My (very short) take on this: I do believe that Krugman has a point, a very good one, when claiming that standard textbook analysis is (almost) all you need to understand the current crisis, and to implement the correct policy solutions.
The point is what we define as “textbook analysis”. Krugman refers to IS-LM models. But these, that starting in the 1980s virtually disappeared from graduate curricula because supposedly too simplistic, not grounded on optimization, not intertemporal, and so on and so forth.
I personally was exposed to these ideas in my undergraduate studies in Italy, and I still teach them (besides using them to discuss the crisis with my students). But they were nowhere to be found during my graduate studies at Columbia (certainly not a freshwater school). None of the macro I studied in graduate school (Real Business Cycle models, or their fixed-price variant proposed by New Keynesians) as interesting as it was intellectually, could give me insight on the crisis. I simply do not need to use it.
The IS-LM model with minor amendments (most notably properly accounting for expectations to deal among other things with liquidity traps) remains a powerful tool to understand current phenomena. The problem is that it is not mainstream at all. What bothers me in Krugman’s post is the word “standard”, not “textbook analysis”.
I just published an editorial on the Italian daily il Corriere Della Sera (in Italian), that summarizes my views on the causes of the crisis and of global imbalances. It is a reprise of one of my first posts, written with Jean-Paul Fitoussi. It is useful to summarize and refresh the argument: