One can access Paul Krugman's Nobel Prize speech at http://nobelprize.org/mediaplayer/index.php?=1072. In it he gives a pretty clear description of the new trade and new economic geography approaches, with some interesting discussion of how this fits with the broad history of urbanization in the US. Unsurprisingly he once again fails to cite important predecessors of these ideas, with him basically deserving credit for linking them and doing a good job of publicizing them with his clear models. The two names not mentioned that most deserved to be were Avinash Dixit, co-developer of the Dixit-Stiglitz model that is the key to "Krugman's" theories, with Krugman briefly noting that the theory ultimately came from industrial organization. The other was the first person to apply the Dixit-Stiglitz model to economic geography, who would be Masahisa Fujita, 1988, "A monopolistic competition model of spatial agglomeration: a differentiated product approach," Regional Science and Urban Economics, vol. 18, pp. 87-13124. Krugman is a better writer than the Japanese Fujita, but Fujita has done far more innovative work in this area than Krugman ever did, which I think Krugman knows as he later coauthored with Fujita, even as he did not cite him in his much cited 1991 paper in the JPE that used the same approach as Fujita. Having Dixit and Fujita share the prize with Krugman would have been appropriate and also given the prize to someone from East Asia for the first time. I hope that Krugman finally gets it right for the written version of his speecch and cites the even longer list than this of people who preceded him and deserve recognition for it by him. The model here is Stiglitz, whose reference list for his Nobel Prize speech paper goes on for 13 pages in the AER.
As for his remarks on the auto industry in Detroit, in the end his only explanation is that wages and medical care costs are too high in Detroit compared to the Deep South (no mention of legacy pension costs). Supposedly he was going to explain the problems of the auto industry in Detroit by his theory, which supposedly explains "agglomeration," but he made no reference to his theory other than a vague statement that economies of scale are declining, which supposedly has been going on since about 1965, according to him. However, how or why they have been declining was not explained by him. This rather puts to shame his bragging that he has explained "agglomeration," in contrast to all those pathetic people prior to him, whom he assiduously avoids citing, except for a couple of ancient scribes who used no math, so he can present himself as the great savior who uses math to lead us all to enlightenment regarding these important matters. If Detroit arose because of the factors laid out in his model, he does not say how this happened nor how they stopped holding so that Detroit is now doomed. Blaming high wage and medical care costs amounts to nothing more than de facto union bashing with no link to any version of his model discernible at all. A pretty pathetic performance all in all, especially after he went after Brian Arthur some years ago in Slate for supposedly overselling his role in describing increasing returns, which took Kenneth Arrow to come in and defend Arthur, noting that he, unlike Krugman, actually cited his appropriate predecessors.