Pattern shifted down, however sensitivity up.
Determine 1: Log of ratio of auto miles traveled to month-to-month GDP (000’s miles divided by bn.Ch.2017$ SAAR) (blue), and 2000-19 stochastic pattern (purple). NBER outlined peak-to-trough recession dates shaded grey. Supply: NHTSA through FRED, SPGMI, NBER, and writer’s calculations.
Nonoverlapping log differenced q/q month-to-month GDP regression on log differenced VMT has a adj-R2 of 0.03 pre-pandemic, and 0.31 put up pandemic (coefficient rises from 0.14 to 0.23).
Given the structural break, it’s unclear how dependable VMT is a number one or contemporaneous indicator of recession (traditionally, VMT, gasoline consumption, and petroleum consumption all predicted the recession of 2022H1).
