In this EconLog post , my Mercatus Center colleague Scott Sumner challenges the new-found attraction that many Keynesians (including Paul Krugman ) have to protectionism, allegedly as a means of helping to combat economic recessions. Here’s a slightly modified version of a comment  that I left at EconLog on Scott’s post:
Another valid response to these Keynesian protectionists is even simpler – namely, there’s no reason to believe that a change in domestic spending patterns that results in more dollars being spent on imports is more likely to result in less domestic spending than is a change in domestic spending patterns the reach and effects of which are confined exclusively to the domestic economy (that is, a change that does not result in more dollars being spent on imports).
Suppose that Americans start buying more steel, plywood, and grapes from foreigners and buy less of these goods from Americans. It’s possible that these foreigners will sit on (that is, neither spend nor invest) X% of the dollars they earn from their sales to Americans and, thus, cause a reduction in aggregate demand in the U.S. But the very same is true if Americans start buying more steel, plywood, and grapes from west-coast American producers of these goods and buy less of these goods from east-coast American producers. The west-coast American producers of these goods might also sit on X% of their new dollar earnings.
What reason have we to suppose that a change in spending patterns that results in more imports is more likely to cause aggregate demand in the U.S. to fall than is a change in spending patterns that does not result in more imports? None that I can see. (Indeed, if someone wishes to get more full-on hydraulic Keynesian about the matter, it can be asserted that citizens of poorer countries have a higher marginal propensity to consume than do Americans and, therefore, a change in American spending patterns that results in more imports from countries such as China and Mexico will increase aggregate demand in the U.S.!)
The bottom line is that even if (contrary to fact) it were true (as Krugman asserts ) that many of the basic laws of economics somehow stop working when there are unusually large quantities of un- and under-employed resources, a change in spending patterns that results in more imports is no more likely to reduce domestic aggregate demand than are changes in spending patterns whose reach and effects are confined exclusively to the domestic economy.
See also this earlier Cafe Hayek post .