Abstract

This dissertation consists of three essays which examine the relationship between the information and asset pricing. The first essay is based on a working paper with Professor Harold Zhang and Professor Yexiao Xu. The second essay is based on a working paper with Professor Harold Zhang and Professor Henry Cao. The last essay is based on a working paper with Professor Yexiao Xu and Professor Harold Zhang.

The first essay applies a matching method to investigate the information content of short sales. Taking into account both the supply and demand effect of short sales, we propose to use the binding ratio, calculated as the short interest ratio divided by the institutional ownership, to measure the negative information of short sales. Our results indicate that the negative information contained in short sales is more pronounced than previously documented. Stocks in the top decile of the binding ratio generate negative abnormal returns as high as -36% per year. However, the underperformance of these stocks only exists in the period before the short interest data is publicly released. Furthermore, we find the main source of the negative information contained in the short sales comes from the earning reports with negative surprises. In addition, we find that the previously documented abnormal return of portfolios constructed using published short interest data is largely due to the differences in liquidity.

The second essay examines the effect of a short-sale constraint on risky asset price in a rational expectations model with asymmetric information. Imposing a short-sale constraint creates two competing effects. On one hand, it reduces the risky asset supply and exerts upward pressure on asset price. On the other hand, it forces investors with negative views on asset payoff to be sidelined. The latter effect can reduce the informational efficiency of asset price, which in turn decreases investors' demand for the risky asset. Consequently, imposing a short-sale constraint can bias equilibrium asset price in either direction depending on which effect dominates. Empirical analysis using short interest and institutional ownership data suggests that an increase in short interest relative to shares outstanding for individual stocks reduces informational efficiency measured by the probability of information-based trading and leads to lower risk adjusted stock returns. The effect of short-sale constraint on return volatility is ambiguous.

The third essay studies this information linkage across US and Mexico markets by examining the Mexico ADR market and investigating the information flow between individual ADR and its corresponding domestic share. First, by applying a VMA-Bivariate-GARCH(1,1) model, we are able to separate the information transmission at market level from that at the firm level, and examine them simultaneously. We find that information transmission is bidirectional at both market level and firm level for majority of companies in our data sample, which is different from other existing literatures. Second, we utilize a statistical method, Markov Regime Switching model, and apply a two-step approach to examine whether the magnitude of information transmission in high volatile state is different from that in other state. We find that the magnitude of information transmission is different at different states for majority of companies in our sample. However, for less frequently traded companies, the difference is not statistically significant.

In summary, this dissertation explores the relationship between information and asset pricing through several different channels, which sheds light on better understanding this subject.