Testing the Forecasting Prowess of Bitcoin Uncertainty in the Predictability of Stock Returns

Abstract

Motivated by the theoretical prediction of a weak link between the cryptocurrency market and stock markets, most empirical literature contends that Bitcoin is a safe hedge for the stock market. Opposing this position is the view that portrays Bitcoin as a regular, high-risk asset, such that when stock prices rise, so does Bitcoin, and vice versa. To test the validity or otherwise of this competing view, we construct a bivariate predictive model to examine the predictive power of Bitcoin uncertainty on stock returns. In contrast to the extant literature, we rely on the novel measure of uncertainty (i.e., the Bitcoin uncertainty indices, hereinafter URCY) and specify a forecasting model. The objectives of this study are to examine (i) the predictive power of UCRY indices on stock returns and (ii) the extent to which UCRY can make accurate out-of-sample forecasts. Using data for the G7 countries, among other things, our findings show that Bitcoin uncertainty indices are negative predictors of stock returns across the countries under investigation. Results also reveal that the indices are accurate and reliable predictors of stock returns in the short-to-medium term. These results are robust to accounting for structural breaks (i.e., the COVID-19 pandemic) and some macroeconomic variables. The policy implications of these results are discussed.