Do Highway Users Pay for the Highway System? Not Even Close.

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.

record_gas_prices_large.jpgDo federal gas taxes really charge "users" of the highway? (Photo: CAP)

But occasionally O’Toole provides an opportunity to discuss some interesting aspects of the transportation planning process and learn from his errors. And so we turn to his latest policy paper, which was released yesterday. Therein, he writes:

The Interstate Highway System accomplished all of this [construction of the system] without any subsidies. Federal highway user fees paid for 90 percent of the cost of the system, and state highway user fees covered virtually all of the remaining 10 percent.

This brings up an interesting question: What is a user fee? Common sense would suggest that a user fee is a fee paid by a user of something in order to use that something. A common example might be a train fare. When one wants to ride a train, one purchases a ticket. One doesn’t purchase a ticket if one doesn’t want to ride the train, and one doesn’t ride the train without a ticket. A ticket is specifically meant to extract a fee from a potential user, that that user might then be allowed to use the train.

So do gas taxes count as highway user fees? Well, one might pay gas taxes even if one never uses highways. You pay the gas tax on gas used to drive down local roads or private driveways, or to power lawnmowers and tractors that never even see publicly-funded blacktop.

And one can use highways without ever paying gas taxes. Anyone able to obtain a vehicle powered by natural gas or electric batteries or canola oil can ride on the federal highway system for thousands of miles and never pay one cent to do so.

So gas taxes are not user fees. Indeed, the lack of actual user fees is one reason American highways suffer from severe congestion problems; when you give away something valuable for free — like scarce highway space — it ends up seriously over-consumed.

As a thought experiment, let’s consider a world in which federal gas taxes functioned more like a user fee. That is, let’s imagine that when drivers fill up, they pay a federal gas tax only on the gasoline consumed while driving on federal highways. That’s still not really a user fee, but it’s a little closer.

Light vehicles traveled a total of around 2.8 trillion miles in 2007, of which about 23 percent were driven on interstate highways, according to the Department of Energy. If we divide the total number of miles driven on interstates by the weighted average fuel economy of cars and light trucks, we find that about 31 billion gallons of gas were consumed on highways in 2007. That’s a lot!

Next, we know that the Federal Highway Administration budget is around $39 billion. If we assume that truck and diesel revenues are unchanged, then we have about $24 billion in highway funding to be covered by a tax on those 31 billion gallons of gas, for an estimated gasoline tax of about 80 cents per gallon.

That’s an extremely rough estimate. In fact, the nation’s light vehicle mileage includes some diesel-burning engines. If we adjusted the calculations to reflect that, the estimated tax rate would be higher. Highway fuel economy is also higher than the average figures, which means that the calculations above probably overstate the gallons of gas burned on highways. This, too, means that the estimated gas tax rate is too low.

An appropriate gas tax rate to cover the annual highway budget — which many argue is far too small — would be on the order of about $1 per gallon.

All in all, this should illustrate that if you set aside all the O’Toole hand waving about trust fund revenues shifted to other modes, you still wind up in a world where federal roads come nowhere near paying for themselves.

One final point: We learned last year that it doesn’t take much of an increase in the price of gas to generate reductions in VMT and increases in transit use. If we adjusted the current price of a gallon of gas to reflect an appropriate federal gas tax, gas would be selling for nearly $3.50 per gallon, on average.

With gas at that price, travelers would drive less and use transit more often. During the gas price spike last year, we saw a number of transit systems enjoy high demand during peak periods, to the extent that fares might easily have been raised to reduce system overcrowding.

In other words, what transit systems can charge riders depends upon what the government is (or isn’t) charging drivers. This is exactly what we’d expect; if Coke began heavily subsidizing its sodas, Pepsi would have to find a way to cut its prices or face going out of business.

What we see then is that there are two funding equilibria. If drivers pay a fair price for the use of roads, then highway revenues rise and transit fares can rise until both modes are full but not congested. This is the high revenue equilibrium.

But if drivers don’t have to pay a fair cost for the use of roads, then highway revenues will be low, roads will be congested, and transit systems will have too little ridership, such that transit systems will be unable to raise fares without losing riders to the already congested roads. This is the low revenue equilibrium, and it’s a bad place to be — inefficient use of all modes, costly road congestion, and a constant shortage of funding for needed infrastructure maintenance and investment.

That’s where we are now.

When Congress finally gets around to crafting a transportation reauthorization, it would be nice if they recognized some of these dynamics. America needs smarter infrastructure investment rules, but it also needs smarter revenue-raising methods. If you get the money the right way, that makes it easier to spend the money the right way.

  • Clutch J

    Your central point– driving is underpriced– is solid.

    I look at it this way: Cars are expensive, while gas (and, in many places, even parking) are cheap, and insurance is only somewhat tied to driving. Thus, once you’ve bought the car, it makes economic sense to drive it. We’d be better off having a transportation system with lower fixed and higher variable costs for users.

  • In Michigan, snowmobiles, ORVs, boats and others consume fair amount of fuel. Only since 1993 has some of that fuel tax been returned through programs such as the Recreational Trails Program (RTP). Even still, RTP is a donor program within the federal transportation bill and still subsidizes other areas.

  • You can argue that the gas tax collected from cars and light trucks who consume it immediately on the Interstate highways does not cover the entireity of the FHWA’s $40 billion bill. Of course your calculation includes only 23% of all VMT traveled on the Interstate Highway system (about 42,000 miles) and not the entire “National Highway System” (about 160,000 miles). Although the NHS was not constructed like the interstate with 90-10 or 50-50 matching grants from the feds, much of the grants funneled to states today (which is what you’re arguing about) end up expanding or repairing the National Highway System, which carries a much larger chunk of total traffic (about 44.3% actually). Even though every one of those VMT cannot be attributed to federal funds, it does significantly reduce your total number below the already low 80 cents a mile (hardly a point which will reshape american society). Also the FHWA does not spend money just on cars, it has pedestrian and bicycle safety programs, pollution mitigation programs and studies, and other non-car functions. This will even further reduce your calculation.

    But most of your argument here is irrelevant because the feds don’t even pay for all the interstate mileage today, and therefore you cannot simply divide Interstate VMT by Fed gas taxes. Most of the NHS and Interstates are supported by state highway maintenance funds, with only occasional grants from the US DOT (the feds contribute about 4 billion a year to interstates and 5 billion a year to the NHS, along with a host of sundry grants, and even these have to be at least half matched by states), and it is overwhelmingly clear that state highway user fees more than pay for highways and interstates.

    Ultimately what bothers me most about the argument you make(and similar ones) is that it seems to ignore its own conclusions. If gas burnt on the Interstate does not pay for the Interstate itself, and yet if the Interstate is not being subsidized by none-user fees (you cannot simply argue from the negative and say that since federal gas taxes do not cover the Interstate then it must be subsidies, obviously the rest of the interstates today are paid by state highway trust funds, which rely on gas taxes), then cars are over-taxed for an inefficient government-provided resource. If the Interstate is inefficient at generating sufficient gas tax revenue, than it is a burden on auto drivers, not non-drivers.

    This would imply that in a world without gas taxes and federal highways funded by them, people would drive even more. That may be a fine arguement to make from a libertarian perspective, but I don’t think it is the argument that most Smart Growthers want to make.

    There may be other reasons to tax VMT or autos (pollution issues, health issues, livability issues), but its time to lay the old federal subsidy canard to rest right here. It simply doesn’t add up.

  • Omri

    What state highway user fees? State highway spending is mostly from sales & income taxes.

  • Steve

    Agreed with the previous post. Your methodology is nonsense and is unnecessarily roundabout in addressing your basic point, which is that highways are priced too low. If you want to argue that gas taxes aren’t paying for roads, you might want to begin by looking at the FHWA budget and where its revenues actually come from. Of course the answer is, from gas taxes. So your “extremely rough” methodology doesn’t reflect that reality mainly because it’s “extremely rough,” a nice way of saying “not very good.”

    You’re confusing two issues here. One is that roads are underpriced, and the other is that they’re subsidized. The first point is valid. But the second one, that they are somehow not being paid for by user fees, when in fact the HTF is essentially a closed system (recent cash infusions notwithstanding), is false.

    Further, one could make the same argument about transit. You cite “train fare” as an example of a user fee. So how about we add up all the transit rides in America and compare that with the operating budgets of all the transit agencies to get a cost per rider. Do you think that might be a bit higher than what folks are actually paying at the farebox? Does that mean that transit riders aren’t paying their fair share?

  • Steve


    Every state in the U.S. imposes a tax on gasoline, and in most cases that tax is far higher than the federal gas tax. As pointed out by Judge Glock, this sizeable sum of money is completely ignored in the author’s analysis, rendering it completely useless.

  • Louis

    All the same, the gas tax is not a user fee, so you ought not to claim that it is. That said, a transit fare is not a true fee either.

    A fee is a charge that is intended to cover the cost of a service rendered. A zoning application fee is intended to cover the marginal labor required to provide the processing of that application.

    A fare is a “nominal fee,” which is something charged to cover part of the service, and serves mainly to dissuade usage of the service.

    Nominal fees should not be considered as part of a pure economic solution, as they are more tied to emotional and behavioral outcomes. Transit fares make riders and non-riders feel like the riders are paying some share, but since they are underpaying anyway, what is the use? Likewise, the gas tax contributes to the myth that road users are completely paying for the road system.

    It’s like the old Groucho Marx story:

    “Would you sleep with me for a million dollars?”
    “Would you sleep with me for ten dollars?”
    -“do you take me for a whore?”
    “We’ve settled that, now we’re just arguing prices”

    Once you begin charging a nominal fare or gas tax, the fare and the funding becomes completely arbitrary and political.

  • “If gas burnt on the Interstate does not pay for the Interstate itself, and yet if the Interstate is not being subsidized by none-user fees (you cannot simply argue from the negative and say that since federal gas taxes do not cover the Interstate then it must be subsidies, obviously the rest of the interstates today are paid by state highway trust funds, which rely on gas taxes), then cars are over-taxed for an inefficient government-provided resource. … This would imply that in a world without gas taxes and federal highways funded by them, people would drive even more.”

    Doesn’t make any sense. With fewer roads, there would be fewer people driving, because congestion would limit the amount people could drive. Imagine that all the freeways in, say, Los Angeles were eliminated and people could only drive on the local streets. There would be so much congestion that speeds would be close to 0 mph, and people would have to find alternatives to driving if they wanted to get anywhere at all.

  • Jeffrey J. Early

    Don’t forget that the Interstate Highway System doesn’t actually get people anywhere: they still need cars. That’s a HUGE difference between a rail system and an Interstate system. This makes it kind of a weird situation because the user fee is actually enormous (car, plus insurance, etc.); these are all costs which people don’t even realize they really have and yet are completely necessary for the Interstate to have any function.

    The average person must spend something like $5,000 annually on their cars if you add up insurance, maintenance, gas, parking, and purchase cost. Imagine what kind of transportation system you could build with that kind of money :-). I’m not actually advocating that, but its an interesting thought experiment.

  • vnm


    (Sorry to be grumpy here, but,) this post would have benefited from putting into writing what the gas tax is today and how much is actually raises per year. Here is my calculation. If there are 31 billion gallons consumed and the gas tax is 18.4 cents per gallon, that’s $5.7 billion raised, a far cry from the $24 billion needed.


    With that kind of gaping hole, that doesn’t seem like a closed system.

    In February, the Congressionally appointed National Surface Transportation Infrastructure Financing Commission, many of our nation’s foremost transportation experts, acknowledged the need for a massive infusion of money into the highway trust fund. It recommended that a mileage based fee system be implemented by 2020, and an immediate 54% increase in the gasoline tax which should then be linked to inflation. Our nation’s highways are clearly underpriced and subsidized.

  • vnm


    The costs you mention (purchase of vehicle, insurance, maintenance) don’t benefit the public. They are transactions between households and private companies for the acquisition and upkeep of private property. All of the user fees associated with mass transit benefit the public because they help public-transit organizations that provide service to all.

    While it is worthwhile to point how expensive it is to become a motorist (and I think even AAA would say that your estimate of $5,000 is quite low), it is false to suggest that those large expenses justify a heavily subsidized road infrastructure at the expense of the general public.

  • Steve

    @vnm, not sure where you got $5.7B but:

    “The Highway Trust Fund, which is primarily funded through federal gas tax receipts, collected $31 billion in revenue between October 2007 and September 2008” (Source: FHWA)

    I happen to agree that our fuel tax needs to be increased enormously to properly price auto transportation and manage demand. However, I also think that sloppy, easily discredited analysis full of phantom numbers does no one any good. I’ll repeat what I said before: pricing transportation to manage demand has nothing at all to do with “paying your fair share,” and auto drivers certainly do pay for the highways they use.

  • rex

    O’Fool lost any credibility many years ago. I think he should seek medical help. If he really believes what he writes, his perception of reality is tenuous at best. Ryan, I am surprised you thought it was worth a read.

  • Bill Nelson

    If “user fees” covered the cost for any mode of transportation, then “public ownership” would not be needed.

    What’s your point, anyway? That since highways are not self-supporting, therefore money-losing transit somehow makes more sense?

    Incidentally, your claim that “transit systems will be unable to raise fares without losing riders to the already congested roads” is baffling. If roads are congested, then by definition, there is no room on them for ex-transit users.

  • darren

    @Steve, vnm and judge glock are right. Streetsblog is using the wrong numbers in this analysis. He computes the mileage and revenues based only on INTERSTATE miles. But the outlays for interstates only from FHWA are less than $6B, which is in spitting distance of the 18.4 cent gas tax.

    Streetsblog should really dig into their numbers a bit further and post a correction, because while the central point and debate are worthy, the numbers are just wrong, and leaving bad analysis out there is a tool from the O’Toole toolbox that shouldn’t be borrowed.

    Now if you want to quantify how unsustainable our surface transportation system is, have a look at the amount of state and local income/sales/property taxes gets dumped into the system…

  • Ryan Avent

    I appreciate the commenters’ thoughts. I perhaps went too far in saying that tax revenues come “nowhere close” to paying for themselves. The point I hoped to make was that O’Toole’s assertion that highways would pay for themselves if gas tax revenues weren’t diverted to transit is not correct.

    And while the analysis is certainly incomplete, it’s not entirely out of order. If you include all highway miles, you still wind up with an estimated gas tax rate around 2.5 times the current one, and that’s to cover a highway budget most regard as inadequate to the task of properly maintaining the roads. Adding in state revenues doesn’t help, since it’s far from clear that state gas tax revenues cover state highway expenses. I’ve previously linked an analysis from the Texas DOT which suggested that state gas taxes would have to be considerably higher to cover the life-cycle expenses — construction and maintenance — of its roads. No Texas road pays for itself with gas tax revenue.

    Meanwhile, we’ve also learned that the transportation budget can be destroyed by rising fuel prices. A commenter above says that trust fund revenues last year were $31 billion (note: that’s not all gas tax; close to half comes from taxes on diesel and other trucking fees). That’s down about $7 billion from 2006. This suggests that if you begin tweaking gas taxes to be more in line with budgets, you quickly find yourself on the wrong side of the revenue curve.

    But commenters are right that much of this is beside the point. It would be most efficient to charge drivers fees for use of scarce road space. This would reduce congestion, raise more (and more dependable) revenue, and better balance travel across modes. That’s the point I closed with, and that’s the point I hope resonates.

  • I still don’t see where Ryan Avent gets his “2.5 times the current” gas tax number which is supposedly required to pay for contemporary highways. If you look at the FHWA link that darren provides, non-highway user fees make up 32.86% of highway disbursements, or about 55.5 billion dollars a year (most of these general funds are from the local, not the federal, level.) If you subtract the 22.5 billion dollars collected from user fees that go to Mass Transit and “non-highway purposes,” you get about 23 billion dollars, or about 16% of all highway disbursements, coming from non-user fees. So the total gas tax (for federal, state, and local governments) would have to rise only 16% to make all highways financially self-sustaining.

    When you consider that most mass transit systems today operate at only a 50% fare-box recovery ratio for operating expenses, and with completely uncompensated capital expenses at least equal to annual operating costs (50% of highway expenses are capital expenses, and this half estimate may be rather low for mass transit because it has significantly higher initial construction costs than highways), that means the average mass transit fare would have to rise almost over 400% to pay for itself, and to end the almost $11 billion dollars a year that go from highway user fees to mass transit. The financial comparison with highway’s 16% increase doesn’t even come close.

    Again, there are other reasons to tax highway users using variable toll pricing and other congestion mitigation strategies, but there is no doubt that highways are infinitely more financially self-sustaining than mass transit.

  • I’m sorry, I meant mass transit fares would have to rise almost 300% to pay for itself.

  • Jeffrey J. Early

    @vnm, I think that we’re in agreement here — I should have put the “user fee” of cars in quotes. My entire point is that the cost of a car is missing from the equation.

    “it is false to suggest that those large expenses justify a heavily subsidized road infrastructure at the expense of the general public.” It sure would be false to suggest that — that’s why people should realize these costs.

  • I’m sorry, I meant mass transit fares would have to rise almost 300% to pay for itself.

    What you’re missing here, Judge, is the value of ridership. Most of the subsidies are required because the buses and trains are running less than half full for a large portion of the day.

    Almost half of the public bus systems in the US can make an operational profit with an average of forty people per bus – without raising fares. Or with an average of twenty people per bus paying double the current fare.

  • Hmmm. Ah yes, the ‘ol excise tax. Yes, let’s price ordinary Americans off the road, so that the rich may better avoid noticing us. Sounds like an awesome idea. I’m sure access to a personal automobile won’t be relevant to those trying to ascend the class-ladder. No, no, a nation where only the wealthy may have this privilege certainly should go over well with 100 million poor people.

  • Ian Turner


    Given that drivers are on average wealthier than non-drivers, wouldn’t it make more sense to stop subsidizing driving altogether? User fees would be one way to accomplish that.

  • Vance makes an excellent point when he says, with some deliberately irony: “I’m sure access to a personal automobile won’t be relevant to those trying to ascend the class-ladder.” The notion that car ownership symbolizes social promotion is one of the most crippling myths of car dependence, especially in places like New York, where other transport options are plentiful. The livable streets movement needs to address this psychopathology by inverting it: Car dependence doesn’t make you look cool or uppercrust. It just turns you into another victim of what JHK referred to as the national automobile slum.

  • Subsidize biodiesel from wastewater for a volume supply of non-petroluem fuel large enough to cover all transportation needs globally, cars, buses, trucks, trains, airplanes and ships at sea from what most towns and all cities have, a wastewater treatment plant.

    What this will do is turn our dependence on oil into a glut of biomass produced fuels, prices below $1/gallon of a renewable resource that grows with population.

    The economics of this will push the green button … until they can have their cars and not pollute the planet we’re not done, people will not give up the car, ever.

    Switching fuel cuts harmful emissions by 60% over gas-ethanol, extract it to burn it up, so is a big step in the right direction on greenhouse emissions, and, this fuel can be produced locally removing the coercive effect of the oil economy from the equation of national budgets.

  • Ian Turner


    Why is technology-specific subsidiy preferably to cap-and-trade with auctioned permits, to a per-tonne carbon tax, or to a petroleum import tariff?



  • The economics of this will push the green button … until they can have their cars and not pollute the planet we’re not done, people will not give up the car, ever.

    I hope you’re wrong, because then we’ll still have plenty of emissions AND sprawl, carnage, obesity and a system that leaves those who can’t drive high and dry. Thanks, Tom!

  • Patrick McMahon

    It would be nice to see a related posting about the economics/sources of funding for local and state roadways, where gas taxes account for an even smaller proportion of the funding sources and there is a higher level of subsidy (by a mix of income, property, and sales taxes).

    Some good work has been done on this point and I’ve included a couple of links below.

  • Nathanael

    BTW, a study in Wisconsin showed that in that state, the state gas tax paid for less than half of the state road budget. The rest was property, income, and sales taxes.

    The same is probably true in most states (Wisconsin is not a particular outlier on either its gas tax or its amount of road construction). So if you add in the states, it’s even more certain that roads are subsidized by non-road-users.

  • MisterBadExample

    I think the research vis a vis the addition of taxes/user fees to fuel is a good model, but there’s one more wrinkle to it. My guess is that drivers in the Old South do more of their travel on roads paid for by the Federal Government. In the 1930’s, states like New York and Connecticut already had decent paved roads; states like Louisiana didn’t. The formula for the New Deal made for federal subsidies for places like Louisiana to get a larger share of federal blacktop. So my guess is that if you adjusted the tax based on which states’ drivers travel most on federally financed highways, gasoline in places like Atlanta would be more expensive than in places like Albany, NY.

  • what ever it is as a citizen we have to play our Tax Lien so that we can contribute too in the development of our country.


Think Roads Pay for Themselves? Think Again

The myth that U.S. roads “pay for themselves” thanks to user fees is a subject that’s likely familiar to many Streetsblog readers — but just how much of the nation’s highway funding is provided by charging drivers? The answer may surprise even active critics of the current asphalt-centric transportation system. Between 1982 and 2007, the […]

Six Lies the GOP Is Telling About the House Transportation Bill

The transportation-plus-drilling bill that John Boehner and company are trying to ram through the House is an attack on transit riders, pedestrians, cyclists, city dwellers, and every American who can’t afford to drive everywhere. Under this bill, all the dedicated federal funding streams for transit, biking, and walking would disappear, leading to widespread service cuts […]