Forecasting Volatility of Oil Prices via Google Trend: LASSO Approach

Abstract

© 2020, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG. Google search volume index has been widely used as a proxy of investor attention. In this study, we use Google search volume index to forecast energy return volatility. In order to find the keywords, we start with glossary of crude oil terms provided by the U.S. Energy Information Administration (EIA) and then add keywords based on Google Search’s suggestions. Then, we arrive at a set of 75 Google keywords as a proxy of investor attention. As there are a large number of keywords to be considered, the conventional method may not be appropriate for the statistical inference. Thus, we propose using the LASSO to deal with this problem. Finally, we also compare the predictive power of LASSO with three types of stepwise method.