The savings problem in the original position: assessing and revising a model

Abstract

The common conception of justice as reciprocity seemingly is inapplicable to relations between non-overlapping generations. This is a challenge also to John Rawls’s theory of justice as fairness. This text responds to this by way of reinterpreting and developing Rawls’s theory. First, by examining the original position as a model, some revisions of it are shown to be wanting. Second, by drawing on the methodology of constructivism, an alternative solution is proposed: an amendment to the primary goods named ‘sustainability of values’. This revised original position lends support to intergenerational justice as fairness.

Keywords:

• Intergenerational justice
• fairness
• the original position
• models
• Rawls
• constructivism
• sustainability of values

Acknowledgements:

I am grateful to Clare Heyward, Maxime Desmarais-Tremblay, Nicolas Wüthrich, Joe Mazor and the anonymous reviewers of this article for their helpful comments on earlier drafts. I also wish to thank the audience of the PhD seminar at the Department of Philosophy, Logic and Scientific Method at the London School of Economics and Political Science for helpful discussion. I am also indebted to the Swedish Research Council for their generous funding of my postdoctoral project in which this article really is a sidetrack.

Notes

1. For a general introduction to ‘justice as reciprocity’ (see cf. Barry 1989, 463–493; Rawls 2009, 190–224); for a presentation of some of the problems it faces in the intergenerational setting (see Gosseries 2009).

2. There are models that do not assume non-overlapping generations, not least ones developed by economists. The best known probably is Paul Samuelson’s (1958) overlapping generations model, which shows that ‘cold and selfish competitive markets’ (Samuelson 1958, 473) do not lead to intergenerational savings without further complements. Also some philosophers have assumed overlapping generations (e.g. English 1977; cf. Gosseries 2009; Mazor 2010; Heath 2013). Among them, it is common to argue that the more ‘realistic’ starting point saves the contract theory from the above-mentioned problem. There is merit to this claim, which is something I will come back to below in the main text. But I will first consider the possibilities of working out intergenerational obligations under the more challenging assumption of non-overlapping generations.

3. ‘Save’ is a technical term in the text, which covers all kinds of positive intergenerational transfers. Apart from direct savings, it includes various investments (e.g. in technology and in infrastructure projects), and also measures aimed at maintaining the existing stock of goods and preventing future threats (e.g. climate change mitigation and measures to control inflation).

4. This problem should not be confused with a more general one, sometimes called ‘the problem of the first generation’. This is a challenge not primarily to reciprocity-based views, but to any inequality aversive view of intergenerational ethics, such as egalitarianism, prioritarianism and utilitarianism. It arises out the same features as noted above, plus the assumption of economic growth. Thus, people can only improve the situation of their descendants (not that of their predecessors), and if they do, then their descendants will, needless to say, end up better off than them. According to the just mentioned ethical views, there are moral reasons to redistribute resources to improve the situation of those in the worst off position. Now, in this setting, this seems to imply that resources ought not to be invested in intergenerational savings at all. That is because to not save seems to prevent the worst off, first generation, from being exploited by having to give to their better off descendants. It thus seems as if not only is it the case that the first generation have no rational reasons to save because of the lack of reciprocity, but also that they neither have any moral reasons to save, or if they do, it is one to not save. The problem can be generalised in the following way: for any subsequent generation, if its predecessors have not saved for it or (even worse) debited it, then it is effectively the worst off generation and so morally required to not save in order to not further worsen its situation.

5. From when it was first presented until just recently, Rawls’s account of intergenerational justice has been discussed and criticised, see e.g. (Hubin 1976; English 1977; Barry 1978; Gardiner 2009, 2011; Heyd 2009). But I will argue that something is missing in the discussion. That is, to develop the extension on the basis of the general approach to normative theorising Rawls calls constructivism. This has not been pursued in the debate; even Rawls himself seems to have missed the potential of his constructivism to facilitate a more natural extension of his theory to the intergenerational setting.

6. Rawls’s presentation of the original position changes somewhat over time (for an overview, see Freeman 2014). For instance, he answers the question of what is modelled differently: at one point, it is ‘the way in which the citizens in a well-ordered society, viewed as moral persons, would ideally select first principles of justice for their society’ (Rawls 1980, 520); at another it is ‘what we [which just before has been qualified with ‘here and now’] think on due reflection are the reasonable considerations to ground the principles of a political conception of justice’ (Rawls 2001, 17). I do not think that these are substantial changes though, just shifts of emphasis.

7. In the introduction, I noted that no one has made use of the constructivist methodology to develop the revisions required of Rawls’s theory of justice in the intergenerational setting. I should, however, add that Attas (2009) at least mentions it. He writes: ‘This kind of tinkering with the details of the original position is in line with Rawls’s general methodological approach of constructivism […] Much of Rawls’s discussion on justice between generations in Theory and the critical literature that addressed it can be seen as an exercise in constructivism: a way of modifying the initial situation so that principles generated will take adequate account of future generations’ interests too’ (Attas 2009, 192). This is not very helpful though. Attas assumes an understanding of constructivism that is much too truncated to provide a useful approach to these questions. If anything, it comes out as question-begging by giving the impression that the ‘tinkering’ is done to produce an already decided outcome.

8. For an early, but still very illuminating discussion of constructivism, see (Darwall, Gibbard, and Railton 1992, 137–144).

9. But see Brian Barry (1989, 506) who claims, in a kind of tu quoque-argument, that the difficulties are equally featured in both the present-time-of-entry and general-assembly interpretations, and that it is ‘a point in favour of’ the latter ‘that it brings out the difficulties graphically’.

10. See also (Heyd 2009, 173). For two different attempts at responding to similar problems on behalf of the general-assembly interpretation (see Barry 1989; Gardiner 2009).

11. It should be noted that the motivational assumption leaves intact another assumption Rawls makes of the parties, namely that they are mutually disinterested, that is ‘conceived as not taking an interest in one another’s interests’ (Rawls 1971, 13). That the parties are seen as heads of families, caring about the well-being of their closest descendants, does not mean that they are assumed to take an interest in any other party’s interests. In other words, they are only assumed to care about their own descendants. Some commentators (e.g. Attas 2009, 197–198) have mistakenly pinpointed mutual disinterestedness as the object for the revision; interestingly, also Rawls (e.g. Rawls 2001, 160).

12. One can also think about people who choose not to have children: are they still required to care about the future? One possible remedy to this is Clayton Hubin’s proposal of a ‘psychological’ rather than motivational assumption (Hubin 1976). That is, to assume that it is common knowledge that people in general care greatly about their offspring. If this psychological fact were known among the parties in the original position, then it would be rational for them to choose a savings principle to prevent the event that they turn out to be a caring parent once the veil of ignorance is lifted. This solution is an improvement, but still problematic in that it tethers justice too closely to the limits of care (cf. Barry 1978, 227–228).

13. One possibility is that even if the motivational assumption does not fully solve the problems considered, it is the best way can hope for, given the circumstances at hand. We should perhaps be content with a statement about the value of maintaining background justice, as David Heyd (2009) argues. This might be right if we consider the prospects of a truly intergenerational contract, grounded on intergenerational cooperation; but it is wrong for the relevant sense of the target problem, as I will argue in the subsequent section.

14. Another possibility is pursued by Daniel Attas (2009). He argues that the solution can be found in the already existing constraint of universality, which, in Rawls’s (1971, 132) words rules out a principle that is ‘self-contradictory, or self-defeating, for everyone to act upon it’, as well as a principle that is ‘reasonable to follow only when others conform to a different one’. The problem with this possibility is that a no-savings choice is not self-contradictory or self-defeating: there is no contradiction in all generations, or all individuals within one generation choosing to not save. Nor is such a choice only reasonable on the condition that previous generations have saved: the prisoner’s dilemma structure of the problem makes it rational for the parties to not save irrespective of what previous generations have done, and that is true for each generation. In other words, each generation can adopt the universal principle of no savings without having to assume anything about what other generations have or have not done.

15. A fourth reason, or set of reasons, concerns what looms in the background of additional constraint, namely: Kantian ethics. Barry (1978) argues that Rawls exhibits his Kantian persona here and that it is in tension with his Humean persona, as manifested in the doctrine of the circumstances of justice. If this is right, there are perhaps other reasons to worry about additional constraint – both because it gives up the reciprocity ground and because it takes on general problems related to Kantian ethics. Think, for instance, of the common critique of empty formalism: Can we derive substantial norms, such as principles of intergenerational justice, just by considering formal features of agency?

16. This is, of course, premised on the specific conception of the person as free and equal, with the two requisite moral powers, Rawls assumes in justice as fairness. There may be reasons to doubt some of the details of that conception, but it would take as too far astray to that now.

17. See also Raz (2003). It should be noted that Scheffler does not relate his proposal to Rawls’s theory. The idea of sustainability of values as a primary good is, though inspired by Scheffler’s book, my own proposal.

18. Which I believe is too strong. There are plenty of values (e.g. the goodness of wine or the beauty of sunsets) that are intrinsically satisfactory (cf. Raz 2003, 128–131) and would continue to be so even in the afterlife scenario. See also the discussion of the first worry about sustainability of values in the main text.

19. I thank an anonymous referee for pushing me to clarify this point.

20. For a more elaborated discussion of a similar point, see Raz (2003, 119–121).