A Free-Market Energy Blog

Do Passenger Trains Save Energy? Another Look

By Randal O'Toole -- July 9, 2009

In response to my recent post about whether passenger trains save energy, a reader asked about the “life-cycle costs of both forms [rails vs. highways] of transportation.”

Such a life-cycle analysis was the focus of a study by researchers at UC (Berkeley), which I mentioned in the previous post (but incorrectly identified as UC Davis).

Their answer is that neither form of transportation is clearly superior to the other; it depends on such things as load factors. Japan’s bullet train from Tokyo to Osaka, which serves more than 60 million people in a 300-mile corridor, is probably more efficient than driving. But the same train from (say) Eugene, Oregon, to Seattle would probably be a huge waste of resources because it would not likely fill enough seats to justify the energy costs of construction.

The UC Berkeley study also demonstrates one of the perils of trying to do a life-cycle analysis. Such an analysis is extremely complicated, relies on an enormous number of assumptions, and one small error can completely cripple the results.

UC Berkeley Study: A Fatal Factual Omission

The UC study tried to consider everything from the energy costs of mining the coal needed to generate electricity to power light-rail trains to the volatile organic compound emissions from cleaning the carpets on the trains to the particulates generated during highway and rail construction.

The results can seem overwhelming. And yet, at first, it seems that clear winners emerge. At average passenger loads, rail transportation consumes the least amount of energy and emits the least amount of greenhouse gases per passenger mile. Driving in a sedan requires about twice as much energy and emissions; buses are somewhere between trains and cars; flying is about the same as driving a car; and SUVs and pick-ups are much more energy intensive than a sedan.

The authors emphasize that your mileage may vary. For example, the only rail transportation systems considered were in the San Francisco Bay Area and Boston. This doesn’t mean that rail transit would save energy in (say) Boise, Idaho, or Madison, Wisconsin.

Since I had already compared the energy costs of operating rail transit vs. driving, I naturally looked at that part of the Berkeley analysis to see how it compared with my own. Although the Berkeley analysis considered the entire life cycle, from manufacturing to parking (though not, curiously, disposal), the largest cost was operations. I quickly noted a discrepancy: the Berkeley estimates of the energy required to operate rail transit were only about one-third of my own. Since we both used the same source data, something was very wrong.

The clue was in the “about one-third.” Electric power is very convenient, but not very efficient because you first have to convert one form of energy — coal, petroleum, water power, etc. — into another. This typically results in huge losses. Transmission of electricity also results in some losses. In 2002, for example, America’s electric industry used 40.3 exajoules of power to deliver 12.5 exajoules of electricity to end users. At best, about two-thirds of the original energy is lost in generation and transmission.

The Berkeley study accounted for about 9 percent power lost during transmission, but failed to account for the much larger losses in generation. Correcting this error would nearly double the life-cycle energy requirements — and associated emissions — of electric rail transit.

The Berkeley life-cycle analysis also failed to take into account the ways in which various forms of transportation interact with one another. For example, major urban areas with bus transit systems typically operate scores of bus lines that connect suburbs with downtown transit hubs. A single rail line might replace a dozen or more bus routes. But that doesn’t mean the transit agency saves energy by running fewer buses. Instead, it turns what were formerly direct bus lines into feeder lines that connect individual neighborhoods with rail stations.

Human Preference

The problem is that people don’t like to change from bus to train. So many people who might have taken a bus downtown instead drive to the train station and take the train. This leaves the feeder buses running nearly empty. After building rail lines, places such as Salt Lake City and Houston have seen dramatic declines in the average number of people riding their buses — which means a large increase in the energy consumed per passenger mile. The result is that the overall transit systems use more energy per passenger mile after opening the rail lines than before.

Even if a complete life-cycle analysis revealed that rail transit is more energy efficient than driving, that doesn’t mean building rail transit is the best way to save energy. As Steve Polzin of the Florida Center for Urban Transportation Research points out in a recent article, America’s auto fleet could become dramatically more fuel-efficient in just a few years. According to table 2.13 of the National Energy Data Book, for example, between 1976 and 1996, the fuel-efficiency of the average car on the road increased by 55 percent. This is possible because the auto fleet turns over almost completely every 18 years.

In contrast, when a city installs a rail system, it locks in a technology for 30 to 40 years. Table 2.14 of the data book says that, between 1976 and 1996 (or 2006 for that matter), the average fuel-efficiency of rail transit actually declined.

It is clear that tremendous advances in auto fuel-efficiency are still possible. The average car on the road today gets about 22 miles per gallon, yet many get well over 40 miles per gallon. If fuel prices rise (or Obama’s fuel-efficiency mandates are successful), the American auto fleet in 2025 is likely to be far more fuel-efficient than the average rail transit line.

Airlines have also rapidly improved their energy efficiency. Table 2.14 of the data book says that, between 1970 and 2006, the energy efficiency of airline travel increased by 3.2 percent per year, while intercity rail travel increased by only 0.8 percent per year. Boeing promises that its next major plane will be 20 percent more energy efficient than its current models, and jet engine manufacturers have set a goal of  doubling fuel economy by 2020.

What this means is that any assessments of passenger rail against driving or flying must compare the efficiency of the planned rail system with efficiencies of cars or planes 20 years from now, not today. Yet the California High-Speed Rail Authority’s environmental analysis assumed that cars and planes would be no more energy efficient in 2025 than they are today. Similarly, the Center for Clean Air Policy’s report calculating that Obama’s high-speed rail plan would save 6 billion pounds of CO2 per year assumed that “relatively low fuel prices and a continuing trend of drivers switching to sport utility vehicles” would lead to negligible improvements in auto fuel economies.

Even if all of the analyses were correct, it still does not mean rail transit is cost effective. For example, some people think we need to cut our greenhouse gas emissions by 50 percent or more. The Center for Clean Air Policy estimates that high-speed rail would reduce emissions by 0.05 percent. Since Obama’s high-speed rail plan will cost roughly $100 billion, cutting emissions in half at that rate would cost $100 trillion. This makes high-speed rail an extraordinarily cost-ineffective way of dealing with climate change.

Conclusion: Life-Cycle Central Planning

Life-cycle analysis is a tool for central planners, and the real lesson here is that central planning is no more effective at saving energy than anything else. The details of the analyses are too subject to manipulation by special interest groups and politicians and planners beholden to those groups. Government should only make sure that the full costs of energy consumption are included in energy prices so that consumers will be able to make their own choices based on their personal priorities.


  1. Tom Tanton  

    Nice piece Randal. I’d add that “ensuring full costs of energy consumption are included in energy prices”–i.e. internalizing externalities–is at least as much a case of impossibility as is life cycle analysis (and the two are certainly related.) I also agree with your assessment of vintage-of-technology issues, but find that one true irony is the gerneral failure to recognize that new fossil fired power plants approach/exceed 55% thermal efficiency compared to the average (>30 years old) system average of “about a third.” The replacement of old plants with new plants across the board would do more than any law or regulation to reduce fuel use in all things electric, including electric trains, but the laws and regulations hold back any such grace, and instead push silly non-solutions that can’t power a train anyway.


  2. BCassidy  

    I accidently posted on your old post comparing trains.

    But I’d like a rail system. Not for energy saving purposes, but just to reduce the time I have to drive. If we had 200+ mph rail system, making it from Houston to Dallas or Baton Rouge would be much quicker, less stressful, and less deadly.


  3. Andrew  

    BCassidy-Unless I’m mistaken, the point of the post is that there is nothing wrong with rail as a personal choice-if people want to use it, they can. But that is very different either mandatory rail travel or public rail systems (the latter choice is essentially the same as the former-everyone is paying for it, so there personal choice is taken from them).


  4. BCassidy  

    Andrew – We have more or less already had our personal choice taken away from us. We have public highways, interstates, ect. The government has made the option of taking a car much better than taking a train. I have ridden a train from Texas to Florida on a few occasions, and the experience (personaly) was a little worse than driving. However, if I was sitting on a train doing 200+ I think I would much prefer it. If we had a better public rail system, then our tax dollars would at least give us a ‘choice’.


  5. Andrew  

    The solution there is in fact to eliminate the public highway system, not to introduce MORE public transportation (Dang that Eisenhower!)


  6. Bill Sell  

    Andrew, you don’t have to take down the highways and rebuild the whole transportation infrastructure. But rail would alleviate the need to constantly expand the highways, and thereby generate more congesting traffic, driving up the need to widen the highways. Some cities have taken down freeways and found less congestion, because that old grid is quite efficient. Efficiencies in locomotion will expand to fill the need in any transport engine. There is no way they can or will be limited to the automobile.


  7. Bill Sell  

    O’Toole has several disconnects in his argument. He surmises that management of buses will not change when a rail is added to the transit mix. This of course could happen in a less than brilliant management system, but if rail is added, it makes sense to look at bus route and tweak them to serve the new configuration. The big claim of bus-over-rail advocates is that rubber tire service is more flexible; O’Toole would like to ignore this claim when it is inconvenient to his argument.

    Rail is not locked-in technology. The rails are, yes in the ground, but the engines will continue to improve, or electricity might as well be added, constantly improving rail efficiency. O’Toole assumes cars efficiencies will improve and assumes rail efficiency will not improve.

    His favorite org, Center for Clean Air Policy, is a market-based advocacy group, and by that commitment to that mission, will argue against public investment, as does O’Toole’s Cato which has faith in market solutions, and is well funded by industries that stand to gain from the status quo.


Leave a Reply