Cross-Sectional Forecast Dispersion: Survey of Professional Forecasters

The cross-sectional forecast dispersion from all surveys since the fourth quarter of 1968 are available in the data sets listed below. The data sets are in Excel format.

Three measures of cross-sectional dispersion are available. We denote these measures (described below): D1, D2, and D3. For each survey date (T), we provide one or more dispersion measures for each quarterly forecast horizon. The horizons correspond to the current quarter (T) and the following four quarters (T+1, ... , T+4).

Not all dispersion measures are available for each survey variable. In particular, we avoid measures that are influenced by arbitrary changes in the scale of the data, such as those that arise when the base year changes in a benchmark revision.

For most survey variables, we use “level” to refer to the level of the variable — for example, the level of real GDP in chain-weighted dollars. However, for CPI and PCE inflation, we define the level as a quarter-over-quarter growth rate, in annualized percentage points, because the forecasters provide their projections for these variables in growth rates.

The documentation referenced below (Section entitled “Cross-Sectional Forecast Dispersion”) explains the organization of the files for dispersion, our computations, and any caveats.

Measure D1: Forecast Dispersion Data for Levels (643 KB; last update: May 10, 2013)
This dispersion measure is the difference between the 75th percentile and the 25th percentile of the projections in levels. This includes CPI and PCE inflation. See the notes above.
Measure D2: Forecast Dispersion Data for Q/Q Growth (276 KB; last update: May 10, 2013)
This dispersion measure is the difference between the 75th percentile and the 25th percentile of the projections for Q/Q growth, expressed in annualized percentage points. This excludes CPI and PCE inflation. See the notes above.
Measure D3: Forecast Dispersion Data for Log Difference of Levels (310 KB; last update: May 10, 2013)
This dispersion measure is the percent difference between the 75th percentile and the 25th percentile of the projections in levels.
Documentation (205 KB, 41 pages; last update: August 10, 2012)
The documentation (Section entitled “Cross-Sectional Forecast Dispersion”) provides the details on our organization of the files for measures of dispersion, our computation, and any caveats.

Return to the main page for the Survey of Professional Forecasters.