Amtrak’s Chicago-DC Capitol Ltd. crosses the Potomac at Harpers Ferry. Photo by Mr. T in DC on Flickr.

Amtrak’s Chicago-DC Capitol Ltd. crosses the Potomac at Harpers Ferry. Photo by Mr. T in DC on Flickr.

Amtrak’s federal grant, constituting just 0.05% of federal spending in 2010, is once again under attack. Its critics perennially point to the railroad’s 24¢ per passenger mile (ppm) government subsidy, compare it to the 2¢ ppm direct subsidy for driving, and call Amtrak a waste.

Comparing these direct subsidies, though, tells only part of the story. When indirect subsidies are considered, Amtrak’s total subsidy comes out to a little less than 44¢ ppm, but motoring’s subsidy rises up to almost 5645¢ ppm.

When considering all of the costs to society the argument for increasing Amtrak’s subsidy (and/or the gasoline tax) becomes clear. This table compares the direct and indirect subsidy of Amtrak versus roads, per passenger mile:

SubsidyAmtrakRoads
Direct subsidy  $0.240  $0.020
Air pollution  $0.081  $0.118
Global warming  $0.072  $0.109
Parking  $0.000  $0.151 $0.041
Resource consumption  $0.008  $0.040
Crash damage  $0.007  $0.037
Congestion  $0.000  $0.023
Lost tax revenue  $0.006  $0.028
Land use  $0.018  $0.020
Noise  $0.006  $0.008
Transportation diversity  $0.000  $0.004
Total  $0.439 $0.557 $0.447

All values are adjusted for inflation and reported in 2010 dollars. Most of the road values are from 2007 but since then, driving is down more than 15% and road user fee receipts are down as well.

Here is how each cost contributes to the total, and how I arrived at each estimates:

Direct subsidy: In 2010, Amtrak received $563 million in operating subsidies and $1 billion in capital and debt service grants while carrying passengers a total of 6.52 billion passenger miles, for a direct subsidy of 24¢ ppm. Meanwhile, in 2007 roads received a $94.6 billion (49% of $193 billion) subsidy to carry automobile passengers (including drivers) 4.24 trillion miles for a direct subsidy of 2¢ per passenger mile.

This is where a lot of people like to end the story. And doing so makes Amtrak seem like a bad deal, but if we consider all the external costs of both, it becomes a very different story.

Air pollution: According to McCubbin and Delucchi (1996), the external cost of air pollution caused by driving was $0.112 per vehicle mile traveled (VMT) in 1990 dollars. According to the Federal Highway Administration, there are approximately 1.59 passengers per vehicle. So dividing the VMT by passengers per vehicle and adjusting for inflation gives a value of $0.118 ppm.

While a similar analysis is not available for Amtrak, we do have the energy intensity ppm for passenger cars (3501 Btu ppm) and Amtrak (2398 Btu ppm) which can serve as a reasonable proxy for air pollution emissions. If anything, this ratio is unfair to Amtrak, since the bulk of its energy use (carrying Northeast Corridor riders) is electricity which produces fewer emission per BTU than gasoline does due to the use of nuclear, hydroelectric, natural gas and renewables. Using this ratio, the cost of air pollution caused by Amtrak is $0.081 ppm.

Global warming: In addition to air pollution, the burning of fossil fuels like gasoline or coal creates an assortment of greenhouse gases which have been shown to add to global warming. According to the Victoria Transport Policy Institute (VTPI), the external cost per vehicle mile of GHG creation is $0.164 in $2007 which is equivalent to $0.109 ppm in $2010. Meanwhile, Carbon Fund calculates rail travel as creating 0.42 lbs CO2 per passenger mile.

The EPA calculates that the annual emissions from a typical passenger vehicle are 5.5 metric tons of carbon dioxide. Dividing that by the 12,000 miles the average car travels in a year (also from the EPA) results in 1.01 lbs per vmt or 0.64 lbs ppm. Using the lbs ppm to create a ratio (0.42/0.64) and multiplying it by the value for cars gives us the value for rail travel of $0.072.

Parking: While some parking costs are paid by users, others are external because, for example, of tax-exempt “free” employee parking or because of municipal parking minimums. Shoup (2005) sets the external cost of parking at $0.22 $0.05-$0.14 per vehicle mile in $2005. VTPI determines a value of $0.062 pvm. Accounting for inflation and passengers per vehicle this becomes $0.151$0.041 ppm. Amtrak, on the other hand, pays for its rail yards where it parks its trains, which means its parking is covered in its direct costs. Its subsidy is $0 ppm.

Resource consumption: The external costs of resource consumption refers primarily to the cost of producing, importing and distributing petroleum that are not passed directly to users. It includes the military costs associated with protecting oil supplies, macroeconomic costs of oil dependence, uncompensated ecological costs (like loss of a species), subsidies for drilling and costs associated with depleting a non-renewable resource. VTPI prices this at 76¢ per gallon or 4¢ ppm (in 2010) for roads.

Amtrak, meanwhile, used 62 billion gallons of diesel in 2009 to carry passengers 5.914 billion miles. Which puts its resource consumption external cost at $0.008 ppm in $2010. [(62B gallons*$0.80 per gallon)/5.914 billion miles]

Damage from crashes: While some crash costs are internal to the user, like insurance premiums or personal property damage, others are external because they’re uncompensated or paid by an external user, such as when a pedestrian is injured in a hit and run crash. Looking at a series of studies and estimates of the external cost of automobile crashes, VTPI determined that the external cost per vehicle mile is $0.055 in $2007 which becomes $0.037 ppm. 

Miller, Douglass and Pindus’ 1994 paper on railroad injuries determines that per passenger mile, train injury costs comprise only one-fifth those of cars. There’s reason to believe that less of the crash cost for rail is external than it is for autos because of Amtrak’s greater ability to absorb such costs than an individual motorist. External Cost of Transport: Accident, Environmental and Congestion Costs in Western Europe puts the ratio for external costs of auto vs passenger rail at a much higher 41:1, so we can safely use the 5:1 ratio, making the external cost for rail $0.007 ppm

Congestion: Traffic congestion carries a very real transport cost that consists of incremental delay, vehicle operating costs (fuel and wear), pollution emissions and stress that result from interference among vehicles in the traffic stream. VTPI estimates the cost of congestion for roads at $0.035 pvm in $2007, which becomes $0.023 ppm in $2010.  Unlike the highway system, most of our rail network is below capacity (Figure 13) which means Amtrak causes very little rail congestion.

While there has been little attempt to quantify the congestion Amtrak might cause, there is plenty about how it reduces congestion on roads and at airports. A USDOT study, in fact, determined that expanded rail service would have a net positive effect on congestion.  And an Amtrak study put the benefit of the reduction of airport congestion along the Northeast Corridor alone at $104 million. Therefore I’ve conservatively set the Amtrak congestion charge at $0.00.

Lost tax revenue: Most of Amtrak’s train routes use railroads owned by freight companies for which Amtrak pays rent and the owners pay property, sales and franchise tax. If roads and highways were similarly owned by private companies, they too would pay billions of dollars in taxes.  TeleCommUnity set the value of the land used for roads in America at $7.1 trillion in $2007. Taxed at the average state property tax rate of 1.38% in $2010, roads would have contributed 2.8¢ ppm to states’ coffers.

Before being made exempt in 1979, Amtrak paid state and local taxes which totaled $14 million in 1979. Adjusting for inflation and dividing by total passenger miles, lost tax revenue from Amtrak is 0.64¢ ppm.

Land use: ecological impacts: Building roads and railroads has an ecological impact that also carries a cost. These land use changes can, for example, cause a heat island effect, sever and fragment wildlife habitat and result in “roadkill.” VTPI puts the cost of these ecological impacts for roads at $0.03 per vehicle mile in $2007 (or $0.0198 ppm in $2010).

Using the rail and road costs ppm defined in a 2005 Swiss ARE paper (from VTPI), we can determine a ratio between road and rail which is 0.7 centimes to 0.775 centimes per passenger kilometer (1.2 per vehicle kilometer/1.59 passengers per vehicle). This means that the ecological impact of passenger rail is 92.75% the value of the one for roads ppm. Multiplying this by $0.0198 gives a value for rail of $0.0184.

Noise: Noise from roads and rail can have real and measurable impacts on nearby land values. VTPI looked at several studies of the external cost of road noise and determined a value of $0.011 pvm ($0.0075 ppm in $2010). A study by INFRAS/IWW placed the ratio of ppm passenger rail noise costs to roadway noise costs at 3.9/5.2 (page 74). Multiplying this ratio times the cost for autos gives $0.0057.

Transportation diversity: Many communities are automobile dependent. This lack of diversity can often result in inefficiency and inequity, such as when people feel the need to drive for very short trips because they can’t easily walk or bike on roads that don’t accommodate that kind of travel. This eliminates options, leads to less physically active (and therefore less healthy) lifestyles, and can often trap people who can’t, for whatever reason, drive. VTPI determines the value of the impact of roads on transportation diversity at $0.007 pvm ($0.0044 ppm). Amtrak doesn’t create the same kinds of dependency, so it’s cost is zero.

I omitted some costs because adequate values were not available for both. This includes the delays, discomfort and lack of access that vehicle traffic imposes on nonmotorized modes; waste disposal; water pollution and hydrological impacts; and other land impacts such as reduced property values, reduced community cohesion, and the costs associated with sprawl. Together these constitute another 4¢ ppm for roads, with most of that cost related to sprawl.

Amtrak is unlikely to have a larger ppm cost for these factors. Amtrak isn’t considered an engine for sprawl. Railroad surfaces are capable of absorbing twice as much rainwater as paved roads —and there is far less land dedicated to Amtrak rail than to roads per passenger mile. So it is reasonable to believe that the addition of these omitted values would not change the relative values between the two subsidies very much or change them to be further in favor of Amtrak.

Based on these numbers, it appears that roads are subsidized at almost 12 1¢ per passenger mile more than Amtrak. To bring the two into balance, the gas tax would either need to be raised from the national average of 48.1¢ per gallon to about $3.30 $0.96 per gallon, or the Amtrak subsidy would need to be increased from $1.565 billion to $2.348$1.630 billion.

Another way to think about this is to combine the subsidized cost with the portion paid by user fees to get a total public cost. (Drivers’ user fee revenue in 2008 was $94.512 billion and miles travelled was 4.9 trillion passenger miles for an inflation adjusted contribution of $0.020.)

Total public cost, unpaid + paid
Amtrak: $0.439 + $0.318 = $0.757
Roads: $0.557 $0.447 + $0.020 = $0.577 $0.467

Which means that Amtrak users are paying 44.5% of their public costs, while drivers pay only 3.6% 4.3% of theirs.

To be clear, this does not mean that motoring is cheaper. Drivers are also paying their own internal costs (purchase and maintenance of a car and insurance) which the IRS estimates to be 51¢ per mile total (or 48.5¢ per mile, less tax). That increases the total cost of driving to $1.062 $0.952 per mile. That’s without considering the travel time costs. Comparing all the costs, Amtrak becomes much cheaper than motoring.

We should not ignore the environmental, political, human and other non-obvious costs of transportation when discussing how to fund it. By focusing only on the direct costs, as many choose to do, we run the risk of making the wrong decisions. While a more thorough scholarly analysis would surely come up with different values and totals than my amateur one, it’s more than possible that Amtrak is the bargain paying most of its own way, and roads are the resource-consuming boondoggle that need to have their subsidy cut.

We've just launched our brand new website and are working out some kinks. Find something that looks like a bug? Please help out by sending us an email with the details!

David Cranor is an operations engineer. A former Peace Corps Volunteer and former Texan (where he wrote for the Daily Texan), he’s lived in the DC area since 1997. David is a cycling advocate who serves on the Bicycle Advisory Council for DC.