Who produces the survey?

The survey is produced by the staff of the Real-Time Data Research Center, located in the Philadelphia Fed’s Research Department.

Is there information available about the history and use of the survey?

Yes. Please see Fifty Years of the Survey of Professional Forecasters.

Where can I find the probability of a recession? What is the Anxious Index?

The survey does not ask the panelists for the probability of a recession. Rather, the survey asks for the probability of negative quarter-over-quarter growth in real GDP for the quarter when we conducted the survey and the following four quarters. There is one probability for each quarter. The Anxious Index refers to the probability of negative growth for the quarter that is one quarter after the quarter when we conducted the survey.

The data for the probability of negative growth are available at Data Files - Survey of Professional Forecasters (RECESS).

The data for the Anxious Index are available at The Anxious Index: Survey of Professional Forecasters.

What is the timing of the SPF?

The survey is conducted each quarter. Over the period that the Philadelphia Fed has been conducting the survey (1990:Q2—present), we send panelists the questionnaire on the day when the U.S. Bureau of Economic Analysis publishes its first estimate for the historical value of GDP. The deadline for returns is about 10 days later. Publication is about three days after the deadline for returns. Most panelists submit their projections over the few days leading up to and including the deadline.

We track the dates of the survey deadlines and publications, but not the dates when the panelists made their projections.

Historical dates for deadlines and publications are available at Historical Dates.

For more information on the survey’s timing, see SPF Documentation.

Are examples available for the survey’s historical questionnaires?

Yes. Please see Survey Form Examples: Survey of Professional Forecasters.

Where can I find the individual projections of each panelist?

The projections of each panelist are available at Individual Forecasts: Survey of Professional Forecasters.

Who are the panelists?
The panelists are professional forecasters who produce projections in fulfillment of their professional responsibilities. They have long track records in the field of macroeconomic forecasting. They are knowledgeable about the variables to be forecast and the relevant economic theories.

Is there any information available on the methods the panelists use to generate their projections?

Yes. Please see Forecasting Methods.

Can you tell me the names and affiliations of the panelists?

The survey is anonymous. We do not release panelists’ names and affiliations with their projections. Each published survey includes an acknowledgment listing the names and affiliations of recent contributors. There are no exceptions to the rule for anonymity.

How are the panelists identified in the online data for individual projections?

We identify each panelist with a confidential ID number. The same ID is assigned to each variable forecast by the same panelist. The survey is anonymous. We do not provide a forecaster’s name and affiliation with the ID number.

Important caveats apply to the ID number. For more information, see SPF Documentation.

Does the Philadelphia Fed keep track of the accuracy of the projections?

Yes. Please see Forecast Error Statistics for the Survey of Professional Forecasters.

Are historical values available for the variables forecast in the survey?

Yes. Please see the Philadelphia Fed’s Real-Time Data Set for Macroeconomists, at Real-Time Data Set for Macroeconomists.

How does the survey define growth rates?

The survey defines growth rates as the discretely compounded rate of growth, expressed in annualized percentage points. The survey does not use continuously compounded growth rates.

For more information, see Data Sources.

How do the panelists report their projections: In levels or growth rates?

Beginning with the 1992:Q1 survey, we permitted the panelists to report their projections in levels or growth rates. When a panelist reports in growth rates, we convert the projection to levels. The variables for which panelists have the choice between reporting in levels or growth rates are nominal GDP, the GDP price index, real GDP and its components (except real net exports and the change in inventories), industrial production, corporate profits, employment, and housing starts.

Over the survey period from 1992 to 2024, 70 percent to 80 percent of the respondents reported their projections in levels for nominal GDP, the GDP price index, and real GDP. Nearly 100 percent of the respondents reported in levels over the survey period 1990 to 1991.

Results are unavailable for other time periods and other variables.

Why are some data missing from the survey’s data files for some variables currently forecast?

Not every variable was entered into the survey on the same date. For example, the survey has always included projections for nominal GDP (or nominal GNP), the unemployment rate, industrial production, and housing starts. Projections for headline CPI inflation were first added into the survey for 1981:Q3. Projections for headline and core PCE inflation were first added into the survey for 2007:Q1.  

Is there a list of all variables currently forecast in the survey?

Yes. Please see Data Files - Survey of Professional Forecasters.

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