by Mario Rizzo
There is a tradition of thought in economics that views the rationality of individual actions as non-falsifiable. There are variations in how this tradition might be justified. These do not concern us to any significant degree here. For concreteness I shall examine the position of Ludwig von Mises (excerpted below) because of the purity and clarity of his argument.
Economists want to abstract from any particular theory of human motivation. In particular, in the early years of the twentieth century, they were keen to distinguish between the subjective theory of value and hedonistic (pleasure-pain) theories associated with Bentham and later W.S. Jevons and F. Y. Edgeworth. So they wanted to say that people choose according to whatever standard they might consider important or on whatever deep basis psychologists might discover. This is not the concern of economists (or so Mises and others argued).
Therefore, when economists speak of people seeking to increase their well-being they mean that they do so in terms of whatever they consider important – pleasure, moral values, long-term interests, short-term fancy, and so forth. Economists also did not want to take a position on how carefully individuals choose what goals they want to attain. Thus, an individual increases his well-being, as he sees it, when he drinks away his paycheck as when he spends it on supporting his family.
Suppose, however, an individual claims that he wants to support his family but uses means that are, objectively, not adequate to the end. Spending on drink does not accomplish the job. So the individual, in this view, is simply mistaken but he is not “irrational.” He uses the best technology he knows (although he is foolish or stupid). But, more likely in this context, he is really aiming at some other goal. He wants perhaps to escape his problems in a quick and easy way even at the expense of his family.
The root of the matter is that we have, by a methodological decision, no admissible way to determine a person’s goals except by reference to his actions. Economists infer the ends from the observed behavior. They also can infer the subjectively-perceived means or technology of attaining the ends from the behavior. At this point inferential picture is complete. Thus, we have entered into a tautology, as Mises readily admits. The effect is dramatic: Individuals are always acting to increase their wellbeing as they see it.
This is perfectly reasonable if the purpose of the economist is simply to analyze (“predict”) behavior. The economist can alter the means-ends framework imputed to the agents until finds a plausible one that rationalizes (literally) the observed behavior. He can even build in error with respect to the means. The agents can make errors with regard to technology or expected market prices of inputs. Yet in a momentary equilibrium they are maximizing their putative objective functions. As such these functions consist simply in momentary objectives and do not necessarily reflect the agents’ conception of their own longer-run welfare.
On the other hand, we cannot use this approach to do normative welfare economics. Consider if we did. We would be incapable of even saying that a person did not maximize his wellbeing as he saw it. Why is that not a good intellectual box? That is because the maximization statement follows definitionally from the assumption that the actor is engaged in action which, by Mises’s definition, means rational behavior. We have simply labeled the behavior rational at the outset: no argument, just an assumption. Thus, we are not saying anything about the subjective well-being of real-world individuals. We are simply drawing out the implications of what we have assumed about them.
It is important to recognize that Mises himself did not rely on the postulate of rationality as the basis of welfare economics (at least not entirely). Mises often judged government interventions in the market, for example, on the basis of how well their outcomes matched the goals of the interventionists. Thus the standard of evaluation is that of the social planner, and not of the economic agents themselves. The purpose of the attribution of frustrated goals to the agents is to show how the market might generate processes that offset or counteract the goals of the interveners. This, of course, is a positive use of the rationality postulate (Cf. Mises, Liberalism, 78 and elsewhere).
(I should also note that some of Mises’s followers have not been so careful. They believe they have found the invincible welfare standard that “justifies” all voluntary action. Murray Rothbard is one example.)
In the extended quotation below I add italics to bring attention to the most important aspects of Mises’s argument.
Ludwig von Mises, “The Task and Scope of the Science of Human Action,” Epistemological Problems of Economics (1933, this essay was probably written in the late 1920s), pp. 34-5.
“Everything that we say about action is independent of the motives that cause it and of the goals toward which it strives in the individual case. It makes no difference whether action springs from altruistic or from egoistic motives, from a noble or from a base disposition; whether it is directed toward the attainment of materialistic or idealistic ends; whether it arises from exhaustive and painstaking deliberation or follows fleeting impulses and passions. The laws of catallactics that economics expounds are valid for every exchange regardless of whether those involved in it have acted wisely or unwisely or whether they were actuated by economic or noneconomic motives. The causes of action and the goals toward which it strives are data for the theory of action: upon their concrete configuration depends the course of action taken in the individual case, but the nature of action as such is not thereby affected.
These considerations have an evident bearing on the widespread tendency of the present age to appeal to the irrational. The concepts rational and irrational are not applicable to ends at all. Whoever wishes to pass judgment on ends may praise or condemn them as good or evil, fine or vulgar, etc. When the expressions “rational” and “irrational” are applied to the means employed for the attainment of an end, such a usage has significance only from the standpoint of a definite technology. However, the use of means other than those prescribed as “rational” by this technology can be accounted for in only two possible ways: either the “rational” means were not known to the actor, or he did not employ them because he wished to attain still other ends perhaps very foolish ones from the point of view of the observer. In neither of these two cases is one justified in speaking of “irrational” action.
Action is, by definition, always rational. One is unwarranted in calling goals of action irrational simply because they are not worth striving for from the point of view of one’s own valuations. Such a mode of expressions leads to gross misunderstandings. Instead of saying that irrationality plays a role in action, one should accustom oneself to saying merely: There are people who aim at different ends from those that I aim at, and people who employ different means from those I would employ in their situation. (Emphases added) ”
 If the Misesian wanted to say that certain behavior did not advance subjective welfare he would have to say that the individual’s behavior was not “action” but simply stimulus-response. Modern psychology, however, recognizes that there are cases in between these two poles.