CEI Comments in Response to OST Regarding Impact of Automated Vehicle Technologies on Workforce
On behalf of the Competitive Enterprise Institute (“CEI”), I respectfully submit these comments in response to the Office of the Secretary of Transportation’s (“OST”) Request for Comments on the Scope of the Study on the Impact of Automated Vehicle Technologies on Workforce (“RFC”).
CEI is a nonprofit, nonpartisan public interest organization that focuses on regulatory policy from a pro-market perspective. This comment letter proposes expanding the scope of the RFC’s Statement of Work to include research questions related to employment impacts of low-cost automated vehicle taxi-style services on low-income, transit-dependent urban populations.
Exploring the Potential Employment Benefits of Future Automated Taxi Services Replacing Traditional Urban Mass Transit Services
In the RFC, OST notes two recent whitepapers examining the labor market effects of automated vehicles (“AVs”). The first, published by the Department of Commerce in 2017, examines potential workforce impacts by separating “motor vehicle operators” from “other on-the-job drivers,” noting that jobs in the latter category employed more than three times as many workers and that “other-on-the-job drivers” could stand to benefit from “greater productivity and better working conditions.”
The second, published by Securing America’s Energy Future in 2018, notes similar potential labor market effects while also highlighting benefits to commuters by automobile, examining potential declines in travel costs related to AV-related traffic congestion reductions and enhanced productivity through the elimination of time-on-task driving.
Both of these whitepapers make important contributions to the preliminary debate over the projected employment impacts of AVs. However, neither address a significant longstanding urban policy issue related to low-income residential location choice, means of transportation to work, and accessibility to metropolitan area employment opportunities.
Since the 1960s, urban economists have attempted to explain why low-income households tend to concentrate in urban cores within metropolitan areas. A popular theory given the state of transportation technology and residential spatial patterns is that the urban poor reside close to central business districts largely because of access to mass transit service and their inability to afford their own automobiles. Recent research has supported this hypothesis, finding “the primary reason for central city poverty is public transportation.”
Very few Americans rely on mass transit, with just 5 percent of American workers aged 16 years and older commuting to work via mass transit in 2017. In contrast, 76 percent of workers drove alone and 9 percent carpooled. Despite this, in 2017, mass transit received 28 percent of total federal, state, and local surface transportation funding—more than five times its commuting mode share and 11 times mass transit’s share of total commuting and non-commuting trips. Thus, the most compelling public interest argument for continued mass transit subsidies is transportation equity for the transit-dependent urban poor.
Unfortunately, even when lavishly subsidized, mass transit poorly serves low-income, transit-dependent urban populations. As explained below, metropolitan area job accessibility by mass transit compares poorly with job accessibility by automobile. But before discussing U.S. metropolitan area job accessibility data and the diminished work prospects of the transit-dependent urban poor, we will first explain the type of travel cost at the root of this problem.
Transportation users face two travel cost budgets: time and money. Given their relative lack of financial resources, the urban poor are more sensitive to financial costs of transportation and less sensitive to travel time costs. A key constraint to travel time is known as Marchetti’s constant. Marchetti’s constant posits that people are willing to spend on average one hour commuting each workday, or 30 minutes each direction. Recent empirical research examining mobile phone data from the U.S., Europe, and Africa supports Marchetti’s theory that this “universal law of commuting” holds across time, space, income, and culture.
The Accessibility Observatory of the Center for Transportation Studies at the University of Minnesota has been publishing empirical research on job accessibility in America’s largest 50 metropolitan areas by car, transit, and walking since 2013 in its Accessibility Across America series. This body of research shows, in keeping with Marchetti’s constant, the average share of jobs reachable by car in 30 minutes from home to work is 47.3 percent versus 1.12 percent by transit (see A-3, Table 3).
However, there are two major caveats. First, high-quality transit allows riders to engage in activities such as reading or napping that are unavailable to auto commuters, who must stay on task while driving, suggesting that quality differences would increase acceptable transit commute times above acceptable driver commute times. Second, most U.S. cities lack extensive and robust mass transit networks, and transit system usage is concentrated in a handful of very large metropolitan areas, namely New York City.
Even a doubling of Marchetti’s constant for mass transit (from 30 to 60 minutes; two hours of total daily commuting) and a one-third reduction in Marchetti’s constant for automobiles (from 30 to 20 minutes; 40 minutes of total daily commuting) does little to improve the standing of mass transit relative to driving in terms of metropolitan area job accessibility. The metropolitan area average for jobs accessible in 20 minutes of driving is 22.68 percent versus 8.1 percent in 60 minutes by transit.
Only in five of the six legacy transit cities, which account for the majority of total transit trips in the U.S., does mass transit outperform driving in this very transit-favorable job accessibility comparison:
- Boston, 10.59 percent versus 9.55 percent;
- Chicago, 7.65 percent versus 6.77 percent;
- New York, 14.55 percent versus 5.42 percent;
- San Francisco, 17.96 percent versus 12.77 percent; and
- Washington, 12.16 percent versus 8.52 percent.
In Philadelphia, the sixth legacy transit city, drivers can access 8.71 percent of metropolitan area jobs in 20 minutes versus the 7.41 percent of jobs reachable in 60 minutes by transit.
As these data indicate, mass transit performs poorly relative to private automobiles. Given the high capital and operating costs of transit relative to autos and the inherent first- and last-mile challenges of radial or grid-based transit networks, it is unlikely mass transit will ever be able to offer service capable of meaningfully reducing this accessibility gap. This means low-income, transit-dependent urban populations are condemned to limited employment prospects until they are able to afford superior transportation service.
Fortunately, AVs may finally be able to solve the “urban transportation problem.” Research published in 2018 by a team of Swiss academics suggests automated driving systems have the potential to reduce taxicab operating costs by 85 percent in urban settings and 83 percent in suburban and exurban settings.
Under this projection, automated taxi service costs on a passenger-mile basis would fall below present costs of providing rail and bus transit, and shared automated taxis are projected to be cheaper even than automated buses. If these cost savings are realized, the presently transit-dependent urban poor would be able to access automobility and reach jobs across their metropolitan areas currently inaccessible by transit.
OST asks in the RFC, “Should the [Statement of Work] be expanded…?” Given the above, the answer is yes. In conducting the AV workforce study, the Department should examine the potential employment prospect gains—particularly those for low-income, traditionally transit-dependent urban residents—in a world of affordable AV services.
We appreciate the opportunity to submit comments to OST on this matter and look forward to further participation.
Competitive Enterprise Institute
. Scope of the Study on the Impact of Automated Vehicle Technologies on Workforce, Request for Comments, DOT-OST-2018-0150, 83 Fed. Reg. 50,747 (Oct. 9, 2018) [hereinafter RFC].
. RFC, supra note 1, at 50,747-48.
. David Beede et al., The Employment Impact of Autonomous Vehicles, Office of the Chief Economist, Economics and Statistics Administration, U.S. Department of Commerce, ESA Issue Brief #05-17 (Aug. 11, 2017), at 1, available at https://www.commerce.gov/sites/commerce.gov/files/migrated/reports/Employment%20Impact%20Autonomous%20Vehicles_0.pdf.
. Richard Mudge et al., America’s Workforce and the Self-Driving Future: Realizing Productivity Gains and Spurring Economic Growth, Securing America’s Future Energy (Jun. 2018), at 25-26, available at https://avworkforce.secureenergy.org/wp-content/uploads/2018/06/Americas-Workforce-and-the-Self-Driving-Future_Realizing-Productivity-Gains-and-Spurring-Economic-Growth.pdf.
. See, e.g., John R. Meyer & John F. Kain, The Urban Transportation Problem (1965).
. See, e.g., Stephen F. LeRoy & Jon Sonstelie, Paradise Lost and Regained: Transportation Innovation, Income, and Residential Location, 13 J. Urb. Econ. 67 (1983).
. Edward L. Glaeser et al., Why Do the Poor Live in Cities? The Role of Public Transportation, 63 J. Urb. Econ. 1 (2008).
. U.S. Census Bureau, 2017 American Community Survey 1-Year Estimates, Table S0802 (Sep. 2018), available at https://factfinder.census.gov/bkmk/table/1.0/en/ACS/17_1YR/S0802.
. Congressional Budget Office, Public Spending on Transportation and Water Infrastructure, 1956 to 2017 (Oct. 2018), available at https://www.cbo.gov/system/files?file=2018-10/54539-Infrastructure.pdf.
. Federal Highway Administration, Person Trips by Transportation Mode, 2017 National Household Travel Survey (Mar. 2018), available at https://nhts.ornl.gov/person-trips.
. Cesare Marchetti, Anthropological Invariants in Travel Behavior, Tech. Forecasting & Soc. Change 45 at 75-88 (1994).
. Kevin S. Kung et al., Exploring Universal Patterns in Human Home-Work Commuting from Mobile Phone Data, 9 PLoS ONE 6 (2014), available at https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0096180.
. Author’s calculation using 2017 auto data and 2017 transit data from the Access Across America series where jobs accessible by mode in 10-minute increments from 10 to 60 minutes are divided by total jobs. This is an analysis of the 49 largest metropolitan areas in the U.S., excluding Memphis, which lacked appropriate transit data for 2017. See Appendix A for the complete dataset.
. Patrick M. Bösch et al., Cost-based analysis of autonomous mobility services, 63 Trans. Pol’y 76, 82 (May 2018).
. RFC, supra note 1, at 50,749.