To Formalize, or Not to Formalize

The holy grail of the economics profession is to be able to write down a closed-form, quantitative, formal model that accurately and precisely explains the evolution of the economy. Even though pretty much everyone recognizes that this is pie-in-the-sky, there remains a bias towards models in which all behavior is explained within the context of the formal model, where only the institutions are exogenous. The result is a divide in the profession between economic historians and a minority of straight economists on the one hand, and the mainstream of the profession on the other hand.

The historians and the minority tend to make narrative and "story" arguments about the evolution of the economy, relying heavily on historical facts and circumstances. They allow for complex patterns of behavior and expectations and naturally incorporate policy changes. While they sometimes make allusion to formal models, they rarely spend much time developing them. They can therefore capture broad trends relatively well, but have some difficulty with detailed forecasts.

The mainstream position, by contrast, develops detailed models that tend to incorporate with certainty (which I take to include stochastic rules) the evolution of behavior and expectations. The resulting models can therefore make detailed and precise quantitative predictions, but the simplistic behavioral rules required to make the models tractable cannot capture economically important behavioral and expectational complexities, stories and paradigms.

The solution, I think, is to recognize that expectations, behavior and policy are difficult to model quanitatively but nonetheless respond to economic as well as social variables. That calls for marrying quantitative formal models with the economic history approach: a good economic explanation should include both a formal model that takes these "soft" variables as exogenous and a narrative that explains how these exogenous variables may evolve with the economy. The narrative and formal model can then feed back on each other with the economist as a guide. To put it another way, the narrative informs structural changes in the formal model, and the results of the formal model feed into the narrative.