Central Banks Aren’t Very Good at Macroeconomic Forecasting

Peter Tulip and Stephanie Wallace of the Reserve Bank of Australia have a great research paper out in which they attempt to mark the RBA’s macroeconomic forecasts to market. Their conclusion is that the RBA is terrible at foreseeing the future:

We use past forecast errors to construct confidence intervals and other estimates of uncertainty around the Reserve Bank of Australia’s forecasts of key macroeconomic variables. Our estimates suggest that uncertainty about forecasts is high. We find that the RBA’s forecasts have substantial explanatory power for the inflation rate but not for GDP growth.

One interpretation of this would be that the RBA is for some reason more skilled at guessing about future inflation. I’d say the more likely difference is that the RBA is targeting an inflation rate. I’m pretty good at predicting what I’ll eat for breakfast, since I’m also in charge of deciding what to eat for breakfast. My forecasts about my neighbors’ breakfasts, by contrast, are terrible.

At any rate, as I wrote reviewing Nate Silver’s book I’m glad that election forecasting made him famous but I hope he’ll move on to other things. It turns out that a crude polling average is a pretty damn good predictor of election outcomes, and people just need to not let wishful thinking blind them to that. But macroeconomic forecasting is a total disaster area. Nobody seems to know how to do it with even a modicum of accuracy, and yet official institutions keep cranking out these forecasts.