Data for: Predicting gasoline prices using Michigan survey data
datasetposted on 18.07.2019 by Hamid Baghestani
Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
Abstract of associated article: This study investigates the predictive power of Michigan Surveys of Consumers (MSC) data for gasoline prices. Specifically, we utilize the MSC data on both expected inflation and consumer sentiment to construct a vector autoregressive (VAR) model for forecasting gasoline prices for 2003–2014. Our findings indicate that the VAR forecasts are superior to the comparable benchmark forecasts obtained from a univariate integrated moving average (MA) model in terms of both predictive information content and directional accuracy. As such, we conclude that the MSC data on both expected inflation and consumer sentiment have significant predictive information for gasoline prices. Further inspection reveals that the VAR forecasts are particularly accurate for the period since 2008, reinforcing the notion that consumers are “economically” rational.