NOTE: The following analysis has been slightly adapted for Webpage presentation on Light Rail Now! from a paper distributed by the authors at a Federal Reserve Bank ("Fed") of St. Louis public event in August 2004, responding to an attack by a Fed official on the MetroLink light rail transit system as a "boondoggle". While the focus is primarily on the St. Louis MetroLink system, the paper provides an illustrative case study of the fallacies in the myths underpinning the common "buy every rider a luxury car instead" argument used by opponents to disparage rail transit investment.

In two separate publications, personnel of the Federal Reserve Bank of St. Louis have recently launched vigorous attacks on US urban mass transit and particularly light rail transit (LRT) systems, together with the suggestion that rail transit passengers could more "cost-effectively" be transported by private motor vehicles, and that rail transit development and expansion in American cities might be replaced by further promotion of highway travel instead. The Regional Economist, a periodical of the Federal Reserve Bank of St. Louis, in its July 2004 edition published an article by Molly D. Castelazo and Thomas A. Garrett titled "Light Rail: Boon or Boondoggle?" Then the Fed published a study by Garrett alone, titled "Light-Rail Transit in America: Policy issues and Prospects for Economic Development", and dated August 2004.

The "Boondoggle" article (with criticisms particularly focused on St. Louis's MetroLink LRT system) has far more of the character of a political tract than a well- reasoned, fact-based analysis conducive to sound public policy decisionmaking. While Garrett's "Light-Rail Transit in America" study is slightly more demure and informative, it shares most of the "Boondoggle" article's fallacies and serious methodological weaknesses. identifying and understanding these fallacies and weaknesses is critical to furthering sound transportation decisionmaking in US cities.

Societal benefits get "disappeared"

The Castelazo-Garrett article is replete with contradictions and non-sequiturs. in their first paragraph, they state, "Whether light-rail transit is a boon or a boondoggle depends on whether the societal benefits of light rail outweigh its costs" – as if it is the intent of the paper to methodically address this issue, examine both costs and benefits, and provide evidence one way or the other. Unfortunately, they do nothing of the sort - the remainder of their paper simply ignores this pivotal question, turning instead to other assertions and themes. Not one shred of tangible information on "societal benefits" (on the basis of which the key question is purportedly to be decided) is presented throughout the article. This issue is simply "disappeared".

Castelazo-Garrett do seem aware of the kinds of issues they should be examining: They mention, in passing, benefits such as Increasing "community well-being", "creating jobs", "boosting economic development and property values", "reducing pollution and traffic congestion", and "providing drivers with an economical alternative to the automobile". Yet no evidence is presented, and no effort is made even to quantify a single one of these benefits in a common measure (e.g., dollars). instead, all these issues, as well as the notion of comparing "societal benefits" to costs, get jettisoned with a cursory dismissal that "Opponents counter that light-rail transit provides little of these benefits to citizens...."

Garrett's independent study avoids this contradiction by simply ignoring the issue of a benefit-cost comparison altogether. At most, Garrett simply discusses an assortment of "Economic issues" (Sec. III), in what is largely an oration of personal opinion while making no serious effort to develop a meaningful, fact-based, professional analysis of the benefits vs. costs of LRT.

LRT's impact on traffic congestion

Curiously, both the "Boondoggle" article and Garrett's independent study acknowledge that LRT does have some kind of positive impact on traffic congestion. The Castelazo-Garrett article dismisses this impact with the disdainful observation that "building light rail is only a short-run solution to the problems of traffic congestion and pollution." Garrett's "Light-Rail Transit in America" study examines this issue in a bit more depth, presenting what Is apparently Garrett's own research into traffic congestion, which actually finds that "light rail may have slowed the growth in roadway congestion in some cities." This leads him to conclude that perhaps there is some very marginal usefulness for LRT.

But what is rather bizarre is the central contention in both papers in regard to traffic congestion. The Castelazo-Garrett paper makes the assertion that "To permanently alleviate the problems of traffic congestion and pollution, policy-makers must address the root cause of both: the inefficient pricing of roadway usage. Traffic congestion and pollution exist because the costs of driving an automobile are artificially low." Similarly, Garrett's independent study argues that "Traffic congestion exists because of inefficient pricing of roadway usage."

This is a somewhat astounding contention, which seems to suggest that, if you could just make roads sufficiently expensive to use, you would force people to stay at home and not create all that congestion. Well, perhaps. But doesn't that seem to run counter to the basic objective of providing publicly sponsored transportation systems to promote mobility, enable workers to get to work, and make it possible to turn the wheels of industry and foster economic growth? What is especially disconcerting is that this is a Federal Reserve Bank representative (in a Fed publication) who is arguing for ratcheting up road-use costs to hold down traffic!

Public transport "profit" and subsidy

Both papers attempt to build a case against LRT (actually, against virtually all publicly provided mass transit) by attacking government supplementary funding ("subsidy") of these services. Thus the authors repeatedly and emphatically disparage St. Louis's MetroLink and other LRT systems for not paying their full costs from farebox revenues alone. Notwithstanding the fact that this is true of almost every public transport system in the world, and that the huge systems of New York, Toronto, London, Paris, Moscow, and elsewhere would have to shut down if they adopted the Castelazo-Garrett profit criterion as their key financial principle, the authors proceed to build their entire denunciation of LRT on the basis of this characteristic.

It should be noted here that there's a compelling reason why transit systems cannot fund their operations purely from passengers' out-of-pocket payments. The competing roadway-based transportation systems – highly subsidized after nearly a century of strong government investment – have been structured to minimize motorists' out-of-pocket costs. The high costs of private motor vehicle travel are covered by a largely unobtrusive umbrella of public and private subsidization as well as the transfer of "external costs" (like accidents and air pollution) to the general public. indeed, Vukan R. Vuchic, referring to a 1994 study by the US Office of Technology Assessment (OTA), observes that

the OTA study ... estimates that car drivers pay about 60 percent of the total cost of their travel. The remaining 40 percent consists of costs of highway construction, maintenance and control (traditionally subsidized by all three levels of government), "free" parking (subsidized by employers, store owners, schools, federal tax laws, and so on), and various social and environmental costs absorbed by society.
[Vuchic, Transportation for Livable Cities, Center for Urban Policy Research, Rutgers, 1999]

Against this heavily subsidized, government-promoted competition, public transport operators find it virtually impossible to charge fares high enough to secure "profitable" operation – especially if social policy goals such as maximizing ridership, reducing motor vehicle traffic, and stimulating downtown revitalization are being pursued.

Mass transport feasibility and public funding

The feasibility of major public investments, from libraries to highways to dams to public transport projects, is typically based on an economic assessment of the project – especially comparing costs to benefits - and a financial assessment of various funding mechanisms to pay for the project. But Castelazo and Garrett hopelessly confuse these methodologies. The feasibility of public transportation is, in effect, evaluated solely on the basis of the level of government funding, or subsidy – i.e., a financial mechanism.

The "Boondoggle" article, for example, asserts that "The economic value that society places on light-rail transit is reflected, in part, by people's willingness to pay for it. This is true for most products and services in the economy. To make a profit and stay in business, private companies must offer a product or service whose production costs are below what consumers are willing to pay for it."

To apply such a yardstick of private-profit-sector business operations to critical public enterprises and projects is an absurd distortion of the realities of economic functioning. The economic value of a library is not determined by how much users are willing to pay to check out books. The economic value of a public school is not determined by how much parents are willing to pay in tuition to send their child to classes. The economic value of a dam is not determined by how much lake users are willing to pay to use the lake, nor by how much those in the flood plain downstream of the dam are willing to pay to avoid being flooded. The economic value of a street is not reflected by how much money motorists and bicyclists are willing to insert in a tollbox or meter to run their vehicles on the street. in essence, by the Castelazo-Garrett "profit" criterion there would be no public schools, no fire stations, no public libraries, very few dams, no sidewalks, and a lot fewer bridges, viaducts, arterials, and other roadway facilities.

Nevertheless, Castelazo and Garrett leap from their private-profit-based frame of reference to a denunciation of LRT, including MetroLink, for failing their "profit" test:

The public provision of light-rail services, in contrast, costs more than consumers are willing to pay. For example, fare revenue covers only 28.2 percent of operating costs in St. Louis, 19.4 percent of costs in Baltimore and 21.4 percent of costs in Buffalo. Nationwide, annual light-rail operating costs ($778.3 million) far exceed fare revenue ($226.1 million); the balance ($552.2 million) is paid for with tax dollars. ... No privately owned system would ever be operated (or even be built) with such a dismal balance sheet.

In his "Light-Rail Transit in America" study, Garrett repeats this argument, arguing that MetroLink's failure to cover its total costs from the farebox "shows the value that residents place on their transit system is much less than the system's operating cost." This leads him to pronounce that "rail transit systems are cost-ineffective...."

Both papers proceed from this conclusion – plus the unfounded premise that the core objective of public transport should be to provide some vestigial service to "the poor" – to promote bus and other non-rail transit modes as supposedly preferable alternatives. The "Boondoggle" article recommends that, "instead of building light-rail systems to provide transportation for the poor, communities could expand bus service, offer more express bus routes or expand on-demand services; these would still realize the benefits of providing public transportation to the poor." And, in his independent study, Garrett makes a similar argument, adding the rather contorted thesis that "Although bus service is also cost-inefficient, it is socially beneficial to have fewer inefficient public transportation systems."

One would infer from this that the authors' preferred public transport alternatives – buses and "on-demand" transit (usually called "demand-responsive" service) – exhibit "healthier", more "cost-effective" farebox revenue performance vs. cost performance according to the authors' stated criteria. This however is far from the reality.

As previously noted, transit's rate of farebox return is actually more a measure of financial performance and indicator of social policy, and not a measure of economic cost-effectiveness or viability. Yet, even by this measure, there is a major inconsistency in both Fed papers' analysis.

At the same time that Castelazo and Garrett chastize MetroLink for achieving a farebox return of "only" 28.2%, they propose substituting bus services and "on-demand" public transportation service, which have a significantly lower farebox return: St. Louis Metro's bus services achieve only a 21.2% farebox return, and their "on-demand" demand-responsive services an astoundingly low 4.2%! Clearly, there is something seriously out of kilter with the authors' methodological approach, which assails low farebox returns for LRT, but then advocates modes with even lower farebox returns.

Indeed, MetroLink's farebox return of 28.2% compares very favorably with the performance of most comparable major cities operating only buses (as of 2002). in the table in Exhibit 1, it can be seen that St. Louis achieves a higher return than in all these cities except for Las Vegas, Milwaukee, and Minneapolis. And certainly, in not one of these cities is transit "paying its way" out of the farebox – every single one of these systems would fail the "profit" test postulated by Castelazo and Garrett.

Exhibit 1
Bus-Only Systems – Farebox Return

In regard to other mass transport alternatives, Castelazo and Garrett imply that nationwide LRT operations, with a national average farebox return of about 29%, somehow stand egregiously outside the norm of urban public transportation in terms of financial performance, and is some kind of failure for requiring supplementary public funding. in reality, all major public transport in the USA requires some form of public subsidy. The average farebox coverage of operating costs nationwide in 2002 (the latest year tabulated by the Federal Transit Administration) is only 36.2% – and that figure is undoubtedly skewed upward by New York City's amazingly remunerative transit system, which dominates US public transit figures, accounting for approximately 30% of all US transit ridership. NYC's system has an unusually high total average farebox recovery rate of 55.0%, with its heavily used subway achieving a 67.3% return, and its bus system 40.9%.. Yet even this comparatively very productive system would flunk the Fed papers' "profit" test!

"Transit for the poor" straw man

Apparently to provide themselves with a more vulnerable target for their attack on mass transport, the authors erect a "straw man" with their implied premise that the pre-eminent objective of LRT is to transport "the poor". Having glibly dismissed all other possible benefits of LRT, the "Boondoggle" article proceeds to establish that "Another justification for expenditures on light-rail systems is that they provide transportation to thousands of low-income individuals who otherwise would find their mobility quite limited." Garrett's "Light-Rail Transit in America" study likewise adopts this premise. With this focus, both Fed papers appear to promulgate the familiar highway industry stereotype that mass transit, including LRT, is primarily a mode of last resort for the lowest economic strata of society.

This allows both papers to launch into an extended polemical exercise in which the authors propose providing hybrid Toyota Prius automobiles as a supposedly more "cost-effective" alternative to providing mobility by LRT. This "Buy 'em a Prius" rhetoric is merely the latest variant in a familiar ploy of many rail critics, who propose that purchasing a motor vehicle (Lexus, Jaguar, BMW, Geo, Prius, etc.) for potential or existing rail transit riders would supposedly be "cheaper" than installing or expanding the targeted rail transit system.

"Buy 'em a car" – rhetorical exercise

During Houston Metro's fiercely contested (and ultimately successful) electoral campaign in 2003 to procure voter approval of an ambitious LRT expansion plan, one critic proposed rhetorically that "we could buy each new rider a top-of-the-line Mercedes (at more than $90,000) and still spend less money." This argument is dissected at some length on the Light Rail Now! website, an analysis which is quite relevant to the similar argument being put forth by the Fed papers. For example, Light Rail Now! asks, "What would be the benefit of buying an expensive new car for new transit riders already attracted from their own cars because of their desire to avoid the hassles of traffic congestion and scarce, expensive parking?"

As the article further points out,

The cost of a car is only a relatively small part of the total expense of adding a vehicle to the already-congested roadway system. To this must be added major additional capital costs such [as] those for roadway lanes and parking spaces to accommodate these additional vehicles. And then there are the ongoing costs of operation, incremental roadway maintenance and management, etc. Incurred in the use of the new luxury car – costs ignored by the perpetrators of this argument.

... Automobile use – including use of the new luxury car postulated in this argument – incurs a host of external costs – accidents, air pollution, undesirable land-use impacts, noise, stormwater runoff problems, etc. – also not addressed in this argument.

... The "buy 'em a Mercedes" argument proposes buying a car for each rider – i.e., promoting 1.0 vehicle occupancy, a traffic policy which runs counter even to the nominal policies of highway proponents and planning and development agencies, which are struggling to increase vehicle occupancy. How does flooding the roadway system with (typically) thousands of luxury automobiles (hypothetically each driven by a transit "new rider" wooed back to the automobile) provide any relief for highway congestion?

To these common fallacies, the Fed papers' "Give 'em a Prius" exercise in sophistry adds another: including major capital grants – intended for a 50-year investment – as a major share of the alleged "Annual total subsidy" to the MetroLink system. Adding in these two large capital contributions to the 2001 operating deficit, of course, conveniently drives up the ostensible huge annual subsidy number, supposedly clinching the Fed papers' argument. By this same token, if you buy a $300 camera, and take a photo, the Fed analysis would presumably proclaim you'd just spent $300 for that snapshot (plus the incremental cost of the film)!

Obviously, capital investments need to be amortized, or in this type of economic analysis, annualized, based on the assumed economic life of the investment. And that's just what we have done in the Appendix of this critique, annualizing the entire MetroLink capital investment since its inception. To this, annual operating costs are added to round out an accurate picture of annual operating performance (see graphs in Exhibits 2-3).

Exhibit 2
MetroLink Operating Cost vs. Motor Vehicle Out-of-Pocket Cost

To effect a fully accurate comparison with the "Buy 'em a Prius" alternative, our analysis estimates some of the plausible major highway, street, and parking costs which would also be incurred. While we've had to simplify this hypothetical exercise with rough, "horseback" assumptions and estimates, it's clear that draconian alterations to the St. Louis roadway system in the MetroLink corridor, and to downtown St. Louis, would be necessary to accommodate the avalanche of nearly 17,000 additional motor vehicles which would be unleashed into traffic – nearly 12,000 of them pouring into the St. Louis CBD. These implications are totally ignored in the Fed "Buy 'em a Prius" exercise, along with extra severe economic impacts such as the relatively high expense of daily CBD parking.

Exhibit 3
MetroLink Total Cost vs. Motor Vehicle Total Cost

In contrast, MetroLink is conveying all these passengers efficiently and cost-effectively into the heart of St. Louis ... with no influx of additional street and highway traffic, no cutting more swaths of concrete and asphalt through the city – instead, MetroLink provides environmentally friendly, quiet electric transport at 28% less cost than what the Fed papers' "Buy 'em a Prius" alternative would entail. in the charts in Exhibits 2 and 3, the two alternatives are compared on the basis of their cost, both per trip and cost per passenger-mile (pax-mile).

Additional fallacies

The Fed papers contain a host of other fallacies, of which just a few will be highlighted here. For example, Garrett's "Light-Rail Transit in America" study appears to present much misleading information on the history and characteristics of American urban electric rail transit systems. in its Section II, "History and Scope of U.S. Rail Transit Systems", Garrett asserts that "c. 1880", "Electric street car began operation. Much cheaper than cable car; cleaner than horsecar. Fares were about 5 cents, about half that of horse-cars. However, electric streetcar did not reduce growing congestion." Garrett, who regards "horsecars on rails" as "the earliest form of light-rail transit in the United States", seems to regard 1880 as approximately representative of the heyday of early "light-rail" (which, he insinuates, carried only a tiny proportion of urban residents). Yet viable urban electric rail traction systems did not appear until 1888 and thereafter, following inventor Frank Sprague's introduction of this technology in Richmond, Va.

Garrett's "overview" also seems to seek to downplay the significance of electric rail transit even in its earliest years. For example, he claims that in 1880, only "about 493,000 [riders] per day" rode on horsecars; thus, he argues, "roughly 3.5 percent of all urban residents rode light rail."

But a far more representative year for collecting early electric rail passenger percentages with respect to America's urban population would arguably be about 1900, and the 1900 census would be a more reliable basis for demographic data. in 1880 most trips were on foot, by horse, or by buggy because the majority of Americans lived in small towns or on the farm; cities large enough to have any public transit apparently accounted for a quite small percent of the urban population of 14,100,000 referenced by Garrett. in 1900, by contrast, some 6 billion rider-trips were carried on America's urban electric railways, amounting to about 18.5 million rider-trips a day, or about 8.3 million persons (calculated via more realistic factors of 325 annual-to-weekday ridership and 45% trips-to-persons). This amounts to over 27% of the total 1900 US urban population of 30.2 million, not the relatively trivial 3.5% Garrett claims.

Also implicit in Garrett's statement cited above is the notion that, even from its inception, public transit has somehow been supposed to reduce the congestion of private surface transportation modes (e.g., buggies, wagons, and eventually motor vehicles). This is a false premise for urban transit's function, particularly since it seems to equate traffic congestion – a condition prevalent in virtually any city – with a lack of mobility. On the contrary, urban mobility should be measured by how many tasks at different locations can be accomplished within a given period of time – and public transport can provide alternative, congestion-free mobility.

What are they thinking?

In fact, Garrett's overall views on congestion, elaborated further in this same report, and taken in context with his other positions, are rather troubling. in Section II, "Economic issues Surrounding L:ight-Rail Transit", Garrett basically argues not only that some level of urban traffic congestion is not only inevitable, but desirable:

It is also important to realize that there is an optimal level of traffic congestion. A roadway with miles of bumper-to-bumper traffic is clearly an overused resource, but a roadway with no congestion at all is an underused resource. Thus, there exists some optimal level of congestion. By having some commuters in heavily congested areas substitute rail for car, it is possible that light rail serves as a marginal reducer of traffic congestion, thereby providing a more optimal amount of highway congestion.

Thus, we have a Senior Economist of the Federal Reserve Bank who is arguing that:

· Private motor vehicles are (supposedly) the most "cost-effective" form of urban transportation;

· But, rather than supporting public transport as an alternative to traffic congestion, it's a better policy to alleviate congestion by imposing cost hikes on the use of roads, so that fewer people can afford to use them;

· However, "optimal" traffic congestion is actually a desirable condition;

· it would be far more "cost-effective" to promote motor vehicle use, attracting transit passengers onto highways, even, if necessary, by providing cars to the poor (thus flooding the congested roadways with even more traffic).

What are they thinking? These bizarre and wildly contradictory premises suggest that something is seriously out of kilter with policy analysis at the Federal Reserve Bank.

Economic development and property values

Garrett acknowledges that LRT's economic development potential is a salient issue; in his "Light-Rail Transit in America" study, he emphasizes, "The most important economic question surrounding light rail is whether it can help foster economic development." Yet Garrett's focus seems to reside almost exclusively on property values as the measure for this impact. in his discussion, Garrett cites eight studies of rail transit's impact on property values, then concludes that "the impact of light rail on property values cannot be generalized." Yet, of the eight studies, seven indicated a predominantly positive effect of rail transit on property values.

Furthermore, in his list of studies on the property-value impact of rail transit, Garrett leaves out one of the most important of all – the Bernard Weinstein/University of North Texas study, performed for Dallas Area Rapid Transit (DART) and focusing on the Dallas LRT system's impact. This found a "jump in valuations around DART stations" to be about 25 percent greater than in the control neighborhoods". in addition, Weinstein's study suggested that DART's LRT might have positively impacted retail sales in Dallas's CBD, which had experienced much stronger retail sales growth (36.2%) since DART's LRT had gone into service.
[Bernard L. Weinstein, Ph.D. et al., "The initial Economic impacts of the DART LRT System", University of North Texas, July 1999]

Interestingly, Garrett (in "Light-Rail Transit in America") does report his findings from his own study of the property value impact of St. Louis's MetroLink LRT, which include "strong evidence of a relatively larger accessibility effect – home prices generally rise as distance to a MetroLink station decreases." Yet Garrett dismisses the economic development influence of LRT as a whole, repeatedly asserting it might perhaps "guide", but does not attract development (or provide other economic stimulus).

On this, as in many other aspects of the benefits of LRT, Garrett is strong on opinion but extremely weak on evidence. On the impact of transit-oriented development (TOD), Garrett cites only one study which "suggests that TOD has brought only modest benefits to transit agencies" – ignoring the growing array of literature on this issue (such as the Weinstein study, cited above, which also addresses TOD around Dallas's LRT stations).

Particularly in St. Louis, LRT is surely as close as one gets in the real world to an example of a successful urban project. As the graph in Exhibit 4 indicates (showing bus, demand-responsive service, or DR, and LRT ridership), the introduction of MetroLink service in 1994 (first full year) succeeded in reversing the long decline of St. Louis transit. Despite the area's previous attempts to spark ridership with innovative express bus services and even early "bus rapid transit" ("BRT"), ridership had relentlessly continued to drop.

Exhibit 4
St. Louis Transit Ridership 1988-2002

As the graph illustrates, MetroLink has successfully reversed that trend. it has attracted upwards of 70 and 80% of its ridership from people in motor vehicles. it has helped revitalize downtown St. Louis, and is continuing to attract TOD and stimulate economic development throughout the corridor it serves. Truly, MetroLink has been a transportation bargain for the St. Louis area, in more ways than one.

APPENDIX

MetroLink Costs

Rather than assume an entire capital cash flow in a single year as an annual cost, our study takes MetroLink's capital costs and annualizes them over the life of the project; only these annualized capital expenses are considered as portions of the system's annual cost, per normal, competent economic analysis. Thus, the capital costs of the MetroLink LRT system as of 2001 were added and converted to 2001 dollars. This produced a total of $845 million. Economic life was assumed as 50 years for fixed facilities, and 25 years for rolling stock. A discount rate of 7% was assumed for the annualization conversion.

Ridership, passenger-mileage, and operating costs for 2001 were taken from the Federal Transit Administration's National Transit Database. The results, with calculations are summarized in the tables in Exhibits 5-7.

Exhibit 5. Ridership & Passenger-Mileage

Exhibit 6. Annualized Costs

Exhibit 7. MetroLink Unit Costs

Motor Vehicle Costs

Both Fed studies claim that private motor vehicles would be less costly and more "cost-effective" for transporting rail riders than mass transit or MetroLink, and the assumption of a strictly private motor vehicle alternative has been maintained for the purposes of this analysis. A motor vehicle-based system substituting for MetroLink would have to take into account the impact on all MetroLink passengers, not merely low-income riders. it was further assumed that one of the key urban development goals of MetroLink – bolstering the economic and social revitalization of downtown St. Louis – would need to be addressed by such an alternative (in addition to the fact that approximately 70% of MetroLink riders apparently have jobs in the St. Louis CBD).

Without MetroLink, there would be a surge of some 4,200 passengers, in 3800 motor vehicles, travelling during the most intense peak hour in the peak direction. Even if free Priuses were provided to low-income riders, the entry of these former passengers, plus other MetroLink riders, into the MetroLink corridor's traffic flow would impose a demand for additional roadway capacity, particularly to handle the extra peak surge. This study assumes that a new 2-lane freeway facility would be provided – 20 miles of a 2-lane facility in semi-urban areas and 8 miles of a 2-lane facility in urban conditions. (These could be HOV or HOT lanes, or general-traffic lanes.) in addition, an additional 2-lane bridge over the Mississippi was assumed.

This flood of thousands of additional vehicles would impose stressful demands on the street system of downtown St. Louis, especially in peak conditions. There are various alternatives for trying to accommodate this traffic smoothly (in the context of reliance on the private motor vehicle system), none of them easy to implement. This study assumed a widening of at least 7 arterial streets to handle peak flows in both directions for a total of one mile. The acquisition of business property on one side of each street via condemnation was assumed to enable this widening, at a cost of $40 million per mile. These roadway facility capital cost assumptions are summarized in the table in Exhibit 8.

Exhibit 8. Roadway Capital Cost Assumptions

The avalanche of additional motor vehicles would impose a need for thousands of additional parking spaces, particularly in the St. Louis CBD. Parking demand was calculated by deriving the number of MetroLink passengers (approximately 70%) who apparently used cars before riding LRT and the number (also approximately 70%) travelling into the CBD. To convert rider-trips to individual passengers, a factor of 45% (recommended by American Public Transportation Association) was used. Vehicle occupancy was assumed as 1.1 for work trips and 1.2 for all others. Bottom line: 11,900 additional vehicles would pour into downtown St. Louis each weekday, and another 5,000 elsewhere would be competing for traffic space and parking in other segments of the corridor.
[Sources: APTA, "Number of People Using Public Transportation", APTA website, Aug. 2004; Paul Weyrich, "Does Transit Work? A Conservative Reappraisal – Part 3: St. Louis", APTA, 2002]

For the 11,900 additional vehicles needing parking in the CBD, the annual cost was assumed to be $1,200, based on an average calculated from prevailing parking costs provided by Metro (July 2004) and converted to 2001 dollars. Parking costs outside the CBD were assumed to be 10% of those in the CBD. (Even a "free" parking space carries an annual cost in terms of amortization of capital expense and ongoing maintenance.)

Rather than adopt the Fed assumption that all low-income passengers would be provided with Toyota Priuses, this study assumed that only the 27% of riders without cars would be so provided – a total of 5,200. The annualized capital cost of these was assumed to be the same as that indicated by Garrett.

The Fed papers also assumed, as the alternative cost of motor-vehicle transport, the entire annual operating and maintenance cost of these "low-income gift" Priuses. However, in this study, the cost of automobile operation was assumed to apply only for the commute trip, averaging about 7 miles in each direction, for all former MetroLink riders "converted" into motorists.. For the hypothetical 5,200 new Prius owners, this cost was calculated at $0.30 per vehicle-mile, based on data in "Cost per Passenger-Mile of an Urban Automobile", Light Rail Now! website, March 2001; no depreciation cost was assumed since the vehicle capital cost is tabulated separately. Likewise, parking cost is not included in this figure since it is separately tabulated in this analysis. For the 11,700 vehicles of the other "reconverted" motorists, who presumably already would own these motor vehicles, both operating and depreciation costs were assumed, at $0.72/mile.

Based on these assumptions and calculations, including a discount factor of 7% to annualize roadway capital costs, results are presented in the tables in Exhibits 9-10.

Exhibit 9. Person-movement & Passenger-Mileage

Exhibit 10. Annualized Costs of Motor Vehicle-Based System

The public typically are not aware of the relatively large disparity between perceived out-of-pocket costs of operating a private motor vehicle, and the total costs (much of which are borne by society at large, partly through governmental and other private subsidies). For the hypothetical MetroLink riders converted to motor vehicle users, these costs are tabulated in the two tables in Exhibits 11-12.

Exhibit 11 Out-of-Pocket Costs of Motor Vehicle System

Exhibit 12 Annualized "Hidden" Costs of MV System

Unit costs, derived from these calculations and estimated trips and passenger-mileage, are shown in the table in Exhibit 13. A dramatic disparity can be seen between out-of-pocket costs, which convey an impression that motor vehicle travel is "cheap", and actual, total costs, which are approximately four times greater.

Exhibit 13. Motor Vehicle Unit Costs

It can also be seen, in comparing total costs of MetroLink LRT and the "Buy 'em a Prius" alternative, that MetroLink is a comparative bargain as a transportation investment in this application (travel to and from St. Louis's CBD) – amounting to a 28% saving per passenger-mile when all costs are considered.

While critics like Garrett denounce St. Louis's MetroLink light rail transit system, it has nevertheless proven extremely cost-effective as a transit operation and popular with the public. Here, Cardinals fans crowd a suburban platform to catch a train to Busch Stadium. Apparently, rail opponents would prefer them all to be jamming into central St. Louis in their motor vehicles.
[Photo: L. Henry]

Lyndon Henry is a transportation planner and technical consultant to the Light Rail Now Project. David D. Dobbs is publisher of the Light Rail Now! website.