Adaptive expectations and reaction to information

Abstract This paper develops a model combining adaptive expectations with noisy signals, and derives three coefficients and one impulse response function (IRF): the Coibion–Gorodnichenko (CG) coefficient capturing consensus under‐reaction to information, the Bordalo–Gennaioli–Ma–Shleifer coefficient capturing individual over‐reaction, the Kohlhas–Walther coefficient capturing extrapolation, and the Angeletos–Huo–Sastry IRF capturing delayed overshooting. There exists a parameter region in which the model reconciles all four moments with the data simultaneously. The model also delivers a testable prediction linking the CG coefficient to variable persistence, distinguishing adaptive expectations from Kalman‐filter updating, and I present supporting evidence for adaptive expectations. The model's fit to survey data is evaluated.