Inferring recession probabilities from the yield curve
Yesterday, we showed that inversions of the US Treasury yield curve are an imperfect but reasonable indicator for future recessions. Typically, the yield curve (as measured by the difference between 10-year and 2-year yields) inverts about one year before the onset of a recession, though lead times can vary dramatically.
Keeping these limitations about the ability of the yield curve to predict recessions in mind, one can calculate the recession probabilities based on the current steepness of the yield curve. Using our preferred measure of steepness, the 10s-2s spread, we have estimated the probability of the US economy being in a recession right now as well as the probability of the economy being in a recession one year from now. Our chart shows the resulting probability functions.
Obviously, a strongly inverted yield curve is a sign of being in a recession, while a positive steepness is associated with a very low recession probability of around 10% historically. What is more interesting is the probability of being in a recession in one year’s time. This recession probability increases steadily as the yield curve flattens and reaches a peak around a yield differential of -100bps. Since very strong levels of inversion are typically only reached towards the end of a recession the probability of being in a recession in one year’s time starts to decline again, once the yield differential drops below -100bps. The fact that the recession probability one year ahead never rises above c. 60% reflects the fact that yield curve inversions have an imperfect track record of signalling future recessions as we have demonstrated yesterday.
Today, the 10s-2s spread is hovering around 10bps. That puts us in a rather unexciting spot where the probability that the US economy is in a recession right now is minimal and the probability of being in a recession one year from now is a mere 30%. Being afraid of a recession in 2019 seems to us overly pessimistic at this point.
Recession probabilities based on steepness of the yield curve
Source: Fidante Capital.