The Information Content of Insider Trading

Submission note: A thesis submitted in total fulfillment of the requirements for the degree of Doctor of Philosophy to the Department of Economics and Finance, La Trobe Business School, La Trobe University, Victoria, Australia.

This thesis examines information content of insider trading in the post Sarbanes-Oxley Act era. In the first empirical essay, I attempt to solve the benefit-cost puzzle of insider trading activity. I show that net insider purchase is a convex function of quarterly earnings surprise decile. The convexity suggests that the magnitude of net insider purchase is positively associated with information asymmetry, and not necessarily with potential abnormal returns. Further analysis shows that most insider purchases are not designed for earning abnormal returns but for signalling firm quality, and for competing against short sellers when short interest is high. In the second essay, I incorporate insider trading into a jackpot prediction model to improve the model’s accuracy ratio, in order to investigate the relationship between lotterylike return probability and overpricing in distressed stocks. Upon finding Conrad, Kapadia, and Xing’s (2014) results are driven by errors, I document evidence that over-weighting predicted lottery-like return probability in distressed stocks is not a plausible explanation for overpricing in distressed stocks. A Fama-MacBeth regression analysis shows that predicted jackpot return probability positively predicts expected returns, while distress risk negatively predicts expected returns. The third essay explores private information conveyed in a long time period of insider silence. I find that investor inattention to insider silence explains the distress anomaly. In a Fama-Macbeth framework, insider silence negatively predicts expected returns. The inattention effect is the most pronounced in large size and mature stocks, indicating rivalry between salient public information and non-salient private information. Empirical results are consistent with theoretical predictions based on the model developed in studies by Hirshleifer and Teoh (2003), Peng and Xiong (2006), and Hirshleifer, Lim, and Teoh (2011).