We live in a world in which events -- whether they be jumps in the price of Microsoft or blizzards in the Midwest or Super Bowl victories for the Cowboys -- are caused by certain things and not by others. The nature of those causal relationships, though, often remains obscure. We may feel comfortable drawing some conclusion about the market from a rise in Microsoft's stock price, but we would probably feel much less comfortable saying that one thing had caused that rise.
More importantly, even if we can state with some certainty why something happened, that leaves us a long way from being able to state with similar certainty what will happen. We can read the past for portents of the future, but we can never be sure that we're looking at the right evidence, which is just another way of saying that we can never be sure we're looking at the right past. Those disclaimers at the bottom of mutual fund ads are not, in the end, there simply to keep the funds from getting sued. Past performance is no guarantor of future performance, either for the market or for money managers. Things change. Things always change.
The problem, then, is that we want -- and have -- to make decisions about the future, but we do so without perfect knowledge. Peter Bernstein's new book, Against the Gods: The Remarkable Story of Risk (Wiley, 1996), takes on this problem by constructing a kind of history of risk management. Tracing the development of concepts of probability, gambling, and risk from the ancient Greeks through the Enlightenment and into the twentieth century, Bernstein illuminates the desperately complicated relationship between predictability and randomness, and in the process provides a convincing picture of just how tricky successful investing is.
Bernstein's book is at times cavalier in its survey of the past, and the briskness with which he summarizes periods like the Renaissance and the Enlightenment leads him into casual generalizations better suited to a fifth-grade social studies book than to a serious history. At the same time, though, his historical perspective allows us to see the dramatically different ways in which cultures have understood the problem of predicting the future. Probability itself, he argues, was not invented as an idea before the Renaissance, while predictability -- which in turn led to insurance, investing, etc. -- emerged only following the Restoration in Britain. In our own century we've seen a resurgence of critiques of predictability and an increased emphasis on the importance of chance and uncertainty. Even chaos theory, which sees connections everywhere, tends to imply that most of those connections will never be understood.
The way we see the relationship between past and future, then, is neither universal nor necessary. Bernstein suggests, though, that the idea that the future could be managed -- which is to say that risk could be taken into account and insured against -- has been crucial to the growth rate of economies in the industrialized world. The point is not that we know what the future will bring, but that we have a real chance of estimating it, and of making decisions accordingly.
Bernstein's discussion of the roots of probability theory is engaging, as is his explanation of mathematical problems. The best part of Against the Gods, though, is his discussion of the application of theoretical concepts to the world of investing. Given his background as an investment consultant, this comes as no surprise, but the richness of these pages is still striking. Bernstein offers us very few conclusions, preferring instead to complicate our picture of what investing entails. He offers a solid critique of efficient markets theory, and uses game theory to help explain why the safest and most rational investment approaches are also the dullest. His explanation of derivatives is a model of succint and insightful prose, and his discussion of chaos theory goes a long way toward disrupting all of our conventional notions of predictability and repetition.
In a curious way, in fact, Bernstein has written a history of risk management that ends by leaving us more aware than ever of the impossibility of fully managing risk or comprehending the workings of complex systems. There's always something just beyond our grasp, something of which we will be unable to make sense. Risk itself, after all, is the product of uncertainty. That said, some risks are better than others.
How, then, can one know which risks are better? We can take a pretty good stab at predicting, for example, where the Dow Jones Industrial Average is going to be in five years, and we can base that prediction on specific reasons. If we could actually foresee those reasons and those results, we could either make an enormous amount of money or protect ourselves against losing an enormous amount of money. Bernstein quotes a fund manager's thoughts about information as it pertains to investing:
The information you have is not the information you want.
But sometimes the information you have is precisely the information you need, and sometimes the information you can obtain is priced perfectly, and you catch a glimpse of what the market is going to do.
The problem is that you don't know you've caught a glimpse until after it's all over. The hope is that the more information you have, the more work you do, and the better attuned you are to the underlying realities of the businesses in question, the better your chances of prediction. Peter Lynch succeeded in part because he was able to understand the importance of consumer brands to the American economy and because he did the research needed -- including a lot of time spent in malls -- to figure out what companies were going to be able to give the consumers what they wanted. Not all of Lynch's investments paid off, and there was -- it has to be said again -- no guarantee that any of them would. He might have done all the right work and still been wrong. Something different might have happened. He might have had the wrong future in mind.
If that is so, the temptation is to say that everything is simply random, and that the idea of successfully predicting what will happen in the market is preposterous. Whether it is called efficient markets theory or Bachelier's Theorem, the point is the same: the market cannot be beaten because the market does not move in a way that is predictable. If you look at a graph of the S&P 500 over the last seventy years, it's hard to argue with this. The index's movement over that long period has been essentially random on a month-to-month, quarter-to-quarter, and year-to-year basis. There's no way to predict what will happen a year from now on the basis of 1996.
What's strange about this is how counterintuitive it feels, especially to investors in a bull market. When things are booming, it's hard not to feel as if one day is building on the previous day, and as if the market has to be higher a year from now. If, a year from now, the market ends up 20%, it's hard not to feel as if you had known what was going to happen. Incorrect predictions are usually forgotten -- people have the tendency to remember the correct ones much more often.
In any case, there are two important caveats to the idea that the market moves randomly. The first, and most important, is that although the market does seem to move at random, its motion is steadily upward. Over the long term, through all the crashes and booms, the S&P has tracked upward. Very few people invest for seventy years, and the idea of never cashing out and dying at 85 with a pile of money is probably not most people's idea of a rich life -- even if it is Warren Buffett's. Still, staying in the market for the long term has been as close to a sure bet as American capitalism has found.
The second caveat is that the motion of individual stocks seems much less random than the motion of the market as a whole. The market is not efficient when it comes to individual stocks. Information is not perfectly distributed. More importantly, the meaning of that information is not ever self-evident. Interpretation -- like Lynch's insight into the importance of brands -- is a crucial part of understanding a company's future, and skill at interpretation is not evenly distributed. The idea of valuation itself is a tendentious and complicated one -- there is no Bible we can use to determine appropriate P/E ratios. Nevertheless, within the context of a marketplace where industry-wide valuations for successful companies are relatively stable, there are stocks that are under- and overvalued, and finding them with some degree of regularity is possible.
Bernstein's book contains no hints on how to become the next Peter Lynch. But it does contain a fascinating discussion of what he calls "nonrational theory" that consists of investigations into the way people actually act, as opposed to the way rational-choice theorists think they will act. These pages of Bernstein's book are filled with intriguing examples, including a study in which people were inclined to say that they would pay $200 to avoid a one-in-a-thousand chance of immediate death but that they would have to be paid $50,000 to take that same one-in-a-thousand chance. Similarly, people were much more willing to risk money they had already won rather than putting up an equivalent sum, though the risks and rewards were identical.
It's not clear whether there's a way to translate into financial success people's propensity to play with found money, their aversion to risk, and their tendency to identify with their investments. But it is clear that opportunities are created by the nonrationality of the market, and that looking for undervaluation is not a fruitless ta