• PAGE ONE
• INDEX
• HOME

Handling a portfolio

Efficient Asset Management

By Richard O. Michaud

Publishers: Harvard Business School Press, US

*Price: $35.

MINIMISING risks to investments and maximising returns on them are two key principles of asset allocation in general and stock portfolio management in particular. In the risk-ridden stock market, efficient asset management requires investing institutions to keep keen watch on the share price movements and deviations in earnings, define the investment policy based on expected returns and deviations (risk), and continuously improve investment value by optimally allocating capital and structuring the portf olio of assets. This branch of financial management has its several specialised techniques and practices evolved through many decades of study and hard-earned experiences of institutional investors and financial economists.

The most popular portfolio structuring technique is Harry Markowitzs mean-variance (MV) efficiency. It is the classic paradigm of modern finance for efficiently allocating capital among risky assets. Markowitz gave the classic definition of portfolio opt imisation. A portfolio is efficient if it has the highest expected (mean) return for a given level of risk (variance) or, equivalently, least risk for a given level of expected return, of all portfolios from a given universe of securities. However, MV ef ficiency has its limitations of instability and ambiguity, because a small input error can result in large errors in optimised portfolio. So, investing institutions have to look for other techniques too to further improve the value of MV optimised portfo lio.

The book under review is a practical guide to stock portfolio optimisation and asset allocation. It discusses the other statistically based techniques. Its contents are: Introduction, Classic Mean-Variance Optimisation, Traditional Criticisms and Alterna tives, Understanding Mean-Variance, Portfolio Review and Mean-Variance Efficiency, Portfolio Analysis and the Resampled Efficient Frontier, Portfolio Revision and Confidence Regions, Input Estimation and Stein Estimators, Benchmark Active Asset Allocatio n, Investment Policy and Economic Liabilities, Return Forecasts and Mixed Estimation, and Avoiding Optimisation Errors.

The first chapter introduces the MV optimisation and its limitations, while the second chapter describes the essential technical issues that characterise MV optimisation and portfolio efficiency.

Chapter 3 carries other authors criticisms of MV efficiency and descriptions of many other alternatives that can be classified in five categories, namely: non-variance risk measures, utility function optimisation, multi-period objectives, Monte Carlo fin ancial planning, and linear programming. It carries also the alternatives own serious limitations. The author avers that none of these alternatives addresses the basic limitations of MV optimisation.

In the fourth chapter the author details the limitations of MV efficiency as an investment management tool, going deeper into its instability and ambiguity features and argues effectively that MV optimisation requires modification and constraints, fallin g in line with Frost-Savarino efficient frontier estimation.

The following eight chapters discuss several statistical approaches for enhancing MV optimisation. While each one of those approaches can help to improve the investment value of optimised portfolios, together they can make a substantial impact on the opt imisation process. This precisely is the focus of the book.

In Chapter 5 the author advises the asset manager to draw statistical inference whether a portfolio needs revision or change in the existing structure before doing a restructuring. This is because those portfolios that are already efficient should be lef t undisturbed in the safe frontier while they are kept under watch. For drawing inference the MV efficiency has to be run through statistical tests. Testing model is given in the book. Other software simulation models are available for buying.

Chapter 6 introduces new tools for portfolio analysis and revision. The most important of them is frontier, which is a computable and practical alternative investment strategy. This, depending upon the quality of the data used, may often be the criterion of choice for defining optimal portfolios in practice. The next chapter is an extension where the author introduces some additional tools for efficiency analysis and revision at the portfolio level and continues the discussion on the statistical charact eristics of resampled efficiency. In these two chapters he explains the procedures using tables, exhibits and appendices. The techniques discussed would immensely help practicing asset managers improve their efficiency.

Improved input estimation is the focus of the next chapter. In all the procedures, the quality of input is paramount. There are many well-known models of estimators. They are listed and discussed here with their advantages and disadvantages. The most wid ely known one is Stein estimator. It is capable of minimising the errors in MV efficiency. Because of its wide acceptance, the author has discussed it in great detail.

The next chapter defines priors and benchmarks of returns. Benchmark optimisation procedure is an MV optimisation that includes a benchmark return for each asset. There are three different benchmark optimisation frameworks commonly used. They are: benchm ark-relative or index-relative, implied return, and economic liability-relative. These are discussed in some detail with illustrations.

The author discusses how to define investment policy and handle economic liability. MV optimisations based on long-term historic return data appear to be the most obvious way to define an optimal long-term asset allocation. For the management of cash flo w and liability risk of corporate investors and fund trustees in relation to economic liabilities towards provident fund, pension fund, gratuity, and the like, continual estimations and re-examinations are necessary. The discussions cover actuarial appro aches and several liability models for handing these requirements.

In the succeeding chapter the author discusses return forecasts with mixed estimation. In the illustrations he mixes forecast returns and uncertainty with historic data. Different formulae and methods for forecasting are mentioned and discussed with appr opriate illustrations.

The last chapter deals with implementation errors. A number of sources of input errors and the techniques to avoid them are discussed. For capturing and enhancing the investment value of optimised portfolios avoiding implementation errors, the asset mana gers should carefully consider investment theories and intuition, investor objectives, forecast return biases, and optimiser behaviour.

It is a well-written book on an important subject of perpetual interest to investors. Brevity of text and the use of many exhibits, tables, appendices, and formulae enhance the quality of the book. However, adding a glossary of terms would help the reade rs who are not well versed in the subject.

P.K. Joy