Thanks to a reader (Dartmouth professor) for the link ;-)
The Times notices the same thing (more politely) about academic economists. Not only can they not predict in advance, but they have trouble adjusting after their priors are harshly contradicted by reality. Must be tough to be in a field that selects for overconfidence in dealing with complex systems and messy data.
Ivory Tower Unswayed by Crashing Economy
...That was before last fall’s crash took the economics establishment by surprise. Since then the former Federal Reserve chairman Alan Greenspan has admitted that he was shocked to discover a flaw in the free market model and has even begun talking about temporarily nationalizing some banks.
...Yet prominent economics professors say their academic discipline isn’t shifting nearly as much as some people might think. Free market theory, mathematical models and hostility to government regulation still reign in most economics departments at colleges and universities around the country. True, some new approaches have been explored in recent years, particularly by behavioral economists who argue that human psychology is a crucial element in economic decision making. But the belief that people make rational economic decisions and the market automatically adjusts to respond to them still prevails.
The financial crash happened very quickly while “things in academia change very, very slowly,” said David Card, a leading labor economist at the University of California, Berkeley. During the 1960s, he recalled, nearly all economists believed in what was known as the Phillips curve, which posited that unemployment and inflation were like the two ends of a seesaw: as one went up, the other went down. Then in the 1970s stagflation — high unemployment and high inflation — hit. But it took 10 years before academia let go of the Phillips curve. ...
More (and better) from UC Davis economist Greg Clark in the Atlantic.
Here's something I wrote in 2005 about "expertise" in certain, ahem, soft subjects. See also here, or under the label expert prediction.