Microfoundations: A Decisive Dividing Line between Keynesian and New Classical Macroeconomics?

It is often argued that what marks the difference between Keynesian macroeconomics and new classical macroeconomics (the first installment of dynamic stochastic general equilibrium models) is the presence of microfoundations. These are deemed to be absent in the Keynesian approach, but central to the new classical one. The aim of my paper is to critically discuss this view. Lucas and Sargent defined the microfoundations requirement as consisting of two elements, optimizing behavior and market clearing. I claim that an alternative, weaker, definition is conceivable, which can be traced back to Hayek and Patinkin. According to them, the microfoundations requirement consists of a single criterion, optimizing planning. This definition, I claim, is better than the new classical one. Next, I examine whether Keynesian macroeconomics, which admittedly does not abide by the Lucas-Sargent definition, does accord with the Hayek-Patinkin approach. My conclusion is that Keynes’s General Theory is indeed microfounded in this sense, although no single conclusion can be drawn for Keynesian models in general.