Sunday, October 07, 2007

An Interesting Question

A discerning reader asks, "what is social utility?"

Good question!

Short answer: There's no such animal as "society" that has utility or well-being. Treating society as thinking, evaluating, choosing entity is a metaphor, and almost always when this metaphor is used it leads to confusion -- the formal name is the fallacy of reification. Society is, literally, a collection of individuals and their relationships, not a big meta-person.

So "social utility" is nonsense on stilts.

Long answer: Here are 5 possible definitions by which social utility or social welfare or the common good might be defined.

1) There is an idea of social welfare, something that can be maximized. At one time it was supposed that this was the sum of every individual's personal utility, a measure of their well-being.

However, the idea is completely wrong, since utility refers to ordinal, subjective (individual) rankings of things. Economist Kenneth Arrow put the final nail in the coffin with his "impossibility theorem," that proved that in any reasonable setting there's no way to generate a consistent, objective "social utility function." More or less, if I prefer A and you prefer B, there's no way to discern that A is socially better than B without making a value judgment.

2) A second idea refers to Pareto optimality. Pareto optimality relies only on individual judgements of their own utility, and so doesn't impose any additional utility standard. Suppose we can take an action, A. If A makes no one worse off, and at least one person better off, then action A increases "social welfare" and should be undertaken; action A is a "Pareto improvement." Presumably everyone unanimously agrees that Pareto improvements should be undertaken, since no one feels worse off by them. The situation in which there are no further Pareto improvements is Pareto optimal.

This is a reasonable standard, but it cannot evaluate situations in which even one person is made worse off, so it is incomplete as a standard.

3) A third idea, the Kaldor-Hicks criterion, refers to social welfare as the totality of effects of some action, event, or state, as measured by different peoples' willingness to pay. Skipping a number of technical complications, this completes the Pareto criterion. If action A is really so good, it would benefit the winners by so much that they could recompense the losers for the damage done, and still have net benefits.

For example, consider the U.S. Sugar Program. It costs roughly $2 billion per year, mostly to consumers. Hence abolishing it would save consumers $2 billion, but this would cost the sugar industry $1 billion. The winners (sugar consumers) benefit so greatly that they could pay the losers (sugar producers) for the losses and still benefit from the move. By Kaldor-Hicks, abolishing the program increases "social welfare." (The payments to losers need not be made.)

(So why isn't the program abolished, since there's $1 billion per year going unclaimed? This raises some interesting conundrums that I'll skip for now.)

4) A fourth idea: the Lockean standard of individual rights starts with the fundamental axiom that each individual has self-ownership, and derives all other property rights from this. These rights define the individual's proper sphere of action. Each individual has the same equal rights, and this set of rights defines a common set of interests, a.k.a. the common good. Anything that promotes this common good could be said to improve social utility or welfare (although most of us who advocate this 4th standard avoid this terminology).

This standard is compatible with the Pareto criterion, and also with a variant of the Kaldor-Hicks notion: if people actually contract (agree) on some action A, it isn't objectionable so long as it doesn't violate rights of those not party to the contract (but if it does, and they are voluntarily brought into the contracts with side payments, it's not objectionable).

5) There are no doubt many other definitions, but so far as I know they mostly depend on taking a particular set of values or ethics and simply defining it as the social welfare standard. This is problematic -- it is really just asserting that one's own ethical system is objectively true, and the rest demonstrably wrong. However, no objective proofs are forthcoming. Presumably everyone agrees that the one true ethical system should be followed. We also agree that this is (insert the system from one element of the following set here: {422 Christian denominations, 523 sects of Islam, 666 variants of Marxism, etc.}). This isn't to say that there's no such thing as objective morality, but rather that almost no one bothers with this very difficult step, and prefers to just call their own preferred standard the obvious common good. It follows that their personal judgment in the matter should law of the land.

Taking this step, imposing a common ethical standard (other than Lockean rights) from above always ends up sacrificing individuals for the alleged "higher good." If it doesn't, it must be respecting Lockean rights.

Popes, kings, dictators, parliaments, philosopher-kings, commissars, etc. follow. "If we only enforce this, my obviously superior standard, Utopia or Nirvana or God’s Kingdom or the Thousand Year Reich or perfect Communist society or some other wonderful thing will follow." What we actually get is the Inquisition, Reign of Terror, Aushwitz, Gulag, Killing Fields, etc. How else to ensure that people follow the higher good?

The argument for Lockean freedom is quite relevant here. As an economist, I think it's great that when individuals are free to choose between, say, Coke, Pepsi, or none of the above. But there's a far more important matter: the single most important set of choices an individual must make concern the choice of a personal ethical system, and all that this entails. If any of us force our own standard on the individual, we deny him or her something fundamental to his or her humanness: she is no longer a free moral agent. There's no such thing as "good personal character" if some external party is determining your actions for you.

Think of forced conversions, to any religion or political philosophy, and how meaningless they are, and how objectionable. It is the rape of the soul.

The Lockean standard, and only the Lockean standard, respects the individual's free moral agency to reach her or his own conclusions re ethics, God, religion, etc., just so long as she or he doesn't interfere with anyone else's identical rights in doing so.

Some occasionally argue that this is equivalent to saying there's no morality and any choice is equally good, but they are entirely mistaken: it's a standard that says that everyone is personally responsible for their own personal choices, and that this should be respected. It doesn't mean that all choices are equally good, but rather that no other individual can legitimately make personal choices for others. That burden legitimately rests on the individual; the one with the greatest incentive to get it right, the most direct knowledge of the situation she or he faces, and who bears most of the consequences.

In sum, in a social setting, the Lockean (liberal) standard defines the legitimate set of property rights, and the Pareto standard then defines what is legitimate under this endowment. (This is, BTW, why Sen’s paradox of the Paretian Liberal makes no sense to me.)

The question is obviously important, but has some particularly important implications for neoconservative philosophy. The doctrines of political philosopher Leo Strauss constitute one of the fundamental bases of neoconservative thought. I'll discuss this in more detail in subsequent posts, but will say a little here, so as not to leave discerning readers in suspense.

Strauss' view of the "social good" is utterly inimical to Lockean individual rights, and proposes a common good or social welfare that falls into category 5 above. While it purports to defend some sort of "freedom," this is defined as the "freedom to do what is good." In other words, some authority (say, one of Plato's philosopher kings) defines what is good. Everyone is then compelled to follow this. Social good is thus maximized, since everyone is forced to be "good." In practice, the philosopher kings are Straussian neocons, of course.

Sounds nutty, no? Maybe "terrifying" is a better word for it. The only difference between this and a dictatorial theocracy, Nazi or Communist dictatorship, or the Reign of Terror under the Committee for Public Safety is the dispute over what standard will be enforced.

I'll return to this theme in the future; there's much more to be said about the dangerous ideas of neoconservatism. And for all its failures under Bush junior, neoconservatism is not going away... it offers too much to the likes of Rudy, Mitt, Hillary, Barack, Fred, and the rest of that power-hungry rabble for that.

Well, a combined definition of social utility that preserves both economic and individual rights attributes is simply the aggregate rate at which individuals create imrprovements in welfare according to voluntary exchanges possible in the marketplace - and we can reasonably estimate this with GDP which isn't perfect, but about as good a measure as has been proposed.

But the "total value" of social welfare isn't so important as the question of "how much will it change upon the enaction of a certain change in law - which can usually be estimated with market elasticity data. If you assume that people will respond on their own to new rules in the marketplace such as to optimize their own subjective utility functions (a reasonable assumption), then you don't require any objective definition of total social welfare - only elasticities collected from the real marketplace. Furthermore you don't have to worry about restrictions on individual rights since the exchanges are all voluntary.
Thanks for the comment, anonymous.

However, RGDP doesn't measure welfare. Simon Kuznets was upfront on this, and any principles text explains it clearly. RGDP is a measure of output, not welfare. These are entirely different concepts.

You are correct that "total value" for welfare isn't so relevant, just as total utility for an individual is meaningless. But that's an issue apart from the notion of aggregated welfare.

And market data simply do not measure welfare, even within a single market. There are various "willingness to pay" sorts of measures, such as CS, EV, CV, but these necessarily assume that $1 in real terms is a constant measure, and we all know it isn't.

As for all exchanges being voluntary... where do you live? I want to move there.

Thanks again for your comment.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?