Christian Schubert (Reference Schubert2015) has reviewed a series of papers in which I have investigated the problem of reconciling normative and behavioural economics and have advocated the use of opportunity as a normative criterion. Naturally, I am flattered that my work has been given this attention. Since Schubert's main criticism of this work is one that has often been put to me informally but has not previously been written down, I welcome the opportunity to write a short Reply.

Schubert's presentation of my work on opportunity is accurate and fair, but his sense of what is more and less central to it sometimes differs from my own. So I will begin with a brief overview of my project, as I understand it. This project is still in progress. The papers that Schubert reviews should be read as interim reports; I make no claim that, in combination, they describe a fully consistent position whose every detail I can now defend. This Reply reflects my current thinking.

My principal motivation has been to try to build a form of normative economics that is consistent both with the liberal tradition of economics and with the findings of behavioural economics. In what I take to be the liberal tradition, there is a presumption in favour of allocating resources through individual rather than collective choice, and thus in favour of the market and market-like mechanisms. This presumption is expressed in its most abstract form in the First Fundamental Theorem of welfare economics, but it also underlies many aspects of applied normative economics. For example, competition policy typically treats competitive markets as normative benchmarks; economists have generally favoured the creation of new market mechanisms, such as carbon trading, to deal with problems of public goods and externalities; cost-benefit analysis is often interpreted as ‘market simulation’. All of this is underpinned by theories that use preference-satisfaction as their normative criterion, on the assumption that individuals act on coherent and stable preferences. However, that assumption is called into question by the findings of behavioural economics. It has long seemed to me that the failures of the conventional theory of rational individual choice are not isolated anomalies or merely the results of errors in decision-making; they are systemic. Economics has to face up to the possibility that the best explanations of human decision-making may have no use for the concept of coherent preferences. If that proves to be the case, must the presumption in favour of markets be given up too?

My project starts from the hunch that the answer is ‘No’. Its guiding intuition is encapsulated in the title of my first paper on reconciling behavioural and normative economics: ‘The opportunity criterion: consumer sovereignty without the assumption of coherent preferences’ (Sugden Reference Sugden2004). Schubert is puzzled that I defend the concept of consumer sovereignty, which he sees as ‘firmly embedded in a welfarist framework of preference-satisfaction’ (footnote 4). But, as far as I know, this usage of ‘welfarism’ was coined by Sen (Reference Sen1979) as part of a philosophical critique of conventional welfare economics. How far the practitioners of welfare economics have actually seen their work as welfarist is open to question. On Sen's reading, the object of welfare economics is to arrive at judgements about the overall goodness of states of affairs for society; welfarism is the assumption that goodness is an increasing function of the utility of each individual, ‘utility’ being interpreted as a representation of a preference ordering. But is this what economists have meant by ‘consumer sovereignty’? Or have they had in mind the tendency of the market to supply each consumer with what he or she wants? The latter seems a more natural understanding of ‘sovereignty’ than the idea that each individual's preferences count positively in some synoptic judgement about the social good. A sovereign does not report a preference to some ethical observer who then considers how much weight to give it: she expresses her will, and her subjects obey. As I have argued in another paper, the core idea to which economists have appealed when they have argued that competitive markets implement consumer sovereignty is not a judgement about well-being: ‘Whatever consumers want and are willing to pay for, whether their reasons for wanting it are good, bad or non-existent, producers will find it in their interests to supply’ (Sugden Reference Sugden and Olsaretti2006b: 217). That the market has this property has been seen as a reason for valuing it.

My project is to decouple the principle of consumer sovereignty from the assumption that individuals act on coherent preferences, and to use that principle as the starting point for normative economics. I take it that normative economics is primarily concerned with public policy about economic matters. Since my approach is contractarian rather than welfarist, it needs a criterion of individual interest that each individual can endorse for use in public policy-making. That criterion needs to be tractable and transparent, and to be applicable across the range of problems that normative economics deals with. The objective of my project requires a criterion that does not assume the existence of coherent preferences. My proposal has been to use a criterion of opportunity, construed as opportunity for each individual, whenever he faces a problem of choice, to satisfy whatever preferences he then wants to act on. Schubert challenges my claim that most people will be able to endorse this criterion as an adequate representation of their interests for the purposes of normative economics.

To support this claim, I have offered a concept of individual identity that is fundamentally different from the multiple-selves models usually used in economics to describe preference inconsistencies. The key idea is that a person thinks of the preferences on which he will act in the future as his preferences, even if he does not know what they will be and even if they will not be what he would currently like them to be. Thus, all the opportunities he will face in the future are opportunities for him as a continuing agent, whether or not his preferences are consistent across decision problems. I have presented this conception of the continuing self both as a theoretical rationalization of the opportunity criterion and as what Schubert calls a ‘normative ideal’ or ‘role model’ – the responsible agent. Schubert (footnote 8) is right to say that this ideal has something of the spirit of Disraeli's maxim. (Never feel obliged to explain your private choices to others, because your reasons are no one else's business. Never complain to others about the outcomes of those choices, because you are responsible for them.) Perhaps it was a mistake to present this ideal as uncritically as I have done, and to link it with a criterion that was to be recommended to everyone. But I have wanted to show that I was using a conception of the self that could be understood morally as well as pragmatically.

Schubert, like many other readers of my work, baulks at this account of the continuing self. (This is the criticism to which I referred at the beginning of this paper – the criticism that I have often encountered but which Schubert is the first to write down.) He characterizes the attitude of the continuing self to its future preferences as ‘unconditional endorsement’, and says that this makes my account ‘incomplete’ and ‘narrow’ (p. 276); and he criticizes me for treating preferences as ‘passions [that] escape theorizing, they are “just there”’ (p. 289). I think this reading of my account is misleading. Although I have often defended Hume's distinction between passion and reason (e.g. Sugden Reference Sugden2006a), you do not have to be a Humean to identify with your future preferences. At the time when you will act on them, what are now your future preferences need not be unexamined desires. If, when that time comes, you are happy to act on unexamined desires, then of course those desires will determine your preferences. But if at that time you choose to deliberate about what you have most reason to do, your future preferences will be grounded on reasons. And if, between now and then, you choose to cultivate preferences that express the kind of person you want to be, your future preferences will be the product of cultivation. The continuing self's endorsement of its future preferences is no more than a delegation of future decisions to itself in the future.

Schubert is particularly concerned with the possibility that people might want to impose constraints on their own future choices. In any theory that values opportunity, opportunities to reduce one's own future opportunities have a paradoxical status. In presenting the responsible agent as a normative ideal, I have pictured this agent as a person who has no desire to restrict his future opportunities. In my formal analysis of the value of opportunity over time, present opportunities to restrict future opportunities are treated as having zero value (Sugden Reference Sugden2007). Hence the distinction, which Schubert sees as essential to my analysis, between self-command (which is consistent with the ideal of the responsible agent) and self-constraint (which is not). Schubert argues that the opportunity criterion will not command general assent because it does not recognize the legitimacy of self-constraint.

In thinking about this argument, it is important to remember that the opportunity criterion is intended for use in guiding public policy. So the important practical question is not whether private acts of self-constraint are worthy or unworthy – a question that leads Schubert to ask, reasonably enough, about the sharpness of the distinction between self-command and self-constraint. The important question is whether public policy should support restrictions of a person's opportunity in one period on the grounds that, in a previous period, that person chose (or at least wanted) the imposition of those restrictions. This question impinges on my opportunity-based justification of the market, because the market cannot be relied on to satisfy preferences for self-constraint. That this is so is implicit in my slogan that the market allows each person to get what he wants and is willing to pay for, when he wants it and is willing to pay for it. In a competitive market, self-constraint technologies will tend to be supplied to those people who are willing to pay for them, but so too will be the counter-technologies that allow people to escape from constraints they no longer wish to be bound by. Thus, there can be cases in which self-constraint is effective only if some voluntary transactions are prohibited.

Schubert argues that there are pervasive human desires for constraints which support ‘subjective coherence’, reduce choice overload, or guard against powerful but transient emotions that can overwhelm rational deliberation.

I have to say that I have not fully understood Schubert's argument about subjective coherence. For example, he appeals to the evidence of cognitive dissonance. I can agree that the desire to bring one's beliefs into line with one's previous choices is a desire for subjective coherence, but I cannot see the connection with self-constraint. I think Schubert greatly overstates the evidence for choice overload. He cites Schwartz (Reference Schwartz2004), whose book is as much an ideological statement as it is a scientific report. (The flavour of this book is captured by its opening claim that when the number of options becomes too large, ‘choice no longer liberates, but debilitates. It may even be said to tyrannize’, and by its less than persuasive opening example of tyranny – Schwartz's (2004: 1–2) complaint against Gap for offering him a choice between too many different styles of jeans.) Schubert also cites a well-known experiment by Iyengar and Lepper (Reference Iyengar and Lepper2000), which later research has not replicated (see Scheibehenne et al. Reference Scheibehenne, Griefeneder and Todd2010). Taken at face value, this experiment suggests that expansions in the range of choice in retail outlets (represented by an increase from 6 to 24 in the number of jams available to be sampled in a supermarket tasting booth) reduce consumers’ motivation to buy from that range (represented by the frequency with which sampled jams were actually bought). But how can this interpretation of an isolated experimental result be reconciled with the overwhelming success of retail business models, such as those of Wal-Mart and Amazon, which depend on offering enormous ranges of choice?

Still, I accept that there are cases in which some individuals would choose to be subject to externally enforced constraints – for example, to counter transient impulses to consume harmful and addictive products. Although such constraints would not be chosen by the idealized ‘responsible agent’ of my model – one might say that a person who feels the need for them is doubting the responsibility of her future self – I do not want to claim that choosing to be constrained is always irresponsible in the ordinary sense of the word. And if individuals really want to impose constraints on themselves, a contractarian analysis has to take account of that fact, whether the underlying motivation is judged worthy or not. But I maintain that these cases are exceptional: for most people, most of the time, the opportunity criterion is a sufficiently accurate representation of their interests, as they perceive them.

As a way of dealing with these exceptional cases, Schubert proposes using an ‘opportunity to learn’ criterion in place of the simple opportunity criterion. He argues that, when assessing institutional arrangements, we should ask ‘Do they allow individuals to try out and learn new preferences?’ rather than ‘Do they allow individuals to satisfy whatever preferences they happen to have?’ And: ‘the continuing agent is interested, first and foremost, in making sure that all her selves be able to enjoy the opportunity to try out new preferences over time’. I am not sure that preserving opportunities to learn new preferences is typically the primary explanation of preferences for self-constraint. (For example, one might have thought that the prospect of dying from liver or lung disease would be more salient as a reason for trying to escape dependency on alcohol or nicotine.) Schubert's view of self-constraint strikes me as more perfectionist than contractarian. But in any case, it seems highly implausible to claim that opportunity to learn can substitute for (rather than merely supplement) simple opportunity as the criterion of individual interest.

Suppose you are setting out to do your weekly shopping. You could go to a well-stocked supermarket, which sells at reasonable prices all the things you normally buy, along with a wide if unexciting range of other things you could try out if you wanted a change. Or you could go to a niche shop that sells only a small range of esoteric products, none of which you have consumed before. First and foremost, which matters more to you: opportunity to get what you want, or opportunity to learn new preferences? Deliberate learning of new preferences seems to me to be a second-order concern. And, although my opportunity criterion does not explicitly value ‘opportunity to learn’, a regime that satisfies it allows individuals to get any good they want to try out, provided they are willing to pay the costs of supply. (If enough people have the desire to sample esoteric products, the market will sustain esoteric retailers as well as supermarkets.) I think Schubert is losing sight of the basic realities of economic life.

I am often puzzled that my critics give so much attention to cases in which individuals might want to subject themselves to constraints, and so little to the much more common cases that provide the motivation for my project – cases in which individuals with no desire for self-constraint simply fail to reveal consistent preferences. I accept that the opportunity criterion may need to be amended in some way if it is to command general assent when applied to self-constraint problems. But the project of reconciling normative and behavioural economics has a much wider scope than the analysis of self-constraint, and one has to start somewhere. I remain convinced that the opportunity criterion is fundamentally sound.