Three Essays on Empirical Asset Pricing

This thesis consists of three chapters. In Chapter 1, I show that returns to the currency carry and momentum strategies are compensations for the risk of US monetary policy uncertainty (MPU). The exposures to the MPU risk explain 90% and 98% of cross-sectional variations in average excess returns of currency carry and momentum portfolios. The betas to the MPU shocks are lower for currencies with higher interest rate or realized returns, leading to a negative and significant price of risk. The information in the US MPU is not subsumed by other types of uncertainties, and controlling for exposures to the MPU risk renders insignificant role of global FX volatility on explaining the carry trade. These new findings can be rationalized by an incomplete market model featuring the financial intermediary with limited risk-bearing capacity and the US interest rate risk or ambiguity. In Chapter 2, I document novel source of time variations in the cross-sectional inflation risk premium. A consumption-based asset pricing model with inflation non-neutrality and ambiguity shows that the investor's fear of inflation model misspecification ties up the inflation beta and ambiguity beta of individual stocks. As a result, the inflation ambiguity premium amplifies or counteracts the cross-sectional inflation risk premium, whose net effect largely depends on the co-movement of inflation and inflation ambiguity, named as nominal-ambiguity correlation (NAC). Empirically, positive NAC at the current quarter predicts in the following quarter a loss of quarterly return of -4.88% (-2.87%) for a zero-investment high-minus-low value-weighted (equal-weighted) portfolio, obtained by sorting on all individual stocks based on their exposures to inflation risk. The time-varying NAC also explains well the dynamics of inflation premium at the industry-level. The ambiguity channel differs from the existing resolutions both theoretically and empirically. In Chapter 3, I provide new empirical evidence on the risk-sharing between FX and stock market. The exchange rate reflects the cross-country differentials of stochastic discount factor. By exploiting the information in the currency returns, I show that the estimated volatility of stochastic discount factor from FX market significantly predicts the aggregate US stock market excess returns, both in- and out-of-sample. The predictability dominates most of the existing predictors. The FX-specific volatility of stochastic discount factor also strongly drives the time-varying risk-return relation in the US stock market.