Asset Pricing: (Revised Edition) [Hardcover]

Given the innumerable finance books available, I find myself constantly trying to separate the wheat from the chaff (and, sadly, finding a whole lot more of the latter than the former). John Cochrane's Asset Pricing (2001, Princeton University Press) is not only wheat, but also perhaps the most finely milled flour baked to perfection into one's favorite dessert, served with a chilled glass of Château d'Yquem. Cochrane identifies his target audience as "economics and finance Ph.D. students, advanced MBA students, or professionals with similar background". Residing in the third camp, I can say from this point of view that this book could have been subtitled, "the Practitioner's Portable Ph.D." Academic researchers, students, and practitioners of finance should all value Cochrane's Asset Pricing enough to own a copy.

Asset Pricing is extremely readable, as Cochrane stresses economic intuition over formal proofs. The book is structured into four parts: 1) asset pricing theory; 2) asset pricing models; 3) options and interest rates; 4) an empirical survey. Cochrane begins powerfully, introducing us to the notion that the consumption-based asset pricing equation, given by an investor's first-order conditions, is the central formulation in asset pricing; market-based models simply consider the market returns specified in the consumption models to be exogenously determined free parameters. Cochrane emphasizes that all factor models are derived as specializations of the consumption-based model, using extra variables to proxy marginal utility.

In Part 1, Cochrane covers the field from the Law of One Price, to the mean-variance frontier, to the CCAPM, the CAPM, ICAPM and APT, covering both discrete- and continuous-time, as well as market- and consumption-oriented approaches. Cochrane begins with a simple concept: that price equals discounted payoff, and claims that this is the core of all asset pricing theory. I found this section to neatly clarify my understanding and perspective of these models. Cochrane argues effectively for the use of contingent-claims budget constraints as our lens rather than the traditional mean-variance frontiers and beta models: "...it has seemed that there are several different asset pricing theories: expected return-beta for stocks, yield-curve models for bonds, arbitrage models for options. In fact all three are just cases of p = E(mx)." Cochrane makes clear in his theorems of chapter 4 that the Law of One Price guarantees the existence of a discount factor, and the lack of pure arbitrage implies that the discount factor must be positive. Furthermore, the absence of arbitrage is the result of a positive discount factor, which is the natural result of any sort of utility maximization. Cochrane provides proofs of these relationships for both complete and incomplete markets. I also learned something new (to me) in Chapter 8: in addition to the famous Roll (1977) critique, which states that testing the CAPM using empirical data is impossible because the wealth portfolio is not observable, there is another basic but profound critique due to Hansen and Richard (1987), regarding the conditional versus unconditional CAPMs, which asserts that tests of the CAPM are doomed since the conditioning information of the agents is not observable.

Part 2 introduces us to The Generalized Method of Moments (GMM) approach to free parameter selection, distribution estimation, and model evaluation. GMM is quite powerful and is becoming increasingly popular in empirical studies; one recent example of applied GMM can be found in Andrew Lo's 2002 paper "The Statistics of Sharpe Ratios" (FAJ 58(4)). Cochrane provides the background and methodology for implementing the GMM approach of Hansen and Singleton (1982). Cochrane also covers time-series and cross-sectional (OLS and GLS) regressions for testing linear factor models, with a special emphasis on the Fama-MacBeth (1973) procedure, as well as Maximum Likelihood, which is a special case of GMM, and closes the section with examples of Monte Carlo and bootstrap simulations. Chapter 16, "Which Method?", highlights both Cochrane's pragmatism and masterful intuition of the subject (which is evident throughout the book); I especially enjoyed his brief commentary on statistical philosophy here.

In Part 3, Cochrane covers option pricing and term structure of interest rate models. Two chapters (17 and 18) is hardly enough to do justice to options pricing, which is better served by a complete text such as Cox and Rubinstein's "Options Markets" or Hull's "Options, Futures, and Other Derivatives", but given the limited space, Cochrane does an impressive job, using the Law of One Price to describe put-call parity, arbitrage bounds, early exercise rules for American options, and the Black-Scholes and Feynman-Kac solutions as well as real options. Chapter 19 is devoted to bond pricing. Cochrane covers bond basics, yield curves, and term structure models. The Cox-Ingersoll-Ross (1985) model and the Vasicek (1977) models are shown to be special cases of the affine class of term structure models, and Cochrane derives all three. He also provides a nice review of the literature of both affine and non-affine models, including as Constantinides' 1992 closed-form solution and many others.

Part 4 provides a well-written survey of the empirical work in the field, specifically on time-series predictability, cross-sectional models and equity premium puzzles, and new variations on the consumption-based models. Cochrane also provides an introduction to continuous-time stochastic processes in the Appendix, which succinctly covers Brownian motion, time-series diffusions and Ito's lemma. Most chapters include several problems at the end, a nice addition for readers who really want to dig in and explore asset pricing directly. Although solutions are not provided in the book, Cochrane's website,

http://gsbwww.uchicago.edu/fac/john.cochrane/research/Papers/,

offers them via e-mail to teachers using Asset Pricing as a class text. The website also offers a preview of the book through page 50, which encompasses the Contents, Preface, and chapters 1 and 2 in their entirety. The website also contains an important errata page describing more than 160 equation typos and errors, additions and clarifications to the manuscript.

Cochrane's experience as editor of the Journal of Political Economy shines through in his clear writing style, and his students at Chicago's GSB, where he is Theodore O. Yntema Professor of Finance, are lucky indeed if this book is any indication of his teaching ability. Asset Pricing is not a book to be missed.

It's probably true that the first book you study about a subject inevitably determines your approach to it afterwards. My first book on asset pricing was Duffie's Dynamic Asset Pricing Theory (2nd ed), and it has perhaps forever biased my judgment. Given this caveat, I wanted to like this book. For econometricians, the stochastic discount approach is increasingly important, and Cochrane's articles are engaging and well written. But, no matter what the blurbs on the back cover of the book say, or what some Amazon reviewers claim, this is a flawed book. It's true that "the hurdles of asset pricing are really conceptual rather than mathematical" (last sentence in the book preface), but this is no excuse for being sloppy, and sloppiness in this book abounds. Assumptions are not clear; theorems are imprecisely stated. Continuous-time formulations pop up without explanation of the variables or of the motivation behind them. Expected-utility derivations are the main tool used by the author, but the connection between no-arbitrage, utility maximization and equilibrium are not clear, and one is led to think that the stochastic discount is unique to this line of reasoning. On the positive side, there are many interesting results and many intuitive explanations. My recommendation: i) read Duffie or Pliska first; ii) take the plunge and download Hansen & Richard's 1987 Econometrica paper (very dense); iii) read Cochrane, but reobtain all the results independently from what you have learned in in i) and ii).

This book presents finance in the modern way: p=E(mx). After having read it, the reader should be able to understand the papers currently published in the field. That's the big advantage of the book, because, in this sense, it is better than Duffie's, Dothan's or Ingersoll's. Be advised that the book is not worried about technicallities or math, but the economics underlying the models in it.

However some deep discussions assumes the reader knows: mean-variance frontier, (C)CAPM, APT, and so on, including the several empirical tests already performed on these models and their results. This is not always true, and the reader can easily get lost.

The author uses graphs to clarify the ideas. It is not always successful. Many graphs are confusing. For instance, the author assumes the reader knows how to add and to subtract vectors graphically, which is really easy if you knew that in advance, but difficult to figure out if you do not.

Also there are several minor mistakes the reader should take care of. I am sure the second edition of the book will correct those mistakes and will make the book a lot better.

I think the part talking about the GMM econometrics very clear and that helps a lot to implement the models presented in it.

I recommend the book, mainly because there is no other book treating modern finance like that. Once you get used to it, you'll see the book is not difficult and very useful.