The importance of driverless trains

As Honolulu is making progress on its driverless elevated rail system under construction, Washington, DC is finally beginning to return to computer operation on its red line after a 2009 crash brought an end to reliance on the computerized system. While the move in DC will facilitate smoother driving and braking, WMATA still relies on train operators in the cabs, forgoing the cost-saving opportunity that driverless systems provide. It’s difficult to overstate the importance of driverless trains in the effort to bring U.S. transit operations down to a reasonable price.

Driverless systems currently operate successfully in cities from Vancouver to Algiers, and the world’s most financially successful intracity transit systems in Hong Kong and Tokyo have embraced the technology. In spite of WMATA’s high profile accident that happened while the trains were computer-operated, a well-designed driverless system is actually safer than human operated one. Driverless systems offer a better ride quality, stay on time, and face a lower marginal cost of extending service hours.

Labor costs make up huge shares of U.S. transit systems. In DC, for example, personnel costs make up 70% of the agency’s operating budget. In 2010, WMATA spent $38 million on the salaries of 611 train operators, and this does not include their retirement and health benefits. In New York, personnel costs make up $8.5 billion of the agency’s $11.5 billion operating costs, and in Chicago labor takes up 73% of CTA’s operating expenses. Obviously not all transit workers jobs can be automated (all of these systems have more bus drivers than train operators) and some operating costs would rise under a driverless system. But taking steps toward reducing labor — that comes at a premium in high-cost-of-living cities where transit is most important — is crucial for reducing transit’s operating costs and making transit systems financially sustainable.

In all sorts of industries automation reduces the cost of goods and services, but transit systems face particularly high returns on automation because several institutional factors inflate their high personnel costs. As Stephen and Alon have explained, union work rules play a role in driving transit costs by requiring eight-hour shifts. Transit agencies face peak demand during the morning and evening rush hours. If transit agencies were run on a for-profit basis, they would staff more bus drivers and  train operators during these times of peak demand and fewer during the work day and night time. However, transit unions’ work rules make it impossible to staff according to demand.

In addition to the premium that transit agencies guarantee their employees through relatively high wages and union work rules, their pensions and benefits make up a large part of their employment costs, and these costs are not transparent. For example, BART, with its high profile strikes this summer, reports a $187 million unfunded pension liability. This means that future operating costs will have to rise to cover benefits accrued in the past. Transit agencies’ pension liabilities are based on their discount rate assumption with 7.5% or higher being typical. If these agencies’ pension funds fail to realize compound annual returns greater than or equal to their assumed discount rates, their pension liabilities will actually be higher than what they report.

A City Lab post posits that mainland cities are unlikely to follow Honolulu’s driverless lead because converting existing trains to driverless would be prohibitively costly. But a commenter on her post points out that the change to a driverless system would be a capital cost, typically covered by the federal government. What better use for the Federal Transit Administration’s Core Capacity dollars than making the transition to driverless trains in large systems? Moving to a driverless system could create a virtuous cycle better service increasing ridership, begetting further service improvements. While making the transition to a driverless system entails short-term political and financing challenges, maintaining bloated operating expenses year after year is an unacceptable outcome.

During BART’s infamous strikes over the summer, a San Francisco tech entrepreneur quipped, “Get ‘em back to work, pay them whatever they want, and then figure out how to automate their jobs so this doesn’t happen again.” His sentiment may come off as cold-hearted, but transit agencies should have one mission: providing adequate transit service at a reasonable cost. Their mission should not be to provide well-paying jobs to workers who might not be able to earn such high wages and benefits in the private sector. While the transition to driverless would be difficult for transit workers and agencies, in the long-run the advantages of substituting relatively inexpensive capital for expensive labor are too high to ignore.