Whether asset returns are predictable is among the most enduring issues in finance. The thesis explores the return predictability through the following two chapters. In Chapter One, I examine the return predictability associated with short interest ratio in the equity market. Specifically, I investigate the relation between short-selling activity and expected returns for stocks with different ratings to assess the implications of financial distress. I find that the predictive power of short interest is concentrated in the worst-rated stocks and the longshort trading strategy generates abnormal profit only for these stocks. The profit is derived primarily from taking the long position in the stocks with the largest decrease in short interest. These firms are more likely to experience a rating upgrade subsequently, and short interest changes predict their future earnings. In Chapter Two, I target the return predictability in the credit market. Specifically, I investigate whether investor sentiment can predict the cross-section of corporate bond returns. To my knowledge, existing proxies for credit market sentiment are all aggregate measures. For the first time, I propose an investor sentiment measure at the bond level and find that it has strong cross-sectional predictive power for corporate bond returns. A long-short portfolio that buys low sentiment bonds and shorts high sentiment bonds generates economically and statistically significant returns. This profitability is robust to various controls. The sentiment index contains rich information for economic fundamentals and perform equally well as existing aggregate measures of Gilchrist and Zakrajsek (2012); Lopez-Salido et al. (2017). x.