Essays in finance: Behavioral explanations of post -earnings announcement drift and the market for Russian government bonds

Existing studies of post-earnings announcement drift use Standardized Unexpected Earnings (SUE) as a measure of earnings surprise. The first chapter asks whether including historical standard deviation of the difference between earnings and their expectations into the definition of SUE requires too much sophistication from an average investor. The main finding is that the initial reaction of stock returns to earnings announcements is closer related to primitive measures of earnings surprise than to the SUE. This tendency is especially strong for stocks owned mainly by individual investors. Consequently, the drift is closer related to the SUE than to the primitive measures of earnings surprise for individually, but not for institutionally owned stocks. This finding explains some cross-sectional variation in drift. I also test the hypothesis of Hong and Stein (1997) that investors' underreaction is related to the slow diffusion of information across investors, and do not find much support for that hypothesis. Finally, I find that trading strategies based on various measures of earnings surprise bring abnormal returns that are positive and stable over time, i.e. relatively riskless. Moreover, the cross-industry analysis of drift shows that it is unrelated to industry betas. This suggests that attempts to explain the post-earnings-announcement drift as a failure to fully adjust for risk are likely to be hopeless.;The second chapter considers two prominent behavioral models that explain the post earnings announcement drift of stock prices, namely Barberis, Shleifer and Vishny (1998) and Daniel, Hirshleifer and Subrahmanyam (1998). In accordance with the BSV predictions, I find that when investors see a trend in earnings they expect this trend to continue in the future. Investors are also persistently more surprised when bad news hits a growth firm than when it hits a value firm. The chapter also looks at stock market analysts' behavior and finds that analysts always expect a trend to be broken in the next quarter.;The third chapter looks at the market for Russian government bonds (GKOs) in 1994–96. I find that the system of primary traders that existed during the whole observation period allowed them to receive average abnormal returns of 1.23% in a day following the auction. This amounted on average to a stream of profits of 8.66 mln US dollars a week. An increase in the number of primary dealers in 1995 reduced abnormal profits, but did not eliminate them. The chapter also presents some evidence (although indirect) that the GKO market was used by the Central Bank as a means of providing implicit subsidies to otherwise insolvent commercial banks.