From a recent Barrons:

Reading the entrails of animals may not be an infallible method of divining the future. And it holds inherent and obvious drawbacks for the fastidious. But, for all that it’s undeniably primitive and ancient, as a guide to coming events the entrails approach is at least as reliable as most of the contemporary tools of prophecy, and the shamans who employed it were as often on target as are any of the present-day soothsayers and their vaunted computer models. So what else is new? We’re glad you asked that question, because there is something new. And that is solid confirmation of the errant nature of forecasting, especially that genre of the black art that traffics in financial and economic auguries. For such confirmation, we are indebted to Pierre Lussier, proprietor of the Investment Information Provider, who sent to us a neat study he did starkly entitled, “Street’s Track Record.” Using a survey by the Federal Reserve Bank of Philadelphia of a variegated mix of duly anointed economists (they nest in brokerage houses, banks, universities, private firms and consulting outfits), along with securities analysts’ earnings forecasts, Pierre has gone back to 1982 and traced the accuracy — or lack of it — of predictions one and two years out on a clutch of economic variables and corporate profits. What he found, not surprisingly, is that the further out the forecast, the greater the forecasting error. The clairvoyants were relatively — and we stress relatively — successful predicting housing starts, unemployment and long-term interest rates. They were pretty awful on inflation and GDP. In a rare nice-guy mood, we would rate the analysts’ projections on profits as middling. Specifically, on housing, the average gap between what the wannabe Nostradamuses said would happen and what really happened was 5% in one-year forecasts, 11% in two-year predictions. On unemployment, the seers averaged a 5% miss on one-year, and 12% on two-year, forecasts. On long-term interest rates, they were off by 10% in their expectations a year hence and 17% in two-year predictions. Their average error guessing short-term interest rates was 14% when they looked ahead one year, which ballooned to 47%, when they attempted to foretell rates two years out. On inflation, the professional economists were off 29% when reckoning the rate a year in the future; they missed by 41% trying to gauge what the rate would be two years out. They were wildly off the mark in their guesses of nominal GDP: by 102% one year ahead and 109% two years away. Analysts, as intimated, fared so-so doing their clairvoyant shtick. Their forecasts of operating earnings a year ahead went awry by 14%; when they pushed their luck and tried to see two years out, they missed, on average, by 25%. What’s the moral? Well, if you’re a forecaster, never look back and keep forecasting. And if you’re a civilian, at the very first whiff of a forecast, reach for a grain of salt; better yet, go buy a Ouija board of your very own.