Krugman's Defense Of Short-Term Protectionism

Many critics of Paul Krugman, including me, have argued that while Krugman may be good on trade, he is horrible on macroeconomics. Now it seems that he is letting his horrible macroeconomics infect even his analysis of trade. In this post, Krugman argues that under "normal circumstances" the standard comparative advantage argument for free trade is the end of the story. But these are not normal circumstances, Krugman argues, and under these circumstances a case exists for short-term protectionism. His argument essentially goes like this:

1) Because other countries will benefit from one country's stimulus package, while not sharing any of the costs in the form of a higher debt, positive externalities for fiscal policy exists, which means that fiscal stimulus will be smaller than it should be.

2) By adopting protectionism, these externalities will be reduced, and so will encourage more stimulus which will boost output everywhere.

This argument has 3 key flaws. First of all, it presumes that the effects of the stimulus really are positive, which is possible in some well designed versions, but is hardly universally true. Secondly, the argument really isn't about the protection per se, but about its indirect effect on the behavior on politicians. The direct effect of protectionism remains negative. And thirdly, it assumes that other countries will "turn the other cheek" and not retaliate. That is a very unrealistic assumption, and a global trade war would have very disrupting effects on global economies. Not just because of loss of comparative advantage effects, but in the short-term (and it is the short-term that the fiscal action is supposed to be beneficial for) also because of adjustment costs as factors of production in many cases can't instantly move from industries focused on exports to industries that sell to buyers that can no longer buy imported goods.