Sunday, April 08, 2007

Economics of transit vs. cars

I came across this pretty interesting analysis of the cost per passenger mile of using a car vs. transit. Cars turn out to be surprisingly efficient, on both a cost and pollution basis. Cars work out to about $0.22 per passenger mile, while buses are about $0.92, commuter rail is $0.65, and light rail is a hefty $2.03. Even the often demonized "highway subsidy" only adds 0.38 cents per passenger-mile, bringing it to $0.224. When all the transit modes are blended together:
"So transit costs 84 cents per passenger mile, nearly four times the cost of driving. Transit fares average 20.4 cents per passenger mile (interesting how the transit industry keeps fares close to the actual cost of driving), so the subsidies are nearly 64 cents per mile."
And after factoring in "social costs" like pollution...
"So transit costs are nearly three time auto costs, while transit subsidies are more than eight times auto costs."
His conclusion:
This doesn’t mean transit is evil. What this should suggest instead is that the transit industry needs to make itself more efficient. This could be done by introducing competition into the industry or, at the least, contracting out transit to private operators who spend about 60 percent as much, per passenger mile, as public agencies. It goes without saying that an efficient transit industry would not sink billions into rail capital improvements when buses can do everything rail can do at a far lower cost.
My thoughts? I do think it's good to debunk the common myths of transit cost-effectiveness and environmental benefits vs. cars. But there are caveats.

First, this is a blended average of transit costs. Even the bus number is mostly based on substantially-empty buses covering local routes. I'm sure nearly full express commuter buses are far more efficient than cars on a per-passenger-mile basis. They also use limited freeway capacity (HOV/HOT) more efficiently during peak hours. So I think that's still a very legitimate transit strategy, albeit one where his suggested privatization could bring improvements, not just in cost, but in schedules, service, and routes.

Transit's core mission to provide mobility to those without it, whether they can't drive or can't afford a car, is also still a perfectly legitimate social good, regardless of this analysis. Yes, it's not cheap (mainly because the buses aren't full enough at the required frequency and coverage), but we've still made a choice as a society to provide it as the transportation option of last resort with comprehensive coverage.

I think it makes a valid point on rail, especially commuter rail in modern, dispersed, auto-centric cities like Houston (remember, that "almost reasonable but still too high" $0.65 number above is for cities designed around commuter rail like NYC, Chicago, Boston, and DC - it would be far, far higher here). My thoughts on Houston's core LRT/BRT plan are more nuanced. The economics clearly don't justify it. So it's essentially a city amenity - like sports stadiums or parks or libraries - that the citizens/voters choose to provide because, well, they want it. For tourists. For the "urban lifestyle." To support a dense core of pedestrian-oriented neighborhoods. To enable errands, lunch, and outside meetings during the day for those who use commuter transit. Whatever the reasons, they voted for it, and the economics are only a piece of the equation, same as for other subsidized amenities like stadiums, parks, and libraries.


At 1:54 AM, April 09, 2007, Anonymous Anonymous said...


No offense, but you should know better than to accept everything you read at face value.

While I don't have the data available off-hand, nor the time to find it, to check the numbers for transit costs, a quick check of automotive costs reveals that these numbers are complete BS.

For starters, we might look at the fact that the IRS credits vehicle mileage to the tune of $0.485/mile, and it has long exceeded $0.30/mile.

Let's take a look at the numbers, though.

If we take a typical sedan getting 25 mpg and a typical fuel price of $2.50/gallon, this gives us a fuel cost alone of $0.10/mile. Given the stop-and-go traffic on Houston highways, many autos are not going to average 25 mpg.

If we look at the typical vehicles on Houston roads, of which about half are pickups or large SUVs, it becomes clear that 25 mpg average is highly optimistic.

For the typical truck and SUV, we can expect about 15 mpg, which gives up $0.16/mile.

Now if we look at the actual cost of the vehicle, if we assume a conservative note of $375 per month and $75/month for insurance, this gives us $450/month in fixed cost. For a vehicle driven 1000 miles/month (12,000 annually), we end up at $0.45/mile + fuel, for a total of $0.55/mile for an auto.

At this figure we see validation for the IRS mileage figure, with a typical IRS conservatism.

This simple calculation neglects tire wear, brake wear, pollution impact, highway subsidies, military costs in the Mideast, etc.

If the car is driven less than 1000 miles per month, the cost per mile goes up further.

If we assume that the vehicle is an old vehicle that is paid off already, then the repair costs go up dramatically, as does the exhaust emissions and pollution impact, as compared to a newer vehicle.

For a city like Houston where auto and truck emissions contribute substantially to pollution loads, especially for older vehicles, it's important to have the newest possible vehicles on the road to minimize our pollution problems.

FWIW, I do agree that transit agencies have a long ways to go to actually come up with public transportation that makes sense economically, from a standpoint of time efficiency, and also from pollution impact.

Best regards.
Sean Thomas

At 6:47 AM, April 09, 2007, Blogger John said...

It's an interesting attempt to attack numbers to some hard to measure items, but not a very good one. Do the pollution costs include all pollution related illness? Do the car subsidies include the cost of storing cars in densely populated areas, resulting in use of prime real estate for fairly unproductive purposes? Etc., etc.

Transit becoming more efficient is certainly a good thing; I'm a bit skeptical about privatization as the way to accomplish that, based on the attempts we've seen out there, which have yet to show much benefit.

At 7:08 AM, April 09, 2007, Blogger Tory Gattis said...

Sean,I'm not really in a position to argue numbers. The stats quoted seem pretty solid. I think the IRS number is oriented towards fully-loaded cost of new cars - maybe even luxury ones. The vast majority of people drive older, less-luxurious cars with far less depreciation. Also, don't forget he's looking at passenger miles, not vehicle miles, and the average car trip has 1.6 passengers. That alone gets your $0.55 down to $0.34.

In any case, all of the numbers are substantially less than transit.

John, his model is what it is. Not sure about the pollution factor, but transit can be just as problematic here (diesel buses, coal-fired plants for electricity). As far as parking real estate, people pay for that (one way or another), so it's not a subsidy, but I'm not sure it's in the model. It said "operating expenses", so that would seem to include explicit paying for parking - but probably not the implicit cost (like when an employer or business pays for it).

As far as privatization, my only support for that is express commuter services (buses, vanpools), and even then with per-psgr-mile subsidies. I agree the local lines and rail would be much more problematic.

At 10:36 AM, April 09, 2007, Anonymous Anonymous said...

For your statistics about cities designed around commuter rail, you forgot an important very high expense - parking. In Chicago, parking downtown is about $250/month.

At 10:56 AM, April 09, 2007, Anonymous Anonymous said...

I should have checked something else before I posted the first time. I don't know where you get $0.65 per passenger mile for commuter rail, but I have to question it. Chicago's commuter rail system (Metra, the suburb-to-city system, as opposed to the intra-city CTA, which is a different agency) provides 2.055 billion passenger miles per year at an operating cost of $554 million dollars, or about $0.27 per passenger mile.

At 11:10 AM, April 09, 2007, Blogger Tory Gattis said...

Matt: I believe the cost estimates include amortized capital costs, not just operating costs.

At 11:35 AM, April 09, 2007, Anonymous Anonymous said...

I typically tend to be highly skeptical of any sort of "analysis" proferred by an extremist like Randal O'Toole.

At 12:31 PM, April 09, 2007, Anonymous Anonymous said...

Where does the total cost come from?

By the time you consider the cost of purchasing a vehicle, housing it, the insurance, gasoline, maintenance expenses, parking meters/garages, etc. and add in environmental costs and taxes to support the never ending desire for more roadways, public transportation looks better and cheaper per capita overall.

At 12:44 PM, April 09, 2007, Blogger Tory Gattis said...

You can take a look at his data sources. They seem like pretty solid govt stats to me.

At 2:19 PM, April 09, 2007, Anonymous Anonymous said...

Randal O'Toole has been debunked so many times its not even funny. Of course he gets things from government stats but then twists them to get the information he wants. Lies, damn lies and statistics. Lots of his rail stuff has been debunked at and Randal O'Toole is a joke paid by the oil companies to come up with this number crunch

At 4:08 PM, April 09, 2007, Anonymous Anonymous said...

Is incremental commute time included in this? I think the average hourly wage in the US is about $17.50/hour. If it takes an extra 30 min/day to commute on the average 20 miles, then you can add $.43/passenger mile to the cost of transit.

At 6:27 PM, April 09, 2007, Anonymous Anonymous said...

Yet another personal anecdote:

My car is fully paid for, 17 years old, and averages 30 mpg for a normal tank's worth of driving. It costs me $633/yr for insurance, $467/yr for gas, and about $1000/yr (some years more, most years much less) for tires, repairs, registration, inspection. I drive about 5000 miles per year. That comes out to $0.42/mile, only $0.093 of which is gasoline.

If I took the bus and train to/from work, that would triple my 12-15 minute commute time and would cost me $0.80/mile to boot.

I'm sticking with my car until our stupid building management lifts its ban on bicycles in the building.

At 10:41 PM, April 09, 2007, Anonymous Anonymous said...

I once had an environmentalist friend in college who refused to get into my car when we were going to the dining hall because she thought it was wasteful and frivolous to take a car when a bus would take us there. She walked over to the bus stop, and the rest of us took off. Thirty minutes later, we greeted her when she came into the dining hall.

I'd love to show her these stats and point out that she actually hurt the environment more by supporting an inefficient public transit system.

At 11:52 AM, April 10, 2007, Anonymous Anonymous said...

*I believe the cost estimates include amortized capital costs, not just operating costs.*

Then the comparison isn't apples-to-apples unless you include the amortized cost of road construction and maintenance on the car line.

However, even though it's beginning to look to me like the basis of this discussion is stacked by making assumptions which put transit in the worst possible light while putting cars in the best, I'll play along...

Metra capital expenditures for 2007 (projected) are about $225 million. Add that in, it brings us to $0.38 per passenger mile - for an actual large-city transit agency which provides over 200,000 people with their primary mode of transportation to work each day, not a strawman.

At 1:23 PM, April 10, 2007, Blogger Tory Gattis said...

I believe the car side does include spending on roads, although it should be noted (and a point I've made before): a road network is non-optional for any city of any kind, if only for fire, police, ambulance, and freight deliveries. Even the most transit dependent cities in the world have a comprehensive road network (albeit a slightly smaller one than auto-dependent cities in lane-miles per capita). So lumping in all roads with car costs is not exactly fair. That's why most local roads are funded by property taxes, not gas taxes - because every citizen requires access to the road network. What we really need to add to cars is the marginal cost of the "extra" road network required above and beyond the basic grid.

The Chicago example is interesting - clearly they're more efficient than the national average - although looking at a one year capex is not the whole picture, since it's almost entirely maintenance, and doesn't include depreciation of the original construction cost or trains. It's like looking at Metro's 2006 line item for the LRT, when the real bulk expense was 2002-2003 construction and buying the train cars. I will admit roads have some similar accounting issues, although less, I think, because govt doesn't pay for the vehicles, and the broad national annual number for road construction/maint is reasonably steady (thus including road renewals/rebuilds), whereas commuter rail construction is much more lumpy. Most lines were build years and even decades ago. Very few, if any, are built today (at least in America). In other words, looking at a recent annual number for all roads is probably representative of overall costs, including capital expense, but not necessarily for heavy commuter rail.

At 9:04 PM, April 10, 2007, Anonymous Anonymous said...

*The Chicago example is interesting - clearly they're more efficient than the national average - although looking at a one year capex is not the whole picture, since it's almost entirely maintenance, and doesn't include depreciation of the original construction cost or trains.*

Actually, Chicago has just finished adding a second track on one line and is in the middle of an engine- and passenger car buying spree, replacing old stock.

At 9:09 PM, April 10, 2007, Anonymous Anonymous said...

Oh, and another line was recently extended. The annual capital expenditure of something north of $200 million is representative of the past few years and of forward projections.

At 6:08 PM, April 17, 2007, Blogger strfsh said...

I'd question the use of "per passenger mile" as a measure. What about "per trip." What if people taking the train aren't going as far on average? Certainly people using mass transit in a central city area may only be traveling a few miles but still get where they need to go and it's "clear on the other side of town" functionally and psychologically. In addition, the fact that they commute to work on transportation probably means they walk on their lunch break and possibly during other trips at their home base. Some studies have shown that a large proportion of VMT comes from non-work trips.

Furthermore, a transit commuter population enables the kind of dense urbanism that allows other people to commute by bike or foot. Personally, the fact that I have public transit (and zipcar) as an option when I need it, allows me to walk and bike 90% of the time whereas if I invested in buying a car I would probably use it more than the 10% that it's really necessary because the marginal cost (mostly gas) is fairly low or at least untallied (increased maintenance); not to mention, I might live in a different location altogether (with free parking) geographically farther from my destinations.

At 11:07 PM, April 20, 2007, Anonymous Anonymous said...

I just hope we're not advocating making policy based on averages. Obviously, there are transit systems that are placed to maximize their ridership per mile, and others that are not. There are cities that are compact and cities that are not. Randall O'Toole's averages continue to just muddy the picture.

Houston needs to keep poking through the best practices of major cities, to cherry pick the best strategies among them. Houston is too big to be one-size-fits-all.

We need high-capacity transit in the urban areas where we have the density to support it. We need to encourage compact development where transit is not yet practical, to give people more options to live closer to where they work. And we need to dramatically improve connectivity for bicycles and pedestrians across the city. Houstonians are clamoring for more options and they won't come from averages.

At 4:29 PM, April 27, 2007, Anonymous Anonymous said...

"Chicago has just finished adding a second track on one line and is in the middle of an engine- and passenger car buying spree, replacing old stock."

So why does Chicago, with all it's rail, have over 7,000 cabs?

At 11:12 AM, May 07, 2007, Blogger Unknown said...

On the cab comment, I'm not sure where I saw this, but transit development is correlated with increases in taxi riding. When cities build effective transit, people become less dependent on cars and walk more, as noted above. But when you are a transit rider, it doesn't mean that you always have to ride the bus. So whenever you need to rush somewhere directly in a hurry or are caught out late somewhere, you hail a cab.


Post a Comment

<< Home