Arbitrage as defined by the authors of the book seeks to take advantage of known price discrepancies typically occurring either in two different locations or at two different points in time. In the latter situation, the arbitrageur often knows the price in the future with some reasonable certainty. The book goes into great detail discussing how Warren Buffett has taken advantage of time-based arbitrage to make outsized profits. It first carefully guides the reader through the basics of arbitrage as it is purportedly practiced by Mr. Buffett, and then devotes many chapters to various spins on the basic idea of time-based arbitrage.
The authors present to the reader a basic arbitrage strategy that ignores transaction costs and taxes. Granted, trading costs have come down tremendously, but unless you are an investor of considerable means, these costs are still somewhat prohibitive. Moreover, taxes reduce the expected return from arbitrage operations, over and above the reduction in return due to the time element and the possibility that the deal under consideration either isn't as certain as one originally thought or just doesn't pan out. Finally, for the overwhelming majority of arbitrage situations, which by the way are very well followed on The Street (thus reducing considerably the possible gain), one still needs to deploy massive amounts of capital in order to make outsized profits- something that still confers to Mr. Buffett a considerable advantage, given the many billions at his disposal.
According to the authors, the two critical variables in any time-based arbitrage situation are the certainty of the deal and the time element. As for the latter, they failed to carefully elaborate on how the time element can greatly reduce expected returns; specifically, the longer one has to wait for the deal to go through, the lower the expected return from the arbitrage operation. The authors generally hide behind the stature of Buffett and Graham to present to the reader a method that basically boils down to an annualized net-benefit analysis, and they exercised great care to hedge their approach by saying that arbitrage operations are more art than science. I believe that they should have disclosed the very real fact that in most publicly accessible arbitrage situations, some investors (almost always the bigger fish at the venerable names on The Street) know more about the deals and know it sooner than others; furthermore, this latter point explains why expected gains have tended to decline over the years. Indeed, they also should have disclosed that there are often a few other variables, some of which may even defy quantification, that tend to rear their heads in such deals, and that each deal almost always has its own unique wrinkles.
Over the course of reading the book, I recognized a few problems, some run-of-the-mill, and some potentially mortal, with the methods outlined in the book. First off, the retail investor could be wrong on the timing of the deal. Granted, this is always an ever-present risk, and is somewhat mundane. Next, you could be wrong on your certainty that the proposed deal will actually be completed. As for this potentially mortal investment risk, many little-guy investors seek to hedge it by taking the temperature of the market with regard to the deal. Also, there may not be enough of these deals of a sufficient quality to be worthwhile. One has to figure that some of these deals probably won't work out, and one has to have a large capital base to make the numbers work after factoring in the likelihood of one or more deals in (investment) play not going as planned. Moreover, for the little guy, as mentioned before, the arbitrage operations have taxation dimensions and transaction costs that together can reduce expected returns to mediocrity. This is no small point, as arbitrage raises activity level and costs but may or may not pay off in additional return.
Additionally, the authors do not touch upon opportunity costs and ways to adjust for it. Finally, the authors look only at the arbitrage transaction, and not, for example, life after the transaction. Here, perhaps a personal anecdote from my own experience can shed light on what I mean. A few years back, Unilever announced its intention to purchase Alberto-Culver (you know, the Alberto VO 5 people). At the time, I happened to have good `ol `Berto in my taxable account and Unilever in my IRA. I immediately recognized this as a better buy-and-hold situation, as the union of the two made for an even better company (Berto paid a rising dividend and sported a clean balance sheet with ample cash and no debt; Unilever had global reach and economies of scale working for it; the two together would have greater earnings combined right off the bat, and production synergies as well as mutual scale economies would tend to magnify earnings over time, and Unilever's reach would extend `Berto's brand in more places). Naturally, the market looked favorably upon the deal, so the share price of `Berto made an immediate and dramatic rise to within a few points of the deal price. So, instead of buying more of `Berto, I bought more of Unilever (as an aside, I initially hated Unilever, as I had planned on milking `Berto for steadily rising dividend checks for years to come). Ultimately, the deal closed, and I made out on my position in Alberto Culver; however, I made out even better, and still make out quite well, with my position in Unilever. Arbitrage is great for one-off gains, but one can't regularly feed off those like fat, rising dividend checks over time.
In sum, the book provides investors with significant sums of money (say, in the 250K range plus) an interesting possibility for making some decent money. However, in the end, the method relies upon the law of averages and engaging in many deals for its overall success, which I hasten to add, is not guaranteed. This book is a quick read on the topic, and potential readers would do well to use it more to learn about the (bare) basics of arbitrage and very carefully consider putting its methods into potentially profitable practice.