The Role of Transit, Shared Modes, and Public Policy in the New Mobility Landscape (2021)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

27 The purposes of this chapter are to address elements of the committee’s charge to consider (1) how shared mobility options “can improve the trans- portation system’s ability to further goals such as accessibility, efficiency, equity, safety, and sustainability”; (2) “the relationship to and impact of these services on existing public transit”; (3) “ways that transit agencies have coordinated with the new mobility providers”; and (4) the roles that shared modes could play “in the provision of transportation services as part of regional transportation systems, and specifically the relationship to ... existing public transit.” The next section sketches out what is known about how shared mobil- ity providers are affecting the social and environmental goals listed above. The section that follows summarizes estimates from surveys, statistical studies, and models on the impacts of shared modes on transit rider- ship. The third section reviews U.S. experience in transit agency coordina- tion with shared mode providers (experience in other nations is described in Chapter 3). The fourth section builds on the models of coordination described in the third section to illustrate how they might apply outside of urban cores where most fixed-route transit service is provided. The final section summarizes the findings the committee draws from the evidence presented in the chapter. SHARED MODE SERVICE TO SOCIAL GOALS An emerging body of evidence suggests how shared mobility can serve social goals such as “accessibility, efficiency, equity, safety, and sustainability.” As 2 Shared Mobility and Public Transportation

28 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY will be described, the contributions of shared modes to social goals vary across modes and they may not always be positive. Conclusions about users of shared modes, changing travel choices, and the effects these may have are often stated tentatively in this chapter because these subjects are dif- ficult to study.1 Data on the characteristics of shared mode users and how these characteristics affect their modal preferences are not readily available. Some studies rely on convenience (non-random) samples, often of early adopters whose preferences may not be representative. Moreover, some of the shared modes themselves have evolved quickly in recent years even as they have grown and increased their appeal to a wider range of users, thus raising questions about the relevance of earlier studies. This section describes what is known in illustrative, broad-brush fashion, drawing on available empirical studies and surveys of North American experience that point to both positive and negative outcomes. The intent is not to provide a comprehensive assessment but rather to point to areas where public policies may be needed to steer the effects of shared mode impacts toward posi- tive contributions. (Policy strategies that governments can advance at the metropolitan area scale in this regard are described in Chapters 4 and 5.) Accessibility and Equity Equity has several different dimensions; this report focuses on equality of access to shared modes across income levels and by travelers with disabili- ties.2 For ridehailing, the most recent estimates of users are based on sur- veys of specific areas and studies of specific cities. For example, a random sample of California ridehail users finds that predominant users tend to be “well-educated independent Millennials and young Gen Xers who do not have children and live in urban areas.”3 The least likely users are less educated individuals living in lower-income households (less than $50,000 annually).4 There is evidence from trip data in some jurisdictions, however, 1 Circella, G., and F. Alemi. 2018. Transport Policy in the Era of Ridehailing and Other Dis- ruptive Transportation Technologies. Advances in Transport Policy and Planning, Volume 1, edited by Y. Shiftan and M. Kamargianni, Elsevier, Chapter 5, pp. 119–144. 2 For discussion of a wider range of equity issues and policy responses regarding usage and availability of shared modes, see Shaheen, S., et al. 2017. Travel Behavior: Shared Mobility and Transportation Equity. Federal Highway Administration, U.S. Department of Transportation. https://www.fhwa.dot.gov/policy/otps/shared_use_mobility_equity_final.pdf. 3 Circella and Alemi. 2018. 4 Circella, G., et al. 2018. The Adoption of Shared Mobility in California and Its Relation­ ships with Other Components of Travel Behavior. National Center for Sustainable Transpor- tation, University of California, Davis. https://escholarship.org/uc/item/1kq5d07p. See also Alemi, F., G. Circella, S.L. Handy, and P.L. Mokhtarian. 2018. What Influences Travelers to Use Uber? Exploring the Factors Affecting the Adoption of On-Demand Ride Services. Travel Behaviour and Society 13. https://doi.org/10.1016/j.tbs.2018.06.002.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 29 that ridehailing companies are improving access in urban neighborhoods outside of the central business district of Manhattan,5 and reaching low- income African American neighborhoods in Los Angeles with relatively little transit service and doing so with higher-frequency service and less cost than taxis.6 (Whether the cost of trips required to make ridehail companies profitable will provide future benefits remains to be seen.) Ridehailing in Los Angeles appears to significantly reduce racial/ethnic discrimination in driver acceptance and response times, compared with significant discrimi- nation by taxi drivers against African Americans compared with whites.7 However, a randomized experiment in Seattle and Boston showed signifi- cantly longer wait times for African American passengers in Seattle and more frequent cancellation of rides for riders with African American sound- ing names in Boston.8 Wait times for ridehail trips for Chicago residents suggest considerable disparities between whites and African Americans and small disparities between non-Hispanic whites and Hispanics, which may be due to the presence of much less service avail ability in Chicago’s African American neighborhoods.9 There is some evidence of ridehail driver discrimination against LGBT riders,10 albeit whether this is less or more discriminatory than taxi services is an open question. Access to ridehail- ing by disabled individuals has long been a source of controversy; ride- hailing companies generally contend that the Americans with Disabilities Act of 1990 (ADA) does not apply to their services. Access to ridehailing also depends on user access to smartphones and bank accounts, which some low-income individuals lack. ADA access issues and how they are addressed in ridehailing–public transit coordination efforts are described in the third section of this chapter and in Chapter 5. Early studies of bikeshare use in individual cities, mostly based on convenience samples of docked bikeshare users, report that users tend to 5 For expanded service provided by ridehailing providers compared to taxis in the boroughs of New York City, see Schaller, B. 2017. Unsustainable? The Growth of App­Based Ride Services and Traffic, Travel, and the Future of New York City. http://www.schallerconsult. com/rideservices/unsustainable.pdf. 6 For service by Lyft across neighborhoods in Los Angeles, see Brown, A. 2018. Ridehail Revolution: Ridehail Travel and Equity in Los Angeles. Doctoral Dissertation. University of California, Los Angeles. https://escholarship.org/uc/item/4r22m57k. 7 Brown. 2018. 8 Ge, Y., et al. 2016. Racial and Gender Discrimination in Transportation Network Com­ panies. National Bureau of Economic Research, Working Paper 22776. https://www.nber.org/ papers/w22776. 9 Irvin, E. 2019. Is Ridehailing Equitably Available Across Chicago? CNT. https://www.cnt. org/blog/is-ridehailing-equitably-available-across-chicago. 10 Mejia, J., and C. Parker. 2018. When Transparency Fails: Bias and Financial Incentives in Ridesharing Platforms. Kelley School of Business Research Paper 18-59. https://papers.ssrn. com/sol3/papers.cfm?abstract_id=3209274.

30 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY be younger, well-educated, middle- to high-income, non-Hispanic white residents of urban areas living in childless households.11 Based on a 2019 representative survey of 15,000 respondents in 18 metro areas, micro- mobility users are younger and somewhat wealthier than non-users, but the genders and races of users are proportional to their representation in the general population.12 This is generally true for users of both shared bikes and shared e-scooters. Although higher-income earners are over represented among e-scooter users, about 40 percent of users have incomes below $50,000. Across age groups, very few users of micro mobility are over age 55, perhaps due to older users’ safety concerns. Reports from pilot programs in Arlington, Virginia; Chicago, Illinois; and San Francisco, California, that included city efforts to encourage equitable access to e-scooters across neighborhoods of varied race and income have had mixed results.13,14,15 A rigorous comparison of the docked and dockless bikeshare systems in San Francisco estimates that dockless bikes are more accessible than docked bikes to residents of areas defined as communities of concern.16 Dockless bikes and scooters left cluttering sidewalks and poorly placed stalls for micromobility parking, however, can impede people using wheelchairs and those with restricted mobility.17 Based on convenience samples, early adopters of traditional carshar- ing (short-term rentals provided by a private company) tended to be well- educated, middle- to higher-income young adults living in households averaging two persons and, typically, without a personal automobile.18 Little is known about the demographics of peer-to-peer carshare users (pri- vate individuals being reimbursed for short-term use of their personal cars by others). Although the barriers to wider use of carsharing and bikesharing 11 Shaheen, S., and A. Cohen. 2019. Shared Mobility Policy Toolkit: Docked and Dockless Bike and Scooter Sharing. https://escholarship.org/uc/item/00k897b5. 12 Populus Groundtruth Survey as summarized in SUMC (Shared-Use Mobility Center) et al. 2019. Transit and Micro-Mobility: Interim Report. TCRP J-11, Task 37. 13 Mobility Lab and Arlington County Commuter Services. 2019. Arlington County Shared Mobility Pilot Evaluation Report. https://1105am3mju9f3st1xn20q6ek-wpengine.netdna-ssl. com/wp-content/uploads/2019/11/ARL_SMD_Evaluation-Final-Report-1112-vff-2.pdf. 14 City of Chicago. 2019. E­Scooter Pilot Evaluation. https://www.chicago.gov/content/dam/ city/depts/cdot/Misc/EScooters/E-Scooter_Pilot_Evaluation_2.17.20.pdf. 15 SFMTA (San Francisco Municipal Transportation Agency). 2019. Powered Scooter Share Mid­Pilot Evaluation. Appendix A: User Survey. https://www.sfmta.com/sites/default/files/reports- and-documents/2019/08/powered_scooter_share_mid-pilot_evaluation_ appendices_ final.pdf. 16 Qian, X., et al. 2020a. Enhancing Equitable Service Level: Which Can Address Better, Dockless or Dock-Based Bikeshare Systems? Journal of Transport Geography 86. https://doi. org/10.1016/j.jtrangeo.2020.102784. 17 Shaheen and Cohen. 2019. 18 Dill, J., et al. 2017. Peer­to­Peer Carsharing: Short­Term Effects on Travel Behavior in Portland, OR. Transportation Research and Education Center, Portland State University. https://pdxscholar.library.pdx.edu/trec_reports/133.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 31 by residents in low-income areas have been well described,19 little empirical evidence is available about the appeal and use of carsharing in low-income communities.20 The equity impacts of microtransit have not yet been documented; ongoing pilot projects and experimentation with microtransit are described in the next section. Efficiency For this report efficiency is defined as having two components: (1) for the individual user in terms of monetary and time costs of travel and (2) for users collectively in terms of travel times and user experience on road networks and transit systems. Regarding individuals, the popularity of, and rapid growth in, ridehailing and micromobility, especially e-scooters, described in Chapter 1 indicates that many users are finding them more attractive ways of traveling than their alternatives, presumably because of their time or monetary advantages. The growth in use of ridehailing since 2012 has been extraordinary. By 2018, total trips on Uber and Lyft in the United States had tripled that of taxis and approached that of local bus transit trips.21 One study estimates that UberX users, collectively, would have been willing to pay as much as $6.8 billion more for their trips than the fares paid in 2015,22 an amount that has presumably grown since then. This estimate of what economists refer to as consumer surplus, however, does not account for costs to the public, such as safety and non-market costs resulting from increased congestion and emissions, as described in the following subsections. Regarding collective efficiency, studies reviewed in the second section of this chapter regarding ridehailing impacts on transit suggest that ride- hailing induces additional trips, draws riders away from transit and active modes, and increases total traffic in congested urban networks in New York City and San Francisco.23 (Ridehail traffic in off-peak periods would not be expected to have the same effects on congestion.) Carsharing and micromobility, also discussed in the next main section, reduce auto trips 19 Elliot, M., et al. 2020. An Evaluation of Free­Floating Carsharing in Oakland, California. Transportation Research Sustainability Center, University of California, Berkeley. https:// escholarship.org/uc/item/3j722968. 20 Hyun, K., and C. Cronley. 2019. Assessing the Viability of Car­Sharing for Low­Income Communities. Center for Transportation, Equity, Decisions and Dollars, The University of Texas at Arlington. https://ctedd.uta.edu/wp-content/uploads/2019/07/Carsharing_HyunCronley.pdf. 21 Schaller, B. 2018. The New Automobility: Lyft, Uber, and the Future of American Cities. http://www.schallerconsult.com/rideservices/automobility.htm. 22 Cohen, P., et al. 2016. Using Big Data to Estimate Consumer Surplus: The Case of Uber. NBER Working Paper. https://www.nber.org/papers/w22627.pdf. 23 See the interpretation of studies summarized in Table 2-1 in the next main section and the subsection that follows on impacts on congestion.

32 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY as well as transit trips. However, for almost all U.S. cities where these ser- vices are available, the scale of these modal alternatives has probably not grown sufficiently large to significantly affect road network travel speeds. To the extent that carsharing and micromobililty encourage users to shift away from transit during its most congested periods, they could relieve overcrowding on buses and trains.24 Safety Commuting by bicycle is associated with reduced risk of cardiovascular disease, cancer, and other causes of mortality, but cycling involves higher injury risks to individuals than walking or riding in an automobile, and considerably higher risks than riding in a bus.25 Moreover, shared bike riders infrequently use helmets that reduce the risk of head injury from bike crashes. Even so, a recent study estimates that New York City’s bikesharing program results in net public health improvement after accounting for the health benefits of active travel that cycling offers, bicycle-involved crashes, and cyclist exposure to vehicle emissions.26 The rapidly growing use of e- bikes would presumably compound the crash injury risks to cyclists due to the higher speeds e-bikes are capable of sustaining,27 but results from studies conducted in Israel28 and the Netherlands,29 where e-bike use is more widespread than in the United States, are inconsistent in this regard.30 24 See the subsection on micromobility impacts on transit in the next main section. 25 Beck, L., et al. 2007. Motor Vehicle Crash Injury Rates by Mode of Travel, United States: Using Exposure-Based Methods to Quantify Differences. American Journal of Epidemiology 166(2). https://www.researchgate.net/publication/6378823_Motor_Vehicle_Crash_Injury_Rates_ by_Mode_of_Travel_United_States_Using_Exposure-Based_Methods_to_Quantify_ Differences. 26 Hou, Z., et al. 2020. Do Special Bike Programs Promote Public Health? A Case Study of New York City’s Citi Bike Bike Sharing Program. Center for Transportation, Environment, and Community Health, University of California, Davis. https://cpb-us-w2.wpmucdn.com/ sites.coecis.cornell.edu/dist/6/132/files/2020/06/UCD_YR3_ZHANG_FINAL_DO-SPECIAL- BIKE-PROGRAMS.pdf. 27 Zmora, O., K. Peleg, and Y. Klein. 2019. Pediatric Electric Bicycle Injuries and Com- parison to Other Pediatric Traffic Injuries. Traffic Injury Prevention 20(5):540–543. http:// dx.doi.org/10.1080/15389588.2019.1608361. 28 Siman-Tov, M., I. Radomislensky, and K. Peleg. 2018. A Look at Electric Bike Casualties: Do They Differ from the Mechanical Bicycle? Journal of Transport & Health 11:176–182. http://dx.doi.org/10.1016/j.jth.2018.10.013. 29 Schepers, J., et al. 2014. The Safety of Electrically Assisted Bicycles Compared to Classic Bicycles. Accident Analysis and Prevention 73c. https://doi.org/10.1016/j.aap.2014.09.010. 30 Fishman, E., and C. Cherry. 2015. E-bikes in the Mainstream: Reviewing a Decade of Research. Transport Reviews, July. Also included in this review are safety studies from China, where e-bike use is the highest across nations, but the results lack transferability because of rare use of helmets and because of the use of a much wider range of powered bikes that are not all defined as e-bikes in Europe or the United States. http://dx.doi.org/10.1080/01441647. 2015.1069907.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 33 The rapid increase in e-scooter use after 2017, followed by a spike in injuries, most commonly head injuries, raised concerns about helmet availability and use.31,32 Early evidence of e-scooter injuries from Austin, Texas, found that one-third of the injuries occurred on the first ride, half of the injuries were severe (most commonly fractures), 15 percent had evi- dence suggestive of severe brain injuries, and riders rarely wore helmets.33 Even so, as a new and fast-growing mode, the risks of micromobility are not well documented. Although considerable research is under way, at the time of this writing, good evidence is lacking about the extent and types of injuries and travel, which makes it difficult to make definitive state- ments about comparative risk. Moreover, the growing use of a plethora of new small, powered personal vehicles, including powered skateboards, e-bikes capable of speeds in excess of 20 mph, and e-mopeds capable of speeds of almost 30 mph, would have different risk profiles. This lack of understanding makes it difficult to characterize risk holistically. In principle, e-scooters draw travelers away from other modes, including walking and non-motorized cycling.34 This would reduce physical activ- ity important to personal health, but they also draw trips away from automobiles, which would reduce emissions harmful to general health and risks to pedestrians, but also expose riders to increased personal risk of crash involvement. Increased ridehail trips for late night weekend trips may be associ- ated with declines in alcohol-involved crashes, as early studies suggested, but a more thorough study found no relationship between the beginning of Uber service in a metropolitan area and downward trends in traffic fatalities.35 A study using a comparison group of cities without Uber and Lyft service, however, finds a 2 to 4 percent increase in overall traffic fatalities after Uber and Lyft begin providing service, with an estimated 31 Aizpuru, M., et al. 2019. Motorized Scooter Injuries in the Era of Scooter-Shares: A Review of the National Surveillance System. The American Journal of Emergency Medicine 37(6). https://www.ajemjournal.com/article/S0735-6757(19)30215-3/fulltext. 32 Namiri, N,. et al. 2020. Electric Scooter Injuries and Hospital Admissions in the United States 2014–2018. JAMA Surgery 155(4). https://jamanetwork.com/journals/jamasurgery/ article-abstract/2758159. 33 Austin Public Health. 2019. Dockless Electric Scooter­Related Injuries Study. https:// austintexas.gov/sites/default/files/files/Health/Epidemiology/APH_Dockless_ Electric_Scooter_ Study_5-2-19.pdf. 34 Populus Groundtruth survey as summarized in SUMC et al. 2019. 35 Brazil, N., and D. Kirk. 2016. Uber and Metropolitan Area Traffic Fatalities in the United States. American Journal of Epidemiology 184(3).

34 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY loss of consumer welfare that matches or exceeds the benefits cited in the previous section.36,37 Personal security of ridehailing received considerable attention in late 2019 with the release of Uber’s first safety report documenting murders, rapes, and sexual assaults in 2018.38 These events were rare—occurring on 0.0002 percent of rides—in which drivers and riders were equally victim- ized and for which comparable statistics are not available from taxi and limousine rides or transit.39 Sustainability This report focuses on the vehicle emissions component of sustainability, and this section briefly summarizes literature reviews of studies assess ing use of shared modes and their effects on automobile emissions. In principle, trips by shared modes that substitute for auto trips would reduce emissions, and growing reliance on shared modes might affect auto ownership and use, but, in practice, the net effects depend on how people adjust their travel preferences and which shared modes they choose. Convenience surveys done in 2008 of round-trip carshare users in the United States and Canada reported reductions in auto ownership and use as a result of becoming carsharing subscribers,40 a conclusion supported in a separate review of the U.S. and international carshare literature.41 This same review concludes that although several studies of one-way car sharing have been done, the travel behavior of users of one-way carsharing remains 36 Barrios, J., et al. 2018. The Cost of Convenience: Ridesharing and Traffic Fatalities. Becker- Friedman Institute for Economics, University of Chicago, Working Paper 2018-80. https://bfi. uchicago.edu/working-paper/the-cost-of-convenience-ridesharing-and-traffic-fatalities. 37 Note that a 2 to 4 percent increase in fatalities resulting from increased vehicle miles trav- eled (VMT) due to ridehailing would result in small numbers of increased traffic deaths in indi- vidual metropolitan areas. For example, in 2018 Uber and Lyft added about 100 million VMT to the Chicago metropolitan area. If all these VMT were a net increase, one would expect about one additional fatality, assuming a fatality rate of about 1 death per 100 million VMT. Chicago VMT estimates taken from Balding, M., et al. 2019. Estimated Percent of Total Driving by Lyft and Uber. https://drive.google.com/file/d/1FIUskVkj9lsAnWJQ6kLhAhNoVLjfFdx3/view. Fatality rate from National Highway Traffic Safety Administration. 2019. Traffic Safety Facts 2017. U.S. Department of Transportation, Table 2. https://crashstats.nhtsa.dot.gov. 38 Conger, K. 2019. Uber Says 3,045 Sexual Assaults Were Reported in the U.S. Last Year. The New York Times, December 6. https://www.nytimes.com/2019/12/05/technology/uber- sexual-assaults-murders-deaths-safety.html. 39 Comparability can be complicated in other ways. For example, crimes can happen at bus stops and walking to and from them, and are not necessarily reported as transit related. 40 Shaheen, S., et al. 2016. Shared Mobility: Current Practices and Guiding Principles. U.S. Department of Transportation, p. 35 and Appendix A, Table 3. https://ops.fhwa.dot.gov/ publications/fhwahop16022/fhwahop16022.pdf. 41 Circella et al. 2018.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 35 unexamined. As noted previously, little is known about peer-to-peer car- share usage. Round-trip carshare membership in North America is asso- ciated with increased use of transit, walking, and cycling, reduced use of automobiles, and lowered greenhouse gas (GHG) emissions.42 Convenience surveys of early North American bikeshare programs in- dicate that one-quarter to half of respondents report driving less as a result of bikesharing.43 Survey respondents participating in city pilot projects to date allowing dockless bikesharing consistently report reductions in trips by motor vehicles.44 A survey of e-scooter users in 18 metro politan areas indicates that 41 percent of the most recent e-scooter trips would have been made by automobile if the e-scooter option had not been available.45 An exception to micromobility sustainability benefits could be for the first generation of e-scooters, which increased life-cycle GHG emissions com- pared to all modes they replaced but automobiles given the short life cycles of first-generation e-scooters and the rebalancing practices of the time.46,47 E-scooter vendors subsequently began upgrading their products for both productivity and environmental reasons.48 The impact of ridehailing on auto ownership and use remains uncer- tain, but, as described in detail in the next main section of this chapter, ridehailing may reduce use of transit, cycling, and walking and has been estimated to increase total vehicle miles traveled (VMT) and emissions in major metropolitan areas.49,50,51 Notably, unless and until vehicles operate 42 Shaheen and Cohen. 2019. 43 Shaheen et al. 2016, Appendix A, Table 6. 44 SUMC et al. 2019, p. 20. 45 SUMC et al. 2019, p. 42. 46 Hollingsworth, J., et al. 2019. Are E-Scooters Polluters? The Environmental Impact of Shared Dockless Electric Scooters. Environmental Research Letters 14. https://iopscience.iop. org/article/10.1088/1748-9326/ab2da8. 47 Rebalancing is the practice of collecting and moving both docked and dockless vehicles, typically at the end of the day, to locations where the majority of trip origins are expected to occur or in compliance with local government regulations. 48 Note that more rugged second-generation e-scooters may have longer life cycles. See Hawkins, A. 2020. Skip Pulls Back the Curtain on the High Costs of Electric Scooter Mainte- nance. The Verge, January 21. https://www.theverge.com/2020/1/21/21072785/skip-electric- scooter-life-cycle-maintenance-costs. 49 Schaller. 2018, Figure 1. 50 Graehler, M., et al. 2019. Understanding the Recent Transit Ridership Decline in Major U.S. Cities: Service Cuts or Emerging Modes? Presented at 98th Annual Meet- ing of the Transportation Research Board, Washington, DC. https://www.researchgate.net/ publication/330599129_ Understanding_the_Recent_Transit_Ridership_Decline_in_Major_US_ Cities_Service_Cuts_or_ Emerging_Modes. 51 Qian, X., et al. 2020b. Impact of Transportation Network Companies on Urban Con- gestion: Evidence from Large-Scale Trajectory Data, Sustainable Cities and Society 55. https://doi.org/10.1016/j.scs.2020.102053. http://www.sciencedirect.com/science/article/pii/ S2210670720300408.

36 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY on electric power, the typical (solo passenger) ridehail trip adds 69 percent more carbon-equivalent vehicle emissions than the trips they replace.52 Not only does ridehailing appear to draw trips away from walking, cycling, and transit, such trips also require more miles traveled than the personal auto trips they replace due to driver deadheading between passenger drop-offs and subsequent pick-ups. (These carbon-equivalent emissions increases compared with solo personal vehicle trips would also apply to taxis.) A critical review of ridehail travel behavior studies on mode shift and extra driving per ridehail trip concludes that ridehailing probably increases VMT and GHGs, though this review notes the uncertainty behind this conclusion and does not estimate the potential magnitude of increased travel on GHG emissions.53 The scale to date of shared modes matters in considering the conse- quences described in this section. As described in Chapter 1, round-trip car- share subscribers represent roughly half of 1 percent of licensed drivers and likely drive far less than the average motorist. Although micromobility trips grew 60 percent between 2018 and 2019, the 136 million micro mobility trips is a tiny percentage of the roughly 370 hundred million annual auto, transit, walking, cycling, and other passenger trips.54 Ridehail trips had grown by 2018 to rival those of bus transportation, which is about 1 per- cent of trips, and therefore have become an increasingly important compo- nent of urban passenger travel. The scale of shared modes would certainly be larger and more consequential in urban areas; their use is growing fast, and use may continue to grow as technology and platforms make it easier to access them. These trends help explain the wide interest of state and local policy makers and planners in shaping the impacts of shared modes on social goals. IMPACTS ON TRANSIT This section focuses on the known impacts of shared mobility services on public transit ridership in North America by synthesizing the results 52 Anair, D., J. Martin, M. Pinto de Moura, and J. Goldman. 2020. Ride­Hailing’s Climate Risks: Steering a Growing Industry Toward a Clean Transportation Future. Cambridge, MA: Union of Concerned Scientists. https://www.ucsusa.org/resources/ride-hailing-climate-risks. 53 Rodier, C. 2018. The Effects of Ride Hailing Services on Travel and Associated Green­ house Gas Emissions. White Paper, National Center for Sustainable Transportation, University of California, Davis. https://escholarship.org/uc/item/2rv570tt. 54 Shared mobility trips from NACTO (National Association of City Transportation Officials). 2020. Shared Mobility in the United States in 2019. https://nacto.org/shared- micromobility-2019; U.S. Department of Transportation. 2017. Total trips from National Household Travel Survey: Summary of Travel Trends 2017, Table 1d. https://nhts.ornl.gov/ assets/2017_nhts_summary_travel_trends.pdf.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 37 reported in the literature to date, starting with estimates of the effects of ridehailing services followed by what is known about the effects of micro- mobility and other shared modes. Ridehailing Assessing the impact of ridehailing on transit ridership is challenging for several reasons. Ridehailing has been growing remarkably fast, while also evolving as it does so, which makes any assessment a snapshot in time of a changing phenomenon. Moreover, despite the remarkable creativity by researchers in studying ridehailing and its impacts in a rapidly expand- ing body of literature, the general lack of detailed data on ridehail trips hampers the ability to estimate ridehail impacts on transit ridership while controlling for other influences. Detailed ridehail trip data are available in North America for only Chicago, New York City, and Toronto, but studies of specific cities or metro areas may not generalize due to the expected vari- ations in impact across different geographic areas. Differences in the density of development and population, walkability, transit availability, climate, demographics, auto ownership, congestion, and the scale of ridehailing in the locales studied could significantly influence study results. With the above caveats in mind, this section provides an overview of study results to date, as summarized in Table 2-1. This table is based on selected studies chosen because of their use of a range of methods. In the discussion that follows, more recent studies are emphasized over the earli- est ones due to the evolving nature of ridehailing. Studies listed are also limited to North America because of concerns that European and Asian results may not be indicative of experience in the United States due to the large differences in the built environments, transit orientation of cities, income levels, and public policies across continents. The committee notes, however, comparable international results in the text when such results are available. Table 2-1 also includes studies that estimate the impact of ridehailing on VMT and congestion because these have direct impacts on transit bus speeds and indirect impacts on bus ridership, the significance of which becomes apparent in the subsequent section. The studies reviewed in this section have limitations of various kinds that constrain the inferences that can be drawn from them. Early intercept studies relied on small samples and were asking questions about use of a relatively new mode to which many users had not yet fully adapted. Not all of the broader user surveys cited in this section are based on random sampling and therefore may not be generalizable to a larger population, and the surveys that do use random sampling may only be representative of the state or metropolitan areas from which they sample. The results of modeling studies on individual large cities or metropolitan areas also may

38 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY TABLE 2-1 Illustrative Studies of Ridehail Impact on Transit and Traffic Author/Year Geography Method Results Rayle et al., 201655 San Francisco Intercept survey n=380 33% of ridehail users would have used transit for most recent trip. About 5% of ridehail trips appeared to link to transit. About 8% of respondents indicated they would not have made the trip had ridehail not been an option (induced travel). Gehrke et al., 201956 Boston Intercept survey n=944 59% of ridehail trips added a new vehicle to the road. 59% used ridehail instead of transit because of time savings. 20% of ridehail trips were pooled with other riders. Henao, 201757 Denver Intercept survey n=311 ridehail riders 22% of ridehail users would have used transit; 4.5% would have used taxis. 12% of users indicate that most ridehail trips would not have occurred without the availability of this mode. Clewlow and Mishra, 201758 Seven U.S. metro areas (Boston, Chicago, Los Angeles, New York City, San Francisco, Seattle, Washington, DC) Online survey, n=4,094 over 2015–2016 6% of ridehail users reduced bus transit and 3% reduced light-rail use. 3% increased subway use. Ridehail users likely contribute to traffic congestion. Availability/cost of parking main reason for choosing ridehail instead of driving. 55 Rayle, L., et al. 2016. Just a Better Taxi? A Survey-Based Comparison of Taxis, Transit, and Ridesourcing Services in San Francisco. Transport Policy 45:168–178. https://www.ce.berkeley. edu/sites/default/files/assets/news/Ridesourcing_Taxi_Transit_SF_TransportPolicy.pdf. 56 Gehrke, S., et al. 2019. Substitution of Ride-Hailing Services for More Sustainable Travel Options in the Greater Boston Region. Transportation Research Record 2673(1). https:// journals.sagepub.com/doi/full/10.1177/0361198118821903. 57 Henao, A. 2017. Impacts of Ridesourcing—Lyft and Uber—on Transportation Including VMT, Mode Replacement, Parking, and Travel Behavior. Doctoral Dissertation. http://digital. auraria.edu/content/AA/00/00/60/55/00001/Henao_ucdenver_0765D_10823.pdf. 58 Clewlow, R., and G. Mishra. 2017. Disruptive Transportation: The Adoption, Utilization, and Impacts of Ride-Hailing in the United States. Institute of Transportation Studies, Univer- sity of California, Davis. https://www.reginaclewlow.com/pubs/2017_UCD-ITS-RR-17-07.pdf.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 39 TABLE 2-1 Continued continued Author/Year Geography Method Results Alemi et al., 202059 California California Mobility Panel study–2018 travel survey Most recent (non-pooled) ridehail trip substituted for drive alone (28%), carpool (17%), bus (8%), light rail (4%), commuter rail (1%), and taxi (26%). About 9% of respondents indicated availability of ridehail induced their most recent trip. Ridehail alternatives and use very different across young urbanites, older car-oriented suburbanites, and older transit-oriented suburbanites. Feigon and Murphy, 201860 Service areas of MARTA, NJ Transit, BART, WMATA61 Surveys of transit users about ridehail trips 11% to 17% of most recent ridehail trips substituted for transit (not counting WMATA due to service issues at the time). 6% to 16% connected to transit (not counting WMATA). Faster trip time (less waiting) main reason for substituting ridehail trip for transit trip. Feigon and Murphy, 2018 Chicago, Los Angeles, Nashville, San Francisco, Seattle, Washington, DC Daily and hourly ridehail trips at zip code level for May 2016 Ridehail complements transit. Three- quarters of ridehail trips occur outside of peak period, though most trips occur in urban core. Three of six metro areas increased transit ridership between 2010 and 2016. Hall et al., 201862 Metropolitan statistical areas in the United States Statistical analysis (differences in differences regression) For the average transit agency, entry of Uber increases transit ridership, with larger effects in large cities and for small transit agencies. However, Uber seems to be decreasing ridership in larger agencies. 59 Alemi, F., G. Circella, G. Matson, and D. Sperling. 2020. Insights from the California Mobility Panel Study: Evolution in the Use, Impacts, and Limitations on the Use of (Shared) Ridehailing. Presented at 99th Annual Meeting of theTransportation Research Board, Washington, DC. Paper 20-02731. 60 Feigon, S., and C. Murphy. 2018. Broadening Understanding of the Interplay Among Public Transit, Shared Mobility, and Personal Automobiles. Transit Cooperative Research Program Research Report 195. http://www.trb.org/Publications/Blurbs/177112.aspx. 61 Metropolitan Atlanta Rapid Transit Authority (MARTA); New Jersey Transit Corpora- tion (NJ Transit); San Francisco Bay Area Rapid Transit (BART); Washington Metropolitan Area Transit Authority (WMATA). 62 Hall, J.D., C. Palsson, and J. Price. 2018. Is Uber a Substitute or Compliment for Public Transit? Journal of Urban Economics 108.

40 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY Author/Year Geography Method Results Babar and Burtch, 201963 Urbanized areas in the United States Statistical analysis (differences in differences regression) For the average urbanized area, Uber entry reduces bus transit ridership by 1.3%, increases commuter rail ridership by 2.7%, and has no significant effect on light-rail or subway ridership. Young et al., 202064 Toronto region Detailed analysis of ridehail trips in 2016 Household Travel Survey About 30% of ridehail trips (primarily in urban core) had cheaper transit alternatives with durations of less than 15 additional minutes. About 27% of ridehail trips (primarily in suburbs) would require 30 or more additional minutes by transit. Li et al., 202065 City of Toronto Models based on 2016– 2018 ridehail trips, transit supply, and ridership Ridehail trips appear to complement rail transit and substitute for bus and streetcar trips in the urban core. Ridehail complements rail mostly in off- peak periods and substitutes for surface transit mostly in peak periods. Dumas et al., 202066 City of Toronto Modeling and analysis of ridehail origin-to- destination trips 2016–2019 Ridehail trip volumes doubled between 2016 and 2018. Ridehail trips about 5% to 8% of total vehicle kilometers traveled (not counting deadheading). So far, only marginal impact on downtown travel speeds. Ridehail commute trips growing outside of downtown, but the same percentage (10%) of ridehail commute trips arrive at subway stops as head directly downtown. 63 Babar, Y., and G. Burtch. 2019. Examining the Heterogeneous Impact of Ride-hailing Ser- vices on Public Transit Use. Information Systems Research 31(3). https://pubsonline.informs. org/doi/10.1287/isre.2019.0917. 64 Young, M., J. Allen, and S. Farber. 2020. Measuring when Uber Behaves as a Substitute or Supplement to Transit: An Examination of Travel-Time Differences in Toronto. Journal of Transport Geography 82:102629. 65 Li, W., et al. 2020. Exploring the Impact of Ride-Hailing on Multimodal Public Transit in Toronto. Master’s thesis available at https://tspace.library.utoronto.ca/handle/1807/98128. 66 Dumas, R., et al. 2020. A Comprehensive Investigation into Ridesourcing Company Activities in Toronto. Presented at the 99th Annual Meeting of the Transportation Research Board, Washington, DC. For report on which this paper is based, see https://www.toronto.ca/ wp-content/uploads/2019/06/96c7-Report_v1.0_2019-06-21.pdf. TABLE 2-1 Continued

SHARED MOBILITY AND PUBLIC TRANSPORTATION 41 Author/Year Geography Method Results Henao and Marshall, 201867 Denver Driver data from 416 ridehail trips and intercept survey of ridehail riders n=311 After accounting for shifts from drive alone, transit and other shared modes, and walking, “ride-hailing leads to approximately 83.5% more VMT than would have been driven had ride-hailing not existed.” Schaller, 201768 New York City Analysis of data provided by ridehailing companies Ridehailing companies added 600 million VMT to city traffic over 3 years. Ridehailing trips added 7% to the total VMT of Manhattan, western Queens, and western Brooklyn. Schaller, 202069 Illustrative large U.S. cities with extensive transit Spreadsheet model Ridehailing adds 2.5 miles VMT for every mile reduced from personal auto and taxi trips (151% more VMT) in large cities due to deadheading, and assumed share of (1) trips pooled and (2) shifts from transit, walking, biking. No discernable or consistent shift in auto ownership in eight large cities or close-in neighborhoods between 2006 and 2017 as ridehail trips grew markedly. Qian et al., 2020b70 Manhattan Ridehail trip trajectory data and models to estimate speeds and emissions From 2017 to 2019, increased ridehail trips “results in an average speed reduction of 11.3% on weekdays, down from 8.37 km/h in April 2017 to 7.42 km/h in March 2019.” This results in “83.7% more NOx, 91% more CO and 95.4% more HC emissions” per ridehail kilometer traveled. 67 Henao, A., and W. Marshall. 2019. The Impact of Ride-Hailing on Vehicle Miles Traveled. Transportation 46(6). https://link.springer.com/article/10.1007/s11116-018-9923-2. 68 Schaller. 2017. 69 Schaller, B. 2020. Does Sharing a Ride with Strangers Take Cars off the Street? Evidence from Uber and Lyft. Presented at the 99th Annual Meeting of the Transportation Research Board, Washington, DC. 70 Qian et al. 2020b. TABLE 2-1 Continued continued

42 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY Author/Year Geography Method Results Erhardt et al., 201971 San Francisco Aggregated ridehail data by time of day and route to estimate congestion impact compared to counterfactual Over the 2010–2016 period, ridehail increased VMT 6% on top of 7% VMT increase due to population and economic growth. Ridehail also increased vehicle hours of delay by 40% and reduced speeds by 7% on top of increases of 12% in delay and 4% reduction in speed due to other factors. Graehler et al., 201972 22 major U.S. cities Statistical analysis of monthly transit ridership data over 2002–2018; ridehail and shared bike operations; population; other factors After ridehail companies enter a market, heavy rail ridership decreases by 1.29% per year, and bus ridership decreases by 1.70% per year. Over an 8-year period, this implies a decrease of 9.8% for heavy rail ridership and 12.7% in bus ridership. The presence of bikeshare operations is associated with increased rail ridership and decreased bus ridership. Hou et al., 202073 Chicago Statistical analyses of ridehail trip data About 26% of ridehail users were willing to pool, of which about 72% of trips involved an additional rider. (About 19% of shared rides were actually pooled.) 71 Erhardt, G., et al. 2019. Do Transportation Network Companies Decrease or Increase Congestion? Science Advances (5). https://advances.sciencemag.org/content/5/5/eaau2670. 72 Graehler et al. 2019. 73 Hou, Y., et al. 2020. Factors Influencing Willingness to Share in Ride­Hailing Trips. Presented at 99th Annual Meeting of the Transportation Research Board, Washington, DC. TABLE 2-1 Continued

SHARED MOBILITY AND PUBLIC TRANSPORTATION 43 not generalize beyond those areas. Finally, inferences in statistical studies about whether or how ridehailing causes changes in transit ridership depend on the rigor of the methodological approach employed and inclusion of variables controlling for independent influences. Substituting for and Complementing Transit Trips Several surveys indicate that ridehail users are substituting ridehail trips for transit trips. For example, ridehail users responding to intercept studies in San Francisco and Denver indicate that 22 to 33 percent of their most recent ridehail trips would have been made by transit if ridehailing had not been available.74,75,76 (About 5 percent of ridehail trips in San Francisco, however, also began or ended at transit stations, implying that they could be first/last mile connections.) A 2016 survey of seven large metro areas reports that 6 percent of ridehail users reduced bus transit, 3 percent reduced light-rail use, and 3 percent increased subway use.77 Results from a California survey that included both urban and suburban areas indicate that about 12 to 13 percent of ridehail users would have used transit if ridehail service had not been available.78 Although suggestive, these results do not estimate the overall impact on transit ridership because the survey population’s representativeness of transit users is unknown and the magni- tude of ridehail users’ reduction in transit use was not measured. Results from these North American surveys, however, are consistent with those conducted in Brazil, Chile, and China.79 The high proportion of ridehail trips on weekend nights in six U.S. metro areas suggests that there would be little substitution of ridehail for transit, as this is a period when transit service is limited if present at all.80 (Ridehail trips for leisure or socializing is the highest-ranked reason that travelers give for using this mode in studies from cities in Brazil, Chile, India, and the United States.81) Even so, roughly one-quarter of ridehail 74 Note that a more complete summary of U.S. surveys of ridehail users finds a wider range of induced trips—0.5 to 12 percent. See Moody, J., and J. Zhao. 2020. Adoption of Exclu- sive and Pooled TNC Services in Singapore and the United States. Journal of Transportation Engineering, Part A: Systems, Table 5. https://www.researchgate.net/publication/342887348_ Adoption_of_ Exclusive_and_Pooled_TNC_Services_in_Singapore_and_the_US. 75 Henao. 2017. 76 Rayle et al. 2016. 77 Clewlow and Mishra. 2017. 78 Alemi et al. 2020. 79 Tirachini, A. 2019. Ride-Hailing, Travel Behavior, and Sustainable Mobility: An Interna- tional Review. Transportation 47(3). https://www.researchgate.net/publication/337311205_ Ride-hailing_travel_behaviour_and_sustainable_mobility_an_international_review. 80 Feigon and Murphy. 2018. 81 Tirachini. 2019.

44 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY trips in these six U.S. metro areas occur during peak periods, the signifi- cance of which depends on their relative magnitude and whether they have transit alternatives. A detailed study that matched ridehail trip origins and destinations to transit stations and lines in Toronto, for example, finds that 30 percent of ridehail trips (mostly in the urban core) had much less expen- sive transit alternatives available with modest time penalties.82 Detailed modeling and statistical analyses come to somewhat differing conclusions about complementarity versus substitutability of ridehailing for transit trips. A modeling analysis for Toronto finds that ridehail trips substi- tute for bus trips while complementing rail trips in the urban core.83 Analysis of ridership trends in multiple metropolitan areas following Uber’s entry compared with those without Uber service provides mixed results: increases in large cities and for small transit agencies but indicators of declines for large transit agencies up to 2015,84 whereas a separate study using a similar method found decreases in bus ridership, increases in commuter rail rider- ship, and no significant effect on light-rail or subway ridership.85 As a further contrast, a statistical analysis of transit ridership trends over 8 years from 22 large U.S. urban areas estimates a statistically significant 10 percent decrease in heavy rail ridership and a 13 percent decrease in bus ridership following the introduction of ridehailing service.86 A shortcoming of studies aggregating and comparing cities is their inability to account for differences in the built environment, extent of transit services, and presence or absence of public policies to influence complementary service between ridehailing and transit.87 These studies diverge on the complementarity versus substitutability of ridehailing for transit, but they are certainly suggestive. Ridehailing users indicate they use transit less, particularly buses, and modeling and statistical studies indicate that ridership on buses is more negatively influenced than on urban rail. Indeed, there is evidence of complementarity with rail transit in some studies, which points to the potential of first/last mile ridehailing trips to rail transit. Public policies to enhance complementary service be- tween ridehailing and transit are important in this regard, as discussed in Chapter 5. Impact of Increased Congestion on Road Network and Bus Speeds Ridehailing could also adversely affect surface transit modes (bus, light rail, vanpools) by increasing congestion and reducing traffic speeds. There are 82 Young et al. 2020. 83 Li et al. 2020. 84 Hall et al. 2018. 85 Babar and Burtch. 2019. 86 Graehler et al. 2019. 87 Tirachini. 2019.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 45 multiple mechanisms by which this can happen. As with taxis, ridehailing trips involve some VMT without passengers as vehicles cruise between drop-offs and pick-ups (though ridehail vehicles may require less cruising between passengers than taxis88). Ridehailing drivers also have commute trips of some distance to the areas where they cruise. A detailed analysis of ridehailing trips in Denver estimates, conservatively, that deadheading in- creases VMT of individual ridehail trips by at least 40 percent,89 which can be important in cities such as San Francisco, where ridehail trips account for at least 15 percent of daily intra-city trips.90 In addition, ridehailing trips apparently divert some share of trips from transit, carpooling, and walking, which adds more automobile trips to the network.91,92,93,94 Finally, the mere availability of ridehailing induces trips that would not otherwise have happened, thereby also contributing to increased VMT; various sur- veys report that 8 to 12 percent of respondents would not have made their most recent trip had ridehailing not been available.95,96,97 These induced trips obviously add utility to the people who make them but can also im- pose social costs when they increase congestion and emissions. To the extent that auto users are willing to pool or split rides, ridehail- ing does not have to increase VMT as much as do solo passenger ridehailing trips. Uber and Lyft offer less expensive fares for those willing to pool with other riders, and this approach is Via’s and other microtransit providers’ market niche. Careful analysis of Chicago ridehailing data suggests that about 19 percent of ridehailing trips are pooled,98 which is consistent with survey results in Boston (20 percent pooling).99 After accounting for dead- heading, mode shift, and pooled rides, ridehailing adds 2.5 miles of VMT for every mile of personal driving and taxi trips reduced in large urban areas.100 Increased VMT due to ridehailing necessarily increases vehicle emis- sions because of increased VMT, but it does not necessarily affect transit 88 For discussion of taxi and ridehail efficiency, see Tirachini. 2019, pp. 18–19. 89 Henao. 2017. 90 San Francisco County Transportation Authority. 2017. TNCs Today: A Profile of San Francisco Transportation Network Company Activity. https://www.sfcta.org/sites/default/ files/2019-02/TNCs_Today_112917_0.pdf. 91 Alemi et al. 2020. 92 Clewlow and Mishra. 2017. 93 Henao. 2017. 94 Rayle et al. 2016. 95 Alemi et al. 2020. 96 Henao. 2017. 97 Rayle et al. 2016. 98 Hou et al. 2020. 99 Gehrke et al. 2019. 100 Schaller. 2020.

46 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY ridership. As noted, most ridehailing trips occur in the off peak;101 thus, the ridehailing impact on transit depends on the time of day it occurs and whether the ridehailing trip has a reasonable transit alternative. Two studies that account for these effects by modeling the impact of ridehail- ing on traffic speeds (in San Francisco and Manhattan) find significant reductions in speeds and significant increases in delay due to sharp growth in ridehailing trips.102,103 A similar modeling exercise in Toronto, however, estimated only marginal impacts of ridehailing trip growth on peak-period downtown traffic speeds.104 These divergent results suggest that the impact of ridehailing on urban traffic speed and delay depends on the share and magnitude of ridehail trips in the peak period and other conditions. For example, because the impact of additional vehicles in traffic becomes highly non-linear as a route reaches its peak volume, pre-existing traffic condi- tions will be important determinants of ridehailing impact. Manhattan traffic, for example, was highly congested before daily ridehailing trips increased to 630,000 in 2018 (an estimated 60 percent increase in a little over 1 year105).106 The larger impacts of ridehailing on congestion in San Francisco than in Toronto may be attributable to the much larger share of ridehail VMT in San Francisco (15 to 20 percent for San Francisco107) com- pared to a much smaller share of ridehail VMT estimated for Toronto (5 to 8 percent108). (Ridehail passengers, of course, can benefit in overall saved time in congested conditions because of the time saving from a door-to-door trip and avoidance of walking to and from transit stops or parking, but still impose social costs by adding vehicle trips to the network in peak periods.) The above estimates are short-run impacts. Over the longer run, ride- hailing trips could continue to increase and further degrade traffic speeds and increase emissions. However, to the extent that the availability of ride- hailing reduces the need for auto ownership (and parking spaces) over the long run by urban dwellers with good access to transit, ridehailing could result in less overall VMT by these households even if they use transit only occasionally.109 Trends over 11 years in eight major U.S. cities with substan- 101 Feigon and Murphy. 2018. 102 Erhardt et al. 2019. 103 Qian et al. 2020b. 104 Dumas et al. 2020. 105 Qian et al. 2020b. 106 Schaller. 2017. 107 Alemi et al. 2020. 108 Dumas et al. 2020. 109 Conway, M., et al. 2018. Trends in Taxi Use and the Advent of Ridehailing 1995– 2017: Evidence from the U.S. National Household Travel Survey. Urban Science 2(3):79. https://www. researchgate.net/publication/327277013_Trends_in_Taxi_Use_and_the_Advent_ of_ Ridehailing_1995-2017_Evidence_from_the_US_National_Household_Travel_Survey.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 47 tial growth in ridehailing, however, provide no compelling evidence to date of declining auto ownership in close-in urban neighborhoods.110 Micromobility In general, both personal and shared bikes, scooters, Segways, and other small vehicles are infrequently used to connect to transit. In large metro- politan statistical areas, only about 0.4 percent of bike, scooter, etc., trips of 1 to 6 km are for the purpose of access or egress to transit.111 Of particular interest is how rapidly expanding bikesharing and scootersharing opera- tions can affect this small percentage of micromobility trips complementing transit trips. Results to date suggest that micromobility both complements and substitutes for transit trips. Docked bikesharing appears to substitute for transit trips in the dense cores of metropolitan areas and to complement transit in less popu- lous neighborhoods.112 For example, detailed analysis of the impact of Washington, DC’s Capital Bikeshare (a docked bikesharing operation) found that from 2010 to 2015, riders using bikesharing stations in the city’s central core substituted bikeshare for rail transit trips while riders using bikesharing stations on the transit system’s periphery complemented rail transit service.113 (The rapid growth in bikeshare use occurred during the same period that Washington Metropolitan Area Transit Authority [WMATA] rail transit ridership was declining. Half or more of Capital Bikeshare users in a 2014 survey indicated that they used Metro rail less often or much less often.) Introduction of docked bikesharing in Manhattan and Brooklyn resulted in about a 3.3 percent decline in bus ridership along routes with multiple docking stations.114 Results from convenience surveys of bikeshare users in Minneapolis-St. Paul, Salt Lake City, Montreal, and Toronto show mixed results: little impact on bus use in Minneapolis-St. Paul and a net increase in Salt Lake City.115 Meanwhile, rail use increased in both cities but declined in Montreal and Toronto. 110 Schaller. 2020. 111 Krizek, K., and N. McGucken. 2019. Shedding NHTS Light on the Use of “Little Vehicles” in Urban Areas. Transport Findings, November. https://doi.org/10.32866/10777. 112 Shaheen, S., and E. Martin. 2015. Unraveling the Modal Impacts of Bikesharing. Access 47. http://www.accessmagazine.org/fall-2015/unraveling-the-modal-impacts-of-bikesharing. 113 Ma, T., and G. Gnapp. 2019. Estimating the Impact of Capital Bikeshare on Metrorail Ridership in the Washington Metropolitan Area. Transportation Research Record 2673. https://journals.sagepub.com/doi/full/10.1177/0361198119849407. 114 Campbell, K., and C. Brakewood. 2017. Sharing Riders: How Bikesharing Impacts Bus Ridership in New York City. Transportation Research Part A 100. https://www.sciencedirect. com/science/article/abs/pii/S0965856416304967?via%3Dihub. 115 Shaheen et al. 2016, p. 27.

48 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY Because e-scooter growth is so recent, its impact on transit use is less understood. The 2019 survey of micromobility users in 18 metropolitan areas cited previously finds that about 18 percent of e-scooter users use them to connect to transit most or all of the time.116 Roughly equal propor- tions of users in medium- and high-density metropolitan areas report less and more use of transit, whereas in low population density areas, a higher percentage of users (28.5 percent) report less use of transit than more use (22.1 percent). (Worth also noting is that, across all three density group- ings, 29 to 32 percent of e-scooter users also report less use of bikesharing.) Less is yet known about the use of shared e-bikes than about the use of e-scooters in the United States, though use of undocked powered bikes reached 10 million trips in 2019, and docked bikeshare systems that include both conventional and e-bikes saw increased growth as well.117 A meta-analysis of 24 e-bike studies from around the world, mostly based on studies of private e-bike use, finds the primary mode substitution of e-bike use is from transit (33 percent) followed by automobile (27 percent), with substitution from transit more characteristic of Chinese studies and substi- tution for autos more characteristic of studies from Europe, Australia, and the United States.118 Carsharing and Microtransit The degree to which carsharing complements or substitutes for transit trips has received little empirical study. A review of the international literature reports mixed results from case studies of Ulm (Germany), Toronto, and London, the results of which depend on the local context and type of carsharing.119 Analysis of conventional round-trip carshare subscribers in North America finds a small increase in the number of respondents who report less use of transit, primarily due to formerly carless households adopting carsharing.120 As mentioned in Chapter 1, microtransit emerged in the United States as private company efforts to experiment with smart- phone apps for connecting with riders and advanced routing algorithms to 116 SUMC et al. 2019. 117 NACTO. 2020. 118 Bigazzi, A., and K. Wong. 2020. Electric Bicycle Mode Substitution for Driving, Public Transit, Conventional Cycling, and Walking. Transportation Research Part D. https://doi. org/10.1016/j.trd.2020.102412. 119 Circella et al. 2018, pp. 7–8. 120 Martin, E., and S. Shaheen. 2011. The Impact of Carsharing on Public Transit and Non-Motorized Travel: An Exploration of North American Carsharing Survey Data. Energies 4(12). https://www.researchgate.net/publication/266891168_The_Impact_of_Carsharing_on_ Public_Transit_and_Non-Motorized_Travel_An_Exploration_of_North_American_Carsharing_ Survey_Data.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 49 minimize traveler trip times. These services might have competed with con- ventional public transit had these efforts succeeded. As described in the next section, many transit agencies are actively experimenting with microtransit technologies and contractors to attract more patrons to transit, typically in low-density areas beyond the reach of fixed-route transit, the empirical results of which have not been reported as of the time of this writing. SHARED MOBILITY–TRANSIT COORDINATION: OPPORTUNITIES AND BARRIERS This section describes pilot transit agency efforts to coordinate with ride- hailing, micromobility, and microtransit in ways that can advance acces- sibility, equity, sustainability, and other social goals. An important set of issues for transit agency coordination with shared mobility providers involves contracting, data reporting, and other requirements established in federal and state law. Transit agencies, and private providers with which they contract, must meet federal and state legal equity obligations that come with providing transportation services to the general public. Among the most relevant federal laws in this regard are (1) meeting the requirements of the ADA in the provision of wheelchair accessible vehicles (WAVs) and equivalent service, including cost and response times,121 and (2) meeting Title VI of the Civil Rights Act, which includes providing ways for low- income minorities without bank accounts or smartphones to pay for such service.122 These obligations are challenges for shared mobility providers such as the major ridehailing companies, which do not accept ADA re- sponsibilities and rely on smartphone apps to arrange and pay for service. Regarding payment options, however, ridehailing companies such as Uber, Lyft, and Via have expanded options recently for payment that do not require access to bank accounts, such as pre-paid debit cards.123 Related to these is the reluctance of ridehailing companies to share data about trips with public agencies, which limits the ability of agencies both (1) to evaluate the effectiveness of their collaborations and (2) to meet federal transit agency reporting requirements to the National Transportation Data- base (NTD).124 Examples in the following section illustrate the principal 121 See Waite, J. 2018. Legal Considerations in Relationships Between Transit Agencies and Ridesourcing Service Providers. TCRP Legal Research Digest 53, pp. 23–24 and 112 regarding equivalent service for WAVs, including response times. 122 Title VI implications for transit agency partnerships with ridehailing companies are com- plex and involve multiple dimensions. See Waite. 2018, pp. 27–29, for a comprehensive review. 123 ITF (International Transport Forum). 2019. Regulating App­based Mobility Services: Summary and Conclusions. ITF Roundtable Report 175. OECD Publishing, Paris. https:// www.itf-oecd.org/sites/default/files/docs/app-based-mobility.pdf. 124 Waite. 2018, p. 115.

50 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY opportunities and barriers to government and transit agency coordination with ridehailing providers. Pilot Partnerships with Ridehailing Companies in the United States General Experience A 2018 survey of 44 public transit agencies125 in 22 states known to be engaged in pilot tests of formal and informal partnerships with ridehail companies found that three-quarters of agencies were motivated by the goal of improving first/last mile connections to transit in urban and suburban contexts, roughly one-third were also motivated by goals to improve or expand paratransit service and reduce its cost, and about 20 percent wanted to provide service to new areas lacking service and to provide service to low-income and transportation-disadvantaged travelers (respondents could select all goals that applied in this survey question).126 Of the 37 agen- cies that responded to the survey in early to mid-2018, 14 agencies were engaged in active pilots; 5 were offering regular, continued service beyond the pilot stage; 14 were exploring or procuring partnerships; and 4 had concluded pilot tests.127 This survey focused on the nature of transit agency efforts to coordi- nate with ridehailing companies and how issues such as ADA and Title VI compliance, data sharing, labor union concerns, and agency liability were being addressed. It did not systematically gather metrics about trips, costs, or other output measures. Key insights from the 2018 survey are described in the following paragraphs. Although this survey did not include results of evaluations, the Federal Transit Administration (FTA) funded several demonstration projects in 2016 through the “Mobility on Demand (MOD) 125 Curtis, T., et al. 2019. Partnerships Between Transit Agencies and Transportation Net­ work Companies (TNCs). Transit Cooperative Research Program 204. Note that this survey was based on those agencies the research team knew or believed were engaged in or explor- ing partnerships with TNCs, which means that they may not have been aware of all such efforts. Moreover, of 44 agencies surveyed, 37 responded, although few of them answered all the questions in the survey, limiting the generalizability of the survey results. The results cited in this section reported as percentages are limited to those questions that had responses from 33 or more respondents. Note that this survey did not capture all examples of projects where transit agencies are or have been partnering with TNCs. See, for example, other formal and informal TNC and transit agency partnerships in Schwieterman, P., et al. 2018. Partners in Transit: A Review of Partnerships Between Transportation Network Companies and Public Agencies in the United States. Chaddick Institute, DePaul University. https://las. depaul.edu/centers-and- institutes/chaddick-institute-for-metropolitan-development/research-and- publications/ Documents/Partners%20in%20Transit_Live1.pdf. 126 Curtis et al. 2019, Figure B-4. 127 Curtis et al. 2019, p. 13.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 51 Sandbox” that focus on, or include elements of, transit agency and ride- hailing company coordination. Most of these demonstrations include inde- pendent evaluations of project outcomes. Five MOD Sandbox reports had been released as of the time of the writing of this report and are cited in this chapter and the next where relevant. Transit agencies addressed the unwillingness of major ridehailing com- panies to provide ADA services and meet Title VI transit agency obligations by bringing in additional partners, typically providers of WAVs or call centers to facilitate rides for users without smartphones.128,129 Other major challenges identified by respondents regarded establishing limits to transit agency liability, ensuring passenger safety, and addressing concerns about impacts on unionized labor, among others.130 For the most part, transit agencies were able to satisfy these latter concerns, although a few proposed pilots did not go forward because of them. Although not identified in this survey, there is also an issue about partnering with private companies that have not yet proven profitable or financially sustainable and which may raise fares or simply end programs at any time.131 Few of the early pilot partnerships succeeded in receiving more data from Uber or Lyft than summary statistics on ridership and costs. Transit agencies wanted to receive such information to understand whether the pilot projects were meeting agency goals, for insight into users of the ser- vices and the purposes of the use, and to report subsidized ridehail trips to FTA’s NTD.132 Other benefits of receiving disaggregated information would be to understand impacts of ridehail trips on traffic congestion and on fixed-route ridership. Ridehailing companies resisted providing such informa tion for fear of divulging traveler privacy and to protect proprietary data. Note that third-party data aggregators, both private and non-profit, have emerged in recent years to serve as intermediaries between public agencies and private providers such as ridehailing and micromobility com- panies.133 These organizations can both receive and hold data from private companies, such that they are not subject to disclosure under public records laws, and can aggregate the data in ways that protect privacy while still 128 Curtis et al. 2019, p. 15. 129 Other shared mode providers also resist being obligated to meet ADA requirements. 130 Curtis et al. 2019, Figure B-6. 131 See, for example, Simons, S. 2019. They Relied on Lyft Rides for Groceries. Now These Seniors Must Find Another Way. WAMU 88.5, October 10. https://www.npr.org/ local/305/2019/10/10/769071292/they-relied-on-lyft-rides-for-groceries-now-these-seniors-must- find-another-way. 132 Curtis et al. 2019, p. 73. 133 Lempert, R. 2019. Shared Mobility Data Sharing: Opportunities for Public­Pri­ vate Partnerships. Translink New Mobility Lab. https://sustain.ubc.ca/sites/default/files/ Sustainability%20Scholars/2018_Sustainability_Scholars/Reports/2018-70%20Shared%20 Mobility%20Data%20Sharing%20.

52 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY providing information that public agencies want to have to under stand the impacts of these shared modes. As would be expected in pilot programs, most of these partnerships were of limited duration, covering limited geographic scope, and involving a relatively small number of riders. Case examples with a broader scope and impact illustrate some initial results and potential for transit agency partnerships with ridehailing companies, as described next. First/Last Mile Partnerships The Pinellas County, Florida, transit agency, Pinellas Suncoast Transit Authority (PSTA), was one of the first transit agencies to develop partner- ships with ridehailing companies, initially with Uber and, later, with Lyft as well.134 PSTA has had three different partnership initiatives, which have provided different kinds of service and raised different kinds of issues (see Box 2-1 for details).135 The first, described in this section, was designed to replace regular fixed-route transit service in lightly populated areas with little patronage with alternatives that were less expensive to provide. The other two PSTA initiatives are described later. The first effort, “Direct Connect,” provided first/last mile connections within certain zones to select transit stops. It illustrates how partnering with TNCs to serve the general public must also involve partnering with other providers to serve federal equity goals and legal obligations. The initiative involved multiple partner- ships, including with Uber, a taxi company (for those without smartphones or credit cards), a provider of WAVs, and a paratransit operator for those requiring door-to-door rather than curb-to-curb services. For the general public and ambulatory paratransit-eligible patrons (those not requiring assistance in accessing or entering or exiting vehicles), PSTA subsidized first/last mile connections by these providers to selected transit stops and stations. The pilot also evolved and grew over time to include (1) expanded funding for passenger subsidies (including equalizing the subsidy for wheel- chair users with that of the general public) and (2) expanded service areas from relatively small zones to the entire county. Over a 9-month period, daily ridership grew from a mere handful to more than 1,200 daily trips before leveling off, thereby providing a considerable expansion in mobility for users of the service.136 134 Unless otherwise noted, the details of the PSTA partnerships with TNCs are drawn from the case study of PSTA in Curtis et al. 2019, pp. 51–55. 135 Curtis et al. 2019, p. 52. 136 Murphy, C., et al. 2019. When Uber Replaces the Bus: Learning from the Pinellas Suncoast Transit Authority’s “Direct Connect” Pilot. Shared-Use Mobility Center. https:// learn.sharedusemobilitycenter.org/overview/direct-connect-what-the-first-transit-tnc-partnership- can-teach-us-pinellas-county-fl-2019.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 53 This PSTA initiative lacks a formal evaluation, but a case study by a third party points out that although Direct Connect reduced PSTA costs compared with the lightly used fixed-route service it had been providing, lack of trip-level data from Uber limited PSTA’s understanding of whether certain neighborhoods could benefit from additional marketing, and lack of user surveys conducted by PSTA limited the agency’s ability to assess progress in meeting other goals of the project, such as increasing transit ridership by partnering with Uber.137 Whereas Direct Connect provided a new service to a considerable number of patrons, it was not possible to determine whether patrons used Uber for a first/last mile connection to the bus line or, instead, to make trips to activity centers that were near transit stops for shopping, recreation, and dining. Other pilots relied on technology to facilitate first/last mile connections. In the Denver, Colorado, region, Uber developed a trip planner within its app that includes trips on the Denver Regional Transportation District 137 Murphy et al. 2019. BOX 2-1 Pinellas Suncoast Transit Authority (PSTA) Ridehail Coordination Initiatives This quoted text describes the three PSTA initiatives: “In its original zone-based ‘Direct Connect’ partnership with Uber and United Taxi, PSTA paid up to $5 [initially $3] toward first-/last-mile Uber and United Taxi rides to and from selected bus stops or transit stations in zones within Pinellas Park and the East Lake area. The program began operating in February 2016 and expanded countywide in January 2017. In April 2018, PSTA removed the zones and added additional locations. PSTA subsequently began a second partnership in August 2016, called ‘TD Late Shift,’ to provide unemployed or low-income residents up to 25 discounted Uber, taxi, or wheelchair transport rides to and from work per month when PSTA fixed-route service is unavailable. In its Public-Private-Partnership for Paratransit Mobility on Demand ( P4-MOD) project, PSTA aims to improve mobility of paratransit customers and operate paratransit trips more cost effectively than its current Demand-Response Transportation (DART) program. According to the transit agency’s MOD grant application, PSTA currently spends $22.50 per ride on its current DART service, which utilizes nearly 10% of the transit agency’s operating budget. PSTA seeks to achieve these goals through centralized dispatch technology that matches riders with TNCs, taxis, or wheelchair vans depending on the rider’s needs, estimated arrival time, and cost.”

54 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY (RTD) and subsequently made it possible for users of its app to purchase RTD tickets.138 The Denver RTD’s own app also allows for booking Uber or Lyft trips. As another example, Dallas Area Rapid Transit (DART) enhanced its GoPass app to allow users to choose a DART microtransit vehicle or UberPool in north Plano (an area within a small city about 20 miles outside of Dallas) for first/last mile within-zone trips.139 The DART project is an FTA MOD Sandbox project that uses UberPool as a supplement to its micro- transit pilot, as discussed in the microtransit section that follows. Late Night Services The second PSTA partnership initiative, “Transportation Disadvantaged (TD) Late Shift,” was developed to provide discounted home-to-work trips for qualified, low-income, late-night service workers during hours when PSTA does not operate. Uber and the other paratransit and WAV providers in PSTA’s Direct Connect pilot also participate in this service. As of mid-2018, this service was highly popular with roughly 4,000 Pinellas County users. The Direct Connect and TD Late Shift efforts both experienced limited data reporting by Uber initially, but by late 2019 PSTA indicated it was receiving the kind of data it needed to assess pilot impacts. Both initiatives included paratransit and WAV providers to meet ADA and Title VI obligations. Complementing ADA Paratransit Services with a Ridehailing Option The third PSTA initiative, P4-MOD, was developed as an FTA MOD Sandbox project, the evaluation of which is not yet available. The aim was to improve paratransit service to qualified riders that required making res- ervations 24 hours in advance by making available on-demand trips by Lyft and the agency’s paratransit service. The PSTA chief executive officer Brad Miller (also a committee member) reported in August 2019 that this service had been very popular. These trips, however, had not replaced traditional paratransit trips but had, instead, expanded the mobility of eligible riders. The successful expansion of mobility from this pilot has been coupled with the unanticipated cost of subsidizing the new demand. Similarly, the Massachusetts Bay Transit Authority (MBTA), the tran- sit provider in the greater Boston area, experimented with strategies to 138 Bliss, L. 2019a. The Uber Transit Convergence Arrives in Denver. CityLab, May 2. https://www.citylab.com/transportation/2019/05/uber-denver-transit-ticket-bus-train-light- rail- fare-app-rtd/588559. 139 Parks, R., and S. Moazzeni. 2020. Mobility on Demand Sandbox Demonstration: DART First and Last Mile Solution. Federal Transit Administration Report 0164. https://www. transit.dot.gov/research-innovation/mobility-demand-mod-sandbox-demonstration-dart-first- and-last-mile-solution-0.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 55 replace paratransit services that require 24-hour advance scheduling with on- demand services provided by Uber and Lyft.140 The success of this pilot led to an ongoing program. The initial goal was to provide paratransit cus tomers with an option that would cost the MBTA less per trip than tradi tional paratransit service. In this partnership, Uber and Lyft contract with other parties to ensure that WAVs are available and to ensure that those without smartphones and credit cards can arrange and pay for trips. The MBTA subsidizes a set number of trips per month. Initially limited to 400 eli gible paratransit users during the pilot phase, the program was sub- sequently opened up to all eligible paratransit riders, but caps were placed on the number of trips available to riders due to limited funding. The program was received positively and by early 2018 monthly pilot program ridership reached 13,000 compared with 150,000 provided through the conventional paratransit program. This service was delivered at a lower cost per trip, thereby enhancing mobility without increasing agency cost.141 Uber and Lyft provide the MBTA with summary statistics on ridership, wheelchair use, and the origin and destination of trips (but only at the zip code level). WMATA has also pilot tested on-demand paratransit service at lower trip costs than traditional paratransit providers.142 WMATA contracted with two technologically sophisticated taxicab providers, one of which had experience with wheelchair accessible service. Although a ridehailing company did not participate, the taxicab providers lowered the per-trip cost compared with the previous ADA paratransit provider, received positive feedback from users, and supplied WMATA with requested data. Provision of Ridehailing Service Instead of Fixed­Route Transit Cities and towns can contract with ridehailing companies to serve their communities rather than providing traditional, fixed-route transit services. Arlington, Texas, is one such city. Located between Dallas and Fort Worth, Arlington voters rejected dedication of local sales tax funds to existing tran- sit agencies operating within the region for the third time in 2017, which caused the city to end operation of its single bus route.143 Instead, it began 140 Curtis et al. 2019, p. 42. For a review of this program and a similar one in New York City, see also Parker, M. 2020. Benefits and Challenges of On­Demand E­Hailing Paratransit Programs: A Case Study of Pilot Programs in New York City and Boston. Presented at the 99th Annual Meeting of the Transportation Research Board, Washington, DC. 141 Curtis et al. 2019, p. 43. 142 Curtis et al. 2019, pp. 66–67. 143 Weinreich, D., and T. Skuzinski. 2020. Transit in Flex: Examining Service Fragmen­ tation in New App­Based, On­Demand Services. The University of Texas at Arlington. https://ctedd.uta.edu/research-projects/transit-in-flex-examining-service-fragmentation-of- new-app%E2%80%90based-on%E2%80%90demand-transit-services.

56 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY offering subsidized rides on Via’s small vans on a pilot basis within a single zone of the city. It subsequently expanded this service in 2020 to cover more than 40 percent of the city.144 Largely funded with a grant from FTA, the city is meeting applicable federal ADA and Title VI requirements. One other known example in North America is the small, exurban town of Innisfil, Ontario. Lacking transit service, the town authorities opted to subsidize Uber trips rather than contract with a transit bus provider.145 For more than a decade, a collaboration of five transit agencies in Denmark has provided a demand-responsive service (FlexDanmark) to coordinate and pool human service, disability, and medical trips, and to replace intermittent fixed-route transit in rural areas for the general public, with contracted trips by personal vehicles, taxis, and mini-buses.146 An independent evaluation of this service is not available, but it reportedly provides 15,000 daily trips and has apparently reduced costs to the transit agencies that operate these services. Micromobility An ongoing Transit Cooperative Research Program (TCRP) study is review- ing pilot efforts by transit agencies to coordinate with cities and micro- mobility providers to facilitate first/last mile connections to transit.147 According to the interim report of this TCRP project and other reports, first/last mile connection efforts have focused on providing docking stations and parking near transit stops and stations,148 regulating sidewalk use and parking, providing bike lanes, and mandating data-sharing requirements that facilitate planning and enforcement. For the most part, these ef- forts are led by local governments rather than transit agencies, since local 144 Mauch, R. 2020. Arlington Expands Its Via Rideshare Service Again. Here Are the Latest Destinations. Fort Worth Star­Telegram, February 10. https://www.star-telegram.com/news/ local/arlington/article240023343.html. Note that although Arlington dropped its one bus line in 2017, it continued to offer a paratransit service for senior citizens. 145 The Innisfil Experiment: The Town That Replaced Public Transit with Uber. 2019. The Guardian, July 16. https://www.theguardian.com/cities/2019/jul/16/the-innisfil-experiment- the-town-that-replaced-public-transit-with-uber. Bliss, L. 2019b. Uber Was Supposed to Be Our Public Transit. CityLab, April 29. https://www.citylab.com/transportation/2019/04/ innisfil-transit-ride-hailing-bus-public-transportation-uber/588154. 146 Comfort, P. 2019. Denmark Transits Form Demand-Response Cooperative to Cut Costs. Metro Magazine. https://www.metro-magazine.com/10002903/denmark-transits-form- demand-response-cooperative-to-cut-costs. 147 SUMC et al. 2019. 148 Hernandez, M., et al. 2018. Public Transit and Bikesharing: A Synthesis of Transit Prac­ tice. TCRP Synthesis 132. Transportation Research Board, Washington, DC. See also Shaheen, S., and N. Chan. 2016. Mobility and the Sharing Economy: Potential to Overcome First and Last Mile Transit Connections. https://escholarship.org/uc/item/8042k3d7.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 57 governments establish requirements that micromobility companies must meet in order to operate. Notably, micromobility companies generally share trip and other data with cities, though a suit Uber filed against Los Angeles raises as yet unresolved questions about what kinds of data can be shared without violating users’ privacy.149 As the TCRP interim report indicates, some transit agencies are actively involved in coordinating with micromobility operators. For example, in Minneapolis, the regional transit agency, city, county, and neighborhood groups are pilot testing providing mobility hubs (micromobility parking, e-scooter and bike recharging, and other features) at metro stations. Oklahoma City gives dockless bikeshare operators incentives (e.g., in- creases in their fleet size) if they work with the local transit agency to locate bike parking areas near transit stops. LA Metro facilitates dockless vehicle parking near transit stops, but establishes specific requirements to ensure vehicles are upright and parked within designated zones and any violations are corrected within 2 hours. Violations of ADA requirements are strictly prohibited. The Dayton transit agency, which already operated the city’s bikeshare system, has also partnered with an e-scooter company to provide rebalancing and recharging. As of 2017, joint marketing between docked bikeshare and transit agen- cies and fare integration between docked service providers and transit was limited.150 There are newer exceptions. DART GoPass app, described ear- lier, is also including a link to local bikesharing operators and other shared mobility providers. Similar examples are occurring in Austin, Kansas City, and Los Angeles. Regardless of the role of dockless systems for first/last mile transit connections, they can serve other social goals, as dockless bikes are less expensive for users than other options and have wider geographic scope, thereby potentially enhancing access for low-income neighborhood residents.151 However, many private dockless bikesharing operators have dropped out of the market or shifted to e-scooters.152 149 Rana, P., and J. Rundle. 2020. Uber Sues Los Angeles Over Data-Sharing Rules: The Battle Could Set the Stage for How Cities Police Mobility Providers While Safeguarding Privacy. The Wall Street Journal, updated March 25. https://www.wsj.com/articles/uber-sues- los-angeles-over-data-sharing-rules-11585104223. 150 Hernandez et al. 2018. 151 See, for example, https://medium.com/populus-ai/measuring-equity-dockless-27c40af259f8 cited in ITF. 2019, p. 41. 152 Kiefer, P. 2019. Why Jump—Uber’s E-Bike Scheme—Failed in Two Cities: Is This the Death-Knell for Dockless E-Bike Schemes? Outside, September 17. https://www.outsideonline. com/2402196/uber-jump-ebike-failure.

58 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY Microtransit Despite the lack of commercial success by private providers, the technolo- gies developed for microtransit were being planned or used by at least 22 public transit agencies as of 2018 using contractors that are experienced with paratransit service or, such as Via, that are also microtransit pro- viders.153 Some transit agencies operate the service with their own labor and vehicles using microtransit software acquired from private vendors. Some transit agencies have offered demand-responsive transit (DRT) in the past and are beginning to refer to this service as microtransit; new technologies offer opportunities to make the services more responsive to customers’ needs in real time. These services are being targeted to the gen- eral public living in low population density areas. These areas, sometimes called “Flex” zones, are difficult and costly to serve with traditional fixed- route transit. This updated version of DRT provides either (1) a first/last mile connection with shuttle buses or vans within a zone or (2) point-to- point service between pick-up points and riders’ destinations. The services often involve some degree of route and service-point deviation to pick up or discharge passengers along the way as dictated by real-time demand. DRT customers can arrange for pick-up from hours to 15 minutes in advance by smartphone, computer, or telephone. The TCRP Synthesis 141 survey indicated that human resource issues and labor union resistance to contracting out service to the general public can be barriers for some agencies.154 Agencies with considerable flexibility and experience in contracting for services were able to move forward. Some agencies whose unions resisted contracting out provided their own labor or ended their pilots. A separate survey of shared mode pilot projects around the country found that, between 2015 and 2018, transit agencies tended to shift away from partnering with ridehailing companies to partnering with microtransit providers for first/last mile or low-density services in order to gain access to more data than ridehailing companies were willing to pro- vide; to substitute vans for automobiles to increase occupancy and reduce emissions per trip; and to ensure that the services did not compete with fixed-route transit.155 153 This paragraph draws heavily on Volinksi, J. 2019. Microtransit or General Purpose Public Demand Response Transit Services: State of the Practice (2019). TCRP Synthesis 141. Transportation Research Board, Washington, DC. http://www.trb.org/Main/Blurbs/178931. aspx. 154 Volinski. 2019, p. 40. 155 Lucken, E., et al. 2019. “Three P’s in a MOD”: Role of Mobility on Demand (MOD) Public-Private Partnerships in Public Transit Provision. Research in Transporta­ tion Business and Management 32. https://www.sciencedirect.com/science/article/abs/pii/ S2210539519300987?via%3Dihub.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 59 The DART MOD Sandbox demonstration project mentioned above combined several features: an expanded DART app (GoPass) that allows users to book and pay for contracted paratransit service, micromobility, and new microtransit service within three zones in the northern part of Plano. Plano has a geographic area of roughly 72 square miles with a population density of about 4,000 people per square mile.156 The new microtransit ser- vice (GoLink) extends beyond the bus line service area within Plano to reach potential patrons without transit service. After the microtransit pilot service was launched, UberPool was brought in as a partner to supplement the microtransit service to avoid overly long response times on GoLink during periods of peak demand. In one microtransit service zone, GoLink replaced the single bus line in this zone. The independent evaluation of this project is not available at the time of this writing, but the agency’s report indicates that the GoLink pilot expanded ridership and reduced per-trip subsidy in this zone. DART was also able to negotiate a data-sharing agreement with its microtransit provider that gave the level of detail the agency needed to evaluate the effectiveness of the pilot. Evidence from recent microtransit pilot projects in Arlington, Texas; West Sacramento, California; and Seattle, Washington, suggests that con- tracted microtransit service to the general public is popular with riders, reduces CO2 emissions, and costs less per passenger trip than traditional paratransit service but more than fixed-route service.157 These passenger satisfaction and per-trip cost results are consistent with the experience of transit agencies that responded to the TCRP Synthesis survey, some of whom have long experience with DRT service. SHARED MOBILITY–TRANSIT SERVICES ACROSS REGIONS This section outlines different roles shared modes could play in urban, sub- urban, and rural areas alone or in combination with public transportation. Drawing from the pilot examples described previously, there appear to be four related, but distinguishable, models for how shared mobility could serve travel across different regional spatial scales in relation to public transit.158 1. First/Last Mile Service that expands the reach of transit ser- vice beyond the typical quarter- to half-mile walking distance to 156 The information provided in this paragraph is from Parks and Moazzeni. 2020. 157 Hazan, J., et al. 2019. On Demand Transit Can Unlock Urban Mobility. BCG Henderson Institute. https://www.bcg.com/publications/2019/on-demand-transit-can-unlock- urban -mobility.aspx. 158 The committee relies here on the useful typology described in Lucken et al. 2019.

60 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY fixed-route transit (e.g., through the connectivity provided by ride- hailing, micromobility, and microtransit both (1) along existing feeder and heavily-traveled transit routes, and (2) to extend beyond fixed-route transit service into areas too sparsely populated to sup- port fixed-route transit). 2. Paratransit through ridehailing and microtransit that turns DRT from a service requiring reservations hours or days in advance to one than can be made as little as 15 minutes in advance. 3. Off-Peak Services that substitute for fixed-route service for low- income riders when transit is not operating, such as providing subsidized late-night ridehailing commute trips for low-income, late-shift workers or late-night guaranteed rides home. 4. Low-Density Services that use microtransit and ridehail pro viders for point-to-point services (without connecting to fixed-route tran- sit) that can entirely substitute for fixed-route transit services in some settings. The next four sections elaborate on these examples in terms of their applicability at different regional scales. First/Last Mile Services This strategy applies wherever fixed-route transit is offered by expanding the geography accessible to fixed-route transit via shared modes. Among 48 pilot collaborations between transit agencies and ridehail and micro- transit providers between 2015 and 2019, about 38 percent are first/last mile services.159 Heavily Traveled Routes All shared modes could contribute to expanding the geographic areas ac- cessible to fixed-route transit for origin-to-destination transit trips. Across all metropolitan areas, walking accounts for 91 percent of the first access modes to transit (see Table 2-2).160 In 2017, bike/Segway/scooter and taxi/ ridehailing modes accounted for only roughly 2 percent of first access modes to transit at the metropolitan area scale. Bikeshares and scooter- shares as first access mode to transit would likely occur in center cities 159 Lucken et al. 2019, Table 6. 160 This estimate, prepared by Nancy McGuckin from National Household Travel Survey data, is based on a very small sample, which in the combined “bicycle/Segway/scooter” cat- egory is mostly bicycle trips and only captures scooter trips for those respondents who wrote in an optional response to the “other” category in the survey.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 61 and inner-ring suburbs along major transit lines. Shared modes would be expected to substitute for some walking trips to transit and could substitute for a share of the estimated 7 percent of personal vehicle trips as first access mode to transit as well. As shown in Table 2-2, personal auto trips to transit stations number in the hundreds of millions annually, though estimates are lacking of how many of these trips could be substituted by shared modes. It could be sub- stantial, however, as 45 percent of city trips are 3 miles or less, a distance at which micromobility could substitute for some personal auto trips.161 A second way that shared modes and transit could reduce congestion and vehicle emissions would be by replacing trips made solely from origin to destination by automobile, as described next. A simulation of morning commute trips in the San Francisco Bay re- gion served by Bay Area Rapid Transit (BART) estimated the potential for commuters who live within 5 miles of a BART station and normally drive alone to work to connect to transit via a ridehailing trip.162 The analysis suggests that single-occupancy vehicle (SOV) commuters who would save in monetary and time costs by switching to ridehailing-transit commutes could generate 40,000 new BART trips daily and avoid half a million miles of solo auto travel. Morning peak-period VMT could be reduced by 2 percent if all the drivers who would benefit switched from drive alone for their morning commute. Two percent of VMT may seem small, but in highly congested 161 Populus. 2018. The Micromobility Revolution: The Introduction and Adoption of Elec­ tric Scooters in the United States. https://www.populus.ai/micro-mobility-2018-july. 162 Alemi, F., and C. Rodier. 2016. Simulation of Ridesourcing and Using Agent­Based Demand and Supply Regional Models: Potential Demand for First­Mile Transit Travel and Reduction in Vehicle Miles Traveled in the San Francisco Bay Area. National Center for Sustainable Transportation. https://ncst.ucdavis.edu/research-product/simulation-ridesourcing- using-agent-based-demand-and-supply-regional-models. TABLE 2-2 National Household Travel Survey 2017 Share for First Access Mode to Transit—All Trips, All Metro Areas Mode Sample Size Estimate (thousands of trips) Share Walk 10,342 8,762,011 91.0% Bicycle/Segway/Scooter 180 115,863 1.2% Personal Vehicle 1,177 675,223 7.0% Taxi/Ridehailing 135 73,187 0.8% All 11,834 9,590,284 100.0% SOURCE: McGuckin, N. 2019. Personal correspondence.

62 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY networks any small reduction in trips will have a disproportionately larger impact on easing traffic flow and reducing emissions. A second simulation in the San Francisco Bay area of how pooled ride- hailing for first access mode to BART could reduce VMT for the morning commute found much smaller effects due to the extra time cost of walking to pick-up points and waiting for a ride to a BART station.163 Even the most optimistic scenarios of pooled ridehailing trips to transit reduced morning peak VMT by only 0.2 to 0.5 percent. The small scale of these impacts is corroborated by a MOD Sandbox evaluation of a pilot project using the Scoop technology platform to encourage carpool trips to BART transit stations to replace SOV trips from home to work.164 Demand for ridehailing or micromobility trips as first/last mile connections to transit may remain quite modest in the absence of congestion and parking fees and because of the widespread availability of employer-subsidized parking, which provides an incentive for employees to drive (either alone or in a carpool) to their workplaces, as discussed in Chapter 5. Lightly Populated Areas Shared modes could also expand the reach of transit by providing links between existing transit lines and Flex zones in low-density suburban areas that lack fixed-route service. Ridehailing and microtransit would be logical modes for such first/last mile trips because of the distances involved, but e-bikes and e-scooters could also be attractive to some travelers in appro- priate weather conditions. Although such service would benefit areas that would otherwise not be readily accessible to transit, the number of such trips would be a relatively small share of transit ridership because the popu- lation densities of such areas are low.165 Another market for first/last mile trips in lightly populated areas would connect passengers living in distant towns and exurbs with commuter rail and bus service, a service that accounts for 6 percent of total transit trips.166 163 Jaller, M., et al. 2019. Estimating Activity and Health Impacts of First and Last Mile Transit Access Programs for Work and Shopping Trips Using Shared Mobility Services in the Metropolitan Area. Center for Transportation, Environment, and Public Health. https://cpb- us-w2.wpmucdn.com/sites.coecis.cornell.edu/dist/6/132/files/2019/03/UCD_YR1_JALLER_ RODIER_FINAL_ESTIMATING_ACTIVITY1-1ngtlv3.pdf. 164 Martin, E., et al. 2020. Mobility on Demand Sandbox Demonstration: BART Integrated Carpool to Transit Access Program Evaluation Report. Federal Transit Administration. https:// www.transit.dot.gov/research-innovation/report-summary-mobility-demand-mod-sandbox- demonstration-bart-integrated-carpo-0. 165 Volinksi. 2019. 166 APTA (American Public Transportation Association). 2019. 2019 Public Transporta­ tion Fact Book, Figure 5. https://www.apta.com/research-technical-resources/transit-statistics/ public-transportation-fact-book.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 63 These modes can also benefit from first/last mile connections via shared modes in home-to-station trips where commuter parking is limited or more expensive than shared mode trips to and from stations. Paratransit Ridehailing and microtransit as real-time demand-responsive paratransit for eligible senior citizens and disabled travelers is not limited by geography. Such pilots represent about 23 percent of transit agency partnerships with shared mobility providers and occur in cities of all sizes as well as in town- ships and exurban counties.167 Examples cited previously occur in urban- ized counties (Pinellas County, Florida) and a metropolitan area of wide expanse across urban, suburban, and exurban areas (Boston’s MBTA). A similar service is also being provided in a very large center city (New York City).168 Very important in terms of equity and providing more than 200 million unlinked trips annually, existing paratransit trips reported to the NTD represent about 2 percent of all transit trips.169 This estimate, however, is an undercount because many social service and health providers use federal and state funds to reimburse eligible clients for taxi and ridehail trips that are not reported to the NTD.170 The scale of such trips is not known but could easily be double the 2 percent of transit trips reported 167 Lucken et al. 2019, Figure 2 and Table 6. 168 Parker. 2020. 169 APTA. 2019, pp. 4–5 and Table 4.3. 170 This estimate of trips understates total demand-responsive trips by disabled or low- income riders. Transit agencies receiving federal funding report to the NTD, but not taxi companies and other specialized transportation companies that provide transportation to individuals with funding from Medicaid, the Veterans Administration, and myriad other federal programs and state agencies. The absolute scale of funding for these activities and the number of trips they support is not known. Non-emergency medical transportation (NEMT), funded through Medicaid, is thought to be the largest source by far of federal support for such trips. Its funding is estimated to be in the range of $3 billion annually, or about 1 percent of total Medicaid spending. NEMT and other human service transportation is provided by taxi companies, other specialized transportation companies, and transit agencies. To provide a sense of scale, $3 billion in Medicaid NEMT would support about 100 million trips annually at the average cost for demand-responsive trips estimated by the American Public Transporta- tion Association. This would represent about one-third of demand-responsive trips reported to the NTD. Even so, adding these trips would be an undercount since dozens of other fund- ing sources and state agencies provide funding for human services transportation that is not reported to the NTD. For sources see Coordinating Council on Access and Mobility. 2019. Inventory of Federal Programs Providing Transportation Services to the Transportation­ Disadvantaged, October. https://www.transit.dot.gov/sites/fta.dot.gov/files/docs/subdoc/261/ ccam-program-inventory-summary-10-2019.pdf. Garrity, R., and K. McGehee. 2014. Impact of the Affordable Care Act on Non­Emergency Medical Transportation (NEMT): Assessment for Transit Agencies. Research Results Digest. Transit Cooperative Research Program. Trans- portation Research Board, Washington, DC.

64 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY to the NTD. The success of this model in pilot tests to date has suggested substantial latent demand that is currently capped due to lack of transit agency funding from local and state jurisdictions to support subsidies to ridehail providers of this service. Off-Peak Services By being a substitute for fixed-route service during off hours, this example would apply in suburbs or urbanized areas with transit service. Off-peak service pilot projects are occurring relatively infrequently (about 8 percent of pilots) and mostly within urbanized areas.171 The successful pilot test described previously in Pinellas County, Florida, occurred in a jurisdiction with a population density of about 1,600 people per square mile, which would seem typical of many suburbs and urbanized portions of close-in counties of metropolitan areas. Low-Density Services Transit agencies are piloting use of microtransit and pooled ridehailing to serve the general population in areas that are lightly populated and expen- sive to serve by fixed-route transit; these trips can be point to point (the service is not designed to provide first/last mile services but some riders may use it for this purpose). These kinds of services apply across all geographic areas and are the most frequently used model of pilot transit agency part- nerships with ridehail and microtransit providers (40 percent).172 Northern California alone has six projects in urban, exurban, and rural settings within and beyond the San Francisco-Oakland urbanized area. In impoverished rural areas of the San Joaquin Valley, shared modes are estimated to be less expensive per trip in roughly half of the census tracts currently receiving intermittent fixed-route services; this share would grow to 78 percent of tracts if the shared mode service was provided by a ride-splitting service like UberPool or Lyft Line and to 90 percent for carsharing.173 An evaluation of a pilot project including carsharing in this area is under way.174 171 Lucken et al. 2019, Table 6. 172 Lucken et al. 2019, Table 6. 173 Rodier, C., and L. Podolsky. 2017. Rural Disadvantaged Communities in California’s San Joaquin Valley: Existing Conditions and Conceptual Program Development. National Center for Sustainable Development, University of California, Davis, Table 2. https://escholarship. org/uc/item/4xp49309. 174 Evaluation of San Joaquin Valley Pilot That Leverages New Technologies to Provide Better Transportation Options to Rural Residents. University of California, Davis. https:// www.ucits.org/research-project/2019-44.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 65 Two cases of complete substitution of microtransit and ridehailing for fixed-route transit cited previously illustrate that they can apply in limited settings, across urban, suburban, and exurban scales. As mentioned previ- ously, Arlington, Texas, a suburban city, offers subsidized microtransit services from Via rather than fixed-route transit. Located in the Dallas-Fort Worth urbanized area, Arlington is a city of nearly 400,000 residents with a population density of about 4,000 residents per square mile. In contrast, the example of Innisfil, Ontario, is an exurban town outside of Toronto that has a population density of only 361 people per square mile.175 CHAPTER FINDINGS 2.1 Shared Mode Service to Social Goals Studies summarized in this chapter suggest that (1) a growing number of individuals are benefiting from the opportunities shared modes provide and (2) the rapid growth in ridehail trips in congested major urban centers may slow traffic speeds in these settings and increase emissions. Most of the individual users of shared modes to date are well-educated adults, mostly non-Hispanic white urban dwellers, and typically moderate- or higher-income earners. Shared mode services are not yet widely accessible across racial or income groups, but ridehail and micromobility service is spreading and state and local policies can hasten this outcome as described later in this report. The effects on safety remain unclear, with the possibility of increased injuries for micromobility users and increased motor vehicle crashes resulting from shifts in travel away from transit and walking to- ward ridehail trips. The effects of shared modes on vehicle emissions may be positive from carsharing, microtransit, and micromobility, but negative from ridehailing. The net social benefits of shared modes have not yet been determined, but public policies can enhance future social benefits and miti- gate costs through policies and strategies described in Chapters 4 and 5. 2.2 Impact of Shared Modes on Public Transit 2.2.1 Impacts of Ridehailing Taking the data and methodological caveats and conflicting evidence men- tioned earlier into account, the evidence to date is not conclusive about the impact of ridehailing on transit ridership, but it does suggest negative im- pacts for surface transit modes and in large urban areas where most transit 175 Statistics Canada. 2017. 2016 Community Profiles. https://www12.statcan.gc.ca/census- recensement/2016/dp-pd/prof/index.cfm?Lang=E.

66 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY trips occur. After only about 8 years since the rapid growth in ridehailing beginning in 2012, at the time of this writing (mid-2020), ridehailing’s impacts on transit are still unfolding. Ridehailing and other shared modes could reduce auto ownership and use for some households176 and this effect could grow as shared modes become more widely relied upon. Although use of ridehailing as the first access mode to transit is currently quite limited, ridehailing could increasingly provide complementary first/last mile connec- tions to transit, particularly rail transit, in the large cities with this option, and perhaps to bus transit as well. Subsequent chapters explore policy strat- egies local governments could employ to support complementary transit and shared mode services. Bus ridership was in decline before ridehailing was introduced but could be further reduced in cities where ridehailing is reducing traffic speeds. In cities such as New York and San Francisco, the substantial volume of ridehailing trips in the peak is enhancing individual travelers’ mobility even as it appears to be degrading traffic speeds and adding to vehicle emissions, which also increases the time penalty for using buses and van pools. Such effects may not be pronounced off peak in these areas or in smaller, less congested cities. If ridehailing use continues to climb, however, the effects of ridehailing on traffic and bus transit service could worsen in many more urban areas. See the review of strategies transit agencies and local governments can use to enhance bus system speeds and service reli- ability in Chapter 5. 2.2.2 Impacts of Micromobility As with ridehailing, micromobility appears to have mixed effects by comple- menting transit in some settings, such as in first/last mile connections to rail transit, and substituting for it in others, particularly in urban cores where shared bikes or scooters can provide faster, more direct trips than transit for some trips. At this early stage following their wide-scale introduction across metro areas, the impacts of shared e-scooter use on transit in early pilots may be overstated because trip costs are being subsidized by private investors, just as is the case for ridehailing. Price increases by e-scooter operators necessary to cover their costs may shift some riders away from using e-scooters.177 Moreover, a high proportion of e-scooter use is recreational or joyriding, which may diminish as the novelty wears off.178 176 Conway et al. 2018. 177 Lazo, L. 2019. That Scooter Ride Is Going to Cost You a Whole Lot More. The Washington Post, October 18. https://www.washingtonpost.com/transportation/2019/10/18/ that-scooter-ride-is-going-cost-you-lot-more. 178 SUMC et al. 2019, p. 47.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 67 2.3 Shared Mode–Transit Coordination 2.3.1 Ridehailing The major barriers for transit agencies identified from the survey and case examples of private mobility providers involved in the pilots reviewed above were (1) reaching successful data-sharing agreements because of ride- hail company concerns about state open records laws and losses of control of proprietary and personally identifying data and (2) ADA compliance (wheelchair access, equivalent response times).179 The lack of pilot data available to some agencies raises issues about how they would measure pro- gram impacts and meet federal reporting requirements about ridership and other statistics to FTA’s NTD. Transit agencies found ways to provide non- discriminatory service (Title VI obligations) by providing options for people without bank accounts and smartphones, and to meet ADA obligations by contracting with wheelchair providers; they were able to do so and keep the pilot project trip costs below those of regular paratransit operations. The pilot projects described in the chapter illustrate potential benefits from coordination between shared mobility providers and transit agencies. They indicate that new technologies and partnerships can both reduce the cost and improve the quality of paratransit services, thereby enhancing efficiency, and the mobility of eligible riders, thereby enhancing equity. Because of the limited data available from ridehailing companies, however, they had not produced information about how these kinds of partnerships are affecting travel behavior and sustainability. Not known is the degree to which ridehailing first/last mile services, for example, were replacing single- occupant auto trips or trips by foot or bike to transit stations, or inducing new trips. Over time, some compromises were found on data sharing by, for example, providing heat maps showing concentrations of services in certain areas but not individual trips,180 but the nature and extent of public agency access to ridehail data remains contentious. See also Finding 2.5 for more discussion about data access and sharing. 2.3.2 Micromobility It appears that, regardless of the direct role of transit agencies, cities and bikeshare operators of station-based bikesharing locate bikeshare stations near transit stations/stops and are experimenting with integrating bikeshare into trip planner apps and transit fare payment systems. Even with the rapid unfolding of dockless bikes and e-scooters, these docked bike operations 179 Curtis et al. 2019. 180 Murphy et al. 2019, p. 25.

68 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY can continue to provide first/last mile connections. The dockless model seems to be evolving quickly and shrinking; the future of dockless bike- share operations outside of cities with substantial transit services is an open question. At the time of this writing (mid-2020), the volatility of ongoing experience with dockless e-scooters and e-bikes makes it difficult to know how these services will evolve. 2.3.3 Microtransit Microtransit/DRT allows transit agencies to provide service in thinly popu- lated areas that are difficult to serve productively with fixed-route services to meet geographic coverage equity goals set by their local jurisdictions and funders. The tradeoff is that, by its nature, microtransit in lightly populated areas costs more per passenger served than fixed-route services. Service to low-density areas will reach fewer passengers per agency dollar expended, and, if agency budgets remain fixed, paying for more micro transit subsidies to serve fewer passengers means less funding for fixed-route services.181 Whether the higher cost per passenger is acceptable depends on the goals that transit agencies are trying to serve and funds they have available.182 Experience and technology may help transit agencies improve the pro- ductivity and cost of this service and serve geographic equity goals, but it involves tradeoffs in resource allocation to fixed-route services and overall productivity. 2.4 Shared Mobility–Transit Services Across Regions Shared mode–transit trips and subsidized shared mode services could apply at all spatial scales and become part of regional transportation systems. Strategies that might increase efficiency and reduce emissions, such as first/ last mile connections, are most likely to occur in center cities and inner-ring suburbs of metropolitan areas where transit is most accessible and where most transit trips already occur, but first/last mile pilot projects are also occurring in exurban and rural areas. Simulations suggest that regional first/last mile ridehailing trips to transit stations during peak periods could reduce peak-period VMT by 2 percent in San Francisco, which is a small but significant enough reduction to ease congestion and reduce emissions. A 181 Walker, G. 2019. The Problem with On­Demand Transit. Shelterforce. https:// shelterforce.org/2019/12/16/the-problem-with-on-demand-transit. 182 The tradeoffs between fixed-route and microtransit services are well framed in the criti- cism of LA Metro’s microtransit pilot and the agency response to it in Linton, G. 2019. Six Months in, Metro/Via Mobility on Demand Pilot is an Expensive Flop. https://la.streetsblog. org/2019/10/09/six-months-in-metro-via-mobility-on-demand-pilot-is-an-expensive-flop. See also Volinski. 2019, pp. 85–87.

SHARED MOBILITY AND PUBLIC TRANSPORTATION 69 range of public policies is discussed in Chapter 5 that may result in a greater shift in demand. Paratransit strategies to enhance mobility and equity could apply at any spatial scale. Although they represent 2 to 4 percent of tran- sit trips, evidence of considerable latent demand suggests that the volume and share of such trips could be larger if transit agencies had additional funding for this purpose. Off-peak services to provide ridesharing when fixed-route transit services are not operating would, of necessity, occur in the areas with fixed-route transit. Low-density services could occur in any geographic setting. Many pilot projects and new programs are under way to test all of these services. 2.5 Data Access and Sharing Data sharing by shared mobility providers, particularly ridehailing compa- nies, remains an issue. Cities want anonymized trip data to enforce micro- mobility parking regulations and to understand, and provide for, demand (adding bike lanes and parking corrals, and providing curbspace for pick- ups and drop-offs where warranted). Transit agencies in partnerships with shared mobility providers need such data to evaluate the effectiveness of their partnerships. Ridehailing companies, in particular, are highly protec- tive of their data for proprietary reasons and to protect their customers’ privacy. There are ways to address this by working with third parties that are not subject to public open records laws and can provide anonymized, privacy-protected data to cities and transit agencies. This model is already being used for the sharing of micromobility data. 2.6 Uncertainty About Private Shared Mode Business Models Uncertainty about the economic viability of the ridehailing business model also raises questions about whether ridehailing companies will have the same cost-saving advantage in the future over current paratransit pro viders. Uncertainty also applies to other private shared mobility providers. The private operations of microtransit, dockless bike and e-scooter sharing, and especially carsharing have been subject to considerable volatility in recent years, with many companies entering and exiting markets on short notice. 2.7 Importance of Evaluation A great deal of innovation is occurring in passenger transportation because of the introduction and growth of shared modes. At the moment, it is much easier to observe that innovations are occurring and that transit agencies are attempting to coordinate with shared modes than to understand the effects these efforts are having on consumer preferences and their social

70 THE ROLE OF TRANSIT, SHARED MODES, AND PUBLIC POLICY benefits and costs. Continued and expanded evaluations of pilot efforts such as those funded through FTA’s MOD Sandbox would be invaluable in this regard. The chapters that follow describe how the innovations occurring in passenger transportation can improve consumer information about travel options that serve both the individual and society and how to use public policies to expand consumer modal choices.