کالیبراسیون تجربی یادگیری سازگار

Adaptive learning introduces persistence in the evolution of agents’ beliefs over time, helping explain why economies present sluggish adjustments towards equilibrium. The pace of this learning process is directly determined by the gain parameter. We document and evaluate gain calibrations for a broad range of model specifications with macroeconomic data, also developing alternative approaches to the endogenous determination of time-varying gains in real-time. Our key findings are that learning gains are higher for inflation than for output growth and interest rates, and that calibrations to match survey forecasts are lower than those derived according to forecasting performance, suggesting some degree of bounded rationality in the speed with which agents update their beliefs.