I was going on about my “don’t say utility” lark online the other day, that is to say, I was arguing neoclassical economists should just stop talking about “utility” because of the potential for equivocation on different definitions of the term and start talking instead about interpersonal preference rankings of individuals, which is the relevant concept of utility for neoclassical welfare economics. As if on cue, the only real response my idea generated was from a fellow wanting to establish that individual preference ranking was actually a philosophically or ethically inferior way of defining utility because ordinal utility leaves out important considerations addressed by cardinal utility. In one way it was a depressing bit of the same old same old of precisely the type my “don’t say utility” challenge was meant to avoid, but in another way I suppose it was also an opportune comment in the sense it gives me a good reason to review ordinal and cardinal utility and at the same time clarify what I’m talking about when I discuss neoclassical welfare economics as it relates to what I call bad economics or sometimes folk economics. Yes, it’s a two for one this week.
I discussed ordinal and cardinal utility relatively recently in my post of October 21, 2020, rather prosaically entitled Ordinal and Cardinal Utility in Neoclassical Welfare Economics, but what the heck, I’ve noticed that by constantly talking over these issues online my thoughts are prone to a certain amount of development over time, a good sign surely, and at the very least I’m always finding new and better ways of saying the same old things. Anyway, let’s have another go at it now, and I’ll add a bit to the end about what I’m discussing when I discuss neoclassical welfare economics and how it relates to ubiquitous bad economics.
I would suggest the tendency among economics to want to talk about different conceptions of utility in terms of the mathematical attribute or characteristic of ordinal and cardinal is an interesting example of how mathematical formalism can complicate consideration of the underlying normative or ethical issues in neoclassical economics. One initial complication is that if we’re going down this road, we must distinguish ordinal and cardinal concepts of utility from ordinal and cardinal utility functions because one may have a cardinal concept of utility but choose to use, or even believe one is restricted to using, an ordinal utility function when discussing or expressing it in the context of neoclassical welfare economics.
What’s important for us in terms of understanding normative neoclassical welfare economics is the concepts, not the potentially variable mathematical expression of those concepts, so let’s talk about concepts rather than functions. If one is using an ordinal concept of utility, then one is talking simply about the preference rankings of individuals. The only thing that matters is the order of the preferences for a given individual. There is nothing behind or beyond the preference rankings of individuals to which one is referring with the word “utility.” One may say the ordinal concept of utility doesn’t refer to anything that actually exists, per se, but is simply a word we can use when describing or discussing individual preference rankings. Under this ordinal conception of utility, an “interpersonal utility comparison” is quite literally nonsense or grammatical gibberish. Consequently, any normative propositions based on an ordinal concept of utility must refer to the preference rankings of individuals, not any of the various things referenced in other definitions of utility as being behind those preferences, such as happiness, satisfaction, etc. It makes grammatical sense to talk about those other things in interpersonal contexts, and one can sensibly talk about degrees of happiness or strength of perceptions of satisfaction and so on, which are attributes of a cardinal rather than an ordinal concept of utility and are indeed inconsistent with an ordinal concept of utility.
In contrast, a cardinal concept of utility necessarily refers to something behind or beyond observed preference rankings of individuals that continues to exist even in interpersonal contexts and to which one can attach some notion of intensity or strength, in old economics typically internal perceptions of satisfaction from preference fulfillment. The other traditional interpretation in old economics, happiness, lifted directly from utilitarian ethical philosophy, is a bit dodgy because preferences in neoclassical welfare economics are not necessarily restricted to things that make one happy in any traditional sense. Using a cardinal concept of utility, one can sensibly introduce ethical propositions relating to whatever is purported to lie behind preferences, satisfaction, happiness defined in certain ways, etc. In terms of normative economics, we’re obviously in a whole different realm from that of an ordinal concept of utility.
A cardinal concept of utility is consistent with neoclassical welfare economics if and only if we have no way of measuring or inferring the strength or intensity of the relevant internal perceptions, thus rendering interpersonal utility comparisons impossible. That’s because the internal logical of the normative or ethical argument in neoclassical welfare economics involving Pareto optimal, economically efficient, perfectly competitive markets being socially optimal requires interpersonal utility comparisons to be either gibberish or impossible.
Both the ordinal and cardinal concepts of utility that work in the context of neoclassical welfare economics lead to neoclassical welfare economics being an ethical half-theory, but they arrive there via subtly different routes. The ordinal concept of utility straightforwardly removes utility from the issue of how to ethically resolve interpersonal conflicts, which is what the lion’s share of ethics, and the evaluation of economic systems and outcomes, is all about. We could never have a full ethical theory based only on individual preference rankings. Such an ethical theory would be an ethical half-theory at best, and would obviously need to be supplemented with ethics relating to the important issue of resolving interpersonal conflict.
The cardinal concept of utility could serve as the basis of a full ethical theory if we could measure or infer the intensity or strength of the relevant internal perceptions and use that to resolve interpersonal conflicts, but we cannot, by assumption or definition. Thus, the cardinal concept of utility that works in the context of neoclassical welfare economics also cannot be used to construct a full ethical theory. It must also be supplemented with ethical beliefs relating to how to resolve interpersonal conflicts in practice. It also leads to at most an ethical half-theory.
If one considers the utter ethical implausibility of accepting the goal of maximizing the relevant cardinal concept of utility if it were possible, one may appreciate that result is likely not coincidental. To see the ethical implausibility, consider a dream where one could access the relevant internal perceptions and actually maximize social utility. Basically, the preferences of the person with the strongest internal perceptions would rule, be they benign or otherwise. That’s why I often point out that even with the potentially relevant cardinal concept of utility, the use of the word “utility” in neoclassical welfare economics should not be interpreted as representing any real or genuine support for utilitarian philosophy but really more of a rejection of it, confusingly expressed, a clumsy attempt to rule out utilitarian ethical philosophy be defining “utility” in such a way that it becomes irrelevant to the lion’s share of situations generating ethical issues that need resolving.
With the cardinal concept of utility, the temptation is always there to “fix” or “improve” neoclassical welfare economics by introducing exogenous ethical ideas from traditional utilitarian ethical philosophy or elsewhere, along the lines I often discuss in the context of “social welfare functions.” The ethical basis of those exogenous value inputs is not developed in neoclassical welfare economics. They’re really just the subjective views of whatever random economist is putting forward those inputs. In that sense the theory itself remains an ethical half-theory. The main point to remember in this context is regardless of how one feels about economists playing ethical philosopher and trying to decide subjective ethical issues that really belong with the people and democratic government, that activity is distinct from the normative argument presented in neoclassical welfare economics, per se. The normative economic argument that leads to bad economics and anti-democracy sentiment is the one involving Pareto optimality, economic efficiency, perfectly competitive markets, etc. It’s distinct from the quasi-economic subjective ethics of free-floating “welfare analysis.” We can clarify the relevant argument by considering the necessary features of any concept of utility that works in the context of that argument, and by talking about preference rank utility versus internal perception utility, or utility that is proposed to exist in some form or not exist, rather than about the mathematical attribute or characteristic of ordinal and cardinal.
If academic economists no longer wish to promulgate the normative or ethical argument presented in neoclassical welfare economics relating to optimal market systems and outcomes, if they no longer see any utility in teaching it (sorry, couldn’t resist), then they should stop teaching it, because it can easily lead to anti-democracy bad economics and folk economics. If academic economists continue to teach it, or worse, try to teach it and make a complete hash of it because they neither know nor care what it actually says, which seems all too often the case, then it will be sensible and indeed necessary for others to try to explain what it really says, how it really works, in order to fight the bad economics based loosely upon it. Academic economists are obviously in the best position to properly explain the argument in neoclassical welfare economics and dispel the confusion that leads to bad economics and anti-democracy sentiment, but if they’re onto other things, more general ethical philosophy and welfare analysis, social welfare functions, models, random optimization problems, stories, and can’t be bothered with anything as mundane as addressing the bad economics, folk economics, and anti-democracy sentiment they did so much to foster, then unfortunately it devolves upon others. You should help me fight bad economics by understanding what orthodox neoclassical welfare economics really says and addressing the incorrect versions of it associated with bad economics.
