Stock exchanges

For a stock market to allocate funds efficiently, stock prices should immediately incorporate all of the information available. If we find that there is a lag between changes in variables that might affect the price of stocks, and the reflection of that change in its price, the market for stocks will be inefficient. This thesis tests the stock markets in six of the largest developed economies for informational efficiency. It tests the stock markets in Canada, France, Germany, Japan, The United Kingdom, and The United States, for the existence of a causal relationship between changes in the money supply and changes in stock prices, and applies the Granger-causality test to perform it. A stock market is informationally inefficient if a causal relationship between changes in the money supply and changes in stock prices is found. In this case, money supply changes could be used to predict movements in the prices of stocks, create profitable trading rules, and help us earn above-normal returns, thus casting doubts on the ability of the stock market to allocate funds efficiently.

This study subjects the world's major stock markets to cointegration tests in an effort to respond to the common investigation that the major equity markets tend to move together with the U.S. market. The relationships between the United States and Canada, France, Germany, Japan, and the United Kingdom will be investigated. Not only will the entire period of 1970-1991.3 be examined, but the pre-1987 crash and the post-1987 crash periods as well. In addition, the Chow test will be employed to probe for any structural change occurrence relating to the worldwide stock market crash of October 1987. Evidence of cointegration is found to exist vis-a-vis the U.S. with the U.K. and Germany for the entire period; however, since the crash, only the U.K. and Japan have exhibited equilibrium relations. Absolutely no cointegration was detected for Canada nor France with the U.S. market.

This thesis discusses the implementation of a feed forward NN using time series model to predict the sudden rise or sudden crash of a company's stock prices. The theory behind this prediction system is Pattern recognition. Pattern recognition techniques for time-series prediction are based on structural matching of the current state of the time-series with previously occurring states in historical data for making predictions. This study reports the result of attempts to predict the Motorola stock price index using artificial neural networks (ANN). Daily data from January 1999 to December 2001 were taken from the NYSE. These data are classified based on criteria of an n% fall or rise of price corresponding to the previous day close price. A novel method using Hurst exponent is used in selecting the data set. These data are fed into a Back Propagated Neural Network. The number of hidden layers and number of neurons are systematically selected to implement a better predicting machine. The implemented model is tested using both interpolated and extrapolated data. Fundamental limitations and inherent difficulties when using neural networks for processing of high noise, small sample size signals are also discussed. Results of the prediction are presented and an elaborate discussion is made comparing the results.

The primary objectives of this study are to investigate the stock market over- or underreaction of various U.S. stock market indexes, the over- or underreaction of a global stock market index, and the over- or underreaction of various-countries' national stock markets relative to a "global" over- or underreaction. The secondary objectives are to investigate the reasons for the U.S. over- or underreaction and for the relative under- or overreaction of individual countries. For six U.S. stock market indexes, we find a one-day stock market underreaction to highly positive and negative news releases. Over a sixty-day interval, we find strong evidence of a stock market underreaction (overreaction) to positive (negative) news. Cross-sectionally, we find strong evidence that investors are more optimistic the larger the recent runup in the stock market index is. Also, investors are more optimistic in periods of high economic growth, whether they are faced with positive or negative information. Focusing on the MSCI World Indexes denominated in both local currencies and U.S. dollars, we find that investors underreact to both positive and negative news in both the short- and the long-run. The last objective of this study was to investigate the relative under- or overreaction of nineteen individual countries in response to a global under- or overreaction. We find that several countries exhibit a one-day underreaction relative to the MSCI World Index on the day following a very large positive or negative movement in the MSCI World Index. Over a sixty-day interval, several countries overreact relative to the MSCI World Index when positive information is released on a global basis but underreact to the MSCI World Index when negative information is released on a global basis. Cross-sectionally, results reveal evidence consistent with a hypothesis where investors are more optimistic with respect to both positive and negative news when there is a speculative bubble in the foreign stock market. We also find that investors in countries with high economic growth rates tend to be more optimistic than investors in countries with low economic growth rates when faced with positive and negative global news arrivals.

This study documents the nature of the underlying information that caused investor overreaction and under-reaction. While research has documented the existence of market overreaction and under-reaction, it has not comprehensively addressed the underlying information releases that caused the extreme price fluctuations. This study controls for the underlying announcements, and finds that the degrees of overreaction and under-reaction vary according to the underlying information releases. A primary contribution of this dissertation is the finding that undefined events are associated with higher degrees of overreaction than defined events. The Wall Street Journal Index was used to determine if each event had an announcement that coincided with it. Defined events are those for which an underlying announcement was found in the Wall Street Journal Index. For undefined events, no announcement was found. This finding supports the theory of investor overconfidence and biased self-attribution by Daniel, Hirshleifer, and Subrahmanyam (1998). This study analyzes the overreaction and under-reaction phenomenon in three areas: international securities, domestic securities, and foreign currency. The international securities analyzed are American depository receipts and international closed-end funds. The domestic securities analyzed are financial and non-financial stocks. In the foreign currency area, currencies are classified into two types: emerging country currencies and industrial country currencies. In all of these areas, controlling for the underlying announcements is beneficial in understanding market overreaction and under-reaction. Finally, cross-sectional regression equations are employed to relate post-event returns or exchange rate changes to different variables, such as initial price change, pre-event information leakage, size (market value), month of the year (December or January), day of the week, and announcement type. There is a substantial amount of evidence that suggests larger initial price movements and prevent information leakage are associated with higher degrees of overreaction, and that the tendency towards overreaction is stronger for undefined events.

The purpose of this study is to extend the research on mutual fund performance persistence to closed-end funds. Because closed-end funds trade at market prices different from net asset value (NAV) they are used to test both the performance persistence of their NAV and market price returns. While research has assessed the performance persistence of open-end funds, it has not assessed the performance persistence of closed-end funds. Yet, the unique characteristics of closed-end funds allow stronger arguments for their persistence than the arguments previously submitted for open-end funds. The characteristics that can potentially cause closed-end fund return persistence to be different from open-end fund return persistence include investor sentiment, restrictions on liquidity of underlying assets and cash holdings, and performance pressure on fund managers. This study also investigates the NAV and market price return persistence of international closed-end funds. This group of funds is of particular interest because investor sentiment determines a large portion of the market price return. In addition, as some international markets are less efficient due to restricted flow of information, fund managers may have an advantage in portfolio selection. Finally, this study examines cross-sectionally whether the persistence measure is related to the fund characteristics size, goal, management fees, turnover, fund family, fund experience, and the stock exchange a fund is traded on. The results show evidence for NAV and market price performance persistence, which is stronger for past winning than losing funds. NAV return persistence is less for foreign than domestic funds, possibly due to exchange rate fluctuations. However, market price return persistence is greater for foreign than domestic funds, which may indicate that price pressure is affecting foreign funds more than domestic funds. Funds with lower expense ratios, funds that are not members in a fund family, and funds traded on the NYSE show more persistence of strong NAV and market price performance. The results imply that investors should benefit from investing in past winning funds. This is particularly true for foreign funds, funds with lower expense ratios, funds that are not members of a fund family, and funds traded on the NYSE.