I would love to see a study that explores whether those taxes cover the negative externalities compared to other forms of transit, because that seems incredibly unlikely.

I've seen several different takes on that. They vary on how they value the externalities, and which components of the transit infrastructure they consider subsidies. For example, you could consider all infrastructure subsidies, or you could claim that basic 1+1 lane roads are essential infrastructure, while everything beyond that is subsidized.

The conclusions vary, except that they generally agree that driving in dense urban areas is heavily subsidized. And if you accept the roads as essential infrastructure claim, driving in rural areas is too heavily taxed.

Gas is the worst example of a subsidy - a better example would be the mandatory parking built everywhere, which is usually free.