Cloudonomics: The Business Value of Cloud Computing, + Website

CHAPTER 13

Fallibility of Forecasting

The world descended into financial chaos in the fall of 2008, driven ostensibly by the subprime mortgage crisis and marked by the implosion of Lehman Brothers in mid-September. Dominique Strauss-Kahn, now perhaps better known for his May 2011 Sofitel New York hotel stay but then head of the International Monetary Fund, predicted a recovery beginning in 12 to 18 months, but this forecast, he admitted, was “plagued with uncertainty.”1 To his credit, at least he acknowledged it was basically a guess.

John Kay of the London Business School has argued that economists rarely disagree significantly in their predictions. The salient differences are not among economists but between economists and reality: “[W]hat [economists] say is almost always wrong . . . the consensus forecast failed to predict any of the most important developments in the economy.”2 No wonder author and economist John Kenneth Galbraith quipped that economic forecasting makes “astrology look respectable.”3

There surely are increasingly sophisticated technologies being used for forecasting—exponential smoothing, chaos theory, state-space models, crowdsourcing, and prediction markets, Box-Jenkins autoregressive moving average models—and many of these can be aided by the immense computational ability of the cloud. However, while these models may often apply, fundamental folly of forecasting is the fantasy that the future can be foretold, often leading to catastrophic flame-outs, such ...