Essays in Empirical Asset Pricing

Abstract

Return predictability has always been an interesting topic and discussed on the academic front. In this thesis, we first study the correlation between return predictability and firm level liquidity. We find that the illiquid firm commands higher excess return. Next, we study the return forecasting ability of the short interest ratio. We find that Rapach et al. (2016) conclusions have certain limitations, and the forecasting ability cannot last more than one year. In addition, we confirm that highly shorted stocks perform slightly worse than the stocks with smaller short positions. Finally, we find that information flow impacts a company’s liquidity volatility and trading volume after the SEC introduces the EDGAR system.

Chapter 2 contributes to the related literature by studying the cross-sectional relation between Amihud (2002) illiquidity measure and expected stock returns using the latest stock data and studying the effect of Regulation Fair Disclosure (Reg FD) implementation on stock-level illiquidity. This chapter focuses on the most commonly used measure, the Amihud (2002). We find that illiquidity has a solid positive cross-sectional relation with future stock returns, which is that illiquid securities command higher expected returns than more liquid securities. Regardless of whether illiquidity is measured using one, three, six, or twelve months of historical data. Reg FD’s implementation positively impacts the firms’ liquidity, especially for small firms. After the implementation, the policy mandates the small firms to establish a sound system and reduce selective disclosure, resulting in a liquidity improvement.

In chapter 3, we study the predictability of the aggregate short interest from the econo metrics and economics views by employing new detrending methods and the tests for dynamic predictive regressions to avoid the unit root. From the econometrics angle, the short interest index (SII) that Rapach et al. (2016) build is not stationary and has a unit root. There is an upward trend in this series. Thus the analysis results obtained by incorporating this series are spurious and not reliable. From the economic angle, the highly shorted stocks consistently underperform the lightly shorted stocks. The lightly shorted portfolio, in which the SIR is less than 2.5% comprises approximate 80% of the listed firms over the market. We doubt the predictive power of SIR could be driven by a small number of firms reporting high short interest. The extremely high SIR value will magnify the predictive power of the aggregated short interest because the number of firms with high SIR constitutes a small percent over the entire market.

In chapter 4, we exploit the quasi-natural experiment created by the roll-out of the EDGAR system to study the causal impact of the additional flow of stock-specific information on firms. We find that this information flow to investors results in statistically significant and economically essential changes in illiquidity and trading volume but not in idiosyncratic volatility. Across firms, illiquidity falls for the smallest firms more than it does for the largest firms. Across industry groups, the mining and manufacturing sectors have the largest decreases in illiquidity and increases in trading volume.