The sum of all fears: Forecasting international returns using option-implied risk measures

Abstract

This paper investigates international index return predictability using daily-updated option-implied information in predictive regressions and out-of-sample forecasts. We document the joint predictive power of a variance risk premium (VRP) proxy (defined as risk-neutral variance minus realized variance), Generalized Riskiness (GR), and higher-order moments for forecast horizons of one day to one year. These four risk metrics, which capture cumulative market “fears,” perform well in the US and in an international sample of countries. The VRP proxy and GR are significant and complementary predictors for several horizons, including under one month (VRP proxy) and longer horizons (GR). Risk-neutral skewness and kurtosis are significant for several countries across multiple horizons. Utility gain calculations confirm the economic significance of these risk-neutral variables across countries.