Forecasting Gasoline Prices Using Consumer Surveys

Abstract

This paper analyzes the predictive power of a new data set of consumer gasoline price forecasts taken from the Michigan Survey of Consumers (MSC). MSC data generally perform as well as a no-change forecast in predicting future gasoline prices, and they substantially out-perform the no-change forecast during the recent economic crisis, during which time they track futures market prices. Finally, the cross-respondent dispersion of the MSC forecasts increases substantially during the economic crisis, paralleling the large increase in price volatility at this time.