Very thoughtful and worth reading, if—as he says—a bit rambling: Fabio Ghironi: Ramblings on Tractability in Macroeconomics: "Tractability is in building models that make it possible to shine a bright light on the mechanisms the model-builder wants to focus on...

...At times, this can be done with models that are simple enough to be solved completely with pencil and paper. At times it cannot. And yet, if sufficient time is spent removing unnecessary ingredients and getting one’s hands dirty with the grease of the machine, even a model for which the computer is ultimately necessary can be tractable, in the sense of delivering results that can be understood and explained.... My most successful work so far has been my 2005 QJE paper with Marc Melitz... the real exchange rate—equation (5) in our paper—that made it possible for us to understand fully the implications of our model for the Harrod-Balassa-Samuelson evidence and to explain them clearly to our audiences and readers. Without that effort, forget about QJE publication and the success the paper had—even if computing tools made it feasible to spit out impulse responses and tables of numbers. Tractability is also tied to the purpose for which I am writing a model, and the model itself is tied to the purpose for which it is written. If what I want to do is to show how one specific mechanism works and how it can, say, contribute to shaping fluctuations, I will try to strip out of the model everything that is not directly relevant to that mechanism, even if that means omitting a lot of stuff that is empirically relevant for other purposes...

Also cf.: Beatrice Cherrier: How ‘tractability’ has shaped economic knowledge: a few conjectures What is the cost of ‘tractable’ economic models?


#shouldread

Comments