Google searches and Gasoline prices

Abstract

We study the relation between gasoline prices and Google searches for the term “gasoline prices”. Utilizing the framework of a vector autoregressive model and Granger causality, we find a two-way relationship between gasoline prices and Google searches. In both directions, the relation is negative, i.e. an increase in one of the variables leads to a decrease in the other variable. Moreover, we find the strongest evidence for causality from Google searches to gasoline prices.