تئوری پورتفولیو و تحلیل عملکرد

Here is a very readable and comprehensive book, produced by practitioner/academics Amenc and Le Sourd (Misys Asset Management Systems/Edhec). It is well suited as an MBA level investments course text, and the material is presented in a very accessible fashion for the practitioner. CFAs will especially appreciate this book as it references AIMR standards throughout (my rough count indicates more than 50 references to AIMR throughout the book, mainly in chapters 1, 2, 7 and 8). Francophiles may also especially appreciate the international flavor of many of the discussions.Chapter 1, “The Portfolio Management Environment”, defines portfolio management and describes how it works in practice. Asset classes, including alternative investments, are first introduced. The distinction between passive and active management is reviewed, and then the investment management process is described in terms of strategic asset allocation, tactical asset allocation and security selection. Performance analysis is then covered with a nice discussion of market efficiency and an excellent review of the performance persistence literature.Chapter 2 focuses on performance analysis, and shows how to calculate both absolute and relative returns. Here, I found the introduction to the “Portfolio Opportunity Distributions” (POD) as developed by Surz to be fascinating, and along similar lines as the recent research by Kritzman and Page (“Asset Selection Versus Security Selection”, 2003) which addresses one of the commonly misunderstood aspects of the Brinson et al 1986 and 1991 model, which decomposes historical performance and concludes that more than 90% of the variation in portfolio returns over time is related to asset allocation. As Kritzman and Page correctly note, “The Brinson et al studies and others like them present a joint test of investor behavior and capital market opportunities. They do not answer the question, which activity is more important: asset allocation or security selection, which is what we propose to do [by measuring the potential for dispersion].” In chapter 7 Amenc and Le Sourd also provide an interesting summary of one researcher’s finding that 98% of subsequent writers on the topic have either misinterpreted or erroneously quoted the Brinson studies. The advantage of the POD approach is that it does not suffer from being a joint test of investor behavior and capital market opportunities, and as Amenc and Le Sourd describe, offers a plausible and useful way to potentially distinguish skill from luck in a manager’s performance. Risk measurement is covered next, including introductory level discussions of VaR, extreme value theory and scenario analysis. The appendices to chapter 2 provide methods for calculating returns and descriptions of global market indices.Chapter 3 covers pre-CAPM MPT (the Markowitz 1952 model), with nice discussions of the critical line method, Wolf’s 1959 simplex method, Sharpe’s 1963 single-index model, and the Elton, Gruber and Padburg 1977 simplified method that employs the Treynor ratio. A useful appendix presents Merton’s 1972 Lagrangian resolution of the efficient frontier. Chapter 4 covers CAPM pretty thoroughly, though Treynor 1962 is inaccurately referred to as a 1961 paper (as most of the literature also erroneously reports), Fama’s 1968 clarification is ignored, and while Amenc and Le Sourd discuss Lintner’s first exposition from February 1965 in the Review of Economics and Statistics, they inexplicably ignore the best Lintner paper (in my opinion), Lintner’s second paper of December 1965 in the Journal of Finance. A nice comparison of the Sharpe, Treynor and Jensen measures is provided here, and more recent concepts of tracking error, information ratio, Sortino ratio, M-squared, Morningstar’s ranking system, VaR and style analysis are also discussed. A very nice section on the analysis of timing is provided, covering the models of Treynor-Mazuy 1966, Henrickson and Merton 1981, Henrickson 1984, and Grinblatt and Titman’s 1989 decomposition of the Jensen measure. I was surprised to find Roll’s 1977 criticism of tests of the CAPM to be presented in this book as a criticism of the CAPM itself. I suspect that Amenc and Le Sourd were lulled by Roll’s introductory remarks, which refer to his argument as a “broad indictment of one of the three fundamental paradigms of modern finance”; I would strongly urge that a careful reading of Roll indicates he is defending CAPM by critiquing the tests rather than the model – for example, on pp. 142-3 Roll 1977 examines the 1972 Black, Jensen and Scholes empirical analysis as follows: “Black, Jensen and Scholes rejected the Sharpe-Lintner theory…[however] …we are entitled to be suspicious of their conclusion. Unless Black, Jensen and Scholes were successful in choosing [the true] market portfolio, their results are fully compatible with the Sharpe-Lintner model.”Chapter 5 examines performance measurement, first presenting ARCH, GARCH and ARMA models, and then presents conditional CAPM (including discussions of conditional beta and conditional alpha). Here again we find the Jensen measure, the Treynor and Mazuy model and the Henrickson and Merton model applied to performance measurement. Also valuable here is a nice presentation of non-market model dependent methods of performance analysis, including the Cornell 1979 measure and the Grinblatt and Titman 1989 and 1993 measures.We find the APT model of Ross 1976 and Roll & Ross 1980 in Chapter 6. Fama-MacBeth’s 1973 procedure is discussed, as well as the models of Fama and French 1993, Carhart 1997, and the BARRA model that eventually grew out of Rosenberg and McKibben 1973. Factor modeling is discussed, with overviews of maximum likelihood and principal components approaches presented. Next is a useful section on applying factor models to portfolio risk analysis, covering commercially available models from BARRA, Quantal and Advanced Portfolio Technology. I would like to have seen a discussion of the Northfield model as well. Sharpe’s 1992 style analysis model is presented, as well as Roll’s 1997 approach. An appendix derives the arbitrage valuation relationship.Chapter 7 covers the portfolio construction process utilizing the approaches of Markowitz 1959, Treynor and Black 1973, Black and Litterman 1991 and 1992, Scherer 2002 as well as others. Amenc and Le Sourd is one of the few good descriptions of Fama’s 1972 selectivity model, though they fail (as, I believe, have all else) to note the identity relationship of Fama’s [Gross] Selectivity measure with Treynor and Black’s 1973 Appraisal Premium metric. Perhaps had Amenc and Le Sourd presented their Figure 7.2, illustrating Fama’s decomposition, to more closely reflect Fama’s Figure 2 this identity would have been apparent to the authors. But Fama illustrated a case in which Net Selectivity is negative, whereas Amenc and Le Sourd illustrate a case in which Net Selectivity is positive, and they do not depict [Gross] Selectivity at all. Brinson’s 1986 model is presented along with a nice discussion of the interaction term. Multiperiod and international aspects of performance attribution are covered well, and Engström’s 2001 and Grinblatt & Titman’s 1989 replicating portfolio techniques are also discussed briefly.The text covers fixed income securities in the final chapter. Yield curve analysis, portfolio construction, and performance analysis for bond portfolios are covered, with introductory discussions of various fixed income models.Amenc and Le Sourd include a bibliography at the end of each chapter, and I found this very useful. A colleague of mine bought the hardcopy, and after reviewing it I decided to purchase the e-book version, which I am enjoying very much. The authors have produced a well-researched and readable exposition that most practitioners would find to be a handy reference, and students would be well served with as a text.