In this recent lecture to the Association of American Geographers, Krugman offers a brief defense of a mathematical model investigation of geography (as opposed to the “discursive” atheoretical style in vogue among non-economist geographers today), and counters the criticism that New Economic Geography is irrelevant to the world today. While NEG generates its main results through transport costs, and does a good job of explaining industrial clusters, the US and Europe do not seem to be defined by industrial clusters as much today as in the past (“Pittsburgh was a steel city, Atlanta today is a…what?”). To explain modern cities, information seems to be the critical element (in addition to consumer cities a la Glaeser). Nonetheless, traditional NEG does do a good job explaining the developing world, such as China; this is unsurprising, since economically these countries look a lot like the US in 1900, which NEG was basically invented to explain.