Austrian Utility Theory (long)

To my mind, the Austrian school’s account of utility is (1) true (2) completely different from mainstream economists’, and (3) very interesting.

What follows is my attempt to explain.[1] Be warned: It’s long. You should also perhaps drop your notions about what “Austrian economics” means. This isn’t the TV pundit version; below the fold, you won’t find a single fully developed argument in favor of the free market. Nothing about business cycles or interest rates, either.

Let’s start with neoclassical utility. For neoclassical economists, utility is the amount of satisfaction taken by a consumer in the act of consuming. The process has three parts: Desire motivates action, and happiness is the result. Alfred Marshall was among the first to make an important simplifying assumption, namely:

[T]hat the monetary measure of the intensity of a desire also accounts for the intensity of the pleasure resulting from its fulfillment…. [I]n Marshall’s thought, desire and pleasure converged in the same concept: utility.[2]

The utility nexus—desire-action-happiness—can be measured using money as a proxy.

This assumption has allowed a great deal of model-building, generating many interesting and plausible claims about happiness, among other things. Mainstream economics still depicts actors as satisfaction-maximizers whose first resort is to the monetized market. It adds epicycles to explain divergences, and the most interesting research today is often precisely on these epicycles.

Mainstream economists long debated whether utility was cardinal or ordinal . In a cardinal account, we can measure magnitudes of happiness and do everyday arithmetic with them, as well as doing the more complex math that proceeds from things like change over time. A scientifically measurable unit of happiness—a util—might even be found to exist, perhaps in neurotransmitter levels or other measurable brain states, and these we might reliably relate to consumption.[3]

Almost everyone would like utility to be cardinal. Why? Because life would appear to be more tractable that way.

In a cardinal-utility world, utilitarianism would be an exact science: If my copy of Human Action endowed me with 9523 utils, and if burning it left me with an added deficit of 123 utils, and if burning it would give a socialist 9647 utils, then the result would be a one-util improvement in human welfare. Obviously we should proceed, assuming the opportunity costs and externalities were negligible or favorable. Marshall’s simplification was an attempt to deliver something at least a little bit like this result.

Ordinal numbers—of the form “first,” “second,” “third,” and the like—work differently. If utility is ordinal, we can only do comparisons within a ranked set, not any kind of arithmetic. It’s like judging a footrace without a stopwatch. A runner may be first, second, or third. But we can’t say how fast they were. We can’t say how much faster the first-place runner was than the second. We certainly can’t say whether the winner in today’s race was faster than the winner in yesterday’s.[4]

That’s what utility is for Austrians. There’s no such thing as a hedometer, a device to measure utility, and without it we are thrown back on rank order. Each of us judges our own desires in a sort of footrace in our minds. Austrians strongly insist on ordinality, and it seems obvious to me that this is correct.

Ordinality imposes huge limitations. Mainstream economists have grudgingly admitted this, and also that preferences are indeed ordinal, but these admissions raise an uncomfortable question: What then are the models modeling? The formulas of mainstream economics almost always do things that you can only do with cardinal numbers: You simply cannot divide third place by second place and come in at one-point-fifth.

The seemingly cardinal quantities of money don’t help. These quantities, remember, are proxies for utility, given Marshall’s simplifying assumption. But dollars are not constitutive of utility, and they are not indexed to it in any precise way. A poor man may desire a dollar much more than a rich man does, and giving him a dollar may make him happier—simply because the dollar is closer to supplying his own most urgent needs. These needs a rich man has by definition satisfied long ago. In choosing who gets a dollar of charity, there would appear to be no question; we must give it to the poor man.[5]

Then again, perhaps the rich man would perversely be made happier: Suppose him greatly eaten by avarice, and, although he does not need a dollar in any conventional sense, he nonetheless makes much greater sacrifices for it—sacrifices of his morals or his friendships, perhaps, that the poor man would never make. These might indicate that the dollar means more to him. (Note that if our hedometer told us to pay the rich man, we might just say so much the worse for cardinalized hedonic utility: let him suffer his well-deserved unhappiness! I am unsure how the hedometer would register my ambivalence here.)[6]

Dollars have numbers, but those numbers float relative to the utility of other things, driven by factors that are not quantified. The very best one can do is measure the tradeoffs an actor makes, then infer that the actor’s knowledge and values brought him to believe that what he hoped to receive in the trade was of more value to him than what he surrendered.

Of course, it would still be great to link the science of dollars and markets to the science of brains and minds. That’s why cardinality keeps coming back into mainstream economics: not because it’s true, but because it would be nice if it were true, and (more charitably) because we can still generate surprising but testable hypotheses by assuming that it is true. To my mind this isn’t necessarily a bad research program, but it does tend very strongly to persuade the practitioner that Marshall’s assumption is not merely convenient, but actually true.

That’s why the Austrian school gives a much more modest—and therefore in its own way more rigorous—account of utility: Stated simply, an agent will act if he believes that his action will remove a felt unease. As Ludwig von Mises puts it:

A man perfectly content with the state of his affairs would have no incentive to change things. He would have neither wishes nor desires; he would be perfectly happy. He would not act; he would simply live free from care.

But to make a man act, uneasiness and the image of a more satisfactory state alone are not sufficient. A third condition is required: the expectation that purposeful behavior has the power to remove or at least to alleviate the felt uneasiness. In the absence of this condition no action is feasible. Man must yield to the inevitable. He must submit to destiny.

These are the general conditions of human action. Man is the being that lives under these conditions.[7]

That sounds vague because it is. It has to be, if it intends to keep within the proper limits of our knowledge. Taken seriously, the Austrian account of utility is a call to a profound, even radical skepticism about economic model-building and forecasting. It also calls into doubt the rational, utility-maximizing economic actor of the neoclassical approach, because feelings and beliefs are incorporated at the very foundation of the Austrian theory of utility.

Following David Hume, Austrians agree that reason is the slave of the passions, and the work of practical reason is to seek out methods of satisfying the uneases that our passions entail. And we may of course settle on unfit means, owing to errors of knowledge. The knowledge problem thus appears at the very foundation of Austrian economics.

The contrast with the mainstream approach is sharp: Mainstream economists often build first-approximation models that assume perfect knowledge, then try to add the imperfections later. The problem is that no one knows what the world would look like if all actors had perfect knowledge. Not even the economist building the model. The best a modeler can do is to assume that his own (imperfect) knowledge constitutes the perfect knowledge that the actors in his model do not yet know. That’s assuming a lot. The actors might really know better. And undoubtedly everyone’s knowledge can improve, often quite radically.

How do we know how strongly we feel “unease” about various prospects? In our speech we may order our feelings however we like, including inconsistently or incoherently. But we order our feelings in our lives by deciding the order in which to act upon them. We cannot do otherwise. Whatever we are doing right now is definitionally the thing that we believe will remove the most unease on net and at the present time.

This isn’t open to argument, by the way: If you believe that you aren’t acting according to your current most-unease-removing course of action, then what you have found is not a flaw in the theory, but a flaw in your self-understanding: If, on reflecting on these matters, you are prompted to adopt a new course of action, then you have gained knowledge about yourself and about which actions are most fit to your purposes. All the theory has done is prompt you to reflect. If you don’t change your actions, then you make an empty claim.

All of our plans aim to remove various things that we feel uneasy about in our lives—things we are dissatisfied with. We are not seeking the things that would make us happier to do, or have, or experience. We all move toward a state of immobile ease, in which there is no need to act at all. Rest is the end of action, in both senses of the word.

Happiness isn’t quite on point. A thirsty man lost in the desert has a first preference of seeking water. My first preference on a Friday night might be to seek a nice cold martini. If I succeed, I may be somewhat happier. He will be somewhat less desperate. But it would be insulting to call him happy. Is there a precise point at which you hold a balance of zero utils? The “removal of unease” paradigm means we don’t have to answer this probably nonsensical question. Both his case and mine can be fit into the paradigm of removing unease, and for the moment no more needs to be said.

A given unease can also be removed in various ways. Some are more fit than others. But fitness is judged by how well the total of felt unease is removed for the individual: Does it remove the intended unease while generating others? If so, are these other uneases relatively more bearable? Or not?

Of course, it may be better to eschew the economic process and conquer the unease directly. Here Buddhism meets theoretical economics, or it could at any rate. Neoclassical economists are embarrassed by Buddhism and other forms of self-denial that don’t promise positive hedonic payoffs. Austrians need not be:

Buddhist: “Happiness is an illusion. To remove worry, remove desire.”
Neoclassical Economist: “It must make you happy to say that.”
Austrian Economist: “Actually, he might be right, depending.”

The Austrian account of utility also handles intransitivity of preference both elegantly and well.

When speakers’ declared preferences are intransitive — that is, when actors say that preference A beats preference B, and preference B beats preference C, and preference C loops around and beats preference A — the problem is not that their preferences actually are intransitive. It’s that to express an intransitive preference is easy, but to act upon it is impossible. One can always express intransitive preferences, just as one can always write intransitive inequalities using mathematical notation. But to act on an intransitive preference would require a time machine and/or some fairly weird assumptions about causality, identity, and multiple universes. Time doesn’t loop in the real world, so expressions of intransitive preference are always merely expressions.

When actors in experimental situations make choices that seem to imply that they hold intransitive preferences, something is often neglected: All actors are necessarily choosing from only one menu at a time, thus they are always—by definition—making choices that are not intransitive at the moment the choices are made. Preferences, however, can and do change over time, even rapidly.

Actions are also not always about preferences for goods; an actor may act just as readily for purposes of signaling. Signaling is also capable of removing a felt unease, albeit of a social or expressive kind, and this type of unease necessarily takes its place somewhere on the hierarchy of things that we worry about and seek to redress. (One might even argue that procrastination, a classic example of intransitive preference over time, is also a form of signaling: Before beginning work, the procrastinator may signal that he is talented enough to afford to procrastinate, in addition to enjoying the actual pleasure of idleness. After beginning work, the procrastinator signals that he is very, very busy, hence deserving praise. Our culture professes to disfavor procrastination, but procrastination is hard to detect, and we all like to say that we’re busy. Signaling wins on the upside and on the down.)

The Austrian theory of utility tidies up a number of other loose ends as well. Acting on the assumption that utils exist, and that our pursuit of them drives economic activity, we might be brought to various absurdities: Is it really true that he who dies with the most utils wins? Is that not a sorry way to live? What do you do with them when you’re dead? Do utils hang about our souls, enriching them in some long-term way? If not, why want them? Can a big util bank compensate for future disutility? How evanescent are utils, and if they are very evanescent, shouldn’t we be ashamed to chase them?

Austrians dismiss these questions as malformed, the product of mistaken premises. Expected utility ex ante is not an expected increase of anything. It is an expected decrease of something—a decrease of our universal, nearly limitless net unease, which may always rise again: The goal of acting man is to deploy his varied resources to remove the unease caused by all sorts of problems, great and small, present and future. Thirst. Hunger. The urge to write. The urge to read. The need for psychological or spiritual fulfillment. The need for transcendence. While the Austrian account of utility is exceedingly modest in some senses, in others it isn’t. It’s a theory of everything human.

Notes

[1] IF YOU ARE AN AUSTRIAN SCHOOL ECONOMIST, please note that I’m self-taught. I’ve read a lot of Mises, Hayek, Rothbard and company, and I’m trying to summarize, with what I trust are a few forgivable extensions of my own. If I’ve screwed up somewhere, I’ll happily revise and give credit where it’s due. I don’t mind being corrected. I welcome it.

[2] Rozenn Martinoia, That Which Is Desired, Which Pleases, and Which Satisfies: Utility According to Alfred Marshall. Journal of the History of Economic Thought 25(3), September 2003.

[3] In our house, we define a util as the amount of utility generated by eating exactly one Reese’s Piece. That this is also the smallest unit of happiness—an idea heretofore unknown to philosophy or economics—is easily demonstrated: No one ever subdivides a Reese’s Piece intending to consume only part of it. They may do so to prove an obscure point in moral philosophy, but this of course brings many utils of its own, obscuring the fractional utility of the consumption as well as calling into doubt that which it attempted to demonstrate. An uncertainty principle may apply.

[4] Note that we can bundle our “second” and “third” preferences, and they may be better considered as a whole than our “first” preference. But this conclusion will not always hold, not in the same way that four plus five will always be greater than six.

[5] There are other reasons why redistribution may not always be utility-maximizing over time. David Schmidtz’s Elements of Justice, ch 27, rightly observes that “large-scale need-based distribution has never been the key to making people in general less needy”; what we find compelling as an act of individual charity may be a poor choice of model for a social order.

[6] New brain scanning technologies and a generous use of survey questionnaires have produced an effort to restore cardinality to the ambitious place it held in Victorian era. Deirdre McCloskey recently attacked this trend from a more-or-less Austrian perspective in, of all places, The New Republic. Whether you agree with her will depend on philosophical commitments beyond the scope of this post.

[7] Human Action, pp 13–14. The idea that man is the being who lives under conditions of unease and constrained choice is, weirdly enough, one of the things that brings me to reject the idea of a personal God. If this is a good definition of a person, then God is clearly excluded.