MA: Harvard University Press.
 Rawls, J. (1999). A theory of justice. Revised edition. Oxford: Oxford University Press.
 Roemer, J. (1998). Equality of opportunity. Cambridge, MA: Harvard University Press.
 Roemer, J.E.; Aaberge, R.; Colombino, U.; Fritzell, J.; Jenkins, S.; Lefranc, A.;Marx, L.; Page, M.; Pommer, E.; Ruiz-Castillo, J.; San Segundo, M.J.; Tranaes, T.; Trannoy, A.; Wagner, G. and Zubiri, I. To What Extent Do Fiscal Systems Equalize Opportunities for Income among Citizens?. Journal of Public Economics 87, 539-565.
 Schneider-Bunner, C. (1997). Santé et justice sociale, Paris: Economica.
 Schultz, T. (1963). Economic value of education. New York: Columbia Univ. Press.
 Sen, A.K. (1985). Commodities and capabilities. North-Holland: Elsevier.
 Sen, A.K. (1992). Inequality re-examined. Oxford: Clarendon Press.
 Sen, A.K. (2000). Social justice and the distribution of income, Ch.1 In: Atkinson, A.B.; F. Bourguignon (eds). Handbook of Income Distribution, vol 1, Norht-Holland: Elsevier.
 Tobin, J. (1970). On limiting the domain of inequality. The Journal of Law and Economics, 13(2), 263-277.
 Trannoy, Alain (1999). L´égalisation des savoirs de base: l´éclairage des théories économiques de la responsabilité et des contrats, In: Meuret, D. La justice du système éducatif. Paris: De Boeck Université.
 Van Parijs, Ph. (1995). Real freedom for all. What (if anything) can justify capitalism?. Oxford: Oxford Univ. Press.
 Van Parijs, Ph. (2004). What is a just educational system? A response to the report ´L´équité des systèmes éducatifs européens. Un ensemble d´indicateurs´. Document de travail de la Chaire Hoover 138, Louvain: Université Catholique de Louvain.
 Waltenberg, F.D.; V. Vandenberghe (2007). What does it take to achieve equality of opportunity in education? An empirical investigation based on Brazilian data. Economics of Education Review, 26(6), 710-724.
 Walzer, M. (1983). Spheres of justice. New York: Basic Books.
 Williams, A.; Cookson, R. (2000). Equity in health, In: Cuyler, A.J. e J.P. Newhouse (eds.). Handbook of Health Economics, vol 1, Amsterdam: Elsevier.
 Woessman, L. (2001). New evidence on the missing resource-performance link in education. Kiel Working Paper, 1051. Kiel: Kiel Institute for the World Economy.
 Woessman, L. (2003). Schooling resources, educational institutions, and student performance: the international evidence. Oxford Bulletin of Economics and Statistics, 65(2), 117-170.