> These shifts in driver population are more than enough to account for the change.

This is an interesting and maybe even plausible hypothesis, but it lacks any data whatsoever for the final claim.