In recent documents, Randal O'Toole – a prominent national critic of public transport (particularly rail transit investments) and Smart Growth policies – presents a ferocious attack on America's major urban rail transit systems, which he characterizes as "Great Rail Disasters":

• Great Rail Disasters
http://americandreamcoalition.org/1-2004.pdf

• Rail Disasters 2005
http://americandreamcoalition.org/RD2005.pdf

Both of O'Toole's "Rail Disasters" tracts (produced under the moniker of his self-styled "American Dream Coalition", which vigorously promotes highways and urban sprawl) have been widely publicized, excerpted, summarized, and disseminated by a battalion of anti-rail extremists, partisan journalists, and academic critics, for whom his fulminations serve as a kind of Holy Writ or magical touchstone to cite as purported "research" validation for extravagant denunciations of rail transit. Yet, when examined in detail, O'Toole's methodology is revealed as a mélange of disastrously flawed methodology and deceptively manipulated and selected statistics.

Absurd measures of "urban livability"

For example, in his 2004 diatribe, "Great Rail Disasters", O'Toole purports to assess "urban livability" via a collection of "indexes" he concocts supposedly relating to the performance of rail systems in US cities. Yet approximately half of O'Toole's "indexes" have nothing actually to do with "urban livability" at all – they're predominantly based on a rehash of worn arguments and "Numbers Voodoo" (deceptively selective and misleading data manipulation) purporting to show that rail systems have failed to reach ridership and budget targets, have lost out in competition with the automobile, etc.

But even if a transit service fails to attract as many passengers as planners had hoped, one wonders: how is this a negative impact on the "livability" of an entire urban area? On the contrary, one would think that getting more people onto transit is an inherently positive impact, even if ridership is not as much as was predicted.

If a library is built, but fails to attract as many patrons as had been originally projected, does O'Toole then believe the library thus has a negative impact on the city? If a highway bridge is constructed, but fails to carry as much traffic as had been predicted, would O'Toole rate this as a "negative impact" on the "urban livability" of the metro area?

Another of O'Toole's "urban livability" measures is traffic congestion. in "Great Rail Disasters", O'Toole attempts to enlist data from the annual Mobility Study performed by the Texas Transportation institute (TTi) to assail the performance of rail transit. Specifically, O'Toole deploys one of the TTi's urban area congestion measures (the Travel Time index) as one of his own "indexes", and then compares cities on the basis of this index. incredibly, he uses this index to ascribe the entire increase in congestion in all the areas running some degree of rail transit (even the relatively short segments in New Orleans and Buffalo) as a measure of the "negative" impact of rail transit in those cities! Presumably, all the private motor vehicles stacked up in congested traffic had either a neutral, or "positive" impact on that congestion?

O'Toole's premise also seems to totally contradict another key premise, which is that rail transit (or any transit) is so minuscule it has no impact on congestion whatsoever. So, how can it have no positive impact because it's so irrelevant to the total volume of traffic, but suddenly be assigned de facto total responsibility for congestion in O'Toole's index?

Basically, with this line of attack, O'Toole deploys a familiar trick favored by many rail critics: posing the absurd expectation that a small amount of rail – in many cases, a single line – in an urban area should magically cause traffic congestion to melt away throughout the entire region, even including areas outside the transit service area. This is a type of "straw man" argument that not only implies supernatural powers to rail transit, but also ignores the very real, and enormous, ongoing process of upgrading and expanding the roadway system in virtually every urban area – a process which inherently induces more traffic and does so at a much higher rate than relatively small rail transit operations can possibly mitigate.

"Housing Affordability" – tangle of contradictions

O'Toole's "Land-Use Regulation and Property Rights" index of "urban livability" represents another example that trenchantly illustrates his deeply flawed methodology. This particular "urban livability" measure is based directly on the National Association of Home Builders' Housing Opportunity Index (HOI), which happens to find that a lot of cities with rail transit have housing "opportunity" (i.e., affordability for median-income families) below the national average. But housing tends to be in relatively short supply in many of the very large, dense metro areas in the nation, most of which happen now to have rail transit; yet, astoundingly, O'Toole's methodology ascribes the resulting low affordability of housing (a result of the housing market) primarily to the presence of rail transit.

The absurdity of O'Toole's method becomes especially clear when one examines the HOI data in detail. For instance, Portland scores low in the HOI – 46.6, meaning that only 46.6% of housing in the Portland PMSA (Primary Metropolitan Statistical Area) is considered affordable to median-income families. O'Toole calculates this to be 30% below the national average, so his "index" assigns a poor score of -30 to Portland's rail transit system – as if the entire housing market were somehow negatively impacted by Portland's MAX light rail transit (LRT) system and the city's streetcar operation! (Since this would actually mean that rail transit is driving housing prices up, this would seem to suggest another huge contradiction with rail opponents' repeated claims that rail transit disastrously pulls property values down ... but, of course, this contradiction, like so many others, does not bother O'Toole.)

Contradictions and flaws continue to mount in O'Toole's "housing affordability" index of "urban livability". While Portland's HOI 46.6 rating seems low, the HOI ratings of two other Oregon cities – Eugene and Salem, neither of which has any rail transit – provide revealing comparisons (but of course are ignored in O'Toole's "research"). Eugene, with an all-bus system, scores even "worse" than Portland – 38.9. And while Salem's rating, 50.4, is only slightly above Portland's, it's still well below the national average – all of which evidence suggests that a tight, expensive housing market is a regional phenomenon having nothing to do with the type of public transport in these cities.

Contrary to Randal O'Toole's absurd contentions, Portland's MAX LRT and streetcar system have helped make Portland one of the most livable cities in the nation.
[Photo: L. Henry]

The case of California bestows even more insight into the absurdity of using a measure based on indexes like the HOI as a yardstick of "urban livability" – and particularly of relating this to the type of public transport in any given city. Judged by HOI, almost every significant city in the state would seem to flunk the "affordability" index. For example, Santa Cruz, with no rail transit at all, scores a puny 8.0, less than the 9.2 in high-density, rail-served San Francisco. San Luis Obispo, which could hardly be called a "rail transit city", scores 13.0, far less than the 34.4 earned by Los Angeles (with rail rapid transit, regional passenger rail, and several LRT lines, clearly a hotbed of rail transit development).

Then there's the case of Texas. Dallas (regarded by rail adversaries as a kind of rail transit Den of Iniquity) scores 70.5 – comfortably above the national average. Yet, while several other major Texas cities (all without any rail transit at the time of O'Toole's rating) are also above the HOI average, all fall below Dallas's level of housing "affordability": Houston 67.8; Austin 67.9; San Antonio 68.5; El Paso 68.8. Fortunately for O'Toole, this all doesn't matter, since his methodology consists of comparing only "rail transit" cities with the HOI national average.

Cherry-picking data for desired results

Most of O'Toole's arguments have been meticulously examined in analyses performed by Todd Litman of the Victoria Transport Policy institute – particularly in the following studies:

• Evaluating Rail Transit Criticism
http://www.vtpi.org/railcrit.pdf

• Evaluating Criticism of Smart Growth
http://www.vtpi.org/sgcritics.pdf

As Litman points out, O'Toole consistently and carefully "cherry-picks" his data and analysis methods to achieve results that buttress his desired conclusions. For example, he notes, O'Toole's recent report "Rail Disasters 2005" compares transit ridership trends between a group of cities that installed or expanded their rail transit systems during the last two decades and a selected set of cities that expanded their bus transit systems. However, says Litman, this is an unfair comparison, because the cities that expanded rail transit were generally large, mature, and slow-growing cities that experienced slow or negative population growth during the broad period O'Toole selected (although in recent years most have gained population), while those selected with all-bus system expansions were rapidly-growing cities experiencing rapid population growth, where in some cases the transit systems grew from tiny or nearly non-existent to medium-size, and thus, not surprisingly, experienced major transit ridership growth. Nevertheless, Litman observes, in absolute terms rail transit cities have experienced significant ridership growth during the last few years, due to a combination of expansion and other ridership incentives.

For example, says Litman, during the two-decade period O'Toole presents, transit ridership in Las Vegas (one of the all-bus cities selected by O'Toole and the one with the largest percentage ridership increase) grew ten-fold … but over this same period, the urban area's population tripled, business activity exploded, and the city expanded to a size where transit has become increasingly important. During the same period, New York City transit ridership increased "only" 15% – proportionately much smaller, but far larger than Las Vegas's in absolute terms. And although annual per capita transit ridership increased in Las Vegas from 95 to 331 passenger-miles, this is still small compared with New York's 1,707 passenger-miles. At this growth rate it will take a century for Las Vegas residents to reach New York's current per capita transit ridership.

Austin provides a good case study of the misrepresentation inherent in O'Toole's "cherry-picking" methodology. In 1983, Austin Transit was still a small, poorly funded city bus system operated by the City of Austin. In 1985, Capital Metro was formed, funded via a full 1% sales tax (a goal local transit advocates had pursued for the past decade or more). Thus, there was an immediate major injection of not just city funds but region-wide funding into regional public transport. This enabled the community to start running regional transit services – thus adding new, previously untapped suburban ridership into the local tally.

Then, in 1989, Capital Metro contracted with the University of Texas to take over UT's previously out-sourced "BRT"-type "shuttlebus" services – nearly doubling total public transit ridership. O'Toole treats this basically administrative measure as if it were proof of the inherent power and attractiveness of buses compared with rail – yet the previously existing UT ridership had simply not been previously counted in national statistics (since federal aid had not been involved).

Capital Metro, to its credit, has pursued an aggressive program of service improvements and facilities and rolling stock upgrades – but it has also kept a very low base fare of $0.50, which now accounts for only about 3% or so of operating revenues. Nonetheless, O'Toole has cherry-picked the Austin system (with a 522% increase in ridership over the 20-year period) for his tally. This practice of "gerrymandering" cases for mining statistical data seems to be a stock in trade of "professional" rail critics like O'Toole.

One can assume that such cherry-picking of cases also accounts for the other all-bus cities O'Toole has "creatively" selected: Eugene, Louisville, Charlotte, Houston, Phoenix, and Raleigh-Durham. It's interesting to note, however, that of all the previously all-bus cities selected, Houston has since started operating rail transit (LRT); Charlotte, Phoenix, and now Austin have new-start rail projects under way; Raleigh is finalizing its rail transit program and seeking federal approval; and Las Vegas is seriously exploring several rail options, including LRT.

Further, as Litman points out, nothing in O'Toole's analysis considers what would have happened in his selected rail transit cities had rail not been available. Referring to his own work at VTPi, Litman notes that "My matched pair analysis shows that rail transit cities tend to have much greater per capita transit use than comparable cities that only have bus transit, indicating that ridership losses would have been much greater in those cities if they had not had rail. [O'Toole's] claims that rail transit fails to increase ridership therefore appear to be wrong."

O'Toole's statistical trickery

In both his "Rail Disasters" tracts, O'Toole makes use of another common statistical trick favored by several anti-rail ideologue "researchers" – purporting to assess rail systems' performance through a period during much of which rail in the target city or cities did not even operate. Typically, this involves going way back to a previous year well before rail construction as a starting point of the "study" period. For example, his "Rail Disasters 2005" comparison starts at 1983 and stretches through 2003 – most new rail projects were operative during only a fraction of this period (most modern rail projects don't start coming online till the late 1980s and the 1990s).

Yet another of O'Toole's methodological contradictions, according to Litman, is that he "criticizes rail as ineffective when its ridership grows less than bus, and [also] criticizes it when its ridership grows faster than bus, for taking away resources." Indeed, says Litman, "several cities that have invested substantially in rail have experienced growth in total transit ridership (Portland is a good example) indicating that rail and bus transit are complementary." Furthermore, notes Litman, there are typically lags of several years between the construction of a rail line and achieving full ridership – which, of course, O'Toole ignores. Described differently, concludes Litman, "O'Toole's analysis favors short-term benefits, while rail investments tend to represent larger but slower benefits."

O'Toole's diatribes also present a deceptive interpretation of transit performance in individual rail transit cities. For example, notes Litman, O'Toole misrepresents the history of transit developments in Los Angeles.

He claims that transit ridership growth in LA is explained completely by intervention by the [LA- based] Bus Riders Union, but ridership seems to have grown when rail transit was introduced. His argument is based on the assumption that buses and rail compete for resources, but voters approved funding specifically for building rail, and the money would have otherwise been spent on highways.

With respect to transit system operations as a whole, O'Toole's methodology relies on obscuring or ignoring important information and critical differences between rail and bus operations. For example, although in some cases, as Litman acknowledges, light rail costs per rider tend to be higher than for buses, "this reflects the fact that rail often uses its own right-of-way, light rail systems are mostly new and still growing ridership, and light rail is built in congested corridors where costs are high."

To provide the same level of service on the same corridor by bus would also be costly, since it would require separate busways, and extra-quality vehicles and stations. Overall, rail transit (including all forms of rail) tends to cost much less per passenger-mile than bus transit, and far less than accommodating the same trips by car.

Rail transit's real impact on congestion

And in regard to rail transit's impact on traffic congestion – which, of course, O'Toole trivializes and dismisses – Litman's research leads to some strikingly opposite conclusions, far more favorable to rail. Referring to a study by Thomas A. Garrett (Light Rail Transit in America: Policy issues and Prospects for Economic Development, Federal Reserve Bank of St. Louis, 2004), Litman relates that Garrett's research identified significant ameliorative effects on traffic congestion in various cities:

Research by Garrett (2004) found that traffic congestion growth rates declined somewhat in some U.S. cities after light rail service began. In Baltimore the congestion index increased an average of 2.8% annually before light rail, but only 1.5% annually after. In Sacramento the index grew 4.5% annually before light rail, but only 2.2% after. In St. Louis the index grew an average of 0.89% before light rail, and 0.86% after. O'Toole [claims] that this proves that light rail does nothing to reduce congestion, but the research in my paper actually shows otherwise: congestion would almost certainly have been worse without rail.

To sum up: O'Toole's "Great Rail Disaster" would seem to be his own deceptively concocted polemical tracts presented in the guise of a "research study". Far from the result of bona fide "research", O'Toole's products appear much more to be ideologically motivated diatribes that amount to an argumentative Train Wreck by dint of their own chicanery, inconsistencies, and hopeless anti-transit fanaticism.

Some material in this article has been adapted and quoted from internet postings of Todd Litman and Lyndon Henry.