How good are economists at forecasting economic variables? That will be the topic for discussion at this week’s Econoblog at WSJ.com (look for it late Wednesday or early Thursday). James Hamilton (co-author of Econbrowser) and I will be discussing some of the forecasting successes and failures of the economics profession, and their implications.
As a preview, I thought it might be interesting to take a look at one simple type of forecast: GDP growth. In October of 2005 the CBO published a report entitled “The CBO’s Economic Forecasting Record,” in which they took a look at the track record of the CBO in its forecasts, and compared their accuracy and bias to the “Blue Chip Consensus” (that’s the average of about 50 private-sector economic forecasts) and White House forecasts.
The following picture presents an interesting summary of the results, at least for GDP growth. It shows the three forecasts of GDP growth, and actual GDP growth.
Note: figures show two-year growth rates; forecasts are for current year and the following year.
First of all, it’s nice to see that the CBO and White House forecasts both tend to stick very closely to average private-sector forecasts. It’s less reassuring to see that all three types of forecasts are not particularly good predictors of actual GDP, however. In fact, the average error of the CBO and Blue Chip forecasts is about 0.97% (note that average GDP growth over the period is about 3.1%). On the other hand, a simple rule of thumb that guesses that next year’s GDP growth will be the same as this year’s GDP growth would have produced an average error of 0.94% over the same period – virtually identical.
That’s pretty disappointing to me. It would be like a meteorologist realizing that, despite decades of studying, theorizing, estimating, and modeling, they can not produce a better weather forecast than to simply guess that tomorrow’s weather will be the same as today’s.
Stay tuned for more on this subject in this week’s Econoblog…
UPDATE: Chart amended for clarity.
wow… great article… thanks