Recession Likelihood

Fiscal Insights Recession Prediction Model

The Fiscal Insights Recession Prediction Model was developed by studying several factors that contribute to the predictability of recessions. Recessions are typically characterized by a slowdown in aggregate economic activity. The National Bureau of Economic Research’s (NBER) Business Cycle Dating Committee studies several factors to retrospectively determine both the occurrence and duration of recessions. In its description, the NBER characterizes business cycles as follows:

"Expansions are the periods between a trough and a peak; recessions are the periods between a peak and a trough. By convention, the NBER classifies the peak month as the last month of the expansion and the trough month as the last month of the recession. Expansion is the normal state of the economy; most recessions are brief. However, the time that it takes for the economy to return to its previous peak level of activity or its previous trend path may be quite extended."

Several methods can be used to project the likelihood of a recession over a given time span. For example, standard time series econometrics tools could be used to forecast macroeconomic aggregates, which could, in turn, be used to determine the likelihood of broad cyclical decline in macroeconomic activity. In contrast, our method relies on treating recessions as discrete events, which can be predicted based on certain factors validated by external research and our own rigorous testing and research.

Our research finds that several factors related to the current trajectory of macroeconomic aggregates affect the likelihood of an impending recession. These factors include data series related to labor markets and productivity. Moreover, research indicates that expansionary and contractionary monetary policy can influence the near-term likelihood of a recession. Finally, we find that the current duration of an ongoing expansion, relative to the expected duration of the business cycle, carries significant predictive capacity.

Our research finds that several factors related to the current trajectory of macroeconomic aggregates affect the likelihood of an impending recession. These factors include data series related to labor markets and productivity. Moreover, research indicates that expansionary and contractionary monetary policy can influence the near-term likelihood of a recession. Finally, we find that the current duration of an ongoing expansion, relative to the expected duration of the business cycle, carries significant predictive capacity.

The figure above shows the model’s performance over time, as measured by the probability of a recession happening over the subsequent 12 months. Two things are notable. First, the probability of a recession tends to increase as the duration of the expansion increases. This defies the age-old adage that expansions don’t die of old age. Second, nearly every recession occurred at a time when the recession likelihood was heightened. To an extent, this reflects the nature of fitting a model to historical data that was used for its own estimation. At the same time, it reflects the precision with which the model was estimated.