When I read Paul Krugman, I first think "that's crazy! He's insane." Then I read it again and that usually (but not always) becomes "that's interesting, but it still wrong." Reading Krugman's February 1st article (spawning a nifty debate here, then here, then here) led me to the typical reaction.
I'm still reading through the responses, but Krugman's basic argument is that fiscal stimulus benefits other countries in addition to the one doing the stimulating (for now, let us assume Keynesian economics works the way Keynesian insist it does). The country being directly stimulated imports from other countries, which help their economy. But these economies share none of the debt generated by the stimulus package. Thus the funding country gets a smaller benefit than they should and we have too few economic recovery acts. Short term protectionism internalizes the benefits to those that are paying for it, thus we get more such packages when we need them.
It's a clear argument but to see why it doesn't work, let's follow the money. Suppose the exporting foreign country is Japan. They now have (say) American dollars. What do they do with this crazy American money? They can't spend it in their country. There are two things they can do with it.
(1) They can import from America. This means that money Keynesians want to stay in the US will come back. It's just adding another step. Instead of Government to Domestic Market, it's Government to Foreign Market to Domestic Market.
(2) They can invest in America. Instead of buying goods, they buy companies (or just capital). Again, the money comes back into the economy, but now there are two steps before it enters domestic consumption (Foreign market to Domestic Company to Domestic Market). Again, it's a wash.
You could argue time is an issue but money moves pretty quickly, especially in the era of Internet and telecommunication satellites. And, because the government's buying international, it's going to get a better deal which will help in the long run (when this huge debt will have to pay off). It's hard to imagine that extra delay is really such a huge problem.
An implicit assumption of Krugman's argument is that exports are inherently good for an economy and imports are not. But that's completely wrong. While Krugman might paint my argument as grounded in theology and not economics, the case for free trade is so strong, and for protectionism so weak, such interpretations are closer to being grounded in laws of economics as irrefutable as gravity or motion.