Is automation a "wonder pill" for reducing transit operating costs? Do automated transit systems have an inherent advantage in dramatically lower operating costs compared with manually operated light rail transit (LRT)?

increasingly, proponents and vendors of automated transportation technologies are emerging to tout the alleged superiority of "driverless" operation. Visions of significant operating cost savings and even "profitable" operation are dangled to bolster arguments – especially from monorail enthusiasts and vendors of various "gadget-transit" guideway systems – that transit agencies considering LRT should switch their sights to AGT (although it requires far more costly, totally grade-separated construction). Some examples:

· "Benson Chin, project spokesperson for [Vancouver] SkyTrain's new Millennium Line, says the agency keeps its operating costs low because the system, like the proposed monorail here, is fully automated."
[Seattle Weekly, 25 September- 1 October 2002]

· [For an automated "Personal Rapid Transit system] "... the annual operating costs are a fraction of those for light rail, streetcars, or bus systems, because typically 80% of the operating costs for these other systems are payroll costs for drivers."
[Forward Quest, "Sky Loop Financial Plan" 2001/02/13]

· "While automated guideway transit systems may or may not be marginally cheaper per mile to build than light rail systems, their significantly cheaper operating costs are what can make them so attractive."
[Letter to the Editor, Christian Science Monitor 26 August 2002]

· "...the largest recurring cost in any enterprise are the salaries and benefits of the employees. An automated, driverless monorail drastically reduces this cost and provides the revolutionary possibility of a profitable transit system."
[Letter to the Editor, Seattle Times 22 August 2002]

· "Monorails never mix with traffic, which makes them safer and more reliable. This also means the system can be automated, which will allow it to operate with little or no ongoing subsidy -- something you don't hear light-rail systems boasting."
[Letter to the Editor, Seattle Times 22 August 2002]

· "A surface rail system will always require drivers .... However, a monorail system, being elevated, can be fully automated, allowing one to run trains without incurring the expense of a driver. A single set of operators can manage an entire network of trains from a central location, helping to reduce labor costs and removing this as a factor in determining what frequency of service and what service hours can be afforded the customers."
[Austin Monorail Project, 27 March 2002]

· "No ongoing tax subsidy. Once the monorail is built, it should be able to support itself entirely from passenger fares and advertising. No ongoing operating tax subsidies should be needed."
[Rise Above it All (Seattle), "Top Ten Reasons Why We Love the Monorail", Nov. 2002]

Certainly, there are advantages to automated operation – for example, automated operation may improve operational efficiency at very close headways. it should also be noted that one of the benefits of LRT, as a guided transit mode, is that it can be partially automated – i.e., automated in totally grade-separated sections (e.g., where several lines might come together at tight headways) and operated manually in lower-traffic sections.

indeed, the experimental operation of a totally automated PCC streetcar was introduced in the 1940s. Furthermore, if desired or necessary, LRT is capable of gradual conversion to fully grade-separated alignments and automated operation, affording transit agencies greater flexibility in staging transit development and incrementally introducing upgrades compatible with financial resources. And, in fact, both the Vancouver SkyTrain AGT and London Docklands AGT (Advanced LRT) represent adaptations of LRT technology to automated, grade-separated rapid transit applications.

However, the assumption that automated transit – such as monorail, Personal Rapid Transit (PRT), or maglev (magnetic-levitation) systems – inherently have significantly lower costs than manual light rail transit (LRT), does not seem supported by evidence. Modern, heavy-duty, urban transit systems in daily, real-world, revenue service are complex operations, not just "horizontal elevators". Many LRT designers and proponents point out that, while automated operation indeed does not require a driver in each vehicle or train, this advantage is typically offset by the need for additional security personnel, technicians, central control personnel, administrative personnel, and other staff requirements.

The pros and cons of the economics of automated vs. manual operation continue to be debated. For example, Dr. Vukan R. Vuchic, currently UPS Foundation Professor of Transportation Engineering at the University of Pennsylvania, acknowledged in his textbook Urban Public Transportation (Prentice-Hall, 1981) that automation could produce peak labor cost savings for fully grade-separated rail transit systems; nevertheless, he noted that automatic train operation (ATO) introduced considerably higher investment costs and system complexity. Thus, "in many medium-size cities it would be more beneficial to use the additional investment spent on ATO for network extensions."

In a paper for the 1988 National Conference on Light Rail Transit sponsored by the US Transportation Research Board (TRB Special Report 221), Gerald D. Fox, then an engineer for Portland's Tri-Met transit agency, compared the cost performance of new automated transit and manually operated LRT systems. Fox assessed automated guideway transit (AGT) systems in Lille (VAL), London (Docklands), and Vancouver (SkyTrain) vs. manually operated LRT systems in Nantes, Manila, Calgary, San Diego, and Portland.

Fox's conclusion: "When new AGT systems are compared with new LRT systems, or when AGT and LRT are compared on identical alignments, it appears that the cost of additional maintenance and supervising staff and additional 'non-staff' budget may exceed the savings that AGT systems achieve by eliminating operators." indeed, Fox's comparison of the two French systems in his study – Lille (AGT) and Nantes (LRT) – indicated that the operating cost per passenger-km was 114% higher for AGT in Lille.

Fox also compared a hypothetical LRT system, based on Portland's MAX, to Vancouver's SkyTrain, using grade-separated operation in the same corridor. His analysis concluded that LRT would cost about 31% less per passenger-km than the SkyTrain AGT.

In a subsequent paper, presented to the 6th National LRT Conference in 1992 (Transportation Research Record 1361), Tom Parkinson, a Vancouver consultant, provided data which compared the SkyTrain AGT very favorably with other North American and Canadian LRT and metro systems. While cost per passenger-mile or pasenger-km was not presented, Parkinson did show that staff productivity, in terms of annual passenger-miles per employee, was about 65% higher on SkyTrain than on LRT in San Diego, 83% higher than on LRT in Calgary, and 115% higher than on San Francisco's BART rapid transit system.

On the other hand, in terms of operating cost, Parkinson showed that while SkyTrain's cost per passenger-mile was about 36.4% of Vancouver's bus operating cost, the similar cost of Calgary's LRT was just 33.3% of the cost of its area bus service. This certainly does not suggest sharp operating-cost advantages for the automated system.

Nor do other available data for AGT (automated guideway) systems in the United States support the contention of monorail and other AGT enthusiasts that these systems provide "significantly cheaper operating costs" than LRT. The following tables show the operating and maintenance (O&M) cost per passenger-mile of selected relatively new LRT and all the currently operating AGT systems in urban revenue service in American cities, as reported to the Federal Transit Administration and included in the FTA's National Transit Database (NTDB, year 2000).

Table 1
Selected Light Rail Systems:
O&M Cost per Passenger-Mile

Table 2
US Urban-Revenue AGT Systems:
O&M Cost per Passenger-Mile

For all of the claims of the economic benefits of automation, and assurances of the "significantly cheaper operating costs" of monorails and other AGT systems vs. LRT, it is difficult not to notice from these two tables of data that each of the totally automated guideway systems has operating and maintenance costs measured in several dollars per passenger-mile vs. cents per passenger-mile for most LRT systems.

it is also useful to contrast the O&M cost of the totally automated Jacksonville monorail (and these other AGT systems) with the O&M cost of the manually operated Seattle monorail of $1.02 per passenger-mile. Clearly, with the automated system reporting a unit O&M cost more than 10 times higher than that of the manually operated system, a convincing case for the dramatic cost-slashing potential of automated operation is not made.

With respect to the Seattle monorail, it should be noted that, while it is a publicly owned facility, it has experienced operating surpluses from fare collections, even returning revenues on occasion to the City of Seattle – although it exhibits O&M costs higher than those of King County Metro's motor buses and electric trolleybuses. This raises some additional points in regard to profitability, which is in large part a function of (1) unique service features, (2) fare policy, and (3) ridership volume.

in the Seattle monorail's case, ridership volume is relatively tiny for a fixed-guideway line – less than 5,700 riders per day. But it's a major tourist attraction itself, connecting two other major tourist attractions – Seattle Center and Seattle's Westlake Mall, near the Space Needle – and providing a unique, direct connector service for tourists wishing to travel between those major activity centers. in addition, administration and operation of the monorail are contracted out to a private contractor, resulting in labor rates significantly below those prevalent for the unionized workforce of the citywide transit system.

Furthermore, the Seattle monorail is certainly not competing with freeways and other highway facilities, so fares for this unique experience can be set fairly high. The average trip length on the monorail is just under a mile, compared to 5.5 miles on Seattle's citywide transit system. For the monorail, the average fare revenue per trip averages $0.91, compared with $0.69 per trip on the city bus system. This suggests that the "profitability" of the monorail is less a factor of operating characteristics than of novelty and the provision of a unique service to a tourist-oriented niche market.

And the disparity per passenger-mile is even more striking: $1.02 for the monorail, vs. $0.13 for the citywide transit system. Basically, per passenger-mile, the monorail service is collecting nearly eight times the fare revenue that the citywide system is. This would lead any wise planner or decisionmaker to wonder whether region-wide commuters, school students, shoppers, and other common, "routine" travellers would really pay such a premium rate of fare for their daily transit rides.

in any case, these data and other results of this analysis do not appear to corroborate the contention that monorails, "PRT", or other automated systems would render phenomenal, or even significant, operating cost savings compared with LRT (such as the surface-routed system in Orleans, at left). On the contrary, the data suggest that fully automated systems may actually, in some cases, incur higher operating costs. Certainly, the evidence is enough to suggest caution in regard to claims of "gadget" mode vendors, "niche" consultants, and enthusiasts, and to promote very close scrutiny of factual data and the context of any alleged "economic miracles" in urban transit.