| Following the 1987 stock market crash, there has been much discussion in policy circles about the use of price limits in financial markets. Proponents argue that price limits reduce overreactions, while skeptics suggest that it harms the price discovery process. As yet, little is known about the effects of such limits on price dynamics, and the few relevant empirical studies that do exist mostly concern U.S. derivative markets. This thesis offers an empirical examination of the effects of daily price limits on stock returns. It focuses on the Egyptian Stock Exchange (ESE), a developing market that has been regulated since 1997 by a symmetric price limit mechanism. The ESE, unlike other stock markets, has unusually tight limit bands making it more appealing and insightful to study.; Chapter 1 critically surveys studies conducted on price limits and trading halts and offers possible reasons for the inconclusiveness of the findings in existing literature. The ESE's background, market development, market description and stylized facts are outlined in Chapter 2.; Chapter 3 investigates the impact of price limits on volatility using a variety of mean and variance specifications in GARCH-type models (GARCH, EGARCH, GJR, and APARCH) and four different error distributions (Normal, Student- t, GED, and Skewed-t). Results from examining a split sample suggest significant changes in the time-varying volatility process. In-sample results, prior to the imposition of price limits exhibit leptokurtosis, but show no sign of the widely cited leverage effect. In-sample results after the imposition of price limits display both leptokurtosis and the leverage effect. Out-of-sample forecasts exhibit the leverage effects, when present, but provide conflicting results regarding the distribution.; The nonparametric policy-shift event studied in Chapter 4 provides evidence that such limits in the ESE do not reduce volatility and may, in fact, have three deleterious consequences: (1) they can be the source of higher volatility on subsequent trading days, (2) they can delay full incorporation of information into prices, and (3) they can interfere with the trading process.; Chapter 5 raises another compelling problem caused by daily price limits in the ESE. Utilizing a logit model to analyze the morning behavior after large overnight price movements, a statistically significant tendency for stock prices to accelerate towards both upper and lower bounds is documented. Previous research has referred to this phenomenon as the “magnet effect” of daily price limits. The results presented in this chapter are the first empirical verification that such an effect is present for both floor and ceiling hits.; Finally, Chapter 6 concludes by summarizing the findings of this dissertation, identifying the caveats, and outlining an agenda for future research. |
