Hysteresis and path-dependence are properties not just of real-world economies, but also of our theories about those economies. Ways of talking and thinking live long beyond the time when anyone remembers why (or even that) we once decided this was a good way to talk and think. The daily business of normal science, at least in economics, is largely about local improvement, refinement, and reformulation. Consequently, starting points matter a great deal. No full understanding of the evolution of macroeconomics since World War II is possible without understanding where we started, and why we started there.

We started of course with the neoclassical Synthesis. Every economist knows that the Synthesis was an historical compromise (some would say an unholy marriage) between the Classical and Keynesian views of the world. The Hicksian IS/LM model became the ground on which the postwar intellectual battles were fought Because the model seemed to suggest that the difference between the Classical and Keynesian views was about slopes of curves and speeds of adjustment, the battles were fought with statistics, and as a consequence macroeconomics became much more empirical than it had been previously.

It is less often remembered that the Synthesis was also a bridge between the old institutionalist economics and the new more formal styles that soon came to dominate. The formulation of the Synthesis as a set of simultaneous equations was seen by some as the mathematical object most appropriate for the institutionalist vision of the economy.