Social Welfare Functions, Again

I know I discussed it relatively recently, but I feel I should say a few more words more about the rhetoric of so-called “social welfare functions,” mostly because so many dismiss the baleful effects of bad economics in the conservative style by waving their hands at it. It’s meant to be a way to introduce interpersonal ethics, equity issues, into neoclassical welfare economics, to make it a full ethical theory rather than an ethical half-theory. In reality, it’s a rather daft, misleading bit of bad ethical philosophy, rhetoric, mostly about deflecting criticism.

If you recall my previous discussion of the odd way the term “utility” is used in neoclassical economics, you may recall the dominant modern interpretation is that it has no content beyond indicating preference ranks for a given individual. There’s nothing “behind” it, only the fact of the ranking. This interpretation is often called “preference rank utility,” although what the word utility is meant to add to that formulation is anyone’s guess and it should simply be “preference rank.” Under that definition, “utility” is undefined in interpersonal contexts. That is to say, it’s grammatical nonsense to talk about how the “utility” of person A compares to that of person B. One has a “utility” defined relative to A, and a “utility” defined relative to B, but there is no “utility” defined relative to both at the same time. Despite this, it’s not uncommon to see economists who profess to be using modern preference rank utility then move on without skipping a beat to adding up “total” utility by using a “social welfare function.” One cannot avoid the implication they simply don’t know what they’re talking about.

If a “social welfare function” is to be sensible, one must use an interpretation of “utility” in which something lies behind individual presence ranks and so exists in interpersonal contexts, as in the older “inaccessible inner perceptions of satisfaction from preference fulfillment” definition. You may recall that older definition, related to utilitarian ethical philosophy in which “utility” typically related to usefulness for an implied end as satisfaction, happiness, well-being, was also formerly used to shut down interpersonal “utility” comparisons to produce an ethical half-theory. The earlier approach to achieving that end was based on interpersonal utility comparisons not being undefined, as under the modern preference rank definition, but being technically and scientifically impossible, there being no way to obtain information on these inner perceptions of satisfaction. Later economists developed a desire to be viewed more as scientists, despite their philosophical and ethical concerns, leading them to want to shed the non-observable “nonsense” of their older definition of “utility” and adopt the new observable preference rank one. In a practical sense both approaches led to the same conclusions because although the “inaccessible inner perceptions of satisfaction from preference fulfillment” was not defined only relative to individuals, individual preference ranks were the only observable information meant to relate to it. 

The funny thing about the “inaccessible inner perceptions of preference satisfaction” utility is that, as ethical philosophy, it’s complete garbage, in the sense few would likely agree with it if they could actually implement it. Say one had a dream where one could access these postulated inner perceptions and one found they were greatest with someone, oh heck let’s say Adolph Hitler, whose perceptions were so much stronger they dwarfed all others. So the ethical thing to do would be to let Mr. Hitler victimize everyone? Would anyone seriously accept maximizing that type of “utility” as sound ethical philosophy if they could implement it? Unlikely. As I like to say, it’s an ethical theory whose utter implausibility is masked by it utter impossibility. It’s not a real ethical proposition but a rhetorical maneuver. However, if one starts with “utility” defined in that way, and has no actual idea the size or intensity of the relevant inner perceptions, one might want to “weight” them according to some formula in order to add them up and get a total “utility” or inner perception across people.

It’s garbage. Bad philosophy. One claimed implausibly to want to maximize a peculiar sort of “utility,” then one basically ignored it and introduced unrelated ethical thinking (ironically possibly involving other takes on utility), and made up conclusions having nothing really to do with it.

It introduces the ethical stylings of philosophically clueless economists, the ones making these social welfare functions, based on who knows what ethical thinking, and sets economists up as ethical arbiters. Why? What does your average economist know about ethical philosophy? 

It does nothing to address the problems introduced by bad economics in the conservative style, as that rhetoric remains untouched. The ethical rationale for interpersonal ethics and equity issues remains external to neoclassical welfare economics, which remains an ethical half-theory.

The odd formulation supports rhetorical arguments in some forms of bad economics in the conservative style in which interpersonal ethics, equity issues, are viewed as irrational, unjustified, arbitrary, mere “virtue signaling,” nothing that should be taken seriously in economic evaluation.

If one is serious about addressing the dodgy normative rhetoric of anti-democracy bad economics in the conservative style, one won’t be distracted by the essentially mathematical issue of economists potentially using “social welfare functions” to rank perfectly competitive market outcomes. 

 

Social Welfare Functions

Maybe this week we can talk about the rhetoric of “social welfare functions” in anti-democracy bad economics in the conservative style and real neoclassical welfare economics. Might be good for a laugh or two.

To understand what’s going on with social welfare functions, one must return to the peculiar, idiosyncratic, suspiciously odd definition of “utility” used in neoclassical welfare economics. The most common modern definition is it’s just another way of talking about individual preference ranks. So if person A prefers good X to Y, we can say “A gets more utility from X than Y,” but it’s meant to add no content beyond that. It’s simply a funny way of talking about individual preference ranks. As such, it is linguistically incorrect to invoke it in an interpersonal context. It doesn’t make sense to say something like “A gets more utility from X than Y and we want to compare that to the utility B gets from whatever.” It’s really a form of equivocating on terms. One is invoking two different “utilities,” one defined relative to A, the other defined relative to B. If one is going to talk about “utility” interpersonally, one needs a definition of utility that goes beyond preference ranks of given individuals as say the older “inaccessible internal perceptions of satisfaction from preference fulfillment,” meant to exist in an interpersonal context.

In their zeal to be positivists, some economists dealing with normative economics propose it doesn’t really matter which definition of “utility” one uses, it’s either defined to be individual preference rank or it’s only inferable by individual preference rank, but of course that’s incorrect. Not only does what one can sensibly say about “utility” depend on how it is defined, but the philosophical arguments in favor of why anyone should care about maximizing it in a social context, indeed what that even means, change depending on the definition. One can’t simply switch back and forth. If we’re talking about preference rank utility, the normative proposition society should strive to “maximize utility” is about how one should act with respect to individuals moving up their own preference ranks with no implications for resolving interpersonal conflicts of preferences. If we’re talking about inaccessible internal perceptions of satisfaction, we’re attaching normative significance to those perceptions so, for example, if one particular person had very strong such perceptions, maximizing “utility” implies that person’s preferences take precedence over all others. Critics will argue the problem doesn’t exist in reality because the whole point is we can’t access inaccessible perceptions, but one can hardly take seriously as an ethical theory we should do something with the proviso we can’t actually do it, and if we could, then we shouldn’t. It’s misdirection.

All that was just to review where the prohibition against making “interpersonal utility comparisons” in neoclassical welfare economics comes from, which informs the whole theory about perfectly competitive market outcomes that is taken up and misused in bad economics in the conservative style. The “social welfare function” is meant to get around that restriction, reintroducing interpersonal ethics, thus changing the entire argument in neoclassical welfare economics that informs arguments in bad economics in the conservative style about not interfering with or distorting market results. However, and perhaps less obviously, discussions of “social welfare functions” can happily co-exist with the rhetoric of bad economics in the conservative style and indeed even be used to support it. Not sure how that works? Allow me to explain.

First, social welfare functions allow the expression of interpersonal ethics, but only in an awkward way unjustified within neoclassical welfare economics itself. As noted earlier, one cannot really even talk about “utility” in an interpersonal context under the usual economic definition. The result is to make interpersonal ethics appear unjustified, arbitrary, a matter of preferring some mostly identical theoretical ciphers over others, suspiciously similar to the notion from bad economics in the conservative style that interpersonal ethics are empty virtue signaling, nonsense. But of course the actual ethical justification has nothing to do with that “utility.” Let’s have an example. Say one supposes A should not murder B. There is no rationale in neoclassical welfare economics to not allow that as we can’t compare the individual preference ranks of the two individuals. However, we could make up a “weight” based on our ethics and express our objection to murder by attaching a greater weight to the preference rank (or really again the strength of the internal perceptions) of the intended victim. But of course, that expression doesn’t express the ethical rationale.

Second, given the ethics of economic power may be context specific, for example, one may accept most goods being allocated on markets if acceptable distribution of economic power but exempt others, say vaccines or housing or whatever, defensible social welfare functions may be quite complex. In that sense, social welfare functions may be used rhetorically as “roadblocks,” a common stratagem of throwing up of technically difficult or perhaps practically impossible demands meant to make changes from some preferred situation difficult, costly, or indeed impossible.

Third, social welfare functions can be used in conjunction with fake indifference to present objections to the status quo based in social welfare functions as less ethically justified in economic theory than the status quo because someone needs to write up, defend social welfare functions. Of course, the same objection could be made to defending any particular status quo, or even The Status Quo in general, no matter what it is. There will always be an implied virtual social welfare function supporting the resolutions of interpersonal conflicts of preferences actually taking place. One might retort, sure, but that’s not why anyone supports any given status quo or even The Status Quo in general, no matter what it is. Indeed. And the proposing of explicit social welfare functions is similarly not the real reason anyone might object to any status quo. One can also find weird hybrid road block arguments, as trying to maintain the policy recommendation from bad economics in the conservative style one should first concentrate on attaining some approximation of a perfectly competitive market and only then introduce a social welfare function. 

Did I break the bank on this one? Sorry. Just trying to make people aware of the sorts of rhetorical games one may find not only in real neoclassical welfare economics but especially in anti-democracy bad economics in the conservative style. Keep your critical faculties close if you venture there.