Subjectivism and Constitutionalism

Abstract

In the light of Giuseppe Eusepi’s career-long interest in the subjectivist element of James Buchanan’s work, my strategy here is to follow Eusepi by outlining and situating the variety of subjectivism discussed by Buchanan and, after some critical discussion, to attempt to track some of the implications of that subjectivism for the idea of constitutionalism, particularly of the contractarian variety, and some of the limitations imposed by subjectivism. To be clear, what follows is not strictly an attempt at the detailed reconstruction of either Buchanan’s or Eusepi’s position, but rather a discussion of subjectivism, constitutionalism, and the linkages between these two themes that is inspired by Buchanan and Eusepi.

1 Introduction

A clear theme running through much of Giuseppe Eusepi’s published work is his engagement with a variety of aspects of the work of James Buchanan. This engagement operates at the historical, methodological, theoretical, and practical levels, with Eusepi acting as commentator, friendly critic, and critical friend.

Somewhat unusually among Buchanan’s commentators, Eusepi’s interest seems to develop in large part from Buchanan’s subjectivism and in particular from (Buchanan 1969 [1999]). This focus on subjectivism as a key component of Buchanan’s contribution persisted right through to Eusepi’s final publication in 2020 (Eusepi, 1987, 1991, 2006a, b, 2008, 2020; Eusepi and Cepparulo 2007). While Eusepi was clearly interested in the historical connection between Buchanan’s subjectivism and the Italian public finance tradition, his interest went far beyond this to include the methodological and theoretical implications of subjectivism and, in particular, its implications for constitutionalism.¹

It should come as no surprise that even if we limit discussion to the field of political economy,² subjectivism comes in many varieties and degrees, both in the history of thought and in more contemporary usage. Buchanan identifies his subjectivism with that of the mid-twentieth century LSE group of economists including Coase and Thirlby which was linked to Austrian economics via Hayek,³ while Eusepi emphasises the link with the Italian tradition and particularly Ferrara. ⁴ (Eusepi, 1991). I will not focus on the historical antecedents of subjectivism of the Buchanan/Eusepi variety here, but rather on the content and implications of the stance on subjectivism discussed by Buchanan and Eusepi.

I can claim some linkage to the relevant subjectivist tradition via Jack Wiseman. As a student of Jack’s in the 1970s it was impossible to avoid forceful subjectivist argument, but also difficult to avoid the conclusion that a thoroughgoing subjectivism places severe limitation on what an economist can say (beyond criticising objectivism) in both positive and normative domains. My strategy here is to follow Eusepi by outlining and situating the variety of subjectivism discussed by Buchanan and, after some critical discussion, attempting to track some of the implications of that subjectivism for the idea of constitutionalism, particularly of the contractarian variety, and some of the limitations imposed by subjectivism. To be clear, what follows is not strictly an attempt at the detailed reconstruction of either Buchanan’s or Eusepi’s position, but rather a discussion of subjectivism, constitutionalism, and the linkages between these two themes that is inspired by Buchanan and Eusepi.

A basic question faced by any subjectivist is, simply, subjectivism about what? A variety of answers relevant to political economy present themselves: one might be a subjectivist about preferences, costs, probabilities. interests, well-being, welfare, or morality. Other relevant possibilities exist, and these various possible subjectivisms may be combined in a variety of ways. Most mainstream economists, for example, would self-identify as subjectivists about preferences, in the sense that they take an individual’s tastes to be idiosyncratic and personal, but revert to a form of objectivism about most, if not all, other items on the list. Buchanan’s subjectivism goes much further than this. He self-identifies as a radical subjectivist, just how radical is part of what is in question in what follows. Whatever the details of any particular position, it should be recognised from the outset that subjectivism about one item on the list, does not entail subjectivism about all items, so that the case for (or against) subjectivism must be investigated at a variety of levels.

To take one step back; there is a familiar distinction between two models of the relationship between desire and value. Things might be valuable because they are desired, or they might be desired because they are valuable. On the first view desire, or preference, or some similar attitudinal placeholder, drives value. On the second view, desire (preference, or some relevant attitudinal placeholder) follows from the recognition of value in the independently valuable thing. Furthermore, the nature of the relationship between desire and value may vary dependent on the type of value under discussion. We might, for example, offer one account of the relationship in the case of prudential value, and a different account of the relationship in the case of moral value. Hume, for example advocates a desire-based account of value for all types of value, while many philosophers would argue for a recognition of independent value approach for at least moral values.⁵

If values (of a particular type) are set independent of desire (or other similar attitudes) those values may be taken to be objective in a fairly straightforward sense. Of course, different individuals may see the value of some particular thing differently, so that there will be some variation in the value recognised by different individuals, but this is not basic.

If values (of a particular type) are dependent on desire (or other similar attitudes) we might take those values to be subjective. But this sense of ‘subjective’ falls rather short of the idea pursued by Buchanan and others. Indeed, the mainstream approach to economics might qualify as subjectivist on this rather broad-brush approach to definition, at least for prudential values.

To zoom in on the distinction that Buchanan and Eusepi have in mind, we need to ask a further question within the broad desire-based account of value. To what extent, if any, are an individual’s desire-based values (of a particular type) capable of being seen as objective data in the social world? As we will see in what follows, it is the answer to this question, for each type of value and what follows from those answers, that distinguishes between mainstream economics and Buchanan’s more radically subjectivist approach.

The remainder of this paper is divided into four parts. The first two outline the radical subjectivists view in the positive and normative domains, respectively, and offer some critical discussion. The third then considers the links between subjectivism and constitutionalism. A final section offers some concluding comments.

2 Subjectivism in the Positive Domain

We begin with choice. Simple individual choice of the type considered in standard consumer theory. This focuses attention on prudential value. Subjectivists of the Buchanan variety insist that while choice is real and objective, in the sense that choices are conceptually observable and, hence, part of the objective data set, any account of the internal, psychological process leading to choice must be hypothetical. The subjectivist points to behaviour as the empirically observable, objective, fact while any psychological account of choice is merely a theoretical model. More specifically, the subjectivist points to the economist’s textbook model of well-defined individual preferences operating in an environment of objective costs as being a misleading fiction. This on three major grounds, relating to preferences, cost, and uncertainty. In this context, subjectivists like Buchanan are subjectivists about each of preferences, costs, and probabilities and so are subjectivists about prudential choice and value.

On preferences, the subjectivist perspective is that while individual choice may be assumed to be purposive, this should not be taken to imply the existence of anything like a complete and stable preference function.⁶ The idea of purposive individual choice is a weakening of the mainstream assumption of rationality. Purposive choice implies that the individual considers the perceived net benefits associated with the perceived alternatives and chooses the ‘best’ option (that which best achieves perceived purposes). The subjectivist emphasises the repeated use of ‘perceived’ in the last sentence, and also emphasises that benefits themselves are best understood in subjective terms – the achievement of personal goals whatever they might be. The subjectivist also argues that choices are time and circumstance specific so that while we may assume that in each choice situation the individual chooses the ‘best’ option, there can be no guarantee that the range of alternative options actually considered is in any sense complete or objectively defined, or that the relevant ‘purposes’ are consistent over choice situations or limited to any notion of self-interest. Opportunity sets are imagined rather than objectively fixed and relevant purposes may be transient and may also respond to concerns far removed from even the most generous interpretation of individual self-interest, including, for example, altruism, ideological preferences, and beliefs about what morality may require, and even to the idea of attempting to change the individual’s own current tastes and attitudes.⁷

On cost, the subjectivist emphasises that the relevant notion of cost is opportunity cost and points out that the appropriate measure of opportunity cost in the case of the choice among mutually exclusive alternatives is the net benefit foregone from the most attractive rejected option.⁸ Two implications follow. First, the relevant metric of cost is the extent to which the best rejected option achieves the individual’s underlying (perceived) purposes, whatever they might be. So that opportunity costs are subjective and knowable only by the individual concerned. Second, since costs are best understood as net benefits foregone, the subjectivist account reduces to the claim that choice maximally serves perceived purposes, given perceived alternatives. Despite the emphasis on subjective costs found throughout Buchanan’s subjectivist writing, the claim that costs must be understood in terms of net benefits foregone puts the real weight of the subjectivist enterprise on the subjective nature of benefits (including unrealised benefits foregone) and, therefore on the variety of individual purposes.

On uncertainty, the subjectivist rejects modelling uncertainty in terms of well-defined and widely shared probability distributions over an objective list of possible outcomes and insists that the future is unknowable so that individuals take an idiosyncratic and essentially imaginative view of the likely consequences of any choice they may make. Individuals gather information and learn, but they never approximate the state assumed in mainstream choice theory. In part this view builds on the account of probabilities as subjective degrees of belief, but it also denies that the degrees of belief of a purposeful individual obey the rules of the standard probability calculus.⁹ In short, individuals’ responses to uncertainty reflect idiosyncratic and inchoate views of probability. Choices do not select a particular future from a well-ordered list of alternatives, rather they create the future.¹⁰

Of course, this subjectivist account of choice is itself a mere model. One might object (on fully subjectivist grounds) that all talk of ‘purposeful choice’, ‘perceived alternatives’, and ‘imagined future consequences’ is just as hypothetical as talk of well-defined preferences. If we are to give up on all psychological models of choice, we are driven to a pure behaviourism. Even so, the subjectivist critique of the mainstream model of choice seems powerful. If one wants to understand actual choices, made by real individuals in real time, focusing on the actual purposes and perceptions of the individual at the point of choice seems an appropriate starting point.

But understanding choice is not the only, or even the most significant, target of the mainstream economic analysis of choice. Choice theory, as standardly deployed by (objectivist) economists is also a crucial building block of predictive economic theory and normative economic theory, and it is in these roles, and the relationship between them, that subjectivists claim to offer a real alternative (Buchanan 1982 [2001]).

A specific psychological model of choice allows the analyst the possibility of predicting choice behaviour to the extent that the model identifies observable variables that contribute to the explanation of choices. Clearly, then, an objectivist model offers more scope for prediction than a subjectivist model whose principal feature is that the key determinants of choice behaviour are unobservable in principle.¹¹ And this may be so even if the objectivist model offers an inferior understanding of choice. When considered as positive, predictive devices, models may be judged by their usefulness regardless of their psychological accuracy. This is the standard point that lies at the heart of what has been termed the ‘so what’ response to the subjectivist critique of mainstream choice theory.¹² Buchanan accepts that there is a place for mainstream economics as a predictive science (even if based on a potentially misleading model). Indeed, he repeatedly attempts to identify two distinct domains: one in which individuals react to changes in their (observable) environment, another in which individuals create their chosen actions.¹³

In the reactive domain, which might include many of the routine choices of standard consumer theory, the mainstream model may be an acceptable (if imperfect) approximation and the available objective data indicted by the mainstream model may be sufficient to provide the basis for useful prediction. But this predictive usefulness should not be taken as establishing anything like the credibility of the underlying psychological model. This point will be important when we consider the normative domain below.

In the creative domain, which might include choices that have a more entrepreneurial flavour, or where the options are not clear-cut, the mainstream model will not be acceptable even as an approximation, so that available objective data will not add significant predictive power.¹⁴ In this arena, the subjective model applies but it does not provide the basis for a predictive science.

While the recognition of reactive and creative aspects of choice seems plausible, the categorical distinction between reactive and creative domains seems oversharp. Much more plausible, surely, would be the suggestion of a continuum of cases ranging from the ideal type of purely reactive choice at one logical extreme (‘men as rats’ (Buchanan 1982 [2001])), to the ideal type of the purely creative choice at the other logical extreme.

Whether we take the categorical or the continuum approach to the reactive/creative distinction, a significant issue is the range of choice situations which fall in the (predominantly) reactive domain. If this range is extensive, then the practical implications of subjectivism in the positive domain may be rather slight. The mainstream, objective model will be a useful predictive tool, and will fail in only a relatively small number of (predominantly) creative cases.

At this point we might recall that we started with ordinary choice situations of the type considered in standard consumer theory. But much of the debate on the subjectivity of costs relates, implicitly or explicitly, to the discussion of the meaning of ‘costs’ on the supply side: in the theory of the firm, rather than in consumer theory. Does the shift from the logic of choice in the case of individuals as consumers and partners in exchange to the logic of choice in the case of firms (and other relatively complex organisations) as producers and sellers have any impact of the subjectivist argument or the likely balance between reactive and creative domains?

Consider a range of choices that we might term professional choices. A professional choice is one made by an individual (so there is no deviation from methodological individualism) but made in the context of that individual’s role in some rule-governed organisation, (a firm, corporation, or institution such as a government department). The point here is to recognise that while the individual might still be understood to be choosing subjectively, she is likely to be acting under considerable constraints imposed by rules, protocols and procedures that characterise the relevant organisation. Furthermore, the relevant rules will, of necessity, be couched in objective terms: setting targets, identifying specific practices, etc. To the extent that the individual has private motivations in relation to such professional choices, they are likely to be dominated by career concerns that strongly recommend playing by the organisational rules. So that professional choices might seem to be largely explicable in objective terms – to be broadly reactive and so be usefully predictable.¹⁵ On the other hand, it is clear that at least some decisions and choices made within organisation will be (predominantly) creative - particularly those that might be termed entrepreneurial, concerning new directions, or constitutional, setting or reforming the internal rules of the organisation. This last point on the creative nature of constitutional choices will be relevant later.

One might expect that the subjectivist would, at this point in the debate, offer a detailed critique of the standard model of costs in the theory of the firm – but this is not generally the route taken by Buchanan or other subjective cost theorists.¹⁶ Instead Buchanan and others typically shift their attention to the critique of the mainstream model as a static equilibrium analysis whilst choice should be seen primarily in a dynamic disequilibrium context.¹⁷ The point here being that, even if objective money outlays are a good proxy for the underlying subjective costs, this will only be true in equilibrium, so that the dynamic, disequilibrium nature of most choice situations renders the objective model less useful.

This may be true. But the addition of the criticism from disequilibrium hardly changes the picture in the positive domain. It remains the case that the focus on objective costs (essentially, money outlays) is useful to the extent that it provides a basis for improved prediction. The existence and extent of that usefulness is an empirical matter which will vary from case to case and time to time and is in no way altered by adding to the list of theoretical critiques of the mainstream model. At the same time, it is not at all clear that the critique from disequilibrium dynamics is necessarily subjectivist in nature.

The shift from rational choice to purposeful choice and the emphasis on idiosyncratic subjective benefits rather than a preference function suggests a move away from Homo Economicus towards Homo Sapiens, away from homogenous agents and towards heterogeneous agents (Eusepi and Cepparulo 2007). There is tension within Buchanan’s work on this issue. The position taken when writing specifically about subjectivism, or when discussing the ‘constitutional attitude’ required within his vision of constitutional contractarianism, does not always carry over in other areas. And this tension is not fully resolved by reference to the distinction between reactive and creative choice.

To the extent that the mainstream model provides a reasonably reliable basis for prediction across a relatively wide range of choices the positive impact of the subjectivist critique is somewhat limited. Subjectivism provides a clear basis for a critique of the logic of the mainstream model but offers rather little in terms of an alternative basis for positive analysis and prediction.

3 Subjectivism in the Normative Domain

In moving to the normative domain, subjectivists, and Buchanan in particular, would claim that the critical value of the subjectivist perspective is enhanced and that there are real implications for the scope and nature of economic policy. In short, it is in the normative domain where subjectivists insist that the mainstream objectivist model goes badly wrong.

The essential point here is that while the claim that mainstream theory should be read on an ‘as if’ basis, and seen as a useful predictive model, may shield the mainstream model from the full force of the subjectivist critique in the positive domain, that ‘as if’ approach is not available if the aim is to provide the basis for normative economics. The normative claims of the mainstream model are, and must be, based on the (at least approximate) accuracy of the claim that preference functions are relatively complete and stable and that these preference functions represent the individuals’ interests, which are in turn the basic input into all-things-considered value.

The basic strategy of the mainstream normative model is to infer preferences from behaviour, and then to infer interests or well-being from preferences. This amounts to a desire-satisfaction theory of well-being or interests. These interests are then taken to be objective in the sense that they are observable and stable, so that they can provide a reliable basis of public policy. The mainstream normative approach is broadly utilitarian in that it takes individual interests, well-being, or utility as the only primary input into all-things-considered value, although it does not necessarily involve the classical utilitarian commitment to sum-ranking whereby the simple aggregation of individual utility yield overall social welfare. The mainstream economic approach may depart from sum-ranking to allow the distribution of well-being across individuals to matter. This strategy allows public policy to be designed and evaluated by reference to a criterion built on the (objectively observed) interests of individuals.

Of course, there are a number of issues to be confronted within this strategy, not least the aggregation issues raised. But the most basic subjectivist criticism relates to the strategy as a whole rather than to the steps within it.

If the very idea of an individual’s complete and stable preference function is a mistaken psychological fiction, as subjectivists believe, there can be no prospect of building a normative analysis on such foundations. This, then, is subjectivism about interests, or well-being, or any other construct aimed at capturing the appropriate concept of normativity at the individual level. This is the starting point of Buchanan’s normative individualism (Buchanan 1991 [1999]).¹⁸

The most basic idea within Buchanan’s normative individualism follows on directly from the discussion of subjectivism in the positive domain, that the notion of ‘that which is best for the individual’ does not exist independently of actual choices made by the relevant individual.

From a subjectivist perspective, a “utility function” as such, does not exist which, even conceptually, could be observed and recognized independently of an individual’s choice behavior. All there is are individual choices, and it is about these choices, not about some alleged relationship to some utility function, that we develop theories. (Buchanan 1991 [1999] p286)

Buchanan is at pains to point out that his position is sharply distinct from ‘epistemic individualism’, the position that grants each individual an epistemic advantage in knowing what is best for them, but which nevertheless allows the objective observation of ‘that which is best’ for any individual, at least to a reasonable approximation.

In essence, Buchanan commits to the extension of the subjectivism about preferences and prudential value to the subjectivism of well-being and the foundations of moral value. This is not to say that Buchanan conflates prudential and moral values. He is perfectly open to the distinction, so that any individual may, in their own mind, identify certain values as moral rather than prudential, and may recognize the possibility of some trade-off between moral and prudential values in at least some circumstances. But Buchanan holds that both categories of value (and any other categories, and the nature of any trade-off among them) are subjective in the sense that they are internal to the individual and that they are only realised in actual choices. This normative subjectivism is opposed to any idea of objective or transcendental notions of well-being or morality.

Normative subjectivism of this sort does not deny that there may be many influences – including objective influences – on the idea of well-being and morality held by any individual. Social and environmental forces will certainly have some impact, and these and other forces may lead to moral views that are relatively widely shared in any community of individuals. But widely shared values are not objective values. And not all values are widely shared. Diversity abounds. The idea that, as a matter of fact, individuals differ in their conceptions of morality and on the demands morality may make on them, as well as in their conception of their own well-being, is clearly descriptively plausible, but normative subjectivists go beyond this recognition of moral diversity in practice to argue that even in principle there is no objective standard of morality or well-being that stands outside of any individual, so that moral diversity is a deep feature of the normative landscape.

The standard argument against such a normative subjectivism begins by pointing out that this is essentially a current desire-based account and claims that current desires are too flimsy a basis on which to build normativity. The initial step is to point out that actual, current, desires may often be seen to be fleeting, mistaken or poorly informed - whims or prejudices – and that this suggests that rather than actual, current desires we might consider somewhat idealised desires (considered desires, rationally endorsed desires, fully informed desires, or some such formulation) as a more promising starting point for normativity. Even after such a move to idealised desires, there are further criticisms.¹⁹

But for subjectivists like Buchanan, even the move to idealised or ‘laundered’ desires is a step too far since, on their account, while current desires may be part of the subjectivists psychological account of purposeful choice, it is actual choices, rather than any notion of preferences or desires, whether actual or laundered, that are basic, and any shift to idealised desires breaks the connection with actual choices that is fundamental to the subjectivist position. For the radical subjectivist, choice is the ultimate standard of value, and this is so across all types of value.²⁰

The radical subjectivist critique of mainstream normative economics views all policy prescription based on claims that the relevant policy serves the interests of most or all individuals as lacking foundation unless that policy is actually chosen by the individuals concerned. Even if it seems clear that some policy goal – some improvement in health, for example – would likely be validated by (almost) all individuals as being in their interest it cannot be claimed that the policy is validated since individuals may differ in how they view alternative policies aimed at the same goal and how they view the opportunity costs of any such policy. Policy without the explicit choice or consent of those affected is, at best, paternalism and, at worst, dictatorship.

If the subjectivist critique of the mainstream economic approach to policy formulation and evaluation is clear enough in outline, the subjectivist alternative is far less obvious. If individual choices, rather than imputed interests or values, are to be the foundation for policy making it might seem that a form of direct democracy with unanimous (or near unanimous) support as a precondition for the implementation of any policy is the only way forward, however impractical this might be. But this is surely too simple a thought. Under Buchanan’s own moral view, it seems that any group would have reason to pursue political ends or policies that benefit its own members at a cost to others, if only they have the power to do so. Nothing seems to require anything like unanimity in the wider society. This is where Buchanan’s subjectivism and Buchanan’s constitutionalism are intended to come together.²¹

Before moving on to consider constitutionalism, it is appropriate to note that the relatively recent growth in behavioural economics represents an interesting sidelight on the subjectivist/objectivist debate in both the positive and normative domains. In the positive domain, it is clear that a main thrust of behavioural economics has been the criticism of the mainstream model of individual, self-interested rational choice and, in particular, the idea of homo economicus as captured in a well-defined and stable preference function. However, it is equally clear that most behavioural economics attempts to provide a richer psychological model of choice behaviour rather than dispensing with such models in favour of a focus on actual choice behaviour. Such richer models may include both subjective and objective elements, but they are normally taken to sufficiently objective to imply a degree of predictability of behaviour, indeed a central claim is that these more behaviourally sophisticated models outperform the mainstream model when seen as predictive science. So, while the critique of the mainstream model from behavioural economics has something in common with the subjectivist critique, there is clear evidence that many behavioural economists see their approach in objectivist terms as a potentially predictive science.²²

Following from this position in the positive domain, many behavioural economists see their role in the normative domain as providing a guide to means of ‘improving’ on the choices that individuals would make if left to their own devices. A prime example being the advocacy of the benevolent paternalism involved in idea of ‘nudging’ behaviour.²³ An exception is provided by Sugden (Sugden, 2018, 2020), who draws inspiration from Buchanan (and from Hume), and certainly operates in an area which places greater weight on actual individual choices.²⁴

4 Subjectivism and Constitutionalism

In a general sense, constitutionalism revolves around the idea that the powers of government can and should be limited by a higher law, and that the proximate legitimacy of government depends on its respecting that higher law. Of course, the overall legitimacy of government depends upon the further specification of the legitimacy of the constitution. For some, a legitimate constitution might be derived directly from the consideration of objective values, but for Buchanan a legitimate constitution must itself be based in choice.

There are different possible interpretations of what it might mean for a constitution to be ‘based in choice’. One possible view is that the constitution should institutionalise and validate individual choice to the greatest extent possible. All forms of coercion and compulsion, by the government or by other agents, should be minimised. For example, individuals should, where possible, be able to opt out of schemes such as social insurance schemes on terms that are fair, so that participation in such schemes can be seen as the result of real choice. Of course, minimal compulsion may be a long way from zero compulsion but call such a minimally coercive constitution a libertarian constitution of choice.

An alternative view is that the constitution itself should be chosen by near unanimity. Under such a constitution, institutions involving compulsion may be legitimate, provided that they have the clear constitutional support of the relevant population. Call such a constitution a contractarian constitution of choice.

Whether a contractarian constitution of choice yields a libertarian constitution of choice must be an open question, and particularly so for a subjectivist. Indeed, for a truly radical subjectivist, who emphasizes the fact of actual choice rather than hypothetical choice involving the use of such theoretical constructs such as preferences, interests, or values in both the understanding of behaviour and the normative validation of outcomes, it would seem difficult to construct an argument as to exactly what constitutional arrangements any set of individuals might actually agree.

Certainly, the constitutional political economy (CPE) approach based in Buchanan’s work, favours the contractarian line of argument. CPE sees a fundamental logical distinction between in-period politics, where existing political rules dictate political outcomes, and constitutional politics where the political rules themselves emerge, and argues that constitutional legitimacy, and hence political legitimacy, depends on constitutional rules themselves emerging from a process of choice.²⁵

While Buchanan is generally consistent in insisting that ongoing active choice is the relevant criteria by which to judge a constitution, he clearly recognizes that, as a matter of fact, constitutional rules are rarely chosen by the populations that live within them. At best, those constitutional rules may have been chosen by some members of an earlier generation.

The fundamental questions are then, why should the current generation be bound by a constitution they had no part in choosing, and does their allegiance depend on the content of the constitution in question? We might identify two broad approaches to addressing this pair of questions. On the one hand, one might argue that there are good reasons to suppose that the current generation would choose a particular constitution, and that this provides the relevant normative support for that constitution. This is a move in the direction of hypothetical choice. On the other hand, one might argue that ‘acceptance’ of an extant constitution by members of each overlapping generation is normatively equivalent to the more active choice of that constitution by successive generations. This is a move in the direction of tacit choice. Neither of these moves seems, at first glance, entirely consistent with Buchanan’s commitments to both subjectivism and contractarianism, and both are subject to Hume’s basic critique.

Buchanan’s tactic for avoiding the difficulty of giving his constitutionalism substantive form while retaining both subjectivist and contractarian credentials involves the argument that in contemplating constitutional matters, individuals tend to step away from their personal, in-period interests (whatever they may be) and focus instead on the structure of their more general interests in relation to feasible political procedures. The argument is that behind the relevant constitutional veil of uncertainty, (almost) all individuals will view the properties of alternative political procedures (whether voting rules, rules on the separation of powers or whatever) through a common lens. This might be viewed as a move that has the effect of introducing a substantive norm of interpersonal respect.

That lens focuses on the trade-off between the benefits of collective action and the cost of coercion. For example, in the case of the choice of a voting rule to be applied across a range of particular issues and over time, Buchanan’s contractarian constitutionalism points to the trade-off facing any individual between having the decision go with their in-period interests (whatever they may be) and having the decision going against them. Sometimes they will expect to be on the winning side, sometimes on the losing, and the choice of a voting rule balances the advantages of winning against the costs of losing.

In essence, Buchanan’s argument is an attempt to retain a commitment to subjectivism at the in-period level while arguing that the nature of the constitutional level of choice renders individual’s interests in constitutional matters both more similar to each other and more objectively accessible, so that the analyst can make at least some reliable predictions as to the constitutional choices that are likely to be made by viewing a representative constitutional agent. There is the further question of what such chosen constitutional rules would look like and, in particular, to what extent they would allow in-period decisions by less than unanimous choice. But that question can be left on one side for our present purposes.

This argument for the harmonization and quasi-objectification of ‘interests’ at the constitutional level involves a move in the direction of hypothetical choice. A move that is justified, on Buchanan’s argument, by consideration of the distinctive nature of the constitutional setting and the veil of constitutional uncertainty. This is a major move for Buchanan. He is not just claiming that ‘interests’ at the constitutional level are sufficiently objective to allow of a predictive science of constitutional choice, whereby the analyst can expect to make reasonably reliable predictions of the choices that would be made by populations of individuals via the analysis of a representative individual at the constitutional level. He is also claiming that the shift to the constitutional level and the quasi-objective constitutional interests of individuals provides a sufficiently strong normative foundation for constitutionalism.

There must be at least some doubt about this.²⁶ It amounts to claiming that there are objective reasons that are capable of being known (at least approximately) in the absence of any actual choice-making behaviour, that these reasons apply to (almost) all individuals, and that these reasons operate specifically at the constitutional level. While the argument relating to the constitutional veil of ignorance attempts to provide a basis for these claims, they are not easily supported. Remember our earlier discussion of Buchanan on the distinction between reactive and creative arenas of choice, where the claim was that it is particularly in the creative arena that the subjectivist critique of objective reasons was at its strongest. It might be argued that the constitutional level of political choice is more creative than reactive, and more creative than the in-period level. More creative since it allows individuals to imagine very different social structures to those that they may have experienced, while in-period choices might be seen as more reactive since they amount to operating within given guidelines. It seems entirely plausible that individuals differ in their constitutional and ideological preferences and attitudes at least as much as they differ in their everyday preferences and attitudes, so that the idea of a representative individual at the constitutional level is at least as difficult to sustain as the idea of a representative individual at the in-period level, and if this is the case, there seems to be no stronger claim for (near) unanimous agreement on constitutional matters than there is in the case of in-period politics.

Put in another way, the difficulty for Buchanan’s argument on the quasi-objective nature of the constitutional ‘interests’ of the representative individual might be due to a rather narrow reading of ‘interests’. If we read ‘interests’ at the in-period level in terms of the narrow self-interest usually associated with textbook versions of mainstream economics, then it seems at least somewhat plausible, when moving to the constitutional level, to assume that constitutional interests must be similarly instrumental and limited to viewing constitutional arrangements in terms of their likely impact on each individual’s ability to pursue her own narrow self-interest. And this seems to be the line that Buchanan takes. But a much broader view of interests, at both the in-period and constitutional levels, seems much more in keeping with the subjectivist rejection of the idea of objective interests and the shift to more generalized purposeful choice. If individuals may have purposes that go beyond narrow self-interest to include, inter alia, different moral beliefs and preferences over political systems, it seems unlikely that the shift from the in-period context to the constitutional context will produce either the harmonization or the quasi-objectification of individual ‘interests’ required for Buchanan’s argument.

Given these doubts about the argument from the veil of uncertainty, there must certainly be at least some remaining tension between the commitment to radical subjectivism and the reliance on essentially hypothetical choice at the constitutional level. However, some glimmer of subjectivist logic may remain. For this glimmer to be amplified, we would need the hypothetical nature of the choice to be specified in as concrete form as possible. I will return to this point below.

The alternative to the move in the direction of hypothetical choice is the move in the direction of tacit choice, effectively substituting acceptance for actual choice. This move might initially seem more in line with the broadly subjectivist spirit since it points to a behaviour (acceptance) as the relevant criterion of constitutional legitimacy rather than some claimed (objective or quasi-objective) understanding of the reasons behind hypothetical constitutional choice. While acceptance of X is not equivalent to the choice of X, it does represent a choice of sorts.

So, is there a form of tacit choice or ‘acceptance’ of a constitution that is robust enough to carry the relevant normative load? The problem here is that mere acceptance seems both too passive and too open to manipulation to carry the normative load required. Most obviously, an oppressive constitution backed by strong enforcement might well generate what passes for acceptance in the population in the sense that the costs of political or constitutional protest may be so high as to effectively block opposition. In such a case it would seem perverse to take the apparent (passive) acceptance of the constitutional order as an argument in favour of its legitimacy. Similarly, even if a constitution includes the specification of processes by which it might itself be amended or reformed, the absence of actual cases of constitutional reform cannot be taken as demonstrating acceptance of the constitution of the required type.

Taking these arguments together, there is at least some considerable tension between the radical subjectivism that insists that actual choice is the only reliable normative benchmark and any plausible form of contractarianism that depends on either hypothetical or tacit contracts, given the obvious problems with actual contract in the constitutional setting. But note that while this line of argument limits the possibility of a constitutionalism that is both radically subjectivist and contractarian, it does not undermine constitutionalism itself. Routes to constitutionalism may remain open if, for example, one relaxes either the subjectivist or the contractarian requirement. Given our present focus I will concentrate on the possibility of a subjectivist constitutionalism by relaxing the contractarian requirement. This takes us into essentially Humean territory, and its attraction is explicitly acknowledged by Buchanan in his appendix to (Buchanan and Tullock 1962 (1999)) which is worth quoting at length:

In this respect our immediate precursor is Hume, who quite successfully was able to ground political obligation, neither on moral principle nor on contract, but on self-interest. Hume did this by resorting to the idea that the self-interest of each individual in the community dictates the observance of conventional rules of conduct. These rules, which may or may not have been formalised in contract, are necessary for the orderly conduct of social affairs. This argument, which does not base political obligation on contractual obligation, allows the primary difficulty of the contract theorist to be neatly surmounted. Not only is it to the initial interest of parties to agree on conventional rules if such rules do not exist, but it is also to the continuing interest of individuals to abide by the conventional rules in existence. (pp314-315).

The shift from strictly contractualist constitutionalism to a more convention-based constitutionalism requires further discussion. First, we might dispute the claim made in the final clause of the above quote. It is certainly the case that it is in the constitutional interests of all that certain rules exist. All have reason to desire that these rules exist and are complied with. But this falls short of demonstrating that such rules can be legitimately enforced.

Second, there is the question of how this move relates to the argument, outlined above, concerning the quasi-objective nature of individual ‘interests’ at the constitutional level. Essentially, the more conventionalist approach tends to separate the positive and normative elements of the account of constitutions from each other. Rather than the normative support for a constitution being baked into the contractual process which generates that constitution, the conventionalist sees conventions arising initially purely as regularities in behaviour which may then gain normative traction by being appropriately aligned with normative attitudes that exist independently in individuals.²⁷ This separation of the account of the emergence of rules and their normative force seems entirely consistent with the subjectivist approach that places few restrictions on the form or content of individual normative attitudes.

Third, it is more generally the case that the Humean approach is broadly consistent with a form of subjectivism. Although Buchanan, in the quote just cited, indicates that ‘interests’ play a part in both the initial emergence of conventional rules and their continued acceptance, these ‘interests’ are not the assumed objectified interests or preferences of standard normative economics, but are much closer to the ‘purposes’ that underlie Buchanan’s subjectivist account of choice. Purposes that are fully defined only in actual moments of choice. Conventions emerge out of, and are sustained by, behaviour rather than hypothetical choice.

Fourth, it is important to note that the more Humean, conventionalist account of constitutional rules is clearly based on individual heterogeneity. This in contrast to Buchanan’s contractual approach in The Calculus of Consent and elsewhere which, as I have suggested, is often formulated in terms of the constitutional choices made by a representative individual. Eusepi recognises this point, particularly in (Eusepi and Cepparulo 2007) where the discussion highlights Buchanan’s use of the assumption of homogeneous individuals and implicitly criticises Buchanan for adopting a Homo Economicus model when his subjectivism should propel him towards a more flexible and heterogeneous model of Homo Sapiens.

Fifth, the conventionalist reading of constitutionalism still suffers from a problem surrounding the acceptance of ongoing conventional rules. The conventionalist account perhaps offers a more active account of ‘acceptance’ than the contractual reading, but the conservative bias in favour of inherited rules is still at least somewhat evident. Indeed, it seems an inescapable part of constitutionalism that if a constitution is to be understood as a system of rules that are confidently expected to extend over generations, there must be some element of conservative continuity.

The strategy of weakening the contractarian focus of constitutionalism while emphasising the heterogeneity of individual (subjective) perspectives, can be seen as one theme in some recent developments of the public reason approach to political philosophy in the post-Rawlsian tradition. This literature argues that in a world of deep moral and political diversity where individuals differ in their conceptions of the good, normative political theory should proceed on the basis of identifying those rules, arrangements, and policies that can be justified to individuals regardless of their particular values.²⁸ This approach allows the analyst to create a variety of types of individual differentiated by their moral and political values and interests and attempt to identify the constitutional rules that might be justified to each type of individual. Gaus and Vallier argue that the broadly liberal set of rules and institutions are most likely to be acceptable to such a variety of individuals.

Note that in some ways this strategy is the opposite of Buchanan’s strategy in relation to the constitutional veil of ignorance. Rather that arguing that the constitutional setting renders the assumption of a representative individual with quasi-objective interests appropriate, the strategy adopted by Gaus and Vallier argues that recognizing the wide diversity of individual values at the constitutional level allows one to create a sort of test bed on which one can test out the acceptability of any particular set of rules or institutional arrangements, seeking to identify the set of rules that would not be ‘vetoed’ by any particular type of individual. Of course, such a test bed approach amounts to a departure from any reliance on the actual choice of constitutional arrangements – so that we are once again concerned with the analysis of a form of hypothetical choice. But the claimed advantage of the approach from diversity is that, at least in principle, the acknowledgment of a variety of individual positions moves the hypothetical nature of the choice toward the concrete, in the sense that the larger the range of types of hypothetical individuals one considers, the more confident one might be that the conclusions apply to any set of actual individuals.

5 Concluding Comments

Radical subjectivism is not a popular position in political economy. There are good reasons for this. In the positive domain, while subjectivism about preferences, costs, probabilities, and prudential value provides a clear understanding of individual choice and an equally clear basis for a critique of many elements of mainstream economic theory, it provides a very flimsy base for an alternative approach to economics seen as a predictive science. In the normative domain, subjectivism about interests, well-being, and the location of the source of normativity similarly raises more issues than it can address and is open to a range of criticisms that begin from the claim that present desires are insufficiently robust to provide the foundations for normative theory. The radical subjectivist approach may attempt to inhabit the space between predictive science and moral philosophy, (Buchanan 1982 [2001]) but it provides little alternative to the former and a very particular morally relativistic approach to the latter.

Against a background in which mainstream economics offered an account of political economy that was generally objectivist, utilitarian and focused on the piecemeal analysis of policy options, Buchanan’s lifelong attempt to build a subjectivist, contractarian and constitutional account of political economy is impressively distinctive. But, as we have seen, while Buchanan’s discussions of subjectivism are important in developing his critique of the mainstream position, they also establish the limits of truly radical subjectivism and exhibit some remaining tensions with his own contractarian constitutionalism. In the final analysis, I would argue that it is Buchanan’s commitment to constitutionalism that takes priority in his overall political scheme. There is no reason that constitutionalism should be based in either subjectivism or contractarianism, so that the commitments to both subjectivism and contractarianism can be weakened somewhat in an attempt to find a normative account of constitutionalism that retains at least some underlying subjectivist and contractarian features. Radical subjectivism may be a step too far, but Buchanan remained convinced of the subjectivist critique of mainstream economics and the need for an approach that was at least more subjectivist.

Buchanan may not succeed in demonstrating the feasibility of a fully subjectivist, contractarian constitutionalism that is broadly supportive of a liberal political order, but his repeated and detailed explorations of the wide variety of issues at play in the interaction between subjectivism, contractarianism, constitutionalism, and a liberal political order provide a rich source of arguments that serve as an excellent guide to the terrain.

I began by noting that much of Giuseppe Eusepi’s career was spent in engaging with James Buchanan’s body of work. I too have spent a significant part of my career engaged with Buchanan’s work. Giuseppe and I focused on rather different aspects of Buchanan’s overall project. Giuseppe concentrating his attention on the subjectivist tradition and concerns central to the theory of public finance including the possibility of a constitutional welfare state (Eusepi, 2006a), and the constitutional status of public debt (Eusepi, 2020); while my own interests moved more towards the connections to more general political philosophy. It is often claimed that the major philosophical influence on Buchanan was Hobbes but, as is made clear in (Brennan and Eusepi 2013, 2016), while there is an underlying Hobbesian element in Buchanan’s approach, and especially in (Buchanan, 1975), there is also much in Buchanan that is distinctly non-Hobbesian, including his approach to contractarianism. I would agree with Kliemt (Kliemt, 2011) that Buchanan’s project might be better seen in broadly Kantian terms with a distinctively Humean gloss.

Despite the detailed differences in our interests, Giuseppe and I interacted on a number of occasions: we met at a variety of conferences, jointly organised a conference, edited a book together, and I was very pleased to play a small role in the European Centre for the Study of Public Choice that Giuseppe led in Rome. Giuseppe was always both interested and interesting. Always the scholar. Always the gentleman.

The author confirms that there is no conflict of interest in relation to this manuscript.