The Morality of Marginalism

In the orthodox rational choice model, risk aversion follows straightforwardly from diminishing marginal utility. If diminishing marginal utility holds in all domains (a reasonable a priori assumption), humans should be risk averse over all domains. Nevertheless, this generalization has many apparent counterexamples, ranging from the trivial (e.g. gambling) to an importance so great that any theory of human action would be remiss to overlook it. In this latter category are those pursuits from which humans typically derive existential “meaning”, pursuits supported by an array of aphorisms (“nothing ventured, nothing gained”; “better to have loved and lost than never to have loved at all”) and moralizing tales of deathbed regrets.

In attempting to explain the difference between these existential pursuits and what may be called “ordinary” risk aversion, we may start from a stylized fact: risk aversion is observed to be strongest in pursuits with money at stake – hence its importance for economics. Marwell & Ames (1981), for example, find that contributions to a public good fund are nearly twice as much (43% versus 84%) when the public good is indivisible than when it is simply a proportional monetary return – a counterintuitive result considering non-fungibility of the indivisible good. It is true that money can itself be an existential pursuit for some (and thus invite risk-loving behavior), but more often existential pursuits are objects – like the indivisible public good – with less easily quantifiable returns: games, career, artistry, or love, where the payoff is self-respect or the acclaim of one’s peers.

The obvious resolution would be to say that existential pursuits are characterized by increasing marginal utility. Naturally, this could only apply to things which could not be purchased for money, which exhibits diminishing marginal utility. This does, indeed, describe the observed pattern: love and friendship, for example, are usually thought to be cheapened by being exchanged for money, regardless of the amount.

Nevertheless, this answer tells us nothing about the world; rather, we attribute increasing marginal utility to those domains in which we observe risk-loving behavior. A more satisfying answer must penetrate the black box to ask: what subjects these domains to increasing psychic returns?

A functionalist sort of answer lies in the fact that risk-loving behavior in certain domains confers positive externalities – or, perhaps, to confer positive externalities is risk-loving behavior so far as its effectiveness depends on similar behavior from others. This is certainly the case in modern economies. Entrepreneurship is a risk-loving activity. The expected value is, for most enterprises, small or negative. Yet what market economy could do without them? Low time preference, likewise, is risky. Savings, in whatever form, can be confiscated, lost, or debased, or the saver could suddenly die before being able to enjoy them. But industrial society depends on a large stock of savings to be intermediated. And in general, excellence is usually a consumption good nowadays – not only unnecessary for survival, but actually costly on net to indulge in. In existential pursuits, virtue must usually be its own reward.

Hayek (1988) put the problem this way:

These rules ‘are not derived from any utility or advantage which either the particular person or the public may reap from his enjoyment of any particular good‘. Men did not foresee the benefits of rules before adopting them, though some people gradually have become aware of what they owe to the whole system. (emphasis in original)

These internalized “rules” are not merely the side constraints we typically think of as the “rules of justice” undergirding a liberal polity. They include, importantly, the aspirational goals which impel men to take risks that benefit the community.

If we are to take preferences as given, that would be that. Risk-loving behavior is rational given these norms. But the norms themselves are not rationally justifiable, at least from the individual’s perspective. “Why should virtue be its own reward?” “Why should I give respect to my hard-working peer?” “If ability is mostly genetic, why try at all?” In short, the problem is: in expected value terms, many of the behaviors upon which social cooperation in an industrial and cosmopolitan society depends are individually “irrational”. In order to do any work, aspirational norms must serve as ultimate ends, not subject to any further justification.

For this reason, these behaviors are reinforced by the array of aphorisms and morality tales. Strive to be excellent. Provide for the future. Take care of your fellows. And just as importantly, praise those who succeed at these things, even if you cannot. The internalization of these norms itself constitutes the increasing marginal utility observed in these domains, leading to the risk-loving behavior that keeps the world turning.

The logic of marginalist rational choice stipulates a difference between preference (the shape of the utility function) and choice (the maximizing point achieved on the function). Preference drives the logic; choice is the object of that logic. Internalized norms are preference phenomena – in particular, preferences exhibiting increasing marginal utility. To return to the first example, ∂²U/∂(Respect)² > 0. This is the feature of preferences that results in risk-loving choices.

Max Weber ([1905] 2001) argued that Capitalist society was characterized by ever-increasing rationalization, one aspect of which is the drive to articulate and justify the norms that constitute one’s ultimate ends. The purpose of articulation is to render an idea consciously manipulable. The articulation of a norm, therefore, is an attempt to make the utility function itself an object of choice. The prototypical act of deliberately changing one’s own preferences is the attempt to solve a time-inconsistency problem by the development of a habit or discipline. This is done with some ultimate end in mind: the norm has not yet been actualized, but it has been internalized as an aspirational goal. In any case, any voluntary solution to a time-inconsistency problem must be driven by some preference exhibiting increasing marginal utility.

The rationalistic impulse is precisely opposite; an anti-habit. Increasing marginal utility cannot be rationally justified; the object must constitute an ultimate end. When this end is itself made subject to conscious choice, its character as an ultimate end is destroyed, and it becomes subject to ordinary decreasing marginal utility. When meaning, ambition, and acclaim, like money, become subject to diminishing marginal utility, the result is pauperism, alienation, anomie, and all the other ills which critics have ascribed to industrial society since its very inception.

This suggests why risk aversion rears its head most strongly in the monetary domain. Monetary calculation is a phenomenon of rationalization, just the same as the ex post articulation and rational justification of norms: both bring into deliberative consciousness what was previously accepted tacitly and without deliberation. The mode of thought required for monetary calculation is much more deliberate and conscious than the rough and fuzzy heuristics applied to barter exchanges. This being the case, “ordinary” risk aversion has nothing to counterbalance it. Since money itself exhibits diminishing marginal utility, nothing which is ordinarily reckoned in money terms can exhibit the increasing marginal utility that constitutes an ultimate end. No doubt this encapsulates the unease behind the literature in moral philosophy concerned with the moral limits of markets (see Brennan & Jaworski 2015 for a survey). The extension of the money economy is at the same time the extension of risk-aversion.

Norms as objects of choice are subject to the logic of diminishing marginal utility and thus can never, on their own, overcome time inconsistency or free-rider problems. For these purposes, the absoluteness of such norms is as important as their content. In truth, such norms can never be more than practically absolute. There is usually some cost beyond which any man will abandon a norm. But even this realization serves to undercut its practical absoluteness. However much men may fail to live up to those norms when faced with the exigencies of life, a society which ceases to regard them as aspirational will find itself increasingly unable to sustain social cooperation without externally imposed direction and hierarchy. This, perhaps, is the main factor behind the rise of leftist ideology and the consequent rise of the welfare and regulatory state in the West. Erasmus has triumphed over Luther: not only is “ought” taken to imply “can”, but (note the popular antipathy toward abstinence-only sex education) there is hardly now a more well-accepted argument for “ought not” than to demonstrate “cannot”. Indeed, the modish “privilege” concept is a reaction to an observed “cannot” which attempts – not to assist the under-“privileged”, who may have internalized but have not achieved, but to deliberately obliterate the “ought” contained in aspirational norms.

To be fair, the rationalistic impulse has allowed us to jettison many harmful norms as well. This was the original foundation of Occidental ascendancy following the industrial revolution. However self-destructive the rationalistic impulse may be, it is not obviously worse than the alternative. The problem, as North (1990, p. 16) emphasizes, is that neither the unconscious evolution nor the rational reconstruction of social norms are subject to anything more than weak and distant feedback.¹ There is little guarantee that a society’s aspirational norms confer positive, rather than negative, externalities on the community. Industrial society and the division of labor have certainly expanded the scope for human excellence and risk-loving behavior. But at the same time, the expansion of the money economy, as one instance of the increased rationalization of modern life, appears to undercut the justification for the sort of risk-loving behavior that was once lauded as virtue.

A Note on Natural and Artifactual Man

Buchanan’s “Natural and Artifactual Man” (1979) makes what can be read as a similar point. His concluding sentence, “man wants liberty to become the man he wants to become” can be read as a paean to the importance of self-direction as an aspirational norm.

Nevertheless, the structure of the argument is somewhat different. Buchanan lauds man’s capacity to choose his own goals, rather than choosing within the context of those goals. This, it might seem, is what I have argued exposes man to time-inconsistency problems and anomie.

Buchanan relates the lecture to his previous work on constitutional choice by likening the ‘artifactual’ act of choosing one’s own goals (as opposed to making choices consistent with given goals) to writing a personal constitution – i.e. writing the rules of the game, as opposed to playing the game. There is a close analogy between knowledge and choice here. As Popper (1979) notes, “Knowledge never begins from nothing, but always from some background knowledge—knowledge which at the moment is taken for granted—together with some difficulties, some problems.” In other words, we can question any particular piece of knowledge only from the perspective of some other knowledge which we take for granted. This cannot, of course, go on ad infinitum: Hayek’s (1952) point that the mind can never explain itself is due to the fact that, at some level, we will not be able to articulate the knowledge upon which we stand in order to question it.

Much the same is true of choice: we always choose within some framework of norms which we must take as given. If we step back and make those norms themselves objects of choice, we make that choice within still another set of norms which we hold as given. There is no terminus from which we can ultimately choose all our norms, as a metaphysical concept of free will would imply. The number of steps back we can take is limited by the articulability of the norms within which we choose: at some level (and this, perhaps, is the point of Stigler & Becker [1977], which Buchanan criticizes), the norms within which we choose cannot be made objects of choice: they are too ineffable.

The distinction between playing a game and drafting the rules of the game takes account of one level of choice. The important point is that drafting the rules is itself a game that takes place within a further (possibly unarticulated or inarticulable) set of rules. The act of personal constitution-writing that he describes takes place within a given framework of aspirational norms. If we follow Buchanan in talking about discrete levels of choice, we may say his personal constitution is one level up. Man chooses who he wants to become only by taking the wider society’s aspirational norms as given. The problems described earlier come when man attempts to step back two levels up, and subject those aspirational norms themselves to conscious choice. What further norms does he rely on then?

The constitutional framework, Buchanan rightly notes, is less an analytical framework than an attitude. It is itself an aspirational norm; a deliberate choice to stop the regress at one level and take the norms further back for granted (and commits to doing this, perhaps, by denying their existence or importance). In an ontological sense, however, a strict division between the constitutional level and the level of ordinary action obscures the point by suggesting that choice in the former happens from a god’s-eye view, or perhaps from behind a “veil of ignorance”. This is why the problem of desiring to quit smoking is better expressed as a time-inconsistency problem than a constitutional decision. Man is not assuming a norm-free vantage point from which he can “choose who he wants to become”. All we need to say is that his long-term interests and his short-term interests are in conflict. His long-term interests, of course, are informed by these aspirational norms. He can make choices consistent with his long-term interests to constrain his short-term interests, but he can only do this by taking for granted a given set of aspirational norms – in this case good health or self-respect.

If this were not so, on what basis would he prefer non-smoking to smoking? On what basis would he prefer anything to anything else? This is precisely the danger of anomie in demanding rational justification too many levels back: these aspirational norms cannot be rationally justified because the norms upon which they are founded are unarticulated, and possibly inarticulable. If they are to be rejected on this basis, man has disavowed his power to aspire, and even to prefer.

On the Moral Education of Economics and Economists

In a personal essay that I can no longer locate, Murray Rothbard asks an interesting question: how could Mises, as a thoroughgoing utilitarian and wertfrei economist, display such moral fortitude in defense of liberalism?

We do not have to accept Rothbard’s ethical framework to wonder at the same question. If there is any discipline that exemplifies the impulse to rationalize and articulate what was tacit and unconscious, and to disregard anything not amenable to this treatment, it is economics. The thrust of marginalist logic is to tear down all categorical imperatives. Given enforcement costs, there is an optimal non-zero amount of crime in a society. Given a time-dated income stream, there is often an optimal non-zero amount of debt to take out. And given the disutility of labor, there is an optimal non-total amount of effort to devote to any particular pursuit.

Where money cannot be exchanged for everything, marginalism purports to apply to every action. Very rarely will a corner solution obtain where the answer is “all” or “none”. Thus, economics education might well be expected to finish the job that monetary calculation started, to wit, the corrosion of norms as ultimate ends.

Marwell & Ames again demonstrate exactly this. No variation on their experiment resulted in less than 40% contribution to the public goods fund except running it on first-year economics graduate students, who contributed only 20% of their tokens. They also note, regarding the follow-up questionnaire, that “more than one-third of the economists either refused to answer the question regarding what is fair, or gave very complex, uncodable responses.” Evidently the ordinary norms of fairness animating everyone else had been at least half eroded, and a number of subsequent papers have confirmed the general pattern.

Perhaps, however, the antidote also lies in economics. These norms have long been recognized in the field, but have been widely underappreciated until the relatively recent advent of New Institutionalism. It is true that marginal logic as such applies always and everywhere, regardless of institutional context. But without internalized norms – what are now being called “informal institutions” – there is no guarantee in a non-integral world that maximizing and marginal behavior will result in an extended order. Indeed, Roger Koppl (2003) for example, has reinterpreted maximizing behavior as a special case of rule-following behavior. Whether or not this formulation serves to make the point more clearly than the formulation of norms as domains of increasing marginal utility, the aspect of “human nature” which requires practically absolute rules and aspirations must not be overlooked. Horwitz (2000) makes clear what’s at stake, putting the matter in Buchananesque terms (though, of course, referring to a level further back than Buchanan himself did):

Human actors require constitutional constraints because we are epistemologically unable to generate social order in any other way.

A proper appreciation for the importance of institutions and norms as ultimate ends can, perhaps, reverse any moral decline conditioned by the study of economics. There are, indeed, things worth disciplining one’s self for, regardless of whether the benefits redound entirely to the self. Perfectionism, for example – even taking pride in one’s work – cannot be rationally justified. But where would the world be without the contribution of such “irrational” perfectionists? Where would economics be if the great economists had heeded their own advice and devoted an “optimal” amount of effort to their pursuits?

We do a disservice to economics students, both intellectually and morally, by presenting marginalism without institutionalism. “Optimal amounts of everything” is a powerful and necessary mode of thought. But if we as teachers of economics are to avoid deserving to be drowned with a millstone around our necks, it must not be presented without the importance of those ultimate ends, norms and institutions, both constraining and aspirational, that undergird social cooperation in the extended order.

1. North, of course, uses the increasing marginal social returns of institutions (including norms) to emphasize path dependence and historical contingency for a society. This is a different point than the one emphasized here, that aspirational norms are themselves constituted by increasing marginal utility for the individual. But, norms being socially conditioned phenomena, we may speak of the same feedback mechanism for both.