Recession Probability
A probabilistic estimate — typically between 0% and 100% — that the U.S. economy will enter a recession within a given forward window, most often 6 or 12 months.
Recession probability models estimate the likelihood of an NBER-defined recession over a forward horizon. The most well-known public model is the New York Fed's, which uses the 10-year minus 3-month Treasury yield spread as a single input; a fully inverted curve historically produces a 12-month recession probability above 30%. Academic models like Wright (2006) add fed funds rate levels and short-rate expectations to sharpen the signal.
More elaborate composite models use 10 to 20 leading indicators — yield spreads, unemployment trend, credit spreads, financial conditions indices, initial claims, housing starts, consumer sentiment — and weight them dynamically based on regime detection. These models trade single-variable clarity for higher hit rates and earlier warning, at the cost of interpretability.
BullrunData's model uses 15 economic indicators with dynamic weight adjustments based on Fed policy stance and market regime. Retrieve the live probability at /api/v1/model/probability and the historical series at /api/v1/model/probability/history.