No, Yogi Berra never said, “It is dangerous to make forecasts, especially about the future,” even though a recent Google search produced several claims that he did. The website Quote Investigator also disproves attributions to Samuel Goldwyn and Niels Bohr, among others, concluding that the comical proverb originated in Danish, author unknown. Whoever first said it, though, made a point worth bearing in mind.
Economic forecasts attract attention at this time of year. Companies have to make assumptions about future demand for their goods or services, on which they budget their head count, capital spending, and more. Thoughtful budgeting is also imperative for governments, at least at the state and local levels, where printing money isn’t an option if revenue falls short of outlays. Also, investment managers’ clients expect them to begin the year by laying out their strategies for the next 12 months. Economists’ predictions of indicators such as gross domestic product are key inputs to all this planning.
But how accurate are their prognostications? I recently took up that question for PitchBook. Each December from 2000 to 2021, Bloomberg News reported one-year forecasts of the change in real GDP by economists at banks, investment managers, independent research firms, universities, and other organizations. The number of economists reported as participating in this survey varied from 33 to 76. I compared the medians of their forecasts with actual GDP.
On average, the median forecast was 0.99 percentage points above or below the actual year-over-year GDP change (which averaged +1.99%). It is often the case, however, that averages conceal as much as they reveal. In December 2019, for example, the median forecast for 2020 was 1.8%. The actual number was minus 2.8%, resulting in a 4.6-percentage-point overprediction.
Nobody could have foreseen the Covid-19 outbreak. But the next year, the surveyed economists underpredicted the GDP change by 1.9 percentage points, nearly double the average forecast error.
Perhaps also buried within those averages was a forecaster who compiled a flawless 22-year record? If so, planners would seem well advised to follow that genius and ignore the group consensus, always keeping in mind that past performance is not a guarantee of future results. An examination of individual predictions in December 2007, though, dashes hopes of finding such a consistently accurate forecaster.
Economists’ forecasts for 2008 ranged from 0.7% to 3.4%. The actual GDP change was minus 2.6%. None got the direction right on the eve of the worst U.S. economic downturn since the Great Depression of the 1930s.
These results may tell us less about economists’ qualifications or diligence than the inherent difficulty of predicting the future level of something subject to as many imponderables as GDP. More problematic was my discovery of a pronounced optimistic bias. The median forecast exceeded the actual GDP change in 18 of the 22 years covered by my analysis.
Planners and portfolio managers should and will incorporate economic forecasts into their decisions. Wise ones, though, will avoid making too big a bet on individuals’ predictions, or the consensus view.
As for the current outlook, in December the Bloomberg-surveyed forecasters collectively predicted a 0.3% GDP rise in 2023. Last month, they boosted that figure to 0.5%. They could be right. Then again, 2023 could be one of those years when they get it badly wrong.
Martin Fridson, CFA, is chief investment officer of Lehmann Livian Fridson Advisors.