Bad News for News Shocks? The Expectational-Stability and Comovement Tradeoff in a News-Driven Business Cycle Model

This paper investigates whether adaptive learning and news shocks, operating in tandem, can generate business cycle dynamics that match the data. In doing so, this paper adopts expectational stability (“E-stability”) as a natural criterion for rationality: plausible equilibria should arise from an adaptive learning formulation where agents form forecasts from a correctly specified model whose parameters are updated in real time. In examining the model’s stability properties, I find that the rational expectations equilibrium (REE) is not learnable for calibrated parameter values capable of generating news-driven recessions. In particular, I uncover a tradeoff between E-stability, or learnability, and the model’s ability to produce realistic business cycles: REE associated with parameter values necessary for generating recessions in response to news shocks are not learnable, and parameter values that imply learnable REE cannot generate empirically plausible recessions in response to news shocks. This finding thus has implications on the plausibility of rational expectations solutions in business cycle models driven by news.