Welcome to the new Independent website. We hope you enjoy it and we value your feedback. Please contact us here.

Book Of The Week: Nothing random about shares

A Non-Random Walk Down Wall Street

by Andrew Lo and

Craig MacKinlay

(Princeton University Press)

THERE IS an old joke amongst economists. Two of them are walking along when they see a $100 bill on the ground. "Are you going to pick that up?" asks one. The other replies: "Certainly not! If it were real somebody would already have taken it."

In a nutshell, that is what is wrong with the theory that financial markets are efficient - that share prices already incorporate all known information and only respond to genuine news. This implies that share price movements are random - they follow a "random walk" which means the price change in one time period is unrelated to the previous period. In other words, if there really were a systematic opportunity to make profits, somebody would already have taken it.

This belief is bred in the bone for most conventional economists. The authors of this book tell the story of the first time they presented evidence that share prices do not follow a random walk, and are, on the contrary, predictable. The discussant of their paper, an eminent economist, concluded that they must have made a mistake in the computer programme. He could not accept that the result might be true, because that would imply tremendous untapped profit opportunities in the stock market.

Professors Lo and MacKinlay conclude in this book, a collection of their research into share prices since 1986: "The fact that the random walk hypothesis can be rejected for recent US equity returns suggests the presence of predictable components in the stock market. This opens the door to superior long-term investment returns through disciplined active investment management."

With all its equations, this book is going to turn out to be a classic text in the theory of finance. But it is also one for practitioners. It is a technical study into how share prices depart from a random walk. It looks, for example, at the extent to which there is "long-term memory" in prices. This is also called the "Joseph effect" - seven fat years followed by seven lean years. The authors find there is some memory, but it is much shorter term than the Old Testament cycle. Stock markets are efficient enough, and clear often enough, not to be predictable over longer periods than weeks or months.

Another chapter looks at how to predict share prices. Variables such as the dividend yield, the interest rate trend and risk characteristics have the most explanatory power. This does not mean that it is easy to take advantage of predictability. On the contrary, it requires great technical expertise and a good computer programme, excellent and up-to-date information, and an awareness of dealing costs and taxes.

Most important of all, however, is the role of risk in investment. The classic efficient markets hypothesis effectively ignores the fact that some investors are more willing than others to take risks. But the central insight of financial economics is the existence of a trade-off between risk and return. The fact that risk-takers earn a higher return, even one that could have been predicted with some probability, does not mean that the market is inefficient.

There is a parallel with the market for new products - the fact that some companies earn huge profits from new inventions does not imply industry is uncompetitive. Rather, without the inducement of profit, the invention would not have been made in the first place.

Diane Coyle