Version classiqueVersion mobile

Arguing about justice

Axel Gosseries
Philippe Vanderborght

Why auntie's boring tea parties matter for the fair distribution of gifts

Robert van der Veen


Het pleidooi van Philippe van Parijs voor een maximaal onvoorwaardelijk basisinkomen stelt voor dat de opbrengst van ongelijk verdeelde giften zoals erfenissen en schaarse banen - wordt afgeroomd door belasting, ten bate van de reele vrijheid van de minstbevoorrechten. Met behulp van Ronald Dworkins egalitaire veilingmodel beredeneert Van Parijs dat de billijkste herverdeling van deze belastingopbrengst iedereen een gelijk aandeel toekent, ongeacht de bereidheid tot het verrichten van werk. In dit hoofdstuk laat ik echter zien dat in zon egalitaire veiling sommige giften een prijs zullen krijgen waarin een vergoeding is verdisconteerd van het werk dat nodig is om de inkomsten van die giften te genieten. Neemt men deze veilinguitkomst in aanmerking, dan volgt dat een billijke herverdeling althans ten dele gebonden moet zijn aan de bereidheid tot werken.

Note de l’auteur

Note portant sur l'auteur1

Texte intégral


1Philippe van Parijs's real-libertarian case for maximum sustainable basic income starts by asserting that the metric governing the just distribution of real freedom should be the competitive value of diverse rent-bearing gifts to which people have unequal access in the real world. Van Parijs shows that predictably taxing all market sources to secure the highest sustainable revenue approximately captures the part of the aggregate value of gifts needed to satisfy the principle of maximin real freedom. He then argues that an equal and unconditional distribution of the tax is the only fair way of sharing out this value. Using a simple example featuring inheritances and scarce jobs, I dispute this last view. When the competitive value of gifts is determined in an equal token-auction by people's evaluations of the time and effort they must spend in order to derive benefit from some gifts (notably, but not exclusively, scarce jobs), the fair way of sharing out the tax yield must include a work-conditional component. This need not rule out a basic income, but in general it does rule out dispensing the largest sustainable one.

Implementing the maximin real freedom objective: an issue of fairness

2Real freedom - the freedom to do what it is one might want to do - is an opportunity concept involving several dimensions. Chapter 4 of Real Freedom For All (Van Parijs 1995) notes that maximizing the real freedom of the least advantaged can be implemented in two ways, looking at two key dimensions: the freedom to earn income, and the freedom to consume free time. The tax yield available for redistribution can be distributed equally as unconditional basic income, or in the form of equal rights to work for those who are able and willing, and generous income grants for those who are willing but unable to work (Van Parijs 1995: 111, 126). If the work-conditional scheme offers more opportunity to earn and less opportunity to consume free time than the basic income scheme does, then the real freedom conferred by each scheme is differently shaped, in a way that makes neither evidently superior to the other in terms of its capacity to deliver maximin real freedom. Both schemes, then, can be optimal from the point of view of real-libertarian distributive justice. The remaining normative issue is to select the fairest of the two. Van Parijs holds that it must be settled on grounds of liberal neutrality, from the perspective of a ’liberal ban on discrimination between conceptions of the good life’ (Van Parijs 1995: 109). As a matter of fairness, the government should adopt a tax sharing scheme that takes a neutral stance towards people with different preferences for earning income and consuming free time, a stance that gives no special privilege to one dimension of real freedom over another.

3But what, exactly, constitutes this neutral stance? Here we must consider the causes which call for redistribution: the arbitrarily unequal distribution of rent-bearing gifts in the real world. To assess the total amount of this rent, we should start by imagining all such gifts being put up in an equal-token auction. In particular, as Van Parijs explained in a recent lecture (Van Parijs 2009: 12), we should

… look directly at jobs and other market niches as incorporating very unequal gifts to which we are given very unequal access by a messy combination of factors. It is these gifts, and not only the much smaller amount that takes the form of donations and bequests that should be made the object of a Dworkin-like auction. This is the key distinguishing feature of the approach proposed in Real Freedom for All.

4Auctioning the gifts means that their values are expressed by a metric of opportunity costs, that is, exchange rates reflect the cost to others of what each gift commands. The tax revenue up for redistribution is to be expressed in this neutral metric. As Van Parijs (2009: 12) puts it:

If and only if this metric is adopted, we get a strong presumption - in the context of a discrimination-free market economy regulated in such a way that prices track opportunity costs - in favour of a universal cash income unconditionally granted to all and paid for out of the predictable taxation of all market activities.

What form of redistribution does liberal neutrality require?

5Van Parijs’s case for maximum sustainable basic income may be clarified by the following argument. The neutral perspective for deciding on the form of redistribution regards the revenue captured by the appropriate tax program as a part of the total value of different gifts determined in a ’Dworkin-like’ auction. From this starting premise, the choice in favor of the basic income regime is spelled out in two steps. First, because the auction metric takes account of everyone’s preferences from a baseline of bidding with equal tokens, it neutrally values the total rent flowing from unequally distributed gifts in the real world where these preferences are held. Secondly, by expressing the tax available for redistribution in this metric, the corresponding gift rent is a sum of values to which no one has any prior claim. Therefore the neutral way of dispensing that sum must be equal and insensitive to people’s preferences. Unlike a work-conditional scheme, maximum basic income meets this requirement of preference-insensitivity. It is therefore the neutral way of achieving maximin real freedom.

  • 1 One might object to the first step of the argument, if the auction participants are differently tal (...)

6For the purposes of this discussion, I accept the reasoning of the argument1 But I reject its conclusion, because I think the starting premise is open to challenge. To explain the challenge, it is helpful to consider an objection which Van Parijs has discussed several times, concerning the inclusion of scarce jobs in the auction. In a recent formulation, he says (Van Parijs 2009: 13):

Needless to say, this assimilation of jobs to gifts is not uncontroversial. Is it not undermined, for example, by the fact that one generally needs to do something in order to get a job and keep it? This undeniable fact does not create a fundamental difference with donations or bequests. Attending politely your aunt’s boring tea parties may be one of the necessary conditions for you not to get forgotten in her will. But this investment of yours does not make you ’deserve’ the whole of the big chunk of wealth possessed by a person to whom you happened to be related. Similarly, the fact that one needs to go to the office every morning and busy oneself once there does not make one ’deserve’ the whole of the salary one is able to earn by virtue of a combination of circumstances most of which are no less arbitrary than the fact that one of our parents happens to have a rich sister.

7This reply to the objection that jobs are not gifts is sound. Surely the scarce job offering a lucrative reward and nice working conditions is a rent-bearing asset, just as is the large bequest which requires you to attend auntie’s boring tea parties before she dies. But the reply also suggests that what one needs to do in order to get the benefits of different types of gift - to fill and keep the job, or to visit your aunt a number of times - must show up in the auction. The prices of work-requiring gifts depend on how bidders assess the work compared to the option of buying gifts which require no work. Now my intuition is that these comparative assessments should be tracked by a neutral scheme for sharing the rent captured by taxation.

8To explain this intuition, I grant that the tax revenue originating from jobs or other work-encumbered gifts is indeed rent in the sense that no one has a prior claim to it. Yet, that revenue is relevantly different from rent which originates from income or wealth coming in just like manna from heaven. Obviously, the people whose job salaries get reduced by taxation first had to perform activities that are of value to others. Something similar holds in the case of the auntie inheritance. That a part of my ’big chunk of wealth’ gets added to the total sum up for redistribution is due to the fact that I engaged (for selfish or other reasons) in a kind of care work which benefits my old auntie, indeed to the point that she wanted to include me in her testament in appreciation of it. Once we view these ’undeniable facts’ to which Van Parijs calls attention in this light, it is reasonable to think that the recipients of redistribution might be asked to perform similar socially useful activities in return for at least part of the benefits to which they are entitled to as a matter of justice. This could mean not sharing out the entire tax revenue unconditionally.

9Let me add that my reason for entertaining this thought is not inspired by a principle of egalitarian reciprocity requiring that potential beneficaries should in general be willing to contribute to the upkeep of redistributive arrangements. Rather, the thought is motivated by the kind of liberal neutrality which leads Van Parijs to introduce the gift auction, the exchange rates of which reflect the cost to others of what each gift commands after equally endowed gift holders have made their final bidding choices. So just like Van Parijs, I proceed from the initial idea that the auction neutrally values the total rent flowing from unequally distributed gifts in the real world. But I add that the outcome of the auction may also tell us more specifically which part of that total rent arises from rewards to gift work. If this can be worked out, then there could be a neutral reason for sharing out the rent taxed away from gift holders in the real world according to conditionalities that reflect the auction’s comparative evaluation of the work which is tied to gifts.

10Having spelled out my intuition that ’comparative assessments of gift work in the auction should be tracked by a neutral scheme for sharing the rent’, it should be noted that once the starting premise of the argument for basic income is accepted, there is simply no room for this intuition. For, following the starting premise, we have to express the Euro amount of tax in the values of the auction by means of the conversion rate of the auction numeraire (the gift in which prices are expressed) to the Euro. It is then impossible to determine what part of that value sum can be attributed to the work required by different gifts. Accordingly, the tax yield gets represented as an undifferentiated amount of gift rent, the size of which is determined by the quantities and prices of all gifts. This naturally suggests that the neutral distribution must grant everyone the per capita share of tax. The two-steps reasoning of the argument merely fleshes out that suggestion.

11Now the challenge I have in mind is not that Van Parijs’s argument for maximum sustainable basic income is invalid. It is rather that its starting premise construes the linkage between the tax yield and the proceeds of the auction in an arbitrary way. For one can just as well express the tax yield as a percentage of the auction proceeds measured in the currency of the real world (here: Euro), rather than as a percentage of the proceeds measured in auction currency. As the example below will show, part of the Euro tax revenue can then be imputed to the rewards of gift work, on the basis of the auction prices. And as I have explained above, this introduces a presumption against equal and unconditional sharing of the tax revenue.

How Euros of gift work depend on auction prices: an example

12Two inheritances of ten thousand Euro are auctioned. The benefits of the first - call it the F-gift - come in free, no strings attached. The F-gift figures as the standard unit of auction prices in this example. The second inheritance is the A-gift. It requires doing a spell of auntie work which most bidders find genuinely boring, and would rather not do free of charge, even after considering the real benefit to auntie’s well-being. The auction price a of the A-gift will then be less than unity. I spend all my tokens to buy the right to one F-gift. You are willing to engage in auntie work at the going price, and spend all your tokens on A-gifts. Having done the work, you end up with 1/a times ten thousand Euro, (1/a – 1) times ten thousand Euro more than I get. This difference represents your auntie work reward. Between the two of us, 1 + 1/a times ten thousand Euro gets added to the auction proceeds. This equals your work reward, plus two times ten thousand Euro free of charge. Next, introduce a third person. She’s willing to busy herself with the work required by a scarce parttime job fetching ten thousand Euro, which is priced in the auction at s units of F-gift currency, with s < 1. She spends all her tokens on these S-gifts. This adds 1/s times ten thousand Euro to the auction proceeds. Her job work reward comprises (1/s - 1) of that. Table 1 below sums up this three-person example.

Table 1. Imputed Euro work rewards when each of the 3 persons spends the equal tokens on a different type of gift, at auction prices a and s

13The notion of Euro rewards for gift work has a significance which is explained by reference to Dworkin’s idea that an equal-token auction satisfies the normatively attractive condition of envy-freeness when bidders’ internal resources are equal, as is assumed here. In the example so far, each of us prefers his or her own chosen bundle of gifts to that of the others, taking into account the characteristics of the gifts and the prices which they command. In particular, the fact that I - buyer of one unit of F-gift - do not envy your bundle of 1/a units of A-gift, and do not envy her bundle of 1/s units of S-gifts, is explained by the comparative judgement that I’d gladly forego the extra income received by each of you two, because I’m not willing to do the corresponding amounts of auntie and job work at the auction prices of these gifts. Likewise, that both you and her are willing to earn the extra income in return for the specified amount of auntie or job work explains why both of you two do not envy my bundle of gifts. Thus the portion of the Euro auction proceeds attributable to work rewards expresses the joint impact of bidders’ decisions to trade off leisure against additional income from work-requiring gifts. Conversely, the portion of the Euro auction proceeds which is not attributable to work rewards expresses the joint impact of bidders’ decisions to trade off additional income against leisure. This division of the Euro auction proceeds is normatively supported by the auction’s envy-freeness.

14Thus it makes sense to model the division by extending the three-person example to a whole population in the simplest possible way. Imagine that each of us (me, you and she) belong to groups making up one third of the population. Members of groups F, A and S each demand only F-, A- and Sgifts. Now, to arrive at the prices a and s at which the auction clears, the quantities of gifts demanded by the three groups must equal the auction’s gift supply. This is so when the per capita available quantities of these gifts are 1/3, 1/3a and 1/3s, respectively.

15Now do some accounting, starting with the auction metric. The value of average auction proceeds expressed in F-units amounts to one unit (i.e. 1/3 available unit of F-gift, to which one must add the value of 1/3a of the A-gift priced at a, and the value of 1/3s of the S-gift priced at s). Note carefully that within the auction metric, there is no way of determining reward rates for auntie or job work. For if we compare an F-bundle to a value-equivalent Aor S-bundle of gifts, then the work reward expressed in value is zero by definition. In the value metric therefore, comparative evaluations of gift work are indirectly incorporated in the prices of A- and S-gifts.

16Next, consider the average auction proceeds in Euro. This figure is obtained by multiplying the available quantities of gifts per head by ten thousand Euro: 10000(1/3 + 1/3a + 1/3s). Following the reasoning above, we introduce the Euro work rewards accruing to members of the A- and Sgroups: 10000(1/a - 1) and 10000(1/s - 1) respectively. Of the average Euro proceeds, 10000(1/3a + 1/3s - 2/3) Euro therefore represents the rewards of gift work (i.e. of 1/3a units of auntie work and 1/3s units of job work required per head). Call this part of average Euro proceeds the work-related gift. The remaining part of 10000 Euro can be regarded as a free gift. It is the part of the proceeds not attributable to gift work. To provide a numerical illustration, take a = 2/3 and s = 2/7. Then average auction proceeds adds up to 20000 Euro. It divides into 10000 Euro of work-related and free gift each.

17On the basis of this accounting, units of auction currency change into Euros by the ratio 10000(1/3 + 1/3a + 1/3s) to 1. So with the numbers above, one F-unit is worth 20000 Euro. Now we have two alternative ways of calculating the tax revenue. Taxing away the spending power from unequally distributed gifts in the real world negatively affects the propensity to save and bequeath, and destroys incentives for creating scarce jobs. So after economic agents have adjusted to the tax rates, the optimal tax program cannot capture the entire proceeds of the auction. To avoid complications involving preference changes during the adjustment period, assume the auction is run every year with the outcome described above. Suppose that 10000 Euro per head are up for redistribution annually, after collection of tax.

The neutral status of maximum basic income: two conflicting views

18What is the neutral distribution of the gift rent captured by tax? On the premise of Van Parijs’s argument, we must start by converting 10000 Euros of tax revenue into the auction metric. This gives us half of a F-unit, 50% of the value of average auction proceeds. Once the tax yield is assessed in the auction metric, it necessarily appears as undifferentiated gift rent. On the two-steps argument, maximum basic income wins. However, if we regard the tax yield as 50% of average Euro auction proceeds (i.e. half of 20000 Euro, given a = 2/3 and s = 2/7), then an equal sharing scheme following the division into the free gift and the work-related gift is the winning candidate. To operate such a sharing scheme requires that the government should ensure opportunities for work closely similar to the types of work of our gift example. The scheme then dispenses 5000 Euros unconditionally, and gives everybody a maximum of 5000 Euros in return for up to half the average amounts of gift work required in the auction situation.

19Since this division of the tax to be shared out depends on the prices of Agifts and S-gifts, the partly work-conditional scheme respects the neutral valuation process of the auction in the following sense. Consider beneficiaries of redistribution who do not choose to perform work similar to gift work, when given the opportunties to do so at the auction’s Euro reward rates. In the context of our example, we may call such people the ’voluntarily unemployed’. They are relevantly like the members of the F-group of bidders in the auction. Since the latter do not envy the gift bundles of the two other groups, they acknowledge that they have no fair claim to the work-related part of the auction proceeds. Therefore, all the voluntarily unemployed can claim under a neutral sharing scheme is the share of tax revenue which corresponds to the free gift part of the auction proceeds. The partly work-conditional sharing scheme follows this judgement. For this reason, I claim, it beats maximum basic income on grounds of liberal neutrality.

The role of the auction

20Two conflicting claims about the liberally neutral status of maximum basic income are on the table. They depend on different ways of presenting the very same facts about how the tax yield can be linked to the outcome of the auction specified in our example. Since nothing substantive turns on these different presentations, it is not obvious which of the two claims one should endorse. But we can make headway on the issue by referring to Van Parijs’s statement about the auction metric quoted earlier: "if and only if this metric is adopted, we get a strong presumption (…) in favour of a universal cash income unconditionally granted to all and paid for out of the predictable taxation of all market activities." How should this statement be interpreted? In light of what I argued above, there are two possibilities.

  1. ”Adopting the auction metric” is narrowly interpreted to mean that the tax yield must be expressed in units of that metric. In this case, the two-steps argument supports the rest of the statement.

  2. ”Adopting the auction metric” is interpreted more broadly to mean that the outcome of the auction, as revealed by its value metric, should be authoritative for deciding how the tax yield should be shared. In this case, the rest of the statement does not necessarily follow.

21In my view, interpretation (2) is a more reasonable way of explicating the normative role of the auction. It also gives purchase on how to decide between the two conflicting claims. Because the tax yield can be linked to auction prices in two different ways, we have to look at what role these prices play in each of them. When tax is expressed as a percentage of the value of average auction proceeds, the role of the auction outcome is very limited. Prices are used only to fix the rate at which Euros of tax convert into auction currency. By contrast, when tax is expressed as a percentage of the Euro value of average auction proceeds, gift prices determine the Euro rewards of different kinds of gift work. This assigns a far more significant role to what goes on in the auction from which we are seeking guidance.

22To bring out this contrast more vividly, consider two variants of our job-inheritance example. They are based on the fact that the same value of average auction proceeds can either be represented by different available quantities of the same gifts, or by a package of different gifts. Recall that our original example was built by assuming that 1/3, 1/3a and 1/3s units of F- Aand S-gifts were available per head respectively, and that the population consisted of three equally sized groups spending their tokens exclusively on one of the three gifts.

23The first variant retains this last assumption, but it assumes that only 1/3 units of each gift are available per head. The auction then clears at prices a = s = 1, hence maximum basic income is the neutral sharing scheme. For with these new quantities, the size of the demand for A- and S-gifts now tells us that both auntie work and job work are held to be pretty attractive ways of spending one’s time. As a result, average auction value - which as before, is one unit of F-gift - converts into 10000 Euro, because the work-related gift component of average Euro proceeds is reduced to zero. In this (unlikely) special case, it does not matter in which currency you express the tax yield. In the second variant, there is ’no such thing as a free gift’. It is constructed by dropping the F-gift, and taking the A-gift as unit of auction currency. Suppose two equally sized groups each purchase only A- and S-gifts, and that . and 1/2s units of these gifts are available per head respectively, with s’< 1. The positive Euro reward of job work depends on s, the price of the Sgift at which the auction clears. Now that work is socially evaluated relative to the auntie work required by the A-gift. Therefore the neutral sharing scheme rejects unconditional distribution of any part of the tax yield. In the other method of calculation nothing changes. The value of average auction proceeds is one unit, so the neutral way of sharing the undifferentiated rent represented by the tax yield expressed in A-units remains the same: equal and unconditional.

Concluding comments

  • 2 For all variants of the original example with auction prices of work-requiring gifts exceeding unit (...)

24The argument of the last section shows why the method by which Van Parijs assesses gift rent does not pick up normatively significant features of the auction valuation process. Therefore I believe it is not the right method. Applied to gifts with different characteristics, the auction is a counterfactual state of affairs. It describes how those gifts get allocated, when equally endowed people with the same preferences for earning income and consuming free time as obtain in the real world bid on the available quantities. The point of the auction is to work out what the fairest way to share out tax revenue in the real world would be, by taking the final expression of people’s free choices under the bidding process as a normative standard of reference. As argued in van der Veen (2004, section 4) we use the auction device by searching for a redistributive scheme, the result of which approximates the reference outcome as closely as possible. The price-sensitive conditional scheme I proposed can indeed perform this job. Equal and unconditional sharing cannot possibly approximate the reference outcome, except in the very special case where all auction prices are at unity, as in the first variant of our example mentioned above2

25Surely we must admit that it is quite hard to ascertain the exact outcome of an imaginary gift auction. Yet, if there are good reasons to think that work tied to some gifts is burdensome, just like attending auntie’s boring tea parties was made out to be in Van Parijs’s story, then the auction device tells us that the non-discriminatory way of redistributing for purposes of real-libertarian justice must include a work-conditional component of some magnitude. This need not rule out an unconditional basic income. But it does rule out dispensing the maximum sustainable one.

26Finally, while I have focused exclusively on gift work in reaching this conclusion, my reasoning can be extended to other relevant sources of cost attached to gifts. Indeed one can make a case for distributing tax revenue originating from gifts with radically uncertain monetary benefits in the form of fair bets, that is, conditionally on willingness to accept uncertainty. The form and size of these bets can be derived from the specfication of the relevant gifts, and from their auction prices. Of course, just as is the case for any particular work-conditional sharing scheme for gift rent recommended by the auction, it remains a separate question of social assessment whether one wants to accept the consequences of such a scheme as a defensible redistribution policy, all things considered. These matters are beyond the scope of this contribution. The extension just mentioned, however, does provide another instance of how the liberal strategy of using the auction device I have defended here is unlikely to issue in a conclusion in favor of maximum sustainable basic income.



VAN DER VEEN, R. (2004), 'Basic income versus wage subsidies: competing instruments in an optimal tax model with a maximin objective', Economics and Philosophy, 20 (1): 147-183.

VAN PARIJS, Ph. (1995), Real Freedom for All. What (if anything) can justify capitalism?, Oxford: Oxford University Press.

VAN PARIJS, Ph. (2009), Basic Income and Social Justice - Why Philosophers Disagree, Joint Joseph Rowntree Foundation / University of York Annual Lecture, March13, 2009.


1 One might object to the first step of the argument, if the auction participants are differently talented in ways that make some gifts unequally accessible because the work tied to these gifts requires special abilities that not everyone possesses (see Van Parijs 1995: 123-124). In this note I abstract from the complications this raises.

2 For all variants of the original example with auction prices of work-requiring gifts exceeding unity, the neutral recipe would be to distribute a larger unconditional basic income than per capita tax revenue, and finance the difference by selling off some rights to work in gift occupations at auction-determined rates of ‘negative reward. Even these (highly unlikely) reference outcomes, however, do not coincide with the simple formula envisaged by Van Parijs. As Kasper Lippert-Rasmussen has pointed out, moreover, something similar follows when one considers auctioning off scarce opportunties for spending free time in a very pleasant way, e.g. surfing at Malibu Beach.

Notes de fin

1 I thank the participants of the Workshop on Political Theory (Aarhus, 12-13 May 2011) for valuable comments on a draft of this paper, especially Matthew Clayton, Kasper Lippert-Rasmusssen, Soren Midtgaard and Lasse Nielsen.

Table des illustrations

Légende Table 1. Imputed Euro work rewards when each of the 3 persons spends the equal tokens on a different type of gift, at auction prices a and s
Fichier image/jpeg, 88k

© Presses universitaires de Louvain, 2011

Licence OpenEdition Books

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search