Here's Dani Rodrik on export subsidies:

[T]he economic case for countervailing duties is extremely weak. (The standard economist's line is that you should respond to other countries' export subsidies by sending them a thank-you note, not by shooting yourself in the foot in return.) But presumably there is some (second-best or political) reason why WTO rules sanction countervailing against subsidies.

This paragraph struck me as delightfully odd. Two facts that don't get along are stuck together and left to eye one another warily:

1. Export subsidies are so widely perceived as harmful to recipient nations that the World Trade Organization, a body whose "overriding purpose is to help trade flow as freely as possible", countenances trade barriers to combat them.
2. Standard economics suggests that an export subsidy represents a windfall to recipient nations, an opportunity and a boon, not a harm at all.

What gives? In the wide and wonderful field of economics, is there no room at all for the commonplace observation that a subsidy often does harm to its recipient?

In everyday life, we know that in order to do more good than harm with a subsidy, we often need to make predictions about how the recipient will respond to the grant. Offering to cover an 18-year-old's college tuition is very different than cutting a no-strings-attached check, even if the money's the same either way. Some 18-year-olds would be better off with the cash than a paternalistic tuition grant. But most probably would not be.

This sort of analysis is a priori out of bounds to any economics that views people as rational maximizers. If the best thing a teenager could do with a couple hundred thousand dollars is to turn it into a four-year annuity for college, the utility maximizing teenager will do that. If she chooses to do something else, by revealed preference, that must be the better choice. Right? No.

We reject this kind of reasoning in real-life, even when considering fully competent adults. If we can understand why it is not nice to put a slice of apple pie on a struggling dieter's plate, or why very low introductory "teaser rates" on a home mortgage can entice borrowers into dangerous situations, why can't we understand that certain kinds of subsidies increase the likelihood an economy will trade short-term gains for long-term harms? Of course we do understand that, even the WTO understands it, but economics as a discipline has a remarkably hard time coming to terms with the intuition. Human choice is endogenous and stochastic, not exogenous and rationally determined.

There's an important distinction between noting that certain subsidies increase the likelihood of bad outcomes for a recipient and suggesting that the provider of a subsidy is therefore culpable for the outcomes. It's usually counterproductive to blame someone else's generosity for ones own poor decisions, even if the generosity was cynical and the bad consequences were anticipated by the donor. A crucial feature of subsidy is that it may be refused. A dieter may prefer not to be tempted by the sights and smells of heaven à la mode. But if a host insists on offering, she should still refuse the pie. At the national level, things are more complicated. An export subsidy likely to cause long-term harm to a nation may unambiguously benefit some individuals. How does a nation "refuse the pie"?

By enacting countervailing import tariffs, according the the WTO. But that seems like poor table manners, like raising a middle finger when a simple "No, thank you" would do. If only they could get over the logic of "might as well eat", economists would have little trouble devising polite but firm ways of saying no.

We'd still miss the pie. But maybe that's for the best.