Predicting time series: Intuitive forecasting


John Beshears, James Choi, David Laibson, and Brigitte Madrian


The goal of this project is to learn how people form expectations: specifically, how they form predictions about the evolution of a time series (e.g., GDP, unemployment, or inflation). Using an incentive-compatible laboratory experiment, we will measure how people use past values of a time series variable to forecast future values of that variable. To the extent that psychological biases influence subjects’ forecasts, their predictions will systematically deviate from rational expectations forecasts. For example, people may anchor on recent shocks and fail to recognize the full extent of long-term mean reversion. Additionally, to the extent that statistical patterns are hard to infer, the predictions subjects generate will contain noise, even if they are not systematically directionally biased.


Beshears J, Choi JJ, Fuster A, Laibson D, Madrian BC (2013). “What Goes Up Must Come Down? Experimental Evidence on Intuitive Forecasting.” American Economic Review Paper and Proceedings. NIHMSID#448056