Over at marginalrevolution.com, Tyler Cowen posted a particularly good entry on famous economists’ famous errors. The average quality of comments on that site is usually higher than on most economics sites, due, in no small part to Tyler’s and Alex’s fairmindedness and perpetual introduction of interesting issues. But this time I thought the average quality of comments was even higher than usual.

I thought I would add a comment on one of the most famous and widely-cited articles on health economics, the 1963 classic by Kenneth Arrow, “Uncertainty and the Welfare Economics of Medical Care.” I seemed to remember a whopper in there about how when you have imperfect information, you necessarily need some kind of government regulation. So I reread the whole article this morning and guess what? I couldn’t find the whopper. Arrow’s article was much more nuanced than I had remembered. So I almost decided not to write this post. But then I thought, “Wait a minute. The main purpose of Econlog is to disseminate good economics. So what if I can’t find errors in Arrow. Isn’t it worthwhile to cite good work also?” I had almost been defeated by selection bias.

His article is well worth reading. In it, you’ll see his proofs that if an insurance company has loading fees (costs besides the payout of benefits), the optimal insurance policy will have a deductible and that if an insurance company is risk averse, it will have some degree of co-insurance in its insurance policies. He also has a beautiful treatment of moral hazard in health insurance, not just in the narrow sense of the incentive not to take care of oneself but also in the wider sense of overspending on medical care.

Arrow also points out that optimality in health insurance requires that higher-risk people be charged higher premiums.

And throughout he writes beautifully.

Arrow also takes seriously Milton Friedman’s proposal, made in his 1962 classic, Capitalism and Freedom, for certification of doctors rather than licensing. Strangely, Arrow doesn’t mention Friedman in this context although he does mention Friedman’s colleague, Reuben Kessel, who also opposed licensing.