ADA Paratransit Service Models (2018)

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.

99 Service Model Types In Chapter 3, ADA paratransit service models were described as having three components: • A management structure, • How work is divided among multiple dedicated carriers (zoned, unzoned, other), and • Whether non-dedicated service providers (in most cases, taxis, and in one case, livery opera- tors) are used to provide ADA paratransit trips. When these three components are considered together, the response to the survey unveils that there are actually 25 different service models represented among the 29 transit agencies that fully or partially completed the survey. This broad spectrum of models reinforces the pervading adage in the paratransit industry that “everybody does it differently.” The 25 different service models from these 29 respondents are reflected in Figures 8–10: • Figure 8 presents the 12 service models in which transit agencies have a direct role in some or all of the four primary call center functions (reservations, scheduling, dispatching, and handling customers’ ETA calls), and in some cases, operations as well. • Figure 9 presents the eight service models in which transit agencies have retained a call (and control) center contractor or a broker to perform some or all of the four primary call center functions. Note that there is an overlap with two service models in Figure 8. • Figure 10 presents the 15 service models in which the dedicated service provider contractor(s) also perform(s) some or all of the call center functions. Note that there is an overlap with eight service models in Figure 8 and Figure 9. In the far-right column of each of the three figures is an indication whether the transit agency has elected to offer an alternative service (a taxi and/or TNC-based subsidy program) for their ADA paratransit customers. This is presented for context only as these alternatives services are not included within the ADA paratransit service model. One other note: While we did collect from the survey data on whether the transit agency or an eligibility contractor performed ADA paratransit eligibility determination/certification, this information is not presented in these figures, as it is largely independent of the service model. Themes unveiled by a closer look at these different groupings of service models follow Figure 10. C h a p t e r 5 Survey Analysis Summary

100 ADA Paratransit Service Models Model Number Locations Reservations Scheduling Dispatching Handling Customers’ ETA Calls Operations (# of Dedicated Service Contractors) Service Area Zones? Integrated NDSPs (% of Total Trips) Alternative Service Offered? 1 Nashville Agency N/A Yes (15%) No 2 Pierce County Agency Agency Agency Agency & DS Contractor (1) No No No 3 Salt Lake City Agency & DS Contractors Agency & DS Contractors (2) Yes (3) No No 4 Austin Agency DS Contractors (2) No Yes (3%) No 5 Houston DS Contractors (2) No Yes (11%) Yes (Taxi) 6 Philadelphia DS Contractors (7) Yes (5) No No 7 Ann Arbor Agency & DS Contractor DS Contractor No No No 8 New Jersey Agency Agency & DS Contractors DS Contractors Agency DS Contractors (6) Yes (6) No No 9 Atlanta DS Contractor DS Contractor DS Contractor (1) N/A No No 10 Las Vegas Agency & DS Contractor No Planned (TNC and Taxi) 11 Kansas City Agency & DS Contractor Call Center Contractor DS Contractors (2) Yes (2) Yes (17%) Yes (Taxi) 12 New York Call Center Contractor Agency DS Contractors Call Center Contractor DS Contractors (13) Yes (17) Yes (30%) via 2 NDSP Brokers Yes (TNC and Taxi) Figure 8. ADA paratransit service models in which transit agencies perform one or more primary functions. Model Number Locations Reservations Scheduling Dispatching Handling Customers’ ETA Calls Operations (# of Dedicated Service Contractors) Service Area Zones? Integrated NDSPs (% of Total Trips) Alternative Service Offered? 11* Kansas City Agency & DS Contractor Call Center Contractor DS Contractors (2) Yes (2) Yes (17%) Yes (Taxi) 12* New York Call Center Contractor Agency DS Contractor Call Center Contractor DS Contractors (13) Yes (17) Yes (30%) via 2 NDSP Brokers Yes (TNC and Taxi) 13 Denver Call Center Contractor DS Contractors (3) No Yes (10%) Yes Seattle DS Contractors (2) No Yes (9%) Yes 14 Washington, DC Call Center Contractor DS Contractors (3) Yes (2) Yes (6%) Yes (Taxi) Planned (TNC) 15 Portland Call Center Contractor** DS Contractor** Yes (3) No No 16 Oakland/East Bay Operational Broker DS Contractors (3) No No No 17 Broward County Call Center Contractor DS Contractors Call Center Contractor DS Contractors (3) No No Planned (TNC) 18 Chicago DS Contractors CC Contractor DS Contractors CC Contractor DS Contractors (4) Yes (3) No Yes (Taxi) * The Kansas City and New York service models are repeated here (from Figure 8) because they also involve a call center contractor. ** In Portland the same contractor that provides the call center functions also provides dedicated service operations. Figure 9. ADA paratransit service models involving a call and control center contractor or an operational broker.

Survey Analysis Summary 101 Direct Transit Agency Involvement Different Levels of Involvement As shown in Figure 8, among the identified transit agencies, 12 of the 29 transit agencies that responded to the survey directly perform all or some of the primary functions in-house. Direct involvement included: • Directly performing all four primary call center functions (reservations, scheduling, dispatch- ing, and handling customer ETA calls) and service delivery (1) • Directly performing all four primary call center functions in-house and operating a portion of the dedicated service (2) • Directly performing all call center functions, but using contractors for all service delivery (3) • Directly performing some of the call center functions, and using contractors for the other call center functions (6). Model Number Locations Reservations Scheduling Dispatching Handling Customers’ ETA Calls Operations (# of Dedicated Service Contractors) Service Area Zones? Integrated NDSPs (% of Total Trips) Alternative Service Offered? 3* Salt Lake City Agency Agency & DS Contractors Agency Agency & DS Contractors Yes (3) No No 7* Ann Arbor Agency & DS Contractor DS Contractor No No No 8* New Jersey Transit Agency Agency & DS Contractors DS Contractors Agency DS Contractors (6) Yes (6) No No 9* Atlanta Agency DS Contractor N/A No No 10* Las Vegas Agency DS Contractor Agency and DS Contractor DS Contractor N/A No Planned (TNC and Taxi) 12* New York Call Center Contractor Agency DS Contractor Call Center Contractor DS Contractors (13) Yes (17) Yes (30%) via 2 NDSP Brokers Yes (TNC and Taxi) 17* Broward County Call Center Contractor DS Contractors Call Center Contractor DS Contractors (3) No No Planned (TNC ) 18* Chicago DS Contractors CC Contractor DS Contractors CC Contractor DS Contractors (4) Yes (3) No Yes (Taxi) 19 Dallas DS Contractor DS Contractor & Subcontractor N/A Yes (70%) Yes (TNC and Taxi) Orange County Yes (23%) Yes (Taxi) 20 San Francisco DS Contractor DS Contractor & Group Trip Subcontractor N/A No Yes (Taxi) 21 Milwaukee DS Contractors Yes (2) No Yes (Taxi) 22 Boston DS Contractors Yes (3) Yes (1%) Yes (TNC/Taxi) Phoenix Yes (4) Yes (<100%) Yes (Taxi) 23 Los Angeles DS Contractors Yes (6) Yes (39%) No 24 Pittsburgh Yes (6) No No 25 Columbus DS Contractor No No No Orlando No No Yes (Taxi) * These service models are repeated here (from Figure 8 and Figure 9) because they also involve dedicated service providers performing some of the call center functions. Figure 10. ADA paratransit service models where primary operations contractor(s) also perform one or more call (and control) center functions.

102 aDa paratransit Service Models Directly Performing All Four Primary Call Center Functions and Service Delivery Of the survey respondents, the MTA in Nashville is the only transit agency that performs all call center functions and operates 100% of the dedicated service provided in-house. MARTA in Atlanta formerly fell into this category, but recently privatized all primary functions except reservations; MARTA was in the midst of responding to union grievances related to this out- sourcing, and may have changed its model further depending on the outcome. Directly Performing All Four Primary Call Center Functions In-House and Operating a Portion of the Dedicated Service Pierce Transit in Pierce County (Tacoma), Washington, and the Utah Transit Authority (UTA) in Salt Lake City perform some or all of the call center functions, while also directly operating a portion of the dedicated service along with other service provider contractors. In the case of Pierce County, the dedicated service operated with Pierce Transit employees consists of a set of attractive runs: straight runs during the day on weekdays only with heavy concentrations of standing orders. In Salt Lake City, UTA has divided up its service area into three regions (North, Central, and South). UTA performs all four call center functions and directly operates service in the Central region (which includes Salt Lake City), and uses two service provider contractors in the northern and southern region, both of whom perform their own dispatching. All other call center functions for the two remote regions are performed directly by the UTA. Directly Performing All Call Center Functions, but Using Contractors for All Service Delivery The transit agencies in Austin, Houston, and Philadelphia perform all call center functions but retain service providers for service delivery. In the case of Austin and Houston, transit agencies’ employees have performed these functions on a centralized basis for years. It is also noteworthy that in Houston, the service providers supply employees in the central dispatch area to provide radio assistance to their drivers as needed. Austin has three dedicated service providers plus an overflow NDSP, while Houston has two dedicated service contractors plus an overflow NDSP. SEPTA in Philadelphia has an unusual arrangement in which its reservations and same-day issue staff are contract employees. In Philadelphia, three service provider contractors provide service within the city, while four additional contractors each provide county-based service for the four suburban counties. Directly Performing Some of the Call Center Functions, and Using Contractors for the Other Call Center Functions In Atlanta, Ann Arbor, Kansas City, Las Vegas, and New Jersey, the reservations function is performed directly by transit agency employees. • In Atlanta (and pending arbitration), reservations is the only primary function that is per- formed in-house. All other call center functions in Atlanta, including responding to custom- ers’ same-day issues, are performed by the service provider contractor. • In Las Vegas, which has a similar split of responsibilities, the reasoning behind retaining the reservations function has to do with control, especially since the RTC has operationalized conditional eligibility. • In Kansas City, transit agency employees, as in Atlanta and Las Vegas, are involved in book- ing reservations (on weekdays only); however, KCATA also retains a call center contractor to perform the other call center functions, including the booking of reservations on weekends and holidays. KCATA specifically decided not to give this scheduling function to its service provider contracts because it felt that it created some conflicts of interest. While KCATA has

Survey analysis Summary 103 two service providers that serve different areas (in different states), 99% of the trips are served by one of the contractors. • In New Jersey, separate staffs perform reservations functions, scheduling, and handling same- day issues. The scheduling functions are collaboratively shared with the schedulers of the six regional service provider contractors, and dispatching is the province of those contractors, a decision made from the inception of NJ TRANSIT’s Access Link because of the vastness of the state and regional differences. It is also noteworthy that, like SEPTA in Philadelphia, NJ TRANSIT uses contract employees as a cost savings strategy but, in New Jersey, they are interspersed with agency employee reservation agents and “service monitors” (who field the same-day issue calls). • In Ann Arbor, AAATA recently changed from one unique service model to another, noting that AAATA’s ADA paratransit and senior paratransit service are merged and that same- day requests are accepted for both. Formerly, AAATA directly operated a fleet of accessible vehicles while a local taxi contractor operated a fleet of dedicated sedans. Depending on what type of vehicle a customer needed, the customers call to book a trip was routed to the AAATA or the contractor. This recently changed to a service model where AAATA performs reserva- tions and scheduling for all advance reservation trips, and the contractor intakes same-day reservations and performs dispatching and service delivery for all trips. • In New York, while the call center manager focuses on reservations and handling ETA calls, and monitors dispatching, it is the NYCT staff who perform scheduling. Advantages and Disadvantages of Using Transit Agency Staff to Provide All or Some Functions The reasons for direct transit agency involvement are fairly straightforward. The following rationale came from one of the above transit agencies. “Performing all functions in-house offers more direct control and flexibility; this enables us to make direct adjustments to the balance of service quality and cost, especially with budgetary limitations in mind.” With contractors, whether they are providing call center functions and/or service delivery, achieving the desired balance can only be done indirectly by way of contractual performance standards, incentives, and liquidated damages, and for unzoned multi-carrier systems, mid- contract adjustments to the package of work assigned to that contractor. In cases in which transit agencies perform all or some functions, such adjustments can be more easily and quickly implemented, even under a union shop. In follow-up interviews with the transit agencies, the consensus was that good dispatching was paramount to both on-time performance and service productivity (it is one of the few things that positively impacts both), and that poor dispatching can easily sink a good schedule. In the infancy of paratransit, in the 1970s and 1980s, the conventional wisdom was to always keep the dispatching function with the operator. If a system used one or more service provider contrac- tor, those contractors would provide the dispatching even if reservations and scheduling were done by a transit agency or broker. DART in Dallas was among the first to test and successfully prove the efficacy of dispatching centrally with transit agency employees. This led to a new consideration of centralized dispatch, which has since been replicated by other transit agencies, whether performed by the transit agency, a broker, or a call center manager. Indeed, among the 20 transit agencies selected for the survey where the entity that provides one or more call center functions is not the same entity that delivers service, half (10) provide dispatch centrally and five of those perform dispatch with transit agency staff. While this has worked well, other systems, such as Pace in Chicago, have purposely avoided such a service model because of the blurred employer-employee relationship that results from

104 aDa paratransit Service Models dispatching by anyone other than the vehicle operations contractor. In Houston, as mentioned, the service providers supply employees in the central dispatch area to provide radio assistance to their drivers as needed. In the cases where transit agency staff field ETA calls from customers, the general feeling is that fielding these calls directly gives the transit agency a better sense of the quality of service. In addition, as a general warning from those transit agencies that have opted to operate some or all of its service in-house: it is very difficult to outsource an operation once it is brought in- house, per 13(c) protections. Atlanta, where MARTA outsourced all but the reservations function to a contractor in 2015/2016, has had to respond to a union grievance related to the outsourcing. Another general downside to providing in-house staff, as voiced by some of the survey par- ticipants, is the hiring and firing limitations that often constrain public entities, especially where staff is represented by unions. In contrast, private sector contractors and especially the large national firms can be more nimble in adjusting staffing levels, even where a union is involved. Some of the survey respondents reported cases in which large contractors have been able to “bor- row” from their other contracts in other cities to help out with temporary needs for additional staff or vehicles, something that cannot be matched if the service is provided all in-house. Use of Brokers and Call and Control Center Contractors Both follow-up interviews with transit agencies who still have brokers, after decades, con- firmed that the brokerage model was originally installed because (a) the transit agency at the time did not have the in-house talent to organize the system; (b) because there were multiple sponsors; and (c) because there was an intent to have multiple carriers. Follow-up interviews with transit agencies who have opted to install call and control center manager models versus brokers all indicated that they wished to take advantage of a neutral, experienced entity performing all or some of the call and control center functions but wanted direct leverage over the carriers by directly contracting with them versus a brokerage model where the broker contracts with the carriers. In these cases, the transit agencies indicated that they had a willingness to directly monitor the carriers with experienced in-house staff to do that job. Examples of Brokers and Call and Control Center Managers As depicted in Figure 11, the basic difference between a broker and a call center manager is in the contracting arrangement. In the case of a broker, regardless of what operational type functions the broker directly per- forms, if any, the defining feature of a brokerage is that the transit agency enters into one con- tract with the brokerage contractor and the brokerage contractor in turn contracts with the service providers. There are two types of brokers: an administrative broker and an operational broker (also sometimes called a “full service” broker). Administrative Broker The major functions include organizing the service delivery network and retaining and over- seeing the contractors, monitoring service, reporting and customer service functions; also, an administrative broker might perform the eligibility function in-house and/or use an eligibility contractor. Otherwise, all the call center functions are vested with the service provider contrac- tors. Pittsburgh is perhaps the best example, and from the survey, the only example of using an administrative broker. A noteworthy aspect about this brokerage, which has been in existence since 1979, is that ADA paratransit trips collectively represent only 31% of the ridership, with

Survey analysis Summary 105 senior, human service agency, and general-public trips making up the balance. The Port Author- ity of Allegheny County holds the primary contract with the current broker, Transdev, and funds not only the ADA but the majority of senior trips as well. However, there are also more than 100 human service agencies, including Medicaid, that have separate agreements with the broker—not involving the Port Authority—to use ACCESS for client transportation. ACCESS organizes Pittsburgh and the surrounding portions of Allegheny County into six zones to which service provider contractors are assigned. Note, too, that while San Francisco does involve an administrative broker, ADA paratransit trips are handled through a turnkey contractor model with the same contractor that is retained as the administrative broker. Operational (or Full-Service) Broker There exists among the survey respondents one example of an operational broker, in Oakland/East Bay. As retained by AC Transit and BART, the East Bay Paratransit Consortium broker, Transdev, performs all four call and control functions. Transdev then contracts with three service providers who share equal unzoned packages of work. The broker’s fixed costs are paid separately from the variable costs, with the fixed costs split into equal monthly payments. BART indicated that it was important to have an operational brokerage contractor that is differ- ent from the service provider so that there is no inherent conflict of interest. If that firm also oper- ates a portion and that portion of the contract is paid based on a per-trip rate, the schedulers could maximize revenue for the firm by scheduling many short and ride-sharable trips to the runs it also operates. If that firm also operates a portion of the service and operations is paid based on an RVH rate, it was foreseen that the broker firm could assign itself the “gravy” runs (Monday–Friday, 8:00 to 5:00), while assigning the less desirable runs to the other carrier(s). In Washington, DC, this came to pass and was one of the reasons why WMATA decided to switch to a more traditional call and control center manager service model in which that contractor was prohibited from also operating. Another specific type of broker that is now emerging with more and more transit agencies using non-dedicated service providers is the NDSP broker. An NDSP broker may be used by a transit agency that wishes to use a great number of different NDSPs and does not want the burdensome administrative responsibility to contact with each NDSP. This may be the case, for example, where taxis are regulated by municipality and there are numerous municipalities encompassed by the ADA paratransit service area. Thus, rather than having to contract with Source: Created by Nelson\Nygaard Consulting Associates. Figure 11. Contracting arrangements of brokers and call center managers.

106 aDa paratransit Service Models numerous NDSPs, the transit agency instead has one contract with an NDSP broker, and that broker is then responsible for ensuring that all NDSP drivers are trained and monitored, and that NDSP vehicles are insured up to NYCT-required insurance levels. From the survey, there is only one example of this, in New York City. NYC Transit retains two NDSP brokers that arrange for “black car” and other livery service providers, and not taxis, to provide non-dedicated service for ADA paratransit trips that cannot be scheduled onto dedicated runs. Call (and Control) Center Manager This service model is distinguished from a broker by the contracting arrangement. As depicted in Figure 11, the transit agency not only holds the contract with the call and control manager but also with the service provider contractors. Otherwise, the call and control manager resembles an operational broker in that it can provide any or all of the four primary call center functions: reservations, scheduling, dispatching, and handling customers’ same-day issues. The “control” in the model title is most often attached to call center managers that also dispatch. This is the case in the Denver and Seattle regions, where First Transit is the call and control center contractor, and in the Washington, DC, region, where MV Transportation is the CCC manager. However, in Seattle, King County Metro recently decided to migrate from a call and control center model to a single turnkey contractor model. This was because the responsibility for ser- vice delivery had evolved to one primary contractor, with a second contractor handling a much smaller piece of the work. Under the current model in Seattle, another interesting design element is that the call and control center contractor, First Transit, also serves as an NDSP broker, as it is the entity that contracts with the taxi companies. Under the new model, it is unclear whether this contracting function will be transferred to King County Metro or the turnkey service provider contractor. King County Metro’s decision is actually in concert with a decision of DART in Dallas from a few years ago. As mentioned, DART had pioneered the concept of centralized reservations, scheduling, and dispatch under a service model that utilized two multiple-service provider con- tractors. However, DART collapsed service delivery into one contract, and eventually the agency decided to change to a turnkey contractor model, also with one subcontractor. Their decision, largely based on reducing cost, followed the same logic as in Seattle: why have a separate call and control center when you have only one, or one primary, service delivery contractor? As with operational brokers, few transit agencies have allowed a call and control center manager to also operate service, because of inherent conflicts of interest. In Washington, DC, WMATA had a history of using a traditional operational broker, with centralized reservations and scheduling and decentralized dispatching. WMATA then changed to a hybrid broker, where the broker was permitted to also operate service. Conflict of interest issues did arise, and WMATA has since changed to a call and control center manager that performs the call and control functions of reservation, scheduling, dispatching, and handling customers’ same-day function. Under this service model, WMATA contracts with two service provider contractors (First Transit and Transdev) to serve the District of Columbia and two counties in Maryland, and with a third contractor (a local carrier) to serve Northern Virginia. Note that Transdev also supplied non-dedicated taxis as well. In Portland, TriMet currently has two contracts with the same firm (First Transit): one to per- form call center functions, and one to provide service delivery. The service area is organized into three regions and at one time there were three service providers assigned to those three regions, all of whom were different from the call and control manager. However, as a result of firm merg- ers and acquisitions, TriMet now has one call and control manager and one service provider contractor who are the same firm (i.e., both contracts are with First Transit). When considered

Survey analysis Summary 107 together, this could be seen as equivalent to a single turnkey contractor. That said, TriMet stresses some advantages to this—having to work with only one company, for example—but keeping the structure open in case procurements in the future result in retaining multiple contractors for different elements of the service. The service models in Broward County and New York City are examples of call center service models where the call center contractor does not provide all four primary call center func- tions. In both locations, the service provider contractors perform their own dispatching. In Fort Lauderdale, Broward County Transit uses a local company (Inktel) for its call center func- tions, and retains two unzoned service provider contractors, with each getting about half the trips. The Broward County call center contractor is paid on a flat per-service-hour rate. The service model in New York differs from Broward County’s service model in two respects. First, scheduling is the province of transit agency staff, who use the scheduling software to batch schedule the trips onto the dedicated runs operated by the 13 service provider contractors, allow- ing the service providers to fine-tune the schedule. Second, unscheduled trips (i.e., trips that fall out of the scheduling process as unassigned) are automatically assigned to the NDSP brokers (as mentioned earlier). To improve service productivity and cost efficiency in Kansas City, KCATA retained a call center manager (RMS), the same organization that provides its paratransit scheduling software, to perform scheduling, dispatching, and the handling of same-day issues. The decision to retain a separate contractor for these functions was seen as preferable to vesting these functions with the primary service provider contractor (creating a conflict of interest) or keeping these func- tions in-house. Note, too, that KCATA opted to keep the reservations function in-house on weekdays with Transdev the service provider covering the reservations function on weekends and holidays, as a union concession. Service is provided by two contractors based on geography, but one contractor (Transdev) provides 99% of the dedicated service. Transdev also provides non-dedicated service. Thus, in essence, the KCATA decision was the exact opposite of the King County Metro and DART decisions under much the same circumstances. Advantages and Disadvantages of Using Brokers or Call (and Control) Center Managers Operational brokers and call center models are mostly used where both of the following con- ditions apply: (1) a transit agency wishes to divide up the work among more than one service provider contractor; and (2) a transit agency wishes some or all of the call center functions to be centralized but does not want to directly perform all or some of those functions. As mentioned, if there is only one, or one primary, service provider contractor, most agencies have opted to just go with a single turnkey contract instead of retaining a contractor for call cen- ter functions and another contractor for service delivery. The exceptions are TriMet in Portland, which continues to have two contracts with the same company, and KCATA in Kansas City, which opts to retain a call and control center contractor despite having one provider contractor serving 99% of the trips. Some of the same advantages that pertain to in-house call and control centers are the same as when these functions are performed by an operational broker or call and control center man- ager. Foremost, these include (1) consistency of booking practices and handling ETA calls and (2) objectivity in controlling the balance between service quality and cost efficiency. It is true that control over this objective is one step removed. However, it still can be somewhat con- trolled by the transit agency through performance standards, and incentives and liquidated damages that are performance based (this is discussed in more detail in the contracting sec- tion later in this report). And as also mentioned previously, an advantage that comes with a

108 aDa paratransit Service Models centralized approach is the ability to shift work among contractors mid-contract and based on performance, especially where the service providers are not assigned to exclusive zones and where the transit agency provides some or all of the vehicles. Brokerage models were much more popular among larger systems in the infancy of the para- transit industry because transit agencies did not have the knowledge base necessary to manage the system and contract with and monitor service providers. As more and more experience has been gained in these areas, other service models have become more popular. A notable excep- tion has been ACCESS in Pittsburgh (an administrative brokerage established for a coordinated system and not just transit agency-sponsored paratransit) whose service model has remained fairly intact since 1979. Besides the coordinated characteristic of ACCESS, the broker from the beginning has entered into 1-year contracts with its service providers. This is an unusual prac- tice that has worked there all these years and which has led to frequent service adjustments, as needed, not to mention keeping costs competitive. While the industry has matured, and with transit agencies having gained much more experi- ence retaining and monitoring the “right” service provider contractors and understanding the benefits of multiple contractors, there has been less of a calling for brokerages and more of an interest in establishing service models that involve a call and control center contractor, especially for transit agencies that wish to have a third-party undertake call and control center functions. This also led to the establishment of call and control center manager service models in loca- tions such as Broward County, Denver, and New York City and, for quite some time, Portland and Seattle as well. The MBTA in Boston was also in the midst of transitioning to a call and control center con- tractor service model, while Pace has instituted a service model in Chicago with centralized scheduling performed by a third-party entity. In the survey and via follow-up discussions, MBTA staff presented some fairly compelling reasons why they decided to change their decentralized service model to a call and control center contractor, even as the quality of service under the decentralized model was top-notch, as confirmed by performance statistics and by a customer satisfaction survey. Certainly, a major reason was to reduce cost. This was seen as coming from: • Productivity improvements via more optimal run structures that would be in the best inter- est of the MBTA versus the service provider, and with the change, taking direct financial advantage of the improved productivity by switching the way in which the service provider contractors are paid—switching from a per-trip rate to a per-RVH rate. • Increased productivity from eliminating duplicative service in the core area. • Taking advantage of contractual latitude to reallocate work among carriers, reallocating more work to lower-priced service provider contractors (as long as they achieve stated service qual- ity standards). • More expansive and objective approach to using NDSPs, which would help reduce the overall cost per trip. • Operationalizing conditional eligibility, a task that would not be as effective if reservations remained with the three service provider contractors. • With the elimination of zones, there is nothing to preclude the MBTA from growing or importing additional service providers to stimulate competition. The MBTA also listed some compelling reasons for the change that were focused not on cost reduction but on customer service: • Customers would experience more consistent approach to call center functions. • Transfers for peripheral trips between zones would no longer be necessary (as zones would no longer be necessary).

Survey analysis Summary 109 • With mid-contract reallocation in part based on service performance, the service provider contractors would have a more compelling reason to focus on the quality of service delivery. • Disruption insurance would be enhanced; i.e., in the event that one of the service provider contractors defaulted, customers would still be calling the same central source (the call and control center manager) with trips then assigned to one of the other service providers or to a NDSP until the defaulting provider could be replaced. There was also a consensus among transit agencies that a call and control manager service model requires a strong transit agency staff and strong contractor. The relationship among the transit agency, the call and control center manager, and the service provider contractors is an interesting one. Unlike a brokerage, there is no direct contractual relationship between the call and control center manager and the service provider contractors, but they must work as a unit. Therefore, while a call and control center manager, especially one that performs dispatching, is controlling the service providers’ drivers, the call and control center manager cannot discipline the service provider contractors or their drivers; discipline must come from the transit agency. If the transit agency is not strong, there is not much the call and control center manager can do and service performance can fall apart. As mentioned later in this report, there are some contrac- tual elements, such as incentives and liquidated damages that are shared by the call and control center manager and the service provider contractors that help them work better as a unit. If the transit agency recognizes that it cannot fill this role, it may decide that a brokerage model may be a better option. As more transit agencies consider the use of NDSPs for ADA paratransit trips, there is one more application for a broker that makes sense—the NDSP broker. The best example of this is in New York City. Certainly, if only one or a few NDSPs are used, as in the case in Nashville, no third party is needed. However, if many independent drivers are used, or the circumstances require contracts with many different entities, it would be administratively simpler to have just one contract with a NDSP broker. Use of Multiple Service Provider Contractors One of the findings from the survey is that larger ADA paratransit systems tend to use con- tractors more than providing the service in-house and that larger systems tend to use more than one service provider contractor. In the survey, 18 of the 29 transit agencies responding indicated that they use multiple contractors. These agencies noted in their response that a service provider contractor has a better chance of succeeding if not spread too thin, and that by having multiple contractors, each can be provided with a more manageable package of work. This, though, must be weighed with whether there is a sufficient volume of work to divvy up. For transit agencies with multiple carriers, service areas are most often organized among providers based on a contracted volume of work as shown in Figure 12. Types of Service Models Using Multiple Service Provider Contractors Unzoned Service Provider Contractors that Perform No Call Center Functions This group includes systems with multiple carriers but where all call center functions are centralized either with the transit agency or with a call and control center manager. With that centralization, there is no real need for zones as a method of funneling customer calls to a certain contractor, as mentioned earlier. These transit agencies stated that the strategy of dividing up the work into packages that are not necessarily geographically constrained enables schedulers to schedule in as productive fashion as possible. In cases where the transit agency also provides

110 aDa paratransit Service Models the vehicles, the split of work serves as an initial split only, noting that the transit agency can (if contractually allowed) reallocate work based on service performance. • The RTD in Denver puts out for bid four packages of service, each equivalent to about 25% of the trips. Proposers may propose to perform no more than two packages. Currently the RTD has three service provider contractors, one with 50% of the service (two packages) and two with 25% each. • The East Bay Paratransit Consortium in Oakland has had three service provider contractors (MV, First Transit, A Paratransit), each with a 33% split. • For its METROLift service in Houston, METRO contracts with two service provider contrac- tors, with one (First Transit) operating mid-sized vehicles (owned by Metro), and one (Yellow Cab) operating accessible minivans owned by Yellow Cab and owner-operator drivers. Before the advent of accessible minivans, METRO had First Transit operating accessible vehicles and Yellow Cab operating sedans. However, a few years ago, the sedan fleet was a replaced with an accessible minivan fleet. • In Austin, Capital Metro retains three service provider contractors for its MetroAccess ser- vice, all of which may be scheduled to serve trips anywhere in the same service area. The largest carrier utilizes vehicles provided by Capital Metro and operates out of a Capital Metro facility and its own facility. The other two smaller service providers each has its own facility and vehicles. Unzoned Service Provider Contractors that Perform Dispatch From the survey, there is only one transit agency, Broward County Transit in Fort Lauderdale, which also uses unconstrained biddable service packages. In this case, it is two service packages evenly split at 50%/50%, but that vests the dispatching with the service provider contractors. All other call center functions are carried out by BCT’s call center contractor, as previously discussed. Zoned Service Provider Contractors that Perform All Call Center Functions Different approaches to this are exemplified by the following locations where the service pro- vider contractors also perform all or some of the call center functions and are assigned to zones or regions. The concept that these multiple contractors are assigned to a zone and also perform all or some of the call center functions is not a coincidence; one of the major functions of a zone, N = 17. Figure 12. Multiple service provider systems’ service area organization.

Survey analysis Summary 111 besides enabling competition and allowing for an enhanced chance of succeeding with a more manageable package of work, is to funnel customer calls to a certain contractor based on where the customer lives. • For years, The RIDE in Boston employed a four-zone system with one service provider (per- forming call center and service delivery functions) assigned to each zone and with the zones overlapping in a core area that included Boston and later some nearby cities. This was in part based on the size of the service area, which was considered too large for one service provider and, initially, a desire to use local carriers. With the default of one of the carriers, and coinciding with the transfer of a portion of another zone to a neighboring system, two of the zones were collapsed into one, leaving three zones with a roughly equal portion of trips. This service model remained in effect for ten years. One of the MBTA’s goals was to stimulate competition by hav- ing a few manageable and equal-sized packages of work, while minimizing as much as possible the average length of intra-zone trips. Customers book their trips through their home zone contractor. Customers living in the overlapping core area call their destination carrier if making an inter-zone trip, and may call any of the three service provider contractors if making an intra- core area trip. Peripheral inter-zone trips do not require a transfer if going right across a bound- ary into an adjoining town; inter-zone trips deeper into the destination zone require a transfer. • MCTS in Milwaukee splits its area into two zones, North and South, roughly reflecting a 55%/45% split in the volume of work. As is the case with many zoned systems, custom- ers book their trips with their home zone carrier, with that contractor performing all major functions. • The Phoenix region has a complex service area, with a turnkey contract with MV Transpor- tation covering the city of Phoenix, and multiple turnkey contracts serving Valley Metro’s region (outside the city of Phoenix). In the Valley Metro area, there are four regions. Two of these regions (East Valley and Northwest Valley) are served under contracts with one service provider contractor (Total Transit); this same contractor also has a separate contract to serve inter-region trips. In the other two regions (Glendale and Peoria), ADA paratransit trips are served by municipal contractors providing public dial-a-ride services. Another interesting element of the Valley Metro area is that Total Transit provides mostly non-dedicated service. • Access Services in Los Angeles has employed a six-zone system for years, with a contractor assigned to each zone that is responsible for call center and service delivery functions. Four of the zones are located in the Los Angeles basin; inter-zone trips between these zones do not require a transfer. The other two zones are geographically removed from the Los Angeles basin and do require a transfer. With those basic differences, there are numerous conditions that make Access Services fairly unique. After a customer calls to book an inter-zonal round trip, the customer is transferred to the other zone’s contractor for intake of the return trip request. Calls do not always need to be transferred though, because two pairs of contractors have combined their call centers. One of these pairs of contractors (San Gabriel Transit and Cali- fornia Transit, owned by the same entity) have also unified their scheduling function, enabling the schedulers to place trips on each other’s vehicles. The other pair of contractors in the Los Angeles Basin (Global Paratransit and MV Transportation) are also collaborating. Global contracts with MV to perform its reservations function, while MV contracts with Global to serve some its trips, particularly in the during weekday evenings and weekends Global also contracts with San Gabriel Transit to access some its taxi subcontractors. For one of the two outlying regions in the “North County” area, the call center and opera- tional functions of the service provider contractor, Keolis, are split to take advantage of a call center that Keolis had already established for other contracts. The contractor in the other out- lying area, Santa Clarita, serves ADA paratransit trips on its general public dial-a-ride system. A particularly noteworthy aspect of Access Services’ system is that the four service provider contractors serving the Los Angeles basin zones each utilize taxi subcontractors to serve about 50% of the trips for which they are responsible.

112 aDa paratransit Service Models • ACCESS in Pittsburgh has also employed zones since its inception. The original concept was to take advantage of local carriers as much as possible. With that in mind, the ACCESS bro- ker contracted with a few local nonprofit agency operators, designing the operators’ zones to cover the areas in which their clients lived. The rest of Allegheny County was carved up to reflect the general service area of larger local cab services. While some zones have been modi- fied and combined over the years, this general structure has been in place since 1979. Zoned Service Provider Contractors that Perform Some Call Center Functions • In New York City, NYCT split up the five boroughs into 17 Access-a-Ride service zones, and has contracted with 13 service provider contractors. The original concept upon which the system is based is to ensure a sufficient amount of coverage throughout the service area and to minimize deadheading to the extent possible; this was done by requiring service provider con- tractors to operate out of facilities located in the zones to which they are assigned. The service provider contractors all do their own dispatching, noting that the call center manager contrac- tor does perform some daily monitoring of how each run is performing and will notify a car- rier’s dispatcher if it appears that any shortcomings of a particular run are not being addressed. • In Chicago, Pace employed a three-zone system for years, with a service provider contractor assigned to each zone—north, central, and south, and with a fourth service provider serv- ing long trips. The three-zone system, however, provided great inefficiencies with the high number of inter-zone trips, especially those going into downtown Chicago in the central zone. There was much deadheading coming back from downtown in the morning and much deadheading going into downtown in the afternoon to pick up the return trips home. Pace had considered a call and control center contractor model but ultimately chose to implement a hybrid model instead. Under the current model, customers, as they did before, book their trips through their home zone carrier. These trips are then scheduled by a third-party contrac- tor, who also serves as a call and control center service contractor for some of Pace’s other sub- urban services. As a result of the scheduling a trip might wind up being carried by the home zone carrier—or not. Dispatching is the responsibility of three service provider contractors, while the scheduling contractor also handles the customers’ same-day issues. • In Salt Lake City, the UTA has a three-region system, with its two contractors serving the northern and southern region and the UTA serving the central region. In the remote regions, dispatching is the responsibility of the service provider contractors, while the UTA directly performs the rest of the call center functions for three areas. • NJ TRANSIT established six zones covering most of New Jersey. A service provider contractor has been assigned to each zone; First Transit currently is the contractor in five of the six zones and a local carrier is the contractor in the sixth zone. NJ TRANSIT’s reasoning for vesting dispatching with the contractors is because of the vast size of each region and each region’s unique characteristics; the reasoning was that dispatchers located within the region would have more familiarity with the region than if they were located with the reservation agents and service monitors at NJ TRANSIT’s headquarters in Newark. For the same reason, the sched- uling function, in a somewhat uncommon arrangement, is shared between NJ TRANSIT’s scheduling staff and each service provider contractor’s schedulers after each daily (live) sched- ule is created, noting that NJ TRANSIT’s schedulers are responsible for the run structures and the scheduling of subscription trips onto the master template schedules. Zoned Service Provider Contractors that Perform No Call Center Functions With call center functions completely centralized, there is no real reason for zones as a method of funneling customer calls to a certain contractor. However, in the following cases, the service area is so large that regional assignments make sense.

Survey analysis Summary 113 • SEPTA in Philadelphia has in effect five zones: one for Philadelphia and four in the surround- ing counties. Within Philadelphia, SEPTA has retained four service provider contractors to provide service anywhere within the city, while in the suburban counties, SEPTA has retained one service provider contractor for each county. There are a total of six companies involved, with two companies operating in one suburban county and in Philadelphia. • In Washington, DC, WMATA has a two-region system. The first region incorporates the District of Columbia and two counties in Maryland. WMATA contracts with two service pro- viders (First Transit and Transdev) to serve this region. The second region covers Northern Virginia, across the Potomac River. For this region, WMATA contracts with a local provider. • In Kansas City, KCATA contracts with both Transdev and First Transit. RMS, the call and control center manager, schedules onto Transdev runs all trips on the Missouri side, intercity trips on weekdays until 7:00 p.m., and all trips on weekends. This constitutes about 99% of the trips. First Transit serves trips on the eastern side of the service area and Independence, Missouri, where it also provides fixed-route and ADA paratransit service. Advantages and Disadvantages of Contracting with Multiple Service Provider Contractors The underlying reasons for using multiple service provider contractors are stated as follows by some of the survey respondents in follow-up conversations. • “Where there is sufficient ridership or a broad service area, dividing up the work into man- ageable pieces attracts more contractors to the procurement, including smaller local and regional carriers who would not otherwise be able to submit a proposal. A competitive multiple-carrier marketplace yields more competitive prices and ultimately provides superior cost containment.” • “We find that the contractors in our system compete with each to see who can provide better service quality. This ‘Keeping up with the Joneses’ competition results in an overall higher level of service quality. With just one contractor, our experience has been that they try to attain the performance goals, but no further.” • “Under a zone-less system where the transit agency owns the vehicles and controls the vehi- cles through scheduling and dispatch, a multiple-carrier service model enables us to adjust work among our service providers up to a point without having to renegotiate the rates. This enables us to ‘right-size’ the work assigned to each of ours carriers. “Also, it would appear that rewarding higher-performing contractors with more business and penalizing lower-performing service providers by taking away business is more effective than the performance-based bonuses and liquidated damages, although these are still impor- tant to emphasize our expectations. And, at a reduced volume of work, we have actually seen the poorly performing service provider do better.” • “Having multiple service providers helps with risk management. We have one carrier who was so poorly performing that they were essentially in default, and had three others available to come to our rescue. This allowed us some breathing room for replacing that contractor.” One of the disadvantages of having multiple service contractors is that the transit agency is essentially paying for duplicative labor. This is certainly the case at the higher levels of manage- ment and mid-level management as well. For each service provider contractor, there will likely be a general manager, possibly an assistant general manager or operations manager, a window dispatch (or pullout) manager or supervisor (per shift and per facility), a safety/training man- ager and training staff, a set of road supervisors, a maintenance manager and lead mechanics, an IT manager, a human resources manager, and a financial manager and administrative support for accounting, payroll, trip reconciliation, reporting, and invoicing (although some of these functions are sometimes provided by corporation staff if the provider is a national or regional

114 aDa paratransit Service Models company). Moreover, if each service provider contractor also provides all or some of the call center functions, there can also be a call center manager, a reservations supervisor, a lead sched- uler, a lead dispatcher, and/or a customer service manager. Lastly, there may also be economies of scale with front-line employees, especially booking agents and operators. The transit agency would also be paying for additional operating facilities, large maintenance equipment, security systems, and other considerations for each service provider contractor, whether the transit agency owns or leases the facilities, or not. In addition, the transit agency would likely devote a contractor monitor to each contractor. From the survey, the stated benefits of using multiple carriers outweigh the shortcomings, especially if the number of contractors can be “right-sized,” so that the transit agency does not pay for more duplication of labor then necessary. At the same time, it is also true that in some locations, the sheer size of the service area and the particular geography limit the options. This is supported in Figure 13. Of the 29 transit agencies that responded to the survey, there were 18 locations that used multiple service provider contractors operating dedicated service. As shown, 12 (67%) of the 18 have just two or three carriers. Of the other locations that had four or more carriers, all were serving high volumes of trips or were serving very large areas. Interest- ingly, Boston and Washington, DC, both in the top five largest systems by ridership and both with large regional service area, have been able to serve their ridership with only three dedicated carriers, noting that Boston had previously used four contractors. Zoned versus Unzoned Service Areas, Centralized Booking, and Transfers Other than right-sizing the number of carriers, another aspect of the service model design is whether or not zones are utilized as a way to organize the supply of service. As shown in Figure 14, 12 (63%) of the 19 who use multiple service provider contractors do have a zone system, while the other seven do not. This seems to indicate that a zoned structure is favored. However, a closer examination can attribute half of the 12 to the split of responsibilities and, in particular, to locations where the service provider contractors are performing reservations and where customers are instructed to call their home contractor based on where they live. As shown in Figure 15, six of the 12 fall into this category. In five of the other six cases, zone assignments make sense given the expansive No. of Service Provider Contractors – Dedicated Service Number of Transit Agencies Locations 0 1 Nashville 1 9 Ann Arbor, Atlanta, Columbus, Dallas,* Las Vegas, Orange County,* Orlando, Pierce County,** Portland 2 6 Broward County, Houston, Kansas City, Milwaukee, Salt Lake City, Seattle 3 5 Austin, Boston, Denver, Oakland/East Bay, Washington, DC 4 2 Chicago, Phoenix* 6 4 New Jersey, Los Angeles, Philadelphia, Pittsburgh 13 1 New York City * Prime contractor also uses a subcontractor. ** Transit agency also operates dedicated service. Figure 13. Number of service provider contractors by location.

Survey analysis Summary 115 service areas and particular geography. In the case of Milwaukee, a convenient north and south split was employed. Looking back at Figure 14 and the seven locations where zones are not used, all seven have centralized booking. The MBTA in Boston will soon join the zone-less locations when the transi- tion to a CCC service model has been completed. There does seem to be a trend in the industry toward the centralization of booking—where multiple service providers are used—for three reasons: • Giving the transit agency more flexibility to adjust the number of carriers and/or adjust the volume that each carrier serves • Giving the transit agency—or the broker or call center manager—more control over the operationalization of conditional eligibility • Reducing costs associated with duplicative call centers. Single Contractor Service Models Types of Service Models Using One Service Provider Contractor As shown in Figure 13, nine of the 29 transit agencies responding to the survey indicated that they contract with one service provider contractor. Many of the nine have some nuances that distinguish them from each other. For example: • In Columbus and Orlando, COTA and LYNX provided ADA paratransit services via classic turnkey single service provider contractor service models. • In Portland, TriMet contracts with the same firm (First Transit) to perform call center func- tions and operations (under two contracts), and while that could change in the future, it essentially functions as a turnkey contract. • Similarly, in San Francisco, SFMTA’s broker (Transdev) also is the sole dedicated service provider. Zoned or Unzoned Service Area No. of Agencies with Multiple Contractors Locations Yes 12 Boston, Chicago, Kansas City, Los Angeles, New Jersey, Milwaukee, New York, Philadelphia, Phoenix, Pittsburgh, Salt Lake City, Washington, DC No 7 Austin, Broward County, Denver, Houston, Oakland/East Bay, San Francisco, Seattle Figure 14. Zoned versus unzoned service areas. Centralized or Decentralized Booking No. of Agencies with Multiple Contractors Locations Centralized Booking 6 Kansas City, New Jersey, New York, Philadelphia, Salt Lake City, Washington, DC Decentralized Booking 6 Boston, Chicago, Los Angeles, Milwaukee, Phoenix, Pittsburgh Total 12 Figure 15. Number of zoned locations by centralized versus decentralized booking.

116 aDa paratransit Service Models • In Ann Arbor, which has a combined ADA/senior paratransit service and where same-day trip requests are permitted for both sets of customers, AAATA retains one contractor to perform all trips. While AAATA intakes reservations and scheduling for all advance-reservation trips, the service provider contractor intakes same-day reservations and performs dispatching and service delivery for all trips. • In Atlanta, MARTA retained MV Transportation to perform all functions except reservations. • RTC in Las Vegas has a service model similar to the one in Atlanta, with RTC employees per- forming reservations and RTC’s sole contractor, Transdev, performing all other call center functions and service delivery. RTC management also stated that it keeps the reservations function in-house for enhanced control, adding that they have operationalized conditional eligibility in Las Vegas and felt that RTC would adhere more closely to those customers’ limits than would a service provider contractor that is paid based on the amount of service supplied. • In Dallas, DART changed from an in-house call center with one service provider to a single turnkey service provider (MV Transportation, which performs all call center functions, and serves approximately 40% of the trips). MV has one subcontractor, Yellow Cab, which serves approximately 60% of the trips through a mixture of accessible dedicated and non-dedicated vehicles. • In Orange County, OCTA has a similar arrangement with MV Transportation performing all call center functions, and operating service on weekdays before 6:00 p.m. and on Saturdays before 2:00 p.m. MV’s taxi subcontractor operates all weekday service after 6:00 p.m.; Satur- day service after 2:00 p.m.; and all Sunday service. • In Pierce County, Pierce Transit has a fairly uncommon arrangement where it performs all call center functions and a portion of the operations in a zoneless service area. Pierce Transit drivers serve about 20% of the trips, while its sole service provider contractor serves the rest of the trips (that are not assigned to the NDSP). In the case of the portion of the service oper- ated in-house, the runs consist of straight runs during the day on weekdays only, with heavy concentrations of standing orders. Advantages and Disadvantages of Using One Service Provider Contractor The above transit agencies that have retained one, or a primary, service provider contractor identified several benefits: • Transit agencies can benefit from, and play to the strengths of, the larger national firms, as they have more experience in performing call center functions and operating paratransit, compared to the alternative of operating service in-house. Using a contractor to operate ser- vice is also less expensive than operating service in-house. • One agency mentioned that this handful of national firms can be more nimble in responding to fluctuations in demand (i.e., that require more staff and/or vehicles) than can the transit agency itself. • Another agency mentioned that national companies have kept up with new and emerging tech- nologies, such as dynamic (same-day) scheduling and providing customers with real-time infor- mation about the location and ETA of the vehicle to which a customer’s trip has been assigned. • Compared to a multi-carrier service model, the transit agencies that have a single contractor also pointed to the benefits of a single point of contact and the economies of scale, lack of duplication, and reduced labor needed for oversight. For large systems though, the disadvantages seem to outweigh the benefits. • Limiting the number of service provider contractors to one may effectively exclude local and regional providers from consideration. This limits competition and tends to remove factors that lower proposed prices.

Survey analysis Summary 117 • Putting “all your eggs in one basket” can be daunting for some agencies. Without a fallback position, the agency has little leverage, other than punitive damages, if service quality does not meet standards, or if the level of service required exceeds the negotiating window. In the case of liquidated damages, some agencies confided that they are sometimes reluctant to penalize their one contractor for fear that reduced payment will cause further degradations in service. If a renegotiation is warranted, a sole contractor “can hold the transit agency over a barrel.” A transit agency’s recourse is to wait until the contract is over or to undertake an emergency procurement, which has its own challenges. • If the contractor defaults for any reason, either on a permanent or temporary basis, it will be harder for a transit agency to replace. In such an event, the transit agency would probably have to perform an emergency procurement, which rarely generates competitive pricing. If the service provider also performs call center functions, there are additional concerns. In particular, the contractor, when considering run structures, driver assignments, trip eligibility limits, and use of non-dedicated service providers, will focus first on what is best for the contrac- tor versus what is optimal for the transit agency. In short, while good partnerships with a single contractor are certainly possible, the fear of such partnerships deteriorating without a good recourse and the lack of competitive pressures being limited to every procurement cycle have influenced the majority of larger firms to establish multiple service provider models. Integrated Non-Dedicated Service Providers Systems Using Non-Dedicated Service Providers for ADA Paratransit As shown in Figure 16, 13 of the 29 survey participants use non-dedicated service providers (NDSPs) in an integrated fashion, meaning schedulers and dispatchers respectively schedule or assign ADA paratransit trips to NDSPs. Access Services in Los Angeles is a striking example of an ADA paratransit system that was largely built around non-dedicated taxi service. In the 1990s, Access Services developed a Location Percentage of Trips Served by NDSPs Phoenix (excluding city of Phoenix) Near 100% Dallas 70% Los Angeles 39% New York City 30% Orange County 23% Kansas City 17% Nashville 15% Houston 11% Denver 10% Seattle 9% Washington, DC 6% Austin 3% Boston 1% Figure 16. Integration of non-dedicated service providers.

118 aDa paratransit Service Models certification program in which any taxi driver on a voluntary basis could opt to participate in a certification program where they would receive the same level of training provided to contractor employee drivers, agree to participate in drug and alcohol testing, and undergo similar back- ground checks. Successful graduates became certified to serve ADA paratransit trips. In addition, the owners of the taxicabs driven by certified drivers agreed to insure the vehicles up to Access Services requirements. (Such requirements are also in place in New York City for the black car operators utilized by the NDSP broker). Currently, Access Services has six zoned contractors. Four of the six contractors serve approximately half of their trips with taxis operating in a semi-dedicated or non-dedicated mode, noting that many of the taxi subcontractors are owned by the contractor. Here, semi- dedicated mode refers to the practice of clustering unscheduled trips (i.e., that could not be scheduled onto a dedicated vehicle run) onto mini-runs of 3–4 hours operated by a taxicab. In addition, in each case, the vehicles are equipped with the same in-vehicle communication equipment. In Los Angeles, the agencies do not financially benefit from daily increases or decreases in NDSP usage because both agencies pay their contractors a rate that incorporates a fixed com- ponent covering fixed costs and a per-trip rate covering variable costs. Both agencies state that the contractors choosing to use taxis at a high level ultimately maximizes the contractors’ profit; the subtext being that the emphasis on minimizing expenses from the contractors results in a higher profit for them, and maximizes service efficiency for the agency. However, the agency does ultimately benefit from this practice financially at re-procurement time, where the contrac- tors using taxis at high levels can propose a competitive per-trip rate. Access Services, with one of the largest ADA riderships in the country and with one of the longest average trip lengths, has a comparatively low cost per trip because of the high use of taxis. The contractors in Los Angeles, who own the taxi subcontractors, also reported that the increase in use of taxis for ADA paratransit trips in part is due to the rising economy and the net loss of employee drivers to other, higher-paying jobs. Even with active recruiting efforts, these contractors have been recently experiencing a net loss of employee drivers, and hence dedicated service capacity. Therefore, more and more trips are assigned to their taxi subcontractors. They also note that while there has been a significant loss of taxi drivers as a result of TNC infusion, with taxi demand down, more and more taxi drivers are opting to work for a TNC or are find- ing other jobs. Of the taxi drivers who remain, though, more and more are interested in being certified to serve ADA paratransit trips. New York City Transit’s Access-a-Ride program also has a unique approach to providing ADA paratransit through NDSPs. In New York, the NYCT has retained two NDSP brokers. The software used by NYCT’s call center manager to schedule service is used in batch mode to schedule onto the dedicated vehicle runs operated by 13 service provider contractors. Unsched- uled trips are automatically assigned to the NDSP broker for subsequent assignment to specific NDSP operators. In cases where dispatching is centralized, NDSPs are typically paid on a negotiated distance- based rate, if not the taxi-meter rate. In some cases, payment is based not on the actual distance travelled but on the distance predetermined by the map underlying the scheduling software (or an equivalent available on the Internet), so that when each trip is assigned, the payment for the trip is predetermined. In this way, the transit agency, call and control center, or broker can prepare the draft monthly invoice based on information it already has. As shown in Figure 17, peak overflow trips are the most common type of trip assigned to taxis serving as a non-dedicated service resource, followed by trips that do not fit into dedicated runs. The low ranking of short trips makes sense, as the removal of short trips from dedicated

Survey analysis Summary 119 service can reduce productivity and actually have a negative impact on overall cost per trip. The lowest-ranking use of NDSPs, to handle re-emerging no-show customers, was surprising. Among the four “other” answers, one explained that taxis are used for all trips, while a second agency answered that taxis are used for all kinds of trips. The other two agencies replied that taxis are used in the context of alternative taxi subsidy programs. Also, while it is not required that NDSPs supply accessible taxicabs if the trips of customers who require accessible vehicles are accommodated by the dedicated service, having a NDSP resource that supplies accessible vehicles provides more flexibility for both schedulers and dispatchers, as discussed in Chapter 3. But, as it turns out, the majority of surveyed agencies that have access to taxis for trips do so with companies that supply accessible taxicabs, as shown in Figure 18. As shown in Figure 19, 56% of the transit agencies responding to the survey involve the use of taxis through direct contracts, with the transit agency, broker, or call center manager directly contracting with taxi companies, whereas 46% of the transit systems that involve the use of NDSPs do so indirectly; that is, their dedicated service provider contractors assign trips they cannot or choose not to serve directly to taxi subcontractors. N = 13. Figure 17. Taxi-assigned trip types. N = 12. Figure 18. Accessible taxicab availability.

120 aDa paratransit Service Models Transit Agency Provision of Assets Types of Assets Provided From the survey, the transit agencies that provide assets for their brokers, call center managers, and/or service provider contractors are identified in Figure 20. Based on the response, 22 transit agencies supply the paratransit software system, while 21 provide some or all of the vehicles. Advantages and Disadvantages of Providing Assets Most of the survey respondents did think it a good idea for the transit agency to invest in these assets, and especially paratransit software and vehicles, if feasible. Some of the transit agencies that do provide some or all of the assets made the following statements in follow-up discussions: • “Our provision of the operations facility, vehicles, software and other assets (e.g., tablets, radio system, telephone system, IVR, etc.) enables us to ‘swap out’ contractors when contracts end (or are not extended) or when a contract ends prematurely.” • “We aren’t very good about acquiring real estate to leasing office supply. In our recent call center manager RFP, we specified that proposers should find the space and provide the tele- phone and radio systems, and that we have the flexibility to take over the facility lease or these fully amortized assets at the end of a contract.” • “Since we own the vehicles and control dispatching (through a call center contractor), we reallocate work based on performance during the contract period, if beneficial.” • “Supplying the majority of vehicles gives us control over the type of vehicles we want in our system, and the retirement schedule. Our contractors make up the difference by supplying the extra vehicles needed.” On the other hand, an in-house operation, or just providing the vehicles, can require an enor- mous capital outlay, not only for the fleet but also for the drivers, staff, and facility, and a transit agency may not have this level of funds available for such an investment. Alternative Service for ADA Paratransit Customers Among the survey participants, the transit agencies in Chicago, Denver, Houston, Kansas City, Las Vegas, Seattle, New York City, and Washington, DC, all have taxi subsidy programs optionally available to their ADA paratransit customers. All indicated that one of the goals of the alternative was to reduce overall paratransit costs. N = 11. Figure 19. Contracting entity for taxis.

Survey analysis Summary 121 For a transit agency to determine whether or not savings have been achieved, it is essential that some tracking take place to better understand the level of diverted trips versus the level of new trips for each customer participating in the subsidy program. In many such programs, transit agencies look at the “before-and-after” level of ADA paratransit ridership to estimate those levels. Customer surveys, while less accurate, have also been used. In Denver, calls requesting Access-a-Cab service, the Denver RTD’s alternative service, come to the same call center as the ADA paratransit trip requests for Access-a-Ride. The Access-a-Cab requests are then forwarded to the taxicab company requested by the customer. This also helps in terms of controlling the ridership on the program if there is a limited budget. But with taking the requests centrally, RTD can get a better sense of the before and after picture of ridership on Access-a-Ride versus Access- a-Cab to determine whether there is a savings. New York City Transit has an even more accurate approach for its alternative taxi subsidy pro- gram. After an Access-a-Ride (same ADA paratransit service name as Denver) call taker records an ADA paratransit trip request for one of its Access-a-Ride customers, the call taker, under certain guidelines, may ask whether or not the customer might be interested in taking a taxi instead, with the understanding that the customer is under no obligation to accept the offer, and it will not impact the customer’s ADA paratransit eligibility in any way. If the customer does accept the offer, the customer receives an authorization number from the call taker, arranges for the taxi trip inde- pendently, pays the full meter fare, and submits through the mail the receipt with the authorization Type of Assets Number of Locations Locations Operations/Maintenance. Facilities 14 Atlanta, Austin, Columbus, Dallas, Houston, Las Vegas, Nashville, NJ, Orange County, Orlando, Pierce County, Portland, Salt Lake City, Seattle Vehicles – Some or All 21 Atlanta, Austin, Boston, Broward Co, Chicago, Columbus, DC, Denver, Houston, Los Angeles, Las Vegas, Nashville, NJ, Orange County, Orlando, Philadelphia, Pierce County, Portland, Salt Lake City, Seattle, San Francisco In-Vehicle MDT/AVL or tablets 9 Broward Co, Columbus, Las Vegas, New Jersey, Palm Beach, Philadelphia, Portland, Salt Lake City, Seattle Paratransit Software 22 Ann Arbor, Atlanta, Austin, Broward Co, Chicago, Columbus, Dallas, DC, Denver, Houston, Las Vegas, Milwaukee, Nashville, New Jersey, Orange County, Orlando, Philadelphia, Pierce County, Portland, Salt Lake City, Seattle, San Francisco Radio system 15 Atlanta, Austin, Columbus, DC, Denver, Houston, Las Vegas, Nashville, Orange County, Philadelphia, Pierce County, Portland, Salt Lake City, Seattle, San Francisco Telephone System 20 Ann Arbor, Atlanta, Austin, Broward Co, Columbus, DC, Denver, Houston, Kansas City, Los Angeles, Las Vegas, Nashville, New Jersey, Orange County, Philadelphia, Pierce County, Portland, Salt Lake City, Seattle, San Francisco IVR 18 Atlanta, Austin, Broward Co, Chicago, Columbus, DC, Denver, Houston, Las Vegas, Nashville, Orange County, Orlando, Philadelphia, Pierce County, Portland, Salt Lake City, Seattle, San Francisco Figure 20. Provision of assets to contractors.

122 aDa paratransit Service Models number. Once the claim has been processed, NYCT sends a check for that amount less a base fare to the customer. NYCT reports that ADA paratransit customers who are posed with this offer accept the offer about 5% of the time. In addition, call takers receiving same-day calls from customers who have been stranded for a while have permission to extend this offer to these customers as well. One of the benefits of this practice is that NYCT knows that these trips would have been taken on Access-a-Ride and hence has more accurate information with which it can estimate savings. Most of these alternative services work the same way: the customer pays for a base fare that enables a trip up to a certain length, with the remainder of that fare subsidized by the transit agency. For longer trips, the customer is responsible for the overage. Historically, these subsidy programs have used paper vouchers, either sold to eligible par- ticipants at a discount (user-side subsidy), or distributed to the taxi companies (supply-side subsidy); however, more and more of these programs have gone cashless with the use of “ghost” cards. Ghost cards are debit cards distributed by financial institutions in partnership with the transit agencies that can only be used to pay for trips on the alternative service. For such cards, the subsidy provided by the transit agency is put on each customer’s card. Alternatively, the NDSP provider can bill the transit agency for its share of the trip fare. The MBTA in Boston and DART in Dallas are currently piloting a TNC-based alternative ser- vice, while WMATA in Washington, DC, Broward County Transit in Fort Lauderdale, and the RTC in Las Vegas are planning similar alternative services. One of the unique elements of both the Boston and Broward County pilot programs is the capping of trips per month based on ADA paratransit ridership. So, for example, if a customer barely uses The RIDE, the MBTA’s ADA paratransit service, the MBTA has capped that customers use of the pilot alternative service to only two trips per month, whereas a customer who use The RIDE daily will have a trip cap of 10, 20, 30, or 40, depending on the extent to which the use The RIDE. Also of note in Boston, in order to accommodate pilot participants who require a wheelchair-accessible van, one of the TNCs sub- contracts with a chair car company, while the other TNC utilizes operators of accessible taxicabs. Methods of Contractor Payment Prior to a discussion of how different types of contractors are paid, Figure 21 presents a “quick reference” on how the transit agencies from the survey use contractors. Brokers—Payment Structures Pittsburgh’s ACCESS is the only administrative broker from the survey. The brokerage man- agement firm (Transdev) is paid based on a proposed budget that is annually renegotiated based on a cost-plus basis. All carrier service costs related to ADA paratransit and senior trips are passed through to the Port Authority, noting that carrier per-trip rates are competitively sought annually. Human service agency costs are invoiced to the sponsoring agency, either based on the zone-based fare of each trip sponsored, or an average cost-per-trip rate that is based on an analysis of the trip costs for the very large ridership agencies, like Medicaid. Annual renegotiations of the administra- tive costs might be based on unanticipated fluctuation in demand based on a cost-plus basis. From the survey, the one operational broker is Oakland/East Bay. Transdev, the broker, is paid for fixed costs separately from the variable costs, which are split into equal monthly pay- ments, with incentives and liquidated damages based on on-time performance, excessively long trips, and productivity. Historically, the most common payment structure for operational (full- service) brokers was a per-trip rate. While administratively simple, this rate structure actually backfired with many brokerage-based systems, as the brokerage firm had a built-in incentive to maximize its profit by minimizing its out-of-pocket costs to its contracted carriers. This resulted

Survey analysis Summary 123 in the infusion of low-cost carriers that, in turn, resulted in low-quality service. This is one rea- son why brokerage models have fallen out of favor among many large systems in favor of call and control center manager service models—the basic difference being that the transit agency controls who the service provider contractors are, not the call and control center manager. Payment Structures for Call (and Control) Center Contractors In the industry, there are some call centers that are paid by the call; however, only one of the transit agencies who responded to the survey pays its call and control center manager in this manner. A wide array of methods of payment for call centers exists, from fixed monthly pay- ments, to variable payment structures based on call volumes and staffing needs. Location Admin Broker Ops Broker Call Center Manager Service Provider Contractors Reserv. Sched. Dispatch ETA Ops Ann Arbor (1) Atlanta Austin Boston Broward County R/S/ETA Chicago S/ETA Columbus Dallas Denver All 4 Houston Kansas City S/D/ETA Las Vegas Los Angeles Milwaukee New Jersey (2) New York City R/S/ETA Oakland/East Bay All 4 Orlando Orange County Philadelphia Phoenix Pierce County Pittsburgh Portland (3) Salt Lake City San Francisco (4) Seattle All 4 Washington, DC All 4 R=Reservations S=Scheduling D=Dispatch ETA=Estimated Time of Arrival Calls (1) For same-day requests only. (2) Shared with agency staff. (3) Call center manager and service provider contracts are all with the same company. (4) Broker and service provider contracts are with the same firm. Figure 21. Transit agency use of contractors for ADA paratransit.

124 aDa paratransit Service Models The seven locations that use a call (and control) contractor are in Broward County, Chicago, Denver, Kansas City, New York City, Seattle, and Washington, DC. • In Broward County, the call center contractor is responsible for reservations, scheduling, and handling customers’ same-day issues but not dispatching, which is the responsibility of the service providers. BCT pays its call center contractor on a flat per-service-hour rate. • In Chicago, the call and control center contractor (SCR) is responsible for scheduling and han- dling customers’ same-day issues. Reservations and dispatching are the responsibility of the ser- vice providers. Pace pays SCR with hourly rates per function for approved staffing structure, with fixed costs paid for separately from variable costs and are split into equal monthly payments. • In Denver, the call and control center contractor (First Transit) is responsible for all four call center functions. The Denver RTD pays First Transit via fixed costs payments. • In Kansas City, the call and control center contractor (RMS) is responsible for scheduling, dispatching, and handling customers’ same-day issues, while KCATA directly performs reser- vations. KCATA pays RMS with fixed costs, separated from variable costs and split into equal monthly payments. • In New York City, the call center contractor (Global Contact Services) is responsible for res- ervations and handling customers’ same-day issues. NYCT pays GCS based on a cost-plus negotiated budget. • In Seattle, the call and control center contractor (First Transit) is responsible for all four call center functions. King County Metro pays First Transit based on vehicle service hours, fixed costs, and pass-through costs. In Washington, DC, the CCC contractor (MV Transportation) is responsible for all four call center functions. WMATA pays MV based on a per-call rate, and tracks the percentage of telephone response times that are less or more than 2 minutes. And in Boston, the MBTA will pay its new call and control center manager, Global Contact Services, based on the proposed and annually negotiated cost-plus fee. Payment Structures for Dedicated Service Contractors Where All or Most Call Center Functions are Centralized with Other Entities From the survey, this group includes: • Ann Arbor, Austin, Houston, Philadelphia, and Pierce County, where the transit agency is responsible for all call center functions. • Salt Lake City, where the transit agency is responsible for all call center functions except for dispatching trips scheduled to contractor runs. • Locations where there is an operational broker or call center manager; as identified above, these include Oakland/East Bay, Broward County, Chicago, Denver, Kansas City, New York City, Seattle, and Washington, DC. All of the service provider contractors that fall into this category are paid based on: • A per-hour rate, where the hour is defined as either a “gate-to-gate” service hour or revenue vehicle hour (first pick-up to last drop-off minus any scheduled breaks). From the survey, this is the type of rate used in Ann Arbor, Austin, Chicago, Denver, Houston, Kansas City, Oakland, Philadelphia, Pierce County, Salt Lake City, Seattle, and Washington, DC. • A split rate consisting of monthly fixed fee, which is based on that year’s fixed costs plus a vari- able rate per hour that covers variable costs. From the survey, this is the type of rate used in Ann Arbor, Austin, Broward County, Kansas City, and Salt Lake City. While the first of the two types of rates is common because of its simplicity, the second rate structure is not that much more complex, and provides some risk control for the contractor in the event there is a fluctuation of needed service, but not beyond the threshold, which triggers a renegotiation. In short, it is fairer.

Survey analysis Summary 125 Transit agencies from the survey that do use the second split rate have suggested that they receive more proposals with less risk cost built in. They also note that they use a procurement document, and sometimes a spreadsheet supplied to the proposers, that rolls up line item costs into either the fixed monthly rate or the variable rate in order to make an apples-to-apples comparison of proposers’ costs and rates that cover fixed costs and variable costs, respectively. Note, too, that only five of the survey respondents that fall into this category pay the carriers by a per-trip rate. These transit agencies have learned that, in comparison to the per-hour rate structures above, a per-trip rate typically does incorporate some risk-related cost. This is because most of a service provider contractors’ cost structure is hourly based. Converting from hourly costs to a per-trip rate in essence requires a proposer to estimate productivity. However, pro- ductivity can be significantly impacted by a great many factors that are beyond the contractors’ control (e.g., ridership changes that result from fare changes and/or other policy changes, the economy, agency controlled scheduling parameters, etc.). With a per-hour rate structure, however, transit agencies in this category also have come to understand that they have to keep a tight rein on the number of hours provided. Under a per- hour rate structure, standard operating procedure is to provide each contractor with an efficient run structure before each contractor’s driver pick. In addition, entities providing centralized dispatching must be the entities that authorize run changes, deletions, and/or additions. Payment Structures for Dedicated Service Contractors That Also Perform All or Most Call Center Functions (Including Scheduling) From the survey, the transit agencies included in this group can be broken down into those where the service provider contractor(s) provide all four primary call center functions and those that provide only a subset of these functions. Those in the first group that retain a single turnkey contractor to perform all four call cen ter functions include Boston, Columbus, Dallas, Los Angeles, Milwaukee, Orlando, Orange County, Phoenix, Pittsburgh, Portland (separate contracts with same contractor for call center functions and service provision), and San Francisco (separate contracts with the same contractor for bro- ker functions and service provision). In the second group are locations where service provider contractors are also responsible for some of the call center functions and notably scheduling: Atlanta, Las Vegas, and New Jersey (scheduling function shared with NJ TRANSIT). Most (eight) of these are paid based on one of the hourly rates discussed earlier, while four have a per-trip rate, covering all costs, or a split rate with a fixed monthly fee that covers fixed costs and a per-trip rate that covers variable costs. While the majority of the survey respondents in this group have an hourly rate structure, the transit agencies that pay a per-trip rate that covers variable costs make a compelling case. • The MBTA in Boston has used a split per trip for several years. Two of the reasons for its use in Boston: (1) one of its three contractors commingles ADA customers with its other (senior) customers on its own vehicles, noting that these seniors are not allowed on the MBTA-owned vehicles supplied to the contractor; the MBTA believed that allocating the per-hour cost between the contractor and the MBTA to determine payment would be challenging; and (2) one of its contractors uses a taxi company subsidiary to serve ADA trips in a dedicated, semi-dedicated, and non-dedicated fashion. Otherwise, the MBTA believes the per-trip rate is superior to an hourly based rate because it encourages the contractors to deliver the goods. This is because its three-zoned contractors are paid by the completed customer trip. With per-trip rates, transit agencies are not overly

126 aDa paratransit Service Models concerned about productivity, but they are concerned about service quality metrics such as OTP, the degree of lateness among late trips, and missed trips. When paying contractors on a per-trip basis, the MBTA found that its contractors (if left unchecked) would jam their run schedules, so as to maximize revenue and minimize costs. This practice resulted in a low OTP and high numbers of missed trips. To counter that con- tractor bias, the MBTA established severe service quality liquidated damages: a late trip fee equal to the contractor’s per-trip rate, and a missed trip fee equal to twice the contractor’s missed-trip rate. This has worked well over the years, creating the desired balance between cost efficiency and service quality. The MBTA also reports that as it moves to a CCC contrac- tor service model, it is changing its service contractor rate structure to a split hourly based rate. • Access Services in Los Angeles has also had a split per-trip rate with its six zoned contractors for years. Central to this decision was the contractors’ substantial use of taxis being operated in a semi-dedicated and non-dedicated fashion. For four of the six contractors, taxi sub- contractors are used to serve approximately 50% of the trips. So, rather than pay for call center functions one way, dedicated service in another way, and taxi trips in yet a third way, Access Services lumps all costs together into a per-trip rate. • DART in Dallas for years had centralized all of its call center functions in-house and had two, and then one, service provider. Similar to Seattle, DART did not see the need of a centralized call center, and so changed its service model to a turnkey contract. But note also that, like Los Angeles, the prime contractor (MV Transportation) proposed to use a taxi subcontractor for a substantial number of trips. So, as with Los Angeles, and for the same reasons, DART opted to pay its contractor based on a split per-trip rate. As the basis of payment of trips served, all taxi trips are passed through to DART by the contractor, regardless of whether it is served by a van or taxi. • In Orange County, OCTA has a similar per-trip payment structure with its primary contrac- tor (MV Transportation) for basically the same reason: MV subcontracts with a taxi company to serve trips on weekdays after 6:00 p.m.; Saturday service after 2:00 p.m.; and all Sunday service. OCTA reimburses the primary contractor based on split fixed-fee and variable rate for service provided on the dedicated fleet, adding on the cost of the supplemental taxi trips based on mileage, as a pass-through cost. • SFMTA in San Francisco contracts its brokerage and service agreements separately. Under the broker contract, the administrative and call center costs are paid based on a negotiated annual budget. Meanwhile, for the dedicated providers, payment is based on revenue vehicle hours and trips served, and non-dedicated taxis are reimbursed based on each meter fare. The common denominator in the above examples is this: if NDSPs are involved to a greater extent, a per-trip rate, either covering all costs or just the variable costs, seems to be the simplest approach. Also of note: In Seattle, where King County Metro is changing its service model from a CCC contractor model to a turnkey contractor model, King County Metro has decided to implement a new mixed payment structure for its turnkey contractor. With the installment of this new ser- vice model, King County Metro is choosing not to pay the contractor based on a per-trip rate, as the transit agencies above have, but will pay the contractor a mixed fee: (1) a cost-plus fee for call center functions; and (2) a split per-hour rate (i.e., the same way it formerly paid its service provider contractors). The novelty of this mixed rate is that it has never been packaged together for a turnkey contract, until this current Seattle procurement. As shown in Figure 22, contracts with dedicated service providers often separate out variable costs from fixed costs with service hours. And regardless of whether the variable rate covers just variable or all costs, the most common variable rate is based on gate-to-gate hours, followed closely by a RVH rate. Less than 20% of the survey respondents use a per-trip variable rate, although the ones that so almost all have contractors who use NDSPs.

Survey analysis Summary 127 While agencies use a variety of performance metrics, on-time performance is the most com- mon metric, followed by missed trip rate and complaint-frequency ratio. The most common performance metrics that are used to trigger incentives and liquidated damages are shown in Figure 23. Among the surveyed agencies, 12 reported some type of incentive for its contractors. Paying Non-Dedicated Service Providers In contrast to Boston, Los Angeles, Dallas, and Orange County, where service provider contrac- tors subcontract to taxi companies for non-dedicated service, several of the survey respondents N = 25. Figure 22. Dedicated service provider basis of payment. N = 25. Figure 23. Dedicated service provider incentives and liquidated damages.

128 aDa paratransit Service Models indicated that they have direct contracts with NDSPs. The transit agencies that directly contract with NDSPs are located in Houston, Kansas City, Nashville, Phoenix, and Washington. Of these, most rely on the charge from the distance-based taxi meter, while others have insti- tuted either a zone-based fare or, as in Austin or Portland, rely on the stated street distance using the map underlying the paratransit software or a navigational app, such as Google. One of the advantages of the latter approach is that the cost of each trip is preset, allowing the transit agency to generate a draft invoice on behalf of the NDSP. In contrast, Valley Metro in Phoenix, and Nashville pay their NDSP based on a per-trip rate. Most of the other transit agencies referenced above that have direct contracts with NDSPs have payment structures that are distance based. In Seattle, where the call and control manager, First Transit, contracts with a Transdev-owned taxi service (and accordingly also falls under the definition of broker where NDSPs are con- cerned), it pays for non-dedicated service based on a flat pick-up fee plus mileage, similar to classic payment structures use in the Medicaid non-emergency medical transportation industry. Other Noteworthy Findings from the Survey Coordinated systems. Almost half of surveyed agencies provide senior transportation in addition to ADA paratransit, and more than a third provide general public transportation, as shown in Figure 24. Ridership by type. Surveyed agencies serve an average of 1.2 million total ADA paratransit customers every year, as shown in Figure 25. Several agencies also provide senior trips and other human service transportation. These latter agencies serve an average of 90,000 annual trips of senior or human service transportation trips. Commingling. Although nearly half of the providers only provide ADA paratransit trips, 28% of agencies surveyed commingle other types of trips with ADA paratransit trips and 28% do not, as shown in Figure 26. N = 29. Figure 24. Types of trips served.

Survey analysis Summary 129 Service provided beyond the ADA minimum requirements. As shown in Figure 27, almost a third of service providers extend trips beyond the three-quarter-mile distance around fixed routes, and 18% work to accommodate as many same-day trips as possible. Just a quarter of ser- vice providers, however, extend reservation hours into the evening beyond normal business hours. Technologies. As shown in Figure 28, IVR technologies are in place for half of service pro- viders. Fare cards and API Links to the Taxi Dispatch System are the least-used technology for service delivery, with only a fifth of service delivery organizations using them. Key performance indicators. Agencies were asked to report their performance via a few key indicators. Figure 29 summarizes the responses. Cited benefits of service models. Figure 30 summarizes the percentage of agencies that identified specific benefits to their service model, with each agency categorized by its service model type. Highlighted cells focus on benefits where more than 50% of the transit agencies with that service model indicated that their service model resulted in this benefit. N = 25. Figure 25. Annual ridership figures. N = 29. Figure 26. Commingling of ADA trips with other trips.

130 aDa paratransit Service Models N = 20. Figure 27. Premium services offered. N = 29. Figure 28. Technology used in service delivery.

Survey analysis Summary 131 Key Performance Indicators High Low Average No. of Agencies On-Time Performance 95% 82% 89% 24 No Show Rate 10% 0.16% 4% 19 Missed Trip Rate 3% 0% 1% 20 Percent of Excessively Long Trips 5% 0% 2% 16 Trip Productivity 2.11 0.79 1.52 20 Figure 29. Key performance indicators. Model In-House Split Structure Administrative Broker Operational Broker Call Center Manager Single Turnkey Contractor Multiple Turnkey Contractors No. of Agencies* 1 11 1 1 8 5 5 Creates cost efficiency by fostering competition 0% 27% 100% 100% 63% 80% 60% Creates cost efficiency through economies of scale 0% 27% 100% 0% 63% 40% 40% Enhances control over service quality 100% 73% 0% 100% 63% 60% 20% Enhances control over productivity/cost 100% 55% 0% 100% 38% 60% 0% Enhances flexibility to respond to service quality issues 100% 45% 100% 100% 50% 20% 40% Helps manage demand and encourage use of alternatives 0% 18% 100% 0% 50% 20% 40% Minimizes risk for transit agency 0% 36% 100% 100% 88% 60% 60% Minimizes risk for contractor(s) 0% 18% 0% 0% 25% 20% 0% Other (please specify) 0% 18% 0% 0% 25% 0% 0% *Please note that Kansas City, New York City, and Pittsburgh’s answers are identified by two different service models, and as such their answers are duplicated. Figure 30. Agencies’ service model benefits identified by management structure/service design.

132 aDa paratransit Service Models Takeaways from Figure 30 include the following: • Among the agencies that perform some or all of the primary functions, the two most cited benefits are the ability to exert control over service quality and cost, respectively. Interestingly, for this group cost efficiency through economies of scale is seen as much less a benefit. Also, minimizing risk does not seem to be as important as other benefits. • Among agencies that utilize an operational broker or call (and control) center manager, mini- mizing risk for the transit agency is the top-ranked benefit, followed by creating cost effi- ciencies through economies of scale by centralizing call (and control) center functions and enhancing control over service quality. • When comparing the responses from the transit agencies with single-turnkey contractor designs versus multiple-turnkey contractor designs, a surprisingly higher percentage of single- carrier systems believed their design fostered cost efficiencies through competition than the transit agencies that had multiple-carrier designs. Here too, in both cases, the use of service provider contractors is seen as minimizing risk for the transit agency. Four agencies reported specific benefits in the open-ended comments. Two of these agen- cies have split structures, one of which reported that having multiple service providers allows for emergency backups, and that incentives and liquidated damages help enforce the contract requirements. The second split structure provider reported it was in the process of reviewing the contract for potential changes. The other two agencies to report open-ended comments both use call center managers. The first reported that the flexibility of the service is driven by short base contracts, which also provides cost effectiveness, while the second agency reported their contract allows for an easy swapout of contract assets. Cited challenges of service models. The most commonly cited challenge of any agency’s contracting process is that it is expensive. Figure 31 summarizes the percentage of agencies that identified specific challenges to their service model, with each agency categorized by its service- model type. The open-ended challenges supplied by the agencies provided further detail on the costliness of the model, lack of available subcontractors, competition vulnerabilities, difficulty Model In-House Split Structure Administrative Broker Operational Broker Call Center Manager Turnkey Contractors Multiple Turnkey Contractors No. of Agencies* 1 11 1 1 8 5 5 Costly 100% 36% 0% 0% 38% 40% 0% Indirect or poor control over service quality 0% 9% 0% 0% 13% 0% 20% Very rigid; difficult – or takes a long time -- to implement needed changes 0% 18% 0% 0% 25% 20% 20% Eliminates options 0% 9% 0% 0% 13% 40% 0% Does not mitigate risks 0% 0% 0% 0% 25% 20% 0% Other (please specify) 0% 55% 0% 100% 38% 40% 20% Figure 31. Service model challenges identified by agencies.

Survey analysis Summary 133 N = 11. Figure 32. Reasons agencies changed their service model. implementing changes with multiple contractors, and contract compliance issues, which were cited by three different agencies. Reasons for service model change. As shown in Figure 32, for the 12 transit agencies that did change their service model in the past 5 years, it appears that control over the balance between cost efficiency and service quality was the most commonly cited reason. A more complete picture as to why these agencies changed their service model can be seen in Figure 33. As shown, some agencies changed their model only slightly, while others drastically altered the way the service is delivered. City Service Model Change Cited Benefits Other To improve cost efficiency (afford more trips) To enhance – and provide more control over -- service quality To better control the balance between cost efficiency and service quality To better manage demand Ann Arbor Brought all advance reservations in-house. Austin August 2012, contracted out all service delivery and vehicle maintenance. October 2016, contract with non- dedicated taxi provider came to its natural end. Competitive contract won by an NEMT provider. Requirements of Sunset Commission related to collective bargaining requirements, competitive contract procurements. Chicago Centralized the trip distribution of trips for the City of Chicago. All trips are batched daily and distributed to all carriers based on percentage of capacity. Dallas Changed the reservations/scheduling and dispatch functions from the agency control to being under the contractor, and implemented use of taxis for a share of the trips taken, rather than being completely run with vans. Figure 33. Service model changes and reasons for change. (continued on next page)

134 aDa paratransit Service Models Figure 33. (Continued). City Service Model Change Cited Benefits Other To improve cost efficiency (afford more trips) To enhance – and provide more control over -- service quality To better control the balance between cost efficiency and service quality To better manage demand Denver Added non-dedicated service providers Broward County Changed procurement from low bid to RFP best value. Changed model to County owning fleet vs. contractor, county provides fuel vs. contractor. Changed payment from per customer trip to fixed/variable payment method. Reduce risk to County by owning rolling stock, if vendor fails to perform. Improve vendor cash flow when factors beyond control can impact ridership totals. Kansas City Changed from one contractor to two contractors and are considering a third. Las Vegas Shifted scheduling from in-house to the service provider. New York City Increased use of non-dedicated service. To improve customer service. Oakland In FY13/14, converted to a centralized dispatching system at the Broker's office, which gave the central dispatchers visibility of the entire fleet at any one time. Previously each provider did their own dispatching and had visibility to only their own vehicles. To provide dispatchers visibility and access to the entire fleet. San Francisco In September 2014. Previously, the paratransit administrative broker was responsible for contracting out all transportation services and could not bid on any transportation service. Now, the company hired to provide for paratransit administrative brokerage also provides some transportation services (different divisions, same organization) while contracting with other service providers as well. Washington, DC Transitioned from a brokerage model to call and control center model with multiple providers and with one also providing non-dedicated service. To ensure operational transparency.