If B, then B

This takes question begging to a whole new level. Usually people just implicitly assume their conclusions. Here Krugman does it explicitly. He starts with the assumption that P is fixed. Then examines what would happen under a given policy and finds that P doesn't change. Wow. Very impressive.
Update: See this Sumner post on this as well. Really hoists Krugman by his own petard. You might say that Krugman is petarded.