On the Very Idea of an Efficient Wage

This paper argues that the standard characterisation of the equity-efficiency trade-off as set out in this symposium by Joe Heath overstates the tension between these two values. The reason lies in the fact that economists tend to take individual labour supply preferences as given, which leads to a superficial analysis of the concepts of reservation wage and of economic rent. The paper suggests that we should instead think of reservation wages as variable and as influenced by social norms. Social norms play a double role in this context. First, they represent a constitutive element of market competition; second, they can be a determinant of income inequalities. From this perspective, a certain share of high reservation wages sustained by contingent inegalitarian social norms should count as economic rent. The last section of the paper strengthens this conclusion further by drawing a parallel between expensive tastes in consumption and a certain class of high reservation wages. To the extent that the latter are underpinned by social norms rather than efficiency considerations, not paying them is both just and efficient.

There are valuable human activities which require the motive of money-making and the environment of private wealth-ownership for their full fruition.[…] But it is not necessary for the stimulation of these activities […] that the game should be played for such high stakes as at present.Much lower stakes will serve the purpose equally well, as soon as the players are accustomed to them.
- Keynes ([1936] 1953, 374)   When assessing the structure of wages paid to individuals in an economy, two considerations come to the fore.On the one hand, we want wages to be efficient in the sense that they contribute to an economy that maximises the satisfaction of preferences.On the other hand, we have the intuition that wages should be just, even though there is disagreement about the precise criterion of justice that should apply.
There is a widespread view, especially among economists, that there exists an important tension between these two desiderata.Joe Heath's paper-which is the focus of this symposium-presents a detailed and eloquent defense of this view.In short, it states that any attempts to render the wage structure more just will undermine the functioning of the price mechanism on the labour market.The role of this mechanism is "to channel labour to its best employment" (Heath 2018, 4).Faced with the trade-off between keeping the efficiency gains from wages set by the market and promoting a more just wage structure, economists tend to favour the former.Heath goes one step further and claims that it would be a misplaced form of "overreach" to ask the market for an outcome-"just" wages-that it is ill-equipped to produce (4).
This paper argues that this standard view of the equity-efficiency trade-off suffers from an important blind spot.At the source of this blind spot lies the fact that the standard view takes as given the labour supply preferences of economic agents.I shall argue that this is a mistake with two important consequences.First, it makes unjustified economic rents appear relatively small compared to a situation in which we take labour supply preferences to be a variable rather than a parameter of the analysis of wages.Second, this perspective leads us to overestimate the rigidity of the equity-efficiency trade-off.Put differently, recognising the variability of labour supply preferences entails the possibility of multiple labour market equilibria, some of which are more just than others.
The first section of the paper outlines the normative machinery required to evaluate market outcomes in general, and labour market outcomes in particular.With regard to the former task, the position defended here takes as given the consequentialist framework of the standard view (section 1.1).With regard to the latter, I argue for a distinction between economic rents and incentive payments as two potential explanations of wage differentials (section 1.2).In section 2, I use this distinction to present an understanding of economic rents that is wider in scope than the standard economic account.Importantly, this requires including the labour supply preferences into our normative analysis of wages as a variable, rather than taking them as a parameter (section 2.1).Finally, I argue that the position defended in this paper is congruent with interpreting some labour supply preferences as instances of expensive tastes, thus undermining the legitimacy of the resulting wage inequalities (section 2.2).Section 3 concludes.

EVALUATING MARKET OUTCOMES
The market mechanism is one mode of social interaction among others.Contrary to its main alternative, where the state allocates resources to certain goods and services, and jobs to certain people, the market is based on the decentralised decisions of individuals.
How do we know whether the particular mode of social interaction we choose for a given context serves us well?We need to explicate the social objectives that it is meant to serve, analyse its effects on other social objectives, and then investigate whether it promotes the overall package of objectives better than other available modes of social interaction.As Amartya Sen puts it: "The moral standing of the market mechanism has to be related to its results and it is, thus, derivative and contingent" (1985, 17).Adopting this consequentialist framework, the section aims to set out an evaluative benchmark that can subsequently serve us to assess the performance of the market.

A Consequentialist Assessment of the Market
The list of justifications that have been presented in defense of the market mechanism is long (see Buchanan 1985).In the present context, we will focus on the efficiency-based justification of the market,1 both because it represents the most prominent justification, and because it is directly relevant to the equity-efficiency trade-off (which is defended by Heath as the standard view).
Formulated in non-technical terms, a system is considered to be efficient if it "channel[s] labour to its best employment" (Heath 2018, 4), which is to say, if it allocates labour to the production of goods and services in a way that maximises the satisfaction of preferences.As this definition illustrates, the task of the concept of efficiency here is to establish a link between the allocation of productive resources on the one hand, and human well-being on the other.
It is important to note that efficiency thus understood is not a social objective in its own right, but rather a placeholder for the maximisation of the satisfaction of preferences (see LeGrand 1990, 561ff.).One might, of course, question why maximising the satisfaction of preferences should be regarded as a social objective worth pursuing; but, for the purposes of this paper, I shall take this as given.Note, also, that the maximisation of the satisfaction of preferences does not necessarily tell us anything about the distribution of well-being.2All that matters is aggregate well-being.Now, even if we grant that the market serves efficiency in the sense of maximising the satisfaction of preferences well, for society this goal is only one among others.The other goal on this list that preoccupies us here is the promotion of social justice.While this paper does not endorse any particular theory of justice,3 I assume that any society will want to promote some notion of equity. 4Without asking of the market that it directly serve this notion of equity-that would indeed be a case of overreach-any impact the market might have on the promotion of this goal is clearly relevant to the overall assessment of our institutional arrangements, including the market.In this sense, and this is where I differ from Heath: the distributive outcomes of the market should be part and parcel of our consequentialist assessment of the market, even if our main reason for adopting the market is grounded in efficiency (see also Dietsch 2010).5 Formulated in Sen's terms, the results of the market by which we judge it should include several dimensions, including the two that preoccupy us here, namely efficiency and equity.
Before turning to a distinction more directly relevant to assessing the labour market, let me emphasise that everything said so far is compatible with another aspect of the standard view on the equityefficiency trade-off.Heath cautions against any conception of a fair wage that applies at the transactional level (2018, 9ff.).The reason for his skepticism of transactional approaches is precisely that they would risk interference with the working of the price mechanism.However, Heath does recognise the possibility of a normative evaluation of wages "in terms of overall system performance" (10).In other words, we can compare institutional arrangements to one another in terms of how well they, taken as a whole, serve our social objectives such as efficiency and equity.6

Economic Rents Versus Incentive Payments
Turning to the more specific question of assessing the distributive effects of the labour market, it is important to distinguish two potential explanations for income differentials.
First, the talented often receive an income premium to incentivise them to put their talent to the use where it is most valued by society.An incentive payment can be defined as "that positive or negative amount, above or below its base rate, which puts a person into her particular job rather than into the job she would perform if all were paid the same" (Lamont 1997, 29).Whether incentive payments are justified or not lies at the heart of the debate between G.A. Cohen (2008) and John Rawls  (1999).Cohen argues that incentive payments represent a violation of justice, especially under conditions where members of society are motivated by an egalitarian ethos.Even if this criticism holds, it is not clear whether it also applies to situations where the preferences of occupational choice of the talented do not match their socially efficient allocation (cf.Lamont 1997, 30).In short, even if the incentive payment to the talented person who wants to be a doctor anyway is unjustified, incentivising the talented person who would rather be a composer to become a doctor might still be justifiable.
However, the question of what constitutes an appropriate return to talent and thus might justify an incentive payment, while it has been one of the central issues in the literature on theories of distributive justice in recent decades, is not the one that will preoccupy us in this paper. 7econd, an "economic rent is earned by a factor input (e.g., capital, labour, etc.) when payment to that factor is in excess of the amount necessary to keep it in its current employment" (Lamont 1997, 28). 8If someone likes their job and would still do it even if their salary were lower, the difference between their actual wage and what is called their reservation wage-that is, the wage necessary to keep them in this job rather than switch to another-is called an economic rent. 9For economists, economic rents represent a good tax base, since removing them, by definition, will not have any distortionary effect on the labour supply of economic agents.From a normative perspective, it is also not clear why individuals should have a desert claim to an economic rent they earn. 10efore introducing a complication in our usual understanding of economic rents, a clarification is in order on the distinction between incentive payments and economic rents.Someone might point out that, in practice, it will often be difficult to disentangle the two phenomena: a particular wage might well contain elements of both economic rent and Note that this definition of the reservation wage incorporates strategic considerations.As Van Parijs puts it in his helpful discussion of different conceptions of factor rent, "[t]he reservation pay (the pay required to attract the factor) can exceed the opportunity cost (and hence the pay required to compensate the factor owner for forgoing the next best option) because the factor owner can credibly threaten to withhold the factor" (1996, 171). 10As Heath (2018, 18) rightly points out, Nozick's famous Wilt Chamberlain example is misleading in that is suggests that Chamberlain's income is a function of his talents rather than of the monopolistic structure of this particular labour market.See also Gauthier (1986, 274).incentive payments.11This is certainly correct, but we are nonetheless in a position to clearly distinguish the two at a conceptual level.First, whereas removing an incentive payment will lead to a "drop in the social product", because it makes the allocation of jobs less efficient, the "removal of economic rents usually causes no reduction in efficiency or the social product" (Lamont 1997, 29-30).12Second, economic rents can exist even when everyone's talents are the same, whereas the notion of an incentive payment would be nonsensical in such circumstances because, as a society, we would have no preference about who occupies what social role.
The standard view on the equity-efficiency trade-off makes two points on the issue of economic rents.First, it holds that making the labour market more competitive is sufficient for eliminating, or at least significantly reducing, economic rents. 13Second, it tends to downplay economic rents as a secondary issue in the more general context of how wages are set. 14his paper argues that this treatment of economic rents is unsatisfactory, because it omits an important factor from the analysis.The source of this omission lies in too narrow an understanding of economic rent.

DE GUSTIBUS DISPUTANDUM EST
Economists tend to take the tastes and preferences of economic agents as given.While they themselves argue against this position in their classic article "De gustibus non est disputandum", Stigler and Becker pithily summarise this widespread position when they state: "Tastes are the unchallengeable axioms of a man's behaviour: he may properly (usefully) be criticized for inefficiency in satisfying his desires, but the desires themselves are data" (1977, 76). 15references play a pivotal role not just in consumption theory, but also on the labour market.Among other things, the labour supply of individuals is shaped by a number of preferences including those on the work-leisure trade-off, the costliness of their consumption preferences and, importantly, individuals' motivation to participate in the labour market in the first place.
The central claim of this paper is that these preferences should not be taken as given, but should instead be brought into the purview of both our economic and our normative analysis of labour markets.Notably, I will inquire into the determinants or the origin of individual labour supply preferences.I shall argue that they are, at least in part, socially determined.If this is so, the next section suggests, this will require us to revise our understanding of the concept of economic rent.An important upshot of this argument will be that the trade-off between equity and efficiency is a lot less rigid than we assume when taking labour supply preferences as given.In other words, if labour supply preferences are somewhat malleable, a more equitable wage structure does not necessarily come at the cost of efficiency.

The Scope of the Notion of Economic Rent
What are the determinants of an individual's reservation wage and, thus, of what we consider the economic rent received by this individual?We can distinguish three such determinants-the first two of which have in common that they see economic rent as a return on scarcity (cf.Van  Parijs 1996, 172).
First, both the wage and the reservation wage of an individual are determined by the (perceived) scarcity of her skills. 16The harder it is to replace the skills of a particular individual, the greater the bargaining power this individual commands, and thus the greater the economic rent or, at the extreme, the monopoly rent, she will receive.Note that what matters for determining wages is perceived rather than actual scarcity of skills.
Second, both the wage and the reservation wage of an individual are determined by the scarcity that arises due to structural features of the labour market in which the individual works.As David Gauthier observes, we sometimes mistake scarcity due to market structure for scarcity of skills: Wayne Gretzky's talents command factor rent because they are scarce, but their scarcity is not a characteristic inherent in his talents, but a function of the conditions of supply, and so of the relations between his talents and those of others, and a function also of the conditions of demand, and so of the relations between his talents and the interest of others in attending hockey games (Gauthier 1986, 274).
Consider some concrete examples for the "conditions of supply" and the "conditions of demand" that Gauthier refers to.On the supply side, barriers to entry to certain markets can create artificial constraints on supply.For instance, in many countries, access to the medical profession is regulated by medical associations, thus handing market power and economic rents to those inside the profession.On the demand side, so-called winner-take-all-markets create situations where small differences in skill among suppliers get amplified into big differences in revenue. 17Classic examples here are sports and music markets: for example, there are plenty of hockey players who are not that much worse than Wayne Gretzky, but whose salaries will be significantly lower.
All of the phenomena discussed in the previous paragraph are examples of market failures.It is hard to overestimate the pervasiveness of market failure in the labour market.Many labour market economists recognise this fact (e.g.Boeri and van Ours 2008), but its implications for the equity-efficiency trade-off are less regularly taken into account.In particular, contrary to Heath's claims (see footnotes 13 and 14), the inherent imperfections of the labour market suggest that economic rents are often not significantly reduced, let alone eliminated, by market forces alone.If this is true, then they should play more than just a peripheral role in the debate about income inequalities.
Third, and this is the determinant missing from Heath's account, wages and reservation wages are determined by social norms governing the remuneration of different roles in society (see e.g.Elster 1989;  Atkinson 1997, 310-311; Piketty and Saez 2003, 34). 18Social norms influence wage-setting in two ways.
First, note that even the first two determinants, that is, scarcity due to skills and market structure, are already infused with social norms.What makes a specific skill scarce and able to earn a high salary depends on the norms of the society in question.For example, it is a socially contingent fact of our society that we value the skill of throwing a ball through a hoop.It is easy to imagine a different human society in which that skill, though distributed in the same way as in our society, attracts a much lower wage.We can also imagine a third society with again the same distribution of skills that values basketball like ours does, but where members are not as preoccupied to see the best players in action; as a consequence, winner-take-all-markets do not take hold.If it is correct that any economic notion of scarcity is already infused with social norms, this will have important ramifications for the meaning of an efficient wage. 19It would mean that identifying an efficient wage independent of social norms is impossible.Any labour-market equilibrium we observe in practice will always be sustained by some set of social norms. 20econd, social norms can perpetuate income inequalities that originate in scarcity due to skills or market-structure.Employers overestimate the skills premium certain individuals have over others; barriers to entry to an industry confer market power to insiders over outsiders; corporate executives exploit inadequate governance structures to get the board to grant them disproportionate pay packages; winner-take-all-markets concentrate market return among a few players; and so on.My claim is that, with time, these labour market imperfections acquire the status of social norms and entrench the inequalities in question.This claim looks like a version of the ratchet effect, that is, "an effect that occurs when a price or wage increases as a 18 Some readers will feel reminded of Cohen's (2008) argument about the importance of a social ethos in sustaining a given distribution of income in society.There are indeed obvious parallels here.However, while Cohen's argument is formulated in the context of a debate about incentives, the present paper focuses on economic rent (see the discussion in section 1.2). 19I thank both an anonymous referee and François Claveau for pushing me to develop the argument along these lines. 20Thank you to the same anonymous referee for suggesting this formulation.result of temporary pressure but fails to fall back when the pressure is removed" (Collins Dictionary, 2018). 21mportantly, instances of the ratchet effect are naturally concentrated at the top end of the wage distribution.Why?Because this is where, by definition, both bargaining power and market power are concentrated, and thus, are able to exploit the imperfections of the labour market in order to extract economic rent.
Presumably, Heath and perhaps some economists, too, will contest one central aspect of this analysis.They will claim that wage increases do not conform to the model of a ratchet effect because they will disappear due to increased competition. 22n response, let me distinguish two versions of my claim.The first, strong version confronts the standard economic perspective head-on and argues that once social norms about wage inequalities are in place, more competition will not necessarily be sufficient to dismantle them.In other words, pace the economic perspective, this position claims that we are indeed faced with a ratchet effect where, even when the initial cause of the wage inequalities disappears, its effects persist.I believe there is something to this strong version of the claim, but I will not attempt to defend it in this paper.
Instead, I shall defend a more moderate version of the claim.As evident from the above definition, the ratchet effect is based on the idea that some cause triggers a wage increase for certain people, and that the wage increase persists even when the cause is no longer there.The moderate version of my claim diverges from the last part of this analysis: economists suggest that wage increases of the kind under scrutiny here are temporary and will disappear under competitive pressure.But what if this competitive pressure is insufficient?At the heart of the moderate claim is the idea that the weak competitive pressure of imperfectly competitive labour markets is insufficient to correct for the market failures of these very markets. 23If it is a lack of competitive pressure on labour markets that gives rise to the wage 21 Economists have studied the ratchet effect extensively in the context of incentive contracts, see, for instance, Laffont and Tirole (1988). 22See, for instance, the experimental findings of Charness, Kuhn, and Villeval (2011).It is worth establishing the link here to Heath's discussion of the snow shovel example (2018, 22-27).Heath rightly points out that the higher price of snow shovels has a signalling effect and, by attracting more suppliers to the market and thus lowering the price again, will benefit both suppliers and customers.On the other hand, economists such as Boeri and van Ours (2008), who take seriously the imperfections of the labour market, are likely to be more sympathetic to my position. 23I should make it explicit that this claim is limited to labour markets.inequality in the first place, whence the confidence in competition to solve the problem? 24f course, it would be absurd to suggest that there is no competitive pressure on labour markets or that it is never sufficient for undermining situations of inefficient rent-seeking.But these claims are not necessary to establish my argument.All that is needed is the claim that market failure in the labour market-under certain circumstances, but systematically-leads to long-lasting wage inequalities.Yet, further research is needed to understand what causes wage inequalities to persist, including the circumstances under which they occur.What this section offers is a rough-and-ready list of four considerations that likely form part of the explanation.
First, ever since John Maynard Keynes ([1936] 1953) drew our attention to the downward rigidity of wages, behavioural economics has added an analysis of elements of human psychology that are consistent with this downward rigidity; in particular the 'endowment effect' (Kahneman, Knetsch, and Thaler 1991), which is based on the observation that people attribute more negative value to losses of what they already possess than they attribute positive value to an equivalent gain.This phenomenon might explain why managers are reluctant to actually cut anyone's wages for fear of a substantial negative impact on motivation.Second, even though Heath is right to point to the conceptual shortcomings of marginal productivity theory (2018, section III), the salary someone receives de facto is often perceived as an indicator of the contribution an individual makes to the organisation that pays them this salary.If this salary goes up, people will perceive them as harder to replace. 25The circularity of this reasoning notwithstanding, it helps to translate the initial wage increase into a wage inequality sustained by social norms.Third, bargaining power breeds other forms of power.As an example, consider the weight of lobby groups from medical associations or groups of other high-earning professions in the political system.Initially, they use their bargaining power to acquire market power, for instance in the form of controlling access to the profession.This market power subsequently spills over into other forms of social power, which then contribute to reinforce the social norm sustaining the wage inequality.Fourth, more generally, the vested interests of powerful groups in society tend to generate ideologies designed to legitimate and protect their interests.Such ideologies represent an important subcategory of social norms about wage inequalities and can contribute to explaining their robustness. 26n sum, I have argued that social norms influence wage-setting in two ways.First, scarcity due to skills and market structure is already contingent on social norms; second, social norms sometimes represent an independent determinant of wage inequalities.If these claims are correct, what are their normative implications?
Let me focus on two points here.First, if the above analysis is sound, it introduces an ambiguity into the notion of the reservation wage and, by extension, into the measurement of what counts as economic rent.If my argument about the role of social norms is valid, then one's reservation wage is contingent on social norms.In other words, the wage necessary to keep me in my current employment is not fixed, but varies with the social norms governing the remuneration of economic activities in my society.Though my skills are the same, my reservation wage will likely be lower in a society in which people in comparable jobs earn less. 27gainst this background, imagine two societies-one with a more egalitarian set of social norms about remuneration, S e , and one with a more inegalitarian set of norms, S i .The reservation wages (RW e ) of the high earners in S e will be lower than the reservation wages (RW i ) of their counterparts in S i , without a loss of efficiency.The claim of this paper is that the difference between the reservation wages RW i and RW e should be considered an economic rent.After all, when looking at S i , society could be arranged differently, building on a different set of social 26 Thanks to Colin Macleod for alerting me to this point. 27Some parts of the economic discipline-decision theory is the obvious exception here-have been reluctant to make conceptual room for the influence of other people's behaviour on our own.Consumption theory is one illustration of this point.Consumption theory usually stipulates that the satisfaction we derive from a good or service is something that is determined by the relation between us and the good or service alone.Some attempts have been made in the history of economic thought to fill this lacuna, but they have fallen by the wayside, presumably in part because they are hard to formalise.Consider, for example, Irving Fisher's (1892) idea that the utility we derive from the consumption of a good is in part a function of what quantity of this good is consumed by others; an idea that also underpins Thorstein Veblen's (1899) notion of conspicuous consumption.A parallel point can be made here in the context of production theory.Our social reservation wage is partly determined by the labour supply preferences of others.norms, without this arrangement being less efficient at allocating human resources to their best employment.
As a real-world example, think of the difference in wage dispersion between the Scandinavian countries, on the one hand, and the United States or the United Kingdom, on the other (e.g.Herr and Ruoff 2014).Or, think of the intertemporal differences in wage dispersion in the United States, contrasting the 1960s to now.Arguably, none of these different socio-economic arrangements are more or less efficient than the other, but they rest on different sets of social norms about remuneration.
The standard economic account does not have the analytical tools to allow for this variability in reservation wages.From the standard perspective, RW i is regarded as a parameter of the analysis.Granted, the standard perspective is right to claim that high-earners in S i are likely to reduce their labour-supply in the short-term if their wage falls below RW i .
However, there is no reason why they could not adapt to a different set of social norms over time and adjust their labour supply preferences to coincide with RW e .This is what I take to be the core claim of Keynes's quote cited in the introduction of the paper: the capitalist game "could be played" with "much lower stakes" than at present without compromising efficiency ([1936] 1953, 374).From this perspective, we may interpret the difference between RW i and RW e as contributing to economic rent, thus leaving us with a bigger economic rent.
In other words, it would be short-sighted to regard the reservation wage as fixed at RW i and treat it as a parameter of our analysis, because this obscures from view the fact that there are equally efficient wage distributions available below this level.What stands in our way to attain any of these distributions are social norms rather than efficiency considerations.It would be preferable to treat the reservation wage as a variable instead.Doing so reduces the rigidity of the equity-efficiency trade-off, because it makes vivid the fact that, while lowering someone's wage below RW i might change their labour supply in the short-term, it is in principle compatible with efficiency considerations based on scarcity, and thus, compatible with an unchanged labour supply in the long term.
This leads me to the second point.Heath indicates that when he refers to 'market wages', he means "the general tendency of markets to push wages toward the level at which supply of labour is equal to the demand of labour in a competitive market, and the price that this implies" (2018, 8).My analysis calls for one contextualisation and one amendment to this statement.Concerning the context, I have already emphasised that real world labour markets suffer from considerable imperfections.Concerning the amendment, my account shows that, in such circumstances, RW i and RW e come apart and that the level of wages considered efficient by the standard economic perspective (RW i ) in fact is one that is sustained by social norms.If this is correct, several things follow: First, imperfectly competitive labour markets do not converge on one efficient market equilibrium.Instead, there is a whole range of possible equilibrium wage levels.
Second, the case can be made that the equilibrium of the real world labour market is sustained by inegalitarian norms about remuneration.It would be a mistake to interpret departures from it as necessarily undermining efficiency.
Third, the equity-efficiency trade-off emerges from the above analysis as less of a constraint.To be sure, it still functions as a constraint, i.e. there is some level of reservation wages, below which we cannot lower the wages of high earners without compromising efficiency (this paper does not claim to identify this level).However, there is room for manoeuvre between RW i and that level.Thus, promoting what we consider to be just wage levels is more compatible with the demands of efficiency than the standard economic view suggests.This is the crucial difference between my paper and Heath's position.
What is more, and this is the fourth and final point, it is not just that efficiency and equity are compatible but, where RW i and RW e come apart, they in fact lead to the same policy recommendation.The equilibrium of actual, imperfectly competitive labour markets is not only unjust, but it is also inefficient in an important sense.The social norms that sustain it allow some members of society to extract economic rent, something that economists rightly condemn.

Labour Supply Preferences as an Instance of Expensive Taste
My aim in this final section is to return to the argument that individuals' reservations wages and their labour supply preferences are just that: preferences.I will make the case that if someone has a reservation wage above RW e , it should be considered an expensive taste.This argument requires two steps.First, I shall draw a parallel between expensive tastes in consumption and expensive tastes in production.Second, I shall argue that interpreting labour supply preferences with a high reservation wage as expensive tastes strengthens the case presented in the previous section, namely that justice requires pushing people's wages closer to RW e .One of the main objections against equality of welfare as a theory of distributive justice is the expensive taste argument (Dworkin 1981).If someone has very expensive preferences, say, they need rivers of champagne and tons of caviar to be happy, then it would be unjust towards other members of society with less expensive tastes to allocate more resources to this person.The underlying idea here is that tastes are a matter of choice, and that we should be held responsible for our choices.Put in economic terms, if I have an expensive taste, I should pay for the opportunity cost that satisfying this preference imposes on others.Champagne is expensive because there is lots of demand and limited supply.Consuming a bottle of this limited supply comes at a high price because it deprives others of the opportunity of consuming that bottle.
Note that we will hold the individual responsible for their taste for champagne even if their preference for champagne is socially conditioned.In other words, even if the reason that my happiness is contingent on champagne consumption lies in the fact that everyone in my social reference group drinks a lot of bubbly, we still in most cases consider that it is my choice to follow this trend and that I could have done otherwise. 28ow, consider the parallel between consumption preferences and labour supply preferences.Just as we want to hold the consumer responsible for their expensive tastes, we want to hold the producer responsible for their high reservation wage.In fact, there plausibly exists a connexion between the two kinds of expensive preferences, in the sense that people with expensive consumption tastes will need a high income to satisfy them.In this sense, high reservation wages can be seen as at least partly derivative of expensive consumption tastes.So, what exactly does it entail to hold someone responsible for their high reservation wage?
It means that if someone requires a high wage to get them to perform a task in society, this is a mere preference and, as society, we have no obligation of justice to satisfy this preference.We should not pay them this high a wage.If this strikes you as a strange way to think in a market context, I should hasten to add two potential caveats: to the extent that we consider wages as a return on scarce talent or skill, we may well consider it just to pay someone a higher wage.Similarly, to the extent that we consider incentive payments justified as a way to coax individuals into using their talents in certain socially useful ways, we may again consider it just to pay someone a higher wage.Recall from section 1.2 that I want to stay agnostic on this last issue in this paper.
However, we do not need to delve into these controversial issues to make the central point here.As the previous section has argued, there are wages on imperfectly competitive markets that exceed returns to scarcity due to skills or market structure-namely those that exceed RW e .If we focus the application of the expensive taste logic to this class of wages, then holding people responsible for their expensive labour supply preference, that is, not paying them this high a wage, is entirely compatible with the demands of efficiency as understood as a return to scarcity.
Just as on the consumption side, it does not matter whether my reservation wage is socially conditioned or not.In other words, the above logic applies even if I have a high reservation wage just because my peers do.
In sum, the argument in this short section suggests that it is time to extend our application of the logic of the expensive taste argument from the consumption context to the production context.Just like my expensive taste for champagne is not a reason to make more resources available to me for consumption, so my high reservation wage above the level required by economic efficiency is not a reason to pay me a higher wage.This analysis corroborates the conclusion of section 2.1, that pushing wages down to RW e , and challenging the social norms that support wages above this level, is not only a demand of justice compatible with efficiency but is actually required by efficiency, too.

CONCLUSION
Contra Joe Heath's paper at the heart of this symposium, I have argued for a less rigid version of the equity-efficiency trade-off.Up to a point, we can promote a more just distribution of wages without undermining the efficiency of the price mechanism.102 The usual way of framing the equity-efficiency trade-off in economics is not sophisticated enough in its treatment of individual labour supply preferences.Taking these preferences as given, it fails to account for the potential variability of reservation wages.To make this point tangible, I have distinguished between two societies, S i and S e , which are characterized (respectively) by inegalitarian and egalitarian social norms about remuneration.
I have argued that the difference between them should be included in our notion of economic rent.Reducing the wages of those privileged members of society for whom this economic rent is substantial, either by compressing pre-tax incomes or by taxing them, is not only compatible with considerations of efficiency, it is even required by them.
The final section of the paper has added an additional reason to think that wage premia above egalitarian reservation wages RW e are unjust.They are parallel in nature to expensive tastes in consumption.If we want to be consistent across the consumption and production sectors of our economy, then we should treat reservation wages in excess of RW e as expensive tastes and refuse to pay them.The position defended in this paper entails that equity and efficiency are more compatible than we tend to think.It agrees with economists about the importance of letting the price mechanism do its work in allocating resources in society efficiently.Where my analysis differs is in highlighting, as many labour market economists do, that real world labour markets are not only governed by efficiency but also by social norms.To the extent that income differentials that we consider unjust are underpinned by the latter, eliminating this injustice does not come at a cost in terms of efficiency.On the contrary, it is even required by it.