Essays on empirical asset pricing

Abstract

This thesis contains two essays on empirical asset pricing. The first essay investigates profitability of multiple trading strategies in foreign exchange market. We find that the proximity of current spot exchange rates to their 52-week extremes explains a significant portion of carry and momentum returns in the foreign exchange market. Anchored carry strategies go long high interest rate currencies that are closest to their 52-week highs and short low interest rate currencies that are closest to their 52-week lows. Anchored momentum strategies go long past winner currencies that are closest to their 52-week highs and short past loser currencies that are closest to their 52-week lows. These "anchored" strategies earn significantly higher returns than their corresponding "residual" strategies—where currencies that are near their 52-week extremes are excluded. These results are robust to simultaneously controlling for the carry spread (i.e. difference between interest rate for long and short portfolio) and the momentum spread (i.e. difference between lagged excess returns for long and short portfolio)—and are thus not driven simply by currencies closer to their 52-week extremes having higher interest differentials or lagged excess returns. We also find that the exposure to various macroeconomic and stock market factors cannot account for the outperformance of the anchored strategies. The second essay studies investor sentiment and the crash risk of anomalies in the stock market. We document that the majority of stock market anomalies exhibit significantly negative skewness following low-sentiment periods and significantly positive skewness following high-sentiment periods. The more negative standardized CVaR of these strategies following low sentiment suggests that crashes are more likely during these times. Thus, left-tail risks cannot account for these strategies' higher returns following high-sentiment periods. In tests of co-exceedances of extreme returns, we find that co-crashes are more likely to occur and more severe following low sentiment, whereas joint euphoric gains are more prevalent following high sentiment. Although diversification across the anomalies enhances the Sharpe ratio, it does not eliminate the crash risk.