Settlement Risk reminds me that heavy truck sales crash during a recession. I was wondering how a 12-month change in this variable correlates with a corresponding change in vehicle mileage (suggested Stephen Kopitz). The last one does pretty lousy.
Figure 1: Probability of a recession using simultaneous 12-month changes in heavy truck sales (blue) and car mileage (yellow). 50% threshold, dotted red line. Peak-to-trough dates as determined by the NBER are in grey.
The McFadden R2 from the heavy truck regression is 28% compared to 7% per vehicle mileage (VMT). One can also simply see, using the 50% threshold, that the truck regression covers all recessions since 1970 without false positives, while the vehicle mileage regression misses all recessions except 2020, unless an absurdly low threshold is used. at 25%. (in this case we have a false positive result for 1995). The results are almost unchanged using Kopitz’s preferred measure, 12-month moving average change. Indeed, to be technical, VMT is a lousy recession coincident indicator.
In neither case is there a recession for May (13% car mileage) or July (4% heavy truck mileage) 2022 (latest data for car mileage and truck sales, respectively).