Essays in Empirical Asset Pricing

Abstract

In the first essay, we estimate liquidity-driven trading volume, denoted as inside volume, based on the joint daily reversal pattern of volume and price. With this measure, we find that firms experience a shock to idiosyncratic volatility display an increase in liquidity provision. Furthermore, the under-performance of high-idiosyncratic volatility firms is limited to those with high inside-volume. The relation between idiosyncratic volatility and liquidity provision is prominent in both over- and under-priced stocks. The results suggest that the idiosyncratic volatility puzzle is largely driven by liquidity provision. In the second essay, I document a discontinuity at zero in the conditional distribution of hedge fund quarterly returns following underperformance and outflow. I propose a dynamic measure, conditional kink (CK), to quantify this quick loss recovery and investigate its underlying mechanism. Contrary to the managerial skill hypothesis, hedge funds with higher CK underperform in the subsequent year. Furthermore, this underperformance only pertains to funds with low governance, suggesting that some fund managers may engage in ill-motivated activities to recovery. The flow-performance relationship indicates that investors do not recognize this adverse behavior, thus highlighting the importance of internal control and information monitoring in the hedge fund industry. In the third essay, we demonstrate a “reinforcement effect” between past returns and media-measured sentiment across several asset markets – liquid individual stocks, international equity markets, and currencies. Reinforcement states when past returns and media sentiment agree signify overreaction, leading to return reversion. Reversion disappears in non-reinforcing states. The effect is driven by the unrelated shocks, rather than correlated comovement, in past returns and sentiment. Sentiment’s return-forecasting strength comes primarily through greater negative return autocorrelation in reinforcement states. The effect is stronger in more liquid assets and using local news outlets. Buying disparaged losers while selling praised winners earns several percent annually. In the fourth essay, I demonstrate the importance of inter-firm political links, measured by common campaign contributions made by firm executives. Price movements of a firm’s stock are predictable based on stock price movements of connected firms. Cross-predictability is strongest among politically connected firms that operate in different states and sectors, suggesting that inter-firm political links are largely overlooked due to limited investor attention. To probe the underlying mechanism, I present evidence suggesting that common sources of political exposure across firms ex-ante cannot alone explain this relationship; instead, political ties play the key role, further synchronizing ex-post political agendas. Using the 2010 Citizens United v. FEC decision as an exogenous shock, I find that cross-predictability is weaker for firms that are restricted from actively engaging in political campaigns. Long-short stock portfolios based on political ties yield risk-adjusted returns of 4%-5% per annum.