The Light Rail Now Project can be contacted at:
Light Rail Now!
Issues of Accuracy and Impact on Congestion
American Public Transportation Association
Rail Transit Conference
Boston, Massachusetts · June 2001
Revised & Annotated May 2006
NOTE: The following is a slightly revised and annotated version of a document originally presented by the authors to the annual Rail Transit Conference of the American Public Transportation Association (APTA), held in Boston, Massachusetts in June 2001. The original document is also contained in the conference proceedings, available in a CD-ROM from APTA. At the time, Lyndon Henry was a consultant with Mobility Planning Associates and Sam Archer served on the staff of Capital Metropolitan Transportation Authority (Capital Metro).
Also at that time, Capital Metro was seriously considering an electric light rail transit (LRT) system. The agency had
recently lost an effort to obtain voters' approval (required by state law) for an earlier LRT plan, but was continuing to
explore the possibility of a modified version of such a system. Two years later, the agency changed course,
discarding its planning for electric LRT in favor of a fuel-powered light railway proposal (called "Urban Commuter
Rail") – a plan which eventually won voter approval in November 2004 – see:
Despite the fact that Capital Metro's plans have changed, Light Rail Now believes that the information, data, and argumentation contained in the original paper are valuable and pertinent to issues that continue to be discussed with respect to the impact of rail transit and the relevance of new rail starts in a diversity of American urban areas. For that reason, it is being published here.
Most transit agencies proposing new light rail transit (LRT) projects, such as Capital Metropolitan Transportation Authority (Capital Metro), in Austin, Texas, must contend with strong criticism from rail opponents. [As indicated in our introductory note, Capital Metro has abandoned its plans for electric LRT in favor of a fuel-powered light railway plan, now in development.] Ridership projections for the new service are particular targets of such criticism.
Two consistent themes of these rail project opponents are that (1) LRT ridership projections are invariably in excess of achieved ridership, and (2) the ridership levels estimated for the proposed new light rail service will, in any case, have negligible impact on traffic congestion and air pollution. These contentions can be refuted by available evidence to the contrary. it is hoped that the information provided in this paper will be of value particularly to other transit systems proposing similar new LRT or other rail transit startups, and anticipating opposition from rail critics using similar arguments.
LRT's Record in Meeting Ridership Projections
While rail opponents argue that LRT ridership forecasts are invariably overestimated, the available evidence indicates otherwise. Opponents' criticisms are often based on various manipulations of data – disparagingly characterized as "numbers voodoo" by some rail supporters – performed upon transit agencies' ridership projections for various early system configurations and assumptions.
As a project proceeds, over many years, from initial sketch planning to final design refinement, configuration details, service assumptions, and ridership estimates typically change dramatically. Furthermore, external factors, such as demographics, job markets, commercial activity, and economic conditions may also change over this period – a circumstance to which even highway projects are vulnerable. Opponents typically seize upon very early ridership projections, made in the rather nebulous, formative stages of project planning, and then contrast these to final performance (although the actual design and service objectives may have changed dramatically in the intervening period). in contrast, a more plausible and honest assessment would compare relatively late projections, made in the final stages of design and service-level preparations, with achieved ridership performance. This best measures projections, based on the actual real-world configuration, to actual performance.
On this basis, LRT's performance is, on the whole, laudable. An examination of light rail ridership forecasts in several cities recently inaugurating new systems or lines indicates that, far more often than not, ridership forecasting models are acceptably accurate, and that new LRT services actually have an excellent "track record" for meeting and even exceeding their ridership targets.
Of six new LRT projects assessed – in St. Louis, Dallas, Portland, Salt Lake City, Hudson-Bergen (New Jersey), and Denver – only one has apparently failed to meet its ridership forecast: the new Hudson-Bergen (New Jersey) LRT line serving communities along the Hudson River, across from New York City. The remaining 5 new LRT lines have met or exceeded forecast ridership:
· in Dallas, ridership on the new Dallas Area Rapid Transit (DART) LRT system was 10% higher than forecast. (1)
· in Portland, by February 2000, ridership on the new Westside MAX LRT extension had exceeded forecasts by 22%. (2) As of July 2000, Portland's Westside MAX ridership was already five years ahead of all forecasts. (3)
· in St. Louis, ridership on the MetroLink light rail system exceeded its initial forecast of 12,000 by carrying 20,000 riders its first year – exceeding forecast ridership by 67%. Today, after 8 years, ridership has reached the 40,000 level, already exceeding the 20-year projection of 37,000. (4, 5)
· in Salt Lake City, ridership on the Utah Transit Authority's (UTA) Trax light rail line – now at an average of 20,000 people on weekdays – has exceeded forecasts by 43%. (6, 7, 8, 9)
· And in Denver, the brand-new Southwest LRT extension to Littleton, Colorado, opened in the summer of 2000, is now carrying up to 14,000 people on weekdays – 67% above the original projections of 8,400. (9, 10, 11, 12, 13)
These ridership accomplishments, in comparison with forecasts, are displayed graphically in Figure 1.
in some cases, higher LRT ridership has presented transit agencies with the somewhat bittersweet problem of endeavoring to cope with passenger volumes far in excess of what was expected from model predictions – in Salt Lake City, for example, and in Denver, where Regional Transit District (RTD) personnel are scrambling to acquire additional rolling stock. (14) Planners have begun to question Federal Transit Administration (FTA) forecasting guidelines and their underlying assumptions (in regard to fares and travel time, for example) which are permitted to be used to project LRT ridership.
Doug Allen, vice president of planning and development for Dallas's DART, complained to the Denver Business Journal: "The federal government wants us to use these models to compare various projects, like light rail and bus lines, the same way". Similarly, Mick Crandall, program director for Utah's Wasatch Front Regional Council, observed that "How people respond to rail is different than how they respond to bus." He elaborated: "There's a dependability on travel time with rail that there isn't with buses. it has its own tracks and you know where it goes." The Journal's reporter summarized, "The problem is that people like light rail a lot more than they like buses – a fact the federal rules don't take into account ...." The reporter concluded by noting that the FTA was "working on the problem" and that Denver's RTD, the Denver Regional Council of Governments, and the Colorado Department of Transportation "just started work on a new computer projection model that takes these factors into account ...." (9)
Regarding ridership on the new Hudson-Bergen LRT line in New Jersey, while some 18,400 daily riders were predicted, apparently only about 8,700 have materialized. Reasons for this serious shortfall are not completely clear. However, it should be noted that a rather complicated fare structure, premium LRT fares compared with buses, the lack of rail-bus transfers, and user-unfriendly ticket machines may function as disincentives to patronage. (15, 16)
LRT's impact on Traffic
However, even if ridership forecasts are accurate, or exceeded by actual rider volumes, another criticism widely deployed by rail transit opponents from city to city is that light rail's ridership impact on traffic would in any case be inconsequential – a claim based on a comparison of projected LRT ridership to total traffic in a vast area. Exploiting this argument, critics have publicized repeated allegations that only a tiny proportion of person-trips would be carried on the proposed transit service. Using this pattern of criticism, rail opponents have assailed not only plans for new LRT systems, such as those of Capital Metro for Austin, but also the ridership achievements of light rail systems in Dallas, Portland, and elsewhere, for having a minimal impact on regional trips, congestion, and pollution.
But there are fundamental flaws in such a comparison, which obscures the fact that LRT indeed does and will have a significant impact on traffic flow in the specific corridors it serves, especially during peak periods of travel. The use of areawide data is virtually meaningless, since a single LRT line (or even a system with several branches) could not plausibly be expected to have an impact on all traffic in a metropolitan region. indeed, it is likely that wide portions of any given metropolitan region will have no transit service at all (e.g., some precincts or counties may not have voted to be included in the transit service area). This underscores the absurdity of a claim that, for example, LRT will carry "only one-half of one percent of all trips" in an urbanized area.
Nevertheless, using this specious argument, rail opponents routinely flog new LRT proposals (such as Capital Metro's) for having supposedly minimal impact on traffic problems. But by this same measure, one would reject all possible mobility improvements, since virtually nothing will significantly "reduce congestion", as most highway and transportation planners now almost universally admit.
Certainly, the amount of traffic carried by a new LRT service is quite small in comparison with the vastness of a large metropolitan area's entire regional traffic -- but why try to compare traffic improvements in a single corridor with travel volumes of the entire region? if that were routinely used as a basis for evaluating projects, not a single bridge, underpass, tunnel, or freeway would ever get built. Like any major transportation facility, LRT and its impact should be evaluated on the basis of the corridor served, not a vast region.
On a corridor basis, there is scattered evidence that new rail facilities do have an impact on corridor traffic (although studies to gauge this are unfortunately not performed consistently). A classic case, of course, is the impact of the Philadelphia-New Jersey Port Authority Transit Corporation (PATCO) rail rapid transit line in decreasing highway traffic counts on Delaware River bridges immediately after its opening in 1969. And in the Washington, DC, area, studies have reportedly suggested that new MetroRail lines have brought about road traffic reductions, at least in the short term.
in regard to new LRT service, an Oregon Department of Transportation study suggested an apparent impact of Portland's Westside MAX LRT line on traffic. The study concluded that increased transit trips in the Westside corridor, stimulated by the MAX extension, tended to keep pace with auto trip growth, with the result that traffic would be significantly heavier without those improvements. Looking at data collected in the periods May 1993, October 1997, and May 1999, researchers found that "transit's share of westbound trips leaving downtown during evening rush hour increased 5 percent, while the share of drive alone auto trips declined 3 percent." (17, 18) it should also be noted that diverting auto trips to transit undoubtedly reduces demand for parking, especially in central-city and CBD areas where land is at a premium and construction is extremely costly.
Traffic impact of a Proposed LRT Line in Austin, Texas
Proposed LRT services may likewise have significant potential impacts on traffic in specific corridors. For example, Capital Metro's proposed LRT service [no longer proposed] was projected to have a potentially very significant impact in the particular corridors it would serve, and especially during peak periods of travel. According to Mike Aulick, executive director of the Capital Area Metropolitan Planning Organization (CAMPO), "The benefits of light rail are most significant during the peak travel period." (19)
According to CAMPO's own analysis, produced by CAMPO's trip- estimation models, Capital Metro's proposed LRT line would carry some 6% of rush-hour trips on i-35, the major north-south freeway (parallel to the LRT corridor) through central Austin. Likewise, LRT would carry about the same proportion on another major north-south highway on Austin's west side, Loop1 (the "MoPac" freeway). Without LRT, CAMPO estimates it will need to widen i-35 to 12 lanes, with land acquisitions forcing the removal of 2 hospitals, hundreds of homes, businesses, and other locations. (19, 20)
While rail opponents disparaged the projected 6% reduction in peakhour traffic on i-35 projected for Capital Metro's LRT service as "meager", a major roadway facility of about the same cost magnitude – the SH 130 tollway estimated to cost about $1 billion – was projected to reduce total i-35 traffic by only about 5%. Yet many of these same transit opponents strongly favored the SH 130 project without criticism.
Strongest LRT impact on Major Arterials
But while LRT may have some effect on freeway traffic, that impact is typically rather minimal, especially since much freeway traffic consists of intercity and truck traffic which is relatively less affected by inner-city corridor traffic improvements. Furthermore, freeways serve as trunks for dispersal of trips between many different origins and destinations in the urban area.
in contrast, the impact of LRT on a major signalized surface arterial, funnelling traffic between particular origins and destinations in the given corridor – including the CBD – can be considerably more substantial. Such an arterial may be immediately parallel to the LRT line, or provide a right-of-way for it. This impact was found for the LRT service proposed for Austin.
Using projections of light rail ridership produced by models developed for Capital Metro, plus straight-line extrapolations of future road traffic, another analysis indicated that the LRT service proposed for Austin would produce very tangible reductions of traffic on at least two of the major arterial roadways on which it would run. According to this analysis, an Austin LRT service would significantly alleviate future traffic flows on two major roadway arterials – North Lamar Boulevard and South Congress Avenue – diverting up to 30% of automobile trips in these corridors to transit. in terms of person-trips, the initial LRT service was projected to carry nearly half of the total traffic flow on North Lamar, the single most heavily travelled arterial, thus providing substantial relief for growing congestion. (21)
Based on previous ridership projections for the first and 18th years of operations, the results of this analysis in the two major corridors can be summarized as follows:
• Light rail would carry 48% of all person-trips, all day
• Light rail would carry 23% of all person-trips, all day
• Light rail would carry 56% of all person-trips, all day
• Light rail would carry 56% of all person-trips, all day
The relative automobile vs. LRT share of person-trips in Austin's North Lamar corridor, projected for the first year of LRT operation, is graphed in Figure 2.
Thus, while a given LRT service (like any other transportation improvement) may have a minuscule impact on areawide traffic, its impact in the corridors it serves may be impressive indeed.
Can Traffic Congestion Truly be "Relieved"?
To a certain extent, in raising expectations of significant "congestion relief" from LRT and other major rail projects, transit and LRT opponents exploit a common fallacy and misconception: That any single transportation facility, roadway or transit, can ever truly "relieve" congestion. it is almost universally recognized, even among highway planners, and throughout the transportation planning profession, that traffic congestion is a fundamental fact of life – basically, it continues to grow with population expansion and the proliferation of motor vehicles. Acceptance of some degree of congestion is actually incorporated into the basic design of urban roadways.
in the United States, the central agency setting standards for the design of streets and highways is the American Association of State Highway and Transportation Officials (AASHTO) – an entity somewhat equivalent to a National Bureau of Standards for roadway design. AASHTO has long accepted the reality that congestion is an unavoidable fact of life, and must be accepted, to some degree, in the basic design of roadway facilities.
Under the heading "Acceptable Degrees of Congestion", AASHTO's basic manual, A Policy on Design of Urban Highways and Arterial Streets (1974), states that "users are realists" who recognize they must share roadways with others. Therefore, says AASHTO, "They will readily accept a moderate degree of congestion." [p. 293]
Therefore, concludes AASHTO, "The matter of deciding upon the degree of congestion that should be used as a goal in planning and designing highway improvements is resolved by weighing the desires of the motorists against the resources available for satisfying those desires."
The bottom line seems to be: There is not enough money to build our way totally free of traffic congestion; therefore, to one extent or another, congestion is built into the basic plan and design of all roadway facilities.
While this is stated in the basic manual for the planning and design of highways in the United States, it is certainly not understood by the average motorist, and the public at large, who probably yearn for the relatively unencumbered conditions they occasionally encounter on some little-used rural highways or brand-new suburban arterials, or for the less congested roadways of their youth. Many rail opponents, therefore, proceed to exploit this fantasy, endeavoring to focus blame on the inadequacies of local planners and decisionmakers, who supposedly have failed to fulfill their responsibilities to the public and have subjected the highway system to years of "neglect". inherent in this ideology is the implication that free-flowing, uncongested conditions will somehow be restored, particularly if precious dollars are diverted away from all other mobility alternatives (eminently mass transit) and marshalled to provide a massive transfusion of funds into the roadway system.
Transportation planners also recognize that population and vehicular growth mean that some degree of congestion will be inevitable – that the best we can hope for is to slow the growth of congestion. CAMPO's Mike Aulick, for example, has himself warned that "As fast as we are growing, nothing is going to reduce congestion. Nothing. Not rail, not (Texas) 130, not U.S. 183-A." (22)
Rail Transit May Slow Congestion Growth
Thus, as various planning and transportation professions have been telling us, congestion may be here to stay. Short of a catastrophic event (like World War 2), it is unlikely that urban residents in North America will ever see the free-flowing traffic conditions of 5 or 6 decades ago. (Whatever roadway seems "uncongested" rapidly fills up, as both motorists and real estate developers rush to take advantage.)
LRT planners and supporters are quick to point out that, since light rail trains bypass traffic lanes stacked up with cars, LRT is an alternative to traffic congestion. And New Urbanists argue that, by attracting clusters of development and promoting better urban design, LRT is a powerful tool in attenuating the sprawl which promotes congestion.
But, while congestion is growing in all cities, there is some evidence that cities with rail transit have a lower rate of congestion growth than cities without rail. Data from Texas A&M's Texas Transportation institute (TTi) indicate that, in large cities with rail transit in major corridors, congestion increases at a 42% lower rate than in non-rail cities. (23, 24 [Also see Study: Rail Transit May Slow Growth in Traffic Congestion.] Thus, it appears that light rail – especially if expanded over time into multiple corridors, and coordinated with major improvements to the overall regional transit network – could indeed make an impact.
in conclusion, the evidence seems to suggest that (1) most forecasts of ridership for new LRT services are reasonably accurate (underpredicting, rather than exaggerating potential ridership), and (2) while the impact of new LRT services on total regional traffic in an urbanized area may be small, the impact in specific corridors served can be quite significant. This impact may be strongest on surface, signalized arterials parallelling the LRT service or providing a right-of-way.
However, data on such impacts are sparse. What is specifically needed is data on impact in a particular corridor, immediately after startup of LRT service. Unfortunately, local agencies rarely seem to undertake a study for that purpose. Sometimes routine traffic-count data for arterials and/or a freeway in the given corridor might enable some kind of comparative before-and-after analysis, but that appears to be fortuitous and rare. it should also be noted that such measurements of peak-period traffic would be most relevant to the issue of traffic congestion.
We would recommend that this kind of study, with actual before-and-after measurement of traffic volumes (including during peak) be consistently done when new LRT (or other major transit) upgrades are put into service. An ideal opportunity for such studies might occur when a new branch or extension of an existing LRT line is opened, e.g., in St. Louis, Dallas, Salt Lake City, Denver, etc. Such a studies would, in our opinion, be very worthwhile in evaluating the effectiveness of these major improvements in the corridors they serve.
1. Dallas Area Rapid Transit (DART) website, 2000.
2. Tri-Met (Portland) website, 22 February 2000.
3. Portland Oregonian staff, "Westside MAX leads way to higher Tri-Met ridership", Portland Oregonian, 27 July 2000.
4. Bi-State Development Agency (St. Louis) website, 2000.
5. Citizens for Modern Transit (St. Louis) website, 2001.
6. Utah Transit Authority website, 2000.
7. RapidTrac staff, "Salt Lake City Light Rail Opens with Great Fanfare", RapidTrac News Update, 3 January 2000
8. UTA staff, "UTA TRAX Ridership Exceeds Projections", UTA News, 21 December 1999.
9. Denver Business Journal staff, "RTD not alone in ridership problems", Denver Business Journal, 26 January 2001.
10. Denver Regional Transit District (RTD) website, 2000.
11. Cathy Proctor, "Light rail draws more riders", Denver Business Journal, 22 December 2000.
12. National Research Center inc., survey conducted in September 2000 for Denver RTD, December 2000.
13. Jeffrey Leib, "Rail-line ridership exceeds forecasts", Denver Post, 17 Oct. 2000.
14. PR Newswire staff , "Sacramento Firm to Build Rail Cars for Denver", PR Newswire, 14 Dec. 2000.
15. Amy Westfeldt, "Light rail ridership less than expected but growing steadily", Associated Press, 1 June 2000.
16. Charles J. Lietwiler, "Light rail takes off in New Jersey", Tramways & Urban Transit, July 2000.
17. Oregon DOT researchers, "Westside Traffic & Transit Report Release Study Shows Transit Keeps Pace With Auto Use", Oregon Department of Transportation, Salem, Oregon, 11 Oct. 1999.
18. G. B. Arrington, "Responding to Wendell Cox", report prepared for Auckland City Council, Auckland, NZ, August 2000.
19. Michael R. Aulick, letter to CAMPO Policy Advisory Committee, 24 October 2000.
20. Michael Aulick et al., "Presentation to Greater Austin Chamber of Commerce Light rail Blue Ribbon Task Force, CAMPO, 6 July 2000.
21. Mobility Planning Associates, "Light Rail Will Make a Difference" (report), MPA, Austin, Texas, 6 November 2000.
22. Kelly Daniel, "At issue in rail debate: Will it ease traffic jams?" Austin American-Statesman, 6 August 2000.
23. Mobility Planning Associates, "Rail Transit Slows Congestion, Says Study", news release, 3 October 2000.
24. Light Rail Progress website, "Study: Rail Transit May Slow Growth in Traffic Congestion", March 2001.
Capital Metropolitan Transportation Authority (Capital Metro), in Austin, Texas, like most transit agencies proposing new light rail transit (LRT) startup projects, has found its ridership projections criticized by opponents. Two consistent themes of these opponents are that (1) light rail ridership projections are consistently in excess of achieved ridership, and (2) ridership levels estimated for the new LRT service will, in any case, have negligible impact on traffic congestion. This paper attempts to refute these contentions by presenting evidence to the contrary.
Regarding the allegation that ridership forecasts are consistently overestimated, an examination of LRT ridership forecasts in a number of cities with relatively new systems indicates that forecasting models are acceptably accurate, and that most light rail startups have an excellent "track record" for meeting and even exceeding their ridership targets.
Rail transit critics also contend that LRT's ridership impact on traffic would be inconsequential – a claim based on a comparison of projected light rail ridership to total traffic in a vast area. Exploiting this argument, critics have publicized repeated allegations that only a tiny proportion of person-trips would be carried on proposed LRT services. Using this pattern of argument, rail opponents have not only assailed new LRT plans such as Capital Metro's in Austin, but also the ridership achievements of light rail systems in Dallas, Portland, and elsewhere.
However, evidence exists that LRT does and will have a very significant impact on traffic flow in the specific corridors it serves, especially during peak periods of travel. For example, in regard to Capital Metro's LRT plan for the Austin, Texas area, an analysis from the Capital Area Metropolitan Planning Organization (CAMPO) indicated that light rail would carry some 6% of rush-hour trips on interstate 35 (parallel to the LRT corridor).
Furthermore, another analysis indicated that the LRT service proposed for Austin would produce very tangible reductions of traffic on at least two of the major arterial roadways on which it would run. According to this analysis, an Austin LRT service would significantly alleviate future traffic flows on North Lamar and South Congress – diverting up to 30% of automobile trips in the corridor to transit. in terms of person-trips, the initial LRT service was projected to carry nearly half of the total traffic flow on North Lamar Blvd., a major arterial, thus providing substantial relief for growing congestion.
The information provided in this paper may be of particular value to other transit systems proposing similar new LRT or other rail transit startups, and anticipating opposition from rail critics using similar arguments.
Light Rail Now! website