… is from pages 7-8 of James Buchanan‘s essay “Economics and Its Scientific Neighbors” (reprinted as Chapter 1 of Jim’s Moral Science and Moral Order, Vol. 17 of The Collected Works of James M. Buchanan):
Contrast this [the development of game theory] with the Keynesian and post-Keynesian attention on macro-economics and macro-economic models. Does this “theory” provide the economist with an additional set of tools? Does it extend the application of the central principles of the discipline? Unfortunately, the answer must be negative here. Precisely because it has divorced itself from the central proposition relating to human behavior, modern macro-economic theory is really no theory at all. It has evolved, and remains, a set of models for the workings of economic aggregates, models that have little predictive value. Lord Keynes, of course, recognized this, and it was for this reason that he tried to tie his theoretical structure to basic psychological propensities. These propensities, which were designed to replace the more simple neoclassical behavioral propositions, have never fulfilled the role that Keynes must have hoped for them, and the modern model builders seem largely to leave even these out of account.