Tuesday 11 August 2009

NZIER on internalities

The latest NZIER Insight tackles the issue of internalities. Economists traditionally count only external costs as policy relevant. Internality theory argues that self-control problems lead people to consume more than they would want to. So, if the problem drinker sincerely wishes he could drink less than he currently drinks, the costs to him of his excess drinking get to count as a policy-relevant cost because government could, in theory, make him better off by encouraging him to reduce his consumption.

If we take internality theory seriously, it's not a call for higher taxes on goods consumed by folks with self control problems. Rather, it's an argument for ensuring that sufficient self-control mechanisms are available. Fortunately, many such mechanisms are available. Other economics blogs have already discussed at length private self-binding mechanisms. If that's insufficient, in the case of alcohol, Pharmac already heavily subsidizes disulfiram. For much less than the cost of a bottle of beer per day, you can buy a pill that will make you violently ill if you consume alcohol.

Given that these mechanisms exist, aren't secret, and, in the case of disulfiram, are already heavily subsidized by the government, I just can't buy internalities as policy relevant for alcohol. It's schizophrenic. Internality theory requires that the person suffering the excess cost really wants to be able to cut back; if we then have evidence that plenty of available mechanisms for cutting back are ignored, I call phooey on counting internalities.

Some folks might well want to go the step further and argue "self-control all the way down" - that the same self-control problems that stop the drinker from going cold turkey also mean that the remorseful drinker always promises to take the disulfiram tomorrow. Well, that really seems observationally equivalent to that the consumer just enjoys his drink but knows that he can reduce social disapprobation by whining about self-control problems. One of the best things about economics, done as economics, is that we tend to ignore the lies people tell about why they do things and instead look to preference as being revealed by action.

Unfortunately, Zuccollo at NZIER muddies the waters a bit in using internalities to justify BERL's shonky analysis of the costs of alcohol use. Yes, if you buy the internalities argument, then there will be some internal costs that get to count as policy relevant. But it isn't all internal costs! Rather, it's just the excess costs that are incurred because of the internality problem: the area above the marginal benefit curve and below the marginal cost curve to the right of the intersection of the two curves. Instead, BERL counted all internal costs as social costs under an assumption of zero benefits to harmful consumers. By contrast, our analysis looking only at external costs makes more sense if "internalities" from excess consumption roughly balance internal benefits (consumer surplus) from prior consumption.

Zuccollo insists that consideration of internalities isn't just disguised paternalism. Yes, in theory, if you buy the theory, the problem drinker welcomes the imposed tax. But in practice it's impossible for the government to implement such a scheme non-paternalistically. As Glen Whitman puts it,
In short, the old paternalism said, “We know what’s best for you, and we’ll make you do it.” The new paternalism says, “You know what’s best for you, and we’ll make you do it.”
Zuccollo notes the problems of implementation given that governments haven't access to our utility functions, saying
Perhaps this is the main reason why internality taxes are not yet gaining widespread popularity in government circles.
Perhaps also botched and shonky applications of theories that remain out of the mainstream canon of economics oughtn't form the basis of a government commissioned economic cost analysis. And perhaps other economists ought to keep an eye out that internality taxes not gain more credence than they're due. Read Whitman, linked above.


  1. TGGP: whatever homunculi float around in the brain fighting with each other, eventually their decision is revealed as an action. If that action is refusing the self-control mechanisms that would have asserted the dominance of some homunculus over others, why should we think that the defeated homunculus should be our preferred spokesman for that individual's utility function?