Content area
Full Text
I. Introduction
According to the efficient market hypothesis (EMH), traders should not be able to make abnormal positive returns using publicly available information. Several studies test it by examining stock price changes immediately following their large price declines. [3] Atkins and Dyl (1990) find market overreaction in daily winners and losers but conclude that such reversals are not significant after controlling for the bid-ask price bounce and transaction costs. In contrast, [5] Bremer and Sweeney (1991) note that for 1962-1986, Fortune 500 stocks that have a large one-day price drop tend to experience a significant abnormal price reversal over the next three days in violation of the weak form EMH.
[6] Cox and Peterson (1994) study NYSE (large capitalization), AMEX, and National Market System stocks (small cap stocks traded on the NASDAQ) for the 1963-June 1991 time period and note the post-price drop three-day reversal. Further, they note that the reversals are greater for small cap than large cap stocks and that the reversal disappears even for the small cap stocks by the October 1987 stock price crash. They also find no evidence that stocks with greater event-day price decline have greater reversal and conclude that their data do not support the [7] DeBondt and Thaler (1985) overreaction hypothesis. They attribute the pre-October 1987 reversals to bid-ask price bounce and illiquidity effects and argue that these effects were more pronounced for small cap stocks and disappeared as the market became more liquid by the time of the October 1987 crash. However, it seems unlikely that the October 1987 stock market crash should coincide with a rise in market liquidity and their conclusion that the small cap reversal disappeared in the post-October 1987 period may be due to their short post-October 1987 period.
[6] Cox and Peterson (1994) also note that the event-day price decline inexplicably, and in apparent violation of the EMH, resumes in the period starting the fourth day after the event and ending 20 days after the event (denoted hereafter as [+4,+20]). Finally, they show, as further evidence against the overreaction hypothesis, that the event stock's performance in the [+4,+20] period is worse; the greater is its event-day price drop. [6] Cox and Peterson (1994) have not been updated to see if the...