Why I Picked On Welfare Economics

As I understand it, economics is a science of efficiency. Now, efficiency is pretty meaningless if we don't first specify what we're interested in the efficient production of. In some contexts, efficiency will measured in terms of the net amount of some plain commodity, such as apples or loveseats or iPods. But it is also possible to adopt a higher-order perspective using some other metric that somehow incorporates other particular commodities. One branch of economics, welfare economics, purports to investigate the efficient production of human well-being. Since the consumptions of apples, loveseats, and iPods presumably all contribute to human well-being, welfare economics is a candidate framework for understanding the nature and requirements of efficiency from that higher-order perspective.

Yesterday via Twitter I claimed (in part) that "sometimes welfare economics is just silly." I caught some flack for this, and rightly so. Since welfare economics is a broad tent encompassing a diverse array of techniques and fundamental commitments, it is probably unwise to generalize about welfare economics as such (although I did say "sometimes"). Moreover, since each version of welfare economics presumably wears its commitments and assumptions on its sleeve, it is always open to the economist who practices it to say, "Hey, don't blame me. I'm just telling you what falls out of the algorithm given the axioms and the assumptions." Finally, since welfare economics, in all its versions, is an account of efficiency only---welfare economics as such entails nothing about public policy, unless the only consideration relevant to public policy is efficiency. Some economists seem to operate as if they believe this (I give an example below). But there is no reason why one should take this as an essential feature of welfare economics.

So perhaps what I should have said is that sometimes welfare economists are just silly. Let me give you an example.

One version of welfare economics contains the following (non-exhaustive) axioms:

1. The relevant notion of efficiency is efficiency in the production of human welfare.

2. Human welfare is defined as the satisfaction of subjective preferences.

3. The amount of welfare one derives from the satisfaction of a given preference is measured by the amount one is willing to pay to have it satisfied. 

It is one thing to use these axioms to determine, for this or that reason, the efficiency of this or that policy. Once again, that would be a "Hey, that's just what falls out of the algorithm" enterprise. The economist who does this certainly need not endorse the implementation of the policy deemed most efficient. S/he may not even believe that human welfare is best defined as the satisfaction of subjective preferences. Still, the model may be useful for this or that purpose, and there it is, wearing its axioms on its sleeve. But some economists do seem to go further, and this can lead to ways of thinking that I do think are misguided (if not "silly"). Here is health economist Mark Pauly in his recent short book on health care reform (pdf):

If anything, among the middle class there is more overuse [of medical care]…than underuse…[But for] lower-income households in America (given moderate risk levels) and for high-risk households (given moderate income), this reasonable state of affairs does not apply. Generous subsidies for better care are needed for them. Fortunately, subsidized insurance can both provide financial protection and reverse moral hazard for community benefit, stimulating the use of care that others value more than the person can or does. (p. 7; emphasis added)

Pauly seems to imply that those who cannot afford decent care ipso facto cannot value it either. In such situations, the best reason to provide it to them is that others in their community would benefit if the poor get health care, since these others care about the poor's ability to get care and are willing to put their money where their cares are. (This is sometimes referred to as a "caring externality"). In my view, this way of thinking completely distorts sensible debates about what we owe to others and why. Of course individuals who can't afford insurance value it. Further, if no one happened to care about low-income folks' access to health care, that would entail nothing about whether or not they are entitled to it. And yet Pauly's remarks are precisely what one would expect from someone who embraced the axioms set out above as partially defining (one sphere of) political morality. When I tweeted that welfare economics was sometimes "silly," I was thinking of this sort of thing. But this sort of thing is not the fault of welfare economics per se, but rather of some economists problematically turning a group of useful techniques into a normative framework.  

One final example of welfare economists behaving badly comes from certain welfare economic analyses of climate change. A central issue in that debate is whether or not the value (or disvalue) of future events should be discounted just by virtue of occurring in the future. One group in this debate says "No" for the simple reason that it is hard to justify saying that a death today is morally worse than a death in 5, 10, 50, or 100 years.* Another group in this debate says "Yes," for the following reason: if we do not discount the badness of future deaths, then we would be required to make very significant sacrifices today to prevent deaths that will occur in the very distant future. This second group believes that the only way to release current generations from the unreasonably punishing duty to prevent indefinitely many future deaths is to say that those future deaths count for less simply because they are in the future. If they count for less, then they have less claim on our attention.

This defense of discounting is a case of economists behaving badly because the punishing duty would emerge only from a framework that made efficiency the sole normative criterion for public policy. If one agrees that efficiency is just one among other important moral values, it will not follow that the punishing duty is a "logical implication" (as one group of economists puts it) of the rejection of discounting  (see this short paper for citations.) The argument in favor discounting does not follow from the science of efficiency as such. It follows only if one transforms that science into a normative framework all on its own. And that, I believe, is tremendously problematic.

So in the end, I recant my claim that "sometimes welfare economics is just silly." But I retain the view that sometimes welfare economists make seriously problematic logical leaps that lead them to unjustified policy frameworks and prescriptions. 

*As Derek Parfit once noted, at a 5% discount rate, one statistical death next year counts for more than a billion deaths in 400 years.

Comments

[Bill Gardner]

Well done, Paul! As you note, "silly" is the wrong way to characterize the problem. Silly suggests that they have missed something obvious. If it were so, there would be little need for moral philosophers.

[Brendan Saloner]

I agree with much of what you say here, but I think it may be unfair to beat up on the use of the word "value" in welfare economics. In the Pauly quotation he is referring to the willingness of low-income households to pay out-of-pocket for certain medical services, and surely we can agree that for whatever reason these households purchase fewer services with their limited incomes than is societally optimal. That does not mean that the reason for providing subsidized services has to be to get closer to the social optimum (although it would be fortunate if it were pareto optimal to offer a subsidy to low-income households on the margin). It also does not mean that these households care less about being healthy than people with more money (in fact, we know that people will consume more beneficial services if they are given a subsidy or more income).

[Paul Kelleher]

Thanks for the comment, B. My emphasis is indeed on "value," but it is most strongly on Pauly's use of "can"--I can make no sense out of that word choice unless he's drawing on a welfare econ framework that would also help explain his reliance on "community benefit" to justify helping those who, by virtue of their circumstances, "cannot" value health care. For example, he makes no indication that the people he's referring to cannot value health insurance because they are weak-willed, or uneducated, or the like.

Thoughts on that reading of Pauly?