What are the odds of a recession coming at some point in the next three years?
In order to more precisely quantify recession risks, we took four economic indicators related to the labor market (the Conference Board’s Current Situation indicator, the Jobs Hard to Get—Jobs Plentiful spread, the unemployment rate, and the 6m rate of change in each month’s initial jobless claims average) and four rates market indicators (the 3m5y curve, the 3m10y curve, the 5y30y curve, and the real Fed Funds rate) and ran probit regressions against the binary condition “did a recession start at some point over the next X years.”
A probit regression is very similar to linear regression, but instead of continuous variables, the regression is run against a binary data point. The result is a probability of a binary outcome based on continuous inputs; given a certain set of data points, what are the odds we enter recession? We note that for the real Fed Funds and unemployment rates, we use February data for March; all other variables are either March or March-to-date.