Dispersion + Uncertainty = anything can happen!

One interesting feature of the FOMC's new reporting procedure (today's inaugural report, contained in the minutes of the October 30-31 meeting, can be found here) is that the Committee reports both the dispersion of forecasts across FOMC members and estimates of the uncertainty associated with the forecasts. I can’t think of any other central bank that does it this way: so far as I know, everybody else reports “a” forecast, and some estimate of the uncertainty associated with it. (For an example, look at the “fan charts” in the Bank of England’s Inflation Report.)

I’m not sure how to think about the interaction of dispersion and uncertainty, and the minutes provide very little guidance on this. For example, if I wanted to get a plausible range of outcomes for 2008 GDP growth should I take 2.15% (the midpoint of the central tendency) +/- 1.3%? Or should I subtract 1.3 from the lowest forecast, and add 1.3 to the highest forecast? This would obviously give me a much bigger range, but I suspect this would exaggerate the “true” forecast uncertainty. (I suppose I should go read the Reifschneider/Tulip paper to see if they address this.) But the point is, by reporting both the dispersion of the forecasts and an estimate of the statistical uncertainty, the Fed has complicated the task of figuring the likelihood of specific macroeconomic outcomes.