Navigation – Plan du site
Contributions to the History of Health Economics

Identity and the Value of Health

Identité et valorisation de la santé
Florence Gallois et Cyril Hédoin
p. 345-366


Cet article propose de prendre en compte l'identité d'un agent économique dans la perspective de la valorisation de la santé sur un plan normatif. Nous arguons qu'une telle valorisation dépend de l'identité des individus. Nous soutenons que dans le modèle d’analyse normative fondé sur le choix social, la valorisation de la santé se fait nécessairement en évaluant son incidence sur le bien-être des individus. Face à ce constat, nous établissons deux conditions d’identité impliquant que le bien-être d’une personne dépend de son identité. Il s’en dégage que l'identité est pertinente pour valoriser la santé. En conséquence, nous suggérons qu’une pathologie affectant directement les caractéristiques identitaires des individus doit se voir accorder une importance particulière dans une perspective normative.

Haut de page

Texte intégral

1Identity is increasingly discussed in economics. This is presumably because issues related to identity are essential for a better understanding both of the economic agent and her behavior, and of the nature of economists’ normative claims about the efficiency of economic institutions. Identity is obviously relevant in many fields of positive and normative economics (Gallois and Hédoin, 2017) but most of them have yet to deal with this issue, especially health economics. Although the health condition seems constitutive of the identity of a person (e.g. we call a person that cannot hear ‘a deaf’), few if any health economists have investigated this topic. This paper explores the role played by identity in valuing health.

2Our main claim is that the valuation of health states depends on what the identity of an economic agent is made up of. We argue that within the so-called social choice model of normative analysis, the valuation of health states is necessarily done by evaluating how they bear on individuals’ well-being. We provide two conditions for identity indicating that individuals’ well-being is dependent on their identity. This establishes that identity is relevant for determining the value of health. As a consequence, we suggest that pathologies directly affecting individuals’ identity are distinctively significant in a normative perspective.

3The paper is organized as follows. Section 1 argues that valuing health within the social choice model entails the elicitation of the judgments on how the health state bears on an agent’s well-being. Section 2 states our two conditions for identity. Section 3 shows what makes identity normatively relevant and examines the ways through which well-being depends on identity. On this basis, section 4 discusses the implications for valuing health, building on several philosophical approaches of health. Section 5 briefly concludes.

1. Well-Being and the Value of Health States

4As a preliminary step of our account of the role of identity in valuing health, we have to reflect on what it means to evaluate health states from the perspective of normative economics. Standard welfare economics evaluates states of affairs on the basis of what Hausman (2010) characterizes as the “social choice model.” We shall argue in this section that valuing health within this framework entails that the only plausible way to evaluate health states must be done through the elicitation of judgments on how these states bear on the agents’ well-being. It is important to acknowledge from the start that we do not argue that it is necessarily impossible to value health independently from its bearing on well-being. A normative analysis is not necessarily tied up to the social choice model and if one is ready to give up the latter, then it may be possible to value health states independently of the agents’ well-being.

  • 1 The distinction between the informational basis and the aggregation rule is quite standard in welfa (...)

5It might be useful to briefly characterize this so-called social choice model of normative analysis. The ultimate purpose of the social choice model is to generate a social evaluation of states of affairs. A state of affairs can be seen as the exhaustive description of whatever is taken as relevant from the normative point of view, e.g. persons’ wealth or health, the number of people alive, the distribution patterns of wealth or well-being, and so on. The social evaluation is the output of the normative analysis. It may take several forms, more often a partial or complete ordering of states of affairs or a social decision function indicating which states it is permissible to choose. The social choice model builds essentially on the distinction between two key concepts: the informational basis and the aggregation rule. The informational basis corresponds to the set of information on the basis of which states of affairs are evaluated. The aggregation rule specifies how the information collected should be used to generate the social evaluation. In particular, it specifies restrictions on the resulting social evaluation given the content of the information basis, e.g. if everyone prefers some state of affairs x to another state y, then y cannot be evaluated as socially better than x.1 In this paper, we will mainly be concerned with the relevance of the identity of the economic agent for the characterization of the informational basis in the valuation of health. The choice of an aggregation rule seems indeed to be more remote from issues related to identity in economics. The relationship between identity and the determination of the informational basis is more thoroughly tackled in the next section.

6The informational basis and the aggregation rule are the two basic ingredients of the social choice model of normative analysis but additional features are generally needed. In particular, within the informational basis, we may need to distinguish between what we can call the ‘raw input’ and its representation. The former corresponds to the brute information, as it is directly available to the normative economist. It may consist for instance of data on the distribution of wealth or health states vectors defined along standard measures. This raw input is however generally not what is aggregated. Aggregation rather proceeds on the basis of the individual evaluations of the states of affairs characterized by the raw input as well as on their representations through some formal devices (e.g. preference orderings, utility functions).

7Consider the following simple example: say we want to evaluate a set of states of affairs according to the amount and the distribution of wealth within them. We collect actual and/or hypothetical data on wealth distribution. This is what corresponds to the raw input of the informational basis. However, what is to be aggregated are not these data but rather the individuals’ evaluative judgments on states of affairs as characterized by wealth distribution. To do so, assume that individuals are able to form some kind of judgment on the possible states of affairs and that these judgments are somehow “well-behaved.” For instance, a basic requirement for a judgment to be well-behaved would be that if individual i judges state of affairs x better than state of affairs y, and y better than state z, then i should judge x better than z. Individuals’ judgments should then be elicited through some procedure. This elicitation procedure should be sufficient to measure individuals’ judgments in a relevant way and to permit their representations through a formal device. Ultimately, these measurements will be aggregated to form the social evaluation. The informational basis is thus obtained through the raw input and the combination of at least two mechanisms: (i) an individual evaluation mechanism through which we figure out that individuals are able to form their judgments on states of affairs on the basis of the raw input; (ii) an elicitation mechanism through which we assume we are able to represent the individuals’ judgments in the required form. Figure 1 below summarizes our characterization of the social choice model:

Fig. 1: The Social Choice Model of Normative Analysis

  • 2 This is a simplification. First, the separation between the informational basis and the aggregation (...)

8How is the value of health states measured within such a framework? This issue mainly concerns the formation of the informational basis and is only peripheral to the determination of the aggregation rule.2 We have to consider the three key features contributing to the formation of the informational basis: the raw input, the individual evaluation mechanism and the elicitation mechanism. Obviously, the relevant raw input consists in health states. The latter formally correspond to vectors of attributes ranging over various kinds of mental and physical abilities and conditions. These health states are the objects of the social evaluation but it is crucial in the social choice model that they are only indirectly evaluated through the individuals’ judgments.

9How health states are evaluated thus depends on the way the individuals form their judgments and how these judgments are elicited, measured and represented. More specifically, what is at stake here is the nature of the judgments that individuals are forming. Hausman (2010) distinguishes three possibilities:

101. Judgments of personal values: how good or bad an individual judges a given health state to be.

112. Judgments of well-being: individuals evaluate health states according to how they bear on their well-being.

123. Judgments expressing preferences over health states.

13As argued by Hausman, a common mistake is to assume that these different kinds of judgments are identical or at least converge. Convergence will actually occur only if (i) individuals’ preferences over health states fully coincide with their judgments concerning the value of health states and (ii) preference judgments are fully determined by well-being considerations. Assumption (i) is clearly doubtful because preferences over health states might be influenced by factors unrelated to the direct value of health states. Assumption (ii) is also debatable given the fact that someone may have preferences that work against one’s well-being. Acknowledging that these three kinds of judgments need not converge toward the same evaluation of health states, it remains to determine both what is and what should be the relevant kinds of judgments for a social evaluation of health states within the social choice model.

  • 3 Respectively, quality-adjusted life year and disability-adjusted life year.
  • 4 One of these procedures consists in asking people to choose between gambles in the following way. A (...)

14Regarding the “is” issue, health economists routinely assume that they are evaluating health states through the elicitation of individuals’ preferences. The standard elicitation procedures used to measure QALYs and DALYs3 are indeed very similar to the elicitation procedures put forward in decision theory to measure preferences over risky prospects.4 There are strong arguments against the idea that health states should be evaluated on the basis of individuals’ preferences and we will not reiterate them here (see Hausman, 2015). This is however somewhat peripheral here because it is highly doubtful that what health economists are actually measuring are really preferences over health states. Indeed, it is more likely that the elicitation procedures of the kind described above are rather measuring how health states are bearing on the individuals’ well-being. John Broome (2002) provides a particularly strong argument for this claim. The basic idea is that as a set of attributes contributing to one’s well-being, one’s health state is not separable in a technical sense. By this, we mean that the contribution of each relevant factor to one’s well-being depends on one health state. In other words, it is not possible to separate the relationship between well-being and health state from relations between other factors and well-being.

  • 5 An even more restrictive assumption would be that the w function has an additive form, i.e. w = g(d(...)

15To understand this point, suppose that a person’s well-being w at some time t depends on a vector of factors (d1d2, …, dnh1, …, hm). Each hi corresponds to a health factor (e.g. hearing, sight) while the di capture all the other factors contributing to one’s well-being (wealth, social relationships…). For the sake of simplicity, we ignore hereafter the temporal aspect. Hence, we have w = f(d1d2, …, dnh1, …, hm). Suppose that health factors jointly contribute to well-being such that g(h1, …, hm), in which case we can write w = f(d1d2, …, dnh). Technically speaking, defining the value of health independently of the other dimensions of well-being would require to assume that the function w is at least weakly separable with respect to h, in the sense that the value of h and the values of all other variable di are independent, i.e. w = f(g(d1d2, …, dn), h).5 If health economists were indeed eliciting preferences over health states, they would then be measuring the value of h independently of other factors contributing to well-being. But this is not actually the case for the obvious reason that the various health factors hi are related in various ways to the other factors di. Indeed, the burden (i.e. cost in terms of well-being) of a given disease (e.g. arthritis) depends on the way it affects other factors (e.g. ability to play music) and the importance of these factors to one’s well-being (e.g. whether one is a professional pianist or not). Formally, what we have is thus rather something like w = f(d1(h1, …, hm), d2(h1, …, hm) …, dn(h1, …, hm), h). Now, suppose we ask someone to compare two health states hA and hB on the basis of a standard elicitation procedure. Because well-being is not separable in terms of health and because preferences are at least partially determined by well-being considerations, what will be elicited is not a preference between hA and hB but rather a judgment captured by the measures w(hA) and w(hB), i.e. health states are evaluated according to the individual’s judgment of well-being.

16On this basis, it is still possible to determine how health states contribute to someone’s well-being. The method for doing so consists in considering the counterfactual, “if she were in perfect health, then her well-being would be…” Formally, if we denote g(H1, …, Hm) the value of being in perfect health, then the counterfactual well-being is defined as w* = (d1(h1, …, hm), d2(h1, …, hm) …, dn(h1, …, hm), H). Then, one’s health is measured by the fraction a = w/w* and hence, w = aw*. Provided that we can define a zero level for well-being, the “health index” a provides a way to determine the direct contribution of health to a person’s well-being, holding other factors constant. In practice, the elicitation of the health index a is hardly possible however. What standard measures like QALYs are actually corresponding to is an index q = w/w** where w** = (d1(H), d2(H) …, dn(H), H). In other words, health economists are more likely to elicit judgments of well-being that take into account both the direct and indirect contributions of health states to well-being. In so far as “ought implies can,” this strengthens the point that health states should not be valued in terms of individuals’ preferences over them.

17We thus take as our working hypothesis that the value of health states is to be measured through their contribution to well-being. In the rest of the paper, we shall argue that an individual’s well-being depends on her identity, defined by a boundary and a narrative condition. As we may expect, this must have some relevance for the evaluation of health states.

2. The Boundary and Narrative Conditions of Identity

18This section argues that the concept of well-being in normative economics necessarily depends on a conception of the identity of the economic agent. This is especially the case within the social choice model of normative analysis. In other words, normative economics at least tacitly depends on an assumption regarding the identity of the economic agent. Indeed, identity is especially relevant for the individual evaluation mechanism leading to the formation of the agents’ judgments of well-being (Fig. 1).

19Identity is a recent issue in economics (e.g. Davis, 1995; Davis, 2011; Kirman and Teschl, 2004; Sen, 1999). The work of Akerlof and Kranton (2000; 2010) has led to a significant reconsideration of the role played by identity in economic relations and behaviors. Akerlof and Kranton’s approach corresponds to an extension of the standard rational choice framework. Each individual is endowed with a utility function representing her preferences over outcomes or states of affairs. The extension lies in the fact that matters related to social identity are directly integrated into the persons’ utility functions. The economic agent is thus coextensive with a specified preference ordering in which social identity figures as an input. Kirman and Teschl (2004) propose a more integrative conception of the agent’s identity by extending Nozick’s (1974) conception of the individual. They distinguish three characteristics that are constitutive of the identity of an agent. The first characteristic refers to the What dimension of identity, i.e. the manner in which the agent organizes her preferences. The second characteristic refers to the Where dimension of identity, i.e. the social group she belongs to. The third characteristic refers to the Who dimension of identity, i.e. her self-perception. The What and the Where dimensions of identity are significant to understand the mechanism through which each individual forms her judgment, either in terms of preferences or in terms of well-being. Nevertheless, taking these questions into account requires neither a major revision of the rational choice framework nor a significant modification in the conception of the economic agent.

20The standard conception of economic agency has however recently been criticized by Davis (2003; 2011) on its very foundational grounds. Davis argues that the definition of economic agency as being co-extensional with a preference ordering suffers from inability to satisfy the two criteria of individuation and re-identification. The first criterion “requires that individuals be distinct and independent … they must hold together as single whole beings and cannot fragment or break up into multiple selves” (Davis, 2011, 9). The second criterion “requires that individuals maintain their individual distinctness across change, especially when this change causes them to share many of their characteristics with others” (ibid., 10). The individuation criterion sets a first bound on any appropriate account of agency, namely to allow us to explain why and how individuals “do not fragment into collections of sub-personal multiple selves” (ibid.). The re-identification criterion sets up a second bound: we must be able to tell how and why individuals “do not become part of the supra-personal social world” (ibid.).

21Davis argues convincingly that the standard conception of agency fails to satisfy both criteria, due in particular to a problem of under-determination: if an agent is coextensive with a preference ordering, then the identification of an agent requires that one is able to identify the corresponding preference ordering; however, such identification seems to presuppose our ability to identify the agent to whom the preference ordering should be attributed. In other words, preference orderings are insufficient to identify agents because the definition of a preference ordering already depends on the determination of who the agent is. This problem clearly makes the standard conception of agency unable to satisfy the individuation criterion. But it is also relevant with regard to the second bound of agency set up by the re-identification criterion: nothing in the formal apparatus of the utility maximization framework forbids one to ascribe preference orderings to collective entities such as firms, corporations or any other kinds of groups (Sugden, 2000). There must then be some way to determine when the preferences revealed by or causing the behavior of some individual pertain to the individual agent herself or to collectives to which she belongs. However, no formal rule for doing this exists. It follows that as time is passing by and the individual keeps on interacting with other persons and various collectives, it becomes impossible to keep track of the (possibly changing) preference ordering that is co-extensive with the individual-qua-agent. The point is that the standard conception of agency fails to satisfy the individuation and re-identification criteria because it lacks an adequate account of what makes the identity of an economic agent or in other words, an account of who the economic agent is.

22As stated by Kirman and Teschl (2004, 63), who the person is refers to “the idea of a self-reflexive human being who has the capacity of actively discovering and consciously creating her identity within a given historical and social context.” Though the Who issue has been largely neglected in economics, we shall argue that two necessary and (together) sufficient conditions of agency must be satisfied by any account of identity within the perspective of being helpful for the normative analysis: a boundary condition and a narrative condition.

  • 6 On the notion of group agency, see List and Pettit (2011).
  • 7 As a member of the party I may claim that we believe that policy P is commendable while at the same (...)

23First, the determination of who the economic agent is depends on a boundary condition: while the agent may be characterized by a set of elements (especially behavioral dispositions and propositional attitudes), the self-identification of an economic agent makes it necessary for her to delineate which are those elements that are constitutive of her agency. A simple example is sufficient to illustrate this point: as a member of a political party, I may entertain a set of attitudes and political views that are more or less convergent with the attitudes and views endorsed by the group as an agent.6 Both the individual member of a political party and the political party itself are agents insofar as they can be ascribed propositional attitudes (i.e. desires, wants, preferences, beliefs, etc.) on the basis of various assumptions regarding their rationality. But the point is that for me to conceive myself as an autonomous agent, I must be able to distinguish my views and attitudes from the views and attitudes of the party I belong to. This is so even (and especially) if I may occasionally endorse the party’s views and attitudes while they are not mine.7

24There are probably several ways to fulfill the boundary condition but actually it is highly likely that only a few of them are plausible. In particular, there is a direct relationship between the boundary condition and the competing accounts of personhood and personal identity that have been developed by philosophers. These accounts precisely attempt to determine what makes a person the same person now as the person she was in the more or less distant past (and will be in the more or less remote future). There are broadly two kinds of accounts of personal identity. Depending on the view defended, it is possible to claim that there are facts that make someone being a determinate person (non-reductionist accounts) or to claim that there are no such facts and that being someone is a matter of degree (reductionist accounts). Though we do not need to a take a stand on this issue here, it seems obvious that the way the boundary condition is satisfied is likely to be relevant in the normative perspective.

  • 8 Obviously, diseases whose main consequence is to deprive persons from their narration abilities (e. (...)
  • 9 Our narrative condition echoes various recent writings on agency and rationality in economics. Two (...)

25The second condition refers to the agent’s ability to rationalize to herself and to others her behavior and attitudes. We call it the narrative condition. An agent may entertain both at a given moment and over time a variety of attitudes that may, at least at first sight, seem to be inconsistent according to various standards (rational, logical, axiological, etc.). Consistency is however a constitutive characteristic of any plausible conception of agency because individuals whose behavior is perceived as totally erratic will have a hard time to interact with others and may even be denied the right to do so.8 The narrative condition does not state that an agent’s behavior and attitudes should be impeccably coherent—such a requirement would force us to the conclusion that there are few if any economic agents in the world (except maybe for insects). It requires rather that behavior and attitudes could be (at least in principle) justified by a set of publicly endorsable and reasonable reasons and values. In other words, the narrative condition imposes the constraint that from the point of view of the agent herself, her behavior and her attitudes must be amenable to a rationalization that would be meaningful for everyone.9 This entails that the agent is endowed with a set of abilities, which are not only cognitive (e.g. having the capacity to remember past choices or to imagine possible futures) but also social (e.g. being able to communicate through a shared language, being able to empathize with others). In this sense, the boundary and narrative conditions overlap but only partially: features that may help to fulfill the boundary condition (say, the fact that there is a psychological continuity between an agent’s different selves) are also essential in the perspective of the narrative condition. But clearly, the narrative condition also entails additional capacities that reduce the set of entities that can count as “authentic” agents. We argue that any theory of agency and identity must account for these capacities. As for the boundary condition, there are probably several different ways of doing this but the narrative condition puts nonetheless significant restrictions on what can be considered as acceptable theories of agency and identity.

3. Identity and Well-Being in Normative Economics

  • 10 Fleurbaey’s (2003) characterization of formal welfarism is actually more precise as it corresponds (...)

26The relevance of identity issues in normative economics is especially salient within the social choice model of normative analysis, particularly in the context of what Fleurbaey (2003) characterizes as “formal welfarism.” Axiomatically, formal welfarism is defined as the set of axiological accounts that relies on social welfare functions satisfying two axioms:10 a Pareto condition and an independence condition. The former states that if a social state A is judged as better to all agents than a social state B, then A should be socially ranked higher than B. The latter roughly indicates that the social ordering of two states A and B should depend only on how the agents rank A with respect to B, independently of any other alternative. Formal welfarism restricts the range of social welfare functions by determining the relevant informational basis of the normative analysis. It is however totally permissive regarding i) what is to be aggregated and ii) the nature of the agency on the basis of which aggregation proceeds. Both are obviously directly relevant for the determination of the informational basis.

  • 11 Hedonic utilitarianism as originally developed by Bentham and others was of course relying on such (...)
  • 12 See Nussbaum (2001).

27Consider the former point first. As it is standard in social choice theory, a social welfare function can be defined as a mapping FU O of a vector of n individual utility functions u = (u1, …, un) defined over some relevant domain U onto a social ordering o O of a set of social states X. The social ordering is thus a function of the possible profiles of individual utility functions figuring in the relevant domain. While the function F is defined over profiles of individual utility functions and thus aggregates utilities, the precise nature of the latter is left undefined. That is to say, individual utility functions may represent almost anything that one deems as being axiologically relevant. To be more specific, individual utility functions may represent the degree of preference-satisfaction as it is standardly assumed in welfare economics. But they may also measure what is sometimes called “experienced utility” in cases where one’s axiology is grounded on a mental state account of well-being.11 Individual utility functions may also be grounded on various “objective list” theories of well-being, including those endorsing a notion of capabilities.12 Finally, they may also represent the agents’ (subjective) judgments regarding well-being. As we argue in the preceding section, the latter is more likely to be the case when the social choice model is applied to the valuation of health. The only constraint is ultimately that the measure satisfies some properties in order to make utilities amenable to a numerical representation. Regarding the second point (i.e. the nature of the agency), nothing in the formalism requires that the subscripts in the vector of individual functions refer to persons qua agents. Indeed, as suggested by Mirrlees (1982), the individual utility functions may be ascribed to sub-personal selves. Alternatively, they may plausibly refer to the well-being of some collectives that are considered as sufficiently homogenous by the analyst.

  • 13 Sumner’s welfarist approach is committed to the view that well-being judgments are actually happine (...)

28In this perspective, an important issue is to determine the implications of the boundary and narrative conditions of identity for the normative relevance of well-being judgments, i.e. for the determination of the informational basis of the normative analysis. In particular, we may ask which properties these judgments should have to be considered as relevant. Given our conditions of identity, we shall suggest that these judgments should be authentic in the sense of Graham Sumner’s (1996) “welfarist” approach. According to Sumner (1996, 139), authentic judgments require information and autonomy.13 The information condition plausibly requires that the person’s judgments are not grounded on false beliefs, ignorance, misperception or self-delusion. A simple example is sufficient to illustrate the deep connection of this requirement with the Who dimension of identity: you are enjoying a successful professional career in your father’s company, which contributes to making your life enjoyable and fulfilling. In part, this is due to your belief that your success is due to your hard work and your competencies and to the fact that having a successful professional career is important to you. Now suppose that you reliably learn that actually your rising to the top of the company was mostly due to your father’s discriminatory intervention. Then this would probably lead you to reassess your whole life and probably negatively affect your judgments. The point is that learning that your previous evaluation was grounded on false beliefs directly conflicts with the narrative condition of identity: until then, you were able to rationalize the shape of your professional life (and your life as a whole) through values and reasons (hard work, talent, perseverance, etc.) that matters to you. Now that you are discovering that your career has nothing to do with them, this narrative is no longer available.

29A similar reasoning applies to the autonomy condition. Autonomy can be regarded as the requirement that the values and reasons on the basis of which you are more or less satisfied with your life should be your own. This does not mean that these values and reasons are not shared across the population or that you should not have acquired them by interacting with other persons. It rather demands that having these values and reasons is not due to mechanisms such as social conditioning or indoctrination that deprive you from your self-assessing reflexive abilities. In other words, these values and reasons are your own because you would be willing and able to defend and endorse them in a publicly deliberative context and under rational scrutiny. Of course, there is no clear-cut method to determine whether or not one’s values are authentic in this sense. Socialization is such that values and reasons are never completely our own. Again however, the point is that autonomy is directly related to both the boundary and the narrative conditions. Regarding the former, what autonomy requires is that the values that are guiding your self-assessment of your life should indeed be your own and be distinguishable from those of other persons or groups. Regarding the latter, you should be willing to rationally endorse the values on the basis of which you self-narrate your life. Otherwise, your narrative could be regarded as fictitious and unreliable by others, and the self-assessment of your life axiologically irrelevant.

30Our claim is thus that the determination of the welfare-relevant informational basis must depend on a conception of identity. Interestingly, this point seems to be valid whatever the specific account of well-being one endorses. Welfare economists generally take for granted that well-being is constituted by the degree of satisfaction of the agents’ preferences: state x is better than state y for agent i if and only if i prefers x to y. Moreover, this account of well-being is generally combined with a choice-based or revealed preference approach. According to this approach, preferences, the satisfaction of which is constitutive of well-being, are revealed by choices. Recent developments in the so-called “behavioral welfare economics” research program (Bernheim, 2016; Bernheim and Rangel, 2009) indicate that such choices are relevant when they are made in the appropriate conditions, in particular when they do not result from errors or “characterization mistakes.” Though behavioral welfare economists do not make use of the notion of identity, it could be argued that the determination of what Bernheim (2016) calls the “welfare-relevant domain” of the normative analysis could depend at least partially on the fact that the choices result from (or reveal) well-informed and autonomous judgments. In this sense, the welfare analysis would be respectful of the boundary and narrative conditions of identity.

31The relevance of the identity conditions is even more transparent if one endorses a view of well-being that does not rest only on choice data. This is the case for instance of the so-called “value-based life satisfaction” account (Tiberius and Plakias, 2010). The value-based life satisfaction account of well-being is firmly a subjective account and makes happiness the key constitutive component of well-being. Since this view does not assume that well-being judgments are necessarily reflected in choice, it makes use of non-choice data to measure well-being, for instance individuals’ self-reports about their happiness and life-satisfaction. Still, as for the choice-based approach, the normative relevance of these judgments stems from the fact that they are not based on errors, mistakes or self-delusion and that they are made by autonomous and informed individuals. Hence, the normative relevance is also established within the value-based life satisfaction approach.

32Ultimately, both the choice-based and the value-based life satisfaction accounts make use of individuals’ well-being judgments to derive (possibly incomplete) preference or value orderings on the basis of which social evaluations are derived. In both cases, the boundary and narrative conditions intervene at three distinct levels of the analysis. First, as indicated above, they contribute to determining the relevant informational basis (i.e. which are the judgments/choices that are part of the welfare-relevant domain). Second, they directly shape the individuals’ judgments and choices. Indeed, one’s identity is constituted by goals and values that guide one’s life. What makes these judgments and choices normatively relevant is the fact that they are the result of the latter. Third, the very ability of forming and pursuing goals and values is a constitutive feature of one’s well-being. Using the formal framework of section 2, what we claim is that the normative relevance of identity is twofold. On the one hand, identity directly determines the shape and the content of the w and w** functions. On the other hand, it indicates which judgments encapsulated in the w functions should be part of the informational basis. The former point explicitly shows that the value of health states for each individual cannot be measured without reflecting on the role of identity. The latter point shows that valuing health to derive a social evaluation must include the normative relevance of identity. Presumably, that makes health-related problems that affect or interfere with goals and values particularly significant in a normative perspective because they potentially interfere with these three levels. In particular, they compromise the ability of individuals to form authentic well-being judgments. The next section considers how philosophical approaches to health deal with this issue.

4. Identity, Philosophical Approaches of Health and Valuing Health States

33Valuing health implies defining what health is and what a healthy individual is. Sciences like biology, physiology or histology may help us to objectivize what health is. We have seen above the importance of taking identity into account in order to define well-being. Defining what a healthy individual is thus also implies taking into account the social and personal identity of this person. This section aims at discussing the philosophical approaches to health with respect to what we have previously shown. We also identify consequences regarding the value of health.

34Our discussion of the social choice model has pointed out that the valuation of health states is carried out through the mediation of individual judgments in terms of well-being. The formation of these judgments is deeply related to the identity of the persons who form them. Consider for instance the case of a deaf person. In a naturalistic perspective (see below), if a person is deaf, her bodily health is not good. Cochlear implants allow one to “cure” deafness by resulting in hearing. A sizable part of the deaf community reacted however with hostility to this technology. They maintain that deafness is not a disability but that deaf people constitute a minority cultural group. A significant part of the deaf community claims that they do not want to use a hearing aid. They argue that deafness can induce disadvantages in the global society but that these disadvantages have institutional causes that can be changed by the way society is organized. For example, if everyone were to learn the sign language, the problem of communication between a deaf person and a hearing person would be solved. Robert Sparrow (2005) points out that “the claim that deaf people are disabled relies on assumptions about prevailing institutional arrangements that are themselves part of what is under contestation in arguments about who is to count as disabled” (137, our emphasis). His arguments suggest that there are both a social as well as a biological aspect to deafness. Indeed a deaf person, evolving in the deaf community, shares a history, a language, schools, and so on with other deaf persons. Sparrow refers here to the where dimension of identity. An agent however also builds the narrative dimension of her who identity with her personal history, i.e. the difficulties she has overcome, the schools she has attended, etc. To sum up, a person with a deafness identity does not consider herself as health-deficient regarding to the social aspect of deafness; however, a deaf person cannot hear where a “normal” person hears.

35As an instance of a naturalistic approach to health, Christopher Boorse’s biostatistical theory of disease (BST) defines health as normal functional abilities regarding the agent’s age and sex. A biological function refers to a causal contribution to a goal, and the global goal of an organism is assumed to be individual and reproductive survival. The prime function of lungs is to organize a process of gas exchange, by extracting oxygen from the atmosphere and transferring it to the bloodstream and by releasing carbon dioxide into the atmosphere. By allowing breathing, lungs contribute to one’s survival. An abnormal breathing (dyspnea) refers to what Boorse calls a disease. According to Boorse’s conception, “health is identical with the absence of disease” (Boorse, 1997, 7-8).

36Boorse notes that cultural differences exist in the definition of normality. For example, dyschromic spirochetosis (a skin disease) is so prevalent in a South American tribe that its absence is taken to be an illness and makes one unsuitable for marriage (and thus reproductive survival, a goal of any organism). In other words, the definition of normal functioning may differ according to biological differences between sex and age, but also according to where the individual lives. The consequence is that the Where dimension of identity determines the value of health state as the social group one belongs to and contributes to defining the normal functioning, at least by fixing criteria for marriage. Still, dyschromic spirochetosis is a pathological condition, which induces skin discoloration and lesions.

37Boorse assumes that health corresponds to a normal functioning of organs regarding a normal (Gaussian) distribution within a population restricted by age, sex and possibly cultural norms. As seen above with dyschromic spirochetosis however, a statistically normal functioning does not necessarily correspond to an adequate functioning. Furthermore, the line between an adequate functioning and a pathological functioning depends on an evaluation of the consequences of a deficit in functioning of standard activities (Schwartz, 2007). From this perspective, Hausman (2012) argues that the efficiency of organs also depends on the environment where the agent is living and on the circumstances. Indeed, organs adjust their activity to circumstances. When John runs in his morning jog, the activity of his lungs adjusts in line with this circumstance. Physical activity necessitates a better oxygenation of his whole body, and John’s breathing frequency thus increases from 12-15 breaths per minute to 35-45 breaths per minute (and so does John’s heart rate). Without this circumstance, such a breathing frequency would not correspond to a functional efficiency of John’s lungs but could reflect a panic attack. During the winter break, John likes to ski. High altitude is a change in environment where everyone (including John) breath faster and deeper (i.e. hyperventilate, a sort of dyspnea) since there is less oxygen. Within this environment, hyperventilation is an efficient functioning of John’s lungs.

38Beside gas exchange (and breathing) an important function of lungs is to filter the air inhaled. This function is crucial to the whole organism, in particular within a polluted environment. During pollution peaks lungs have to filter more atmospheric pollutants in order to be efficient. In the long run this filter function can however induce a disease, like lung cancer. In other words, a normal and efficient reaction to the environment may lead to a pathological condition. Boorse’s biostatistical theory (BST), even enlarged by Hausman (2012) through his functional efficiency theory of health, is thus not satisfactory to evaluate a health state. Organ functioning depends on the circumstances in the environment and the evaluation of health depends on the social and individual perceptions of health and diseases/disabilities.

39Nordenfelt’s holistic theory of health (1995; 2007) enlarges the BST approach. He assumes that health is not a purely biological concept, but also a social concept (1995, 129). He considers “a human being as a socially integrated agent who performs a great number of daily activities and is involved in many personal and institutional relations” (1995, 35). This capacity to perform activities is the core of Nordenfelt’s health theory. Thus “A is completely healthy if, and only if, A has the ability given standard circumstances, to reach all his or her vital goals” (Nordenfelt, 2007, 7). The notion of “his or her vital goals” means that a person can have other goals than survival and reproduction. She may want to get a job, visit her friends, shop, walk with 6-inch heels, travel and so on. Vital goals are linked to the agent’s identity. If Mary wants to be a nurse and Thomas an astronaut, it is because they have a different identity. The notion of standard circumstances refers to a cultural norm. Health thus depends on the community in which the agent lives, in other words, on his Where identity. Within Nordenfelt’s conception of health, a deaf person is healthy if her deafness let her reach her vital goals and if she lives in a community where speaking and hearing is not necessary since she can use sign language.

  • 14 Health states considered worse than death have negative values.
  • 15 The same reasoning is used with other methods applied to evaluate health states: HUI(3) and GDB 201 (...)

40Performing some daily activities is a prevalent and vital goal. The ability to perform such activities is commonly used in health measurement schemes like EQ-5D that includes mobility, self-care and usual activities. This ability is also measured through criteria of pain/discomfort and anxiety/depression. Answers allow one to measure health states, with quality weights valued between 0 and 1 and 1 corresponds to full health and 0 to death.14 This permits one to measure health through by single unit, the QALY. Since health cannot be defined and compared in a strict biological manner, health state measures and burden of disease measures are based on the agent’s judgments about well-being and abilities (cf. section 2).15 Indeed, while health is linked to well-being, its effects on well-being cannot be separated from the effects of other factors on well-being (Broome, 2002). Hausman (2015) in particular identifies 5 factors that influence the value of a token health state regarding its consequences on well-being: (i) geography, which refers to the environment that can improve or reduce the burden of a specific disability; (ii) technology and architecture, which refer to the ways through which a disability can be partially repaired by acting on the couple ability-environment; (iii) occupational and recreational structures, which refer to common activities in a society, e.g. deafness is more limiting in a speaking society; (iv) culture and social norms, which refer to the judgment of others on a disease/disability, e.g. dyschromic spirochetosis is a more serious health problem in our society than in the South American tribe where this skin problem is a norm; (v) individual tastes and objectives, which refer to Nordenfelt’s vital goals, like a finger injury that could end the career of a pianist. Of these five factors, three are directly related to the individual’s identity. In other words, the burden of a disease/disability differs according to the Where and the Who dimensions of identity. The Where dimension of identity can increase or reduce the burden of a disease/disability while the Who dimension of identity can be compromised. Measuring the health state/burden of disease thus implies that one takes into account the agent’s identity and distinguishes between acute and chronic diseases, as they will not affect an individual’s identity in the same way.

  • 16 This loss in well-being may increase if she suffers from a painful phantom limb sensation, but the (...)

41The issue of the agent’s identity is more relevant in the case of chronic diseases, when her health state implies a change in her ability to reach her vital goals and her social environment. For instance, the amputation of a limb is a sudden change in the health state of an individual with long-term consequences. After an amputation she loses her ability to reach her vital goal: if the amputation concerns a leg, the individual cannot walk and as a result she cannot accomplish self-care and normal activities. With the loss of her leg, she feels an important loss of well-being since she cannot participate in common activities that used to build her identity.16 A marathoner cannot run anymore and won’t be able to improve her personal record. A mother won’t be able to take care and to raise her children. Both also may struggle with body image issues. In other words, neither the marathoner nor the mother can preserve the narrative condition of her identity as before the amputation. However when fitted with appropriate prosthetic, both of them can pursue their goals. Since they can move on their own, they can perform the common activities of daily living. The boundary condition of identity is reestablished. They can also reach the goals of their narrative identity and perhaps find new goals. The mother can go to the park and play with her children. The marathoner can run again. But she cannot race with the same competitors and has to race with persons suffering of the same sort of disability, changing her Where identity.

42An agent cannot always adapt herself to a chronic condition. Alzheimer’s Disease (AD) is a chronic neurodegenerative condition that progressively induces a loss in the identity of the person. It affects about 6% of people 65 years and older (Burns and Iliffe, 2009). The prevalence of AD is growing in every Western country as the population is ageing. The most common symptoms are a memory loss, language problems, disorientation, a loss of motivation resulting in not managing self-care and behavioral issues. In the first stage, a short-term memory loss appears. The agent cannot remember recent events resulting in a loss of his capacity to narrate her whole history. The narrative condition of her identity is affected. At the intermediate stage (medically named “early stage”) older memory and implicit memory (body memory) are affected. The individual becomes unable to perform the activities of daily living and she suffers from speech difficulties. All these symptoms badly affect her identity. The loss of short-term and long-term memory affects her narrative condition. Frequently, the person fails to recognize close relatives, resulting in a loss of her individuation in the community. She also faces speech difficulties, which leads to communication difficulties. These symptoms affect the where dimension of her identity. They may also affect the well-being of their relatives, progressively facing the loss of the person they knew and loved. In the final stage the person becomes completely dependent upon caregivers (Burns and Iliffe, 2009) resulting in the loss of the boundary condition of her identity. Memory decreases again and the speech difficulties increase but, she can often understand and return emotional signals (Förstl and Kurz, 1999). Her identity still persists even if AD highly reduces her capability to act as an agent with an identity.

43Evaluating the health state of a person with AD is complex since she cannot clearly express her own well-being. She cannot perform daily activities nor reach her vital goals any more. The few possibilities of adaptation are only temporary since the progression of the disease cannot currently be stopped. The disease progressively affects all the dimensions of the agent’s identity, resulting in a terrible health state that progressively and irremediably reduces until death. This suggests that pathologies affecting individuals’ identity have a distinctive significance in the perspective of valuing health in the sense that the ability to realize their identity by forming and pursuing goals and values is itself a constitutive aspect of well-being.

5. Conclusion

44This paper argues about the importance to reflect over what constitutes the identity of the economic agent from the perspective of health valuing. Standard normative economics is used to value health through preference elicitation procedures. These procedures do not however, reflect that judgments on well-being are related to the way each person conceives her life and makes it valuable, i.e. her identity. We propose an account of the normative relevance of identity by identifying two conditions: the boundary condition and the narrative condition. Within this framework, we show that valuing health through well-being judgments within the social choice model of normative analysis depends on identity. On the one hand, well-being judgments are shaped by individuals’ identity. On the other hand, their normative relevance depends on the fact that they are authentic, meaning that they reflect individuals’ goals and values. A survey of philosophical approaches of health leads us to consider that health is at least a capacity for an agent to realize her identity by reaching her goals by herself in the context of her social environment.

45These results suggest that public policy should evaluate the health state (and thus value health) by taking identity into account. Regarding our given examples, it clearly appears that Alzheimer’s disease is more severe than amputation (corrected by an adequate prosthesis) regarding the capacity of an agent to realize her identity. Indeed, an amputation (corrected by a prosthesis) may—although not necessarily—change the agent’s identity. Finally, Alzheimer’s disease gradually affects all the dimensions of a person’s identity. Our analysis is not intended to mean that public health policies should necessarily focus primarily on such pathologies affecting persons’ identity. Our point is rather that these pathologies have a distinctive significance in the perspective of valuing health because they compromise the individuals’ ability to reflect on their goals and values. Incidentally, neurodegenerative diseases like Alzheimer’s are extreme cases that point to a set of unresolved issues for any account of identity taking the boundary condition and especially the narrative condition seriously. In particular, it might be suggested that the narrative condition includes a form of “path-dependence” according to which an agent’s set of possible identities are progressively narrowing due to the accumulation of past commitments. Such kind of path-dependence is partially linked to the reflexive abilities of the agent (which make her able to rationalize her past commitments) and directly leads to the persistence of identity through time. In the case of neurodegenerative diseases, these reflexive abilities progressively disappear, which indicates that identity weakens rather than persists. The normative importance and meaning of this path-dependence phenomenon is an important subject worth further (and future) considerations.

We thank the two anonymous referees for the insightful and challenging comments on a previous version of this paper. We also thank Pierre Van Zyl.

Haut de page


Akerlof, George A. and Rachel E. Kranton. 2000. Economics and Identity. The Quarterly Journal of Economics, 115(3): 715-753.

Akerlof, George A. and Rachel E. Kranton. 2010. Identity Economics: How Our Identities Shape Our Work, Wages, and Well-Being. Princeton: Princeton University Press.

Bernheim, B. Douglas. 2016. The Good, the Bad, and the Ugly: A Unified Approach to Behavioral Welfare Economics. Journal of Benefit-Cost Analysis, 7(1): 12-68.

Bernheim, B. Douglas and Antonio Rangel. 2009. Beyond Revealed Preference: Choice-Theoretic Foundations for Behavioral Welfare Economics. The Quarterly Journal of Economics, 124(1): 51-104.

Boorse, Christopher. 1997. A Rebuttal on Health, What is Disease? In James M. Humber and Robert F. Almeder (eds), Biomedical Ethics Reviews. Totowa, NJ: Springer, 1-134.

Broome, John. 2002. Measuring the Burden of Disease by Aggregating Well-Being. In C. Murray, J. Salomon, C. Mathers and A. Lopez (eds), Summary Measures of Population Health: Concepts, Ethics, Measurement and Applications. Geneva: World Health Organization, 91-113.

Burns, Alistair and Steve Iliffe. 2009. Alzheimer's Disease. British Medical Journal, 338.

Davis, John B. 1995. Personal Identity and Standard Economic Theory. Journal of Economic Methodology, 2(1): 35-52.

Davis, John B. 2003. The Theory of the Individual in Economics: Identity and Value. London: Routledge.

Davis, John B. 2011. Individuals and Identity in Economics. Cambridge: Cambridge University Press.

Fleurbaey, Marc. 2003. On the Informational Basis of Social Choice. Social Choice and Welfare, 21(2): 347-384.

Förstl, Hans and Alexander Kurz. 1999. Clinical Features of Alzheimer's Disease. European Archives of Psychiatrys and Clinical Neuroscience, 249(6): 288-291.

Fumagalli, Roberto. 2013. The Futile Search for True Utility. Economics and Philosophy, 29(03): 325-347.

Gallois, Florence and Cyril Hédoin. 2017. From Identity to Agency in Positive and Normative Economics. Forum for Social Economics. 1-17.

Hausman, Daniel M. 2010. Valuing Health: A New Proposal. Health Economics, 19(3): 280-296.

Hausman, Daniel M. 2012. Health, Naturalism, and Functional Efficiency. Philosophy of Science, 79(4): 519-541.

Hausman, Daniel M. 2015. Valuing Health, Well-Being, Freedom, and Suffering. New York: Oxford University Press.

Kirman, Alan and Miriam Teschl. 2004. On the Emergence of Economic Identity. Revue de philosophie économique, 9(1): 59-86.

List, Christian and Philip Pettit. 2011. Group Agency: The Possibility, Design, and Status of Corporate Agents. Oxford: Oxford University Press.

Mirrlees, James A. 1982. The Economic Uses of Utilitarianism. In A. Sen and B. Williams (eds), Utilitarianism and Beyond. Cambridge: Cambridge University Press, 63-84.

Nikolajsen, Lone and Troels S. Jensen. 2006. Phantom Limb. In S. McMahon and M. Koltzenburg (eds), Wall & Melzack's Textbook of Pain. Edinburgh: Churchill Livingstone, 961-971.

Nordenfelt, Lennart. 1995. On the Nature of Health, an Action-Theoretic Approach. Dordrecht, Boston & London: Kluwer Academic Publishers.

Nordenfelt, Lennart. 2007. The Concepts of Health and Illness Revisited. Medicine, Health Care and Philosophy, 10(1): 5.

Nozick, Robert. 1974. Anarchy, State, and Utopia. New York: Basic Books.

Nussbaum, Martha C. 2001. Women and Human Development: The Capabilities Approach. Cambridge: Cambridge University Press.

Ross, Don. 2005. Economic Theory and Cognitive Science: Microexplanation. Cambridge, MA: MIT Press.

Schwartz, Peter H. 2007. Defining Dysfunction: Natural Selection, Design, and Drawing a Line. Philosophy of Science, 74(3): 364-385.

Searle, John R. 2010. Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press.

Sen, Amartya. 1999. Reason before Identity: The Romanes Lecture for 1998. Oxford: Oxford University Press.

Sen, Amartya. 2002. Rationality and Freedom. Cambridge, MA: Harvard University Press.

Sparrow, Robert. 2005. Defending Deaf Culture: The Case of Cochlear Implants. Journal of Political Philosophy, 13(2): 135-152.

Sugden, Robert. 2000. Team Preferences. Economics and Philosophy, 16(02): 175-204.

Tiberius, Valerie and Alexandra Plakias. 2010. Well-Being. In J. Doris and The Moral Psychology Research Group (eds), The Moral Psychology Handbook. Oxford: Oxford University Press, 401-431.

Tuomela, Raimo. 2013. Social Ontology: Collective Intentionality and Group Agents. Oxford: Oxford University Press.

von Neumann, John and Oskar Morgenstern. 1944. Theory of Games and Economic Behavior. Princeton: Princeton University Press.

Haut de page


1 The distinction between the informational basis and the aggregation rule is quite standard in welfare economics and social choice theory. Bear in mind however that it is mostly a heuristic. In practice, it is not always possible to disentangle them. This is particularly obvious when one looks at the details of the various axiomatizations in social choice theory. Several commonly used axioms seem to determine both the informational basis and the aggregation rule at the same time. The various Paretian and independence axioms are cases in this point.

2 This is a simplification. First, the separation between the informational basis and the aggregation rule is not that radical (see the preceding footnote). Second, it seems obvious that the content of the informational basis may be ethically significant for the choice of the appropriate aggregation rule.

3 Respectively, quality-adjusted life year and disability-adjusted life year.

4 One of these procedures consists in asking people to choose between gambles in the following way. Arbitrarily set the values of perfect health and death to 1 and 0 respectively. Then, ask someone whether she would prefer being in a given health state (e.g. having asthma and otherwise in perfect health) or rather being in perfect health with probability p and dead with probability 1-p. If the former (resp. the latter), then the value of asthma for this person is superior (inferior) to p. The value p* for which one is indifferent between the sure health state and the gamble corresponds to the value of the former. Of course, this procedure is isomorphic to von Neumann and Morgenstern’s (1944) account of how to measure preferences over lotteries in decision theory. An individual’s preferences over such gambles determine an interval cardinal scale through which health states can be (intra-personally) compared. Aggregating the judgments elicited on this basis then forms the social evaluation.

5 An even more restrictive assumption would be that the w function has an additive form, i.e. w = g(d1d2, …, dn) + h.

6 On the notion of group agency, see List and Pettit (2011).

7 As a member of the party I may claim that we believe that policy P is commendable while at the same time I believe that P is not the best option. On the distinction between I-mode and we-mode attitudes, see for instance Tuomela (2013) and Searle (2010).

8 Obviously, diseases whose main consequence is to deprive persons from their narration abilities (e.g. Alzheimer) are particularly burdensome in this perspective. We return to this point in the next section.

9 Our narrative condition echoes various recent writings on agency and rationality in economics. Two otherwise quite diverging accounts defending similar views are Sen (2002) and Ross (2005). The former emphasizes that rationality cannot be reduced to mere consistency but also depends on the reasonableness of one’s preferences. What is reasonable is actually largely determined in public and deliberative settings. Ross explicitly states that agency is grounded on narration abilities. Public scaffoldings, starting with language, are essential in this endeavor because they facilitate coordination between agents but also within agents, i.e. between the agent’s selves.

10 Fleurbaey’s (2003) characterization of formal welfarism is actually more precise as it corresponds to what is known in the literature as the strong neutrality axiom. The latter is itself obtained by combining the axiom of Pareto indifference and the axiom of independence of other alternatives. These details are not relevant here however.

11 Hedonic utilitarianism as originally developed by Bentham and others was of course relying on such a mental state account. Nowadays, mental state accounts are essentially endorsed by behavioral economists who consider that “experienced utility” rather than “decision utility” is what normatively matters. See Förstl and Kurz (1999); Fumagalli (2013) for a discussion.

12 See Nussbaum (2001).

13 Sumner’s welfarist approach is committed to the view that well-being judgments are actually happiness judgments. Hence, he equates well-being with authentic happiness. As we explain below, we do not need however to take a stand on the specific nature of well-being judgments.

14 Health states considered worse than death have negative values.

15 The same reasoning is used with other methods applied to evaluate health states: HUI(3) and GDB 2010.

16 This loss in well-being may increase if she suffers from a painful phantom limb sensation, but the frequency and intensity of a phantom limb pain decline over time (Nikolajsen and Jensen, 2006).

Haut de page

Table des illustrations

Légende Fig. 1: The Social Choice Model of Normative Analysis
Fichier image/png, 63k
Haut de page

Pour citer cet article

Référence papier

Florence Gallois et Cyril Hédoin, « Identity and the Value of Health », Œconomia, 8-3 | 2018, 345-366.

Référence électronique

Florence Gallois et Cyril Hédoin, « Identity and the Value of Health », Œconomia [En ligne], 8-3 | 2018, mis en ligne le 01 septembre 2018, consulté le 16 janvier 2019. URL : ; DOI : 10.4000/oeconomia.3840

Haut de page


Florence Gallois

Université de Reims Champagne-Ardenne, Regards,

Cyril Hédoin

Université de Reims Champagne-Ardenne, Regards,

Articles du même auteur

Haut de page

Droits d’auteur

Licence Creative Commons
Les contenus d’Œconomia sont mis à disposition selon les termes de la Licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Haut de page
  • Logo Association Œconomia
  • Logo CNRS
  • OpenEdition Journals