Financial Econometrics: Problems, Models and Methods.
The past twenty years have seen a remarkable growth of financial markets, paralleled by an equally remarkable growth of books and articles dealing with the movement of prices and returns. A distinct characteristic of this literature has been a growing sophistication of statistical tools used in the analysis of data. Financial Econometrics represents one of the latest additions to this literature.
With its emphasis on time series analysis, it complements the encyclopedic book on The Econometrics of Financial Markets by J.Y. Campbell, A.W. Lo, and A.C. Mackinlay, also published by Princeton University Press, in which the knowledge of time series analysis is assumed. Both books are designed as graduate textbooks for the emerging field of financial econometrics.
According to the authors, the objective of the book is to report on the current state of scientific advancement in the development of econometric methods in finance. The book is intended for graduate students in statistics, mathematics, economics, and business who are interested in financial applications. It is also supposedly useful for applied researchers employed by banks and financial institutions. The underlying philosophy of the authors is that econometric methods need to be adapted to the problem and data under study on a case-by-case basis.
This is justified by the proposition that statistical models are simplified images of reality and, therefore, are by necessity misspecified. Since specification errors differ from problem to problem, different problems are best served by different models. A crucial task of financial econometrics is to keep the specification errors under control. For this reason the authors do not provide a customary computer disc containing a set of financial series used for empirical illustrations but employ a variety of data sampled at different frequencies.
The text consists of sixteen chapters and almost twenty-four pages of references. The introductory chapter provides a brief discussion of assets and markets, of rudimentary financial theory, and of financial data. The topic of the second chapter is the first-order autoregressive model, AR(1), and its generalization in the form of an autoregressive moving average (ARMA) model of BoxJenkins fame. The third chapter extends the discussion to multivariate time series models represented by vector-autoregressive (VAR) and vector-autoregressive moving average (VARMA) models. The usefulness of these models for efficient portfolios is a part of the discussion.
In Chapter 4 on simultaneity, recursivity, and causal analysis the authors come close to the traditional econometrics, except for a section on the capital asset pricing model (CAPM). Chapter 5 proceeds to the currently popular time series models with unit roots and cointegrated processes.
The topics of Chapter 6, namely autoregressive conditionally heteroskedastic (ARCH) and stochastic volatility models, lie at the heart of the book. This is, in my opinion, the best chapter of the book. It relates the volatility of asset prices to risk management, and it justifiably presents the ARCH models as a scientific breakthrough that allowed researchers to come up with empirical evidence against the presumption of unpredictability of returns.
Chapter 7 contains a review of various prediction schemes such as the adaptive and the rational expectation models. Special attention is given to the fact that there are multiple dynamic equilibria compatible with the capital asset pricing model, including the speculative bubble effect. After that, the authors turn to intertemporal equilibrium models, especially the consumption-based capital asset pricing model (CCAPM). They also describe the associated generalized moments method (GMM) of estimation.
Chapters 9 and 10 deal with dynamic factor models and dynamic qualitative processes and are not particularly inspiring, partly because of the esoteric nature of the topics and partly because of a not very illuminating presentation. Diffusion models, presented in Chapter 11, are used to introduce the reader to the basic concepts of derivative pricing and continuous time modeling. The authors derive the Black-Scholes formula of derivative pricing from the underlying process of geometric Brownian motion. The possibility of jumping prices due to a discontinuous component is also considered. The difficulties of estimating diffusion models due to the fact that prices are not observed continuously are examined in Chapter 12. In general, there are two approaches to estimation in this case. One is based on exact methods such as MLE or GMM, which involve problems of identification. The second approach is based on the method of simulated moments, which involves various approximations.
Chapter 13 is devoted to a detailed presentation of the famous Black-Scholes model of option pricing, which is being criticized for its assumption of time-invariant volatility. In Chapter 14 the authors delve into the area of high-frequency data, believed to be "one of the most promising fields of future research in financial econometrics" because of its potential to offer insights into the dynamics of price patterns that are observed at much shorter intervals of time.
The subject of market indexes, described as approximations to the market portfolio, is taken up in Chapter 15. The discussion covers major U.S. indexes such as those developed by S&P and Dow Jones, as well as major indexes used in the UK, France, Japan, and Canada. The failure of some market indexes to provide good approximations of the market portfolio is duly noted. The final chapter is devoted to the subject of management of extreme risk caused by the existence of fat-tailed distributions of prices. In the presence of such distributions the use of volatility as a measure of risk can be misleading since some of these distributions have infinite variances. As an alternative the authors propose a measure called "value-at-risk," which determines the minimum capital required to cover a financial loss with a fixed probability of occurrence.
Following the standard teaching practice, each chapter starts with an outline of the topic to be discussed and ends with a summary of the foregoing discussion. Both the introductory and especially the summary parts are generally excellent. The middle parts dealing with the subject matter itself are written at a level typical of articles in high-ranking professional journals. The early chapters on basic time series analysis are designed for the novice in the field but progress rapidly to higher levels of sophistication. The subsequent chapters rely strongly on the reader's mathematical and statistical skills by presenting recent results from the literature without shying away from complications.
The book was developed from lectures at various French, Swiss, and Canadian universities in courses which clearly made high demands of the students. The readers who have solid backgrounds in mathematics and statistics and who are willing to work diligently through the presented material will appreciate the comprehensiveness of the content of the book and the frequent useful insights provided by the authors.
|Printer friendly Cite/link Email Feedback|
|Article Type:||Book Review|
|Date:||Apr 1, 2002|
|Previous Article:||Economics at work--the Greater Cleveland Growth Association. (Economics at Work).|
|Next Article:||Principles of Forecasting: A Handbook for Researchers and Practitioners.|