CHAPTER-3 REVIEW OF LITERATURE
Portfolio analysis: example for the Warsaw Stock Exchange
International Advances in Economic Research, May, 2008 by Dorota Maria Kwiatkowski
Department of Econometrics and Statistics, Warsaw Agricultural University, u1. Nowoursynowska 166, 02-787 Warsaw, Poland
H.Markowitz (Journal of Finance, March 1952) created foundations for the portfolio management theory and for the method of an effective selection of assets. Thus, the major problem is to determine the number of securities and their shares in the portfolio value, which depends on (1) the investor's attitude toward risk; (2) the investor's return expectations; and (3) the period of investments. The aim of this research is to construct portfolios under different expectations of the hypothetical investors, i.e., (A) the portfolio with the equal shares of all securities; (B) the portfolio that is efficient in Markowitz sense; (C) the minimum variance (risk) portfolio; and (D) the minimal risk portfolio that is constructed assuming the desired rate of return. The efficiency of the constructed portfolios is evaluated in terms of Trey nor, Sharpe, and Jansen coefficients as well as employing expected and actual rates of return. To determine the minimal number of securities in the portfolio, the risk of 20 portfolios that consist of different number of stocks is evaluated. We conclude that the portfolio containing five elements assures the biggest decrease of risk. Thus, the constructed portfolios include securities of five companies listed at the Warsaw Stock Exchange. These companies composite (among others) two major market indexes (WIG and WIG20), and represent different economic branches, i.e., metals, oil and gas, banking, hotels, and construction, since we expect that their rates of returns are not strongly and positively correlated, which is important in the risk diversification. The investigation is provided for the actual data regarding daily prices in the period 2002-2006.
All constructed portfolios are characterized by positive returns and they guaranteed the risk premium. The expected rates of return are smaller than the actual ones. Portfolio D keeps the first position regarding expected and actual rates of return, but it is characterized by the highest risk. Portfolio C is characterized by the smallest risk but also the smallest return. It is worth mentioning that the expected rate of return of portfolio A is not much smaller than applying more sophisticated methods to the portfolio construction, and the real rate of return places portfolio A on the second position in the portfolios ranking. Analyzing efficiency of the constructed portfolios and employing well-known measures, we notice that Trey nor and Jensen coefficients properly point out the portfolio with the highest efficiency, while the Sharpe coefficient misclassified constructed portfolios. Published online: 30 March 2008, [c] International Atlantic Economic Society 2008
Effect of portfolio weighting on investment performance evaluation: The case of actively managed mutual funds
Journal of Economics and Finance, Spring 2007 by Block, Stanley B, French, dan W
Abstract
Among the factors influencing investment performance measurement is the weight dedicated to each security. This paper develops metrics for measuring the extent of equal weighting and value weighting of a portfolio. A sample of 506 actively managed mutual funds shows that funds tend to be equally weighted to a greater degree than they are value weighted, implying that investment performance based solely on a single value-weighted benchmark may not adequately identify excess performance. We propose a two-factor model utilizing both a value-weighted and an equally weighted index and show that the model provides a better fit than the single-- index model. (JEL GI)
Introduction
Over the decades there has been much debate about the ability of mutual funds to outperform the market when performance is measured with Jensen's (1969) alpha. Early research by Friend, Brown, Herman, and Vickers (1962), Sharpe (1966), and Jensen (1968) indicated that mutual fund managers not only have difficulty beating the market but frequently perform at a level inferior to the market. Although some later studies such as Alexander and Stover (1980), Kon (1983), Chang and Lewellen (1984), and Ippolito (1993) have found results more favorable to funds, the average fund still appears to show no above-normal performance. For example, Volkman (1999) found that while the average fund had no ability to select undervalued stocks and a negative ability to time the market, a few individual funds did display a persistent ability to select undervalued investments. Malkiel (1995) found that survivorship bias is more important than previously realized and concluded that funds have in aggregate underperformed benchmark portfolios even before considering fund expenses. Carhart (1997) controlled for common factors influencing returns and found that they generally explained persistence in performance. Carhart's only unexplained persistence existed in significant underperformance of the worst funds. A strong debate continues over the methodology of measuring and comparing returns. As early as 1970, Friend, Blume, and Crockett warned about using a benchmark that effectively tricks the alpha calculation by over (under) weighting small-firm returns. During the same time period, Carlson (1977) further warned about drawing conclusions that were specific to the time period, type of fund, or choice of benchmark. Later research by Chang and Lewellen (1985), Admati, Bhattacharya, Pfleiderer, and Ross (1986), Lehman and Modest (1987), and Daniel, Grinblatt, Titman, and Werners (1997) stresses the importance of factors such as benchmark selection, survivability, portfolio composition, and non-CAPM return-generating factors when measuring fund performance. There is another factor important to the performance evaluation issue, the weighting of individual securities within the portfolio. The weight that a portfolio manager assigns to a given security in a portfolio can make a contribution to return that is just as important as the security selection and investment timing decisions. Because stock indexes, such as the S&P 500 Index, that are commonly used for performance evaluation are often value weighted (market-cap
weighted), their use as benchmarks for evaluating non-value-weighted portfolios may fail to adequately identify fund performance. Strongin, Petsch, and Sharenow (2000) show that an actively managed portfolio's performance is determined not by the success of its managers' security analysis but rather by high concentration of risk in a value-weighted benchmark. To what extent do portfolio managers tend to equally weight or value-weight their portfolios, and how do their weighting choices affect investment performance evaluation? The intent of this paper is to address these issues. In the next section, we briefly summarize some important work in investment performance evaluation based on the capital asset pricing model. In the third section, we examine two popular mechanical schemes for portfolio weighting: equal weighting and value weighting. The fourth section presents the derivation of metrics to measure the extent to which a portfolio is tilted toward equal or value weighting and then, using a sample of actively managed mutual funds, applies these metrics to show how fund managers typically weight their portfolios. In the fifth section, we analyze weighting's effect on investment performance evaluation by computing various measures of investment performance for the mutual fund sample. The paper concludes with a summary.
Summary
When evaluating the performance of equity mutual funds, it can be important to use multiple indexes in order to consider the effects of both small and large firms on portfolio performance. For example, using an index such as the S&P 500 does not fully capture the "market" because of the omission of many smaller firms from the index. In addition, the S&P 500, like most other popularly used market indexes, is a value-weighted index.
Application of Markowitz portfolio theory in the oil and gas industry Gary L. Cartwright, Devon Energy Corp., Oklahoma City
Oil&GasFinanciaJournalSeptember01, 2008
volume4, issue
Author: Gary Cartwright
Modern portfolio theory is designed to optimize return for a given level of variance across a spectrum of investment opportunities. The ability to monitor and optimize a portfolio gives rise to the ability to manage a portfolio toward certain goals and objectives. The basic premise of portfolio theory is that the variance of returns for a portfolio of risky assets is a function not only of the variance of each individual asset, but also of the covariance [or correlation] between each asset (variance of return, or standard deviation of return, can be considered a measure of economic risk). When multiple risky assets are held within a portfolio, it can be expected that some properties will increase in value while at the same time others will decrease in value. By holding risky assets in groups, some of the risk of each asset may be reduced or eliminated through the process of diversification. Additional potential arises from understanding the relationship among the respective projects relative to success or failure. If two projects are negatively correlated, i.e. if success in one project is associated with failure of the other, and failure in one is associated with success in the other, this significantly reduces the risk of double failure across the two projects. If independent, the full value of diversification is achieved; if negatively correlated, natural hedges reduce the risk of complete loss potentially to zero.
Optimizing portfolios using extensions of Markowitz theory
September 23, 2009 Posted by Geordie in For Developers. A seminal application of OR techniques to finance was by Harry Markowitz (1952, 1987) when he specified portfolio theory as a quadratic programming problem (for a survey of this theory, see Board, Sutcliffe and Zambia, 1999). Participants in financial markets usually wish to construct diversified portfolios because this has the substantial advantage of reducing risk, while leaving expected returns unchanged. The objective function for the portfolio problem is generally specified as minimizing risk for a given level of expected return, or maximizing expected return for a given level of risk. While returns produce a linear objective function, risk is modeled using the variance, leading to an objective function with quadratic variance and covariance terms.
Portfolio performance evaluation: old issues and new insights
M.GrinblatandS.Titman UCLA Anderson Graduate School of Management, Los Angeles, CA 90024, USA AbstractThis article presents a model that provides insights about various measures of portfolio performance. The model explores several criticisms of these measures. These include the problem of identifying an appropriate benchmark portfolio, the possibility of overestimating risk because of market-timing ability, and the failure of informed investors to earn positive risk-adjusted returns because of increasing risk aversion. The article argues that these need not be serious impediments to performance evaluation.
