Paul Krugman adds a few more thoughts via email related to the recent trade policy discussion (fully covered here):
Paul Krugman: Another thought or two on distribution and trade policy:
The problem of losers from trade isn't new, obviously, either as a fact or concept. But if you look at the history of trade policy - say, in Matt Destler's book it's hard to avoid the sense that the issue has gotten bigger and harder. His final chapters have a definite sense both of nostalgia for the good old days and foreboding.
I'd put it like this: in the old days, when GATT negotiations were mainly with other advanced countries, the groups hurt tended to be highly specific and local - the left-handed widget makers of Northern South Dakota, worried about competition from their counterparts in Upper Lower Swabia. Economists could in good conscience argue that while individual groups were hurt by trade liberalization in their specific sector, the great majority of Americans benefitted from general trade liberalization. And politicians made trade deals by packaging together the interests of exporters, to offset the parochial interests of import-competing industries
But now we're talking about broad swaths of the population hurt by trade. It's a good bet that almost all US workers with a high school degree or less are hurt by Chinese manufactured exports, at least slightly. You could in principle put together win-win packages - say, trade liberalization together with an increase in the EITC paid for with higher taxes on high-income Americans, who come out winners from trade. But the reality is that we don't make those deals.
For those who like their jargon, by the way, I'm basically saying that the right model for thinking about this has gone from many-good specific factors to Heckscher-Ohlin.
I don't have answers to this. The moral case for open markets is their importance to poor countries: America would do OK even in a highly protectionist world, but Bangladesh wouldn't. The domestic politics of trade, however, are now very hard, and getting harder.
Update: Brad DeLong follows up with his solutions to this problem.