The Fortune Sellers: The Big Business of Selling and Buying Predictions. - book reviews

Business Economics, Oct, 1998 by R.M. Monaco

By William A. Sherden, New York, John Wiley & Sons, 1998, pp. 308, $29.95.

The Fortune Sellers is a sharply worded, often sarcastic attack on forecasting that also at times is entertaining and informative. William Sherden has created a comprehensive reference for cost-cutting managers, doubtful clients, and those who simply want to take a poke at "experts." If you're a forecaster, the book will help prepare you for arguments you're likely to face. But forecasters aren't treated very sympathetically. Sherden wants to show that virtually all forecasting effort is wasted, and at times it's even dangerous. In general, he presents a strong but, in the end, incomplete case.

The book has nine chapters, with an introduction and conclusion sandwiching seven chapters that examine forecasting in different subject areas. Sherden examines the subjects in a progression of "hard science" to complete "pseudo science." He tackles weather forecasting, economic forecasting, stock market forecasting, population forecasting, technology forecasting, social forecasting, and corporate gurus. The order indicates where forecasting is most legitimate. Weather forecasting places first because it is based on physical laws and because forecasts have improved over time, showing "scientific" progress. None of the other fields, according to the author, can claim either.

In later chapters, Sherden widens his sights. In the first four chapters, he mostly evaluates explicit numerical forecasts: where the hurricane hit, how fast GDP grew, how much the stock market rose, whether birth rates rose or fell. In the last three chapters, he evaluates more qualitative predictions. In the chapter on social forecasting, Sherden illustrates the harm caused by bad "predictions" by looking at the consequences of ideas such as Manifest Destiny, the Domino Theory (with respect to Southeast Asia in the 1960s), and the inevitability of communism. The same chapter examines the records of trendspotters such as John Naisbitt and Faith Popcorn, who, Sherden argues, are mostly, if not entirely, wrong. The last chapter debunks theories of management science gurus like Tom Peters (In Search of Excellence) and Bob Waterman. In most cases, Sherden dismisses these management scientists as either providing commonsense or complete nonsense.

The book continually returns to three ideas, First, all forecasters, regardless of field, are only about as accurate as an appropriately chosen "naive" forecasting system. For example, in economic forecasting, Sherden suggests that an appropriate benchmark might be forecasting next quarter's GDP growth by using the previous quarter's GDP growth. The book itself was born out of an evaluation that Sherden conducted of inflation forecasts, where he concluded that nobody outperformed the simple rule that tomorrow's inflation would be equal to today's inflation. The evidence that Sherden cites in his first four chapters is quite good, and, in general, fairly convincing.

Sherden's second theme follows from the first. Because forecasters aren't more accurate than a naive system, forecasters are charlatans-albeit sometimes unwitting charlatans-who are mostly wasting their clients' money. His clearest examples appear in the chapter on stock-market forecasting, where he points out that, although market gurus don't outperform the market, they do extract fees. Correct predictions are mostly the result of luck. He examines Elaine Garzarelli's predictions, showing that, after her one spectacularly correct call of the market crash of 1987, she has performed worse than a system based on coin flipping.

Sherden's third idea explains why forecasters are bound to fail. He suggests that it's largely because all interesting systems are "complex, nonlinear, and chaotic," and therefore "unforecastable." These are the buzzwords of the relatively new field of complexity, whose spiritual home is the Santa Fe Institute. Sherden's argument relies heavily on the complexity theorists' idea that systems with interacting parts generate unexpected system-wide outcomes, and that these effects are nonlinear. A quantitatively small cause can, and will, lead to unexpectedly large consequences. He cites Adam Smith's Invisible Hand (in which a society of individuals interested in only increasing their own welfare interact to produce increased welfare for all) as an example of a complex system generating unexpected results.

In his conclusion, Sherden advises users to consider five qualities that people can use to discriminate good forecasters from charlatans. Good forecasts are based on "hard" science, done with "sound" methods, by forecasters with "credible" credentials and "proved track records." In addition, he advises forecast users to make sure that they don't accept forecasts simply because they like a particular outcome, i.e., users need to be aware of their own biases.

The Fortune Sellers presents a strong but not airtight case against forecasters. Some of the claims are wrong and some of the arguments don't hold water. Consider these two examples, drawn from the chapter on economic forecasting.