Albert Hirschman is one of the most interesting economists of the 20th century, and not just because of his excellent last name. He was influential in development economics (unbalanced growth, linkages), organizational theory (“Exit, Voice and Loyalty”), and the the history of economic thought (“The Passions and the Interests”). A bit closer to my interests, Hirschman contributed an interesting essay to one of the rarest of academic species: the important edited volume, in this case, The Political Power of Economic Ideas: Keynesianism across Nations. Hirschman’s chapter argues that Keynesian economics may have been less influential in the post-WWII US than expected (given its policy dominance 1938-1945, in the “second New Deal” and WWII itself) because Keynesian economists redirected their attention to the world stage, leading influential missions to Japan, Germany, and eventually almost everywhere. Thus, Keynesianism came to dominate the world, even as it never fully took root in the US and faced lots of challenges.

One of Hirschman’s best arguments concerns the connections between macroeconomic policy making and “surprise”. Hirschman notes the arguments of other scholars in the edited volume that public understanding of a policy is possibly one predictor of its policy success – to implement new policies, you have to convince the public of how and why they work, and overturn existing understandings of the old policy. But, Hirschman argues that this can create a tension, especially when macroeconomic policies emphasize expectations as a source of problems in the economy:

To work at all, a policy must first be at least minimally understood, but it becomes unsustainable if it is understood too well, in the sense that the operators will neutralize it by anticipating its effects. In other words, the public’s understanding of the policy must be neither inadequate nor excessive, but since the understanding presumably passes from the former of these negative conditions to the latter, the viability of any policy is necessarily limited in time. Recent experience suggests that this is not an entirely unrealistic interpretation of macroeconomic policy making in a decentralized economy. (Hirschman 1989: 355)

I think part of the “rational expectations” move in recent economic theorizing is directly trying to address this problem: to come up with a macroeconomic policy that “works” even when fully understood and predicted by the relevant actors whose behavior you are trying to predict and control. But, these policies may still make strong assumptions about the knowledge of economic actors (in this case, assuming a maximal understanding which may not obtain). So it’s still a fascinating problem, and I recommend Hirschman’s take on it for anyone interested.