Analyzing the Effectiveness of Commuter Benefits Programs (2005)

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.

36 CHAPTER 3 UNDERSTANDING THE IMPACTS OF TRANSIT BENEFITS PROGRAMS Chapter 3 of this report discusses the impacts of transit benefits programs on employee travel behavior and transit agencies’ ridership, revenues, and costs. Limited data were available to evaluate vanpool and other financial benefit pro- grams, and these results are discussed only briefly. This part of the report is designed to document the experiences of various transit benefits programs across the United States and to shed light on factors that influence the effectiveness of transit ben- efits programs. It is hoped that this information will help tran- sit agencies and other organizations set realistic expectations for potential program impacts. Chapter 3 of this report focuses on two types of impacts asso- ciated with transit benefits programs: (1) impacts on employee travel behavior and (2) impacts on transit agencies’ systemwide ridership, revenues, and costs. IMPACTS ON EMPLOYEE TRAVEL BEHAVIOR Understanding the impacts on employee travel behavior of transit benefits programs is key to quantifying a wide range of effects associated with these programs, including employer parking cost savings, employee commute cost savings, increases in transit ridership, and reduced air pol- lution and greenhouse gas emissions. The primary question this study addresses is: To what extent do employees increase their use of transit when transit benefits are offered? To the extent that increases in transit ridership do occur, transit agencies and others also want to know the following: • What trips are affected? Do transit benefits recipients increase only their commute trips, or do they increase their noncommute trips as well? This is important because there is typically excess capacity on transit ser- vices during noncommute periods. • To what extent do new transit riders shift from drive- alone commuting? This is important since there would be little reduction in traffic and emissions if new transit riders previously walked, bicycled, or carpooled to work. • What factors affect the level of travel behavior change? This information is important for agencies pro- moting these programs to better understand what level of employee response to anticipate from their programs. This information also can help agencies better under- stand how to design employer programs and target marketing efforts to maximize ridership gains. Data Sources and Approach Two primary data sources were used to answer these questions: (1) surveys conducted by transit agencies, com- muter organizations, and other agencies in regions with tran- sit benefits programs, either published or unpublished (referred to as “surveys”), and (2) worksite trip reduction reports from regions with mandatory commute trip reduction (CTR) programs (referred to as “data sets”). Figure 5 displays a map showing the locations where these surveys and data sets were obtained, and each type of data is described briefly below. Survey Data from Transit/ Commuter Organizations Survey data were collected from transit agencies, commuter organizations, and third-party benefits providers around the country, as well as through a review of literature (source infor- mation for survey data is provided in the notes to Table 4). In total, the research team identified 21 surveys conducted in 12 regions from 1989 to 2004 (the same survey was adminis- tered in Philadelphia, Pittsburgh, and Harrisburg in 1993, but is counted as three surveys; Philadelphia, Harrisburg, and Pitts- burgh are counted as separate regions; Montgomery County, MD, and Washington, DC, are considered one region; and San Francisco, CA, and San Jose, CA, are considered two regions) that contained some quantitative results on the travel impacts associated with implementing a worksite transit benefit. The surveys included both published and previously unpublished results. All of these surveys focused exclusively on transit ridership; vanpool ridership was not discussed. Most of the surveys were conducted prior to 1998, when tax law changes enabled employers to let employees set aside income on a pre- tax basis for transit or vanpool benefits; therefore, most of the surveys address only employer-paid benefits. Table 4 pro- vides information on survey locations, dates, and number of worksites represented and on other survey characteristics.

37 veys provide information on the share of employees who are aware of or have access to a transit benefits program; these surveys can also provide information on the extent to which employees say they would participate. How- ever, these surveys usually do not provide information on travel behavior changes that occur in response to implementation of a transit benefits program. In some cases, agencies shared their raw data with the research team, which allowed the team to perform its own cal- culations and analysis to more readily compare results among different surveys. In other cases, the research team received only a written summary of results and could not perform fur- ther analysis. Appendix A explains how various metrics of travel behavior changes were calculated based on available data. Appendix B contains a summary of primary travel met- rics from all of the surveys in tabular form for quick reference. Some surveys asked more detailed questions about the level of employer-paid benefit and the number of trips made on transit for commuting and noncommuting purposes, and some provided more detailed breakdowns of results by geo- graphic area (i.e., urban or suburban worksite location). A summary description of all 21 surveys can be found in Appendix C. (Report appendixes are published as TCRP Web-Only Document 27. To access this web-only docu- ment, go to www4.trb.org/trb/onlinepubs.nsf and click on “TCRP Web Documents.”) It should be noted that although these surveys represent diverse geographic areas, the surveys tend to be concentrated in large metropolitan areas. Also, the transit benefits program type is skewed heavily toward areas with universal pass pro- grams and voucher programs. Eight of the surveys—those in San Jose (1), Portland (2), Denver (2), Los Angeles/UCLA (1), and Minneapolis/St. Paul (1)—involved universal pass pro- grams. Eight of the surveys—those in Philadelphia (3), San Francisco (1), and New York (4)—involved voucher programs. Conventional monthly pass programs are barely represented. As seen in Table 4, the scope of each survey varied widely. Some surveys cover one employer (e.g., a 2001 survey of employees at the University of California at Los Angeles and a 1990 survey of employees at the Port Authority of New York and New Jersey), some surveys cover a large number of employers (e.g., a 1994 survey of 50 employers in the New York metro area and a 2001 survey of 94 employers in Mont- gomery County, Maryland), and some surveys are designed to be representative of commuters at a regional level (e.g., a 2001 State of the Commute Survey in the Washington, DC, area; a 2003–2004 survey of commuters in the New York metropolitan area). The surveys generally fall into three categories, each of which provides a different type of information: • Surveys of transit benefits recipients. These surveys focus solely on people who receive a transit benefit. These surveys provide useful information on the share of recipients who say they are new to transit, increased their use of transit, or reduced driving to work; however, these surveys do not provide information on worksite mode shares or information on how many people do not participate in the program. • Before and after surveys of all employees at partic- ipating worksites. These surveys provide important data on mode shares because they involve surveys of all employees. As a result, these surveys provide more complete information on how employees change their travel behavior. Before and after surveys may be more reliable in estimating changes in travel behavior than surveys of transit benefits recipients, which simply ask about previous travel behavior. On the other hand, a long time period between the before and after surveys—a long time being more than one year—could mean that other factors are influencing changes in mode split. • Surveys of commuters in general. These are typically random phone surveys of the general public. These sur- Atlanta Minneapolis/St. Paul San Jose Denver 2 surveys New York 4 surveysHarrisburg Los Angeles Portland 2 surveys Philadelphia 3 surveys Pittsburgh Washington, DC 2 surveys Montgomery Co., MD San Francisco Survey data from transit/commuter organizations Survey data from mandatory CTR program areas Tucson (Pima Co., AZ) Washington State Figure 5. Locations of surveys collected for analysis.

yevruS noigeR etaD yevruS detcudnoC roF/yB saW ohW deyevruS rebmuN deyevruS syevruS devieceR esnopseR( )etaR fo epyT retummoC tifeneB deyevruS erofeB dna ?retfa no ofnI xat-erP ro r-eyolpmE paid? seussI ataD esoJ naS 1 –699 1 799 aralC atnaS yellaV noitatropsnarT ytirohtuA )ATV( 1 llA seeyolpme ssaP ocE ta sreyolpme 6 ta 063,8 sreyolpme ( 029 11 lasrevinU )% ssap oN seY emussA( )diap yb detanimod stnednopser yevruS neewteb shtnom neT .reyolpme eno yluJ( syevrus retfa dna erofeb 1 yaM–699 1 .)799 RO ,dnaltroP 1 –799 002 1 tcirtsiD dyolL noitatropsnarT tnemeganaM noitaicossA )AMT( 2 llA seeyolpme tropSSAP ta ni sreyolpme dyolL tcirtsiD ta 399,5 24 sreyolpme 002 ni( 1 )yevrus lasrevinU )%36( 677,3 ssap erofeb eht neewteb sraey lareveS oN seY ,elihwnaeM .syevrus retfa dna evah smargorp gnitummoc rehto syevruS .detnemelpmi neeb od tub srebmem AMT tneserper laudividni emas eht tneserper ton taht setamitse AMT eht ;sreyolpme eht 1 000,2 dedulcni yevrus 799 ta seeyolpme 1 .sreyolpme 5 RO ,dnaltroP 1 –899 1 999 teMirT 3 llA seeyolpme gnitapicitrap teMirT ni ylhtnom dna sessap troPSSAP ta 333,7 23 1 sreyolpme ylhtnoM A/N ,ssap lasrevinU ssap egarevA seY seY 1 erofeb neewteb sraey 3. yb nwodkaerB .syevrus retfa dna rehtehw dna tifeneb fo level ssap lasrevinu ni detapicitrap .margorp -nO revneD ,gniog detroper fo sa 3002 revneD lanoigeR noitatropsnarT )DTR( tcirtsiD 4 llA seeyolpme ssaP ocE ta sreyolpme ta 794,5 73 sreyolpme 1 085, rof %3.92( %4.82 ,erp )tsop rof lasrevinU ssap ni stniop tnereffid tcelfer syevruS oN seY ehT .sreyolpme tnereffid rof emit detcudnoc saw yevrus tsrif gnitnemelpmi erofeb yletaidemmi detcudnoc saw rehto eht ;margorp wol llarevO .retal shtnom 6 wol stcelfer yliramirp etar esnopser .sreyolpme egral ta etar esnopser :noitacol yb tuo nekorb sesnopseR .nabrubus dna ,egnirf nabru ,DBC revneD 1 DTR revneD 399 5 llA seeyolpme ssaP ocE ta sreyolpme ,7 1 #( 03 fo sreyolpme )A/N .8( 775 1 lasrevinU )% ssap oN oN emussA( )diap DBC sedulcni ylno yevruS yevrus wol yreV .seeyolpme .etar esnopser ,notgnihsaW CD 002 1 natiloporteM notgnihsaW fo licnuoC OO stnemnrevo 6 modnaR fo elpmas deyolpme snosrep eulav-derotS )A/N( 002,7 A/N drac sedivorP .yevrus enohpeleT oN oN noitamrofni tnacifingis yllacitsitats fo esu dna gnitummoc tnerruc no edivorp ton seod ;smargorp etummoc ni egnahc no noitamrofni .roivaheb ,notgnihsaW dna CD erehwesle 1 lareneG 399 gnitnuoccA )A G G( eciff 7 laredeF seeyolpme eht dnuora yrtnuoc CD ni %57( )aera ortem ta 000,95 1 05 seicnega elpitluM A/N sepyt elbissop llA( seY oN -reyolpme )diap yb nwod nekorb ton serugiF sreyolpme laredef 57 fO .noiger ruof ,stifeneb tisnart gnidivorp 2$ tser eht ,.om/06$ dedivorp 1 .om/ TABLE 4 Characteristics of surveys from transit agencies, commuter organizations, and others 38

(continued on next page) yevruS noigeR etaD yevruS detcudnoC roF/yB saW ohW deyevruS rebmuN deyevruS syevruS devieceR esnopseR( )etaR fo epyT retummoC tifeneB deyevruS erofeB dna ?retfa no ofnI xat-erP ro r-eyolpmE ?diap seussI ataD 002 selegnA soL 1 fo ytisrevinU ta ainrofilaC selegnA soL )ALCU( 8 llA seeyolpme ALCU ta 21,1 ta 94 1 reyolpme lasrevinU A/N ssap llA( seY seY -reyolpme )diap .detroper ton etar esnopser yevruS /silopaenniM luaP .tS tisnarT orteM 3002 9 llA seeyolpme ssaporteM ta sreyolpme ta 005,73 6 sreyolpme edtneserper lasrevinU A/N ssap ,puorg regral derevoc yevrus elihW oN seY eht ylno no atad dedivorp orteM .sreyolpme xis tsegral ATRAM 3002 atnaltA 10 seeyolpmE gniviecer ATRAM sessap 1 88,3 1 ta 78 sreyolpme ylhtnoM )%42( 043,3 ssap tuB( seY oN ton stluser )detarapes .etar esnopser wol ylevitaleR erawaleD 0002 aihpledalihP lanoigeR yellaV gninnalP noissimmoC )CPRVD( 11 seeyolpmE gniviecer kehCtisnarT #( 572,2 fo sreyolpme )A/N .etar esnopser wol ylevitaleR seY oN rehcuoV )%83( 568 aihpledalihP 1 CPRVD 699 12 seeyolpmE gniviecer kehCtisnarT #( 000,5 fo sreyolpme )A/N 1 oN oN rehcuoV )%43( 676, emussA( )diap .etar esnopser wol ylevitaleR 2 ta 005 sreyolpme ;)tset-erp( ta 000,4 34 sreyolpme 31 – )%36( 4 dna ;tseterp ( 683 1 – )%6 yevrus lanif ,aihpledalihP ,hgrubsttiP 13 1 399 CPRVD 14 seeyolpmE gniviecer kehCtisnarT ,aihpledalihP ni 63( ,hgrubsttiP ni 4 )grubsirraH ni 3 oN oN rehcuoV emussA( )diap dna tseterp( yevrus fo segats owT .)yevrus lanif naS ocsicnarF noitatropsnarT noissimmoC (MTC)15 seeyolpmE gniviecer retummoC kcehC ot 006,3 ta 005.4 932 sreyolpme 1 –04( 008, )%05 oN oN rehcuoV emussA( )diap morf deviecer erew sesnopseR 1 94 setar esnopser htiw ,sreyolpme .sreyolpme rellams morf tsehgih esaercni pirt ni seicnapercsiD .sliated rof yduts esac ees ;serugif :noitacol yb tuo nekorb sesnopseR .nabrubus dna nabru retneC tisnarT 4002 kroY weN 16 modnaR fo elpmas seeyolpme #( 050,3 fo sreyolpme )AN/ tuB( seY oN rehcuoV /AN ton stluser )detarapes esnopser yevruS .yevrus enohpeleT evah stluser tub ,dedivorp ton etar .%5 -/+ fo ecnedifnoc fo level a kroY weN 1 retneC tisnarT 499 14 seeyolpmE gniviecer kehCtisnarT ,8 1 ta 57 05 sreyolpme ,4 1 5( 07 1 oN oN rehcuoV )% emussA( )diap yb diap .om/54$ tuoba fo egarevA .reyolpme Harrisburg Metropolitan1994 TABLE 4 (Continued) 39

yevruS noigeR etaD yevruS detcudnoC roF/yB saW ohW deyevruS rebmuN deyevruS syevruS devieceR esnopseR( )etaR fo epyT retummoC tifeneB deyevruS erofeB dna ?retfa no ofnI xat-erP ro r-eyolpmE ?diap seussI ataD kroY weN 1 retneC tisnarT 099 14 seeyolpmE gniviecer kehCtisnarT troP ta fo ytirohtuA JN/YN ta 548 1 reyolpme oN oN rehcuoV )%26( 625 emussA( )diap ot ylno nevig tifeneB lla ton ,ffats lairaterces/lacirelc $ egarevA .seeyolpme 1 diap .om/5 .reyolpme yb kroY weN 1 retneC tisnarT 989 14 seeyolpmE gniviecer kehCtisnarT ta 006,4 1 39 sreyolpme oN oN rehcuoV )%05( 023,2 emussA( )diap .enoN yremogtnoM DM ,ytnuoC 002 1 yremogtnoM ytnuoC retummoC secivreS 17 seeyolpmE setiskrow ta gnitapicitrap erahSeraF ni repuS dna erahSeraF .smargorp tub ,A/N latot 49 sreyolpme 1 elpitluM )A/N( 527, sepyt elbissop -derots( ,drac eulav ylhtnom )ssap llA seY reyolpme diap nehw detcudnoc erew syevruS ;smargorp eht denioj sreyolpme 002 ni syevrus pu-wollof 1 repuS( 2002 dna )erahSeraF( setiskrow erahSeraF .)erahSeraF ;om/05.23$ ot pu hctam ytnuoc teg hctam teg erahSeraF repuS ni esoht tsael ta edivorp yeht fi 46$ fo .om/56$ 1 Unpublished data provided by Scott Haywood, VTA. 2 Lloyd District Transportation Management Association. 2001 Accomplishments. Available online at www.1dtma.com/pdf%20files/2001%20survey%20summary.pdf. 3 Unpublished data provided by Tony Mendoza, Planner IV, Tri-Met, on January 17, 2002, on Transp-TDM listserv. Listserv postings available through www.cutr.usf.edu/index2.htm. 4 Unpublished data provided by Denver RTD staff. 5 Howell Research Group. ìEco Pass Effectiveness Study.” Prepared by for Regional Air Quality Council. November 1993. Unpublished study provided by Denver RTD staff. 6 LDA Consulting, et al. State of the Commute 2001: Survey Results from the Washington Metropolitan Region. Prepared for Commuter Connections. Publication Number 22604, July 2002. 7 General Accounting Office. Mass Transit: Federal Participation in Transit Benefits Programs. GAO/RCED-93-163. September 1, 1993. 8 Brown, J., D. B. Hess, and D. Shoup. “Fare-Free Public Transit at Universities: An Evaluation.” Journal of Planning Education and Research Vol. 23, pp. 69–82, 2003. 9 Unpublished data provided by Robert Gibbons, Metro Transit. 10 Center for Transportation and the Environment. Evaluation of the Effectiveness of Programs Contained in the Framework for Cooperation and Reduce Traffic Congestion and Improve Air Quality—Phase Three: February 2003 Discount Transit Pass User Survey Final Report. Prepared for Georgia Department of Transportation. Available online at www.tdmframework.org/reports/files/FY2002FnlRptAppII.pdf. 11 “TransitChek Research 2000, Summary Highlights.” Unpublished data provided by Stacy Bartels, Delaware Regional Valley Planning Commission. 12 “TransitChek User Survey: Summary of Results.” Unpublished data provided by Stacy Bartels, Delaware Regional Valley Planning Commission. 13 One survey was conducted covering worksites in all three of these metropolitan areas in Pennsylvania. Breakdown of the number of employees surveyed at employers in each region not provided. 14 U.S. Department of Transportation, Federal Transit Administration. TransitChek in the New York City and Philadelphia Areas. FTA-MA-26-0006-91-1; DOT-VNTSC-FTA-95 11. October 1995. Available online at www. fta.dot.gov/library/program/tchek/TransitChek.html. 15 Oram Associates. “Impact of the Bay Area Commuter Check Program: Results of Employee Survey.” Prepared for Metropolitan Transportation Commission. 16 ORC Macro. “Commuter Benefit Impact on Transit Use: A TransitChek Study” (MS PowerPoint presentation). Prepared for TransitCenter, Inc., August 2004. Unpublished study provided by Transit Center staff. 17 Unpublished data provided by Montgomery County Commuter Services staff. TABLE 4 (Continued) 40

41 There are other potential limitations of the data as well. The sample size was small in some surveys. Further, there is a potential for bias in the selection of worksites; it is possible that worksites that were surveyed tended to be the more suc- cessful worksites in terms of increasing transit use. Finally, there is a potential for bias in survey response; it is possible that more employees who switched to transit were likely to respond. This bias is especially a concern because survey response rates were low—in the 10- to 60-percent range—in most surveys. Differences in questions among the different survey instru- ments also mean that results across surveys are not fully com- parable. For instance, some surveys asked about the “primary mode of commuting,” whereas others asked whether employ- ees “ride transit.” The responses to these questions may dif- fer because an occasional transit user would probably say “no” to the question of whether he or she uses transit if the question about the primary mode of commuting is asked, but would respond “yes” to the question about whether he or she rides transit. Other surveys ask about commuting behavior over the course of a week to capture variations in transit use more accurately. Finally, the before and after surveys do not provide infor- mation on factors other than the implementation of a transit benefits program that might be influencing transit use over the period between the two surveys, such as implementation of other employee commute programs (e.g., rideshare match- ing or a telecommuting program) or changes in employment demographics. Data Sets from Regions with Mandatory CTR Programs The second set of data analyzed includes employer trip reduction reports from three regions where certain employers are subject to mandatory CTR programs—Southern Califor- nia, Tucson (Pima County, Arizona), and Washington State. These data sets provide a wealth of information on individual worksites, including the number of employees at the worksite, worksite location, and commuter programs that are offered (ranging from financial incentives to nonfinancial incentives such as preferential parking and telecommuting). These data sets also provide reported mode share data at various points in time. As a result, these data sets enable more detailed analysis than the surveys conducted by transit agencies and other orga- nizations. Specifically, these data sets allow an assessment of the independent effects of a transit benefits program: those effects that are separate from the effects of other factors that might also be influencing travel behavior. The data sets are as follows: • In Southern California, all worksites of more than 250 employees are required to implement a CTR program (from 1988 to 1996, the program covered all employers over 100 employees). The Southern California data, obtained from the South Coast Air Quality Management District (SCAQMD), consist of 33,092 total records from 7,626 employer worksites, covering the years 1988 to 2001. Each record represents information from a specific worksite for a specific year, including benefits offered and mode split. There can be multiple records from a spe- cific worksite over several years or several records for a company with multiple worksites. • In Tucson (Pima County, Arizona), all employers over 100 employees must have trip reduction programs. The Tucson data were obtained from the Pima County Association of Governments and consist of 1,438 total records from 317 company worksites covering the years 1996 to 2001. • In Washington State, the state’s CTR law currently covers employers over 100 employees in nine counties. Data were obtained from the Washington Department of Transportation (DOT) and consist of 2,444 total records from 1,038 company worksites. Data were collected every two years, in 1995, 1997, 1999, and 2001. In all three data sets, not every worksite is represented in the survey for every year. Note on Terminology and Measurement Commuter benefits recipients: employees who receive com- muter benefits through their employer. At most worksites, employees have to opt to participate; however, under some pro- grams, all employees receive benefits, whether they use them to ride transit or not. Transit riders: employees who use transit. At worksites offer- ing a commuter benefit, transit riders are a subset of commuter benefits recipients (in the case of programs where employees have to opt to participate, all recipients may be transit riders). Mode share: the percentage of all persons using a particular mode (e.g., transit, carpool, or walking) to make a trip. A change in the percent of travelers using a mode is a mode shift. Percent change in use of a mode: the percent change shows the increase or decrease in use of a mode, calculated on the basis of starting mode share. For example, an increase in tran- sit mode share from 20 to 22 is a 10-percent increase in transit use (2 divided by 20). Change per 100 employees: In order to provide a consistent way to compare the absolute change in use of a mode, changes are represented in terms of a worksite of 100 employees (equiv- alent to the percentage point change in mode split). For exam- ple, an increase in transit mode share from 20 to 22 percent is a 2-percentage point increase, or an increase of 2 transit riders per 100 employees.

The key advantage of the mandatory CTR program data sets is the great level of detail, particularly with regard to the full range of commute programs offered at the worksite. For instance, the trip reduction program data sets show not only that an employer implemented a transit benefits program but also what other programs the employer has in place, as well as any other programs that were implemented or eliminated at the same time. Another advantage is the detailed data on mode shares; the trip reduction data provide information not only on transit use, but also on use of other modes, including driving alone, carpooling, vanpooling, bicycling, walking, and telecommuting. As a result, the data sets provide more detailed information on shifts in travel behavior. Finally, the research team believes the data sets are less likely to be biased toward worksites with successful transit benefits programs because the worksites were not surveyed directly by the tran- sit agencies, and the focus of the programs is on commute trip reduction, not transit ridership. On the other hand, there are reasons why worksites subject to mandatory CTR programs may not be representative of employers that typically implement transit benefits programs. The worksites subject to mandatory CTR programs each have 100 or more employees (250 or more in Southern California after the late 1990s) whereas anecdotal evidence suggests that in some metropolitan areas small companies have been more likely to offer transit benefits to employees. Moreover, work- sites mandated to have trip reduction strategies in place may have different motivations for making the transit benefits pro- gram succeed than employers that voluntarily introduce a tran- sit benefits program. Many of the worksites that implemented transit benefits in the mandatory CTR areas had low transit mode shares and were located in suburban areas that may not be well served by transit; in contrast, nationally, it appears that most transit benefits programs are implemented by worksites in downtown areas or other areas well served by transit. Many of the CTR records go back to the early 1990s and therefore may not reflect federal tax law changes, such as the option for employees to pay for their transit expense on a pre-tax basis (first allowed in 1998, after passage of the Transportation Equity Act for the 21st Century, TEA-21), greater tax-free lim- its (e.g., TEA-21 raised the tax-free limit for transit and van- pool benefits from $65 per month in 2001 to $100 per month in 2002), and increased general awareness of transit benefits programs. Finally, the data sets from the three areas also have varying degrees of quality control problems (e.g., dollar val- ues in the same field ranging from pennies to tens of thousands of dollars). Two approaches were used to analyze the data sets: • Regression analysis. The research team initially per- formed a regression analysis using the Southern Califor- nia data set to determine the influence of transit benefits and other financial incentive programs on employee travel behavior. The reported VTR, which represents the number of vehicle trips per 100 employees, was the dependent variable, with independent variables repre- senting each of the major types of financial incentives and nonfinancial transportation programs that can be offered at a worksite. The goal of this analysis was to identify programs that have a statistically significant effect on VTR. The results, however, did not reveal a statistically significant effect for transit benefits or most other finan- cial incentives, and some incentives showed an unex- pected positive sign (signifying that implementing a pro- gram actually increases VTR). The lack of significant variables may relate to a lack of several ideal determi- nants in the regression model and potential problems with the data set, or it could be that none of these incentives has a strong correlation with travel behavior. Results of the regression model are available in Appendix D. • Descriptive analysis. To produce results similar to those provided by the survey data from transit agencies and commuter organizations, the research team conducted a descriptive analysis of the data sets from all three regions, isolating worksite records in which an individual worksite either introduced or eliminated a transit benefit, vanpool benefit, or other financial incentive (such as carpooling incentives and parking cash out). The goal of this analy- sis was to examine changes in VTR and transit and/or van- pool mode split. Because the data sets were large, there were enough records to examine changes in each of these transit benefits programs. In this analysis, if a worksite implemented a benefit and later eliminated it, or vice versa, the worksite might be represented multiple times. For transit benefits only, the data records were further broken down. The first and largest set consisted of all records in which a worksite implemented a transit bene- fits program. A second set, a subset of the first, consisted only of worksites that implemented a transit benefits pro- gram without implementing or eliminating any other commute programs at the same time. The goal of analyz- ing this “control” subset was to isolate the effects of the transit benefits program from effects that might be occur- ring because of other programs. In the case of the large Southern California data set, the research team also looked at a third data set: worksites that implemented a transit benefits program in conjunction with supporting benefits (marketing and guaranteed-ride-home). The analysis approach and results for the mandatory CTR program data are described in more detail in Appendix D. Effects of Transit Benefits Programs on Employee Travel Behavior Transit Benefits Increase Transit Ridership in Most Circumstances Transit Ridership Generally Increases 10 Percent or More at Participating Worksites. The surveys suggest that implementing a transit benefits program typically results in 42

43 increased employee use of transit. Virtually every survey pro- vided evidence that transit use increased on implementation of a transit benefits program. However, the percentage increase in transit use varied dramatically among surveys, as shown in Figure 6. Note that for some surveys multiple figures were reported, representing different sets of worksites (e.g., CBD, urban, and suburban). Note also that only surveys that were performed before and after the implementation of transit ben- efits or those surveys that asked about previous commuting behavior are included in Figure 6. Other surveys did not con- tain sufficient information to determine the percentage change in transit ridership. More than half of the surveys reported an increase in tran- sit riders between 10 and 40 percent, and nearly one-quarter reported increases of more than 60 percent. Two surveys— one in San Jose in 1997 and one in Atlanta in 2003—suggest that transit ridership more than doubled after a transit benefits program was implemented. In contrast, the data sets from mandatory CTR program areas—Washington State, Southern California, and Tucson (Pima County, Arizona)—indicate very small changes in transit ridership on average, with a very slight decline in Tucson (Pima County, Arizona), and increases of only 3 percent in Southern California and 6 per- cent in Washington State. Understanding the Presentation of Results The effects of transit benefits programs are presented from three different perspectives: the change in the number of transit riders per 100 employees, the percent change in transit riders, and the share of transit benefits recipients who are new transit riders. An example of how these effects are related is the following: an increase in transit mode share from 8 to 12 percent is an increase of 4 transit riders per 100 employees (12 minus 8), or a 50 per- cent increase in transit riders (4 new transit riders divided by 8 at start), and signifies that 33 percent of transit benefits recipi- ents are new to transit (4 out of 12 total transit benefits recipi- ents). These metrics provide different perspectives on the data and are useful because different surveys are framed around these issues. In all cases, the range of results from the surveys and data sets is reported without calculating an average impact for all regions. This approach was chosen for two reasons: (1) the number of employees and employers in each survey varied considerably, which makes an average not entirely meaningful, and (2) the average is not applicable as a guide for what a “typical” region or employer can expect, as regional situations and individual worksite characteristics vary so widely. 0% 20% 40% 60% 80% 100% 120% 140% San Jose (1997) Atlanta (2003) Los Angeles (2001, UCLA) Portland, OR (2001) Harrisburg (1993) Denver (2003, Suburban) San Francisco (1994, Suburban) Portland, OR (1999) Philadelphia (1993 pre-test) Philadelphia (1993 final) Denver (2003, CBD) Pittsburgh (1993) Philadelphia (1993) Denver (2003, Urban Fringe) Denver (1993) New York (2004) San Francisco (1994, Urban) Minneapolis/St. Paul (2003) Philadelphia (2000) Washington State (1998-00) Southern CA (1991-93) Tucson (Pima Co., AZ) (1998-99) Percent Increase in Transit Riders Note: For the three CTR regions, the percentage changes reflect results for all worksites, not for the control group. These are more comparable to the survey results, since the surveys did not control for whether other benefits were implemented. Also, the figure for San Jose reflects the increase in the number of employees who reported riding transit (even occasionally), not the average daily mode share for transit. See Appendix C for details. Figure 6. Estimated percentage increase in employee transit use at participating worksites.

These differences likely reflect a variety of factors. Gen- erally, the largest percentage gains were in the less transit- intensive regions where starting transit mode shares were rel- atively low. For example, the 1997 San Jose survey and the 2001 Los Angeles area (UCLA) survey showed some of the highest percent increases in transit ridership (156 percent and 72 percent, respectively) and indicated lower starting transit mode shares (10.7 percent and 7.6 percent, respectively) than most of the other surveys. The smallest percentage gains (found in the mandatory CTR areas) were also associated with generally low initial transit mode share (3.5 percent in South- ern California, 4.6 percent in Tucson [Pima County, Arizona], and 14.0 percent in Washington State), but, in these cases, very minimal changes in transit use were found on average. In areas with relatively low starting transit mode share, moderate increases in the number of transit users can result in substantial percent increases in transit riders. As a result, the surveys with the largest percent increases in transit use did not always correspond to those that saw the largest increase in the number of transit riders (measured per 100 employees offered the benefit). There Is a Wide Range in the Number of New Tran- sit Riders per 100 Employees. More important than the percent increase in transit use is the actual number of new transit riders that can be expected at a particular worksite. Seven surveys conducted by transit agencies or other orga- nizations and the three mandatory CTR program data sets contained mode share data from before and after implemen- tation of a transit benefits program. According to these sur- veys, transit mode shares increased by nearly 2 to 17 per- centage points on average, meaning that a worksite with 100 employees that offers a transit benefit might expect the equivalent of 2 to 17 employees to switch to riding transit full-time. The data sets from the mandatory CTR program areas, however, reported on average less than 1 new transit rider per 100 employees. Figure 7 shows the average transit mode share reported before and after implementation of the transit benefit in the before and after surveys. Results for each individual worksite varied widely, with some showing no change or a small decline in transit use, and others show- ing large increases. Some Programs Primarily Serve Existing Transit Riders. The data from the mandatory CTR regions, on aver- age, showed transit benefit programs having little impact on transit use, signifying that, in most cases, the transit benefits program simply served existing transit users. It is important to note, however, that a transit benefits program that does not increase transit ridership, may be viewed as successful in other ways. For example, the increased convenience of receiving transit passes at the workplace and the lower costs of transit use to employees might help to support retention of existing riders and increased satisfaction with transit. The program also might be supporting increased off-peak use of transit by existing transit commuters (not captured in most surveys, which focus on commute travel). However, these effects could not be quantified in this study. 44 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% So u th er n CA (av g. 19 91 -93 ) Tu cs o n (P im a C o., A Z) (av g. 19 98 -99 ) W as hi ng to n St at e (av g. 19 98 -20 00 ) Lo s A ng el es (20 01 , U CL A) Sa n J os e (19 97 , d ow nto wn ) M in ne ap ol is / S t. Pa ul (20 03 ) Po rt la nd , O R (19 97 -20 01 ) W as hi ng to n, D C (19 93 ) D en ve r (20 03 ) D en ve r (19 93 ) + 5.5 + 16.7 + 1.8 + 15.0 + 7.8 + 8.0 + 0.1 + 0.9 -0.6 + 11.7 Before After Mandatory CTR Program Data Surveys from Transit Agencies and Others Figure 7. Average transit mode shares before and after implementation of a transit benefits program.

45 The research team explored the CTR data further by exam- ining a “control” group of worksites that implemented a transit benefits program with no change in the types of other commute programs—like reported availability of rideshare matching and telecommuting—being offered (the employer still might have made a change in the level of incentive or manner in which a program was implemented. This analysis simply looked at worksites that did not change their report- ing of whether or not a program was offered). The purpose of this analysis was to examine the independent effects of the transit benefit. All three CTR regions showed very small changes in tran- sit use on average for the “control” groups, and, in some cases, there were slight declines in transit use; these effects were the opposite of what the research team expected. The Southern California control group contained 57 records, and showed, on average, a decline in transit mode share from 4.5 to 3.4 per- cent. The Tucson (Pima County, Arizona) and Washington State control groups also showed small changes: an increase from 3.6 to 4.5 percent in Tucson (Pima County, Arizona) (sample size: three worksite records), and a decrease from 0.5 to 0.1 percent in Washington State (sample size: one worksite record). In addition, the research team looked at worksites in the three CTR regions where a transit benefits program was elim- inated with no change in other commute programs being offered, and the team similarly found, on average, small changes in the number of transit trips per 100 employees. Although transit mode share declined from 4.5 to 3.4 percent among the Southern California worksites that added a transit benefit with no change in other programs being offered, tran- sit mode share declined from 6.0 to 5.2 percent among South- ern California worksites that eliminated a transit benefit with no changes in other programs being offered. See Appendix D for more detail on these effects. There are several potential reasons why the CTR data showed much smaller effects than most of the other surveys. Differences in worksite characteristics may be an important factor. Several explanations that relate to worksite character- istics and program design were identified based on conversa- tions with staff from agencies supplying the mandatory CTR program data, including differences in transit availability, employer payment toward the transit benefit, and the range of available worksite commuting programs (8). (These factors are documented below in the section on “Factors Affecting Employee Travel Behavior Response.”) Moreover, differ- ences in survey design may also be a factor. For instance, in the mandatory CTR programs, an employer makes a commit- ment to implement transportation programs in its trip reduc- tion plan, but there is the possibility that the employer did not carry through on its commitment, particularly when there are multiple years between trip reduction reports. The three data sets from the mandatory CTR program areas also have vary- ing degrees of quality control problems, and the surveys that focus on transit benefits programs also have several potential problems, including low response rates in many surveys, which could indicate that responses are biased toward those individuals who changed their travel behavior. Individual Worksites Within a Region Differ. In addi- tion to recognizing the differences in results among the sur- veys, it is important to note that results for individual worksites varied even more widely. All of the information presented above represents the average impact found in each survey and is not representative of all worksites in a given region. For instance, data collected from Denver in 2003, which included before and after mode transit share data for employers that implemented an Eco Pass, show that average transit mode share increased from approximately 37.7 to 49.4 percent, implying that transit use typically increased by about 11.7 new riders per 100 employees. The figures were calculated using unpublished data provided as of April 2003 by the Denver Regional Transportation District for 37 employers and were developed by weighting mode shares for individual employers by the number of employees at the worksite (for example, mode share at a worksite with 100 employees is weighted 10 times as much as a worksite with 10 employees). As shown in Figure 8, there was a wide range of mode share changes among the individual worksites, from a small decrease to increases of over 30 percentage points. Among the majority of employers that reported an increase in transit use, the increases in transit share ranged from 1 to 45 percentage points, not counting one small employer (with three employees) that reported going from zero to 100 percent of employees using transit. Among the three largest employ- ers in the survey (each with over 1,000 employees), transit share increased by 11 percentage points (from 13 to 24 per- cent) at a suburban university campus, by 10 percentage points (from 77 to 87 percent) at a downtown financial services com- pany, and by 6 percentage points (from 23 to 29 percent) at a federal government agency in the suburbs. One mid-sized employer (with over 400 employees) in the CBD saw transit share increase by 34 percentage points (from 65 to 99 percent), showing that larger increases are possible. All of the Denver worksites that reported a decrease in tran- sit riders were small employers (each with 38 employees or less) where differences in survey response rate between the before and after surveys could have been responsible for the reduction. All of the worksites that reported no change were small employers with 100 percent of employees report- ing using transit, so no gains in transit use were possible. Another example of the wide range of results among work- sites can be seen in the information provided by Metro Tran- sit in Minneapolis on the six largest employers enrolled in the Metropass program (the information is based on surveys of ridership before and after implementation). Among these six employers, each of which had over 2,700 employees, the weighted average transit mode share increased by 1.8 per- centage points, from 17.0 to 18.8 percent. However, among individual employers, there were considerable differences, as

shown in Table 5. In general, the employers with the lower starting transit mode shares (approximately 11 percent and lower) saw little change in transit use; the employer with about 30 percent of employees taking transit saw a small increase; and the employer with the largest starting transit share saw a relatively large increase—nearly 12 new transit riders per 100 employees. The Southern California mandatory CTR data set is another case where the average impact was small (about 0.1 new tran- sit riders per 100 employees) but there was a wide variation in effect among worksites. As shown in Figure 9, about 1 out of 20 worksites (50 out of 943 worksites) in the Southern Cali- fornia data set saw transit use increase by 5 or more transit rid- ers per 100 employees after implementation of a transit bene- fits program. Overall, about 44 percent of worksites saw an increase in transit mode share, 40 percent saw a decline in tran- sit mode share, and 16 percent saw no change. Initial mode shares by groups are graphed in the U-shaped line in Figure 9. The largest changes in transit mode share—increases or decreases of over five employees per worksite—were associ- ated with higher initial mode shares. Worksites with no change had an average initial transit mode share of zero. Worksites in Tucson (Pima County, Arizona) and Washington State (not shown in Figure 9) showed a similar pattern: no or very low transit ridership before the introduction of benefits was asso- ciated with little or no change in ridership, and higher mode shares were associated with larger increases or decreases. See Appendix D for more detail. Employee Turnover and Other External Factors Affect Travel Behavior as Well. As Figure 9 illustrates, a large number of worksites in the Southern California data set actu- ally saw a decrease in transit mode share after implementation of a transit benefit. The same was true for the Tucson (Pima County, Arizona) and Washington State data sets. Other data sets that included individual worksite data, such as the data sets of Eco Pass employers in Denver and of Metropass employers in Minneapolis, also showed that some employers saw a decrease in transit mode share. There is no reason to believe that implementing a transit benefits program should result in a reduction in transit use, all else being equal. It seems likely that other factors must have been responsible for the reductions in transit use and may be partially responsible for some of the increase as well. 46 7 5 9 5 4 7 0 2 4 6 8 10 decrease no change 1 to 9% point increase 10% to 19% point increase 20% to 30% point increase 30+% point increase N um be r o f D en v er - ar ea E m pl oy er s Change in Transit Share Figure 8. Number of Denver-area employers by change in transit share (based on RTD survey data as of April 2003). Transit Mode Share Number of Employees Before After Increase in Transit Riders per 100 Employees Increase in Transit Ridership 5,535 56.2% 68.0% 11.8 21.0% 2,712 30.0% 32.0% 2.0 6.6% 14,123 10.6% 10.0% –0.6 –3.0% 4,942 7.7% 8.0% 0.3 4.0% 5,382 6.8% 7.0% 0.2 2.4% 4,815 4.0% 4.0% 0 –1.0% TABLE 5 Transit mode shares for the six largest employers participating in Metropass in the Minneapolis/St. Paul Area

47 In any before and after survey, there is the potential for a wide range of factors to influence the results, such as changes in the worksite employment profile (for example, if the com- pany hires a new set of employees or lays off employees between the “before” survey and the “after” survey), changes in transit services at the worksite (for example, if the transit agency implemented additional services or reduced service levels), or changes in transit fares or parking prices. These factors may be particularly important in influencing results for any surveys that are conducted many years apart. Moreover, if employee turnover or other factors might result in changes in mode share of plus or minus a few percentage points for any given mode, these changes may be very notable when starting mode share is very low. For example, if a worksite of 100 employees has 3 transit riders, and 1 rider leaves the com- pany to be replaced by an employee who drives, the overall transit mode share declines from 3 to 2 percent, resulting in a 33 percent reduction in transit use. In the Southern California trip reduction program data, 57 worksites that implemented a transit benefit without imple- menting or removing any other transportation benefits showed on average no impact in increasing transit ridership; in fact, on average, there was 1 fewer transit rider per 100 employees. Of these 57 worksites, over half (32) showed positive or negative differences within 1 percentage point, which may not reflect any real behavior change. Six worksites reported substantial decreases in transit ridership (a reduction of more than 5 tran- sit riders per 100 employees)—the reverse of what would be expected, all else being equal. Of these six worksites, two reported substantial increases in employment (one reported a more than 35-percent increase and the other reported a near doubling in employment) and one reported a substantial reduction in employment (a reduction of over 20 percent). Of the others, two worksites had trip reports several years apart, which may mean that the worksite did not continue to offer the transit benefit by the time the mode share impact was mea- sured, although this is unknown. As a result, it appears that employee turnover and other factors may explain the large drop in transit use. External factors similarly might explain some of the increases in transit use that occurred at other worksites. Overall Findings on Impacts on Transit Use. Overall, the results from the various surveys suggest that a transit ben- efits program can produce increased transit use in some cir- cumstances as well as an increase in new transit riders. It is important to recognize the context within which a transit ben- efits program is implemented. Many factors can produce a change in mode share and VTR. A transit agency can affect or control some of these (e.g., introduction of an incentive and changes in transit services), but not others (e.g., changes in gas prices and employee turnover). Various factors can affect employee travel behavior, regardless of whether or not a tran- sit benefits program is implemented. 37 141 199 154 195 167 50 17.9% 6.5% 1.9% 0.0% 1.5% 3.1% 10.0% 0 50 100 150 200 250 < 5% point decrease 1% to 5% point decrease 1% to 0% point decrease 0% to 1% point increase 1% to 5% point increase > 5% point increase Number of Worksites 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20% Starting Transit Mode Share zero change Figure 9. Number of Southern California worksites by change in transit share and initial transit mode share after implementing a transit benefit.

Transit Benefits Programs Attract New Transit Riders and Increase Use by Existing Riders Transit Benefits Programs Generally Attract Some New Transit Riders. As discussed above, in many regions transit benefits program result in new transit riders. Another way to examine transit benefits programs is to look at the portion of transit benefits recipients who are new riders. Based on sur- veys conducted by transit agencies and others, typically 10 to 40 percent of transit benefits recipients were new to transit, as shown in Figure 10. (Figure 10 includes only those surveys with information on the percentage of new riders.) The areas with large existing transit mode share, such as Philadelphia and New York, tended to have the largest share of recipients who were existing transit riders. The data suggest that even in very transit-intensive areas, new riders can still be added. Areas with relatively low starting mode share or very large increases in transit use, like San Jose and Atlanta, saw the largest portion of recipients who were new transit riders. In the mandatory CTR program areas, very little new transit rid- ership was reported, and so a very small share of transit bene- fits recipients were new to transit: approximately 6 percent in Southern California and 3 percent in Washington State. (The data from Tucson [Pima County, Arizona] showed no increase in transit use on average after implementing a transit benefit). Although the data from a 2001 State of the Commute Sur- vey in the Washington, DC, area is not displayed in Figure 10 because the wording of the question leaves room for other fac- tors, results of this survey indicated that 48 percent of Metrochek users were “influenced by” receiving the benefit, suggesting that up to 48 percent of Metrochek recipients are new riders. This figure is at the high end of the range show in Figure 10, but may reflect the fact that the tax-free limit rose to $100, and many Metrochek recipients are federal employ- ees who receive fully paid benefits. Alternatively, a substantial portion of the Metrochek users who say they were “influenced by” the benefit could be existing riders who use transit more frequently. Increased Transit Use Also Comes from Existing Tran- sit Riders. Because transit benefits programs often result in increased transit use, transit agencies and others are interested in examining which employees increase their transit use and what the patterns of increased use are. As part of this overall question, transit agencies and others are interested in discov- ering the extent to which increases in transit use are due to new transit riders or to existing transit riders who begin riding more often. Several surveys conducted by transit agencies and other organizations asked transit benefits recipients whether they rode transit more often after receiving the benefit. (There were no data addressing this issue from the three mandatory CTR regions.) As shown in Table 6, up to 35 percent of transit ben- efits recipients reported increasing their use of transit; this 48 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% San Jose (1997) Atlanta (2003) Los Angeles (2001) Portland, OR (2001) Harrisburg (1993) Montgomery County, MD (2001) Washington, DC (1993) Denver (2003) Philadelphia (1993 final) Philadelphia (1993 pre-test) Pittsburgh (1993) San Francisco (1994) Philadelphia (1996) Denver (1993) New York (2004) Minneapolis/St. Paul (2003) Philadelphia (2000) Washington State Southern California % New Riders % Previous Transit Riders Figure 10. Percentage of transit benefits recipients who are new transit riders, across transit agency and other surveys.

49 includes both previous riders who increased their frequency of use as well as new riders. Transit Benefits Recipients Ride More Often for Commute and Noncommute Purposes An important question for transit agencies and others is whether transit benefits programs result in increases in transit use for both commute and noncommute trips. This question is important because many transit agencies are very interested in increasing off-peak ridership in order to utilize existing capacity, particularly if the transit agency is at or near capac- ity during peak commute hours. One of the advantages of having employees get monthly or annual passes from their employers is that the pass can be used for any trip—weekday peak, weekday off-peak, and weekend. Although most studies of transit benefits’ impacts focus on commute travel, the four New York area surveys and the San Francisco survey asked about increases in transit ridership for both commute and noncommute trips. As shown in Table 7, all of these surveys suggest that employees took more transit trips for both purposes. It is notable that of the five surveys listed in Table 7, three showed increases in noncommute trips that were nearly as high as the increases in commute trips, and two surveys—New York 2004 and New York 1994—showed greater increases in noncommute trips. It is not clear to what extent these patterns hold in other regions; after all, New York and San Francisco have very high transit ridership. In these two regions, a large portion of transit benefits recipients were already using transit for commuting, and employees may have been more likely to consider transit as an option for other trips. In many smaller metropolitan areas, transit service is infrequent during off- peak periods and may not be as likely to attract as many new riders. Still, these figures suggest that in areas with high tran- sit ridership, transit benefits programs can be effective in encouraging increased transit use, and some of the increased use could be for noncommute trips. Six surveys asked specific questions about the number of new trips taken by people who reported increasing their tran- sit use. Table 8 provides a summary of the average number of new transit trips per week per employee receiving a transit benefit in these six surveys. Most of the surveys show an increase in transit trips for commute and noncommute purposes; however, they also show a wide range in average number of new transit trips—from 0.42 to 3.24 new transit trips per week. The lower increases tended to be in the New York region. Because the New York region tends to have far higher transit ridership than the rest of the country, we would not expect ridership to increase as much with the addition of a transit benefit. Also, in the 1989 and 1990 surveys, the average commuter benefit level was $15 per month, whereas in the 1994 survey, the average benefit was $45 per month. Thus, the larger increase in average number of new transit trips in the 1994 survey may be due to the higher level of benefit received. Note that these figures represent the average among all tran- sit benefits recipients, not just the recipients reporting increas- ing their transit use. Because the number of recipients report- ing increased transit use makes up less than one-third of all recipients in most cases (as shown earlier in Table 6), among the people who do increase their transit use, the actual number of new trips taken per week is several times larger (e.g., if one-third of employees reported increasing their transit use, and the average number of new transit trips per week per recip- ient is 0.42, this means that people who increased their use of transit typically added about 1.26 new transit trips per week [0.42 trips/week among all recipients × 3]). Transit Benefits Programs Reduce Vehicle Travel In Most Cases, the Majority of New Transit Riders Pre- viously Drove Alone to Work. Where did new transit riders come from? Of the many transit agency surveys and CTR data sets, 12 provided information on the percent of transit benefits recipients who are new riders and previous SOV commuters. Just over half of these surveys found that between 90 and 100 percent of new transit riders were previous SOV com- Region Survey Date % of All Recipients Who Increased Transit Ridership % of All Recipients Who Were New to Transit % of All Recipients Who Were Previous Riders and Increased Frequency Philadelphia 2000 35% 8.5% 26.5% Philadelphia 1996 32% 23% 9% Denver 1993 19.4% 15% 4.4% TABLE 6 Changes in frequency of transit ridership Region Survey Date % of All Recipients Who Increased Their Transit Ridership Commute Noncommute San Francisco 1994 34% 29% New York 2004 10% Over 24% New York 1990 22.7% 21.8% New York 1989 16.5% 14.0% New York 1994 11.0% 15.0% TABLE 7 Changes in frequency of transit ridership for commute and noncommute trips

muters, as shown in Figure 11 (note that in Figure 11, survey results from one survey conducted in Philadelphia, Pittsburgh, and Harrisburg are reported separately, but there was a small sample in each individual area). The Los Angeles survey is a somewhat unique case because it covered only one employer, UCLA, and the transit benefit provided was a universal pass on the Santa Monica Municipal Bus Line. Presumably, many of the people taking advantage of the transit benefit lived fairly close to campus and did not drive alone because of the park- ing shortage on campus. At UCLA, the drive-alone share was only 46 percent before the introduction of the pass, which helps to explain why only 31 percent of new transit riders previously commuted by SOV. Based on the surveys in Figure 11, which found that transit ridership increased by 2 to 17 riders per 100 employees and more than half of new transit riders previously commuted by SOV, one could anticipate a reduction of about 1 to 9 SOV users per 100 employees at worksites that implement a transit benefits program. Driving to Work Typically Goes Down. Data from sur- veys with mode share before and after implementation of a transit benefit confirmed that SOV commute mode shares had fallen, as shown in Table 9. For three of the surveyed regions, SOV commuting declined by at least 20 percent. In the Los Angeles area survey (UCLA), the share of drive-alone commuters for transit service area employees fell from 46 to 42 percent; however, carpooling/vanpooling declined more dramatically, indicating that more of the new transit riders likely came from carpools or vanpools rather than SOVs. In the three mandatory CTR regions, SOV commuting, as well as car- pooling and vanpooling, remained relatively stable on average after the introduction of transit benefits. 50 Region Survey Date Avg. # of Transit Trips/Week Avg. # of New Transit Trips/Week Before After Commute Noncommute Total New York 1990 – – 0.28 0.14 0.42 New York 1989 – – 0.31 0.14 0.45 New York 1994 – – 0.32 0.44 0.76 Denver 1993 6.6 7.8 1.20 – – Philadelphia 1993 7.8 10.3 – – 2.50 San Francisco 1994 – – 2.07 1.17 3.24 Dash = not available. TABLE 8 Increase in number of transit trips, average across all transit benefits recipients 96% 96% 100% 100% 96% 96% 73% 60% 56% 31% 29% 90% 0% 20% 40% 60% 80% 100% So u th er n Ca lif or ni a (av g. 19 91 -93 ) W as hi ng to n St at e (av g. 19 98 -20 00 ) Po rt la nd , O R (20 01 ) D en ve r (19 93 ) H ar ris bu rg (19 93 ) Sa n J os e (19 97 ) A tla nt a (20 03 ) Ph ila de lp hi a (19 93 b) W as hi ng to n, D C (19 93 ) Ph ila de lp hi a (19 93 a) Lo s A ng el es (20 01 , U CL A) Pi tts bu rg h (19 93 ) Figure 11. Percent of new transit riders who previously commuted by SOV.

51 The average impact, however, can mask wide variation among individual worksites. For example, in the Southern California data set of worksites subject to mandatory CTR programs, VTR went down on average from 80.3 to 79.1 after implementation of a transit benefits program. Without con- trolling for the introduction or removal of other incentives, about 58 percent of the worksites in Southern California reduced their VTR following the introduction of a transit ben- efits program, as shown in Figure 12. Nearly one-quarter (over 200 out of the 943 worksites) saw a reduction of more than 5 vehicle trips per 100 employees. All worksites either increased or decreased their VTR; there were none with zero change. The worksites that saw an increase in transit use after implementing a transit benefit nearly always saw a reduction in VTR. Overall, the worksites with the largest increases in transit share tended to see the largest reductions in VTR, as shown in Table 10. Factors Affecting Employee Travel Behavior Response As discussed in the previous section, in most regions transit ridership increased when transit benefits were introduced, but transit use did not increase in all circumstances for individual worksites. Most notably, the worksites in mandatory CTR areas on average showed considerably smaller impacts than other worksites surveyed by transit agencies and other organi- zations. A wide range of results have been reported, and vari- ous factors affect the level of impact. Transit agencies, commuter organizations, and others who promote transit benefits programs need to understand the fac- tors that influence travel behavior response in order to better target limited marketing resources to worksites that can see the most substantial impacts, to encourage practices that support transit use, and to better be able to gauge expected reductions in parking demand and traffic. Although data are limited, factors that appear to have the largest impact on employee travel behavior in response to receiving a transit benefit include the following: • Location and transit availability. Urban areas show greater overall increases in transit use; suburban areas with adequate transit services often show greater percent- age increases. No change in transit use may be expected in suburban areas with very limited transit services. • Level of employer payment. Employer-paid transit ben- efits show greater increases than pre-tax benefits, but the impact of the actual dollar amount paid by the employer is inconclusive. • Other worksite commute programs. Employers who implement transit benefits programs in conjunction with supporting programs, like marketing and guaranteed- ride-home, see greater increases in ridership than those SOV Mode Split % Change Carpool/Vanpool Split % Change Region Year Before After Change per 100 Employees Before After Change per 100 Employees San Jose 1997 75% 60% –15 –20% – – – – Portland, OR 1997–2001 60% 45% –15 –25% 16.0% 10.0% –6 –37.5% Denver 1993 40% 32% –8 –20% – – – – Los Angeles1 2001 46% 42% –4 –9% 16.0% 9.0% –7 –44.0% Southern CA2 Avg. 1992–1994 70% 68% –2 –3% 20.6% 22.3% 1.7 8.3% Tucson2 Avg. 1998–1999 76% 78% 2 3% 13.3% 12.8% 0.5 3.8% Wash. State2 Avg. 1998–2000 65% 64% –1 –2% 16.0% 16.0% 0 0% Dash = not available. 1 For Los Angeles (2001), calculations are based exclusively on the respondents living in the service area for the transit provider because overall figures were not reported. 2 For the three mandatory CTR data sets, the analysis results are reported for all worksites over the entire data set that implemented a transit benefits program regardless of changes in other programs. The results for worksites that implemented a transit benefits program with no changes in the types of other programs being offered were relatively similar. TABLE 9 Changes in SOV and car/vanpooling mode shares Large Increase (5+) Large Decrease (<-5) Modest Decrease (0 to -5) Modest Increase (0 to 5) 0 50 100 150 200 250 300 350 400 Change in VTR N um be r o f S o u th er n Ca lif or ni a W or ks ite s Figure 12. Number of Southern California worksites that implemented a transit benefit, by change in VTR.

who simply implement the transit benefit by itself. Work- sites that offer a large number of competing benefits are more likely to see smaller impacts on transit use. Although the type of transit benefits program offered by the transit agency (e.g., monthly pass, universal pass, or voucher) and the availability of free or subsidized parking at the work- site may affect employee travel behavior response as well, insufficient data were available to analyze their effects. How- ever, parking pricing is likely correlated with measures of worksite location and transit availability (e.g., suburban areas with limited transit service are most likely to offer free parking). Largest Increases in Transit Use Typically Occur in Urban Locations An important factor in increasing transit use is the avail- ability of transit services. Transit services tend to be most con- centrated in downtown areas, and worksites in these areas tend to have higher starting transit mode shares than worksites that are not as well served by transit. As a result, it is somewhat dif- ficult to separate out the roles of location, transit availability, existing transit ridership, and parking price in affecting employee travel behavior. Only two surveys, Denver (2003) and San Francisco (1994), provided information on the impacts of transit benefits pro- grams by geographic location within the region. Both studies indicate that there is strong potential for ridership growth in response to transit benefits in both urban and suburban areas. Both surveys showed a much more substantial percentage increase in transit ridership in suburban areas as compared with downtown/CBD locations; however, both surveys also suggest that a larger absolute number of new transit riders (per 100 employees) occurred in the downtown/CBD areas. The Denver RTD survey provides the most detailed data, tracking employer location based on three different service level areas (SLA): SLA A, well outside the Denver CBD, rep- resents suburban areas; SLA B, just outside the Denver CBD and including the Boulder CBD, represents urban areas, and SLA C represents the Denver CBD. As shown in Table 11, in the Denver area, the CBD worksites that implemented Eco Pass saw an average of approximately 16 new transit riders per 100 employees, for a 22-percent increase in transit use; how- ever, the suburban worksites saw an average of approximately 9 new transit riders per 100 employees, representing more than a 50-percent increase in transit use. This difference in percentage change and absolute increase in transit riders is not necessarily surprising because down- town areas have the best transit services, and it may be possi- ble for greater numbers of nonriders to switch to transit. The big difference is the starting mode shares—the very low starting transit mode share in suburban areas means that com- parable increases in transit ridership show up as very large percentage gains. It is also worth noting that these figures reflect occasional ridership, not daily ridership, as they include all Eco Pass users, regardless of the frequency of their ridership. Employees who ride transit even once per week are included. A typical daily mode split would proba- bly show somewhat lower figures. The 1994 San Francisco Commuter Check survey also found that suburban recipients showed greater percentage increases in transit ridership than their urban counterparts did, which may stem from the fact that existing transit mode share in suburban areas is low. The suburban areas saw a slightly larger increase in the average number of new transit trips per recipient (see Table 11); however, the suburban worksites likely had a much lower share of employees participating in Commuter Check than the urban worksites, so the actual num- ber of new transit trips per 100 employees would be higher in the urban area. Of course, at suburban worksites with little or no transit ser- vice, introducing transit benefits may have little or no impact on employee transit ridership, and worksite transit availability may be an important reason that the data from mandatory CTR programs showed little or no change in transit use on average. Employers seldom implement transit benefits unless they have reasonably good transit access and/or a contingency of exist- ing transit riders who can take advantage of the benefit; these characteristics often need to be in place in order for the employer to see value to implementing the benefit program 52 Change in Transit Mode Share Number of Worksites Average Change in Transit Mode Share Average Change in VTR > 5 point decrease 37 –9.3% 6.1 1 to 5 point decrease 141 –2.0% 2.9 < 1 point decrease 199 –0.3% –1.0 No change 154 0.0% –1.7 < 1 point increase 195 0.4% –2.6 1 to 5 point increase 167 2.2% –3.4 > 5 point increase 50 8.0% –6.5 TABLE 10 Change in VTR broken down by change in transit mode share at Southern California worksites implementing a transit benefit

53 (for more information on factors that affect an employer’s like- lihood of implementing a transit benefits program, see TCRP Report 87: Strategies for Increasing the Effectiveness of Commuter Benefits Programs [1]). However, in regions with mandatory CTR programs, employers sometimes implement benefit programs to demonstrate a “good faith” commitment toward reducing vehicle trips, even in cases where the benefit program may not be well matched for the worksite. Staff from SCAQMD, in particular, noted that a number of employers with little or no transit access developed transit benefits pro- grams to show an effort toward trip reduction goals, but these efforts would not be expected to succeed. Moreover, most of the trip reduction program data analyzed for this study was from the early 1990s, a time when there were considerably fewer transit possibilities available to employees than cur- rently. Staff at the Pima County Association of Governments noted similar issues. The worksites that implemented transit benefits programs in the mandatory CTR program areas were primarily those with very low starting transit ridership (less than 5 percent transit mode share), which is not the key market for transit benefits programs. In the cases of the Southern California and Tucson (Pima County, Arizona) mandatory CTR data, average start- ing transit mode shares were 3.5 percent and 4.6 percent, respectively. These mode shares likely reflect locations that are not conducive to transit; for example, worksites served by a limited number of transit lines and/or infrequent service, where carpooling, vanpooling, or flexible work arrangements tend to be more viable options. The Southern California trip reduction records show that of 943 worksites that implemented a transit benefits program, only 32 were located in downtown Los Angeles; the rest were in more suburban locations (of the 57 worksites that implemented a transit benefits program with no other changes in the types of programs being offered, only one was located downtown). The CTR areas saw on average very little increase in tran- sit use, which suggests that in circumstances with very limited transit service, it is difficult to achieve increases in transit use. In cases where transit services are limited, large increases in transit use are unlikely. On the other hand, in cases where the worksite is well suited to transit, implementing a transit bene- fit can result in substantial increases in transit use. As noted earlier, the CTR data from Southern California suggested that the worksites that had little or no change in transit use after implementation of a transit benefit were the ones that had no, or very few, existing transit riders, whereas those that had sub- stantial increases were those with the highest starting transit mode shares. Employer-Paid Transit Benefits Increase Transit Ridership More than Employee-Paid Pre-Tax Benefits Very little research has been conducted on differences in the effects of employer-paid and employee-paid, pre-tax pro- Transit Mode Split Location Before After Increase per 100 Employees % Increase in Number of Riders % of All Recipients Who Are New Riders % All Recipients Who Increased Their Transit Ridership # New Trips/ Week Denver (2003) SLA C (CBD) 72.5% 88.7% 16.2 22.2% 18.2% – – SLA B (Urban Fringe) 63.0% 74.1% 11.1 17.6% 15.0% – – SLA A (Suburban) 17.4% 26.5% 9.1 52.7% 34.5% – – Average (All Employers)1 37.7% 49.4% 11.7 31% 23.6% – – San Francisco (1994) Urban – – – 14% 13% 25% 3.03 Suburban – – – 40% 29% 48% 3.74 Average (All Employers) 1 – – – 21% 17% 31% 3.24 Dash = not available. 1 The average is calculated on the basis of data for all employers; because there are different numbers of employers in each category, the average is weighted towards the larger categories. TABLE 11 Change in transit ridership by location—Denver RTD and San Francisco surveys

grams on traveler response. Even though employee-paid, pre- tax programs can result in tax savings, there are reasons to believe that employer-paid programs are more effective in encouraging increased transit use. Notably, employer-paid programs are easier for employees to understand and can be easier to access, particularly if the employer provides transit passes to all employees (such as through use of a universal pass). Employee-paid programs require enough of a commit- ment to transit for the employee to set aside his or her own income on a pre-tax basis; the employee actually receives less money in his or her paycheck but receives the convenience of getting a transit pass from the employer and saves taxes. As a result, it is expected that employee-paid programs might sup- port increased use of transit by existing transit riders, whereas employer-paid programs might do a better job of encouraging new transit riders. The 2004 survey of commuters in the New York metropol- itan area asked current drive-alone commuters if they would switch to transit at various hypothetical benefit levels. At a pro- posed $50 employer-paid benefit per month, 40 percent of drive-alone commuters said they would switch to transit. At a proposed tax savings of $33 per month from a pre-tax benefit program (which requires reserving $100 per month on a pre- tax basis), 37 percent of drive-alone commuters said they would switch to transit. The differences in these figures are negligible, but may reflect that the survey respondents were focusing on the dollar savings given a hypothetical situation. In a real-world setting, other factors come into play, such as whether employees understand how much they will save in taxes through a pre-tax program. Only two surveys contained real-world information on whether the transit benefits provided to employees are employer-paid or pre-tax: Philadelphia (2000) and Portland (1999). (The reports from the mandatory CTR regions did not track this. The pre-tax option became available only in 1998; all of the surveys conducted prior to that date reflect only employer-paid benefits, and several of the later surveys also include only employer-paid benefits. The Los Angeles (2001) survey of UCLA’s BruinGo program and the Montgomery County, Maryland (2001) survey of worksites participating in the county’s programs (FareShare and SuperFareShare) include only employer-paid commuter benefits). The 2000 Philadelphia survey compared the percentage of transit bene- fits recipients whose employers paid the transit benefit with the percentage of transit benefit recipients whose employers offered employee-paid, pre-tax transit benefits in three cate- gories: benefits recipients who increased their transit ridership, benefits recipients who were new transit riders, and benefits recipients who increased the number of transit trips they took per week. The 1999 Portland survey examined transit ridership increases and compiled results based on whether the employer paid. As shown in Table 12, for each of these measures, an employer-paid transit benefit produced a greater increase in transit ridership than a pre-tax benefit by a fairly substantial percentage. Notably, in the Philadelphia (2000) study, over three times as many employees reported being new transit riders with the employer-paid benefit (13.2 percent) than those employees with employee-paid pre-tax (3.8 percent). These figures imply that employer-paid transit benefits can attract more new tran- sit riders, although the survey does not report on several other factors that may also influence the results, such as the location of the worksites or starting mode shares. The survey also does not report whether those employees shifted from SOV com- muting or another mode and does not explain the difference between increasing “transit ridership” and increasing the “number of transit trips per week.” Transit ridership increased among employers with both employer-paid and employee-paid, pre-tax benefits recipients in the Portland (1999) survey, but the increase was greater among employers with employer-paid benefits. Much of the increase among employee-paid, pre-tax benefits recipients may be due to occasional transit riders who set aside money on a pre-tax basis and, as a consequence, ride more frequently, rather than people who are totally new to transit; however, no information is available to confirm this. Increased Employer Payment May Have Larger Impact—Data Are Inconclusive Although one might expect that an increased employer con- tribution would yield a greater increase in transit ridership, 54 paid paid Average for All Recipients1 Philadelphia 2000 % of all recipients who are new riders 13.2% 3.8% 8.5% % of all recipients who increased their transit ridership 42.0% 23.0% 35.0% % of all recipients who increased the number of transit trips per week 12.6% 8.1% – Portland 1999 % Increase in transit ridership 34% 24% – Region Year Measure Employer- Employee- Dash = not available. 1 These figures represent the average for all transit benefits recipients (both those receiving employer-paid benefits and those receiving employee-paid pre-tax benefits). These figures are not the average of the individual figures for employer-paid recipients and employee-paid recipients since there are a different number of recipients in each of these categories. TABLE 12 Comparison of employer-paid and employee-paid pre-tax transit benefits recipients

55 survey results are inconclusive on this issue. Three surveys directly address the question of whether the level of the employer-paid benefit affects travel behavior; two examine actual behavior, and one examines a hypothetical situation. Only one of the two behavior-based surveys, Portland (1999), suggests that the amount the employer pays affects the extent to which employee transit use increases, as shown in Table 13. The Portland (1999) study divided results into three bene- fit levels: no paid benefit (pre-tax only), 40- to 60-percent employer-paid benefit, and 90- to 100-percent employer-paid benefit. The survey also looked separately at employers par- ticipating in the PASSport program, a universal pass typically fully paid by the employer. The percentage of transit ridership increased as the benefit level rose, and the PASSport employ- ers showed the highest average increase in transit ridership. These results should be viewed with some caution, however, because the location of the employers is not known, and loca- tion could have a bearing on the percentage change in transit ridership. The 1994 San Francisco survey, on the other hand, found no correlation between the level of the employer-paid commuter benefit and the percent of employees reporting an increase in transit work trips. Accounting for urban/suburban location, $20 Commuter Checks showed just as much impact in employee mode shifts to transit as $30 Commuter Checks, suggesting that the fact that an employer offers a benefit has a much greater effect on transit ridership than the level of the benefit. The study authors speculate that most recipients induced to take new tran- sit trips were relatively infrequent riders who plan to ride only once or twice per week and for whom, therefore, the difference between a $20 and $30 benefit would be negligible. Larger dif- ferences in the employer payment, however, may make a dif- ference. The 1993 Government Accountability Office (GAO) study of federal employees in the Washington, DC area and elsewhere asked about current transit ridership, based on a $21 commuter benefit, and whether employees would switch to transit if they received a $60 benefit. Results suggested that the mode split for transit could increase from 31 to 49 percent. However, additional research on this issue is needed, given that this was a hypothetical analysis. Differences in the level of employer payment may be partly responsible for differences in results found in the mandatory CTR areas in comparison with the other surveys. Most of the data available from the mandatory CTR program areas was from the early to middle 1990s, a time when the tax-free limit for transit benefits was considerably lower than under current law. In 1992, the Energy Policy Act expanded the definition of qualified transportation fringe benefits to include transit and vanpool benefits, and imposed a tax-free limit of $60 per month on these benefits; prior to 1992, tax- free transit passes were limited to a de minimus level (up to about $21 per month). In 1995, the tax-free limit increased to $65 per month. In 2002, the tax-free limit increased to $100 per month. The average employer payment in the Southern California data set, one of the mandatory CTR areas, was only $28 per month. It is understandable that such a low employer payment might not encourage a notable increase in transit use, partic- ularly in an area with limited transit services. In contrast, many of the surveys conducted by transit agencies and others were conducted at worksites where the employee received a high-value transit pass. Seven of the surveys were conducted in regions with universal pass programs, a program in which an employee receives an annual transit pass that is typically employer-paid. Implementing a Transit Benefit with Supporting Benefits Results in Greater Impact Implementing supporting programs, like a guaranteed-ride- home program and on-site marketing, can result in larger Region Year Level /Value of Commuter Benefit Increase in Transit Ridership % Employees Reporting Increase in # of Transit Work Trips % Employees Using Transit or Saying Likely to Ride Transit Portland 1999 No benefit 24% – – 40–60% paid 31% – – 90–100% paid 46% – – PASSport (universal pass– usually 100% employer-paid) 57% – – San Francisco $20 per month – 35% – 1994 $30 per month – 30% – Over $30 per month – 38% – 1993 Existing $21 per month – – 31%Washington, DC and elsewhere Proposed $60 per month – – 49% Dash = not available. TABLE 13 Comparison of level of employer-paid commuter benefit

impacts on transit use than simply implementing a transit ben- efit on its own. A guaranteed-ride-home program, also some- times called an emergency-ride-home program, helps support transit use because it helps employees get over the fear of being stranded in the event of unexpected overtime or a fam- ily emergency that would require the employee to leave work during noncommute hours. A guaranteed-ride-home helps employees to set aside one of their biggest concerns about using transit and can be particularly important in locations where transit services to a worksite run solely during commute hours. On-site marketing, through transit fairs and other events, also helps to support a transit benefits program by mak- ing employees who may not have used transit in the past more aware of available transit services and how they operate. The Southern California mandatory CTR program data set contained sufficient records to separate out cases in which the only change in a benefits package was implementation of a transit benefit from cases in which a transit benefit was imple- mented along with supporting programs. The analysis sup- ports the theory that implementing transit benefits in conjunc- tion with supporting benefits—namely, internal marketing programs and guaranteed-ride-home programs—seems to be more effective than implementing them alone (see Table 14). In the cases in which transit benefits were implemented alone, with no other change in commuting programs, transit mode share on average actually declined; however, when the transit benefit and supporting programs were implemented together, with no other change in commuting programs, transit mode share increased by 10 percent. Follow-up conversations with staff at agencies responsible for CTR programs supported this theory; staff said that they had seen “synergy” among benefits programs. However, the analysis does not control for other factors, such as location, so it is not clear if these two groups of employers are comparable in all respects. Figure 13 shows the breakdown within these two groups based on the number of employers who saw increases, decreases, or essentially no changes (−1% to 1% change) in transit ridership. For the group of employers without sup- porting programs, the pattern follows a bell curve, with the most frequent response being no change (−1% to 1% change) in transit ridership. For the group with supporting programs, the number of employers that saw modest increases (1% to 56 Transit Mode Share Impact of Implementing Number of Employers Before After % Point Change (Change per 100 Employees) % Change in Transit Ridership Transit benefits without supporting programs1 57 4.5% 3.4% –1.1% –24% Transit benefits with supporting programs in the same period1 23 4.9% 5.4% 0.5% 10% 1 The worksites could be offering a variety of other competing or supporting programs. The focus of this analysis is on the change in programs being offered from one period to the next. TABLE 14 Average changes in transit mode split and transit ridership for Southern California employers implementing transit benefits programs with and without supporting programs Transit Benefit Implementation with No Other Programs 6 7 34 8 20 5 10 15 20 25 30 35 < 5% point decrease 1% to 5% point decrease -1% to 1% change 1% to 5% point increase > 5% point increase 0 6 8 8 1 0 5 10 15 20 25 30 35 < 5% point decrease 1% to 5% point decrease -1% to 1% change 1% to 5% point increase > 5% point increase Transit Benefit Implementation with Supporting Programs Figure 13. Number of Southern California worksites by category of change in transit mode share, with and without supporting programs.

57 5% increase) was the same as the number of employers that saw no change (−1% to 1% change). There were no employ- ers in this group with large (more than five percentage points) decreases in transit ridership. This seems to indicate that implementing transit benefits with supporting programs had some, albeit small, impact in these circumstances; if they had had no impact, we would expect a random distribution, as we see in the first group. Impacts of Competing Transportation Programs Are Unclear Although it appears that supporting programs can help boost transit ridership, some employer-based commuting programs, like ridesharing or telework programs, may also compete against transit benefits programs. More programs may mean more competition for a finite market of potential “switchers.” For example, if a worksite only offers the tran- sit benefit, there may be a net increase in the number of peo- ple using transit and a reduction in vehicle trips. If, on the other hand, the employer offers both a transit benefit and a telework program, the worksite might see fewer employees take advantage of transit because some employees who would have switched to transit choose to telework (although the worksite might see a larger total reduction in driving trips). Thus, transit benefits programs may be more effective at increasing transit use when there are fewer other “com- peting” incentive programs; however very limited data on this factor were available. The availability of competing programs may be partially responsible for the small increases in transit use that were found on average in mandatory CTR areas. Because worksites in mandatory CTR areas are required to reduce employee trip making, many of these employers offer a large number of transportation program options for their employees. Options such as ridematching, flexible work hours, compressed work weeks, and telecommuting may encourage employees to switch to these options and reduce the impact of the transit benefit in terms of increasing transit use. For example, in the Tucson (Pima County, Arizona) data set, all 21 worksites that implemented a transit benefits program also offered rideshare matching services, and over half offered telecommuting. In contrast, in the areas surveyed by transit agencies and others, it is likely that the employers did not promote options like ridesharing and telecommuting as aggressively, and transit benefits were more likely to be viewed as the primary trans- portation benefit. On the other hand, among Southern California worksites that implemented a transit benefits program with no other change in programs being offered, those that saw the largest increases in transit mode share did not differ greatly from others in terms of the types of programs being offered. This finding suggests that other factors were probably more impor- tant than the existing benefits profile. Effects of Program Design Inconclusive The available data do not reveal whether the type of employer program (i.e., universal pass, monthly pass, or voucher) affects the extent to which the program brings in new transit riders. As noted above, only one of the surveys— Portland (1999)—provided any comparison of results from a monthly pass program and a universal pass program. The figures suggest that the universal pass program may pro- duce a greater increase in transit ridership than a monthly pass program. However, no information is available on the locations of the employers or other factors that might affect the ridership response. Although some of the largest impacts on transit use were reported in regions with universal pass programs, comparing results across different surveys should be done with caution, given that only a few types of programs are represented and the locations of the worksites being analyzed differ. As noted earlier, most of the surveys available are in areas with universal pass programs (7 surveys) and voucher programs (12 surveys), with monthly pass programs barely represented (see Table 4). The locations of the surveys may also have a large effect on results; we would expect that the level of tran- sit service and existing transit use would affect the extent to which programs are able to generate new riders. Four of the surveys come from the New York metro area, which has a very different transit profile than any other metropolitan area in the United States. Impacts of Vanpool and Other Financial Benefits on Employee Travel Behavior The three mandatory CTR regions provided data on the impacts of vanpool benefits and other financial benefits, which include transportation allowances; parking cash-out programs; and financial benefits to bicyclers, walkers, or carpoolers. These benefits programs generally showed sim- ilar patterns as the transit benefits programs—relatively small changes on average in relevant mode shares (vanpool mode share for vanpool benefits and transit, carpool/vanpool, bicycling, and walking mode share for other financial ben- efits). The worksite records are characterized by a wide range of effects, with some worksites showing reductions, and others showing increases in these mode shares. For instance, in Southern California, both introducing and elim- inating financial benefits (with no changes in the types of other worksite programs being offered) were associated with increases in carpool mode share, and it may be that other factors—such as an enhanced high-occupancy vehicle network throughout the region—had an impact on encour- aging ridesharing. These findings suggest that the effects of these transit ben- efits programs vary considerably at different worksites based on specific worksite factors (e.g., location, employer commit- ment, and types of other programs being offered) or perhaps

that the effects of these programs are overshadowed by other external factors (e.g., employee turnover and changes in transportation costs). Additional detail on the findings on van- pool and other financial benefits from CTR areas is included in Appendix D. Comparison of Study Findings with Other Literature These findings on the travel effects of transit benefits are generally consistent with previous research on the effects of employer-based TDM programs. The literature on the factors that affect transit mode share and employee travel behavior is too extensive to be completely reviewed here, but the follow- ing briefly compares several of this study’s findings with con- clusions from other literature. Employer-Provided Transit Benefits Usually Increase Employee Transit Ridership and Reduce Driving to Work Transit benefits, whether pre-tax or employer-paid, lower the cost of transit. It is thoroughly consistent with the litera- ture, and with economic theory, to find that decreased transit costs increase transit ridership. The finding from this research that SOV use typically declines by up to 20 percent after implementing a transit benefits program is firmly within the range of effects reported in the literature on the potential of employer-based TDM programs. For instance, another TCRP project (B-4, “Cost-Effectiveness of TDM Strategies”) evalu- ated some 50 employer-based TDM programs in the United States and estimated that the average reduction in vehicle trips among these “successful” programs was 15.3 percent (9). A synthesis of TDM experience for the U.S. DOT concluded that “with the right mix of strategies, a TDM program at individ- ual employment sites could reduce vehicle trips by as much as 30 to 40 percent. . . . In almost all cases, however, one major conclusion stands out—some level of incentive or disincentive must be present to encourage automobile users to change their travel behavior” (10). In 2001, EPA analyzed the effects of commuter benefits using its Commuter Model, a tool designed to estimate the travel and emissions effects of employer-based TDM pro- grams based on findings from the TDM literature (11). Although the model is not specifically designed to analyze commuter benefits programs, the model was run using a lower price for transit and vanpools to simulate the way that com- muter benefits programs reduce employee transit and vanpool costs. According to the model results, a $20-per-month, employer-paid benefit shows a 2.8- to 4.6-percentage-point reduction in SOV use (depending on starting mode share), and a $40-per-month, employer-paid benefit shows a 7.3- to 10.5-percentage-point reduction in SOV use. These effects are consistent with the research findings based on the surveys con- ducted by transit agencies and other organizations. Although the average subsidy in most of the surveys is unknown, it is likely that the average would be $65 or under because that was the tax-free limit when most of the surveys were conducted. Transit Availability and Urban Location/ Design Influence the Effectiveness of Employer-Based Programs The literature generally supports the research finding that transit availability and land use patterns are important factors in the effectiveness of transit benefits programs. An analysis of the effects of land use and TDM strategies on commuting behavior, relying on SCAQMD data, found that land use and urban design characteristics influence mode choice and the effectiveness of TDM strategies. The data revealed that “when financial incentives are present, the greatest reduction in the drive alone share is realized in areas with an aesthetically pleasing urban character . . . This appears to be a result of the availability of alternatives modes (e.g., transit service) and the quality of the environment” (12). Employees Are More Likely to Increase Transit Use with Employer-Paid Benefits Most of the literature used to estimate the effects of incen- tive programs on travel behavior is based on pricing studies (transit fare prices or parking prices) and does not directly address the question of who pays. This study suggests that it may be an important factor if the employer pays, which is con- sistent with literature that generally finds that employer com- mitment to a program is an important determinant of employee travel response (see, for example, the work of Weber, Nice, and Lovrich [13]). It may be that making a financial commit- ment sends a powerful message to employees, and, that once that message is sent, the absolute amount is somewhat less important. Data Gaps on Travel Impacts In terms of understanding travel impacts, the research revealed several gaps in knowledge, as well as observations regarding how these impacts are tracked. These gaps and observations are discussed below. Relatively Little Information Has Been Collected on Travel Impacts in Many Regions Although this report compiled data from 21 surveys and 3 mandatory CTR programs, the total number of regions represented is only 14 (Southern California is represented in both data sets). Several major metropolitan regions where transit benefits are available did not have available survey 58

59 data (e.g., Boston, Houston, and San Diego), and many mid- sized regions that might be able to shed light on how effec- tive transit benefits are in less transit-intensive environments also did not have available data. Quality of Survey Data Is Uncertain Thirteen surveys provided information on their response rates. These ranged from 8 to 63 percent, with both an aver- age and a median of 38 percent. (The mandatory CTR regions did not have information available on their response rates but generally require at least an 85-percent response rate.) These rates are not sufficiently high to ensure that the surveys accu- rately reflect the behavior of all employees. For example, in surveys of all employees, it is possible that employees who receive transit benefits and ride transit are more likely to respond than drivers because they see the topic as more per- tinent. This would introduce bias into the survey responses and skew the results toward suggesting higher transit rider- ship than actually exists. This bias would probably constitute less of a problem in surveys in which the only employees sur- veyed are those receiving transit benefits. Surveys Do Not Provide Comparable Information Many of the surveys, because of the respondent pool and the types of questions asked, provided relatively little usable infor- mation. The research team contacted several agencies that col- lected some information via surveys, but the information was ultimately deemed unusable because there were too many gaps, because the questions asked were not germane to the research problem, or because the information was provided only anecdotally with no supporting evidence. Of the surveys that were incorporated into this report, several asked hypo- thetical questions (which are generally felt to provide less valuable information than behavioral questions). The discrep- ancy in survey design makes it difficult to compare results across regions. Additional Information Is Needed on Several Factors That Influence Travel Response Although the available data do provide an indication of the factors that influence the level of travel response, additional data would be helpful to provide stronger evidence about the impacts of these factors in different circumstances. In particu- lar, more information is needed on the following: • Effect of program design. The survey results we were able to obtain are primarily from areas with universal pass and voucher programs. Only one of the surveys— Portland (1999)—provided any comparison of results from monthly pass programs and universal pass pro- grams. It would be desirable to obtain more detailed data on employers that have implemented different types of programs (i.e., monthly passes, universal passes, and vouchers) in specific metropolitan areas in order to com- pare the impacts of programs across different types. • Employer location and transit service levels. Only two surveys allowed an examination of the impact of transit benefits programs by employer location (e.g., downtown or suburb). It would be helpful to have more detailed information so that impacts could be compared among, as well as within, regions. It would also be useful to have better data to understand the role of transit service levels (e.g., a suburb with very good transit service versus a suburb with limited transit service). • Employer-paid versus employee-paid, pre-tax impacts. Only two regions collected data from both employers offering employer-paid benefits and employ- ers offering employee-paid, pre-tax benefits. This is due in part to the fact that many of the surveys reviewed were conducted before 1998, the year when legislation was passed that allowed pre-tax benefits. However, even among more recent surveys, data are not always col- lected on whether the employer contributes to the cost of transit benefits. Although results confirmed the hypoth- esis that employees are more likely to switch modes with employer-paid benefits, this finding would be far more robust if it was supported with data from multiple areas. Overall, it is notable that relatively few transit agencies have conducted surveys or evaluations to assess the impacts of their transit benefits programs on transit ridership and vehi- cle travel. The transit agencies that were most likely to have conducted surveys were those with universal pass programs because surveys often play an important role in determining the price paid by the employer. However, for other transit agencies and organizations that play a role in promoting tran- sit benefits programs, surveys can play a valuable role in determining the effectiveness of the program in meeting goals such as increased transit ridership, reduced vehicle travel, and reduced parking demand. IMPACTS ON TRANSIT AGENCIES’ RIDERSHIP, REVENUES, AND COSTS The second component of this research focuses on how tran- sit benefits programs affect transit agencies in terms of rider- ship, revenues, and costs. This section builds on the research conducted on employee travel behavior impacts and addresses the following questions: • How much systemwide ridership and revenue come from transit benefits programs? The share of overall ridership and revenues that comes from employer pro- grams affects the extent to which these programs can help

retain and attract riders and yield cost savings to the tran- sit agency. • Do transit benefits programs increase transit rider- ship and revenues? Research on the impacts of transit benefits programs on employee travel behavior (dis- cussed earlier) suggests that transit benefits programs can increase transit ridership. This section explores the extent to which transit ridership and revenues increase and how program design affects revenues per rider. • How much do transit benefits programs cost to administer? Administration costs include staff time for employer outreach as well as marketing and other costs. • Are there differences in revenue, ridership, or cost characteristics among different program types? If dif- ferent types of programs (e.g., universal passes or monthly passes) generate different levels of revenues per rider and have different costs, it is useful for transit agen- cies to understand these effects so that they can offer the program options that best meet their agencies’ goals. • How do transit agencies rate the success of their tran- sit benefits programs? Answers to these questions are designed to help transit agencies and others (1) assess whether employer transit ben- efits programs are effective in attracting riders and revenue and what an effective program costs, as well as (2) determine which programs would be most suitable for meeting their goals and objectives. Data Sources and Approach The results summarized in this section are drawn from interviews conducted with seven transit agencies selected to provide a range of modes, geographic areas, ridership, and employer programs. (Although an eighth agency, Capital Metro in Austin, was also interviewed, the program was far smaller than the others and comparable data for many pro- gram attributes was not available. Therefore, data findings are not presented in the body of the report, although the pro- gram is described in Appendix F.) The seven transit agen- cies whose interview results are included in this study are the following: • Washington Metropolitan Area Transit Authority (WMATA), Washington, DC; • Metropolitan Atlanta Rapid Transit Authority (MARTA), Atlanta, GA; • King County Metro, Seattle, WA; • Regional Transportation District (RTD), Denver, CO; • Metro Transit, Minneapolis/St. Paul, MN; • Santa Clara Valley Transportation Authority (VTA), San Jose, CA; and • Valley Metro, Phoenix, AZ. The agency locations also correspond with places where survey data on travel impacts were available. The interviews were conducted using an interview guide, and interviewers asked follow-up or clarifying questions when necessary. In some cases, the persons interviewed sent additional informa- tion following the interview. A copy of the interview guide is available in Appendix E, and case study write-ups of the tran- sit agency programs are in Appendix F. Background information on the seven agencies such as loca- tion, modes, service area population, and other characteristics is provided in Table 15. The transit agencies are diverse in terms of region (representing the Mid-Atlantic, Southeast, Midwest, Mountain West, Southwest, and West Coast) and in modes (three are bus-only agencies, two have light rail, and two have heavy rail). Average weekday ridership ranged from just over 100,000 to over 1 million, and all types of transit benefits programs are represented: universal pass, monthly pass (dis- counted and nondiscounted), and stored-value card/voucher. Effects of Transit Benefits Programs on Transit Agencies Many Agencies Offer Multiple Types of Transit Benefits Programs Although the research team anticipated that most transit agencies would have only one transit benefits program in place, of the seven interviewed, four had multiple programs. Types of employer programs offered included monthly passes, stored-value cards, universal passes (for more information on universal pass programs see “Unlimited Access” [14]), and vouchers (which can be traded in for transit fare media or use on vanpools). Generally these situations have evolved in response to employer demands and available technology. As Table 15 shows, three of the seven agencies have only one employer program, and King County Metro has seven. Table 16 provides additional information on one of the more complex types of programs in terms of pricing, the universal pass. Universal pass programs are generally defined by three elements: (1) they function as an annual pass (valid for a full year of service); (2) they are priced based on a requirement that passes be purchased for all employees; and (3) the price of each individual pass is deeply discounted, based on the recog- nition that not all employees at the worksite will actually use transit daily. However, in practice, such programs vary con- siderably and do not always follow these prescriptions. Transit Agencies Generally Track Program Participation Employer Participation. The number of employers par- ticipating appears to be one of the measures agencies track most closely. Most employer programs serve several hundred employers, although the figures varied widely among transit 60

tisnarT ycnegA emaN aerA ecivreS sedoM noitacoL noitalupoP 1 egarevA yadkeeW pihsrediR 1 eraF launnA seuneveR 1 margorP emaN margorP epyT raeY nageB 000,003,3 sub ,liar yvaeH CD ,notgnihsaW ATAMW 1 eulav-derotS kehcorteM 000,000,573$ 000,003, rehcuoV/drac 1 399 eulav-derotS stifeneB tramS cinortcelE/drac rehcuov 0002 sub ,liar yvaeH AG ,atnaltA ATRAM 1 $ 054,035 000,003, 101 ATRAM 000,003, pihsrentraP margorP ssap ylhtnoM emulov htiw tnuocsid 1 299 ylno suB AW ,elttaeS 1 ssap lasrevinU ssaP xelF 000,004,87$ 000,043 000,057, 1 399 ytnuoC gniK orteM ssap lasrevinU ssaPU 1 099 ssap lasrevinU ssaPoG 1 799 tnemngisnoC ssaP liateR ssap ylhtnoM 1 779 liaM/enohP margorP – ssap ylhtnoM retummoC rehcuoV sunoB rehcuoV 1 599 sulP sunoB srehcuoV sdraweR margorp 1 699 ,74$ 000,072 000,003,2 sub ,liar thgiL OC ,revneD DTR 1 ssap lasrevinU ssaP ocE 000,00 1 99 1 /silopaenniM tisnarT orteM NM ,luaP .tS ylno suB 1 deifidoM ssaP orteM 000,000,56$ 000,442 000,006, ssap lasrevinu 1 899 – ssap detnuocsiD !skroWtisnarT sub ,liar thgiL AC ,esoJ naS ATV 1 000,007, 1 – ssap lasrevinU ssaP ocE 000,007,53$ 000,68 ylno suB ZA ,xineohP orteM yellaV 1 000,053, 1 rof "drac tiderC" sulP draC suB 000,006,82$ 000,70 sub 1 99 1 – ssap ylhtnoM teltuO etavirP .elbaliava ton = hsaD 1 002 ,esabataD tisnarT lanoitaN eht morf era serugif llA 1 .tnemucoDnepO?ataDDTN/scoD/fsn.emohdtn/DTN/moc.margorpdtn.www ta elbaliava , TABLE 15 Summary characteristics of transit agencies examined and their employer programs 61

62 agencies. WMATA’s Metrochek program is the largest in terms of employer participation, with over 3,300 employers participating. Table 17 provides figures on employer partici- pation, the approximate number of transit benefits recipients per employer, and the percentage of employers who pay for the benefit or offer employee-paid, pre-tax deductions. The average number of participating employees per employer varies widely among transit agencies and pro- grams. The lowest is at Valley Metro, with approximately 35 transit benefits recipients per employer on average. The highest is for the university programs at King County Metro, with 6,000 riders per participating employer. The universal pass programs tend to have several hundred employees using the benefit per employer (on average, 250 for Flex Pass, 210 for Metropass, and 490 for the VTA Eco Pass), which sug- gests that universal pass programs typically serve large employers. Most other programs have 20 to 100 employees on average participating per employer. Who Pays for the Benefit. The share of employers that fully cover the cost of the transit benefit appears to vary con- siderably among agencies, although many agencies do not collect this information. Four of the seven agencies inter- viewed had complete or partial information on whether their Understanding the Presentation of Results Although information is summarized for each transit agency, in general, the reader is cautioned against comparing ridership, rev- enues, and costs among different transit agencies because the tran- sit agencies and the environments within which they operate dif- fer so greatly. Although the data presented in this report are useful as benchmarks, they are a snapshot in time for individual agencies. It should be recognized that the sample size is small, and a range of factors could affect these metrics for individual transit agencies. Transit Agency Name Program Name Purchase Requirements Universal Pass Cost (Annual per Employee) Regular Pass Cost Employer Requirements on Sharing Cost FlexPass All employees $50 to $400, based on zone; may be determined individually for non- zone or employers over 500 employees; incentives provided in first 1–3 years $396 to $1,584, depending on zone and peak vs. off-peak Employer must pay at least 50% of costs King County Metro UPass1 Interested employees/ students Approximately $280 ($70 per quarter, sold on a quarterly basis) $396 to $1,584, depending on zone and peak vs. off-peak Students pay $35/quarter; Faculty/staff pay $48.96/quarter; the University of Washington pays rest RTD Eco Pass All employees $50 to $228, depending on service level area and # of employees $1,050 (Regional Valupass) None2 Metro Transit Metropass Interested employees $756 ($63 per month) for participating riders; new riders can be added without additional costs during the year $504 to $1,140 (based on monthly fares of $42 to $95) None2 VTA Eco Pass All employees $7.50 to $120, depending on area and # of employees $577.50 (regular) or $990 (express) None2 1 Both the UPass and Go Pass programs allow participation by students and employees. Only employees are eligible for the tax benefits associated with a transit benefits program because students are not considered employees under the tax code. However, the research team used data available on these programs to the fullest extent possible, separating students from employees when that information was available. 2 “None” regarding requirements on cost-sharing means that the transit agency does not require employers to pay a minimum dollar amount for their participating employees. In theory, the employer could ask employees to pay any portion of the cost of the universal pass, up to the full cost. In reality, most universal pass costs are provided by the employer, especially when the employer is required to purchase passes for all employees. RTD and VTA do not track this for participating employers; information on Metropass employers is available in Appendix F. TABLE 16 Universal pass program comparison

63 participating employers paid for transit benefits or whether they allowed employees to pay using pre-tax dollars. The data show wide differences in the share of employers paying for the full cost of the benefit. In the Metropass program in Minneapolis, the vast majority of participating employers offer a pre-tax program, and only 12 percent pay the full cost. MARTA has a more equal proportion of employer-paid to employee-paid, pre-tax benefits, although the agency was not certain about what some share of employers offer. WMATA reported that 55 percent of employers offer fully employer- paid benefits, and this high share is explained in part by the fact that federal executive agency departments are required to pay the entire cost. King County Metro’s Flex Pass reported the highest percentage of employers fully subsidiz- ing the program, 75 percent (this agency requires employers to pay at least half the cost of the pass). King County Metro also provides employer incentives during the first few years, which may accustom employers to paying the full cost; fur- thermore, large employers in the region are subject to manda- tory CTR requirements. Employer Participation Trends. For most transit agen- cies interviewed, the trend in the number of employers partic- ipating has been upward, even with the downturn in the econ- omy in the early years of this decade. Four of the seven transit agencies provided data on employer participation over time, which is displayed in Figure 14. These data show that employer participation in the King County FlexPass has grown steadily, whereas employer participation in Metro Transit’s programs plateaued for a year, and then continued to increase. The number of employers participating in transit benefits pro- grams at both RTD and VTA has fallen from previous highs. RTD staff attributed the decline to fare increases and changes in policy that made it less attractive for small employers to participate (the average number of transit benefits recipients per employer for the RTD Eco Pass is still relatively small at 50 employees, compared with many other universal pass pro- grams). According to VTA, the largest factor in the decrease in employer participation has been the poor economic climate. However, RTD’s and VTA’s total number of employees participating in their employer programs has increased % of Employers Who Transit Agency Name Program Name Number of Participating Employers Approx. # Recipients per Employer1 Pay Full Amount Pay Portion (Combination Benefit) Pre-tax Only Metrochek 3,349 55 – – – SmartBenefits 623 30 – – – WMATA Total 3,972 502 55% 10% 35% MARTA Monthly pass Over 3003 100 20%4 20%4 30%4 Flex Pass Over 2003 250 75% 25% UPass and GoPass 8 campuses 6,000 0% 100% 0% Retail programs 700 to 800 15 – – – Voucher programs 540 – – – – King County Metro Total 1,400 to 1,500 1005 – – – RTD Eco Pass 1,041 50 – – – Metropass 72 210 12% 18% 70% Metro Transit TransitWorks! 515 20 – – – Total 587 452 – – – VTA Eco Pass 87 490 – – – Bus Card Plus 331 35 – – – Private Outlet 198 – – – – Valley Metro Total 529 – – – – Dash = not available. 1 Calculated based on average number of employees who use the transit benefit at least occasionally, divided by number of participating employers. 2 This figure is representative of the entire transit agency program and not a sum of the figures from the component programs listed above. 3 Some of the MARTA and King County Metro contracts represent more than one employer (such as sales to a transportation management association that distributes passes to multiple employers). 4 Percentages do not sum to 100 because MARTA was uncertain about what 30 percent of the employers offer. 5 Total employees per employer excludes voucher programs. TABLE 17 Data on participating employers (as of 2003)

steadily, demonstrating that a decline in employer partici- pation does not necessarily mean fewer transit benefits recipients. The number of employees receiving transit bene- fits depends on the size of employers participating (in terms of numbers of employees) and the share of employees who par- ticipate at those worksites (which may depend on the level of employer payment, type of program design, or other factors). Transit Benefits Users Can Make Up a Substantial Share of System Ridership Employee Participation. Employees participating in tran- sit benefits programs make up a substantial portion of total transit ridership for many transit agencies. For the agencies interviewed, the percentage of all riders using employer tran- sit benefits programs was estimated by the transit agencies at between 5 and 25 percent. The total number of employees receiving transit benefits through an employer program ranged from 12,000 to over 200,000. The highest percentages of tran- sit riders who participate in employer-sponsored transit bene- fits programs were at WMATA (approximately 25 percent of transit riders), Valley Metro (about 22 percent), and King County Metro (20 to 22 percent of riders). WMATA is the largest transit agency in terms of total daily ridership and attracts a large number of federal employees who receive full employer-paid benefits. Valley Metro, in Phoenix, is the small- est of the seven transit agencies interviewed in terms of total systemwide ridership, but has the largest number of staff working in employer outreach (including rideshare programs), so the program’s success may stem in part from this intensive effort. Table 18 provides figures on ridership for each transit ben- efits program and the percent of total system riders using transit benefits. Employee Participation Trends. Employee participation in transit benefits programs has been increasing for nearly all of the agencies that provided historical participation trends. Even when employer participation has declined or plateaued, employee participation has consistently increased. Five agencies had trend information on the number of employees participating in transit benefits programs; this information is graphed in Figure 15. Three of the agencies offer universal pass programs, which track the number of employees at partic- ipating employers. While generally not all universal pass recip- ients ride transit, the figures assume that all of King County’s UPass program employee participants ride transit, since stu- dents, faculty, and staff are allowed to opt out of the program. Most striking in the employee participation trends is the large jump in participation in WMATA’s transit benefits pro- gram between 2000 and 2001. Two substantial reasons for the large increase between 2000 and 2001 at WMATA were the increase in the tax-free limit from $65 to $100 and implemen- tation of an Executive Order signed by President Clinton that requires federal government agencies to fully pay for transit benefits up to the tax-free limit for all interested executive branch employees in the Washington, DC, region. VTA, MARTA, and RTD have shown much steadier increases in employee participation over time. VTA and MARTA reported being affected by economic downturns, and all three had fare increases (or in the case of MARTA, a reduction in the employer discount that made employers’ costs higher). The strong employee participation figures seem to indicate that 64 0 200 400 600 800 1000 1200 1400 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 King County Flex Pass RTD Metro Transit VTA Figure 14. Number of employers participating in transit benefits programs, 1992–2003.

65 the programs are fairly resilient in the face of financial obsta- cles for employers. Participation in King County’s UPass has been quite steady, but the program only serves the University of Washington, and, therefore, it may have reached its satura- tion point among potential recipients. Transit Benefits Programs May Have Contributed to Ridership Growth Most transit agency staff involved in the transit benefits pro- grams believed that these programs have contributed to rider- ship growth. Given a scale from “significant” to “no impact,” most of the transit agencies characterized their transit benefits program as having a “significant” effect on peak-period rider- ship and a “minor to moderate” or “moderate” effect on off- peak ridership. (See Table 19.) The agency staff who market these programs clearly feel that the programs are not only serv- ing existing transit riders but are also encouraging additional transit trips. The agencies, however, often did not have con- crete data to assess impacts at a quantitative level, and staff involved in managing the program are personally invested in it and may not be able to accurately judge ridership impacts at a systemwide level. Moreover, the agencies’ assessments of ridership increases do not necessarily match the percentage of total riders using transit benefits. For instance, WMATA, with the largest absolute number and percentage of riders participat- ing in the transit benefits program, rated the program only “moderate” in increasing ridership, whereas many of the pro- grams with much smaller shares of ridership through employer programs rated their programs as “significant” in increasing ridership. Although a larger number of employees participating in a transit benefits program does not necessar- ily indicate that the transit benefits program increased rider- ship by a larger margin, the participation level does provide some indication of the maximum level by which the program might increase ridership. Differences in perceptions about impacts may reflect different expectations of what the pro- gram is meant to accomplish. Transit Agency Program Name Number of Participating Employees % of All Riders Using Employer Passes1 Metrochek 189,067 – Smart Benefits 18,933 – WMATA Total 208,000 25%2 MARTA Partnership Program 30,700 <10%2 Flex Pass 38,000 to 40,000 (est.) 6% to 8% UPass and GoPass 48,6003 >10% Retail programs 10,000 to 14,000 (est.) 3%4 Voucher programs – – King County Metro Total 95,000 to 103,000 20% to 22% RTD Eco Pass 52,700 (est.)5 12% to 21%6 Metropass 15,000 7% TransitWorks! 12,000 5% (est.) Metro Transit Total 27,000 12% (est.) VTA Eco Pass 42,800 (est.)7 5% Bus Card Plus 12,189 11% Private Outlet 12,000 (est.) 11% Valley Metro Total Over 24,000 22% Dash = not available. 1 Estimated by transit agency staff, unless otherwise noted. 2 Estimated based on National Transit Database ridership figures for FY 2001. 3 UPass ridership is lower during summer quarter; approximately 26,000. 4 Estimated based on King County Metro staff estimates for other programs. 5 Estimated ridership based on survey figures showing that 67 percent of eligible employees participate (see survey write-up in Appendix C). 6 14% of bus riders, 12% of light rail riders, and 21% of skyRide riders. (No numbers were given, so an overall total could not be estimated). 7 Estimated ridership based on survey figures showing that 36.4 percent of eligible employees participate (see survey write-up in Appendix C). TABLE 18 Employee participation in transit benefits programs (as of 2003)

66 Note: WMATA participation figures were estimated based on revenues; see case study in Appendix D for details. MARTA data estimated based on number of annual cards sold. 0 50000 100000 150000 200000 250000 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 N um be r o f E m pl oy ee s WMATA VTA RTD King County UPass MARTA Figure 15. Trends in employee participation in transit benefits programs at five agencies. Impacts on Ridership Transit Agency Name Program Name Peak Off-Peak Overall Metrochek WMATA Smart Benefits Moderate No impact Moderate MARTA Partnership Program Significant Moderate Significant Flex Pass Significant Minor to moderate Significant UPass and GoPass Significant Minor to moderate Significant Retail programs Significant Significant Significant King County Metro Voucher programs Staff characterized program as contributing to maintaining ridership RTD Eco Pass Significant Moderate Moderate Metropass Significant Moderate Significant Metro Transit TransitWorks! Significant Moderate Significant VTA Eco Pass – – Moderate Bus Card Plus Valley Metro Private Outlet Moderate Minor Moderate Dash = not available. TABLE 19 Transit agency perceptions of ridership impacts of transit benefits programs

67 It is difficult to develop quantitative estimates of the extent to which the transit benefits programs have affected overall transit ridership at agencies over time because it is impossible to state what ridership trends would have been if such pro- grams were not in place. As noted earlier, four transit agen- cies interviewed—WMATA, MARTA, RTD, and VTA— provided trend data on employee participation in transit benefits programs. Based on data on total transit system rider- ship from the National Transit Database (NTD) and available survey data on the share of transit benefits recipients who are new to transit or who increased their transit use, an estimate of the contribution of the transit benefits program to total system ridership can be developed. Estimates for these agencies sug- gest that the transit benefits programs may have been respon- sible for a substantial portion of ridership growth between 1997 and 2001 (the most recent year for NTD data on rider- ship). It should be noted, however, that limitations in survey data (i.e., small sample sizes, low employee response rates, and surveys that were conducted many years in the past) cre- ate a high degree of uncertainty in these estimates. For WMATA, there was a noticeable increase in overall transit ridership in 2001—a 118-percent increase—which cor- responds with the steep increase in the number of employees participating in the transit benefits program. Over the period of 1997 to 2001, the number of weekday rides on WMATA ser- vices increased by nearly 187,000, whereas the number of transit benefits participants increased by 127,100 (between 2000 and 2001, overall weekday riders increased by about 130,000, while commuter benefits participants increased by about 65,000). Assuming that about one-quarter of transit ben- efits recipients in the Washington, D.C., area are new riders, based on the 1993 GAO survey of federal employees, and that the average recipient might take up to two transit trips per day, this suggests that perhaps up to about 60,000 new transit rid- ers over this period can be traced to the transit benefits pro- gram. If this were the case, the transit benefits program may have accounted for about 34 percent of the ridership growth. However, the survey data may not reflect the actual ridership patterns of transit benefits recipients over the 1997-to-2001 period. The more recent State of the Commute survey (Washington, DC, 2001) found that approximately 48 per- cent of people who use Metrochek say that they “were influ- enced by” it, which could mean a number of things, from rid- ing transit more often to continuing to stay on transit (not switching to driving alone); this survey also includes non- WMATA riders (e.g., riders on suburban bus services). The results of this survey may indicate that with up to $100 per month available now, an even higher portion of Metrochek users are new riders or more frequent riders. At RTD, the number of employees participating in the Eco Pass program increased from 1997 to 2001 by approximately 25,400, whereas overall ridership during that period increased by 29,600 rides per day. The ongoing RTD survey of employ- ees at employers participating in the Eco Pass program (Denver, 2003) suggests that 24 percent of all recipients are new transit riders. As a result, the employer program may have accounted for about 6,000 new riders per day, or, assuming up to two transit trips per day, up to nearly 42 percent of the over- all growth. At VTA, the gains may have helped contribute to an increase in ridership. Between 1997 and 2001, the number of weekday rides on VTA services increased by about 13,000 trips, whereas the number of Eco Pass participants increased by about 26,400 (the number of VTA Eco Pass par- ticipants was estimated based on the total number of employ- ees eligible for Eco Passes [based on the employee population working for participating employers] multiplied by .364—a VTA survey showed that 36.4 percent of eligible employees hold Eco Passes). A VTA survey of employees at six partici- pating employers (San Jose, 1997) found that about 61 percent of Eco Pass recipients are new transit riders. As a result, the employer program may have accounted for about 16,000 new riders. However, several factors make this estimate fairly uncertain: the small sample size of the 1997 survey (only six employers), the expansion of both light rail and bus service from 1997 to 2001, and the strong employment during that period. So although the Eco Pass program may be one of sev- eral factors responsible for the overall growth in VTA rider- ship, it is difficult to say which factors were most important. For MARTA, an assessment of the impact of the transit benefits program could not be made because the data on tran- sit pass program participation provided by the agency cover the years 2001 to 2003, whereas the data on overall system ridership from NTD are currently only available up to 2001. Transit Benefits Programs Can Make Up a Substantial Portion of Revenues Total revenues associated with employer sales can be a sig- nificant portion of total transit agency revenues. As shown in Table 20, the percentage of total agency revenues associated with employer sales for the seven agencies examined is esti- mated to range from 5 to about 40 percent of total revenues for each transit agency. Metro Transit and King County Metro report the highest shares of revenues from employer sales, 42 percent and 35 to 41 percent, respectively. WMATA fol- lows with about 30 percent of total revenues coming from its employer programs. These are significant shares of total rev- enues, which may have implications in terms of the efficiency of distributing fare media and reducing the costs of individual transactions. Overall, revenues tend to be related to the size of the transit agency and costs of fare media. Four of the seven transit agencies reported that they believe their transit benefits programs increase revenues, whereas three of the agencies felt that the programs have a neutral or unclear impact. The agencies reporting neutral or unclear impacts are all agencies with universal passes, where the cost of the passes is discounted to employers and often is designed so that the employer does not pay more than it would to cover

existing transit riders. In contrast, to the extent that a monthly pass program increases the number of employees using tran- sit, it should result in increased revenues. For stored-value card programs, an increase in the number of employees using tran- sit or an increase in the frequency of transit use by existing rid- ers should result in increased revenues. For all of the programs with data on revenues (provided by the transit agency or developed by the research team based on data from the NTD or the transit agencies), the estimated share of transit agency revenues from the transit benefits pro- gram equaled or exceeded the share of system ridership from the program. These figures suggest that employer programs are not losing potential revenue. Although in most cases, the share of ridership and revenues was similar, in a few cases, the estimated share of revenues far exceeded the estimated share of ridership. The largest differential—an estimated 25-percent share of revenues and only an estimated 7-percent share of ridership—was from the Metropass program in Minneapolis/St. Paul. This disparity is somewhat surprising because the Metropass program is designed to be revenue neutral. However, there are several possible explanations for the disparity: (1) a portion of riders receives discounted fares (i.e., older people, students, and people with disabilities); (2) fares within the CBDs of Minneapolis and St. Paul are 50 cents, as compared to the usual $1.25 local fare; (3) employer programs are geared toward full-fare paying commuters, who often travel longer distances and pay higher fares; (4) some employees may sign up for the program because it is generally inexpensive for them, but they do not ride very frequently; or (5) there are differences in the data reporting between the ridership and revenue figures, so these figures are not totally comparable. Transit Benefits Program Costs for Agencies Can Vary Considerably The costs associated with operating and marketing a transit benefits program for employers were estimated based on the transit agencies’ estimates of staff time and other resources, such as marketing and fulfillment budgets. Table 21 provides a summary of these figures for the seven transit agencies interviewed for this study. It also provides estimates of costs as a portion of revenues from the program, and annual costs per rider, analyses which ideally could be used to assess how efficient these programs are in compari- son to other marketing efforts. Given limited data, however, such comparisons could not be made. Each of the major com- ponents of agency costs associated with transit benefits pro- grams are described below. Staff Time. Staff time differed greatly between programs, from 1 FTE at MARTA to 5.2 to 6.6 FTEs at King County Metro (staff requirements change throughout the year). The number of staff is not correlated with ridership or revenues; rather, the number of staff required to administer a single pro- gram appears to be tied most directly to program type. With one exception (the King County UPass program), regardless of ridership or revenues, universal pass programs seem to require a minimum of 2.5 staff. The RTD Eco Pass program has 3.6 FTEs, but it handles far more employers (over 1,000) than the other universal pass programs (which enroll several hundred employers). Presumably the number of staff required for universal pass programs is due to the complexity of these programs; compared with monthly pass programs, universal pass programs require more time with employers, more sur- 68 Transit Agency Program Name Annual Revenue in $ Million % of Revenue from Program Agency’s Perception of Impact on Revenues Metrochek $177.0 Smart Benefits $13.8 30% Increase WMATA Total $190.8 30% MARTA Partnership Program $20.0 11% (est.) 1 Increase FlexPass $6 to $7 8% to 10% UPass and GoPass $10.7 14% Retail programs $9 to $12 13% to 17% Voucher programs $6.7 2 N/A Increase King County Metro Total $25.7 to $29.72 35%s to 41% RTD Eco Pass $8.1 17% Unclear Metropass $15.1 25% Neutral TransitWorks! $10.0 17% (est.) Metro Transit Total $25.1 42% (est.) VTA Eco Pass $1.7 5% Neutral Valley Metro1 Bus Card Plus $3.6 N/A Increase 1 Only the Bus Card Plus program is included here because information was not available for the Private Outlet program. 2 Commuter Bonus Vouchers not included in total because they may be spent on other fare media, which could result in double counting. TABLE 20 Estimated revenues associated with transit benefits programs (as of 2003)

69 veys, and more frequent repricing. Less complex programs seem to require fewer staff. With the exception of Valley Metro, monthly pass programs used one to two FTEs. Broadly speaking, staff members operating a universal pass program serve fewer employers, but they serve more employees. King County Metro’s retail programs, Metro Transit’s TransitWorks! program, and Valley Metro’s Bus Card Plus cover from 80 to 375 employers per FTE and from 3,000 to 6,000 employees per FTE. In contrast, King County Metro’s FlexPass, RTD’s Eco Pass, and VTA’s Eco Pass cover only 30 to 300 employers per FTE, but these programs cover between 14,000 to 17,000 employees per FTE. This seems to point to different strategies depending on the employer pool: Many small employers may be served more efficiently with a monthly pass program, but a universal pass program can reach more employees through large employers. The exceptions are MARTA and Metropass. In the case of MARTA, the discount structure makes it more attractive to large employers because there is no discount available until an employer purchases 1,000 passes. Metropass is unusual in that it does not require employers to purchase passes for all employees, so it probably achieves lower penetration into the potential employee market. See Appendix F for additional figures on the number of employers and employees served per FTE. The two most “efficient” programs in serving the largest number of riders with the smallest number of staff are proba- bly not widely replicable. The King County Metro UPass pro- gram, which serves students, faculty, and staff at the Univer- sity of Washington, has upwards of 40,000 participants, yet requires less than one FTE for administration. This is proba- bly because (1) it is a long-established program with an employer motivated to promote ridership, and (2) the large market of potential users includes not only employees but also students (who pay for the program partially through student activity fees). WMATA has four staff for over 200,000 partic- ipants, making it highly efficient in serving both employers and staff. There are probably two reasons for this high level of efficiency. First, other organizations in the region assist heav- ily in marketing efforts, including the Metropolitan Washing- ton Council of Governments’ regional Commuter Connections program and Commuter Connections employer representa- tives in each of the local jurisdictions. Second, the federal gov- ernment’s executive branch is required to provide transit ben- efits to all its employees, and federal employees account for three-quarters of participating employees. Transit Agency Program Name Staff Time (FTE) Marketing Budget Other Costs Total Estimated Costs1 Costs as % of Revenue Annual Costs per Rider Metrochek Smart Benefits 4 $300,000 Not specified – – – WMATA Total 4 $300,000 $510,000 0.3% $3 MARTA Partnership Program 1 $0 2 $83,000 0.4% $3 FlexPass 2 to 3 Under $5,000 $142,000 2.4% $4 UPass and GoPass .2 $0 $14,000 0.1% <$1 Retail programs 2 $0 $115,000 1.1% $10 Voucher programs 1 to 1.4 $0 Not specified $81,000 1.4% N/A King County Metro Total 5.2 to 6.6 Under $5,000 $352,000 1.2 to 1.3% $3 RTD Eco Pass 3.6 $25,000 $18,500 (fulfillment) $293,500 2.4% $6 Metropass 2.25 $87,500 $225,000 $312,500 2.1% $21 TransitWorks! 2 $0 $150,000 $150,000 1.5% $13Metro Transit Total 4.25 $87,500 $375,000 (salaries) $462,500 – $17 VTA Eco Pass 2.5 $26,550 $240,000 (salaries) $266,550 11.1% $6 Valley Metro3 Bus Card Plus 4 $04 Not specified $360,000 10.0% $30 Dash = not available. 1 Includes staff time, marketing, and fulfillment. Staff time was calculated based on figures of $47,250 per staff FTE and $67,250 per managerial FTE. These figures include salary and benefits and rounded up to the nearest thousand dollars. In all cases, we assumed one manager per separate program and the remainder staff. 2 Marketing for the Partnership Program is part of overall transit marketing budget; exact figures not available. 3 Only the Bus Card Plus program is included here because information was not available for the Private Outlet program. 4 General marketing budget of $650,000, but not for these programs. TABLE 21 Estimated costs associated with transit benefits programs (as of 2003)

Marketing. Marketing budgets also covered a wide range, from no separate budget to $300,000; some agencies did not have a marketing budget for transit benefits broken out sepa- rately from general transit marketing. The power of a transit agency’s marketing budget can be stretched depending on other partners in the region. All seven agencies had other pub- lic- or private-sector entities helping market transit benefits to employers. The budget differences may be explained by targeted versus general marketing strategies, effectiveness of specific campaigns, and general awareness of transit bene- fits within a region. The differences also may also be due to agencies defining their budgets differently. Fulfillment. Most transit agencies, when asked about a ful- fillment budget, said that they considered fulfillment part of the salaries paid to employees and did not have separate figures available. Only three agencies had separate budget items for fulfillment, ranging from $18,500 to $375,000. Of those three, two included salaries in their figures. Several agencies men- tioned related costs such as printing and software, but they could not provide specific figures. Cost Savings Have Not Been Quantified One of the potential advantages of an employer transit ben- efits program for transit agencies is the potential to reduce the costs associated with cash handling for individual fare trans- actions. Although the transit agencies generally felt that some of these cost savings might be achieved through their pro- grams, none of the agencies was able to quantify these savings or supply a per-transaction cost of accepting cash payments. Therefore, the agencies’ responses in Table 22 are the best information available, and those impressions may or may not accurately reflect the magnitude of the cost savings. To the extent that employer programs capture a large share of total transit agency revenues, it is expected that these programs should reduce the costs associated with cash handling for indi- vidual fare transactions. Several agencies commented that they believed annual pass programs were useful in holding down costs because they reduce the number of passes to be printed and distributed per year. However, the agencies did not have comparative data for annual and monthly passes. Metro Tran- sit thought that the TransitWorks! program did not reduce transactions because participants would have been participat- ing in monthly pass programs anyway as opposed to individ- ual daily cash transactions. Two agencies said that specific programs reduced cash han- dling to a high degree. King County Metro made this comment in regard to their monthly pass programs, which sell approxi- mately 46,000 passes per month to employers and to retail out- lets who sell them to individuals. Most passes are distributed through retail outlets, and employers can participate on gener- ally the same terms as grocery and drug stores that sell them to patrons. WMATA said the same about its Smart Benefits program, in which transit benefits can be downloaded directly by the employee onto a stored-value card. Both of these pro- grams reduce pass distribution costs. Ridership, Revenues, and Costs Differ by Program Type Ridership, revenues, and costs differ across agencies. It is also interesting to note some general differences between uni- versal and monthly pass programs, both of which are fairly common program types. Table 23 compares selected program characteristics from the three conventional universal pass pro- grams (King County Metro’s FlexPass, RTD’s Eco Pass, and VTA’s Eco Pass) and the three conventional monthly pass pro- grams (MARTA Partnership Program, King County’s con- signment retail program, and Metro Transit’s TransitWorks! program). In this comparison, it appears that universal pass programs are more effective than the monthly pass programs at serving a larger number of employees by focusing on larger employers. However, in relation to the monthly pass pro- grams, the universal pass programs often require more staff to administer, are more complex, and are generally designed to be revenue neutral. In contrast, monthly pass programs are more effective than universal pass programs at increasing rev- enues and reaching many employers, but they tend to serve a lot of small- to moderate-size employers. These results generally reflect program design; universal pass programs are generally designed to appeal to larger employers and achieve greater ridership gains by requiring that passes be given to all employees. The comparison con- firms the effectiveness of this strategy and perhaps points to dif- ferent approaches based on the types of employers to be served. Universal pass programs seem to make sense for large employ- ers located where there is existing transit capacity. Monthly pass programs favor smaller employers and are more effective in bringing in revenue per rider. It would be useful to confirm these conclusions with employee survey data to see if there is a difference in the percentage of employees who switch modes based on program type; however, such data are not available. 70 Transit Agency Program Name Reduces Cash Handling? WMATA Metrochek Moderate Smart Benefits High MARTA Partnership Program Moderate King County Metro FlexPass Moderate UPass and GoPass Moderate Retail programs High Voucher programs Moderate/ high RTD Eco Pass Moderate Metro Transit Metropass Moderate TransitWorks! Not at all VTA Eco Pass Low Valley Metro Bus Card Plus Moderate TABLE 22 Transit agency perceptions of the extent to which transit benefits programs reduce cash handling

71 The differences associated with different types of pro- grams may indicate that agencies can combine universal pass and monthly pass programs to reach a wider variety of employers. Both King County Metro and Metro Transit offer both universal passes and a monthly pass program, and they receive the highest proportion of revenues through employer programs (over 40 percent). However, the pro- portion of their ridership that comes from transit benefits recipients is in the middle of the range for this group of agencies (18 to 22 percent and 12 percent, respectively). Given that neither transit agency operates a rail system, and that neither system is located in a dense and transit-rich East Coast city, this may point to an effective strategy for transit agencies in similar circumstances. Transit Agency Perceptions of Transit Benefits Programs Definitions of Success. Transit agency staff members who work on transit benefits programs were asked whether they would rate their programs as “very successful,” “some- what successful,” or “not successful.” Responses are shown in Table 24. Five agencies rated their programs as “very suc- cessful.” Some of the reasons cited are the following: • Increased ridership and revenues. These were cited in some way by all of the agencies, indicating that increased ridership and revenues were clear goals of the programs. • Congestion and air pollution reduction. One transit agency cited these benefits specifically because its region has experienced a huge growth in traffic congestion. • Good relationships with the business community. Sev- eral agencies mentioned that the program provided them with entrée into the local business community and allowed them to develop relationships that helped create a new constituency for transit and an additional avenue for marketing. Program Characteristics Universal Pass1 Monthly Pass2 Pricing Structure Complex—price is negotiated or tiered based on location of employer Generally simple and standardized, although may involve discounts for larger purchases of passes Size of Employer Generally Served Generally serve employers that are moderate to large in size (average of 50 to 490 employees per employer). Typically serve employers that are relatively small to moderate in size (average of 15 to 100 employees per employer) Number of Employers/ Employees Generally cover fewer employers (80 to 1,000)3 but more employees (40,000 to 50,000) Generally serve more employers (200 to 500) but fewer employees (12,000 to 30,000) Staffing 2.5 FTEs or more to administer 1–2 FTEs to administer Ridership Account for 5 to 15 percent of total ridership Account for 3 to 10 percent of total ridership4 Impact on Revenues Generally designed to be revenue neutral Generally designed to increase revenues when ridership increases 1 Table based on general indicators from three universal pass programs—King County Metro’s FlexPass, RTD’s Eco Pass, and VTA’s Eco Pass. 2 Table based on general indicators from three monthly pass programs—MARTA Partnership program, King County Metro’s consignment retail program, and Metro Transit’s TransitWorks! program. 3 The Denver RTD program has over 1,000 employers, but the other two have far fewer (80 and 200). Because the Denver program requires 3.6 FTEs to administer, the number of FTEs required to serve employers works out about even. 4 The percentage of ridership for Atlanta was not available from MARTA staff; we estimate it at less than 10 percent. TABLE 23 Comparison of universal and monthly pass programs Transit Agency Name Successful? WMATA Very MARTA Very King County Metro Very RTD Mixed Metro Transit Very VTA Very Valley Metro Somewhat TABLE 24 Transit agency perceptions of transit benefits program success

• Mechanism for rider feedback. Several agencies men- tioned that the programs give them a way to gauge rider response, especially when staff members have direct contact with employees. • Improved planning. One transit agency noted that employer sales aided in the planning process, giving some indication of ridership trends in the near future. • Customer loyalty. Finally, for several agencies the pro- gram provided a means to build customer loyalty. One transit agency noted, in particular, that it was proud of its flexibility in meeting customer needs, and another agency noted that the benefits provided by employers are per- ceived as having real value. The two agencies that reported their programs were only “somewhat successful” or “mixed” cited the following reasons: • Low participation. One transit agency said that the num- ber of employer and employee participants was below expectations. However, the agency noted that it was pleased that some employers had stayed with the program even in an economic downturn. • Difficulty with recruitment. One transit agency noted that it works in a difficult situation, in which transit has a poor reputation and employers are not receptive to their program. As one staff member put it, “People look at us in horror” when the agency suggests that employees switch to transit. • Unclear financial impact. Denver RTD, who reported their success as mixed, noted that their main reason for this uncertainty was not knowing whether the program was correctly priced. Their concern is that employers are being undercharged for the services their employees con- sume. RTD anticipates solving this problem with a smart card system to track ridership, but financial issues mean that procuring such a system may be several years off. Problems Encountered and Resolved. Agencies were asked open-ended questions about whether having a transit benefits program for employers had created or solved any problems for their transit agency. Some responses indicated that the transit benefits program created problems with fraud, employee/operator confusion, and crowding. More detail on each issue is provided below: • Fraud. Several agencies with universal passes indicated that they had encountered problems with employees attaching real stickers to ID cards for employers not par- ticipating in the program or employees loaning their cards to friends or relatives to ride free. • Employee/operator confusion. One transit agency with a large number of passes and programs reported that they sometimes encounter problems with explaining their sys- tems to staff in other departments, operators, and cus- tomers. However, the programs have developed to fill specific employer needs, and the flexibility is seen to out- weigh the problem. • Crowding. One popular universal pass program brought in so many riders that extra vehicles had to be added to respond to increased demand. Agency Responses to Programs. Several agencies indi- cated that they made changes to either their operations or to the employer program itself in response to employer and employee demand. These included changes to the following: • Routes and service. Several agencies added stops or made minor modifications to routes to better serve the employees at newly participating employers. In one case, ridership demand grew so much that more vehicles had to be added to routes. • Program operations. One transit agency added an option for pass holders to ride another transit provider, but this was later discontinued. The same transit agency also created an upgrade option for express bus routes. • Payment options. One transit agency discontinued voucher denominations that were infrequently purchased and added another credit card to their list of payment options. Other agencies are looking at online enrollment and reenrollment. Data Gaps on Transit Agency Impacts Many transit agencies had relatively sparse data on the effects of their employer programs on ridership, revenues, and costs. In order to assess the typical effects of transit benefits programs on transit agencies, additional data would be help- ful. Moreover, although agencies provided their impressions on the success of their programs, in order to better gauge suc- cess at meeting specific objectives, individual agencies should collect additional data on the following topics: • Program enrollment and revenues. Although every transit agency had good data on the number of employers enrolled, not every transit agency could identify the num- ber of employee participants. For instance, the agency may only know the number of stored-value cards or vouchers that are sold if these are used, but not how many employees are using them (e.g., an employee may receive one or more $20 vouchers). Likewise, transit agencies should be able to track the amount of revenue received from these programs in order to make a comparison with program costs and thereby determine the program’s effectiveness. • Intensity of transit ridership. Not every transit agency had information available on the level of transit ridership associated with transit benefits users. For instance, in the case of universal pass programs, employees may not ride 72