Collective Myopia and Habit

We characterize the equilibrium dynamics of a class of linear, incomplete-information models that feature forward-looking behavior, recurring shocks, slow learning, and rich higher-order uncertainty. We present an observational equivalence result that recasts this dynamics as that of representative-agent model featuring two distortions: extra discounting of the future; and anchoring of the current outcome to the past outcome, as in models with habit or adjustment costs. This provides a unified micro-foundation of various bells and whistles that the literature has used in order to match salient features of the macroeconomic time series. Furthermore, the as-if distortions are predicted to be more pronounced at the macro level than at the micro level, helping explain the disconnect between micro and macro estimates of consumption habit. We finally illustrate the quantitative potential of our insights in the context of inflation.