Matt Yglesias is one of the best mainstream bloggers on land use/transportation that I know of, and, as one blogger who I don't recall right now once said, his urban planning and transportation posts could be blogs in their own right. However, it's puzzling that in an article for Cato Unbound, he comes up with such a pathetic rejoinder to the O'Toole/Cox/Poole vulgar libertarian transportation cabal, who don't seem to have ever met a road they didn't like:
Or consider the fact that Randall [sic] O’Toole is indignant about the prospect of public expenditures on mass transit systems, but appears to have little to say about public funding of highways. This, too, looks more like a case of narrow business interests than sterling free market principles.
While Yglesias' instincts are right – current transportation markets in America are highly distorted – the reason they're distorted has little to do with the ways highways are financed. Based on some basic figures, Randal O'Toole concludes that the vast majority of road funding – over 80% – comes out of user fees. Now, of course there're still some subsidies there, but it's really nothing compared to the subsidies that mass transit systems receive, which in America never even come close to covering operating costs, nevermind capital expenditures. Now, there are some problems with the 80% number, such as the government's favorable access to bond markets and the legacy of infrastructure that wasn't paid for with user fees, but all in all, it's hard to argue that roads have a subsidy advantage over mass transit.
However, that's not to say that Yglesias doesn't have a point when he says that libertarians and conservatives have blind spots when it comes to how they see transportation. But the real government benefit that the road/car system has over mass transit is density: there are innumerable regulations at every level of government in the United States which favor low-density, single-family detached housing over the denser forms that dominated non-rural areas before the 20th century. Successful roads as we have in America require this low density to remain (almost) financially solvent – it would be very difficult to cope with people's road needs if they were allowed to build as densely as they would without maximum density zoning rules and minimum parking regulations.
As a thought experiment, imagine your local town/neighborhood with twice the density. Chances are, the roads would quickly become very congested. They would have to be widened, which would require money, and even more money than normal, because the government would have to purchase valuable land next to existing roads. (That is, assuming that eminent domain is not used.) The gas tax would have to be raised, and soon the costs would get out of hand. On the other hand, mass transit would become more profitable rather than less, because much less track needs to be laid to satisfy the same demand, and mass transit systems have much more excess capacity than roads. If densities are limited, though, then this alleviates both stress on roads that go through valuable urban property (which are expensive and difficult to widen) and forces people to drive farther, thus paying more in user fees.
There's a legitimate case to be made against American transportation and land use policy, but condemning highway subsidies ain't it.