Study: Most Roads Don’t Pay for Themselves

Most American roads — even the most highly trafficked — are financial losers. That’s a major finding from a new study by the Center for American Progress [PDF].

Four out of 10 American highways don't generate enough revenue to pay for maintenance. Photo: Wikipedia
Four out of 10 American highways don’t generate enough revenue to pay for maintenance. Photo: Wikipedia

A financial analysis by the think tank found that about four out of 10 U.S. highways don’t carry enough traffic to generate sufficient revenue to pay for their maintenance — let alone construction.

CAP analyzed individual road segments from around the National Highway System. Using publicly available traffic data, researchers were able to estimate how much revenue each segment generated in terms of user fees paid by drivers, namely state and federal gas taxes. Those totals were then compared to average maintenance costs, assumed to be two resurfacings and one major reconstruction over the course of 30 years.

That just six in 10 highways passed such a low test should be a wake-up call, CAP authors say. For one, the cost analysis did not include initial construction costs or inflation. Including a modest annual 1 percent inflation adjustment on the cost of construction would have increased the share of roads that failed to cover costs by 9 percent.

CAP’s study only examined national highways, which host far more traffic than the average road. Roads on the National Highway System represent only about 5 percent of America’s total road network, but carry 55 percent of all vehicle traffic. Meaning the financial returns on local roads, which generate fewer trips and less fuel use than highways, are much worse.

The study should help dispel the false notion that roads pay for themselves, write authors Kevin Degood and Andrew Schwartz. It should also inspire us to rethink the way we disperse funding for roads versus transit, they say. (At the federal level, the split is about 80-20.) In most cases, the argument that roads are self-sustaining is a myth.

35 thoughts on Study: Most Roads Don’t Pay for Themselves

  1. America: “Tax money spent on things I don’t like is a subsidy. Tax money spent on things I do like is just my tax dollars at work. Everyone is subsidized but me.”

  2. This story is misleading because in the study 39% were a loss, 49% were a surplus, and 12% were break-even. So assuming the surplus was of the same magnitude as the loss, the system as a whole was in surplus. I don’t expect every road to pay for itself, but I expect the system as a whole to do so, and from the numbers that seems to be the case.
    However, it flips when you add that 1% inflation, and this further neglects the rental value of the land the roads are on. Additionally it assumes the gas tax is covering only road costs, while the gas tax further needs to cover externalities of driving, like pollution, noise, and other forms of damage. So the more you dig into it, the more of a black hole it all becomes. But the shallowest level of analysis, assuming these revenues are all for road maintenance, doesn’t look as bad as you describe.

  3. The author ignores a basic argument that is valid also when analyzing transit: transportation networks have value that is greater than the sum of its individual parts. The Interstates that go through the Rockies are a lightly traffic, but they guarantee a wider uninterrupted network, and that long-distance road traffic won’t put up with the atrocious mountain roads that preceded them.

    As for local roads, they can encompass a variety of topologies, from local arterials to connectors or local roads giving direct access to somewhere (a town, a subdivision, an industrial park or facility, a recreational spot etc).

  4. A good point, and also relevant to Amtrak. Frequently its long-haul service is criticized, but providing a national network allows for a more seamless (if slow) trip from origin to destination.

  5. Glad to see an analysis of highway finances being done, likely for the first time ever in a truly neutral way. We can quibble about the meaning of the results, but hard data has been almost impossible to come by when comparing roads to transit in recent times because of the almost religious reverence given to the highway by car users.

  6. Most street space is local, however, with no toll or gas tax revenue to support it. Some of that street space was required before the onset of motor vehicles — there has to be some access to buildings after all. But thanks to motor vehicles streets built later have had to be wider, and have fewer buildings per linear foot. And motor vehicles captured all the pavement for movement and parking.

  7. I don’t expect every road to pay for itself, but I expect the system as a whole to do so, and from the numbers that seems to be the case.

    No it doesn’t.

    Third, this analysis adjusts the life cycle cost estimate for each roadway down to reflect the share of roadway expenditures covered by user fees. States fund their roadway investments with revenues from multiple sources. As a result, user fees in the form of gas taxes and tolls cover only a portion of the annual total. Fiscal performance calculations reflect this reality by multiplying the life cycle cost estimate by the share of roadway expenditures covered by user fees. For instance, if a state derives 75 percent of its highway funding from user fees, the life cycle cost estimate would be reduced by 25 percent. By adjusting the life cycle cost downward, the analysis remains neutral about how a state chooses to fund its roadways. In effect, this downward adjustment means that roadways are only being held accountable for the share of life cycle costs intended to be covered by user fees. Again, this reduces the amount of driving needed for a roadway to break even or generate a surplus.

    So if a segment of roadway costs $2000/day to maintain, but it’s in a state that only funds 50% of it’s highway costs out of user fees, then it only needs enough traffic to generate $850-1150/day to break even. At $1151 it would be in surplus by their methodology.

  8. You’re right they only focus on the portion intended to be covered by user fees, but this article says 40% failed to cover expenses, and that this was an indication of something being wrong. My point is this logic is flawed. If you told me 40% of the bus lines in San Francisco failed to break even, I’d tell you that your logic was flawed because more in surplus than in deficit is a net win. Your point is good but this article didn’t make it.

  9. There are two subsidies to consider. First, the estimated cost of roadway maintenance in the aggregate is not covered by user fees, defined as tolls and estimated revenue from gas taxes. Second, considering only the proportion of costs covered by user fees, financial performance varies from one road segment to another, giving a cross subsidization of sorts. Roadway maintenance costs are taken from DOT, they’re a 30 year lifecycle cost. Divide by the number of days in 30 years and you have a daily cost. The gas tax revenue estimate is derived by taking traffic volume over the segment (including an estimated split between light/heavy duty vehicles), assume average fuel efficiency for each of the two classes of vehicles, and assume all gasoline burned was purchased in state. That together with toll revenue gives you an estimate for daily total segment revenue. Registration fees and the like are not considered user fees. Neither is bond revenue backed by future gas tax receipts.

    The issue of intramode subsidization is rarely discussed. It’s absolutely the case that some bus routes in San Francisco financially perform worse than others. There’s a myth that roads pay for themselves, or at least the direct cost of maintenance/construction. Some do, many could with a moderate increase in user fees directly tied to that road segment. Some could never cover their costs. With buses, the situation is similar, some could pay for themselves, some never could. A poor performing bus coverage route provides a social service. It’s legitimate to ask what social service is provided by a poor performing highway segment. It might provide some worthwhile social service. Maybe not. It’s worth asking the question.

  10. The study estimates gas tax revenue by taking the traffic volume over a road segment, multiplying by the length of the segment and the average fuel efficiency of light/heavy duty vehicles to estimate how much fuel was burned by vehicles using the road. Assume all gas was purchased in state, look up the tax rates, and there’s your estimate. It didn’t take actual gas tax receipts. So by this methodology local roads would produce some small amount of gas tax revenue, even if none of it was spent on them.

    Also, the study managed to say 99% of highways/principle arterials in NJ were in surplus. 97% in NY. It says toll revenue is a user fee, but I’m not sure how it allocates it among roads in the state.

  11. It is misleading if you want it to be…but only in favor of roads. FHWA’s own statistics show the highway system to be pretty deep in the red and heavily dependent on external sources of income.

    And, as has been repeatedly pointed out, FHWA highways are a fraction of the road network.

  12. IIRC, pump tax revenues tend to be dedicated to FHWA and various other large highways. That means, if you don’t use FHWA highways because you’re one of those Galtian super-titans whose entrepreneurial will doesn’t need to be driven on BIG GOVERNMENT HIGHWAYS, the state is still COERCING you to pay for OTHER PEOPLE who use those highways.

    (Pardon me. It’s late. I hope my faux RWA outraged rant was up to par with the real McCoy.)

  13. Amtrak is different. It operates just the Northeast Corridor tracks, where the reasoning applies. Elsewhere, it is a guest of freight railroads. We shouldn’t mix availability of infrastructure (roads, highways, dragged canals, runaways, railroads) with vehicles that use them.

  14. Actually, outside of the NE Corridor, Amtrak does indeed have some operations on Amtrak- or government-owned trackage. The two examples I can think of right now are a good portion of its Michigan Services and the Pacific Surfliner.

  15. Sorry about that, let me give it another read…

    …I reread it and found this, “four out of 10 U.S. highways don’t carry enough traffic to generate sufficient revenue to pay for their maintenance”

    and also this, “Most American roads — even the most highly trafficked — are financial losers.”

    Correct me if I am wrong, but if 4 out of 10 roads don’t have enough traffic to generate enough money to pay for their maintenance, then most roads are financial winners.

  16. Correct me if I am wrong, but if 4 out of 10 roads don’t have enough
    traffic to generate enough money to pay for their maintenance, then most
    roads are financial winners

    No, because you’re making the same mistake again. The 4 in 10 figure refers to a particularly atrociously performing set of highways.

    – Most highways don’t cover their costs
    – Some, 4 in ten, perform so badly they don’t EVEN cover their maintenance costs.

    The other highways that don’t cover their costs do, at least, cover their maintenance costs, but they don’t cover:

    – Construction
    – Land loss/property tax loss
    – Externalities.

    Try looking at private railroads for an instance of what constitutes a profitable transportation system. Railroads have to cover all of the above, even paying property taxes on their rights of way. Highway users don’t cover any of those costs.

  17. You know what else doesn’t pay for itself? Schools. Medicare. Medicaid. The criminal justice system. Lets get the young, old, poor, sick and imprisoned working harder so they pull their own weight.

    Also: Transit. Since most transit agencies do not pay their operating costs 100% from user fees + any dedicated tax source, they also should be questioned harshly.

    Discussing roads (and transit) this way is falling into the tea party trap–no one should get anything from the government that isn’t fully paid by those who us it, and those who don’t shouldn’t have to pay at all.

  18. Yes, your last two sentences are exactly what we should be asking because many transit advocates will say the same thing about “unprofitable” transit routes. If it’s good for the goose it’s good for the gander. I’ll add that we might also ask does a poor performing highway segment perform some worthwhile social function which can’t be performed in a more cost effective manner by some other mode?

    In the end we must think of transportation in general as an enabler of other activities which generate profit or serve a larger function, not as something which must always generate profit on its own. The only key is are we using the most efficient, cost effective means of transport possible for the given set of circumstances. In very rural areas where other forms of transport simply aren’t cost effective, a highway subsidized by taxpayers may well make sense.

  19. Why do you think that is? Well, I’ll tell you. Provided a sufficient volume of passengers or goods are being moved, no mode is more efficient in terms of energy, labor, or capital than rail. That’s why some railroads can make a profit without subsidies. The only key is they need volume to do so. The takeaway from this is basically whenever we have the requisite volume, the vast majority of the transportation system should be rail, not road. You may still need road in spots for its greater flexibilility, or for the proverbial last mile, but in general if the volume exists the vast majority of passenger miles or ton miles should be by rail. This holds true whether you still need subsidies or not. If you use rail, in general those subsidies per unit moved will be less than using road, at least until volumes fall below a threshold.

  20. You don’t even have to go so far as replacing a highway with a bike path or railroad, sometimes it’s just a matter of not overbuilding the highway in the first place. Or going with a cheaper option for the next life cycle.

    And yes, transit service should face similar scrutiny.

    You’re right, it’s fine if transportation facilities aren’t profitable. But that’s not license to waste money endlessly. There’s a balance between providing a certain minimum level of service for all, and spending dollars where they do the most good.

  21. Yeah, your comment was just the most convenient jumping off point–mostly a comment on the article.

  22. Austerity is my middle name. That’s why I always think we should lean towards the least expensive option which meets the needs in terms of capacity and speed. Yes, the problem isn’t highways but the fact that many are overbuilt. Same thing with some “gold-plated” transit systems.

  23. As if transit isn’t over- (and frivolously) scrutinized, it’s silly to subject it to the same scrutiny. Transit is mostly paid for by money sourced locally, even if it usually isn’t profitable. The consequences of over-supplying transit or overbuilding transit are typically low, and could even be inadvertently positive. “Oh shit, my bus is too frequent, I have no time to read at the bus stop,” said nobody ever.

    Sometimes local preferences or even design challenges should override fiscal concerns. Never you mind that the people who “scrutinize” transit often think they’re being fiscally conservative when they’re really just coming up with a way to spend more on an inferior service, or maybe think transit is just a second fiddle option to keep some people out of their way on roads.

  24. Needs can often be expected to change within the lifespan of a project, so it might be a bit imprudent to reflexively lowball.

    Capacity and speed might be competing goals too. Transit buses often meet capacity needs, but they’re pretty much never speed demons – until you consider grade separation, at which point you just may as well go with rail even if it buys you capacity you’ll never need. At least, that’s true unless the design reasons to do otherwise are pretty compelling.

  25. There are other compelling reasons to go with rail besides capacity. Mainly those are labor and energy savings over running buses.

    In the final analysis transit planning is a really complex process. Yes, you may end up needing more capacity than in the beginning as areas near transit become more dense. That’s probably a good reason to overbuild slightly, or to at least acquire land adjacent to the ROW in case you need to add tracks later.

    You could even apply similar thinking to roads, although with motor vehicles the old adage you can’t build your way out of congestion rings true. Probably there’s a stronger case to be made for not overbuilding roads than transit as any excess capacity generally just ends up encouraging more auto use.

    What transit really doesn’t need is excess money spent on gold-plated stations. In the end stations don’t move people, trains do. A station should be functional, able to handle the expected passengers, and that’s all.

  26. Rail’s labor saving is generally because of a higher ratio of users to labor. Though I don’t suppose there is any inherent disadvantage to rail serving low-volume service, it probably makes conversion of low-volume service to rail hard to justify financially.

    Seems like it doesn’t take many users to overwhelm a road with lots of “extra capacity.” The thousands of people using a given lane could fit on a minor bus route pretty non-disruptively. The marginal cost of more road capacity is incredibly high, often very damaging, whereas the marginal cost of more transit capacity is often just more vehicles using the same ROW.

  27. I’m looking at a lot of the comments here, and I can’t help but think a lot of us can’t see the forest for the trees.

    Maybe the problem isn’t subsidies as much as it is a double standard that is applied to transit, in particular rail transit.

    Transit is supposed to be “profitable,” particularly if it’s on rails, or it’s supposed to go away. No such standard is applied to the road system, or for that matter airports and even hiking trails.

    We have people squawking about the “wasteful” subsidy to Amtrak, yet we subsidize roads to the tune of at least 40% overall for the country and don’t say a thing, while Amtrak’s operating cost recovery is now over 90%.

    In Washington, D.C., there have been complaints about the Metro costing too much (and it’s pricey to build, no doubt), but Metrorail moves twice the passengers at half the cost of Metrobus, and Metrobus doesn’t have the maintenance of a rail system, its stations, its signal system, and the rest to pay for.

    This is not new. Check out this movie from the 1950s, particularly at about 13:30. The story hasn’t changed in 50 or 60 years:

  28. There is the other thing that there really is no precise data about road waste. This stuff about FHWA roads only scratches the surface because many or most roads aren’t FHWA. Users typically still pay something indirectly to someone (e.g., property taxes), but there is no guarantee that use is directly related to what you pay.

  29. This doesn’t surprise me at all. Roads with one lane each way are reasonably efficient for their purpose. Even up to two lanes each way might be useful. After that, you are into severely diminishing returns; nothing should ever be wider than two driving lanes each way.

  30. New York State has some of the narrowest “arterials” in the United States. Which is why they are able to cover their costs. Wide roads are much much more expensive.

Leave a Reply

Your email address will not be published. Required fields are marked *


More Money Won’t Fix U.S. Infrastructure If We Don’t Change How It’s Spent

“America’s infrastructure is slowly falling apart” went the headline of a recent Vice Magazine story that epitomizes a certain line of thinking about how to fix the nation’s “infrastructure crisis.” The post showed a series of structurally deficient bridges and traffic-clogged interchanges intended to jolt readers into thinking we need to spend more on infrastructure. The idea […]

Actually, Highway Builders, Roads Don’t Pay For Themselves

You’ve heard it a thousand times from the highway lobby: Roads pay for themselves through “user fees” — a.k.a. gas taxes and tolls — whereas transit is a drain on the taxpayer. They use this argument to push for new roads, instead of transit, as fiscally prudent investments. The myth of the self-financed road meets […]

The Great Traffic Projection Swindle

This is the final piece in a three-part series about privately-financed roads. In the first two parts of this series, we looked at the Indiana Toll Road as an example of the growth in privately financed highways, and how financial firms can turn these assets into profits, even if the road itself is a big money loser. […]

Highway Boondoggles: Widening I-70 in Denver

In a new report, Highway Boondoggles 2 (the original came out in 2014), U.S. PIRG and the Frontier Group teamed up to profile the most wasteful highway projects that state DOTs are building. Streetsblog will be serializing the case studies in the report. Today, we look at the widening of I-70 in Denver, a project with a […]

A Few Words on Transportation User Fees

We tend to have a few good laughs when Randal O’Toole fires up his Cato computer and weighs in on transportation issues. It’s hard to take seriously a man who thinks that having the government tax people to build something which it then gives away for free is the libertarian ideal. Do federal gas taxes […]