Implementation and Outcomes of Fare-Free Transit Systems (2012)

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.

9 IntroductIon This chapter summarizes findings from a literature review related to the subject of fare-free public transit. A TRIS search was conducted to aid the review, using key phrases such as “free transit,” “fare-free public transit,” “no-fare transit,” and “free transit demonstration.” Internet searches applying the same terms were used to discover newspaper articles or other information that might be written by reporters or bloggers interested in this subject. A review was also conducted of any similar research listed in the TRB’s Research in Progress database. Finally, any white papers or agency reports identi- fied by project panel members or discovered through inter- views with managers of fare-free public transit systems were also reviewed. Fare-free public transit has been discussed and considered ever since the federal government became involved in pro- viding capital assistance to local public transit agencies in the 1960s (1, 2). The discussion continues to the present day through Internet blogs posted by passionate transit users and policy analysts who debate why, as a public service, transit is treated differently from other public services such as librar- ies and parks, and whether the charging of fares on transit is simply rooted in the origins of transit systems when they were private, for-profit companies (3). The purpose of this report is not to explore all sides of the debate regarding the philosophy of providing fare-free public transit. As the title of the report clearly states, it is intended to review the implementation and outcomes of fare-free public transit systems. Chapters three and four provide information received directly from representatives of the dozens of agen- cies currently providing fare-free service. However, there have also been reports produced over the years that provide valuable information and insights regarding the experiences of those who have either implemented, or considered imple- menting, fare-free transit (see Table 1). The primary concerns of those who consider implement- ing fare-free transit are: • Whether it is cost-effective to eliminate the fare collec- tion process, • The effect fare-free transit has on ridership and system capacity, and • The effect fare-free transit has on service quality and customer satisfaction. cost-EffEctIvEnEss of ElImInatIng thE farE collEctIon ProcEss Passionate advocates of fare-free public transit argue that the following costs associated with fare collection can exceed the amount of money actually collected (4): • Purchasing and maintaining fareboxes and automated ticket vending machines • Provision of secure money counting rooms, equipment, and cameras • Services to pick up and deposit money securely • Accounting and auditing expenses • Production/purchase of fare media such as passes and smart cards • Commissions to third-party vendors and the staff effort to work with them • On-board fare inspectors • Staff time involved with analyzing modifications to fares and the necessary public hearings • Lost time and productivity for bus trips as a result of having to collect and explain fares. Those advocates also believe that most transit managers do not really know what the total cost of collecting fares is at their agencies. That may or may not be true, but there is sufficient evidence that the cost of fare collection has been examined through research and by a number of agencies. A report reviewing transit systems in Washington State noted that the net cost or income of fare-free transit is an important aspect of a fare-free policy (5). By eliminating fares, the rev- enues collected are reduced to zero. The costs related to fare collection can also be eliminated, potentially cancelling out the loss of revenue. The Seattle bus tunnel and Island County Transit are provided as examples. In both cases the costs of fare collection were greater than or equal to the revenues collected, meaning there was no net income from collecting fares. The costs of fare collection vary widely among public transit agencies. TCRP Report 32 (6) documents that some agencies spend less than 1% of their total fare revenue col- lected on fare collection and related costs. The average for all agencies that responded to that report’s survey was 6.2%. For bus systems, the average was 3.4% for smaller systems and 4.0% for larger systems, although it could range from 0.5% to 22%. Based on 1990 operating statistics for Washington State systems, the gross farebox recovery ratio of most transit systems was below 10%, with only three having a recovery ratio higher than 20%. chapter two lItEraturE rEvIEW

10 In general, the smaller the system, the more likely the net revenue of collecting fares is closer to zero. Many of the small transit agencies that responded to the questionnaire for this TCRP project reported they did not do any formal analysis to determine the cost-benefit of charging a fare. For these small urban and rural systems, it was an easy decision to forego passenger fares owing to the minimal revenues they expected to receive versus the perceived costs associated with collecting fares. Small systems in resort areas respond- ing to this project’s survey indicated that it was imperative to their towns’ economic success to provide fare-free transit, even if fares could create net revenue for the system; that is, it was more important for the towns they serve to remain competitive with other resort communities by providing a convenient service to visitors and an affordable mobility option for relatively low-wage service employees. For some university-dominated towns, it was a perceived matter of equity to allow nonstudents to also board for free, particu- larly when fare-paying nonstudents might represent only a small percentage of all passengers. In the case of Chapel Hill Transit, the administrators of the University of North Carolina believed they were spending an inordinate amount of time with the paperwork involved with subsidizing passes for their students. A fare-free system pre-paid by students that provided them with universal access virtually eliminated all university administrative tasks other than writing a check a few times a year to Chapel Hill Transit. This agreement also negated the need for Chapel Hill Transit to purchase addi- tional equipment to read university ID cards. Although they did not do a specific cost-benefit analysis, they believed that foregoing farebox revenue would result in very low net costs because the additional funding they could receive from both state and federal formula grants would be increased as their ridership increased (C. Elfland, Associate Vice-Chancellor for Student Services, University of North Carolina, personal communication, Apr. 18, 2011). In 2008, in a study conducted by Lane Transit District (LTD) in Eugene, Oregon, staff determined that the cost of fare collection was between $100,000 and $500,000 per year Service Area Dates of Demonstration Population of Service Area Results Asheville, North Carolina 08/06–11/06 70,000 58.5% increase in ridership; some problem riders, schedule adherence suffered, retained an increase of 9% in ridership after demonstrations. Austin, Texas 10/89–12/90 500,000 Credited for ridership increases of 30%–75%; reports of disruptive teenagers and driver complaints. Increased operating costs, but successful in promoting ridership. Chelan–Douglas Counties, Washington 12/91–09/00 100,000 Ridership exceeded forecasts by a factor of 4. Policy ended when state funding source was eliminated by voters. Denver, Colorado (off-peak hours only) 02/78–01/79 1,500,000 Reported increases in ridership of 36% to 49%, although inconclusive because of changes in service made during experiment; decreased schedule reliability, crowding. Mercer County, New Jersey (off-peak hours only) 03/78–02/79 300,000 Ridership increases of 25%–30%; 45% of buses ran late, extra buses required, driver complaints, problem riders. Milton, Canada 06/07–12/07 54,000 Ridership increased 63%; some increased rowdiness among young passengers, but 99% of customers “satisfied” or “very satisfied.” Salt Lake City, Utah October 1979 910,000 13% increase in ridership. Topeka, Kansas May 1988 120,000 Ridership increased 86% and 6% increase in ridership was retained after demonstration. TABLE 1 RESULTS OF SySTEM-WIDE FARE-FREE PUBLIC TRANSIT ExPERIMENTS

11 (although it would appear closer to $100,000) compared with the $5 million in revenue that was collected (7). They found that no employees were dedicated solely to fare collection functions. These employees had several duties, and conse- quently, eliminating fares would not result in the elimination of jobs. For example, a customer service representative sells fare instruments, but also conducts trip planning for tele- phone callers and for walk-in customers. If the sales function were eliminated, those hours might be required to serve pas- sengers in the Customer Service Center, particularly if rider- ship increased as a result of free fares. This same conclusion was reached in reverse by Link Transit in Washington State when they converted from fare-free service to charging a fare in 2001. Link’s manager reported in a telephone interview that the agency was able to spread the responsibility for the fare collection process among many employees and that the cost to the agency was believed to be minimal (see the case study in chapter four of this synthesis). LTD’s fare collection system used very basic farebox tech- nology. The success LTD has had in transitioning passengers to pre-paid fare instruments has meant that cash fare custom- ers represent between only 20% and 30% of total ridership. The less cash that is handled, the lower the cost of the fare collection process, and the less delay there is in the boarding process. LTD empties fareboxes only three days a week. The staff report acknowledged that fare collection costs could be much higher at agencies that use more advanced collection technologies or use honor systems that require fare enforce- ment personnel. It also noted that the cost of fare collection at some small systems that might not receive much in fares could be a much higher percentage of overall revenue, mak- ing it more rational to establish fare-free policies. If LTD became fare-free, the report estimated it would lose between $4.5 and $4.9 million dollars in revenue, without an iden- tifiable alternative source of funds to replace that revenue. This would require a 20% reduction in service at the same time the agency would experience a substantial increase in demand. The report did not estimate the cost of increased service, because LTD had no identifiable funds to pay for it. The San Francisco Municipal Transportation Agency (Muni) utilized a consultant to conduct a detailed analysis of the cost-effectiveness of converting to a fare-free system in 2007–2008 at the request of Mayor Gavin Newsom (8). The study concluded that the costs of fare collection amounted to $8.4 million of the Fy 2006 Operations and Maintenance Budget. This represented 7.5% of the $111.9 million Muni collected in fares. There would be a reduction of 91 full-time employees, representing approximately 2% of the total staff if fares were discontinued. However, the study also examined the results of other free-fare experiments conducted in places such as Austin, Texas, and Denver, Colorado, and developed projections on what their additional costs would be based on three different scenarios of ridership increases. The most likely scenario—a 48% increase in ridership—suggested a probable $69 million increase in the annual operating bud- get would be required to handle the increased demand for capacity. When coupled with the foregone revenue previ- ously collected, the agency would need to find an additional $184 million dollars a year to operate the system. Making matters more challenging, the San Francisco Municipal Transportation Agency would have additional capital costs of $519 million to procure the vehicles, facilities, and infra- structure needed to accommodate the substantial increase in ridership. In 1999, Mayor Vera Katz of Portland, Oregon, requested that a group of citizens, assisted by Tri-Met staff, research the role that making the transit system free might play in helping to keep the area from strangling on auto traffic. At the time of the study, Tri-Met recovered approximately 20% of its operating expenses through fares. The report that sum- marized the financial impact of converting to a fare-free sys- tem noted that the agency would lose $41 million in fares, and need an additional $8 million for operating expenses and $5 million for capital expenses to accommodate the addi- tional passenger demand (9). In summary, an additional $54 million in revenue would be needed to replace foregone fares and handle new demand. Surprisingly, the report did not estimate how much the agency might save by eliminat- ing the cost of fare collection, although its estimate of total costs may have accounted for what savings the agency might realize. The group developing the study researched the pos- sibility of imposing a regional parking tax, but found there were a number of legal, institutional, and economic issues that would be difficult to overcome (see Table 2). Advance Transit in Hanover, New Hampshire, serv- ing small urban and rural areas, has been providing fare- free service since 2002 in the Upper Valley region of New Hampshire and Vermont. Respondents to this project’s sur- vey indicated that a number of transit systems that provide fare-free service are challenged from time to time to justify their continued use of the fare policy. In 2008, the CTAA produced a report that analyzed the cost-benefit of changing Advance Transit to a system that charged a fare (10). The capital costs to outfit their fleet of 33 buses with fareboxes would have amounted to $407,550 (which could be amor- tized over more than 20 years at approximately $20,000 per year). Other one-time costs such as the time to create the policy, hold public hearings, and inform the public about the change were estimated to be $33,900. The estimated cost for ongoing fare collection functions per year (not includ- ing amortization of the new fareboxes) was $53,354. These costs would be offset by the new fares collected. A $0.50 fare would generate an estimated $90,688 a year, whereas a $1.00 fare would produce annual revenue of $145,600, and a $2 fare would produce $175,550. Hence, fares collected would exceed the annual cost of collecting the fares, but only barely in one scenario. The highest estimate for revenue to be collected would represent only 4% of a total annual operating budget of $4.3 million. To date, Advance Transit remains a fare-free service.

12 In Summit County, Colorado, the general manager reported that recent cost-benefit analyses have been undertaken to deter- mine the feasibility of implementing a fare system. These have focused on the infrastructure costs of implementing the fare collection system including fareboxes, money counters, and retrofits to facilities to count and store money that was esti- mated to cost $1 million. The general manager provided an undocumented estimate that the annual ongoing costs would be approximately $225,000 to pay for four employees responsible for farebox maintenance, counting and account- ing for money, and providing security. This would represent 16% of the $1.4 million they estimate a $1.00 fare would gen- erate annually. The Aspen Transit Development Plan produced in 2009 reviewed what the financial impact of establishing a $1 fare would be (11). After careful consideration was given to the number of passengers who ride at a discount and the number of riders that would be lost as a result of the institution of a fare, it was estimated that a $1 fare would generate $447,300 annually. The report noted that there would be some new administrative costs, primarily as the result of the need for marketing and fare media production and distribution. It was also estimated that it would require only two hours per day of one person’s time to count and account for fares. All of these functions were to be absorbed by existing staff. The purchase of 16 fareboxes for its bus fleet was estimated to cost up to $144,000. However, the major cost concern was the effect collecting fares would have on buses’ ability to maintain route schedules. The report calculated the increased dwell time resulting from fare collection would accumulate to between two and four minutes per one-way trip. It was noted that an additional bus would need to be put into ser- vice on up to five routes to maintain the posted headways, or the buses would need to run less frequently. Because the cost to add even one extra bus a year to help routes maintain schedule would be $476,000, the report concluded that estab- lishing a fare would not be cost-effective if current levels of service were to be maintained. It was recommended that fares be established only as a last resort. Fare-free transit is also present in European cities and has been subject to scholarly investigation over many years. In an article written in 1973 entitled “Free Public Transport,” the authors look at the projected costs associated with fare- free transit for several German cities, noting that these costs would range from 22 million Deutschmarks (approximately $15 million) in the city of Kassel to 350 million in a city as large as Hamburg (12). The study took into account lost farebox revenue, remaining advertising revenue, increased capacity required during peak periods, savings from the elim- ination of fare collection, and savings from greater productiv- ity of buses as travel time improves owing to less congestion. The net costs were seen as substantial burdens to municipali- ties and the report casts doubt that the German government would be willing to fill the revenue gaps that fare-free transit would produce. In 2008, the Public Works Department of the city of Ham- ilton, a city of approximately 500,000 in Ontario, Canada, prepared a report for the Public Works Committee of the city addressing the potential of offering fare-free service or some Transit Agency and Year of Analysis Savings from Eliminating Fare Collection Functions Costs of Lost Revenue, New Service, and Additional Vehicles and Facilities Estimated Cost of Implementing Fare-Free Policy Lane Transit–Eugene, Oregon (2008) $100,000–$500,000 $5 million in lost fares $4.5–$5 million in net new costs per year Muni–San Francisco, California (2008) $8,400,000 $112 million in lost fares $72 million for increased service $512 million in capital expenses $184 million in net new operating expenses per year Tri-Met–Portland, Oregon (1998) (not provided, but possibly accounted for in costs column) $41 million in lost fares $8 million for increased service $5 million for additional vehicles $49 million in new operating expenses per year Hamilton, Canada (2008) (not provided, but possibly accounted for in costs column) $900,000 in lost fares $30 million for additional service $30.9 million in additional operating expenses per year TABLE 2 PROjECTED COSTS OF IMPLEMENTING A FARE-FREE POLICy

13 other forms of fare discounts (13). The report stated that, based on a conservative estimate of a 20% increase in rider- ship and the elimination of fares, the increase in its operating budget expenditure would be in the order of $30.9 million per year. This would require an additional tax per household of about $161 per year based on a residential assessment of $250,000 in 2008 dollars. In addition, a capital expenditure in the order of $5 to $10 million for fleet expansion and facil- ities accommodations would be required. EffEct farE-frEE PublIc transIt has on rIdErshIP and systEm caPacIty People may argue about the pros and cons of fare-free transit, but none of the literature reviewed for this project questions the fact that ridership will increase when fare-free policies are implemented. No matter what types of experiments, dem- onstrations, or permanent programs have been implemented, public transit systems have experienced significant increases in ridership when implementing fare-free policies. To estimate the ridership impact of changes in levels of public transit fares, including deep discount fare policies, many transit operators over the years have used the “Simpson– Curtin Rule” as the standard to measure the relationship between fares and ridership termed as “elasticity.” This rule estimates that a 10% fare increase will result in a 3% drop in ridership (denoted as -0.3). Conversely, a 100% decrease in fares (fare-free) would be expected to result in a ridership increase of 30% (13). TCRP Report 95: Traveler Response to Transportation System Changes noted that limited data, including some of which are contradictory, suggest that the ridership responses to fare decreases do not differ significantly from rider responses to fare increases. A review of 23 fare changes in United States cities, selected for similar size, found that the fare elasticities were almost identical for fare increases and fare decreases (14). Dargay and Hanly (1999) studied the effects of U.K. transit bus fare changes over several years using sophisticated statistical techniques to derive elasticity values. They found that demand is slightly more sensitive to rising fares (-0.4 in the short run and -0.7 in the long run) than falling fares (-0.3 in the short run and -0.6 in the long run), and tends to be more price sensitive at higher fare levels (15). In 1991, APTA staff produced a report to verify the accu- racy of the Simpson–Curtin elasticity equation (16). An advanced econometrics model was used to review the results of fare increases and decreases at 52 transit agencies, examin- ing the ridership performance 24 months before a fare change and 24 months after a fare change. The model attempted to isolate the impacts of the fare change from other factors such as employment trends, fuel costs, and labor strikes. APTA’s study showed that transit riders react more severely to changes in fares than the Simpson–Curtin rule would predict, and that their reaction varies depending on the size of cities and time of day the fare change is applied. The fare elastic- ity was found to be -0.36 for systems in urbanized areas of more than one million population, whereas it was -0.43 in urbanized areas with less than one million population, indi- cating that travelers in large cities are less sensitive to fare increases. Further, the average peak hour elasticity was found to be -0.23, whereas the off-peak elasticity was -0.42, indi- cating that peak hour commuters are much less responsive to fare changes than transit travelers during off-peak hours (16). These elasticities can vary significantly depending on local circumstances such as income, driving conditions, level of transit service, and the location of work places in relation to the population. Hence, it should not be a surprise that public transit agencies that offer fare-free service might experience a wide range of ridership increases. However, these analyses still do not fully account for increases experienced by fare-free transit systems that go well beyond these elasticity estimates, such as the 58% increase in Asheville, North Carolina (17), the 86% increase in Topeka, Kansas (18), or the 200% increase reported by the island of Hawaii in response to this project’s survey. An intriguing possible explanation is offered by Hodge et al. (5). In their 1994 report, they postulate that standard elasticity formulas might not apply in the same way when fare-free policies are implemented. They note that there is not just a financial cost associated with transit fares, but a psychological cost associated with the farebox. The removal of the farebox can eliminate a barrier in the minds of potential passengers, many of whom might see the farebox as a source of con- fusion and possible embarrassment. The limited capabili- ties of most fareboxes to accept common forms of payment such as credit cards and/or the requirement to have exact fare can certainly discourage passengers. The report prepared for Portland provides a wonderful hypothetical analogy: “The problem with fares is simple: imagine the result if people had to put $1.40—exact change please—in a farebox in their car each time they wanted to take a trip” (9). The first experiments in fare-free transit were conducted in the late 1970s in Mercer County (Trenton), New jersey, and in Denver, Colorado. These demonstration projects were funded in part by the Urban Mass Transportation Adminis- tration. They were instituted to be in effect only during the off-peak hours between 10 a.m. and 2 p.m. and after 6 p.m. and all weekend because of unused capacity and the thought that marginal costs would be minimal. Peak period fares remained the same. The Denver experiment was more dif- ficult to analyze because the transit agency also implemented major route restructuring during the experiment, had insuffi- cient pre-demonstration data, and changed the off-peak hours during the experiment. The experiment in Mercer County led to a significant increase in ridership during the off-peak peri- ods, with a 25% to 30% increase attributed to the removal of the fare. In all, the fare-free demonstration attracted approxi- mately 2,000 new riders per day. Sixty-nine percent of the new trips were previously made by another mode—half by

14 automobile and one-third by walking. It was estimated that the fare-free off-peak transit service reduced Trenton’s typi- cal weekly 21 million vehicle-miles traveled by 30,000 miles per week (19). The Topeka Metropolitan Transit Authority instituted free fares for one month on the bus system serving Topeka, Kan- sas, during May 1988. Compared with May 1987, ridership increased 83.2% on weekdays, 153.4% on Saturdays, and 93.3% overall. Ridership increased 156% on the downtown circulator route. Only one bus a day was added to address problems of overcrowding, indicating that smaller systems carrying lighter loads of passengers can accommodate rather large increases in ridership without needing to provide addi- tional capacity (18). The next substantial experiment in fare-free transit was implemented in Austin, Texas, and conducted from Octo- ber 1989 to December 1990. This experiment was not lim- ited to off-peak hours. The entire system became fare-free every hour and every day of the week. Ridership exploded, increasing 75% during the demonstration period, although some increased service might have also contributed to a por- tion of that increase. This experiment was not funded by the federal government, and no formal report that provides in- depth analysis is available. However, staff from that time reported that additional equipment was required to carry the heavier loads. Even with additional buses placed into service to help accommodate the new demand, the average cost per rider decreased from $2.51 prior to the fare-free experiment to $1.51 during the 15 months of the experiment. That the average cost per rider rose to only $2.18 in the year after the fare-free program was terminated indicates that some of the new passengers gained during the experiment con- tinued to ride once it concluded (20). Templin, a health resort town located in Brandenburg, Germany, with approximately 14,000 inhabitants, modified their small bus service to be fare-free on December 15, 1997. Since then public transportation has been free for everybody. The declared goal of the fare policy was to reduce auto- mobile usage and its collateral effects such as noise, pollu- tion, and the risk of accidents. Within a year after the transit scheme’s introduction, transit ridership had increased by almost 750%—from 41,360 to 350,000 passengers per year. Two years later, in 2000, ridership was above 512,000—more than 12 times its original amount. The study documenting this fare-free program did not include information on how many more buses were required to carry this substantial increase in ridership. It was more interested in determining the effec- tiveness of the policy’s ability to reduce auto trips. A study carried out on behalf of the Federal Ministry of Transporta- tion investigated transit ridership before and after the fare- free program by surveying passengers (21). The study found that the vast majority of new transit riders were children and adolescents. When asked what means of transportation would be replaced, most people answered they would substitute pub- lic transportation for nonmotorized travel. The study found that 35% to 50% of transit passengers would walk less, 30% to 40% would replace bicycle rides, and 10% to 20% would reduce automobile trips. However, it was unclear whether this referred to the driver or the passenger (22). Perhaps the most astonishing example of successful fare- free transit was implemented in Hasselt, Belgium. In 1997, this financially challenged and car-choked city of 70,000 determined it would completely modify its approach to trans- portation (23). Working on the assumption that you will not get people out of their cars without providing a comprehen- sive public transport system alternative, Hasselt transformed its two-line bus service to a nine-line service, covering every district in the city; and committed to half-hourly service dur- ing the day and a night bus that served every stop in the city. On day one—july 1, 1997—the numbers of passengers rose from the usual 1,000 to 7,832. Ridership increased more than 1,200% by 2001. A ring road near the inner city was con- verted to a pedestrian corridor, and parking in the inner city was restricted. Big car parks were banished to the edge of town, and parking priority within town was given over to resi- dents and the elderly. Parking was allowed for a maximum of one hour. The maximum speed in town was reduced to 30 km. Clearly, more equipment was needed and provided for this major modification to the transportation system of Hasselt. The council was in deep debt in the mid-90s and the radical approach was partly prompted because it could not afford a new ring road. Improving the bus service and making it free was less expensive. In 1996, there were only three bus routes with approximately 18,000 service hours/year. By 2003, the city expanded service to offer 11 routes with more than 95,000 service hours/year. Service frequency now ranges from 5 to 30 minutes throughout the city (see Table 3). Clearly, Hasselt anticipated the need for considerably more transit service with the implementation of free fares and a desire to totally modify its transportation services. The transit system in Hasselt cost local taxpayers approximately $1.9 million in 2006, amounting to 1% of its municipal bud- get and making up about 26% of the total operating cost of the public transit system. Fortunately for Hasselt, the Flem- ish national government covered the rest (approximately $5.4 million) under a long-term agreement (24). Asheville, North Carolina, conducted a totally unrestricted fare-free promotion for three months in 2006. Ridership increased by approximately 60% during the promotion. In spite of the significant increase in ridership, insufficient capacity was not cited as a major problem. However, based on surveys, existing customers were not happy with the crowded buses; that issue represented 21% of all complaints by the 45th day (25). The city of Milton, Canada, near Toronto, was the first municipality in Canada to provide fare-free service for an extended period of time. In 2007, public transit was made

15 free to all users during the midday off-peak time (9:00 a.m. to 3:00 p.m.) from june through December. Ridership increased an average of 63% over the seven-month period. The report did not include information on additional costs or equip- ment needed. Two private companies agreed to pay for lost revenue and additional costs; therefore, the main focus of analysis was on effects on ridership. On-board surveys were conducted during the demonstration and found that of the 80% of riders who used the bus at least two times per week during the fare-free demonstration, 86% would continue to use it as often even after fares were reintroduced. However, only 33% of senior riders indicated that they would continue to use the service as frequently after fares were reintroduced, suggesting that seniors are generally more sensitive to cost increases (26). As noted earlier, in 2008, the city of Hamilton reviewed the potential impacts of providing fare-free transit in the ninth-largest city in Canada. Although the report noted there was no Canadian system-wide experience to draw from, it estimated that ridership increases would conservatively reach 20%, but might reach as high as 50% depending largely on the level of congestion and parking policies adopted (13). This same report included an appendix of a case study of Cha- pel Hill, North Carolina, that included a memorandum pre- pared by the town manager of Chapel Hill in October 2002. In january 2002, Chapel Hill Transit finalized agreements with local universities and townships to offer fare-free pub- lic transit service to all passengers in their service area. The town manager’s report noted that ridership on the fixed-route services had increased by 43% from january 2002 through September 2002. Although the city manager’s report also noted that service hours were increased 11%, the primary reason for the increase in ridership was clearly the fare-free policy (27). One of the most recent instances of implementing fare- free public transit has been in the city of Changning, China, a municipality of approximately 53,000 people located in the central portion of the country. In july 2008, the city began providing fare-free service on the three routes serving the city. Based on information in a paper submitted to TRB in 2010, ridership increased from 11,400 a day to 59,600 per day, representing an increase of almost 550% in less than two years. It was not completely clear from the paper if any service hours were added to handle the additional demand, although it appears likely that it would have been reported if more service hours or buses were added. The paper indicated that an additional 7 million yuan (approximately $1 mil lion dollars) was spent on the program, presumably to replace fares previously paid by passengers (28). Apparently it is the only fare-free public transportation offered in China, and various observers question whether it is something the city can finan- cially sustain given so many other priorities, including health care, education, and housing (29). For the time being, the economy and the public appear to support the fare-free service in this city, small by China’s standards; however, observers believe the concept would not be so feasible in larger cities in the country. EffEct farE-frEE PublIc transIt has on sErvIcE QualIty and customEr satIsfactIon As noted earlier, fare-free transit will attract more passengers to a public transit system. In some experiments, the increased number of passengers not only tested the capacity of the buses, but also the ability of the buses to stay on schedule. When fare-free transit is introduced, the time for each individ- ual passenger to board is reduced, because they do not have to take the time to pay a fare. On average, taking into account that some passengers pay with cash and others with some form of pass, it takes a passenger between 3.0 and 3.5 sec- onds to pay their fare when they board (30, 31). In addition, it is possible that passengers who do not pay a fee can board through all doors, saving additional time. However, because fare-free transit will attract many more passengers, the bus is likely to make more stops than it would if fares were charged. The time a bus takes to decelerate to enter more bus stops and accelerate to regain cruising speed can eliminate any sav- ings from reduced dwell time gained from the elimination of collecting fares (32). Schedule adherence is subject to being negatively affected by a significant number of people riding the bus a short distance who might have otherwise walked Location and Population Description of Program Effect on Ridership Templin, Germany (14,000) Small transit service in health resort town Increase from 41,360 passengers per year to 512,000 per year in two years Hasselt, Belgium (70,000) Total change in transportation policies restricting cars and increasing transit Increase from 1,000 per day to 13,000 per day within four years Changning, China (53,000) Eliminated fares without adding service Increase from 11,400 per day to 59,600 per day within two years TABLE 3 RIDERSHIP RESULTS OF TOTALLy FARE-FREE PROGRAMS OUTSIDE NORTH AMERICA

16 (33). In the fare-free demonstration conducted in Trenton, New jersey, between 5% and 15% more buses entering the downtown area were found to be overcrowded during the time fare-free service was provided. In addition, the number of buses running behind schedule increased to 45% (19). Dur- ing the fare-free experiment in Asheville, North Carolina, the major complaint of riders was poor reliability. Travel time was estimated to have increased by several minutes per hour because of the increased number of stops and longer dwell times associated with the 58.5% increase in ridership (25). It would appear that the potential negative impacts of fare- free transit on schedule adherence could be mitigated to a degree without degrading service frequency or adding to costs by a judicious reduction in the number of bus stops (32). Con- versely, it can be noted that those transit agencies in resort and university-dominated communities that responded to this project’s survey indicated that there would be no way for them to keep their schedules without a fare-free system. Transit agencies in these communities often have bus stops with substantial numbers of passengers boarding, and the boarding process would take much longer if each passenger had to pay a fare or show a pass. Fare-free transit will please many passengers and frustrate others. In Asheville, several reported that some younger peo- ple refused to give up seats for more elderly customers. There was an initial drop in handicapped utilization. A few women reported being uncomfortable with what was described as a rougher than normal customer group; however, no reports of any actual physical abuse were made concerning these fears (25). Because ridership escalates when a fare-free policy is implemented is the clearest indication that passengers, as consumers, appreciate the reduced costs. The seven-month experiment conducted in Milton, Canada, included a survey of passengers that indicated that 99% of all respondents were either “satisfied” or “very satisfied” with the program. The NSI Research Group found that 75% of transit users had a favorable or very favorable reaction to the elimination of fares during the Austin experiment (34). However, that same experiment also was subject to complaints by the system’s bus operators who complained vehemently about excessive rowdi- ness among younger passengers and what they believed were conditions that jeopardized their safety and that of their pas- sengers (20). Similar concerns indicating a decline in morale were expressed by bus operators during the Denver and Tren- ton demonstrations (35). It can be noted that many respon- dents to the survey for this project stated that they believed their bus operators viewed fare-free transit very favorably, and would gladly trade the need to deal with a few more undesirable passengers for being relieved of the duty of col- lecting fares with the attendant fare disputes. In short, fare-free policies have the potential to either improve or detract from the quality of service. As a report on fare-free public transit systems prepared for the Washington State DOT concluded, smaller communities are more likely to encounter fewer problems and more success, as are tran- sit agencies and communities that are committed to the con- cept owing to concerns over the environmental impacts of transportation or social equity (5). The authors of that report noted the importance of instituting education programs to deal with middle and high school students in particular. They also noted that although some larger communities such as Austin might have found it overwhelming to deal with younger students (the former general manager noted how school buses would ride empty while students chose to ride the public buses) (A. Kouneski, General Manager, Austin Transit System, personal communication, june 28, 2011), other communities such as Logan, Utah, and Whidbey Island saw serving youth as one of the agencies’ primary missions. Fare-free public transit relieved parents of the responsibility of serving as chauffeurs, and allowed students to access the many resources in their communities (5). Based on the results of the survey for this project, there are no communities larger than 175,000 residents in the United States that provide fare-free public transit throughout their entire system, nor were any others found in the rest of the world. The primary reasons appear to be the difficulty in find- ing funds to replace the revenue they would lose through the farebox and the additional expenses they would incur in maintaining service quality for greater demand. The literature search has also shown that commuters in private vehicles are not attracted in large numbers to fare-free public transit. Absent other types of transit-supportive policies such as restricting parking, the vast majority of commuters will con- tinue to prefer driving. Hence, without disincentives to using private vehicles, minimal gains toward the goals of reducing congestion and air pollution would usually be expected. However, there are dozens of smaller communities through- out the nation that have implemented fare-free public transit. They are identified in the next chapter, along with the reasons why they have found fare-free public transit to be a posi- tive service in their communities. Other communities such as State College, Pennsylvania, with a regional population of approximately 80,000 in an area dominated by Pennsylvania State University, have hired consultants to review the feasi- bility of establishing a fare-free system for its entire service area (36). The city of Longmont, Colorado, a community of approximately 90,000 people outside of Denver, has made application to the Denver Regional Council of Governments’ Congestion Management Air Quality Regional TDM funding pool in the amount of $300,000 for a two-year fare-free tran- sit demonstration project. Funds would be used to plan for the demonstration, prepare ordinances to deal with disruptive passengers, market the program, and pay the Regional Tran- sit District as a replacement for fares that would have been collected at the farebox (S. McCarey, Alternative Transpor- tation Coordinator, Boulder County Transportation, personal communication, june 23, 2011).

17 The general manager of the Duluth Transit Authority in Duluth, Minnesota, a community with a regional popula- tion of approximately 280,000 on the western most point of Lake Superior, has also indicated that it is strongly con- sidering a fare-free system following review of the total cost of the fare collection process against the amount of revenue being received. Cash fares have become a smaller part of their revenues because of a prepaid program with the University of Minnesota–Duluth (D. jensen, General Manager, Duluth Transit Authority, personal communica- tion, Apr. 20, 2011). Should Duluth proceed with a fare-free system, it would become the largest community, in terms of population, to have such a policy in place. The Corvallis Transit System in Oregon (one of the case studies in chapter four) was con- verted to a fare-free transit agency in February 2011 (37). A bibliography summarizing many of the reports noted in this literature search is included as Appendix C, and the reader is invited to read them for additional details on fare- free experiments and those agencies that analyzed the feasi- bility of establishing fare-free public transit.