Recent Decline in Public Transportation Ridership: Analysis, Causes, and Responses (2022)

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.

63   Shared electric motorized scooters (i.e., shared e-scooters), such as the one shown in Figure 7-1, are one of the fastest-growing micromobility modes in the United States. There were approximately 88.5 million shared e-scooter trips in the United States in 2019 (NACTO, 2020), a 230% increase over 2018. Despite their popularity in nearly every major American city and many other cities around the world, the impact of shared e-scooters on public transit ridership is largely unknown. Similar to other forms of micromobility, like bike-sharing, there are a number of ways that shared e-scooters could potentially impact transit ridership levels. One notable value proposi- tion frequently promoted by the operators of shared e-scooters is that they can be used as a complement to transit by way of solving the “first-mile/last-mile problem.” The idea is that shared e-scooters help make transit more accessible to people whose origin or destination is beyond a short walking distance to a transit stop or station, and therefore, the presence of shared e-scooters could increase transit ridership. Alternatively, given easy access to shared e-scooters, transit riders may substitute their normal transit trips with shared e-scooter trips, which could lead to a decrease in transit ridership. A third possibility is that shared e-scooter users are riding for other purposes, such as recreation, that would neither complement nor substitute transit trips and thus have a negligible effect on transit ridership. By identifying the overall effect of shared e-scooters on transit ridership, transit agencies and city officials can more effectively develop policies to incorporate shared e-scooters into their transportation systems. 7.1 Objective of Shared E-scooters Analysis Research to date has been inconclusive, with different studies having divergent conclusions as to whether shared e-scooters are a substitute or a complement to transit, with survey results somewhat more frequently showing a complementary relationship (City of Bloomington, 2019; City of Chicago, 2020; SFMTA, 2019; City of Santa Monica, 2019). Considering these mixed findings from prior studies, this study evaluates both local and express bus ridership in Louisville, Kentucky, for the period of February 2016 through December 2019 in order to assess the following three research questions: 1. Do shared e-scooters decrease ridership on local bus routes? This would occur if current transit riders primarily substitute their transit trips with shared e-scooter trips. 2. Do shared e-scooters increase ridership on local bus routes? This would be the case if they predominantly provide first-mile/last-mile access to new local bus trips. 3. Do shared e-scooters increase ridership on express bus routes? This would happen if shared e-scooters provide first-mile/last-mile access to new express bus trips. It should be noted that shared e-scooters could not replace express bus trips in Louisville since express bus trips are typically much longer distances than would be comfortable for most to C H A P T E R 7 The Impact of Shared E-scooters on Bus Ridership in Louisville, Kentucky

64 Recent Decline in Public Transportation Ridership: Analysis, Causes, and Responses travel by e-scooter. Additionally, one of the trip ends occurs outside the shared e-scooter service area, since shared e-scooters—like those in Louisville—typically operate in a small geographic area in the center of most cities. Thus, this study did not explore the possibility that shared e-scooters could decrease express bus ridership. 7.2 Why Louisville as a Case Study? As a typical medium-sized city with bus-only transit, Louisville is an excellent case study to evaluate the impact of e-scooters on transit ridership. Louisville was selected as a case study for three reasons: 1. Limited changes to the transit system occurred during the period of February 2016 through December 2019. 2. Shared e-scooters have relatively high ridership levels in Louisville. 3. Louisville is one of a few cities that requires shared e-scooter operators to report trips through Mobility Data Specification and makes an anonymized version of the data publicly available (Github, 2020). These three reasons allow for a research design that could quantitatively explore the impact of shared e-scooters on transit ridership. 7.3 Data Sources This study used data from different data sources, as shown in Table 7-1. The unlinked bus trips per route and VRH were obtained from the Transit Authority of River City (TARC). The shared e-scooter trips data were obtained from the Louisville Metro Open Data repository (2020). Other variables—like population, employment, and weather—were obtained from publicly available data sources. Figure 7-1. Shared e-scooter.

The Impact of Shared E-scooters on Bus Ridership in Louisville, Kentucky 65   7.4 Assigning Shared E-scooter Trips to Transit Routes The shared e-scooter trip data used in this analysis have the start and end points but not the complete route of the shared e-scooter trip. Therefore, this study proposed an assignment method to attribute shared e-scooter trips to transit routes as described in this section. To assign shared e-scooter trips to transit routes, a tight transit catchment area of 0.1 mile was first drawn around each bus route. Then, using this catchment area and shared e-scooter trip start and end points, different variables were defined for each bus route, as shown in Figure 7-2 and described as follows. 7.4.1 Assignment of Shared E-scooter Trips to Local Bus Routes Three different variables were defined for local bus routes to assess the first two research questions: 1. Do shared e-scooters increase ridership on local bus routes? 2. Do shared e-scooters decrease ridership on local bus routes? The first variable defined for local routes was the shared e-scooter substitute trip count. This variable was defined to assess the possibility that shared e-scooters are used to substitute transit trips, in which case they will result in reduced bus ridership. Shared e-scooter trips were defined Bus VRH TARC Shared e-scooter trips Louisville Metro Open Data repository Population One-year American Community Survey Employment Bureau of Labor Statistics Annual median income of individual ($) One-year American Community Survey Weather data (average temperature, precipitation, snowfall) National Oceanic and Atmospheric Administration Variable Data Source Unlinked bus-trips TARC Table 7-1. E-scooter analysis data sources. Shared e-scooter substitute trip: longer than 0.1 miles and both the shared e-scooter trip start and end points were located within the catchment area for a specific local bus route. Figure 7-2. Assignment of shared e-scooter trips to bus routes.

66 Recent Decline in Public Transportation Ridership: Analysis, Causes, and Responses as substituting if the shared e-scooter trip distance was greater than 0.1 miles and both the shared e-scooter trip start and end points were located within the catchment area for a specific local bus route. The total count of competing shared e-scooter trips was determined for each local bus route for each day during the period of analysis. Figure 7-3 shows an example of the shared e-scooter trip assignment method used for one day’s worth of trip data for one local bus route. In Figure 7-3, the blue line represents the local bus route, and the orange line shows the catchment area on either side of the bus route. The shared e-scooter trip origins (start points) are shown as red dots, and the destinations (end points) are shown as green dots in Figure 7-3. The inset map shows the trips that would be counted for this day and this route, which is a total of 12 shared e-scooter trips. For these trips, the origin and destination (which are shown as “linked” in the inset) are both within the catch- ment area for the transit route, implying that these shared e-scooter trips could have replaced transit trips along this route. The second variable was the shared e-scooter first-mile connector trip count. This variable was defined as the number of shared e-scooter trips that have destinations within the bus route catchment area. The assumption is that users could ride a shared e-scooter to get to a local bus stop, as shown in Figure 7-4. This variable was used to evaluate the possibility that shared e-scooters are used as a first-mile connector from the trip starting point to a local bus stop. Figure 7-3. Example of shared e-scooter trip assignment to a local bus route.

The Impact of Shared E-scooters on Bus Ridership in Louisville, Kentucky 67   The third variable was the shared e-scooter last-mile connector trip count. This variable counts the number of shared e-scooter trips that have origins within a bus route’s catchment area. The assumption is that riders could take a shared e-scooter from the bus stop to their final trip des- tination, as shown in Figure 7-5. This variable was used to evaluate the possibility that shared e-scooters are used as last-mile connectors from a local bus stop to the user’s trip end point. 7.4.2 Assignment of Shared E-scooter Trips to Express Bus Routes Shared e-scooters could complement express routes, as they offer a solution to the first-mile/last mile problem that riders may experience between downtown express bus stops and their trip end points (e.g., place of work). Therefore, the shared e-scooter complement trip count was defined for express routes only to assess the third research question. A shared e-scooter trip was defined as complementary for an express bus route if the shared e-scooter trip origin was located within an express route bus stop catchment area during morning hours, or if the shared e-scooter trip destination was located within the catchment area during evening hours. The assumption is that commuters would exit the express bus downtown in the morning and then could begin a shared e-scooter trip for the last portion of their commute. In the evening, commuters could ride a shared e-scooter from work to the express bus stop for the first portion of their evening commute and then board the express bus for the remainder of their trip home. Morning trips correspond to shared e-scooter trips whose start times were between 4 a.m. and 10 a.m., while evening trips correspond to shared e-scooter trips whose end times were between 1 p.m. and 8 p.m. The morning and evening hours selected correspond to the hours of service for the express bus routes in Louisville. Figure 7-6 shows an example of the shared e-scooter trip assignment method used for one day’s worth of trip data for one express bus route. 7.5 Results of Shared E-scooters Analysis Using multiple fixed-effects regression models, shown in more detail in Appendix G of TCRP Web-Only Document 74, this study assessed the three research questions by testing a variety of models using different types of data and/or variables: weekday data, weekly data, monthly data, and an expanded monthly model with other non-related potential demand forecasting explanatory Figure 7-4. Shared e-scooter first-mile connector trip count. Figure 7-5. Shared e-scooter last-mile connector trip count.

68 Recent Decline in Public Transportation Ridership: Analysis, Causes, and Responses variables (e.g., population, income, weather). In all the models, the variable being explained is the total unlinked bus trips per route. A summary version of the most important model results is shown in Table 7-2. Model 1 in Table 7-2 evaluates the first research question (Do shared e-scooters decrease ridership on local bus routes?) and the third question (Do shared e-scooters increase rider- ship on express bus routes?). The results of Model 1 reveal that, unsurprisingly, the number of VRH is a significant positive predictor for bus ridership, as suggested by the significant positive coefficient. This coefficient indicates that each additional VRH on a route is expected to increase bus ridership by 31 trips per month, holding all else constant. In terms of the research questions being explored, Model 1 indicates that there is no statisti- cally significant evidence that shared e-scooter trips are being used as a substitute for bus trips. On the contrary, Model 1 suggests that the shared e-scooters complement trip counts for express bus routes; shared e-scooter trips are a statistically significant but very modest predictor for bus ridership. The shared e-scooter complement trip count coefficient indicates that each additional shared e-scooter trip within the catchment area of an express bus route is expected to increase bus ridership by 0.66 trips, holding all else constant. However, this result should be interpreted with caution. In Louisville, all express routes terminate in the same geographic area near down- town, which resulted in similar counts of shared e-scooter trips along all express routes. These Figure 7-6. Example of shared e-scooter trip assignment to an express bus route.

The Impact of Shared E-scooters on Bus Ridership in Louisville, Kentucky 69   similar counts limit the variability of this variable in the model, which is one of the limitations of this experimental design. Therefore, the relationship between express bus ridership and shared e-scooters requires further study. Models 2 and 3 in Table 7-2 assess the possibility that shared e-scooters increase ridership on local bus routes. Model 2 assumes that bus riders will use shared e-scooters as first-mile con- nectors to go from their trip start points to bus stops, while Model 3 assumes that bus riders will use shared e-scooters as last-mile connectors to go from bus stops to their trip end points. The results of these models show that both the shared e-scooter first-mile trip connector count and the shared e-scooter last-mile trip connector count are not significant predictors for bus ridership in Louisville. Those findings suggest that there is no statistically significant evidence that shared e-scooter trips are being used as first-mile/last-mile connectors to local bus routes in Louisville. Also, both models show that the shared e-scooter complement trip count is a significant predictor of bus ridership on express routes as indicated by the significant positive coefficient. This coef- ficient indicates that each additional shared e-scooter trip within the catchment area of an express bus route is expected to increase bus ridership by 0.65 trips, holding all else constant. This finding is consistent with Model 1, but again it should be interpreted with caution due to limitations in the modeling methodology. The results of all three models suggest that population and employment have a positive impact on ridership, but it is not significant. The reason behind this is likely the small change in popula- tion and employment during the study period, which only lasted about three years. This finding is consistent with prior studies that found changes in population had modest effects on bus rider- ship in the short term (Berrebi and Watkins, 2020; Ederer et al., 2019). Models 1–3 also show that income has a significant negative impact on bus ridership. Similarly, all three models suggest that rainfall and snow have significant negative impacts on bus ridership. These findings are expected since less transit usage is expected during rain and snow. These findings are consistent with the outcomes of prior studies (Brakewood et al., 2015; Ngo, 2019; Owen and Levinson, 2015). Dependent Variable: Unlinked Bus Trips per Route per Month (Model 1) (Model 2) (Model 3) VRH 31.0*** 31.0*** 30.9*** Shared e-scooter substitute trip count (local routes) -0.05 Shared e-scooter first-mile connector trip count (local routes) -0.14 Shared e-scooter last-mile connector trip count (local routes) -0.13 Shared e-scooter complement trip count (express routes) 0.66** 0.65** 0.65*** Population and employment (1,000s) Not Significant Annual median income of individual ($) Significant Negative Effect Average temperature (ºF) Not Significant Rainfall (inches) Significant Negative Effect Snowfall (inches) Significant Negative Effect Route Controlled for differences between routes Month Controlled for differences between months Number of observations 1,980 Note: Full model results are shown in Appendix G of TCRP Web-Only Document 74. Variable significance: ***p-value < 0.01; **p-value < 0.05; *p-value < 0.10; no star = not significant Table 7-2. Impact of shared e-scooters on unlinked bus trips model results.

70 Recent Decline in Public Transportation Ridership: Analysis, Causes, and Responses 7.6 Conclusions, Discussion, and Implications of Shared E-scooters Analysis As shared e-scooters are gaining popularity in the United States, their impacts on transit ridership are still largely unknown. This study conducted an empirical analysis to explore the impacts of shared e-scooters on bus ridership using Louisville as a case study. The results of this study suggest that shared e-scooters do not have a significant impact on local bus ridership, either as competitor or as a first-mile/last-mile complement. This finding could possibly be explained by three factors. • First, transit and shared e-scooters are often used for different trip purposes. Transit in Louisville is mainly used for work and school; a recent survey indicates that 70% of the transit trips are for work and school (Copic, 2019). However, prior studies of shared e-scooter trip patterns indicate that shared e-scooters might not be used for commuting and are likely used primarily for recreation (Caspi et al., 2020; Noland, 2019). These different purposes suggest that shared e-scooters are not used to substitute transit trips in Louisville. • Second, the average trip length between these modes also suggests they are being used for different trip lengths. The average transit trip length in Louisville is 4.2 miles compared to 1.2 miles for shared e-scooters. • Third, transit users in Louisville are typically minorities and have lower household incomes (Copic, 2019). This is likely different from typical shared e-scooter users; surveys from dif- ferent cities suggest that shared e-scooter users are more likely to be white and have higher household incomes (SFMTA, 2019; Mobility Lab, 2019). This suggests that transit and shared e-scooters are used by different demographic groups, which might limit the interaction of these two modes. Overall, this implies that shared e-scooters are not one of the primary reasons for declines in local bus ridership. Last, this study also explored the relationship between shared e-scooters and express bus routes in Louisville and found that shared e-scooters may complement express bus routes as first-/last-mile connectors. However, this finding should be interpreted with caution because of the similar counts of shared e-scooter trips along all express routes, which is one of the limi- tations of this experimental design. Therefore, the relationship between express bus ridership and shared e-scooters requires further study. However, transit agencies and shared e-scooter operators could explore ways to integrate these two modes to offer better service for their users. There are many ways that such an integration could occur, such as developing multimodal trip planning platforms and price bundling of the two modes. Shared e-scooters are not one of the primary causes of bus ridership decline in Louisville.