The Washington, DC region is great >> and it can be greater.


The Transit Trust Fund: a 21st century solution

Cities and towns all over the United States are demanding more transit infrastructure. But a lack of funding has stymied transit expansion. Finding a solution to this problem is essential.

Photo by Daniel Greene on Flickr.

Since its invention the mid-19th century, transit has been a powerful economic development tool. Some of our most celebrated cities grew up around their rail systems: the New York City Subway, Los Angeles's extensive pre-war streetcar system, and most recently, the Washington Metro was a main ingredient in our region's dramatic revival in the 2000's.

If there is so much demand for more transit, and it builds healthy cities and towns, how do we fund more? The answer lies in harnessing transit's power to increase land values.

The Federal Transit Administration's New Starts Program does not have enough money to fund every transit project currently being planned. It can take decades for a new project to make it through the funding pipeline. The Silver Line extension of the Washington Metro, for instance, has been in the works for approximately forty years. The Purple Line in Montgomery and Prince George's counties has been in planning since 1986.

In order to be eligible for federal funding, all projects must meet very strict criteria. These criteria are very tilted towards fulfilling immediate need rather than planning livable, vibrant, economically healthy towns/cities in perpetuity. Yet, using transit infrastructure as a planning tool is exactly how New York City grew into a world capital a century ago.

The FTA's New Starts program receives a very small fraction of the federal transportation budget, which mostly consists of gas tax receipts. The gas tax was conceived as a mechanism to have motorists pay for road costs. A car owner has to buy gasoline in order to use the roads. As more roads were built, more motorists drove more miles. The government then used the higher gas tax receipts to build more roads, causing higher gas tax receipts, etc. It was a very intelligent system when it was conceived. The gas tax's purpose was to provide funding for roads. It was never meant to provide construction money for transit. As long as transit funding is connected to the gas tax, its construction will remain anemic.

Transportation infrastructure (including roads) is not a profitable endeavor in the United States, except for some purely freight uses. The only time in history it was profitable was between the Civil War and World War II. A company would build a streetcar or subway line from a city center to land that was too remote to be worth anything to someone who commuted to work on foot. When streetcar service began, the value of the land surrounding the stations greatly increased because it was then connected to the city center. The streetcar builder would then sell the greatly appreciated land and make a profit. While our remote places are now dozens of miles more remote than they were in the late 19th and early 20th centuries, new transit infrastructure can still dramatically increase property values.

The creation of the gas tax was a clever way to repurpose some of the economic activity created by new road infrastructure towards funding roads. Similarly, a better way to fund new transit infrastructure would be to capture some of the new economic activity created by transit. It worked a century ago when done by private companies. It will work in the near future when done by the citizenry through its elected government. In Arlington County, Virginia, 33% of all tax receipts are collected in the Rosslyn-Ballston Corridor, which takes up only 8% of the county's land area. The positive difference in tax receipts between what the Rosslyn-Ballston Corridor now generates compared to what the dying strip malls that used to line the corridor would be generating is immeasurable.

We should set aside a small fraction of all tax receipts from new transit-oriented development into a Transit Trust Fund. There will be more transit-oriented development as more transit is constructed. Tax receipts from new transit-oriented development that go into the Transit Trust Fund would then be used to build more transit. The new transit would then generate more transit-oriented development, further adding to the fund. It will create a positive feedback loop and a snowball effect similar to how the gas tax once filled the Highway Trust Fund. Just as all levels of government were once incentivized to build more roads to increase gas tax receipts so they would have more road money, government would then be incentivized to build a starter transit line so they could increase the money in the Transit Trust Fund, and so on.

Once the Transit Trust Fund becomes large enough, the federal government could start offering to contribute some large fraction of the construction costs towards building transit projects, subject to well-defined metrics. Back in the 20th century, the federal government provided 90% of the funding for the construction of Interstate Highways. Unsurprisingly, lots of Interstates were built over three decades. It is probably not realistic to expect the federal government to be so generous with new transit projects in the near future. A vastly different fiscal situation, combined with higher materials and labor costs implies a different construction timeline. The first decade would be much slower as the fund fills up from the first tax receipts from the first projects that are currently being planned.

Transit-oriented development that has already been built should NOT be subject to the Transit Trust Fund. Most local and state governments around the country are in trying fiscal situations because of the ongoing Great Recession. I doubt that most could afford even the small fraction of tax receipts that would go into the Transit Trust Fund. Rather, I propose that new transit-oriented developments around new transit stations should pay into the Transit Trust Fund. There should be no subtraction from any local government's already tight budget.

The expansion of our road network in the 20th century provided lots of construction contracts, in addition to more economic activity related to building cars and car-dependent sprawl. The Transit Trust Fund would similarly provide decades of transit construction contracts, in addition to more economic activity related to building transit vehicles and transit-oriented development. The former was a novel idea in the mid-20th century. While we now understand that its model of land use is unsustainable, it provided a backbone for economic growth for decades. While the details of the Transit Trust Fund would have to be ironed out in the political process, the model would work on the local, state and/or federal levels. The Transit Trust Fund would incorporate the most successful infrastructure funding strategies of the 20th century and put them towards a sustainable 21st century template for decades-long economic growth.

Cavan Wilk became interested in the physical layout and economic systems of modern human settlements while working on his Master's in Financial Economics. His writing often focuses on the interactions between a place's form, its economic systems, and the experiences of those who live in them. He lives in downtown Silver Spring. 


Add a comment »

What money, specifically, are you proposing to skim from? Property tax payments? Business receipts? Rental / mortgage payments? TOD is a planning concept, not a revenue generating entity by itself.

by Lou on Aug 13, 2010 11:03 am • linkreport

As an concept, there are some large holes. Basically you are asking the federal government to come in and tax land? Not sure that is every going to fly.

As a rhetorical device, also not helpful. Way to far up the ladder.

I'd start by saying future FTA decisions need to look at how much gasoline would be saved by investing in transit. That is somewhat included already -- environmental, etc -- but it does seem as if the overwhelming criteria is will this project pay for itself over time?

Also, directly make it a criteria that the transit investment would LOWER the amount of cars on the road and relieve congestion.

by charlie on Aug 13, 2010 11:11 am • linkreport

This could be done on a strictly local or regional basis.

A certain percentage of tax receipts within a certain distance of transit stations could go into a fund to build or maintain transit.

I don't agree that it should apply only to new development. Transit is part of the fabric of life in Metro DC. We should have dedicated funding. It makes sense for the areas that benefit from transit to pay for it. It would only be to their benefit as service would improve.

by Steve on Aug 13, 2010 11:26 am • linkreport

If you exempt existing projects, the first generation of new projects would be paying into the fund without getting anything back out, and unless the tax increment on that first generation was really high, it would take a long time for the fund to get to a decent level. I'd note the gas tax didn't exempt drivers using gas on existing roads, and I'd suggest the same logic applies here: increment taxation of existing projects should provide the seed money for the first generation of new projects, which will prevent you from having to excessively tax the first generation of new projects in order to get the ball rolling.

by BrianTH on Aug 13, 2010 11:58 am • linkreport

Can anyone think of any state legislatures that have looked at something like this? Since it implies property taxation, it seems ready-made for large states with several metro areas, like California or Texas. But, as mentioned already, if it does involve property tax it can't be done at the Federal level.

One possibility would be a national sales tax which from time to time is trotted out as an alternative to the income tax. Were that ever to get off the ground, it would be a good idea to lobby for this system to be introduced contemporaneously.

by Packherd on Aug 13, 2010 12:07 pm • linkreport

While this could be a concept towards solving some of the capital needs for rail transit, it doesn't do much for bus, paratransit, or vanpool. Additionally, while the capital expense of the construction is large, the operating expenses are what continue to grow over time. This concept wouldn't do anything to solve the issue we have now with operational funding support.

by Chris on Aug 13, 2010 12:37 pm • linkreport

Charlie, you'll almost never be able to lower the number of cars on the road with any infrastructure project of any kind, unless you demolish car-oriented land uses and return them to pasture.

No matter how much density you create in one area on a rail line, the homes in the suburbs/exurbs that are only dense enough for cars/roads will continue to exist and contribute to congestion identically.

At best you could hope to keep congestion the same (make it no worse), by building densely around new rail lines and NOT building any more car-dependent exurbia. This is what has happened in Arlington. LOSs are almost the same as what they were 25 years ago, only we've added almost a hundred thousand people into the mix.

If Arlington's R-B corridor hadn't been built this way, either the region wouldn't have grown, or more people would live further out, and the traffic WOULD have been far worse.

by Joey on Aug 13, 2010 1:25 pm • linkreport

Nothing in principle would prevent you from incremental taxation of property along a bus route, which could go to operating expenses. I think you would have to set the incremental rate pretty low in cases where there was no dedicated infrastructure, but theoretically you could have a sliding scale of rates (e.g., the lowest rate for just a basic bus route nearby, the highest rate for a grade-separated dedicated-ROW station nearby, something in between for streetcar and BRT routes without grade-separated dedicated-ROWs--you get the idea). That would apply the logic behind this notion not only to entirely new systems, but also to upgrades.

by BrianTH on Aug 13, 2010 1:28 pm • linkreport

Why levee a portion of local tax receipts to the Federal government so that you can just distribute it back to the localities? I agree that we need to decouple transit funding from the constraints of the gas tax and Highway Trust Fund, but you need to find a revenue stream that actually comes from private enterprise, like the gas tax does, not from local tax revenue. Do this and you also eliminate having to tip-toe around dicey local budget conditions.

More importantly, if you can devise a functional value-capture method, it would be best implemented at the local level to best take advantage of the incentive you point out. If a city can be sure that it will receive all of the financial/tax benefits from building new transit infrastructure, they have a strong incentive to build said infrastructure.

by Erik W on Aug 13, 2010 2:16 pm • linkreport

Erik, my framework would work for all levels of government either independently or in cooperation with each other.

When I mentioned using a small fraction of tax receipts, I meant all taxes. That includes sales, property, income, whatever. Maybe a $.005 sales tax. Maybe something else. That's why all levels of government have budget analysts. They'd do a better job than I would on that exercise. I thought about this in aggregate. As I mentioned, specific legislation would have to done later. Step one is to get a visionary framework.

The big problem why we haven't had any progress on transit funding is because we haven't done anything to address the fact that we don't dedicate any of the economic activity that's due to transit to more construction.

by Cavan on Aug 13, 2010 3:14 pm • linkreport

Great summary of the problem and history, though I'm not on board with this solution.

I encourage folks to familiarize themselves with Los Angeles' Measure R, which passed with 64% of the vote in '08, and provides long term dedicated transpo funding (transit and road).

We need to decouple transit from the "needs to pay for itself" mentality, it never will. But like other forms of govt spending, the net benefits are greater than the net costs, it's just we can never directly collect all the net costs from beneficiaries since the distribution of benefits ends up being very very broad.

A decent analogy is military spending (put aside pro/con military feelings for a moment), you can never attribute what share of benefits accrue to which individuals, but we can all agree that it's a net benefit that we deter foreign invasions and confiscation of property. We pay for all military costs out of the General Fund, and don't expect, say, oil companies to pay a proportional fee for the implied protection their U.S. flagged vessels enjoy when passing through the straits of Hormuz or Malaga. We all pay, even though the benefits accrue more to some than to others.

I love the discussion of value capture, and I believe some level of it is necessary to incite new TOD and balance the private gain created by public investment. This is where the conservative in me comes out, does the years long New Starts cue actually delay transit projects? Wouldn't a local approach like Measure R with no hope of additional federal funding incent more municipalities to take an up or down vote on transit without getting in a long line, hat in hand, for the years away hope of getting the feds to pay for the project?

by Wil on Aug 13, 2010 3:21 pm • linkreport


You can't exactly do that ... automobile use pays for itself plus for much of mass transit via the part of motorist taxes siphoned off for projects that benefit mass transit. (E.g., A part of my Dulles Toll Road tolls are going toward paying for the Silver Line Extension.) But of course it can be argued that the reason my mode of transit which does pay for itself is being asked to help pay for other modes of transit is that if these other modes of transit did not exist, there might be too many people on the road to let my mode of transit exist. So, while you might not want to look at whether a particular mode of mass transit can pay for itself (because as you said it probably never will), you still do need to look at the whole picture in terms of all modes of transit combined paying for themselves. For example, you might want streetcars serving downtown Silver Spring, but you don't want to make it more difficult for motorists to go there when changing things to make it easier for mass transit to serve the area. That would be killing the golden goose that lays the eggs to pay for that mass transit.

by Lance on Aug 13, 2010 4:19 pm • linkreport

Lance, you are incorrect about roads paying for themselves. I addressed that in the post. The gas tax once paid for roads but has not kept up with inflation over the years. Consequently, road funds are now taken out of general funds.

That is why relying on the gas tax for transit construction is not a good idea. Hence why I'd write a post about a NEW idea.

You are simply incorrect.

by Cavan on Aug 13, 2010 4:31 pm • linkreport

Excellent write-up.

The biggest challenge for state and local governments is that transit construction isn't always where "transit construction money" goes.

The government worker pension crisis (baby boomers who went to work as governments expanded in the 1970's and beyond are retiring and collecting their state or local pension) is staring to hit, and is already crippling California and many other states. And - as was the case with BART - the unions and their allies strongly support the bond initiatives and reforms but at a high cost. The cost is sometimes reform itself, and always involves salaries, health and pension contributions, and work rules.

Bay Area voters have passed several transit bond issues, and have gotten far less bang for their buck than was promised. Even the alternative media has been questioning where all the money went, and their answers have been disheartening.

So while it is good to support these ideas, it is critical to understand that it's a lot easier to pass legislation than it is to implement it in a way that brings the desired result.

by Mike S. on Aug 14, 2010 1:29 pm • linkreport

@Cavan, The gas tax once paid for roads but has not kept up with inflation over the years. Consequently, road funds are now taken out of general funds.

This claim has come up before, so I posed the question to a DDOT person I know from the neighborhood who would know whether it is true. He said that there's a small grain of truth in it in that over the last couple years the gas tax has indeed not kept up ... But that a very small increasein the tax would close that gap ... and more importantly that historically people using cars for transportation do indeed pay for themselves. He added that part of the reason it 'pays for itself' is that people drive themselves and they buy their own cars. I.e., The direct contribution from the motorist is large enough in proportion to the total costs, that the government's share of the costs can easily be covered via gas taxes and the like,

by Lance on Aug 14, 2010 3:51 pm • linkreport

I'm not sure this is a "solution" although it's a contribution to the discourse in any case.

Most areas served by transit benefit from the proximity, regardless if the property was constructed overtly as TOD. And the increase in property value is captured in part through property taxes, and presumably income and sales taxes generated through increased use, reduced vacancy, etc. Separating out some of the tax revenue for transit, at least in DC, and in the region happens anyway, through the way the WMATA Compact is funded. Passing specific targets might be better.

In any case, I think a transit withholding tax comparable to that collected in certain counties in Oregon is far better and would generate more money. This is what pays for the fareless square in Portland. In a paper I wrote a few years ago, I estimated that it could generate $200MM/year in DC. Of course, the feds would have to agree to it being assessed on federal jobs, which is likely to be a sticking point.

P.S. Lance, name the DDOT person. Just because the person works for DDOT doesn't make them credible. The points you recount are at odds with the facts on the ground, plus the needs for proper maintenance of failing infrastructure, and of course even the gas tax doesn't take into account the cost of US Military forces oversees, maintaining access to foreign oil. The necessary gas tax to pay for road maintenance, infrastructure improvement, road extension, the other stuff (bike and ped, other transpo enhancements) and the military would be at least $4.60/gallon. So I don't see how what this DDOT person says is credible.

From the 2003 report Improving Efficiency and Equity in Transportation Finance:

"... revenues from fuel taxes have for three decades been rising more slowly than program costs as legislators become ever more reluctant to raise them to meet inflation."

37 years of significantly declining revenues from the gas tax is a long time.

by Richard Layman on Aug 14, 2010 6:54 pm • linkreport

I should have written "overseas" and the last sentence would be more accurate if I had instead wrote:

37 years of declining revenues, with revenues declining precipitously over the past 10 years, along with a simultaneous significant increase in demand for infrastructure replacement (in particular, bridges) is in fact a crisis, and not a problem that miraculously developed over the past couple years.

by Richard Layman on Aug 14, 2010 8:28 pm • linkreport

@Richard Layman

Wow. Transit payroll taxes. Those are interesting and I had no idea they existed. Those would work at the Federal level and they are ready-built for an opt-in process. No pressure, no commitment. I like this a lot.

Go Oregon!

by Packherd on Aug 14, 2010 9:01 pm • linkreport


Please see, which are the latest highway revenue and disbursement statistics available as compiled by the Federal Highway Administration.

Total federal, state, and local highway user revenues (including motor fuel taxes, vehicle taxes, and tolls) for 2008 were $122 billion. Of these, $15 billion went to mass transit, $4 billion went to costs of collection, and $8 billion were used for other non-road purposes, leaving $94 billion for road purposes.

But total expenditures on road building, road maintenance, highway law enforcement, and administration at all governmental levels was $182 billion. So even if none of the road taxes and tolls had been diverted to mass transit or other purposes, and their collection had cost nothing, the highway user fees would still account for only 67% of highway costs.

If you go to, which presents summary statistics from 1921 to 1995, and compare the first line for each year (which is highway user fee collection, before collection costs and non-road use diversions) and the last line (which is total disbursements for roads), you will see that the last time highway user fees would have covered the costs of roads even before diversions was 1953.

by rock_n_rent on Aug 15, 2010 9:03 am • linkreport

@Richard and Rock_n_Rent,

The disconnect, as explained to me by this DOT person (yep, I used the wrong acronymn earlier) is that these numbers don't reflect all the costs paid directly by drivers. Because a driver drives him/herself and purchases and maintains their own vehicle, the share of the costs provided by governments (fed or local) toward the real costs of moving people around is relatively small. I.e., the majority of the costs of using our road network are born directly by the users of that network ... even before you count the gas taxes they pay. So, the share of the costs of the system paid out of 'general taxes' and the like is a tiny percentage of the full costs involved.

Compare that to how mass transit if funded where only a very small percentage of the costs of building and operating the system is actually borne by the persons using it, and you see the problem of comparing 'apples' to 'oranges' ...

Bottom line is that if we had mass transit only, we couldn't afford it. It would be cost prohibitive ... Our getting around like we do is very dependent on the average person being able to 'do their own driving' vs. paying more taxes for the government to do it for you ...

by Lance on Aug 15, 2010 10:53 am • linkreport

*Compare that to how mass transit is funded

by Lance on Aug 15, 2010 11:18 am • linkreport


What you're saying doesn't make sense. How does a driver paying for their own car help pay for infrastructure? It doesn't.

The Texas DOT did a study looking at user fee generation and life cycle costs of roads (construction, maintenance, etc for a 40 year timeframe) and found that the best roads only generated about 50% of the revenue needed to cover costs. Some were as lot as 16%.

The Asset Value Index, was developed to compare the full 40-year life-cycle costs to the revenues attributable to a given road corridor or section. The shorthand version calculates how much gasoline is consumed on a roadway and how much gas tax revenue that generates.
The Asset Value Index is the ratio of the total expected revenues divided by the total expected costs. If the ratio is 0.60, the road will produce revenues to meet 60 percent of its costs; it would be “paid for” only if the ratio were 1.00, when the revenues met 100 percent of costs. Another way of describing this is to do a “tax gap” analysis, which shows how much the state fuel tax would have to be on that given corridor for the ratio for revenues to match costs.

Applying this methodology, revealed that no road pays for itself in gas taxes and fees. For example, in Houston, the 15 miles of SH 99 from I-10 to US 290 will cost $1 billion to build and maintain over its lifetime, while only generating $162 million in gas taxes. That gives a tax gap ratio of .16, which means that the real gas tax rate people would need to pay on this segment of road to completely pay for it would be $2.22 per gallon.

This is just one example, but there is not one road in Texas that pays for itself based on the tax system of today. Some roads pay for about half their true cost, but most roads we have analyzed pay for considerably less.

by Alex B. on Aug 15, 2010 11:24 am • linkreport

@Alex What you're saying doesn't make sense. How does a driver paying for their own car help pay for infrastructure? It doesn't.

It's the 'network' I'm talking about. Not the 'infrastructure' as in 'roads'.

Let's pretend we're talking about computers. There was a time when all computing was done on main frames. (Main frames = mass transit). They were very expensive, so expensive that many companies actually rented time on these mainframes rather than purchasing them out right. Then some decades later they came out with Personal Computers ... and a method of connecting them ... thus forming a 'network'. (Personal computers = personal transportation = eg 'car').

In the days of mainframes only, it was thought that the capability and availability of 'computing services' would always be limited by the fact that you needed hardware the size of a small room for it to work. Hardware that cost many millions of dollars ... and that got paid for by the bank or the government or whatever entity was using it.

When the personal computer came about, you could now concentrate computing power where and when you needed it ... and it didn't require the many millions of dollars of investment ... and, it could be paid for by the person/company/whatever actually using it. People got more choices as to how much computing power they wanted to use ... and when ... And they got to pay the bill for that choice by purchasing their own personal computers (personal computers = cars) and paying for a network connection ('network connection = Internet access) ...

If we'd tried to prop up mainframe computing by raising more taxes to pay for it under the assumption that it was somehow better for people to 'share' a large machine than to pay directly for their own machines, do you think we'd be better off now? Would you be running to your local IBM office to 'blog' now?

Most of the costs involved in using personal computers and the world wide computer network (i.e., the Internet) is born by the users. The tax dollars that go toward operating and maintaining the Internet (i.e., the network) are a miniscule part of the whole where the vast majority of the costs are paid for by the user. The same applies to the network used by drivers. The costs of building and maintaining actual physical roads is tiny compared to the full costs of maintaining and operating a non-mass-transit transportation network ... with the vast majority of the costs being borne by the users of that network ... i.e., the drivers.

by Lance on Aug 15, 2010 12:02 pm • linkreport


O.k., I see your point. To put it quantitatively, the government subsidy for roads is about $60 billion per year now. We have 3 trillion vehicle miles traveled per year now, so the government subsidy is around 2 cents per mile. Given that, for income tax purposes, the government accepts an average cost of 50 cents per mile for vehicle travel (similar to AAA calculations of "true cost" of driving), the subsidy amounts to about 4% of total costs.

Per the American Public Transportation Association's 2010 Fact Book (at, total operating and capital costs for all 7,700 public transit operations in the country were about $54 billion for 2008. Fare collections were about $12 billion, implying a government subsidy of $42 billion, or 78%.

Given that there were 55 billion passenger miles on public transit in 2008, total pre-subsidy costs amount to about 98 cents per passenger mile, or nearly twice the unsubsidized cost of 52 cents per mile forf private automobile travel (this ignores that some automobiles carry more than one passenger; on the other hand, the private vehicle cost does not include the cost of parking). I have to admit, I was really surprised by this number. Is this due to lack of economies of scale in the U.S. - that is, are European transit operations more efficient on a per-km basis? Perhaps the additional personnel costs for what is in effect a chauffeur become insignificant only when spread over sufficient passengers, and in the U.S. we are paying too many drivers to drive around mostly-empty buses. But in any case, I will accept that, in fact, on a pre-subsidy basis, in the U.S. mass transit is more expensive per passenger mile than private automobile travel.

From a microeconomic standpoint, though, these numbers also point out that subsidized mass transit is still a better deal than driving (22 cents per mile for mass transit, vs 50 cents per mile for driving). So the APTA is still correct that you can save money by ditching your car (which I have done). Lots of people do not believe this, because they ignore the capital costs of a vehicle as sunk costs and dismiss depreciation (which spreads the capital costs over the use of the vehicle) as an accounting gimmick; thus, most people think that driving costs them only 10 cents a mile or so.

But yes, Lance, I will now accept that I can save money by ditching my car only because you still have your car and are subsidizing me. Thank you!

by rock_n_rent on Aug 15, 2010 12:16 pm • linkreport

Lance, I completely understand what you're saying. The government isn't paying for vehicles or operators.

I'm not sure how that helps your point - roads have managed to keep all those costs 'off the books', so to speak, and are still way behind on 'paying for themselves.'

The bottom line is that the mantra of requiring something to 'pay for itself' isn't all that useful when crafting a real funding mechanism. It's a somewhat useful-if-inaccurate political one-liner, that's it.

by Alex B. on Aug 15, 2010 12:22 pm • linkreport

@rock_n_rent But yes, Lance, I will now accept that I can save money by ditching my car only because you still have your car and are subsidizing me. Thank you!

That was an excellent analysis you did ... And in such a short period of time! I'm sincerely impressed. And btw, I thank you for availing yourself of mass transit, because by you and others doing that, the roads are less crowded for me ... and I can better estimate my times of arrival on them ... and for that it is worth me (and other drivers) subsidizing mass transit for those routes (and times) most heavily traveled. (The fact that current operating costs for mass transit reflect the 'low hanging fruit' of routes and times with the most demand would indicate, at least to me, that the numbers you came up with for mass transit only get worse the more mass transit we build because it involves the 'less traveled' commutes and times ...)

by Lance on Aug 15, 2010 12:37 pm • linkreport


I decided to play around with the APTA numbers for 2008 a bit more. The following are computed statistics, rather than those provided directly by the APTA.

If you look at subsidized cost per passenger mile, it is remarkably consistent across mode: 22 cents per mile for buses, 20 cents per mile for commuter rail (i.e., VRE or MARC), 22 cents per mile for heavy rail (i.e., Metrorail), 18 cents per mile for light rail (which includes streetcars). The only outlier is paratransit, which charges on average 35 cents per mile.

There is much more variability for unsubsidized costs (operating and capital expenses). Commuter rail is the most efficient, at 64 cents per passenger mile, and heavy rail (i.e., Metro) is close behind, at 73 cents per passenger mile. Buses are more expensive, at $1.04 per passenger mile (as expected, given higher ratio of drivers to passengers; the additional personnel cost outweighs the benefit of not having to pay for track costs). Light rail comes in at a whopping $2.35 per passenger mile (this is not a one-year anomaly; in 2006, the cost was $2.18 per passenger mile, and in 1995 it was $1.24 per passenger mile ($1.75 in constant 2008 dollars, calculated using U.S. CPI). It appears light-rail is disadvantaged by both high driver/passenger ratios and high capital costs. Paratransit is again the most expensive, at $4.03/mile.

So it appears that having a system with TOD near metro or commuter-rail stations (which I believe is an approach widespread in Europe) is not that much more inefficient than private driving, while buses add inefficiency and light rail is very inefficient.

by rock_n_rent on Aug 15, 2010 1:57 pm • linkreport

@rock_n_rent, Thanks for the additional analysis. I had a question on your previous analysis. You said We have 3 trillion vehicle miles traveled per year now, so the government subsidy is around 2 cents per mile. Given that, for income tax purposes, the government accepts an average cost of 50 cents per mile for vehicle travel (similar to AAA calculations of "true cost" of driving), the subsidy amounts to about 4% of total costs.

That 50 cents per mile for vehicle travel ... Does it include the total cost of ownership for the roads? (I.e., the data you gave me earlier) ... or just the total cost of ownership for the vehicles traveling those roads? If it doesn't, then your denominator gets larger (by the total cost of ownership of the roads )... and the 4% comes down ... Right?

by Lance on Aug 15, 2010 2:20 pm • linkreport


The 50 cents per mile includes the full cost of gas (including excise taxes) as well as motor vehicle registration fees and taxes, and thus the portion of road costs that is in fact paid for by highway user fees. It does not include the 2 cents per mile government subsidy for roads.

So the subsidy is 2 cents out of 52 cents, which (within rounding) is still 4%.

Another way of putting it is as follows: per the FHWA (at, total motor fuel consumption in 2008 was about 175 billion gallons. Over 3 trillion vehicle miles, this works out to 17 mpg average fuel efficiency. Thus, in order to eliminate the direct government subsidy for roads, the gasoline excise tax would need to increase by 34 cents per gallon (ignoring costs of collection - those would add maybe another 2 cents per gallon).

by rock_n_rent on Aug 15, 2010 3:24 pm • linkreport

Interesting discussion.

@Packherd -- thanks. MTA in NYC just added a withholding tax as part of their funding stream due to the massive drop off in tax revenues from real estate recordation taxes. But Oregon was first and only for a long time. I haven't looked into the NY particulars:

@Lance -- this is also interesting. WRT the issue of personally owned automobiles being cheaper than a mass transit-based paradigm, it's hard to say whether or not that is true, because the cost of a reusable resource like a train car is a lot less than each user otherwise buying, maintaining, and replacing automobiles throughout their lifetime. It's like the argument about treating health care as a public good because by doing so, administrative burden is significantly reduced. I'd have to sit down and try to work through some numbers, but hey, I am not that motivated. (Other more pressing stuff calls.)

It's nice to see your recognition that transit use in places where it really is used, such as DC, has significant positive impacts on congestion reduction. I am amazed myself, witnessing the relatively free flowing streets (with exceptions, particularly at certain times of day of course) so much throughout the day that I can only accord to the success of the transit system. It's great to see when transit works. It is frustrating that it can work like this only in a few places. And it is hard work trying to get other places to make the transit system work comparably. I failed miserably in my attempt to reshape the Transportation Element in Baltimore County's Master Plan around significant changes in the transit network there (and TDM policy too) for the City and County, with the County adding those proposals to their Master Plan...

by Richard Layman on Aug 15, 2010 3:36 pm • linkreport

For clarification in my understanding of this discussion - is the money raised from the gas tax used for local roads? I thought it was just for the highway trust fund.

If the gas tax is just for highways, seems that this discussion needs to add the local road numbers into the calculations as that is a lot of local subsidies for roads not being raised from the gas tax. If the tax is used for all roads, I learned something new today and my comment should be deleted :).

by timfry on Aug 15, 2010 9:56 pm • linkreport


I admit that the Federal Highway Administration's Highway Statistics ( has a lot of numbers in it that can make you blurry-eyed; I'm an accountant by profession, so lots of numbers actually make me giddy with excitement :).

To break down the numbers for 2008: the feds collected almost $37 billion, most of it from the federal excise tax on gasoline (18.4 cents per gallon), but some of it on special excise taxes on trucks, trailers, and tires. The states (plus DC) collected about $80 billion. This includes state-level gasoline excise taxes that ranged from 8 cents per gallon in Alaska to 37.5 cents per gallon in Washington State and averaged 20.48 cents per gallon. In addition, states levy a number of other taxes and user fees on motor vehicles (for example, registration fees, license plate fees, and property taxes) as well as tolls; together these amount to about half of the states' receipts. Finally, local governments collected around $5 billion; this includes both local excise taxes and local tolls. This all adds up to $122 billion.

Once you subtract out diversions to mass transit funding and other non-road uses, and then add in general fund appropriations and similar non-road funding, the total amounts available were $42 billion at the federal level, $98 billion at the state level, and $53 billion at the local level, for a total of $193 billion. As you can see, this means there was relatively little additional funding at the federal and state levels, but at the local level, there is a humongous subsidy from non-road-user sources.

Of the $193 billion, about $11 billion (mostly the proceeds of newly-issued bonds) were placed in reserves. In addition, the feds transferred $37 billion to the states, who in turn transferred a bit more than $12 billion to local governments.

The bottom line is that the feds spent $3 billion, the states spent $114 billion, and local governments spent $65 billion, for a total of $182 billion.

So, the $122 billion raised from highway users is not just gas taxes, but also motor vehicle taxes and tolls at all levels of governments, and the $182 billion spent is also for all levels of governments (from DOT research to local cops and snowplows). The overall shortfall is $60 billion, which is not the shortfall for the federal highway trust fund, but the shortfall for the entire road system across all levels of government.

Thus, if one wanted to make the road system truly self-supporting, one way would be to increase the federal gasoline excise tax by 34 cents per gallon (from 18.4 cents to 52.4 cents), and then transferring these revenues down to the state and local levels.

Note that this will cover the direct costs of the road system; it will not cover parking facilities, the social costs of pollution and highway accidents, or the possible government subsidies for securing petroleum supplies that others want to include in gasoline taxes.

by rock_n_rent on Aug 15, 2010 11:26 pm • linkreport


I also thought that the economies of scale of mass transit (using the same vehicle to transport numerous people) would mean that mass transit would be cheaper to operate than individual automobiles. This is why I was so surprised when I looked at the numbers on actual expenses of the country's public transit agencies as compiled by the APTA (which is trying to put public transit in the best possible light, and thus is not likely to be inflating these figures).

From the 2010 Fact Book at, the following numbers can be calculated for 2008:

If you take total operating expenses and capital expenses per mode, and subtract out all salary and fringe expenses, and then divide by total passenger miles, then the total cost per passenger mile (i.e., for capital costs, materials, supplies, and fuel) is 48 cents per mile for buses, 41 cents per mile for heavy rail/subways, 40 cents per mile for commuter rail, and $1.96 per mile for light rail/streetcars.

I found AAA's study on the costs of driving (including depreciation and finance costs to account for capital expenses) at This has numbers for 10,000 miles per year, 15,000 miles per year, and 20,000 miles per year. Per this table from the Federal Highway Administration (, the average passenger car/SUV/pickup truck/van in 2008 drove 11,432 miles. Adjusting the AAA average number for this mileage, one gets 64 cents per mile (hmm, so the IRS is short-changing you by allowing you to deduct only 50 cents per mile for business travel). From the same FHWA table, one gets that there were 1.64 passenger miles for each automobile vehicle mile traveled, so the average costs of owning and operating a car are 39 cents per passenger mile. To this needs to be added the 1 cent per passenger mile (=2 cent per vehicle mile) government subsidy for the road system, as well as some additional amount for the average cost of parking, which is not included by the AAA (but needs to be for an apples-to-apples comparison, because public transit agency expenditures include both capital and operating expenses for rail yards, bus depots, and the like).

Parking in downtown DC is now going for, what, $200 a month? Spread over 11,432 miles per year, that would be 21 cents per mile. This is probably on the high side for an average cost across the country, as land costs on average are probably lower than in downtown DC, and most parking also does not involve the erection of parking structures.

So, on a passenger mile to passenger mile basis, the material costs are 40 cents per mile for commuter rail, 41 cents per mile for subways, 48 cents per mile (the road subsidy per passenger mile is negligible) for buses, and somewhere between 40 and 61 cents per mile for automobiles, depending on the average cost of parking. So mass transit appears to offer only marginal economies of scale in terms of capital and non-labor operating costs.

Then, on top of that, with mass transit you have to add the labor costs, which (from the APTA fact book) amount to 24 cents per passenger mile for commuter rail, 32 cents per passenger mile for heavy rail, 40 cents per passenger mile for light rail, and 57 cents per passenger mile for buses.

So, once you factor in the costs of having a chauffeur rather than driving yourself, unsubsidized mass transit is more expensive in terms of capital and operating expenses per passenger mile than a private automobile system, at least in the U.S. today. This of course only goes to the direct expenses of each mode of travel and ignores externalities from accidents, pollution (including global warming), and the aesthetics of suburban sprawl, as well as the opportunity costs of doing remunerative work while being transported in mass transit rather than sitting there and swearing at traffic.

by rock_n_rent on Aug 16, 2010 12:32 am • linkreport

@r_n_r -- Not having your facility with numbers, I would offer the point that as long as the automobile-based mobility system is subsidized, it's hard to split out the real costs of each respective system.

30 people have to spend at least $20,000 per car to get around, which equals the cost of 1 bus, which moves far more than 30 people each day. etc.

But transit works best in tight links between destinations and activity centers. When you figure out the sprawl-polycentrism nature of much of it, transit often doesn't compare favorably. OTOH, in DC, transit has significant ROI for individuals (and other stakeholders--city, property owners, developers, businesses, etc.), compared to say Oklahoma City, or even other jurisdictions in the region. We can thank the city's design (L'Enfant) for the success of transit here, along with the continued vibrance and relevance of the central business district, which in turn helps to retain and attract residents, who are able to use transit in trips under 7 miles pretty efficiently...

wrt the other question about gasoline excise taxes, there are two types, federal, which funds federally-funded infrastructure and state-local, which funds state-local infrastructure. E.g., State of Maryland spends some of the money on state roads and some goes to the local jurisdictions for projects.

Of course, the federally funded infrastructure funding is mostly managed at the state level on federal roads (interstate highways, etc.) and projects by the state DOT with the oversight and management of the Federal Highway Administration and other agencies of the US DOT.

Most states have raised their tax over the years, while the federal rate has not changed. Note that Virginia hasn't raised their fee for many years.


Somewhere there is a table that lists when these rates (at the state level) were set, but I can't find it at the moment, and have to go...

by Richard Layman on Aug 16, 2010 12:55 pm • linkreport

Excellent article and many excellent responses. I have a few points.

First, as mentioned by others, TOD does not generate “surplus” revenues. Businesses adjacent to transit may do more business. However, additional profits that might result are siphoned off by the landowners. Landowners will raise rents up to the point where a TOD business is only marginally more profitable than a similar business at a more remote location. The goal of value capture should not be to tax “economic activity” (which when taxed tends to go elsewhere) but to tax transit-created land values. If land values are not taxed, they end up as a windfall to wealthy landowners. (An interesting but often misunderstood fact is that taxes on land values result in lower land prices and tend to incentivize development where land values are high – which is next to Metro stations and where we want development to occur.)

Second, auto use is heavily subsidized. Federal gas taxes have paid for interstate highway construction, but that is only a small part of the costs imposed by autos. State gas taxes do not make up the difference. For example, a very large share of the police, fire and EMS budgets of most localities are spent managing traffic and responding to crashes. Parking lots and garages are expensive – yet usually provided “free” to customers who drive (or their parking ticket is “validated”). The costs of parking lots and garages (and parking ticket validation) are then passed along to all customers – including those who walk or take transit. I don’t recall any store ever validating my transit pass. And finally, what is the cost of over 40,000 people killed each year in auto crashes (and countless more injured)? We would not tolerate that “cost” from transit or Amtrak or the airlines.

For more information, see

by Rick Rybeck on Aug 16, 2010 1:47 pm • linkreport

I forgot to note that the Metropolitan Washington Council of Governments has posted an article that I wrote about value capture. If anyone is interested, they can view it at

by Rick Rybeck on Aug 16, 2010 1:58 pm • linkreport

Thanks Rick. Sadly, my points on this didn't go anywhere in Baltimore County (I called for creating a transit value capture district and re-routing the light rail from Falls Road to serve Towson and then returning to the current alignment in Lutherville, with 7-8 stations in between). The "new" transpo element in the master plan is pretty average. And my point that as long as the transit _network_ in the Baltimore region doesn't really exist, it's hard for property adjacent to transit to develop the necessary level of added value that justifies significant intensification and mixed use projects.

by Richard Layman on Aug 16, 2010 8:48 pm • linkreport

Add a Comment

Name: (will be displayed on the comments page)

Email: (must be your real address, but will be kept private)

URL: (optional, will be displayed)

You can use some HTML, like <blockquote>quoting another comment</blockquote>, <i>italics</i>, and <a href="http://url_here">hyperlinks</a>. More here.

Your comment:

By submitting a comment, you agree to abide by our comment policy.
Notify me of followup comments via email. (You can also subscribe without commenting.)
Save my name and email address on this computer so I don't have to enter it next time, and so I don't have to answer the anti-spam map challenge question in the future.


Support Us