We went searching for an economist who came close to predicting the surprisingly low number of jobs created in May – 38,000. We couldn’t find one.
Bloomberg surveyed 81 economists on their predictions for the nonfarm payroll number for May, the closely watched figure that represents the net number of new jobs. The average prediction was for 161,000 new jobs, which turned out to be 123,000 too high. The forecasting department at National Bank of Canada came closest, predicting 90,000 new jobs. Next was University College, Dublin, at 92,000, followed by German online banking firm Berliner Sparkasse, at 100,000. The top three guesses came not from renowned Wall Street banks but from non-US organizations.
Ten big Wall Street firms whose predictions get a lot of ink – Goldman Sachs (GS), Citibank (C), Bank of America Merrill Lynch (BAC), and so on – ranged from forecasts of 125,000 new jobs (JP Morgan Chase (JPM)) to 165,000 (Goldman). That’s a narrow band of predictions that averaged 146,000, not far from the average for the broader group of 81.
We point this out because investors (and journalists) tend to put way too much stock in predictions. Most of the time, the economy changes in measured ways that are fairly easy to predict, because next month’s change is likely to look a lot like last month’s. That’s one reason forecasts are usually tightly clustered in a range that mostly carries forward the current trend. For all the money, effort and person-power that goes into forecasts, they rarely vary by much.
The difficulty of forecasting becomes apparent when something unusual happens, which clearly seems to have happened in the labor market in May. Everybody was aware of a big strike at Verizon that was expected to pull down the overall number by 30,000 or so. Something more than that was obviously going on, and what, exactly, may not be apparent for months. If ever.
One other area where the perils of forecasting are on full display is the abrupt change in oil prices during the last two years. Last year, when crude prices were around $53 a barrel, we surveyed forecasters to see if anybody had come close to predicting that oil prices would plunge by 50% (and eventually, by nearly 80%) from the highs of 2014. Economist Gary Shilling actually predicted that oil could fall as low as $10, which was the low end of the range of forecasts. And he seems to have overshot, since the low was around $26, in February. The high end? Roughly $200 (an OPEC official), or about $175 higher than the price oil hit when it bottomed earlier this year.
Forecasters occasionally get wild moves right, but some of those folks are people always predicting disaster, who tend to be right about as often as a broken clock. More impressive are people who make money predicting things others fail to foresee – and who also get the timing right. Hedge fund manager John Paulson did just that when he saw the housing bust forthcoming around 2006, and bet massively against subprime mortgages, netting himself a huge $4 billion win. Luckily for economists, there’s usually not that much money on the line when they tell us what’s going to happen in the future.
Rick Newman’s latest book is Liberty for All: A Manifesto for Reclaiming Financial and Political Freedom. Follow him on Twitter: @rickjnewman.