How to Review a Rail Project

Light rail is good for two things and two things only: getting federal pork barrel and creating an excuse to rezone neighborhoods to higher densities. Otherwise, all of the benefits claimed for light rail are either outright lies or could be gain at a much lower cost by expanding bus service. Light rail is an obsolete technology and cities that invest in it are wasting their taxpayers' money.

Heavy rail and commuter rail may make a little more sense in the right situations--but those situations are very rarely found in U.S. cities, and all of those already have rail lines. Proposals to add new rail service in cities that do not already have it are as foolish as light rail.

Fortunately, all of the information needed to show the foolishness of rail proposals can easily be found in the documents that cities are required to prepare to get federal funding for rail projects. These documents include environmental impact statements and the new starts criteria submissions. Whenever these documents claim some benefit for rail, there are two questions you should always ask:

  1. What is the cost of achieving that benefit?
  2. How much would it cost to achieve the same benefit by simply improving bus service?

Here are some of the things you should look for when examining either of these documents:


There are numerous alternatives to rail, most of which center around improved bus services or highway capacity expansions. The Federal Transit Administration requires agencies to consider at least three alternatives:

  1. A do nothing or "no-build" alternative;
  2. Transportation system management (TSM), which usually consists of low-cost improvements to bus service such as increased frequencies or express buses;
  3. The rail alternative.

This limited selection far from exhausts the possibilities for alternatives to rail. Based on this short list, planning documents will often say things like "the rail alternative consumes the least energy" or "produces the least air pollution." But that is only when compared with just the other two alternatives. Numerous court cases have made it clear that if other alternatives could produce greater environmental benefits for the same or lower costs, they should also be considered in environmental impact statements.

Although the TSM alternative includes some bus improvements, it usually does not include running buses as fast and frequently as rail, so it will not attract as many new riders as light rail. Although rail advocates frequently argue that there is something "special" that makes rail more attractive to riders, the computer models used to estimate the effects of alternatives are based solely on frequencies and speed. So an alternative that proposes to run buses on rail schedules should predict as many new riders at a much lower capital cost as rail.

If you are reviewing a draft environmental impact statement (EIS) or making comments prior to preparation of a draft EIS, you should insist that a full range of alternatives be considered. At the very least, one alternative should consider the costs, ridership, and benefits of running buses with the same frequency and stops as the proposed rail line.

If you comment on the draft EIS and ask that reasonable alternatives be considered that are not included in the final EIS, you have the right to go to court and ask that the EIS be rewritten to include those alternatives. At the very least, this will help delay the project while you muster political opposition to it. But even if you are too late to influence the EIS, you can use the information in the alternatives that were developed by the agency to show that spending money on buses is far more cost effective than spending it on rail.

New Riders

Rail proponents often brag about the total number of riders they expect to use their rail lines. But the total is not important; what is important is the number of new riders--that is, the number of riders who would not otherwise have taken a bus or other form of transit. If the documents you are looking at do not separately display new riders, they can easily be calculated by subtracting the number of transit riders under the no-build alternative from the number under the rail alternative. Generally, only a small fraction of total riders will be new riders.

Capital and Operating Costs

The capital cost should be one of the most readily available numbers in the documents. The capital cost of the no-build alternative is probably zero, but if not it should be subtracted from the cost of the rail alternative so that the cost you are using is appropriate for the number of new riders. The same should be done with operating costs.

In estimating capital costs, transit agencies often play games with inflation factors. They may advertise that a particular rail line will cost $400 million "in 1998 dollars." In fine print, they say that, due to inflation, the actual cost will be $500 million "in year of expenditure dollars." But inflation has been running at around 2 percent per year, while the inflation factors they build in are often more than 5 percent per year. This adds tens of millions of dollars to the cost of their project and provides a healthy cushion for covering up cost overruns.

It is worth noting that the rail operating costs leave out some of the significant costs of a rail line. Most importantly, rail roadbeds must be reconstructed every twenty years or so. Overhead wires and electrical facilities will also need updating. These costs can be huge, reaching perhaps a quarter or more of the original construction cost. You probably won't find them in any documents, but you should know that they exist and you can ask the agencies for estimates of how much they will be.

FTA Cost Effectiveness

The Federal Transit Administration requires each transit proposal to include a calculation of cost effectiveness. This calculation involves three steps:

  1. The capital costs are annualized, meaning that a standard amortization formula is used to convert it to the amount of money that would have to be paid each year as if it were a loan being repaid.
  2. The annualized capital costs are added to the operating costs to get total annual costs.
  3. The number of new riders is divided into the total annual cost.

The result is a cost per new rider. If the documents have three alternatives--no build, TSM, and rail--then three different costs per new rider can be calculated:

  1. The cost of TSM vs. no build;
  2. The cost of rail vs. no build; and
  3. The cost of rail vs. TSM.

The FTA requires that the cost of rail be compared with no build and TSM; you will probably have to calculate the cost of TSM with no build yourself. Table one shows these calculations for the Hiawatha light-rail project proposed for the Twin Cities.

Table One: FTA Cost Effectiveness for Hiawatha Light Rail

(Ridership in millions, costs in millions of dollars except for cost per new rider in dollars)

                               Alternative           Cost Effectiveness
                                                    TSM    Rail    Rail    
                         No Build   TSM    Rail    - NB    - NB    - TSM   
Annualized capital cost    $6.43   $8.75  $46.40   $2.32  $39.97  $37.65
Annual operating cost     191.14  194.32  204.34    3.18   13.20   10.02
Total annual cost         197.57  203.07  250.74    5.50   53.17   47.67
Annual ridership           72.46   73.17   75.37    0.71    2.91    2.20
Cost per new rider                                 $7.75  $18.27  $21.67

Source: Metro Transit, "New Starts criteria materials for the Hiawatha Corridor Light Rail Transit project," September 1, 1999, St. Paul, MN.

The table shows that the cost of using TSM to attract one new rider is $7.75, while the cost of light rail per new rider is $18.27. Note that, if TSM were implemented, the cost of adding any new riders with rail jumps to $21.67. Still, rail is expected to attract four times as many riders as bus improvements. The obvious question: Why not spend four times as much money on bus improvements--which would still be well under half the cost of the rail line--to attract all of those new riders?

Calculating the cost per new rider still ignores the question of why we need to increase transit ridership at all. If the goal is to provide transit service to people who cannot drive or cannot afford autos, then agencies should adopt the most cost effective means of doing so. Rail lines serve only very narrow corridors in urban areas, while bus improvements can be made over much larger proportions of the urban areas at a lower cost.

Travel Times

Before 1997, the FTA cost-effectiveness index included another component: the amount of time saved by transit riders who switched from bus to rail. The Federal Transit Administration allowed transit planners to assume that this time was worth so many dollars per hour and to subtract that value from the cost of the rail line before dividing by the number of riders. Since 1997, time saved has been a separate calculation.

To boost transit ridership, many cities have halted road expansions with the hope that increased congestion will make transit more attractive. But buses are caught in the same congestion as autos. This gives rail an apparent advantage when rails are not running in the streets. In Portland, a proposed light-rail line was expected to take 28 minutes to go from end to end. Express buses currently go the same distance in just 24 minutes. But planners predicted that increased congestion, due to their failure to add road capacity to meet demand, would increase bus times to 40 minutes. Thus, they were able to claim that light rail saved transit riders' time.

What transit agencies usually do not calculate is the time lost by auto users due to rail. In the case of light rail, this can be significant because light rail lines frequently cross auto traffic. Even if every auto loses only a few seconds, the total can add up to far more than the savings to transit riders simply because there are so many more auto passengers than transit passengers. Losses will be especially large when existing lanes of roads open to autos are replaced by light-rail tracks.

Even if rail does not interfere with street traffic, building rail lines poses an opportunity cost for auto drivers because it spends money that could otherwise be spent on road expansions. The only light-rail environmental impact statement in recent years that considered highway construction as an alternative was written for Salt Lake City's Interstate-15 corridor. It found that one dollar spent on freeway expansion would reduce congestion and save as much time as nine dollars spent on light-rail construction.[1] The state ended up both expanding the highway and building light rail, but the light rail is proving irrelevent in reducing congestion is its total ridership is less than the increase in transit ridership that planners anticipated.

Congestion Reduction

A more pertinent question to ask is how many autos will rail take off the road during rush hour? The documents may not break ridership down by time of day, but even if every new rider rides during rush hour, the number of cars taken off the road during the morning or afternoon rush hour will only be half of the number of new riders because the new riders will go round trip. In addition, the average car during rush hour may have more than one occupant (1.1 is a conservative number), so the number of cars taken off the road is a little less than half the new riders. The annual cost of taking one car off the road during morning or afternoon rush hour will be very high.

Some environmental impact statements may measure congestion using levels of service (grade A through F). When light-rail alternatives are projected to produce level of service F on some roads, they are clearly not reducing congestion.

Another way of looking at congestion is to compare the annual growth in regional auto traffic with the expected reduction in auto trips. Daily traffic in the Portland area is growing by about 380,000 auto trips per year, meaning that each day Portlanders take about 1,000 more trips than they did the day before. The proposed north light-rail line is expected to add 18,100 new riders, so the line reduces auto traffic by, at most, 18,100 trips per day. In other words, on the nineteenth day after the line is opened for operation, congestion will be back to pre-light-rail levels. Since the light-rail line is expected to cost hundreds of millions of dollars and take years to build, this is not a very good investment.

Neighborhood Redevelopment

Transportation planners will often admit that rail does nothing to reduce congestion, but say that the benefits for neighborhood redevelopment make it worthwhile. Yet the evidence is clear that very little redevelopment will take place without additional subsidies (see the Neighborhood-Redevelopment Myth). A little research should reveal whether your city or metropolitan planning agency has plans to offer such subsidies to developers.

Air Pollution

Rail transit is supposed to produce less air pollution than autos. In fact, Diesel-powered commuter rail transit can easily produce more air pollution than cars, particularly if the Diesel locomotives are not brand new and well maintained. Electric-powered rail transit may produce less air pollution, but the differences are often insignificant and could easily be overwhelmed by calculation errors.

Portland's north light rail is predicted to reduce hydrocarbons, carbon monoxide, and nitrogen oxides by less than a tenth of a percent. Planners did not calculate the effects of low-speed driving due to congestion caused by the light rail; this could easily produce more air pollution than the line is expected to save.

Energy Savings

Rail advocates often brag about how much energy rail transit saves over cars. But they fail to consider the energy cost of building rail transit lines. Since these lines are used by so few people, the energy cost per rider is high.

Portland's Metro says that the north light-rail line "would consume the least energy" of the alternatives it considered and that it would save "the equivalent of 7,875 gallons of gasoline per day." But it also calculates that the energy cost of building the line is equal to nearly 32 million gallons of gasoline. That means that it will take more than 170 years for the daily savings to equal the cost of construction.


Heavy rail is a pretty safe method of transportation because it is separated from pedestrians and autos. Commuter rail safety depends on local circumstances. But light rail, which often runs in streets and crosses the paths of pedestrians, is much more dangerous than buses or autos.

Auto critics frequently point to the total number of deaths due to auto accidents each year. While these deaths are certainly tragic, when measured on a per-passenger-mile basis, autos are much safer than light rail and buses are safer still. Before 1998, fatalities per passenger mile from light rail were higher than those of both autos and buses. Portland's new westside light-rail line, which opened in September, 1998, dramatically increased light-rail fatality rates by killing five people in a little over a year.

Highway Alternatives

To compensate for the cost of gold-plated rail lines and cost overruns, rail proponents tend to exaggerate the cost of highways. When addressing these claims, it is important to remember that, except for the New York City subways and port authority (PATH) trains, no rail transit system in America comes close to carrying as many people per mile as a single lane of a typical urban freeway. This means the appropriate comparison is not the cost of a freeway but merely the cost of a single lane (or less) of a freeway.

To support a proposal for Portland's proposed $1.5 billion south light-rail line, rail proponents asked the state Department of Transportation how much it would cost to build a freeway along the same route. Even though Portland's existing light-rail line carries less than a third as many passenger miles as a single lane of a Portland freeway, they asked the department to calculate the cost of a brand-new six-lane freeway. This cost totaled to $565 million, which was still only about half the cost of light rail.

To make the cost higher still, the department added the cost of rebuilding an entire second freeway, part of a third, and a new freeway bridge across the Willamette River--none of which had anything to do with the proposed light-rail line. This brought the total cost to $3 billion.[2] Light-rail proponents--including Metro and the Oregonian--repeatedly stated that this would be the cost of highways instead of light rail.[3] In fact, Metro's own documents showed that the cost of adding a single lane each way onto already-existing highways would be less than $100 million.[4] At less than a tenth of the cost of light rail, those two lanes would carry far more people than a rail transit line.


Rail projects tend to be so expensive and produce so few benefits that it is generally easy to demonstrate to anyone not motivated by pork or high-density zoning that they are not worth building. Efforts to critique rail projects will be most credible if they use the data collected and calculated by rail planners against them. That credibility will be put to the best use if critiques, with brief executive summaries, are widely distributed to the media and other opinion leaders.


1. Utah Department of Transportation, I-15/State Street Corridor Draft Environmental Impact Statement (Salt Lake City, UT: Utah DOT, 1990).

2. ODOT, "Six-Lane Freeway Scenario for the Inner Sunrise Corridor" (Salem, OR: ODOT, 1995).

3. Metro, "Benefits of the South/North Light Rail Project" (Portland, OR: Metro, 18 March 1998), 2 pp.

4. Metro, Interim Regional Transportation Plan, p. 5-20.

Thoreau Institute Home | Vanishing Automobile Web Tools