Comparative Review and Analysis of State Transit Funding Programs (2006)

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.

6This section is intended to serve two separate but parallel purposes. The first part of this section provides a framework for creating peer groups. The idea here is to present a tool that can be used by readers to formulate their own peer groups to serve their analysis objectives. The framework itself is pre- sented, followed by examples of potential peer groupings that serve varying purposes.Also shown are the peer groups selected by the project panel for the purposes of this research. The second part of this section shows the analyses per- formed on the peer groups selected by the project panel. Some of the most relevant data for each of the peer groups is shown, followed by analyses across peer groups and analyses within peer groups. Each of these analyses provides comparisons of data, from the Survey and other data sources, which serve to highlight key issues in transit funding nationwide. 2.1 Framework for Creating Peer Groups In this section a framework for creating peer groups is provided that can be used to conduct more meaningful peer analyses using data from the Survey. The framework includes suggested peer groups, as well as the peer groups selected by the project panel for analysis in this report. In all cases the peer groups would be used to compare the data available in the Survey—state transit funding, per capita state transit fund- ing, transit funding sources, and transit funding expenditures. This framework enables the creation of peer groups for the purposes of comparing any of those data. The framework has three basic steps. The first step in cre- ating a set of peer group states for analysis is to determine the objectives of the analysis, or the types of measures being com- pared. In general, all peer group analyses are going to com- pare groups of similar states. Determining the objective of the analysis will lead to the second step: determining the metrics for formulating the peer groups. The metrics are determined by figuring out which similarities the states in an individual peer group should share. Finally, the third step is to develop the peer groups based on the metrics chosen and their rela- tive importance. Figure 1 shows how this overall framework can be visualized. In the case of the primary set of peer groups desired by the project panel, the objective was to provide a general basis of comparison between states. The Survey as it is displayed today typically lists states in alphabetical order, which does not pro- vide a good basis for comparison. The objective was to group states into sets of peer groups that provided for a meaningful comparison. Several different analysis objectives can affect the creation of peer groups of states to compare the Survey data. For exam- ple, to determine major differences in federal and state transit funding, the key appropriate metric would probably be levels of federal transit funding. States would be grouped by their level of federal funding and then state funding would be com- pared within and between peer groups. Different objectives require the use of different metrics. For example, if the objective is to compare state transit funding between “transit-dependent” and “non–transit-dependent” states, various metrics that could create such a category would need to be considered. Some potential metrics follow: • Percent transit journey to work • Percent zero vehicles in household • Percent urban square miles • FTA flex funds One could develop peer groups on the basis of any one of these metrics individually. However, in many cases it is useful to develop peer groupings on the basis of a set of metrics that together provide for a more meaningful comparison, assign- ing weights to each metric. In such a case the next step is to determine the relative importance of each relevant metric. In the previous example, transit journey to work is probably the best indicator of transit use, with zero vehicles in household S E C T I O N 2 Peer Group Framework and Analysis

close behind. The other two metrics are probably not direct indicators of transit use but are definitely related to transit use. They would probably be weighted less in creating peer groups. Therefore, the following relative weights might be assigned: • Percent transit journey to work (45%) • Percent zero vehicles in household (35%) • Percent urban square miles (10%) • FTA flex funds (10%) Finally, after the metrics and their relative importance are determined, the next step is to create the actual peer groups. When multiple metrics are used, a statistical package such as SAS is useful. The research team used this package to score each state’s individual metric by quartile. However, in the absence of such a package, one can still formulate the groups if there are not too many variables by scoring each state using the following formula: Where M = Metric W = Weight N = Number of Metrics The states can then be sorted by their individual score. Met- rics may not always be in the same units; for example, three of the above metrics are percentages while one is a dollar amount. In the absence of a statistical package, all metrics must be con- verted to the same units. In this case, the dollars could be eas- ily converted to a percentage by measuring FTA flex funds for each state as a percentage of the national total. In other cases, one might have to reverse the direction of a metric. For exam- ple, if urban and rural states are being grouped, one of the met- rics might be number of rural square miles. However, another metric could be urban square miles. Although these are both important metrics, they measure opposites and therefore one must be reversed in the computation of the score. A sample of fictional scores appears in Table 8. Peer groups should ideally be developed based on natural clustering of scores. In these sample scores, Hawaii, Oregon and Maine probably would be grouped together. Virginia, Maryland, Colorado, Texas and Florida would also logically fall into the same group. Connecticut and Illinois would be grouped together, although they might be separated if Score M W M W M W1 1 2 2 n n ( ) ( ) ( )K there were other states below Connecticut that were closer to its score. Once the states are scored and sorted, one can easily create peer groups of states based on those scores. Some subjectiv- ity is inevitable at this point, but the goal should be to create less than arbitrary cutoff points so that the formulated groups are as distinct from one another as possible. Here is where the objectives may come back into play, as they can help deter- mine how many peer groups are needed and how different they should be. 2.1.1 Peer Group Possibilities In this section are outlined potential peer groups that would enable different types of useful comparisons. For each peer group, its value, limitations, and required data sources are identified. Note that peer group formulation is an inherently subjective process. For each set of groups the research team has made several decisions that could arguably have been made another way. These potential peer groups would provide for useful analysis, but in the process of creating one’s own peer groups one could just as easily devise something different and equally useful. After the various peer group possibilities are presented, the actual groupings selected for analysis in this report are discussed. Potential peer groups are presented by the following themes: • Geographic • Population demographics • Urban/rural • Income • Transit services These themes are intended as a sample of potential ideas for how peer groups could be organized. The potential peer groups are based on actual statistics and could be used for analysis if so desired. However, the idea behind this section is to provide a framework for peer group creation that allows 7 Objectives→ Metrics → Peer Groups Figure 1. Overall framework. Table 8. Sample of fictional scores. State Score Illinois Connecticut Virginia Maryland Colorado Texas Florida Hawaii Oregon Maine 0.91 1.12 1.25 1.25 1.43 1.57 1.75 2.20 2.25 2.25

political influences in state transit funding. However, it is lim- ited because many of these geographic distinctions are rela- tively arbitrary, no matter how many data points are used to refine their differences. Table 10 is an example of how one could formulate such peer groups based on the method just discussed. One judg- ment call made above is that Utah and Colorado were both placed in the Southwest group instead of the Western group, because the demographic data that was used showed that they had more in common with the Southwest. Population Demographics Another simple but useful potential set of peer groups could assemble groups of states that have similar population numbers. (See Table 11 for the value, limitations, and data sources for this potential set.) The objectives of the analysis in this case would be to see how transit funding differs among states with similar levels of population. This set would use only one statistic—population. Despite its simplicity, however, the comparisons enabled by this set of peer groups can still be useful. In particular, such comparisons could highlight the large differences in transit spending between states of similar population levels. However, other relevant demographics could be added to this formulation process depending on objec- tives. For example, one might want to know how states with similar levels of population and age of population differ in terms of transit funding. Another possibility is to include race or income indicators to create groups of demographically similar states. Table 12 groups the states into four peer groups by popu- lation. Some outliers are obvious here. For example, the Dis- trict of Columbia and Rhode Island, despite having small populations, are likely to have very different characteristics than their “peers” in terms of transit funding because they are very urban. Urban/Rural Character One of the most logical ways to group states is urbanization. (See Table 13 for the value, limitations, and data sources for 8 Value: May provide insight into geographic differences in state transit funding. Limitations: Inherent arbitrary nature of geographic divisions. Data Sources: U.S. Census. Table 9. Geographic peer groups—summary characteristics. Northeast Mid-Atlantic Southeast Midwest Southwest Western Pacific Maine West Virginia Alabama Minnesota Arizona Montana California New Jersey District of Columbia Mississippi Illinois New Mexico North Dakota Washington NevadaNew York North Carolina Kentucky Tennessee Iowa Oklahoma Texas South Dakota KansasPennsylvania Virginia Ohio Oregon Connecticut Maryland Louisiana Wisconsin Colorado Idaho Hawaii Massachusetts Delaware South Carolina Michigan Utah Wyoming Alaska Vermont Arkansas Missouri Nebraska New Hampshire Florida Indiana Rhode Island Georgia Table 10. Geographic peer groups. users to create their own peer groups that are most relevant for the analysis they are conducting. Geographic One of the most obvious and useful sets of peer groups could be organized along geographic lines. (See Table 9 for the value, limitations, and data source for this set.) This peer group set would not necessarily require any data and could be done sim- ply by looking at a map and making judgment calls based on objectives. However, one could use data analysis to create geo- graphic groups that correspond to some extent to demographic similarities. Such an analysis could be used to determine how to group borderline states. For example, to determine whether Ohio has more in common with the states of the Northeast or the Midwest, one could select particular characteristics and weights such as the following: • Total population over 65 (20%) • Total population 18 to 64 (20%) • Total population under 18 (20%) • Percent transit journey to work (20%) • Percent zero vehicles in household (20%) These data taken together will provide a good amount of information about the age and size of the population, as well as an indication of its likelihood to use transit. The resulting peer groups would allow the comparison of transit funding levels between well-defined geographic sectors. This set could be valuable in showing how state funding differs by geogra- phy and therefore perhaps in showing the relative cultural or

this potential set.) The objective here could be to determine how states with similar levels of urbanization stack up against one another in terms of transit funding. The following cate- gories and weights could be used: • Percent urban area population (16.67%) • Percent urban square miles (16.67%) • Urban square miles (16.67%) • Rural square miles (16.67%) • Percent urban vehicle miles traveled (VMT) (33.3%) This set of peer groups would highlight the differences in state transit funding between predominantly rural and pre- dominantly urban states. The obvious limitation of these data is that they emphasize area over population. Although VMT captures population to some degree, a state with a very pop- ulous but geographically small urban area could potentially be grouped into a rural category if the bulk of the area of the state is rural. Table 14 shows the urban/rural peer groups. The middle group is intentionally large so as to create greater balance on both the urban and rural sides of the table. Note that the “rural square miles” factor had to be inverted to create these groups. Income Peer groups can be created based on any chosen demo- graphic. One that might be particularly useful is income, because of its strong relationship to transit use. (See Table 15 for the value, limitations and data sources for this potential grouping.) Age and race might also be interesting, but income has a more direct link to transit across all states. The objective could be to determine whether states with similar income levels would also have similar levels of state transit funding. The following categories and weights could be used: • Percent below poverty (33.3%) • Percent household income below $30,000 (33.3%) • Average weekly wage (33.3%) Note that “below poverty” and “household income below $30,000” are two different demographics that provide differ- ent information. The poverty line is determined based on the number of people in a household and thus captures poverty at the individual level. Using straight income provides a measure of low-income families who may or may not be in poverty. Also note that the “average weekly wage” factor had to be inverted to create the scoring system for these groups. This set of peer groups would highlight differences between how state transit spending changes with the prevalence of lower-income individuals. One limitation of these groups is that they only provide information about one specific demographic char- acteristic, which narrows the usefulness of the comparison. Also the groups are necessarily limited by the types of variables used to assess income. Table 16 shows the income-based peer groups. Transit Services Another interesting set of peer groups could use other avail- able transit-related data to group states. This set would enable a comparison of how state funding relates to other available measures of transit use. Table 17 shows the value, limitations, and data sources of this set. The categories and weights could be as follows: • Percent transit journey to work (25%) • Percent zero vehicles in household (25%) • Total FTA funding (25%) • FTA flex funds as a percentage of FTA funds (25%) 9 Value: May provide insight into how states with similar populations differ in their transit funding. Limitations: Population may not be as relevant a factor as other demographics in examining transit funding. Data Sources: U.S. Census. Small Medium Large Extra-Large Alaska Arkansas Alabama California Delaware Connecticut Arizona Florida District of Columbia Iowa Kansas Colorado Georgia Hawaii Indiana Illinois Idaho Kentucky Louisiana Michigan Maine Mississippi Maryland New Jersey Montana Nebraska Nevada Massachusetts New York New Hampshire Minnesota North Carolina North Dakota New Mexico Missouri Ohio Rhode Island Oklahoma South Carolina Tennessee Pennsylvania TexasSouth Dakota Oregon Vermont Utah Washington Virginia Wyoming West Virginia Wisconsin Value: Contrasts state transit funding for predominantly rural and predominantly urban states. Limitations: Ignores the role of population density. Data Sources: U.S. Census. Table 11. Population demographic peer groups—summary characteristics. Table 13. Urban/rural peer groups—summary characteristics. Table 12. Population peer group.

population, prevalence of elderly, and transit funding. Both federal and state transit funding levels were included as met- rics, even though the groups might also be compared in terms of these statistics. The idea is that their inclusion helps to mag- nify the importance of differences in transit funding within peer groups. The research team had the advantage of being able to use a statistical package, SAS, to analyze the data. SAS took each data point for each state and placed it into one of four groups. The groups were formulated in a straightforward statistical manner, with the data point placed into groups based solely on its percentile. States under the 25th percentile went into a group numbered 0, states between the 25th and 50th percentile went into the next group numbered 1, and so on. The result was a grouping for each data point for each state. From there the research team took three separate approaches to formulating peer groups. First, all of the group numbers were averaged across all data points for each state, which created distinct groups with the following obvious cut- off points: those averaging below 1, those between 1 and 1.5, those between 1.5 and 2, and those above 2. The second potential set of peer groups was formulated by first grouping different metrics into subgroups. For example, all age-related characteristics were grouped as one, and all FTA data were grouped together. Then averages were computed for each group, and those were averaged together. For the final poten- tial set of peer groups, the research team chose to include data points that it determined to be not only the most useful data, but also those likely to complement one another. For example, a state with a high urban-area percentage is likely to have a low number of rural square miles. Therefore only urban-area percentage and not rural square miles was included in the cal- culation. This last approach yielded both a four-group and five-group set. 10 Most Rural Rural Middle Urban Most Urban Montana New Mexico Wisconsin Arizona District of Columbia Wyoming Iowa Kentucky Delaware Hawaii North Dakota Maine Minnesota Georgia Illinois South Dakota Alaska South Carolina Nevada California Rhode Island Idaho Michigan Connecticut Nebraska Arkansas Kansas Oklahoma Ohio Florida Vermont New Hampshire Pennsylvania Maryland West Virginia Colorado Virginia New York Oregon Missouri Massachusetts Mississippi Alabama New Jersey Indiana Louisiana North Carolina Texas Washington Tennessee Utah Table 14. Urban/rural peer groups. Table 15. Income peer groups— summary characteristics. Value: Contrasts state transit funding based on household income and wages. Limitations: Ignores all other demographics. Data Sources: U.S. Census. The resulting set would group states in terms of their depend- ence on, and federal funds for, transit. This set would isolate state transit funding as a factor compared to other key tran- sit indicators. The major limitation of course is that this set is likely to be redundant; it will probably provide some cases of outliers that have very different state transit funding levels than their peers, but in most cases one would expect a con- forming alignment. Table 18 shows the devised peer groups based on transit services. The use of the “total FTA funding” metric makes the higher groups lean a bit towards bigger population states; although the District of Columbia and Nevada are in the higher groups. 2.1.2 Selected Peer Groups This section shows the peer groups selected by the panel. The panel was interested in peer groups that would cut across several categories. Table 19 shows the value, limitations, and data sources of the panel-selected peer groups. The objective was to create similar groups of states for the purposes of com- parison. The similarities were intended to be robust across a wide range of characteristics and at least tangentially related to transit funding or use. Therefore, the research team pre- sented the project panel with equally weighted metrics deal- ing with urbanization, racial diversity, income, transit use,

Delaware and Rhode Island both operate service statewide, whereas Maryland and Massachusetts focus their state oper- ations in their largest cities (Baltimore and Boston, respec- tively). New Jersey has one statewide transit operator (New Jersey Transit) that is controlled by the state.When these states are listed in their original peer groups, they are denoted with an asterisk. 2.2 Peer Group Analysis This section shows an analysis of the set of peer groups chosen by the project panel. In the last section the focus was on peer group creation, but now that a set of peer groups has been chosen, this section will focus on exactly how to use the groups to create a useful set of analyses. The goal is to better understand how these groups compare to each other and how states within each peer group compare to one another. Several sets of comparisons are shown in the various sub- sections that follow. First basic statistics for each peer group are presented, which show the score that placed each state in that peer group along with each state’s state transit funding, per capita state transit funding, federal transit funding, and per capita federal transit funding (all data are from 2004). The idea behind this comparison is twofold: (1) it provides a more detailed look at how the peer groups were formulated, so that readers can tell how closely related a state is to others within its peer group and (2) it allows for rudimentary comparisons against four key indicators of transit funding levels. After the peer group statistics, some fundamental com- parisons across peer groups are presented. Four metrics were chosen for these comparisons: state funding, per capita state funding, federal funding, and state and federal funding shares. These metrics are most relevant for showing differences in state funding levels, which is the primary goal of this research. Federal data are shown primarily for comparison purposes. Finally, comparisons within the peer group are presented. For each peer group key data from the Survey itself are com- pared, including total state and federal funding, per capita state 11 Table 16. Income-based peer groups. Montana New Mexico Arizona District of Columbia Iowa Kentucky Delaware Hawaii Maine Minnesota Georgia Illinois North Dakota South Dakota Alaska South Carolina Nevada California Rhode IslandMichigan Connecticut Nebraska Arkansas Kansas Oklahoma Ohio Florida VermontNew Hampshire Pennsylvania Maryland West Virginia Colorado Virginia New York Oregon Missouri Massachusetts Mississippi Alabama New Jersey Indiana Louisiana North Carolina Texas Washington Utah Highest High Middle Low Lowest Wisconsin Idaho Tennessee Wyoming Table 17. Transit services peer groups—summary characteristics. Value: Contrasts state transit funding for states grouped by other transit-related statistics. Limitations: Provides for only one specific useful comparison Data Sources: U.S. Census, FTA. The research team presented each of these potential sets to the project panel, and feedback was strongly in favor of the final set using five peer groups. Therefore, the analysis pro- ceeded using this set of peer groups. The following metrics were used to choose the peer groups (all weighted equally): • Percent urban area • Percent urban VMT • Percent Hispanic/Latino • Percent African-American • Percent household income below $30,000 • Percent transit journey to work • Total population • Percent population over 65 • Percent population disabled • FTA urbanized area formula • FTA flex funds • State transit funding Table 20 presents the results of this chosen grouping method. It also would be useful to separately take into account the states that own or operate transit service statewide. There- fore these states were separated out into their own addi- tional peer group for the purposes of this analysis. Note that this peer group is not exclusive of the others and comprises Delaware, Rhode Island, Maryland, Massachusetts and New Jersey. Not all of these states fit neatly into this category.

Virginia, Wisconsin and Minnesota all have much higher total and per capita funding levels than others in this group, despite their being at opposite ends of the spectrum in terms of peer group scores. As shown in Table 24, in Group 4, Maryland and Massa- chusetts, which have state-operated transit systems, are the two top states for state funding. They also rank highest on a per capita basis if the District of Columbia is excluded. (The District of Columbia is unique in being the only “state” con- taining 100% urban area.) Table 25 shows the last peer group, Group 5. New York and New Jersey have similar per capita funding levels even though New York has more than twice the state funding and about four times the federal funding. Group 5 can almost be thought of as three sets of pairs—New York/California, Illinois/Pennsylvania and Texas/Florida—as each of these couples share similar fed- eral and state funding levels within the peer group. The only outlier in this sense is New Jersey, which again may be explained by its status as a statewide transit-operator state. Finally, as shown in Table 26, the transit-operator state peer group shows high variability within itself, with state funding ranging from around $37 million (Rhode Island) to around $1.3 billion (Massachusetts). The main characteristic these states have in common in terms of funding is that all are rel- atively well funded on a per capita basis. This disparity would be expected as they are all from different peer groups; however, it is unexpected because they tended to be outliers within those groups and may have been expected to be outliers in a similar manner. 2.2.2 Analyses across Peer Groups This subsection presents a cross-sectional peer group graph- ical analysis. It investigates differences among the peer groups along the following dimensions: • State funding levels • Per capita state funding 12 Table 18. Transit services peer groups. Lowest Low Middle High Highest Nebraska Alaska Hawaii Arizona California North Dakota Arkansas Kansas Louisiana Georgia District of Columbia South Dakota Rhode Island Minnesota Nevada Texas Massachusetts West Virginia Montana Alabama Ohio New Hampshire Oklahoma Colorado Maryland Vermont Delaware Indiana Connecticut New Jersey Wyoming Idaho Kentucky Florida New York Maine Iowa Michigan Oregon Pennsylvania Mississippi North Carolina Missouri Washington Illinois New Mexico South Carolina Utah Virginia Wisconsin Tennessee Table 19. Panel-selected peer groups—summary characteristics. Value: Groups states by multiple factors that are all likely to be related to transit funding. Limitations: Includes transit funding itself as a statistic, thus slightly diminishing the value of the resulting comparisons. Data Sources: U.S. Census, FTA, Survey. and federal funding, state funding sources, and state funding expenditure categories. These comparisons allow states to com- pare their funding data to those of similar states. 2.2.1 Peer Group Statistics Table 21 shows the first peer group. The peer group score represents an average of the percentiles of the factors consid- ered in formulating the peer groups.Vermont is a clear outlier in this group, showing a much higher level of state funding both per capita and on a total basis. However, another inter- esting outlier is Alaska, which is a zero-funding state yet has a much higher level of federal funding than any of its peers. Table 22 shows Group 2. Note that Delaware, the only state- wide transit operator in this group, is a clear outlier among its peers. It has much higher state funding, both per capita and total, than any of its peers, and much lower federal funding. Except for Delaware, Iowa has the highest overall funding in this group. The group also has two states that do not provide funding (Hawaii and Utah), both of which receive relatively high levels of federal funding. In Group 3, as shown in Table 23, Rhode Island’s per capita funding is high compared to its peers even though it ranks lower than the mean in total state funding and is the lowest in federal funding. Rhode Island’s uniqueness may be explained by its statewide transit operation. However, Connecticut,

13 Table 20. Devised peer group formulations. Group 1 Group 2 Group 3 Group 4 Group 5 New Hampshire Nebraska Minnesota Louisiana Texas South Dakota West Virginia Wisconsin Maryland New Jersey North Dakota Iowa Alabama Washington Pennsylvania Vermont Kansas Colorado Ohio California Montana Mississippi Oklahoma Tennessee Illinois Maine Hawaii Rhode Island Georgia Florida Alaska Delaware Indiana Massachusetts New York Idaho Arkansas Missouri Michigan Wyoming Kentucky Nevada North Carolina New Mexico Oregon Arizona Utah South Carolina District of Columbia Connecticut Virginia Table 21. Group 1 data. State Peer Group Ranking State Funding (Thousands) Per Capita State Funding FTA Funding (Thousands) Per Capita FTA Funding New Hampshire 0.42 $225 $0.17 $6,516 $5.01 South Dakota 0.42 $996 $1.29 $3,777 $4.90 North Dakota 0.50 $1,546 $2.44 $4,891 $7.71 Vermont 0.50 $6,103 $9.82 $12,667 $20.38 Montana 0.58 $390 $0.42 $2,812 $3.03 Maine 0.67 $505 $0.38 $14,330 $10.88 Alaska 0.75 $0 $0.00 $35,880 $54.74 Idaho 0.75 $312 $0.22 $11,444 $8.21 Wyoming 0.83 $2,466 $4.87 $4,215 $8.32 Group 1 Average 0.60 $1,394 $2.18 $10,726 $16.78† † The average per capita federal funding represents a weighted average by population. Table 22. Group 2 data. State Peer Group Ranking State Funding (Thousands) Per Capita State Funding FTA Funding (Thousands) Per Capita FTA Funding Nebraska 1.00 $1,500 $0.86 $8,156 $4.67 West Virginia 1.00 $2,294 $1.26 $11,680 $6.43 Iowa 1.08 $8,600 $2.91 $33,553 $11.36 Kansas 1.08 $6,000 $2.19 $21,182 $7.74 Mississippi 1.08 $800 $0.28 $14,638 $5.04 Hawaii 1.17 $0 $0.00 $39,384 $31.19 Delaware* 1.33 $72,000 $86.71 $3,919 $4.72 Arkansas 1.42 $2,800 $1.02 $19,142 $6.95 Kentucky 1.42 $1,400 $0.34 $39,859 $9.61 New Mexico 1.42 $2,402 $1.26 $27,354 $14.37 Utah 1.42 $0 $0.00 $54,227 $22.70 Group 2 Average 1.22 $8,891 $8.80 $24,827 $24.58† * Operates a statewide transit system. † The average per capita federal funding represents a weighted average by population.

14 Table 23. Group 3 data. State Peer Group Ranking State Funding (Thousands) Per Capita State Funding FTA Funding (Thousands) Per Capita FTA Funding Minnesota 1.50 $214,255 $42.00 $161,613 $31.68 Wisconsin 1.50 $109,078 $19.80 $65,885 $11.96 Alabama 1.58 $0 $0.00 $19,978 $4.41 Colorado 1.58 $0 $0.00 $122,712 $26.67 Oklahoma 1.58 $2,750 $0.78 $28,461 $8.08 Rhode Island* 1.58 $36,840 $34.09 $13,259 $12.27 Indiana 1.67 $36,201 $5.80 $65,326 $10.47 Missouri 1.67 $6,600 $1.15 $95,664 $16.62 Nevada 1.67 $125 $0.05 $52,256 $22.38 Oregon 1.67 $31,445 $8.75 $158,439 $44.08 South Carolina 1.75 $5,864 $1.40 $28,051 $6.68 Connecticut 1.92 $200,167 $57.13 $67,759 $19.34 Virginia 1.92 $140,100 $18.78 $123,435 $16.55 Group 3 Average 1.66 $60,263 $14.60 $77,141 $18.68† * Operates a statewide transit system. † The average per capita federal funding represents a weighted average by population. Table 24. Group 4 data. State Peer Group Ranking State Funding (Thousands) Per Capita State Funding FTA Funding (Thousands) Per Capita FTA Funding Louisiana 2.00 $4,963 $1.10 $70,321 $15.57 Maryland* 2.00 $789,511 $142.05 $75,132 $13.52 Washington 2.00 $29,150 $4.70 $278,772 $44.94 Ohio 2.08 $18,100 $1.58 $173,992 $15.18 Tennessee 2.08 $38,532 $6.53 $59,619 $10.10 Georgia 2.17 $4,858 $0.55 $141,942 $16.08 Massachusetts* 2.17 $1,291,363 $201.26 $192,082 $29.94 Michigan 2.17 $209,652 $20.73 $113,314 $11.21 North Carolina 2.17 $154,680 $18.11 $77,570 $9.08 Arizona 2.25 $20,068 $3.49 $177,116 $30.84 District of Columbia 2.25 $208,253 $376.23 $286,676 $517.91 Group 4 Average 2.12 $251,739 $70.58 $149,685 $41.96† * Operates a statewide transit system. † The average per capita federal funding represents a weighted average by population. Table 25. Group 5 data. State Peer Group Ranking State Funding (Thousands) Per Capita State Funding FTA Funding (Thousands) Per Capita FTA Funding Texas 2.33 $27,741 $1.23 $310,673 $13.81 New Jersey* 2.42 $837,476 $96.27 $530,201 $60.95 Pennsylvania 2.42 $785,151 $63.29 $410,761 $33.11 California 2.50 $1,317,934 $36.72 $1,229,826 $34.26 Illinois 2.50 $778,700 $61.25 $515,894 $40.58 Florida 2.58 $96,504 $5.55 $268,159 $15.41 New York 2.67 $1,811,372 $94.21 $2,103,584 $109.41 Group 5 Average 2.49 $807,840 $51.22 $767,014 $48.63† * Operates a statewide transit system. † The average per capita federal funding represents a weighted average by population.

15 Table 26. Transit-operator state peer group data. State Peer Group Ranking State Funding (Thousands) Per Capita State Funding FTA Funding (Thousands) Per Capita FTA Funding Delaware 1.33 $72,000 $86.71 $3,919 $4.72 Maryland 2.00 $789,511 $142.05 $75,132 $13.52 Rhode Island 1.58 $36,840 $34.09 $13,259 $12.27 Massachusetts 2.17 $1,291,363 $201.26 $192,082 $29.94 New Jersey 2.42 $837,476 $96.27 $530,201 $60.95 Transit-Operator State Average 1.90 $605,438 $112.08 $162,919 $30.16† † The average per capita federal funding represents a weighted average by population. $1,394 $8,891 $60,263 $251,739 $807,840 $605,438 $182,701 $0 $100,000 $200,000 $300,000 $400,000 $500,000 $600,000 $700,000 $800,000 $900,000 Group 1 Average Group 2 Average Group 3 Average Group 4 Average Group 5 Average Transit- Operator State Average All States Average Figure 2. State funding across peer groups (2004 funding in thousands of dollars). • Levels of federal transit funding • State versus federal funding shares • Funding sources • Funding expenditures Figure 2 compares the peer groups to one another on the basis of state funding for transit. As expected, the larger state peer groups have more funding on average. The figure also tells us that transit-operator states are on average funded at a much higher rate than any peer group other than Group 5, and they are well above the average for all states. Figure 3 tells a slightly different story. It shows that per capita state funding tends to increase with increasing “size”but only to a point. The largest, most populous and most heavily transit-dependent states have lower per-capita spending on transit than those slightly “smaller” than them. However, note that the very high per capita spending for Group 4 is attributable mostly to three “states”: Maryland, Massachusetts and the District of Columbia. Maryland and Massachusetts are both transit-operator states, which seem to have higher per capita funding than other states by a large margin. The District of Columbia is not only unique in that it is the only 100% urban “state” but it is also on the cusp of being in Group 5. If we remove the transit-operator states from all peer groups, Group 4 still has a higher per capita spending ($48.11) than Group 5 ($43.71). If we then proceed

to remove the District of Columbia, Group 4 shrinks to lower than Group 3. Therefore it is important to be cautious when interpreting these statistics. The only statement that can be made with certainty from this figure is that per capita spending tends to increase with peer group “size” and is highest for statewide transit operators. Figure 4 shows federal transit funding. The striking thing about this figure is how the Group 5 states are dramatically out of line with the level of funding for other states. Transit- operator states and Group 4 states are on par with the average level of federal funding. Figure 5 shows the percentage of total transit funding for each peer group that is attributable to state and federal sources. The trend here is that as peer group “size” increases, so does the state burden for transit spending. This trend is bucked only by Group 5, but even that group is well ahead of Groups 1, 2 and 3 with more than 50% of its transit funding coming from state sources. By comparison, the Group 1 states have a little more than 10% from state sources. Note also that transit- operator states bear a much heavier burden on average than any other peer group. Figure 6 shows the funding sources across all peer groups. Every peer group except Group 1 is dominated by general fund revenues, while Group 1 relies more on registration/ license/title fees than general funds. The use of motor vehicle/ rental car sales tax is much more prevalent among Group 2 than any other group, while Group 3 shows the most preva- lent use of both interest income and a general sales tax. Figure 7 shows funding expenditure categories across peer groups. There seems to be a relative lack of variation across peer groups, but Group 3 contributes by far the most to oper- ating costs exclusively. Group 5 offers the most flexibility in its expenditure, allowing more than 40% of its funds to be allocated to either operating or capital. 2.2.3 Analyses Within Peer Groups This subsection provides comparative data within each of the peer groups. Specifically, analyses are provided (in graph- ical format) on total transit funding, state and federal per capita funding, sources of state funding, and state funding category expenditure. • Figures 8 through 11 provide graphical analyses for Group 1. • Figures 12 through 15 provide graphical analyses for Group 2. • Figures 16 through 19 provide graphical analyses for Group 3. • Figures 20 through 23 provide graphical analyses for Group 4. 16 Figure 3. Per capita state funding across peer groups. $2.18 $8.80 $14.60 $70.58 $51.22 $112.08 $28.26 $0.00 $20.00 $40.00 $60.00 $80.00 $100.00 $120.00 Group 1 Average Group 2 Average Group 3 Average Group 4 Average Group 5 Average Transit- Operator State Average All States Average

17 $10,726 $24,827 $77,141 $149,685 $767,014 $162,919 $164,473 $0 $100,000 $200,000 $300,000 $400,000 $500,000 $600,000 $700,000 $800,000 $900,000 Group 1 Average Group 2 Average Group 3 Average Group 4 Average Group 5 Average Transit- Operator State Average All States Average Figure 4. Federal transit funding across peer groups (2004 funding in thousands of dollars). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Federal Share State Share Group 1 Average Group 2 Average Group 3 Average Group 4 Average Group 5 Average Transit- Operator State Average All States Average Figure 5. State versus federal shares of transit funding.

18 %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 rotarepO tisnarTslatoT 5 puorGslatoT 4 puorGslatoT 3 puorGslatoT 2 puorGslatoT 1 puorG slatoT etatS dnuF lareneG xaT saG xaT selaS raC latneR/elciheV rotoM seeF eltiT/esneciL/noitartsigeR sdeecorP dnoB xaT selaS lareneG emocnI tseretnI rehtO Figure 6. Funding sources, all peer groups. %0 %02 %04 %06 %08 %001 rotarepO tisnarTslatoT 5 puorGslatoT 4 puorGslatoT 3 puorGslatoT 2 puorGslatoT 1 puorG slatoT etatS latipaC gnitarepO htoB/rehtiE Figure 7. Funding expenditures, all peer groups.

• Figures 24 through 27 provide graphical analyses for Group 5. • Figures 28 through 31 provide graphical analyses for the transit-operators group. Figure 8 compares total levels of federal and state funding for transit for Group 1. Most of the states in Group 1 have rel- atively low levels of total transit funding; Alaska’s funding is the highest at a little more than $35 million, even though the state does not contribute any transit funding. As shown in Figure 9, for most states in Group 1, total fund- ing is in line with per capita funding. As shown in Figure 10, Group 1 states used a fairly wide vari- ety of funding sources. Four states—South Dakota, Vermont, Idaho and Wyoming—use only “other” sources, such as state highway funds, trust funds, miscellaneous revenue, fees, lottery funds, taxes, tolls, and other assessments. As shown in Figure 11, Montana reported on 19.2% of its funding. Among the other states, Idaho, Wyoming and New Hampshire all devote a substantial portion of their funding to capital, while Maine and South Dakota focus exclusively on operating funds. As shown in Figure 12, in Group 2, Delaware—the only statewide transit operator in the group—has the highest total funding, almost all of which comes from the state. Utah and Hawaii have comparatively high federal funding, but no state funding. Iowa and Kansas are two of the biggest contributors on a state level within this group. As Figure 13 shows, Delaware’s per capita funding of more than $90 per person is much higher than any other state in the group. Hawaii and Utah have relatively high per capita fund- ing with only federal funds. Unlike other peer groups, the “other”category did not dom- inate funding sources in Group 2 (see Figure 14). Arkansas and Iowa rely exclusively on a motor vehicle or rental car sales tax. No state uses more than one funding source. As shown in Figure 15, in Group 2, New Mexico had a high proportion of funding to “other” expenditures. Within Group 3, as shown in Figure 16, Minnesota is highest in state, federal, and overall funding. Connecticut and Virginia are also well above average, especially in terms of state funding. Wisconsin and Rhode Island (a transit- operator state) have fairly even splits between federal and state funding; the other states rely more heavily on federal funds. Figure 17 shows that Connecticut has the highest per capita funding in this group, followed by Minnesota and Oregon. Rhode Island is also one of the states with higher per capita funding. The average in Figure 18 is skewed to some extent because two states do not provide state funding, and four states did not provide information on their sources. Rhode Island and 19 $0 $5,000 $10,000 $15,000 $20,000 $25,000 $30,000 $35,000 $40,000 gvAYWDIKAEMTMTVDNDSHN gnidnuF etatS gnidnuF laredeF Figure 8. Group 1 state and federal total transit funding (in thousands of dollars).

20 0$ 01$ 02$ 03$ 04$ 05$ 06$ gvAYWDIKAEMTMTVDNDSHN gnidnuF etatS gnidnuF laredeF Figure 9. Group 1 state and federal per capita funding. Figure 10. Group 1 sources of state funds. %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvAYWDIKAEMTMTVDNDSHN rehtO dnoB sdeecorP /noitartsigeR eltiT /esneciL seeF xaT saG dnuF lareneG oN sdnuf

21 Figure 11. Group 1 state funding expenditure categories. Figure 12. Group 2 state and federal total transit funding (in thousands of dollars). %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvAYWDIKAEMTMTVDNDSHN htoB/rehtiE gnitarepO latipaC oN atad oN sdnuf $0 $10,000 $20,000 $30,000 $40,000 $50,000 $60,000 $70,000 $80,000 gvATUMNYKRAEDIHSMSKAIVWEN gnidnuF etatS gnidnuF laredeF

22 0$ 01$ 02$ 03$ 04$ 05$ 06$ 07$ 08$ 09$ 001$ gvATUMNYKRAEDIHSMSKAIVWEN gnidnuF etatS gnidnuF laredeF Figure 13. Group 2 state and federal per capita funding. Figure 14. Group 2 sources of state funds. %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvATUMNYKRAEDIHSMSKAIVWEN rehtO rotoM latneR/elciheV xaT selaS raC dnuF lareneG oN sdnuf oN sdnuf oN atad oN atad

23 Figure 15. Group 2 state funding expenditure categories. Figure 16. Group 3 state federal total transit funding (in thousands of dollars). %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvATUMNYKRAEDIHSMSKAIVWEN rehtO htoB/rehtiE gnitarepO latipaC oN sdnuf oN sdnuf $0 $50,000 $100,000 $150,000 $200,000 $250,000 $300,000 $350,000 $400,000 gvAAVTCCSROVNOMNIIRKOOCLAIWNM gnidnuF etatS gnidnuF laredeF

24 0$ 01$ 02$ 03$ 04$ 05$ 06$ 07$ 08$ 09$ 001$ gvAAVTCCSROVNOMNIIRKOOCLAIWNM gnidnuF etatS gnidnuF laredeF 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% gvAAVCSROVNOMNIIRKOOCLAIWNM emocnI tseretnI selaS lareneG xaT xaT saG dnuF lareneG oN atad oN atad oN sdnuf oN sdnuf oN atad oN atad Figure 17. Group 3 state and federal per capita funding. Figure 18. Group 3 sources of state funds.

South Carolina rely exclusively on the gas tax, Nevada on interest income. As shown in Figure 19, Nevada uses its state funding exclu- sively on capital projects; all other states in this group use their funding for operating or both operating and capital. As shown in Figure 20, the two clear outliers in Group 4 are statewide transit operators—Maryland and Massachu- setts—followed by the District of Columbia, an entirely urbanized area. Of the rest, Washington and Michigan have the highest funding. Washington is mostly federally funded, while Michigan is more heavily state funded. Figure 21 shows that the District of Columbia has the highest per capita funding, with nearly $900 allocated per person (including federal and state funding). Massachusetts and Maryland are the next highest. For the states for which data were available, as shown in Fig- ure 22, only Maryland (a transit-operator state) had multiple funding sources. Other states relied primarily on the general fund (Ohio, Georgia and DC), the gas tax (Tennessee), or “other” sources (Louisiana and Arizona). As shown in Figure 23, Georgia used almost all of its funds for capital expenditures, while all other states had more mixed spending patterns. Figure 24 shows that New York, as might be expected, leads Group 5 by a considerable margin in both state and federal funding. In addition to total funding, as shown in Figure 25, New York is also the highest in per capita funding for Group 5. New Jersey, a transit-operator state, has the second-highest per capita funding. While California’s total funding is second to New York’s, on a per capita basis it ranks fifth. As Figure 26 indicates, most states in Group 5 did not have complete information available about the percentage of funding obtained from various sources. (Table 27, pre- sented in Section 3.3, shows that most of these states have multiple funding sources, but the percentages were not specified.) As shown in Figure 27, most of the states in Group 5 had some combination of capital and operating expenditures. Figure 28 shows that all statewide transit operators rely much more heavily on state funding than federal funding and do so to a much greater extent than other states. As shown in Figure 29, the per capita funding among transit- operator states is more comparable than total transit funding. Data on sources of state funding were available for only three states in the transit-operator group; therefore, whether a different funding source pattern exists for this group than for the states as a whole is difficult to determine, as Figure 30 demonstrates. While Rhode Island used almost all of its funding for oper- ating costs, Figure 31 shows that the other transit operators had a variety of spending patterns. 25 Figure 19. Group 3 state funding expenditure categories. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% MN WI AL CO OK RI IN MO NV OR SC CT VA Avg Other Either/Both Operating Capital No funds No funds

26 $0 $200,000 $400,000 $600,000 $800,000 $1,000,000 $1,200,000 $1,400,000 $1,600,000 gvACDZACNIMAMAGNTHOAWDMAL gnidnuF etatS gnidnuF laredeF Figure 20. Group 4 state and federal total transit funding (in thousands of dollars). Figure 21. Group 4 state and federal per capita funding. 0$ 001$ 002$ 003$ 004$ 005$ 006$ 007$ 008$ 009$ gvACDZACNIMAMAGNTHOAWDMAL gnidnuF etatS gnidnuF laredeF

27 Figure 22. Group 4 sources of state funds. Figure 23. Group 4 state funding expenditure categories. %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvACDZACNIMAMAGNTHOAWDMAL rehtO dnoB sdeecorP /noitartsigeR eltiT /esneciL seeF rotoM latneR/elciheV xaT selaS raC xaT saG dnuF lareneG oN atad oN atad oN atad %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvACDZACNIMAMAGNTHOAWDMAL rehtO htoB/rehtiE gnitarepO latipaC

28 $0 $500,000 $1,000,000 $1,500,000 $2,000,000 $2,500,000 $3,000,000 $3,500,000 $4,000,000 $4,500,000 gvAYNLFLIACAPJNXT gnidnuF etatS gnidnuF laredeF 0$ 05$ 001$ 051$ 002$ 052$ gvAYNLFLIACAPJNXT gnidnuF etatS gnidnuF laredeF Figure 24. Group 5 state and federal total transit funding (in thousands of dollars). Figure 25. Group 5 state and federal per capita funding.

29 %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvAYNLFLIACAPJNXT rehtO lareneG dnuF oN atad oN atad oN atad oN atad %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvAYNLFLIACAPJNXT rehtO htoB/rehtiE gnitarepO latipaC Figure 26. Group 5 sources of state funds. Figure 27. Group 5 state funding expenditure categories.

30 $0 $200,000 $400,000 $600,000 $800,000 $1,000,000 $1,200,000 $1,400,000 $1,600,000 gvAJNAMDMIRED gnidnuF etatS gnidnuF laredeF 0$ 05$ 001$ 051$ 002$ 052$ gvAJNAMDMIRED gnidnuF etatS gnidnuF laredeF Figure 28. Transit-operator group state and federal total transit funding (in thousands of dollars). Figure 29. Transit-operator group state and federal per capita funding.

31 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% gvAJNAMDMIRED rehtO sdeecorP dnoB /noitartsigeR seeF eltiT /esneciL /elciheV rotoM selaS raC latneR xaT xaT saG dnuF lareneG oN atad oN atad %0 %01 %02 %03 %04 %05 %06 %07 %08 %09 %001 gvAJNAMDMIRED htoB/rehtiE gnitarepO latipaC Figure 30. Transit-operator group sources of state funds. Figure 31. Transit-operator group state funding expenditure categories.