Utility theory in neoclassical thought has always been subject to the criticism that it is nothing more than mere tautology. Utility is said to explain consumer behavior. But is utility itself observed? Or, if it cannot be observed, how is it operationalized? The answer generally is -- through observing consumer behavior. Thus the criteria for knowing both the explanans and the explandum are exactly the same. Hence, one cannot be said to explain the other. Samuelson expressed his concern on this question as follows:
Prior to the mid-1930s, utility theory showed signs of degenerating into a sterile tautology. Psychic utility or satisfaction could scarcely be defined, let alone measured...Just as we can cancel two from the ratio of even numbers, so one could use Occam's razor to cut utility completely from the argument, ending up with the fatuity: people do what they do. (Samuelson, 1972)
(Samuelson, believed that he his theory of revealed preference "cut utility completely from the argument"without generating a tautology. I'm not so sure. That can be revisited perhaps elsewhere)
In considering the issue of the tautological character of utility, I was struck recently by Mark Blaug's take on the concept. In discussing the problems that utility theory attracts by making it the 'explanation' (his quotes) of consumer behavior, he proposes a solution. The key to understanding utility correctly, he claims, is to recognize its different usage in the theory of consumer behavior and in welfare economics. In the former, it does not "explain" choices, it only describes them:
In the theory of consumer behavior, utility is a quantity that it is useful to regard the individual as maximizing in the interpretation and prediction of behavior. A utility function is nothing more than a way of describing an individual's preferences between various real and hypothetical alternatives. Such a function no more 'explains' an individual's choices than a production-transformation curve 'explains' the state of technology. (Blaug 353)
So a utility function is not intended to explain. It merely describes. However, says, Blaug, matters change when we turn to welfare economics. There, utility does indeed have some kind of independent ontological existence. It is a quantity that must be maximized for people to be better off.
In welfare economics, however, utility is a quantity that 'should' be maximized or that society 'should' help him to maximize. Here, utility is indeed a quantitative concept, whereas in theory of consumer behavior it is strictly speaking no quantity at all but simply a choice indicator....As soon as this distinction is firmly grasped, most of the criticism that was leveled against utility theory as a theory of consumer behavior falls to the ground.
I find it difficult to be as sanguine as is Blaug on this question. Welfare economics is predicated (in part) on the theory of consumer behavior. The former could not exist without the latter. If utility has no separate ontological existence apart from its manifestation in consumer behavior, then how does it suddenly acquire such existence when we move to the realm of welfare economics?
And if it doesn't, then how does welfare economics acquire any content? Welfare economics relies on a concept of consumer welfare defined in terms of maximization of utility (given budget constraints). To say that consumer surplus (for example) is greatest under competitive conditions is to say that the consumer would lose utility under monopolistic conditions. But if utility merely describes or interprets behavior, then there is nothing to be gained or lost. Blaug's solution, though seemingly saving a theory of consumer behavior from the "tautology" critique, appears to do no such thing for welfare economics. Without such a solution, welfare economics has no foundation.
It is possible that I am reading Blaug incorrectly, however. The problem is his bouncing back and forth between the terms "behavior" "choices" and "preferences." Which is it that utility describes? His clearest statement is the last, i.e., preferences (though he also uses the other two). If, indeed, a utility function is a means for describing preferences, then, in Blaug's view, the utility function does explain behavior -- just not the preferences themselves (which it merely describes).
But then we run back into the problem stated at the outset. How do we operationalize those preferences absent behavior? Blaug provides no answer. In which case it cannot be said to explain behavior; in which case it is a poor theory of behavior -- and, even more so, of welfare.
No doubt I am oversimiplifying the matter in countless ways, and therefore many objections can be made. I'd welcome discussion of them (or I will raise them in another post).