Fabozzi et al., The Basics of Financial Econometrics

Don’t be intimidated by the title. The Basics of Financial Econometrics: Tools, Concepts, and Asset Management Applications by Frank J. Fabozzi, Sergio M. Focardi, Svetlozar T. Rachev, and Bala G. Arshanapalli, with the assistance of Markus Höchstötter (Wiley, 2014) is a remarkably accessible book. Yes, it has its fair share of math, but the math is pretty straightforward—what you would expect to encounter in a rigorous undergraduate statistics course.

In fifteen chapters the authors cover simple and multiple linear regression, building and testing a multiple linear regression model, time series analysis, regression models with categorical variables, quantile regressions, robust regressions, autoregressive moving average models, cointegration, the autoregressive heteroscedasticity model and its variants, factor analysis and principal components analysis, model estimation, model selection, and formulating and implementing investment strategies using financial econometrics. Five appendices review the basics: descriptive statistics, continuous probability distributions commonly used in financial econometrics, inferential statistics, fundamentals of matrix algebra, model selection criterion (AIC and BIC), and robust statistics. All told, the book is a little over 400 pages in length.

What contributes to the accessibility of this book is that it centers on “how to construct asset management strategies using financial econometric tools.” It is at its core a “how to” book. The authors discuss “all aspects of this process, including model risk, limits to the applicability of models, and the economic intuition behind models.” (p. xiv)

Let me focus here on a single modeling problem that I think is generally underappreciated. In physics “data are overabundant and models are not determined through a process of fitting and adaptation.” By contrast, “from the point of view of statistical estimation, financial economic data are always scarce given the complexity of their patterns.” Moreover, since “financial data are the product of human artifacts, it is reasonable to believe that they will not follow the same laws for very long periods of time. … The attention of the modeler has therefore to switch from discovering deterministic paths to determining the time evolution of probability distributions.” But, again, “financial data are too scarce to allow one to make probability estimates with complete certainty. (The exception is the ultra high-frequency intraday data, five seconds or faster trading.)”

The authors conclude that “as a result of the scarcity of financial data, many statistical models, even simple ones, can be compatible with the same data with roughly the same level of statistical confidence.” (pp. 292-93) I don’t know whether to find that result heartening or depressing.

Financial professionals of all stripes will profit from this book. It both explains and challenges—a winning combination.