A while ago I did a series of posts about the confusion I contend is caused by the idiosyncratic concept of “utility” in neoclassical welfare economics. I’d like to re-visit that issue this week and propose, or rather re-propose, a simple solution.
The essential problem with how the term “utility” is used in neoclassical economics is the ease with which one can equivocate on the idiosyncratic, heavily attenuated concept of “utility” meant to serve in the distinctive ethical half-theory of normative neoclassical welfare economics on the one hand, and other common and rather broader definitions of “utility” from ethical philosophy meant to serve in full ethical theories on the other hand. That problem is further complicated by equivocation involving two different idiosyncratic definitions of “utility” specific to economics itself. Let’s start with the latter issue.
The now standard, conventional, most common explicit interpretation of “utility” in the context of neoclassical welfare economics is that it is simply a funny word we can use to discuss the preference rankings of individuals. The other, older, typically implicit interpretation of “utility” is that it refers to subjective perceptions of satisfaction from preference fulfillment. In practice, the two definitions amount to the same thing because according to the subjective perception of satisfaction interpretation of “utility,” the only acceptable observable or empirical indication of such perceptions is, yes, the preference rankings of individuals. However, the distinction does, of course, have conceptual and philosophical consequences in the context of normative economics.
Under the preference rankings of individuals interpretation, talking about interpersonal utility comparisons is grammatical nonsense because the word “utility” is undefined in that context. It would be like talking about the intensity of a square root or some other bit of nonsense. Under the preference ranking definition of “utility,” “utility” doesn’t exist, per se. It’s not a thing. There is no “utility.” How can we say that? Because if it referred to anything that actually existed, whatever it was would still exist in an interpersonal context, and we would be able to talk perfectly sensibly about interpersonal utility comparisons, which we cannot do if we define “utility” in terms of the preference ranking of an individual. In contrast, under the definition of “utility” as an inaccessible internal perception of preference satisfaction, the phrase “interpersonal utility comparison” makes grammatical sense because, of course, the internal perceptions to which “utility” would then refer are actually meant to exist in some sense, albeit only as subjective perceptions, and they wouldn’t simply vanish into a puff of smoke when another person arrives on the scene. However, under the perception of satisfaction from preference fulfillment interpretation of “utility,” interpersonal utility comparisons are impossible because the perceptions involved are meant to be inaccessible and subjective in some way. So, again, the practical results are the same.
The funny word to refer to preference rankings of individuals interpretation of “utility” is certainly the standard, formal, most commonly accepted explicit definition, but one commonly hears statements that at least appear to imply the perception of preference satisfaction interpretation of “utility” involving, for example, interpersonal “utility” comparisons being impossible rather than undefined, distributing “utility” using distributional policy, weighting and combining the “utility” of different people, conflating “utility” with money or total economic output, and so on. Similarly, one hears people talking about cardinal versus ordinal “utility” functions with little confidence they understand the implications of those mathematical expressions for the underlying meaning of “utility” and the consequences for the normative evaluation of those concepts. I’ve discussed the issue previously, but briefly, a cardinal utility function obviously is clearly inconsistent with a preference ranking definition of “utility” while an ordinal utility function leaves the underlying concept of “utility” indeterminate.
Moving to the other definition, one philosophically interesting feature of the subjective perception of preference fulfillment interpretation of “utility” is that it clearly bears some relationship to the more traditional definitions or interpretations of “utility” one finds in ethical philosophy, which typically have to do with human welfare or happiness. However, what makes this interpretation stand out from a philosophical perspective is that accepting the goal of maximizing “utility” defined in terms of these subjective perceptions would not only be absurdly ethically controversial but entirely ethically implausible. One can establish that result easily enough by setting up a little thought experiment, a dream let’s say, where one postulates access to the normally inaccessible subjective perceptions and then works out the consequences of trying to maximize them. The problem occurs when one entertains the possibility of what I usually call “superconductors of utility,” individuals who can generate perceptions of such passion or strength or whatever governs the size or level of intensity of such internal perceptions they dwarf the perceptions of others.
Incidentally, I should point out that the need to continually discuss two different and conflicting definitions of a word or concept that is really at the heart of neoclassical welfare economics, a theory that is at least nominally all about maximizing total social “utility,” really demonstrates the sloppy, non-rigorous aspect of orthodox neoclassical economics when it comes to the all important normative content. It’s almost as though economists working in that tradition don’t really care what the underlying concepts are, let alone if they’ve properly evaluated the normative arguments they’re making. They’re clearly quite interested in manipulating the mathematical expression of those concepts, but again in practice the math in this case works out the same, so one infers it may appear to them not very important to sort out, which is rather comical given the two interpretations lead to two entirely different theories in a normative or ethical sense. Indeed, I’ve had other economists suggest to me “utility” is basically undefined in neoclassical economic theory and can mean whatever one likes it to mean, which again is comical to me because one could never establish the core result of neoclassical welfare economics, the ostensible ethical superiority or optimality of a Pareto optimal, economically efficient, perfectly competitive market system or outcome, based upon maximizing an undefined concept, that is, maximizing whatever one feels inclined to maximize. No serious, honest philosopher would proceed with an ethical or normative argument while purposefully refusing to define the terms in an obvious attempt to complicate and confound his or her readers. The refusal of the economics profession to conclusively take a stand on what “utility” is actually meant to represent very nicely demonstrates why economists have no business dabbling in ethical philosophy or normative issues.
I’ve discussed all this material in previous posts, so I won’t belabor it again here. No, what crossed my mind the other day, which is again an idea I’ve discussed previously and not a new idea by any means (see, for example, my post Fixing Bad Economics By Removing References To Utility, September 23, 2020), is that there seems a pretty easy way to eliminate much of the confusion caused by the use of the term “utility” in neoclassical welfare economics, which is to simply stop using the term and instead replace every instance of “utility” with “preference rankings of individuals,” which is its conceptual or practical equivalent, depending on which definition of “utility” one is using.
Instead of saying individuals maximize their “utility,” say individuals express preferences.
Instead of talking about an economic system that “maximizes total social utility,” talk about a system consistent with certain ethical or normative propositions relating to how one should act toward other individuals expressing their preferences in certain context. (And in case you don’t get where I’m going with that from my previous posts, by “particular contexts,” I mean above all else the context in which those preferences do not conflict with the preferences of anyone else, a one person world, as opposed to the context in which they do conflict and we need to get into resolving interpersonal conflicts of needs and desires.)
Instead of talking about social welfare functions that weight the “utility” of different people, talk about social welfare functions that weight the preference rankings of different individuals.
Instead of talking about aggregating “utility” across individuals, talk about aggregating the preference rankings of different individuals across individuals.
Instead of talking about expected “utility,” talk about expected rank in the preference rankings of a given individual.
I contend this simple act alone, which could easily be made with a minimum of fuss, would do a great deal to clarify the ethical half-theory structure of normative neoclassical welfare economics and distinguish that unique, idiosyncratic structure from the more conventional, full or complete ethical theory structure of proper utilitarian theories one finds in ethical philosophy. Let’s call it the Don’t Say Utility Challenge. Economists, and purveyors of bad economics, I wonder, can you do it? For the next month? The next week? How about the next five minutes?
Did I sound a bit sarcastic just then? You know although I’m quite serious when I recommend this approach to intellectually serious readers of economic theory, I don’t expect purveyors of bad economics to entertain the idea for even a moment. Why? Because bad economics is to a large degree an exercise in rhetoric rather more than a serious intellectual endeavor. It is useful for some people in a very practical, financial sense, and as one might expect it is very generously financially supported by them, precisely because of the confusion and darkness it generates about the normative aspects of economic issues, the ability bad economics has to pull the proverbial wool over some people’s eyes and submerge important and relevant ethical issues relating to evaluating market systems and outcomes. We should all work to fix bad economics, but as this business with the funny and continued use of “utility” in economics demonstrates, we probably shouldn’t expect too much help from purveyors of bad economics.
