Safety and Statistics:
A Glimpse at What’s Clouding the Issue of Light Rail

Questions of long term developmental policy are inherently suited for political conflict because they involve differing goals—in Minneapolis, the differences between left wing and right wing visions of the city in 30 years are brought to the surface in the light rail debate—and also because there is valid controversy over what goals can be attained by which methods. Some questions appear much more straightforward. Yet, the public is exposed to countless claims, either contradictory or just seemingly contradictory, regarding the benefits or lack thereof of light rail.

A community group in favor of building a new line in Los Angles County writes “Which feels safer to you: your teenager going downtown or to the beach in a car – or in a light rail trolley? Remember, 1,766 people were killed by traffic in Los Angeles County in 1999, including 214 pedestrians. Light rail is safer than automobile traffic.” A writer for the Thoreau Institute quips “Is light rail safe? Absolutely not. Fatalities -- mostly to pedestrians -- per million passenger miles are much higher from light rail than from buses or automobiles.” And again, “Fatalities per passenger mile suffered on urban freeways are half those on city streets and just a third of those caused by light rail.” In contrast, a Salt Lake newsletter claims that “Federal statistics show light-rail systems have far fewer accidents and fatalities per passenger than public bus systems.” An opposition group to the Kansas City light rail system puts forth the claim that “Every rail line has suffered numerous accidents and fatalities…For example, the Blue Line in LA County…a 22-mile commuter rail line, has averaged some ten to twelve fatalities every year. Since the rail line opened, there have been approximately 60 deaths…The last accident killed a 63-year-old woman who was cut up into three pieces…Why would anyone bring any additional public safety challenges to the public?” A fact sheet from makes the irate claim that “A chart of fatalities related to light rail lines (presented in an anti-light rail paper) distorted numbers and drew a simplistic conclusion that light rail is unsafe in Los Angeles. There was no mention of the fact that about 1/3 of the fatalities were suicides.”

The issue of whether light rail or automobiles are safer is one you might think would be fairly easy to resolve based on freely available federal traffic statistics. At worst, you might get a result something like “light rail is more likely to injure people but cars are more likely to kill people.” Yet, in this debate even the simplest facts are far from clear cut. Take a heavily visited transportation site closely affiliated with the Wendell Cox Consultancy. The Urban Transit Fact Book created by this web site includes the following bar graph:

Intuition tells most people that forms of transport which employ tracks and professional operators and are therefore less prone to human error should be much safer overall. Yet in this graph and on tables provided with it, the figures clearly show that transit by light rail is about 75% more likely to result in a death for every mile traveled. The information here is not hearsay; it is “Compiled from United States Department of Transportation data.” Why, then, are there people claiming that light rail is safer?
The release of this sort of statistical information to the public from the United States Bureau of Transportation falls to the Bureau of Transportation Statistics. The National Transportation Statistics 2000 report from that bureau provides extensive information on transportation fatalities and passenger miles traveled on various modes of transport, among many other things. A glimpse through this report, however, reveals how many distinctions there are with the potential to confuse the issue. To start with, as noted above, if we wish to obtain a meaningful statistic on safety we must decide whether to be concerned with all injuries or only fatalities, or to put a weight on each, or count the total number of accidents. We must decide whether to look at vehicle miles or passenger miles (for most purposes, passenger miles are more relevant). We must decide whether to include statistics for light rail, heavy rail, or both, and likewise whether light trucks and/or heavy trucks are to be included along with passenger cars in determining how safe the highways are. We must decide whether to include only fatalities or injuries pertaining to occupants of the vehicles, or also include people killed or injured by collisions with vehicles. We must decide whether to count only “accidents” (which result from vehicular misuse or malfunction) or to count all “incidents” (which include such occurrences as station fires and maintenance mishaps).

The Urban Transit Fact Book chooses wisely to look at fatalities per passenger mile rather than per vehicle mile. In examining fatalities (deaths) only, it avoids the ambiguity of the term “injury”, but runs into the difficult point that since light rail is not being implemented very widely at present, and its routes are far less extensive than roads, there are relatively few statistics available at present, and this problem is greater for fatalities than for injuries or accidents. U.S. deaths due to light rail rose by 366% between 1997 and 1998; this statistic may prompt the thought that some horrible policy changes took place in that year until we learn that the relevant figures are 3 and 14 respectively. Accordingly, by examining fatality rates and making careful choices about what exactly to report, one can construct widely different views on the relative safety of light rail.

Not surprisingly, the Urban Transit Fact Book chooses to count all incidents rather than only accidents, even though most incidents aside from accidents are not attributable to the mode of transport involved. This decision boots the total LRT fatalities between 1990 and 1997 by over 25%. More reasonably but also in support of its agenda, the Fact Book reports data of non-occupants as well as occupants. (Recall the Thoreau Institute’s concern with unreported pedestrian deaths.) Anti-rail assessments often emphasize that while light rail may be safer for its passengers than are automobiles, it is less safe for pedestrians and cyclists. This concern is relevant for someone deciding whether to support the construction of light rail in his or her city, and like the difference between accidents and incidents, the low number of LRT fatalities make the decision whether to include pedestrian deaths a highly important one.

The strangest aspect of the Urban Transit Fact Book’s data is a disclaimer at the bottom: “Urban Auto and Truck passenger miles estimated using ratio of urban vehicle miles to total vehicle miles.” Strange to begin with because it isn’t at all obvious how knowing what percentage of automobile and truck use takes place in urban areas will help us calculate how many people are in the vehicle for the average trip (which is, after all, the difference between vehicle miles and passenger miles). Stranger still because the information in question is freely available from the same Bureau of Transportation Statistic report as the information on fatalities. Moreover, while the Fact Book does not directly report its estimates on passenger miles, the statistics given by the Bureau’s report are significantly different from what those estimates would have to be to give the results listed. Furthermore, while the Bureau’s report (Table 2-1, Transportation Fatalities by Mode) lists numbers in excess of twenty thousand annual fatalities for passenger car occupants alone, the Fact Book lists numbers well under twenty thousand for combined Truck and Auto fatalities. It isn’t revealed from which Department of Transportation report this data was obtained, but given the unusual method used for estimating passenger miles and the utter straightforwardness of Table 2-1, the natural inclination should be to distrust these figures. The following simple table incorporates fatalities resulting from accidents (not incidents) on light rail and passenger cars, including non-occupant fatalities, and using figures from the National Transportation Statistics 2000 report:

  90 91 92 93 94 95 96 97 90-97 Average
Car (thousands) 24.1 27.4 21.4 21.6 22.0 22.4 22.5 22.2 23.0
Light rail 5 11 6 14 10 10 5 3 8
Passenger miles (100 millions)
Car 22800 22000 22100 22100 22500 22700 23400 23900 22700
Light rail 5.71 6.62 7.01 7.05 8.33 8.60 9.57 10.35 7.91
Fatalities per 100 million passenger miles
Car 1.06 1.24 0.96 0.98 0.98 0.99 0.96 0.93 1.01
Light rail 0.88 1.66 0.86 2.00 1.13 1.16 0.52 0.29 1.06

According to this analysis, light rail is indeed less safe than automobiles, but only barely. Given the low number of light rail fatalities and the fact that LRT as a mode of transport is just beginning to make a comeback in the context of smart growth, it is quite conceivable that the results could be reversed (or alternately, augmented) in a matter of years. Furthermore, there are many more sides to the safety debate, among other things including similar analyses using injuries and total accidents. The lessons to be learned are twofold: firstly, at present the death rates for light rail and automotive travel are comparable, but secondly, any analysis based on fatalities alone as they relate to LRT should be scrutinized carefully. The broader lesson is that in a complex, real world issue, there are usually more angles to any aspect of the problem than are apparent at first glance. Overall judgments are not to be made lightly.