Predicting stock returns has significant implications for asset allocation, investment performance, and testing market efficiency. To this end, we examine whether U.S. stock returns and volatility can be predicted from a comprehensive set of financial and economic uncertainty indicators as well as migration-related uncertainty measures. We employ the nonparametric causality-in-quantile approach which is robust to misspecification errors since it captures nonlinearities in returns distribution. Our decision to use this approach is motivated by the presence of nonlinearity in our examined series, suggesting that the Granger causality test based on a linear framework is likely to suffer from misspecification. Our findings reveal that aggregate economic policy uncertainty (EPU) together with its different sub-components possess predictive information for U.S. stock returns and volatility barring few cases. In general, the prediction is strongest for returns volatility than for returns. Moreover, we document the ability of the recently developed migration-related EPU and migration fear measures in predicting financial market volatility. Our study therefore, provides evidence that level of aggregate and sub-components of policy uncertainty tends to cause stock market returns, and primarily, volatility.
|Journal||Frontiers in Finance and Economics|
|Publication status||Published - 3 Dec 2017|
- economic policy uncertainty
- stock prices
- nonparametric quantile causality