Volatility forecasting: the role of internet search activity and implied volatility

Recent empirical literature shows that internet search activity is closely associated with volatility prediction in financial and commodity markets. In this study, we search for a benchmark model with available market-based predictors to evaluate the net contribution of internet search activity data in forecasting volatility. We conduct in-sample analysis and out-of-sample forecasting analysis robust to window size in multiple markets for robust model validation. The predictive power of internet search activity data disappears in the financial markets and substantially diminishes in the commodity markets once the model includes implied volatility. A further common component analysis shows that most of the predictive information contained in internet search activity data is also present in implied volatility, while implied volatility has additional predictive information that is not contained in the internet search activity data.