Book Reviews: Predicting Presidential Elections and Other Things & Freakonomics (Revised Ed.)
Prof. Ray Fair, an economist at Yale, originally intended to title his book either as "What Can Econometricians Do?" or "Econometrics Made Easy." As he retells in the Acknowledgments section, his kids talked him out of those titles. That's a shame because those titles are actually better descriptions of what Predicting Presidential Elections is really about: What creative econometrics can reveal about the world told in an accessible manner.
That is not to say that Prof. Fair's book doesn't live up to its title; it does. Most of the first half of the book carefully explains how he came up with an econometric model to explain the results of presidential elections. He then uses this model to predict the results of the 2000 and 2004 elections. (He has subsequently made predictions for presidential elections since then at his website.)
But the book doesn't stop there. Prof. Fair also discusses econometric models of extramairtal affairs, college grades & class attendance, performance in marathons (Prof. Fair is an amateur runner), more traditional economic topics like interest rate & inflation, and a very interesting chapter on the econometric valuation of fine wines. This is econometrics at its best; taking fairly basic tools from econometrics and applying them creatively to explore interesting questions.
What one gets out of Ray Fair's book is a sense that one has really learned econometrics. The book does a good and accessible job of explaining what statistical regression analysis (the cornerstone of econometrics) is as well as t-tests for significance and other fundamental statistical issues.
Another book that explores the creative possibilities of econometrics is Freakonomics, a surprising best-seller. Much has already been written about this book -- most of it pro rather than con -- so I don't have much to add. I will note that the 'Revised Edition' has just been published. It is basically the same as the original except: (a) some parts have been re-written to incorporate some new facts that have been brought to the authors' attention, and (b) some of the more interesting posts from the popular Freakonomics Blog have been included in book form.
I have one mild criticism of Freakonomics in comparison with Ray Fair's book: Unlike Prof. Fair's book, Levitt & Dubner's book doesn't attempt to give the reader a sense of having learned what econometrics is even though that is what is at the heart of the book. There is little or no mention of regression coefficients, statistical significance tests, etc. In fact, there is hardly a number in that book (unless you count the page numbers)!
Perhaps Levitt & Dubner were heeding the apocryphal advice rumoured to have been given to physicist Stephen Hawking that there is some proportional drop in sales with every mathematical equation that is added. But, surely it wouldn't have harmed sales of this best-seller too greatly if just one or two mathematical expressions were thrown in for intellectual credibility's sake? Levitt & Dubner can hardly claim that it is too difficult to write an accessible guide to econometrics ... Ray Fair has proved otherwise with his fine book, which is a kind of Freakonomics with the math (in digestable doses) included.
It's a shame that readers of Freakonomics -- unlike readers of Predicting Presidential Elections and Other Things -- don't get a sense that they've learned at least a bit about econometrics since Steven Levitt is one of of the most (if not the most) creative econometricians around. My advice to readers of Freakonomics is that -- if they want to take the next small, baby step towards understanding the tools that Prof. Levitt used to arrive at his interesting observations about the world -- they read Prof. Fair's book (perhaps along with Philip Hans Franses' A Concise Introduction to Econometrics: An Intuitive Guide) for a relatively easy introduction to what econometricians can do.