The Economist magazine once computed just how far the IMF’s forecasts were off on average over the period 2000–2014.12 For two years from the time of prediction (say, the growth rate in 2014 predicted in 2012), the average forecast error was 2.8 percentage points. That’s somewhat better than if they had chosen a random number between–2 percent and 10 percent every year, but about as bad as just assuming a constant growth rate of 4 percent.
