Before 7 of 8 US recessions, next-quarter GDP decline odds were below 25%
probability of a decline next quarter, Philadelphia Fed survey
In August 2007: 17%, and the recession began 5 months later; in August 2019: 14%, 7 months later.
More
The Survey of Professional Forecasters: each quarter since 1968 the Federal Reserve Bank of Philadelphia asks several dozen economists at banks, firms and research centres for their forecasts of growth, unemployment, inflation and the probability of a downturn. Here: the average probability of a fall in GDP next quarter (the “anxious index”) in each survey.
The line is the anxious index in each survey since 1968. Red bands are recessions as dated by the NBER. Dots mark the survey in the quarter before the NBER peak quarter. The dashed line is 25%.
- Recession from 1969-12: survey 1969Q3 probability of a decline next quarter
24.3%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 1973-11: survey 1973Q3 probability of a decline next quarter
12.2%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 1980-01: survey 1979Q4 probability of a decline next quarter
70.1%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 1981-07: survey 1981Q2 probability of a decline next quarter
20.2%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 1990-07: survey 1990Q2 probability of a decline next quarter
17.9%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 2001-03: survey 2000Q4 probability of a decline next quarter
10.9%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 2007-12: survey 2007Q3 probability of a decline next quarter
17.0%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
- Recession from 2020-02: survey 2019Q3 probability of a decline next quarter
14.2%
Federal Reserve Bank of Philadelphia, Survey of Professional Forecasters
The anxious index is the participants’ mean probability that US real GDP falls in the quarter after the survey (RECESS, next quarter). Recessions: NBER peak and trough dates; “a quarter before” is the survey in the quarter preceding the peak quarter.
A quarterly GDP decline is not an NBER recession. This is a timing comparison, not a calibration test. Testing probabilities requires all quarters, including quiet ones, and outcomes with the same event definition. NBER dates recessions after the fact.
A question for this picture
How well calibrated are these probabilities for quarterly GDP declines across all quarters?