Getting Highway Numbers Right: The Tax Foundation’s Response

On Thursday, January 17, the Tax Foundation (TF) issued a paper arguing that only 32 percent of state and local highway costs were paid out of user fees, while the remaining costs came from “general funds.” In a post here, I pointed out that, actually, user fees for highways cover 76 percent of the costs of roads and most of the remaining 24 percent come from interest on user fees before they are spend and bond sales that will be repaid out of user fees.


TF replied, saying “O’Toole conflates taxes and fees.” In fact, TF specifically said that state gas taxes are user fees, but somehow defined federal gas taxes as “general funds.” I simply argued that, to be consistent, TF should count federal gas taxes as user fees as well.


TF went on to say, “O’Toole suggests we include federal gasoline tax collections in state‐​local revenue.” Again, TF said that federal gas tax collections are “general funds” and I disagreed with that statement. If state gas tax collections are user fees, then federal gas tax collections are too. They are certainly not general funds, any more than state gas taxes are general funds, since federal law dedicates them to transportation projects and mostly to highways.


TF said, “O’Toole suggests that we include motor vehicle registration taxes and fees, but not the associated expenses” such as highway patrols. In fact, I said nothing about the associated expenses because, for the most part, those expenses are already included in the reported $155 billion cost of highways.


TF said, “O’Toole suggests that we include state and local bond sales for road construction, which would double‐​count revenue.” But nothing I said would double‐​counting revenues. What I said was that bond sales for highways are not, in any sense, “general funds” if they will be repaid out of user fees.


TF said, “O’Toole suggests that we include $13 billion in “investment income” on state‐​local gasoline tax and user fee revenue, but that is not a net interest figure.” What TF means is that some of that interest might come from investments of non‐​user fees, which is true. But since user fees cover the vast majority of state and local road costs, interest on those user fees makes up the vast majority of interest. Yet TF counted all interest as “general funds.”


TF said, “O’Toole suggests that we use Federal Highway Administration data rather than U.S. Census Bureau data. We have no evidence that the U.S. Census Bureau is unreliable in this area.” I suggest that the fact that the Census Bureau, which uses secondary data, differs from the Federal Highway Administration, which uses primary data, is itself evidence that the Census Bureau data are unreliable.


In sum, by TF’s own definition of user fees as being gas taxes and tolls, something like 55 percent of the cost of roads is collected in user fees. Adding vehicle registration fees brings this to 76 percent, and most of the rest is covered by bonds that will be repaid by user fees and interest on those user fees. TF’s reply failed to address my main point, which is that none of these revenues can be considered “general funds.”