GEORGIA DOT RESEARCH PROJECT 17-01 FINAL REPORT A STATEWIDE ASSESSMENT OF PUBLIC TRANSIT FUNDING NEEDS FOR COUNTIES TRENDING URBAN IN GEORGIA OFFICE OF PERFORMANCE-BASED MANAGEMENT AND RESEARCH 600 W. PEACHTREE ST. NW ATLANTA, GA 30308 GDOT Research Project No. 17-01 Final Report A STATEWIDE ASSESSMENT OF PUBLIC TRANSIT FUNDING NEEDS FOR COUNTIES TRENDING URBAN IN GEORGIA By Laurie A. Garrow Professor and Thomas Douthat Research Engineer and Sara Douglass Lynch and Anna Nord Master's Students Georgia Institute of Technology Contract with Georgia Department of Transportation In cooperation with U.S. Department of Transportation Federal Highway Administration February 2020 The contents of this report reflect the views of the authors who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Georgia Department of Transportation or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. 1.Report No.: 2. Government Accession No.: 3. Recipient's Catalog No.: FHWA-GA-20-1701 4. Title and Subtitle: 5. Report Date: A Statewide Assessment of Public Transit Funding February 2020 Needs for Counties Trending Urban in Georgia 6. Performing Organization Code: 7. Author(s): 8. Performing Organ. Report No.: Laurie A. Garrow, Thomas Douthat, Anna Nord, RP17-01 and Sara Douglass 9. Performing Organization Name and Address: 10. Work Unit No.: Georgia Institute of Technology 790 Atlantic Drive 11. Contract or Grant No.: Atlanta, GA 30332-0355 PI#0015650 12. Sponsoring Agency Name and Address: 13. Type of Report and Period Covered: Georgia Department of Transportation Final; August 2017February 2020 Office of Performance-based Management and Research 14. Sponsoring Agency Code: 600 W. Peachtree St. NW GDOT Atlanta, GA 30308 15. Supplementary Notes: Prepared in cooperation with the U.S. Department of Transportation, Federal Highway Administration. 16. Abstract: This study predicts how spatial and temporal population changes from the 2010 and 2020 decennial censuses will impact funding for the Federal Transit Administration (FTA) 5311 rural and FTA 5307 urban transit programs, both in Georgia and nationwide. Binary logit models and geographic information system (GIS) methods were used to predict which areas of the U.S. will be classified as rural, small urban, or large urban after 2020. This information was then used to forecast funding requirements for the FTA 5311 and FTA 5307 programs after the 2020 decennial census and to identify rural areas that could become enveloped into large areas after the 2020 decennial census. The latter is important because rural transit agencies that shift to large urban areas after the 2020 decennial census will lose their ability to use federal funding for operating expenses for two years due to the "100 bus rule." Results show that amount of additional funding needed for the FTA 5307 small urban will be $344M$411M, representing a 28%33% increase over current levels. This report includes a set of appendices that any government agency can use to understand what the potential changes in FTA funding after the 2020 decennial census mean to their constituents. These appendices include the predicted changes in FTA 5311 and FTA 5307 funding for each state, as well as the list of urban clusters that are predicted to grow into small urban areas and/or be absorbed into large urban areas. 17. Key Words: FTA 5311 rural funding, FTA 5307 urban funding, rural and small urban transit, trending urban, demandresponsive 18. Distribution Statement: 19. Security Classification (of this report): Unclassified 20. Security Classification (of this page): Unclassified 21. Number of Pages: 22. Price: 160 TABLE OF CONTENTS LIST OF TABLES ........................................................................................................... iv LIST OF FIGURES ......................................................................................................... vi LIST OF SYMBOLS AND ABBREVIATIONS ......................................................... viii EXECUTIVE SUMMARY .............................................................................................. x 1 INTRODUCTION................................................................................................. 1 2 LITERATURE REVIEW .................................................................................... 3 2.1 U.S. Population Trends ..................................................................................... 3 2.2 How the FTA 5311 and 5307 Programs Relate to the Decennial Census ................................................................................................................. 5 2.3 How Funding Gaps Can Occur After the 2020 Decennial Census................ 8 2.4 Summary........................................................................................................... 11 3 DATA AND METHODOLOGY ....................................................................... 12 3.1 Step 1: Map 2000 Block Level Data into 2010 Geographies ........................ 14 3.2 Step 2: Estimate Block-Level 2020 Populations Using the Shift-share Method .............................................................................................................. 15 3.3 Step 3: Obtain and Clean Variables for Binary Logit Models .................... 16 3.4 Step 4: Estimate Binary Logit Models ........................................................... 25 3.5 Step 5: Predict Urbanization Scenarios ......................................................... 29 3.6 Step 6: Calculate New Urbanized and Non-Urbanized Population and Land Area ......................................................................................................... 32 3.7 Step 7: Obtain Other Variables Used for FTA 5311 and 5307 Formulas ........................................................................................................... 33 3.8 Step 8: Predict Funding................................................................................... 38 3.9 Summary........................................................................................................... 46 4 RESULTS ............................................................................................................ 47 4.1 Results for Georgia .......................................................................................... 47 4.2 Results for States Nationwide ......................................................................... 54 4.3 Additional Analysis Conducted for Georgia ................................................. 81 4.4 Summary........................................................................................................... 90 ii 5 FINDINGS AND RECOMMENDATIONS ..................................................... 92 5.1 Findings ............................................................................................................ 92 5.2 Recommendations ............................................................................................ 94 REFERENCES................................................................................................................ 96 APPENDIX A: State Binary Logit Models................................................................. 101 APPENDIX B: Rural Areas Predicted to Merge with Urban Clusters or Urban Areas................................................................................................................... 111 APPENDIX C: Supporting Tables for Predicted Changes in 5311 and 5307 Funding Allocations .......................................................................................... 121 APPENDIX D: Supporting Tables for Predicted Changes in 5311 and 5307 Funding Allocations by County in Georgia .................................................... 150 iii LIST OF TABLES Table Page 1. Definition and Descriptive Statistics of Variables Used in State Binary Logit Models ......................................................................................................... 17 2. Definition and Descriptive Statistics of Additional Variables Used in Georgia Model ...................................................................................................... 18 3. Variables Used to Predict 5311 and 5307 Appropriations.............................. 36 4. FTA Tables Used to Predict 5311 and 5307 Appropriations.......................... 37 5. Example 5311 Appropriation Calculation for the State of Georgia .................. 39 6. FTA Values Used for 5311 Appropriation (FY19)............................................ 39 7. Inputs for the 5307 Appropriation Calculation for the State of Georgia ........... 41 8. FTA Values Used for 5307 Appropriation (FY19)............................................ 42 9. Binary Logit Model Results for Georgia .............................................................. 49 10. Predictions of Areas in Georgia That Will Need to Transition to New FTA Funding ................................................................................................................. 52 11. Operating Funding Gaps for Georgia Counties Trending Urban.......................... 54 12. Urban Clusters Predicted to Merge Under Scenario 2B ....................................... 59 13. UCs Predicted to Grow into Small UAs Under the 50% Model........................... 62 14. UCs Predicted to Grow into Small UAs Under a 75% Model.............................. 63 15. Small UAs Close to Growing into a Large UA Under 50% Model...................... 64 16. Small UAs Close to Growing into a Large UA Under 75% Model...................... 64 17. Counties that Grew to over 50% Urbanized Population Scenario 1A ............... 70 18. Predicted Changes in 5311 and 5307 Funding After 2020 (Assumes FY19 FTA Data Values) ....................................................................................... 81 19. Georgia's Funding Outlook After 2020 (Assumes FY19 FTA Data Values) ....... 82 iv 20. Funding Outlook for Large and Small Urban Areas in Georgia (Assumes FY19 FTA Data Values) ....................................................................................... 83 21. FY19 5311 Appropriation Levels for Counties That Do Not Have Transit Service................................................................................................................... 84 22. FY19 5307 Appropriation Levels for Counties That Do Not Have Transit Service................................................................................................................... 86 23. Comparison of FY19 5311 Appropriation Levels Across Counties .................. 88 24. FY 5307 Appropriation Levels for Counties That Only Offer 5311 Transit Service ...................................................................................................... 89 25. FY19 5307 Appropriation Levels for Counties That Only Offer 5307 Transit Service ...................................................................................................... 90 v LIST OF FIGURES Figure Page 1. Percent Urbanized Population in 2010 by State in the U.S. ................................... 4 2. Percent Urbanized Land Area in 2010 by State in the U.S..................................... 4 4. Flow-chart Illustrating Overview of Methodology............................................... 13 5. 5311 Formula Grant ........................................................................................... 34 6. 5307 Formula Grant ........................................................................................... 35 7. Probabilities of Urbanized Areas in Metro Atlanta 2020 ..................................... 50 8. Urban Clusters and Urbanized Areas Expected to Merge with Atlanta After the 2020 Census .................................................................................................... 51 9. Contiguous or Near-Contiguous MPOs and Urbanized Areas in Metro Atlanta After 2020 Census.................................................................................... 53 10. FTA 5311 Apportionment Quotient for 2010 by State ...................................... 66 11. Percent Change in Non-Urbanized Population Under Scenario 1A by State Between 2010 and 2020........................................................................................ 68 12. Percent Change in Urbanized Population Under Scenario 1A by County Between 2010 and 2020........................................................................................ 69 13. Percent Change in Non-Urbanized Land Area Under Scenario 1A by State Between 2010 and 2020........................................................................................ 72 14. Percent Change in Land Area and Population Quotient Under Scenario 1A for the FTA 5311 Formula by State Between 2010 and 2020 ........................... 73 15. Percent Change in Non-Urbanized Population Under Scenario 2B by State Between 2010 and 2020........................................................................................ 76 16. Percent Change in Non-Urbanized Land Area Under Scenario 2B by State Between 2010 and 2020........................................................................................ 76 17. Percent Change in Urbanized Land Area under Scenario 2B by County between 2010 and 2020 ........................................................................................ 78 vi 18. Percent Change in Land Area and Population Quotient under Scenario 2B for the FTA 5311 Formula by State between 2010 and 2020............................ 79 19. FY19 5311 Appropriation Levels for Counties That Do Not Have Transit Service................................................................................................................... 85 20. FY19 5307 Appropriation Levels for Counties That Do Not Have Transit Service................................................................................................................... 86 21. FY19 5311 Appropriation Levels for Counties That Have Only 5311 Transit Service ...................................................................................................... 87 vii LIST OF SYMBOLS AND ABBREVIATIONS ARC BEA DOT FAST FTA GIS LEHD LODES LSAD LU MAP-21 MPO MSA NHGIS NTD PM PSQM RTAP SAFETEA-LU STIC "Section" Atlanta Regional Commission Bureau of Economic Analysis Department of Transportation Fixing America's Surface Transportation Federal Transit Administration Geographic Information System Longitudinal EmployerHousehold Dynamics LEHD OriginDestination Employment Statistics Legal/Statistical Area Description Large Urban Moving Ahead for Progress in the 21st Century Metropolitan Planning Organization Metropolitan Statistical Area National Historical Geographic Information System National Transit Database Passenger Mile People per Square Mile Rural Transportation Assistance Program Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for User Small Transit Intensive Cities viii SU TIGER UA UACE UC VRH VRM Small Urban Topologically Integrated Geographic Encoding and Referencing system Urbanized Area Urban Area Census Urban Cluster Vehicle Revenue Hour Vehicle Revenue Mile ix EXECUTIVE SUMMARY In the United States (U.S.), there are two main sources of transit funding administered by the Federal Transit Administration (FTA). The amount of FTA funding a transit agency receives depends, in part, on whether the agency serves an area that is designated as rural, small urban, or large urban. These designations are defined using the most recent decennial census. Between 2000 and 2010, the percentage of the U.S. population residing in urban areas increased by over 12% (U.S. Census Bureau 2011b). Population forecasts suggest these trends will continue and will be reflected in the 2020 decennial census. This project examines how spatial and temporal changes in the U.S. population will impact funding for transit systems in the U.S. after the 2020 decennial census. We use binary logit models and geographic information system (GIS) methods to predict spatial and temporal population changes between 2010 and 2020 and identify which areas of the U.S. will be classified as rural, small urban, or large urban after 2020. We then use this information to forecast how appropriations for FTA 5311 (rural) and 5307 (urban) formula funding programs could change after the 2020 decennial census. These forecasts are summarized in Table ES-1. The estimates in Table ES-1 assume that the current appropriation formulas and FY19 FTA data values are used; the FTA data values convert each input used in the appropriation formula into a dollar amount, e.g., each rural person translates to $4.72 in appropriation dollars. Under these assumptions, the total amount of federal funding needed to support transit after 2020 using today's funding appropriation formulas is basically unchanged; x however, there would be large changes within the individual programs. These changes are reflective of population trends in the U.S.: the outward expansion of cities and overall decline of rural populations results in additional transit support that is needed for small urban areas (defined as those with populations between 50K and 200K) and large urban areas (populations between 200K and 1M). TABLE ES-1 Summary of Predicted Changes in FTA 5311 and 5307 Funding After 2020 Funding Source Current Predicted (Population) Appropriation* Appropriation Difference % Difference 5311 rural (<50K) 629M 483 to 505M -124 to -146M -20 to -23 5307 small urban (50K200K) 5307 large urban (200K1M) 5307 large urban (1M+) 402M 839M 3.38B 550 to 608M 148 to 206M 1.035 to 1.044B 196 to 205M 3.00 to 3.06B -316 to -358M 37 to 51 23 to 24 -9 to -11 TOTAL 5.25B 5.13 to 5.16B -118 to -71M -1.4 to -2.2 *Note: The numbers reported on the table above do not include the 5340 growing states portion in the totals. As part of this study, we also conduct an in-depth analysis of those rural areas in the U.S. that are trending urban and show how the rapid low-density urbanization of places that were previously designated as rural is not fully contemplated in current transportationplanning regulations. Due to the geographic expansion of metropolitan areas, many cities and counties that were classified as rural (or non-urbanized) in the 2010 decennial census could become enveloped into large urban areas after the 2020 decennial census. This is important because rural transit agencies that shift to large urban after the 2020 decennial census will lose their ability to use federal funding for operating expenses for two years after the appropriation of funds based on the 2020 decennial census and will see significant xi reductions in years three and beyond (FTA 2015). The "100 bus rule" creates this effect as it caps federal funding for large urban systems, whereas rural systems can use all of their federal transit funding to help cover operating expenses (FTA 2015). The loss of operations funding could be challenging for rural transit systems, especially for those that do not receive any local funding support. The ultimate goals of this research are to: (1) help rural transit agencies, state departments of transportation, and metropolitan planning organizations prepare for potential funding changes after the 2020 decennial census; and (2) promote regulatory reform that more fully considers the "trending urban" issue when considering federal funding for transit operating expenses. To help facilitate these goals, this report includes a set of appendices that any government agency can use to understand what the potential changes in FTA funding after the 2020 decennial census mean to their constituents. These appendices include the predicted changes in FTA 5311 and FTA 5307 funding, as well as the list of urban clusters that are predicted to grow into small urban areas and/or be absorbed into large urban areas. xii ACKNOWLEDGMENTS We thank the Georgia Department of Transportation for their support, particularly Carol Comer, Leigh Ann Trainer, Nancy Cobb, Kaycee Mertz, Sunil Thapa, and Supriya Kamatkar for their helpful input. We are also grateful to those who participated in the American Association of State Highway and Transportation Officials (AASHTO) Council on Public Transportation & Multi-State Transit Technical Assistance Program (MTAP) 2019 Winter Meeting in Santa Fe and to Yvette Taylor, Robert Buckley, and Federal Transit Administration (FTA) Region IV staff who provided input into this report. Finally, we offer our thanks to Sharon Dunn who copy-edited this report. xiii 1 INTRODUCTION Urbanization in the United States (U.S.) has greater impacts on federal funding for public transit than may be evident. The funding implications of urbanization include shifts in overall rural transit funding by states, a reduced number of permitted expenses for the transit agencies (i.e., a loss in operating expenses, such as fuel or operator salaries), and increased reporting requirements to the National Transit Database (NTD) (FTA 2015). These implications have the power to present serious challenges to current rural transit systems that will be located in newly urbanized areas after the 2020 decennial census. This report examines these implications and, particularly, addresses urbanizing rural areas and how public transit funding through the Federal Transit Administration's (FTA) 5311 and 5307 formula funding programs will be affected as a result of urbanization. Transit systems located in fast-growing non-urbanized1 areas, and non-urbanized areas that are subject to envelopment by adjacent urbanized areas are of particular focus for this analysis. To date, very few research publications have explored the intersection of urbanization of rural areas and federal funding for transit. As such, it is our hope that the research findings presented in this analysis will be useful in highlighting issues that urbanization can have on FTA rural transit funding and help states better prepare for potential changes after the 2020 decennial census. 1 In this report, we use the terms rural and non-urbanized interchangeably. 1 This report contains five chapters. Chapter 2 explores urbanization trends in the U.S. since 2000; clarifies definitions of "urban" and "rural"; and provides background on FTA's 5311 and 5307 formula funding programs, and describes the allocation process for each. Chapter 3 describes the data and methodology we used to forecast land use changes and resulting implications on statewide transit funding. Chapter 4 presents the results, which include a forecast of how state-level appropriations for FTA's 5311 (rural) and 5307 (urban) formula funding programs will change after the 2020 decennial census and funding gap estimates for individual transit providers and counties in Georgia. The analysis is accompanied by several technical appendices of results that will be of particular interest to state agencies for understanding how population trends in their states could impact their constituents. Chapter 5 summarizes the key findings and offers recommendations for regulatory changes that would help transit agencies make a smoother transition from a system that operates in a rural area to one that operates in a large urban area. 2 2 LITERATURE REVIEW This chapter reviews population growth trends in the U.S., describes key differences in funding for rural and urban transit systems, and explains how operating funding gaps can occur for rural transit systems absorbed into large urban areas after the 2020 decennial census. 2.1 U.S. Population Trends In the first decade of the twenty-first century, the total population in the U.S. grew by 9.7% (U.S. Census Bureau 2011b). Additionally, the overall urbanized population in the U.S. grew from 79.0% to 80.7% and the overall rural population decreased from 21.0% to 19.3%. Also, during this decade, there was a 19% increase in overall urbanized land area (U.S. Census Bureau 2015). Maps showing the distribution of urbanized population and land area in 2010 are illustrated in Figure 1 and Figure 2, respectively. The states with the highest percentage of urbanized population include New Jersey (92%), Rhode Island (90%), Massachusetts (90%), and California (90%), with Vermont (17%), Wyoming (25%), Montana (26%), and Maine (26%) as the states with the lowest percentage of urbanized population. The percent of urbanized land area by state is markedly lower than urbanized population, meaning that urbanized population is concentrated to geographic areas within the state. Urbanized land area ranges from less than 1% in 10 states (most of which are in the western U.S.) to over 37% in the eastern states of New Jersey and Rhode Island. 3 Sources: Mapchart.net 2019; U.S. Census Bureau 2010, 2017. FIGURE 1 Percent Urbanized Land Population in 2010 by State in the U.S. Sources: Mapchart.net 2019; U.S. Census Bureau 2010, 2017. FIGURE 2 Percent Urbanized Land Area in 2010 by State in the U.S. 4 This trend of urbanization in the U.S. is not new; in fact, the U.S. has been urbanizing since 1830, with a short respite between 1930 and 1940 (Boustan, Bunten, and Hearey 2013). These urbanization trends are expected to continue through 2020 and beyond. The objective of this project is to predict: (1) the changes in urbanized population and land area that will be reflected in the 2020 decennial census, and (2) how funding for FTA's 5311 and 5307 programs will also change. 2.2 How the FTA 5311 and 5307 Programs Relate to the Decennial Census The FTA provides funding for public transit systems through the Fixing America's Surface Transportation (FAST) Act, signed into law in 2015 (FTA n.d. (b)). Through the FAST Act, eligible entities can apply for dozens of competitive or formula grants (FTA n.d. (a)). Two of the largest programs are the FTA 5307 Urbanized Area Formula Funding program and the FTA 5311 Formula Grants for Rural Areas. The FTA and the U.S. Census Bureau use similar criteria to define rural and urbanized areas, but there are important distinctions between their definitions. Using block-level geography, the U.S. Census Bureau defines areas with a population under 50,000 as urban clusters (UCs) and areas with a population over 50,000 as urbanized areas (UAs) (U.S. Census Bureau, 2011a), and all others as rural. The FTA defines eligibility for 5311 (rural) and 5307 (urban) programs using the most recent decennial census. Eligibility for the 5311 (rural) grants includes those areas with populations less than 50,000; this includes UCs and rural areas, so some places classified as UCs by the U.S. Census Bureau are rural for the FTA. The FTA defines eligibility for the 5307 (urban) grants as those areas classified as contiguous urbanized areas with populations greater than or equal to 50,000 (49 USC 5 5302(23)). Any area with a population less than 50,000 is classified as non-urbanized (49 USC 5311) (FTA 2018, 2019g, 2019h). So, the primary programmatic binary is urbanized and non-urbanized (smaller than 50,000), but the 5307 program also distinguishes between small urbanized areas (small UAs) with a population of at least 50,000 and 199,999, and large urbanized areas (large UAs) with populations of 200,000 or more (FTA 2017, 2018). The shift from rural to urban after a new decennial census can present planning challenges. Reporting requirements are markedly higher, and the ability to use FTA funding for operating expenses is more limited for systems serving urbanized areas with populations over 200,000 than for 5307 small urbanized area systems and non-urbanized areas (FTA 2017, 2018). Further, because 5311 and 5307 small urbanized funds are appropriated to state governors while 5307 large urbanized funds are apportioned directly to regional recipients (such as metropolitan planning organizations [MPOs] or operators), rural transit operators that are absorbed into a large urban area need to coordinate directly with the local MPO. With respect to reporting requirements, the 5307 and 5311 programs require different levels of reporting to the National Transit Database [NTD]. As of FY18, all transit systems, regardless of type of funding, are required to report operational, service, fleet, and asset management information to the NTD (49 CFR 5307). Reporting requirements are fewer for 5311 recipients and are typically completed by the state department of transportation (DOT), whereas 5307 recipients usually self-report their data directly to the NTD. This level of reporting requires extensive metrics tracking and a dedicated staff to compile and 6 submit the data, which could be taxing on a rural transit system that is newly urbanized if staff resources are limited. With respect to operating expenses, in the 5311 program rural transit operators are permitted to use up to 100% of their FTA funding on eligible operating expenses. Under the 5307 program, recipients are not permitted to use FTA funds for operating expenses except under the stipulations set forth by the "100 bus rule" that was introduced under the Moving Ahead for Progress in the 21st Century (MAP-21) legislation passed in 2012 and slightly modified with the FAST Act passed in 2015. Grants under the 5307 (urban) program can be used to finance operating costs for public transportation systems that "operate 75 or fewer buses in fixed route service or demand responsive service, excluding complementary paratransit service, during peak service hours, in an amount not to exceed 75% of the share of apportionment which is attributable to such systems within the urbanized area, as measured by vehicle revenue hours" (FTA 2017). A similar rule applies to systems that operate a minimum of 76 and maximum of 100 buses; in this case, no more than 50% of the share of apportionment can be used toward operating costs. Total urban vehicle revenue hours (VRHs) are used to determine the portion of funding that can be used toward operating expenses. Note that federal regulations explicitly dictate that the share of apportionment is to be determined using VRHs based on reporting to the urban National Transit Database two years beforehand. This is the regulatory language that will give rise to operating funding gaps after the 2020 decennial census. 7 2.3 How Funding Gaps Can Occur After the 2020 Decennial Census Funding gaps after the 2020 decennial census can occur due to distinctions between the generation (or apportionment) and the allocation of funds. The total amount of 5311 funding a state DOT receives for its rural transit system depends on various factors defined in the 5311 formula grant program, i.e., funds are apportioned based on population, land area, low-income population, and vehicle revenue miles (VRMs) criteria. However, the state DOT determines how to allocate these funds to particular transit agencies and can use different factors than those used in the allocation formula.2 A similar system exists in the case of small urbanized systems, but large urbanized systems are different because their funding also depends on historical VRMs reported to the NTD, and they are restricted in how much they can spend on operations by the 75 bus rule or 100 bus rule. To understand how a funding gap occurs after the 2020 census, consider a rural transit system that learns its previously rural, or non-urbanized, service area has merged with a large urbanized area. The transit system can no longer apply for funding under the 5311 program administered by its state DOT and must apply for funding under the 5307 program and coordinate with its local MPO or other regional transit agency (FTA 2016, 2018). The MPO is responsible for determining how to allocate 5307 funds to eligible recipients. If the MPO has a policy in which it allocates funds based solely on the urban 2 For example, consider two transit systems that generate the same amount of 5311 funds and provide the same number of trips. One agency serves a population that is 100 miles from a major medical facility, whereas the other agency serves a population that is about five miles from a major medical facility. Logically, operating costs will be much higher for the first agency, and the state DOT can allocate more funds to this transit system (even though the two systems would have generated identical funds under the formula). 8 funds that the agencies had generated for the 5307 program, the formerly rural transit system--although now permitted to receive 5307 funds--would not be eligible under the MPO's policy, since it has not yet generated any funds for the urban program.3 Importantly, even if the MPO allocated funding to the transit agency as soon as it became eligible for 5307 funds, the transit agency could not use any of its 5307 funds for operating expenses. The 5307 formula program uses (urban) VRHs to determine the amount of 5307 funding that can be used toward operating expenses, but since the transit system has not yet provided service under the 5307 program, it does not have any urban VRHs to report. Consequently, the formerly rural transit system would need to provide service for two years before it could receive 5307 operating funds. This two-year lag occurs until the NTD is able to certify the accuracy of urban service data and calculate the amount of apportionment that can be used toward operating expenses. Thus, a system that operates in urbanized areas in 2020 would submit these data to the NTD in 2021, and the certified NTD data could then be used in the 2022 allocations. We classify those rural transit systems that will have their service areas transition to urban areas after the 2020 census into one of three categories. High-risk transitions occur when a rural transit system is absorbed into a large urbanized area. Transit systems in this category are at risk of losing federal operating assistance for two years (due to the 75 or 100 bus rule) and experiencing a reduction in operating assistance after year two. Mediumrisk transitions occur when a small urbanized transit system grows into a large urbanized system. Transit systems in this category are at risk of seeing a reduction in operating 3 In 2018, this is how the Atlanta Regional Commission (ARC) distributed 5307 urban funds within the Atlanta metro area. (Atlanta Regional Commission 2018). 9 assistance (as competition for large urbanized funding tends to be higher than for the rural and small urbanized program, and only small urbanized systems are eligible for Small Transit Intensive Cities [STIC] funding). Low-risk transitions occur when a rural system grows into a small urbanized system. Transit systems in this category may still use their FTA 5311 funds for operating expenses. Figure 3 shows the different types of transitions. Rural Small Urbanized Large Urbanized FTA 5311 Pop. < 50K All FTA funds can be used for operating expenses (with match) State DOTs usually administer funds Simplest NTD reporting requirements FTA 5307 50K Pop. < 200K State DOTs usually administer funds NTD reporting similar to rural systems Eligible for STIC funds Rural areas or UCs grow into small urban areas Low-risk transition FTA 5307 Pop. 200K Transit agency usually receives funds directly; must coordinate with local MPO Extensive NTD reporting requirements 75 and 100 bus rules determine if 5307 funds can be used for operating expenses (with match). Max amount determined by 49 USC 5307(a). Small urban areas grow into large urban areas and/or merge with a large urban area Medium-risk transition Rural areas or UCs merge with a large urban area High-risk transition FIGURE 3 How the 2020 Decennial Census Can Impact Funding for Rural and Small Urbanized Transit Systems 10 2.4 Summary The continued outward expansion of urban areas, combined with the distinct eligibility criteria for FTA 5311 (rural) and 5307 (urbanized) programs could lead to systematic funding challenges throughout the nation after the 2020 decennial census. If a system loses its 5311 rural funding, it can theoretically just transition into 5307 funding after it is urbanized. Although this sounds like a simple and feasible solution, there are several limitations: 1. The MPO, which now determines allocation for the newly urbanized transit system, could choose not to allocate any funding to the operator "because it had not yet generated any funds for the urban program." 2. Although the MPO did allocate funding to the newly urbanized transit system, the transit agency could not use any of its 5307 funds for operating expenses. This is because the system has not yet generated any urban vehicle revenue miles, which is one of the inputs that is used to determine how much funding can be used toward operating expenses. 3. Additionally, it takes two years for the NTD to certify and adjust funding after receiving reported urban vehicle revenue miles from the newly urbanized system (FTA 2015). These rules and regulations leave transit systems in newly urbanized areas in a tight spot, unable to return to 5311 funding or initiate 5307 funding. 11 3 DATA AND METHODOLOGY This chapter reviews the data and methodology we used to forecast population and land use in 2020. We used these two metrics to quantify likely implications for transit funding after the 2020 decennial census. First, we identified those areas throughout the nation that we consider to be at high risk or medium risk for losing federal funding. High-risk areas include urban clusters (classified as "rural" by the FTA) absorbed into a large urbanized area. Medium-risk areas include small urban areas that grow into or merge with a large urban area. Next, we performed a deeper assessment of county-level funding implications for the state of Georgia, and predicted operating funding gaps for specific counties and transit operators that are in areas trending urban. Finally, we forecast overall changes in the state-level 5311 (rural) and 5307 (urban) FTA appropriations. Figure 4 provides an overview of our methodology. Because the service area (or counties) served by individual rural transit providers is not collected as part of the NTD, we could only calculate overall funding levels for the 5311 and 5307 programs at the state level. However, we did have this information for Georgia, and, thus, could predict overall funding levels and operating funding gaps for individual counties and rural service providers. Because the outward growth in the Atlanta metro area is spilling over into other metro areas, we visually inspected the results from the binary logit models and assigned new urban blocks to the most appropriate metro area using a grandfathering clause the U.S. Census Bureau applied in 2010. The key distinction between the left and right sides of the flowchart in Figure 4 is that we were able to do a more refined analysis for Georgia. 12 Step 1: Map 2000 block level data into 2010 geographies Step 2: Project block population for entire U.S. to 2020 using shift-share projection method Step 3: Obtain and clean variables for regression models Step 4A: Run regression models by state to predict urban/rural classification for blocks in 2020 and obtain summary statistics Step 4B: Run more refined regression models for Georgia to predict urban/rural classification for blocks in 2020 and obtain summary stats Step 6A: Sum, by state, predicted urbanized and nonurbanized population and land area under each scenario to obtain inputs for FTA 5311 & 5307 formulas Step 7A: Obtain other variables used for FTA 5311 & 5307 formulas Step 5A.1: Apply four scenarios for merging based on probability (50% or 75%) and distance to an existing urbanized area (0 or 0.5 miles) Step 5A.2: Identify UCs, by state, that will remain UCs, grow into UAs, or merge with another UA Step 6B: Sum, by county, predicted urbanized and nonurbanized population and land area under each scenario to obtain inputs for FTA 5311 formula Step 7B: Obtain other variables used for FTA 5311 & FTA 5307 formulas Step 5B: Identify UCs in Georgia that will remain UCs, grow into UAs, or merge with another UA by visually inspecting blocklevel regression results Step 8A: Predict, by state, 5311 & 5307 appropriations after 2020 census FIGURE 4 Step 8B: Predict, by county, 5311 & 5307 appropriations and funding gaps for high-risk transitions after 2020 census Flow-chart Illustrating Overview of Methodology 13 At this stage in the discussion, it is helpful to distinguish between the binary logit model used for calibration (or obtaining binary logit model coefficients) and the binary logit model used for prediction (or forecasting whether a tract will be urban or rural in 2020). Note that in the calibration model, we used data from 2000 and 2010 (as all of the variables are known). In the forecasting model, we used data from 2010 and 2020. However, since the variables for 2020 are not known, we needed to prepare forecasts for the 2020 variables in order to use the binary logit model. The steps shown in Figure 4 before the binary logit model were focused on preparing data inputs. We discuss each of the modeling steps shown in Figure 4 in detail in the sections that follow. The majority of the data cleaning and analysis steps described in this chapter were conducted in R Studio, and complemented by ESRI's ArcMap 10.5.1 (R Core Team 2018; ESRI 2017). 3.1 Step 1: Map 2000 Block Level Data into 2010 Geographies To predict whether a block will be urban or rural after the 2020 decennial census, we estimated binary logit models that were based on 2000 and 2010 data. That is, we first predicted the land use status (rural or urban) as of 2010 by using 2000 data as an input. However, because the geographies representing blocks between decennial censuses change, we first needed to map 2000 data into 2010 geographies using a cross-walk file from the National Historical Geographic Information System (NHGIS) (Manson et al. 2019). The cross-walk file essentially provides information on how blocks from the two censuses relate to each other. We used the `stringr' package in R for mapping the 2000 block-level data into the appropriate 2010 geographies (Wickham 2018). Excluding 14 Puerto Rico and the Island Areas, in 2000, the U.S. had a total of 8,205,582 blocks in the U.S., whereas in 2010, it had a total of 11,078,297 blocks (U.S. Census Bureau 2001; 2011a). 3.2 Step 2: Estimate Block-Level 2020 Populations Using the Shift-share Method After binary logit models were estimated, we used the coefficients from these models, along with updated variables, to predict the probability that a block will be urban in 2020. Thus, as part of the initial data pre-processing, we needed to estimate block-level 2020 populations so that we could forecast input variables related to population and population densities as of 2020. The shift-share projection method is a type of ratio time-series model that is used to project population (or employment) to a given year for a geographic area using a larger geographic reference area (Smith, Tayman, and Swanson 2001). Equation 1 is the formula for the shift-share method, where P = population, i = smaller area (census block), j = larger area (census block group), z = number of years in the projection horizon, y = number of years in the base period, b = base year, l = launch year, and t = target year. (1) To obtain the 2020 projected block population, the census block group populations for 2015 and 2020 were purchased from ESRI and used for the larger geographic reference area (ESRI 2015). Since block-level population data are not available in between decennial census years, we first projected the data to 2015 using a base year of 2000 and a launch year of 2010 (z = 5 years; y = 10 years). We then used the output from the 2015 projection 15 to obtain a 2020 block-level projection using a base year of 2010 and a launch year of 2015 (z = 5 years; y = 5 years). One caveat of the shift-share projection method is that blocks with declining or slowgrowing population during the base period can result in a negative population projection (Smith, Tayman, and Swanson 2001). To correct for these blocks with negative population, the negative population was summed by the block group, and was then subtracted evenly from the blocks with population greater than `0'. 3.3 Step 3: Obtain and Clean Variables for Binary Logit Models Extensive literature is available related to predicting land use (which in our case is predicting whether a tract will be urban or rural after the 2020 decennial census). Table 1 shows the variables that we included in the binary logit models for all of the states, and Table 2 shows two additional variables we included in the Georgia model. Both tables show descriptive statistics for these variables for the state of Georgia. The final model for land use change included variables measuring population density in a block, the distance of the block to urban areas (UAs and UCs), the number of jobs in the census tract, whether a block was nearest to a UA or UC, and the distance to primary and secondary roads. Because population densities and the distance to urban areas were not normally distributed, we created dummy variables for different density levels and distances. We used the log of the distance of each block to city centers (referred to in other studies as activity centers) and all primary and secondary roads, but not local streets. 16 TABLE 1 Definition and Descriptive Statistics of Variables Used in State Binary Logit Models Variable 2000 urban rea / urban cluster (UA/UC) 2010 UA/UC Closest urban is an urbanized area Distance to nearest urban area (UA or UC) Distance from block border to closest road Jobs in census tract (2010) Population density Definition and Descriptive Statistics Indicator variable equal to 1 if block was classified as an urban cluster or urban area in 2000, 0 otherwise. We use the population of contiguous urban blocks to calculate if the block belongs to an urban cluster or area. Of 333,150 blocks, 147,039 (40%) were urban in 2000, after removing blocks that were water and rural protected areas. Definition same as for 2000 UA/UC (above), but for 2010. Of 333,150 blocks, 167,987 (45.9%) were urban in 2010. Definition is based on the 2000 and 2010 UA/UC categories (above). Indicator variable equal to 1 if the closest planar distance of a block is to either an urban cluster or urbanized area. In the global model, 162,905 (44.5%) blocks were in this category; in the rural-only model, 50,331 (29.3%) blocks were. Distance (in miles) to the nearest UC or UA as of 2000. We use the centroids of blocks and urban areas for calculating distances. In the global model: urban area (N=147,039, 40.0%); rural and (0,1] miles from UA (N=39,092, 10.7%); rural and [12) miles (N=17,746, 4.8%); rural and (2,3] miles (N=15,563, 4.3%); rural and (3,4] miles (N=14,948, 4.1%); rural and 4+ miles (N=131,762, 36.1%). In the rural model: borders an urban area (all have a distance of (01] miles) (N=38,318, 22.3%); (12] miles (N=17,746, 10.3%); (34] miles (N=15,563, 9.1%); 4+ miles (N=100,108, 58.3%). In the global model, the mean closest distance to a road is (1.41), with a max of (11.43), min (0), and std. dev of (1.5) miles. In the rural-only model, the mean is (1.3), with a max of (6.7), min (0), and std. dev of (1.54) miles. Because these data are skewed, we logtransformed them. In the global model, the mean is 879, with a max of (40,117), min of (0), and std. dev of (3.79); in the rural-only model, there is a mean of (5.74), max of (5,022), min of (0), and std. dev of (49.7). Because these data are skewed, we log-transformed them. Number of people per square mile (PSQM) at the block level in 2010. In the global model: 500999 PSQM (N=23,227, 6.3%); 10001499 PSQM (N=17,256, 4.7%); 15001999 PSQM (N=14,361, 3.9%); 20003999 PSQM (N=38,520, 10.5%); 4000+ PSQM (N=32,876, 9.0%). In the rural-only model: 500999 PSQM (N=9,943, 5.8%); 10001499 PSQM (N=5,088, 3.1%); 15001999 PSQM (N=3,160, 1.9%); 20003999 PSQM (N=5,133, 3.0%); 4000+ PSQM (N=3,211, 1.9%). 17 TABLE 2 Definition and Descriptive Statistics of Additional Variables Used in Georgia Model Variable Atlanta MSA Definition and Descriptive Statistics Indicator variable equal to 1 if county was in the Atlanta metropolitan statistical area (MSA) in 2010, 0 otherwise. In the global model, 89,925 blocks (24.6%) were part of the Atlanta MSA, representing 29 counties; in the rural-only model, 30,145 blocks (17.6%) were in the Atlanta MSA. Blocks in these counties can be urban or rural. Savannah MSA Indicator variable equal to 1 if the county was in the Savannah MSA in 2010, 0 otherwise. In the global model, 7,851 (or 2.1%) of blocks were in the three counties comprising the Savannah MSA; in the rural-only model, 2,635 (1.5%) blocks were in this MSA. 3.3.1 Population Density Block population density was included in the model because of its importance in the U.S. Census Bureau's rules in determining a block's urban/rural classification. Block densities were calculated by dividing the population of the block by the block's total land area in square miles for 2010. This was completed through the use of several functions, including mutate, join, group by, and summarize, within the `dplyr' packages (Wickham, Franois, et al. 2018). These densities were then classified into the following groups: 500 to less than 1000 people/mi2 (psm) 1000 to less than 2000 psm 2000 to less than 4000 psm Greater than or equal to 4000 psm These groups follow the aforementioned population density U.S. Census Bureau thresholds for urban blocks: a block is urban if it has a population density of 1,000 psm or is located near an urban core and has a population density of 500 psm (Federal Register 2011). If a 18 block fell into one of the categories, it was coded as `1' for its respective category and as `0' for the other categories. For the forecasting model, we used the projected block-level population estimates (calculated in step 2) for the corresponding 2020 fields. 3.3.2 Distance to an Existing UC or UA Along with block density, a block's distance to an existing UC or UA was used in the binary logit model. The distances for each block to the closest UC or UA was executed using the Near Analysis tool in ESRI's ArcMap 10.5.1 (ESRI 2017). Distance was calculated as the distance between the block's centroid to the border of the UA/UC, rather than to the center of the UA/UC. The distances (in miles) for each block were then classified into the following groups: Rural block and less than 1 mile from a UA/UC Rural block and less than 2 miles from a UA/UC Rural block and less than 4 miles from a UA/UC Urbanization is not always contiguous, and can be segmented by roads, commercial development, or other structures (Ratcliffe et al. 2016). To account for this fact, the U.S. Census Bureau has a rule to account for these "jumps" and "hops" in urbanized land area. Jumps refer to areas spanning 2.5 miles along a road corridor, while hops refer to areas spanning no more than 0.5 mile. Under 2010 U.S. Census Bureau criteria, non-contiguous areas were subject to these rules, allowing for multiple hops, but no hops after jumps (Federal Register 2011). 19 As with the density categories, each block was assigned a `1' in its respective distance category and `0' for the other categories, making it binary for the binary logit input. Distances less than 3 and 4 miles were grouped because they are outside the distance of hops and jumps, but near enough to be vulnerable to conversion and urbanization. The calibration binary logit models used the UA/UC classifications from 2000, while the forecast binary logit models used the 2010 classifications. 3.3.3 Closer to a UA versus a UC In calculating the distance variables, the classification of the nearest area was also obtained with the use of the Near Analysis tool within ArcMap 10.5.1 (ESRI 2017). This is another variable that is used in U.S. Census Bureau criteria for defining urban and rural classification at the block level. That is, whether the nearest area was classified as a UC in 2000 or a UA in 2000 (2010 for the 2020 prediction model). The numbers `75' and `76' represent UAs and UCs, respectively, under the Legal/Statistical Area Description (LSAD) Codes. Each UA and UC in the U.S. is assigned a unique 5-digit Urban Area Census (UACE) Code, and is assigned an LSAD classification by the U.S. Census Bureau every decennial census. If the nearest area was listed as a UA (or LSAD 75), the variable was coded as `1', whereas blocks that were nearest to a UC (or LSAD 76) were coded as `0'. The assumption here is that if a block is closer to a UA rather than a UC, it is more likely to transition urban. The calibration binary logit models used LSAD classifications from 2000, while the forecast binary logit models used 2010 LSAD classifications. 20 3.3.4 Distance to Primary and Secondary Roads A block's proximity to the nearest primary and secondary roads is not a U.S. Census Bureau criterion for determining urban/rural classification, but was used as another indicator of urbanization. The nearest distance to either a primary or secondary road was also calculated using the Near Analysis tool in ArcMap 10.5.1 (ESRI 2017). The 2016 primary and secondary roads shapefile (local roads were not included) for the entire U.S. was downloaded from the TIGER/Line Shapefiles database (i.e., the Topologically Integrated Geographic Encoding and Referencing system) for the U.S. Census Bureau using the `tigris' R package (Walker 2018). The data were then written to a shapefile using the `sf' R package (Pebesma 2018). Data for 2016 were used rather than the 2015 roads dataset because there were no data in the 2015 file for Georgia. The assumption was made that the road network in 2020 will be similar to that in 2016. A maximum search distance of 10 miles was used in generating the near table in ArcMap. The output provided a value in miles for every block's distance to either a primary or secondary road. The distances were then natural log transformed to create binary inputs for the binary logit models. We used these logged distances for both the calibration and forecast binary logit models, as we did not expect these to fluctuate (or many new roads to have been built over the decade). 3.3.5 Census Tract Jobs Total employment at the tract level was used as a proxy for land cover (which reveals nonpopulated urbanized areas such as airports or industrial parks). The land cover shapefile has not been updated since 2011, so it was not used in the analysis for potential lack of non-representativeness of the current land cover. Instead, the employment variable was used to predict non-residential urbanization. In other words, employment data can reveal 21 tracts that contain activity, but that may not have population within the blocks that comprise the tracts. The employment data used for the binary logit models included the total number of jobs at the U.S. Census Bureau tract level, and were retrieved from the Census Bureau's LEHD (Longitudinal EmployerHousehold Dynamics) OriginDestination Employment Statistics (LODES) datasets using the `lehdr' R package downloaded through the help of the `devtools' package (Wickham, Hester, et al. 2018; Green and Mahomoui 2017). Tracts are only included in the LODES dataset if the tract contains at least one job. The 2010 data were used in the calibration model. The 2010 and 2015 LODES data were used to project total tract jobs to 2020. In 2010, a total of 72,527 tracts (99.3% of all U.S. tracts) contained jobs compared to 2015, in which 72,585 tracts (99.4%) of all U.S. tracts contained jobs (U.S. Census Bureau 2015). The 2010 jobs data were available for all states except for Massachusetts, for which LODES data begin in 2011. Because of this, 2011 jobs data for Massachusetts were used in place of 2010. Similarly, 2015 jobs data were available for all states except for Wyoming, for which LODES data are only available through 2013. To obtain an estimate of 2015 jobs for Wyoming, the state employment growth rate between 2010 and 2015 (5.4%) from the Bureau of Economic Analysis (BEA) was applied to the total number of jobs in the tract (U.S. BEA 2017). To project employment data for each tract to 2020, the crude growth rate was first calculated. For the 58 tracts in the dataset that grew from containing zero jobs in 2010 to containing one or more jobs in 2015, a total of `1' was assigned to the 2010 tract in order to calculate the growth rate. Next, the 75th percentile for the growth rate was obtained as 22 0.379. For tracts with a growth rate within 37.9%, a compound interest rate formula was applied to project jobs to 2020 (Equation 2). The formula is as follows: 1 (2) where A = jobs by tract in 2020, P = total jobs by tract in 2010, r = calculated growth rate in 20102015, n = total times growth rate is compounded (1), t = number of years (1) (Stapel 2012). A value of `1' is used for n and t because the growth rate is already based on a five-year period. For tracts with growth rates outside of the 75th percentile, the compound interest rate formula was not applied, as the formula would yield an unrealistic projection for 2020 jobs for those tracts that had dramatic increases or decreases in jobs. Instead, for these tracts, the total number of jobs in 2015 was either doubled or halved depending upon if the growth rate was positive or negative, respectively. Finally, after obtaining a projection for total number of jobs in 2020 for all 72,585 tracts, the variable was log transformed. While the projected jobs variable was used in the 2020 state binary logit models, the 2010 binary logit model utilized the logged 2010 jobs dataset. 3.3.6 Urban or Rural Classification for Census Blocks Census 2010 block population data included an urban/rural classification variable (`URBRURALA'), which indicated if a block was considered to be urban or rural. The 20002010 cross-walk file from NHGIS was used to determine if a 2000 block was urban or rural in 2010 geography (Manson et al. 2019). In completing the cross-walk, two new population variables were created to yield the total urban and total rural population within 23 a block. Logic statements were used to classify the block as either urban or rural, as follows: (1) if the total urban population within the block exceeded the total rural population, then the block was coded as urban in 2000; (2) if the total rural population within the block exceeded the total urban population, then the block was coded as rural in 2000; and (3) if there was a `0' population value for both the urban and rural variables, the distance variable was used to provide a urban or rural classification. If the distance of the block to an existing UA or UC was 0 miles, then the block was coded as urban; if the distance was greater than 0, then the block was coded as rural. The 2000 data were used in the calibration binary logit model as the dependent variable. The 2010 data were used to measure the forecasting accuracy of these binary logit models and as input into the forecasting model (used to predict urban status in 2020). 3.3.7 Metropolitan Statistical Area (MSA) Growth The Georgia urbanization model included the Atlanta and Savannah MSAs, which house the state's fastest growing counties. This, in turn, improved the model's explanatory power for urbanization for the blocks with the MSAs. We included an indicator variable equal to `1' if the county was in the Atlanta MSA in 2010, `0' otherwise. Similarly, we included an indicator variable equal to `1' if the county was in the Savannah MSA in 2010, and `0' otherwise. Because these are county-level measures, we used the same set of variables for both the calibration and forecasted models. 24 3.4 Step 4: Estimate Binary Logit Models To predict whether a block would be urban or rural in 2020, we used binary logit models, which are mathematically equivalent to logistic logit models, to estimate the probability a block will be urban in 2010. Formally, we define: 1, 2010 0, 2010 We use a binary logit model to predict whether a block is urban or rural in 2010. The binary logit models a choice among J=2 alternatives where we define alternative 1 to be urban and alternative 2 to be rural. The utility obtained from alternative i J is , which is decomposed as where is defined as the representative utility and is defined as an error term. Further, we may assume that 0 (or that the utility for the rural alternative is zero) since utilities are nominal. That is, for identification purposes it is necessary to set the utility of an alternative to a constant. Then, the probability of choosing alternative i can be represented as a binary logit model where the difference in 's is logistic: 3.4.1 Step 4A: Estimate Binary Logit Models for Each State We estimated a binary logit model for each state using the following utility equations: 25 UC or UA in 2000 Closest urban is an urbanized area in 2000 Rural and 0,1 miles from UA in 2000 Rural and 1,2 miles from UA in 2000 Rural and 2,3 miles from UA in 2000 Log of distance to road in 2010 Log of number of jobs in tract in 2010 population density 500,1000 in 2010 population density 1000,1500 in 2010 population density 1500,2000 in 2010 population density 2000,4000 in 2010 population density of 4000 or more in 2010 0 (3) The models were estimated for each state using the variables described above to predict if a block was urban or rural in 2010 (a known variable from the 2010 decennial census). Each of the state models was fitted to accurately predict the urban 2010 variable. Accuracy for all models was 90% or greater (see Appendix A for model accuracies by state). After obtaining accuracy for each of the models, the urban 2020 variable was predicted using a combination of 2010 and forecasted 2020 variables. We developed a set of urbanization scenarios, described in Section 3.5, based on whether the probability a block would be urban in 2020 was at least 50% or at least 75%. 26 The binary logit models were tailored to each state to ensure that the coefficients were monotonically increasing or decreasing (e.g., we would expect that the probability of urbanization would decrease as you move further outside of an urban area). In addition, for the state of Georgia, we ran an additional binary logit model that included two additional variables for whether a block as of 2010 was in the Atlanta MSA or the Savannah MSA. We used the `glm' and `predict' functions included in the `stats' package in R Studio for this part of the analysis (R Core Team 2018). To evaluate the accuracy of each state's binary logit models, several statistics were generated (see Appendix A): 1. The results from each binary logit model (coefficients, t-statistics). The `jtools' package was used to produce the model summary statistics (Long 2018). 2. The model fit, including the pseudo R-squared (R2) value. The pseudo R2 value can be interpreted as explaining the amount of variation in the data explained by the value (UCLA 2011). 3. The accuracy of the model, which was produced using the `caret' package in R. (Kuhn et al. 2018). After we had run the 2010 binary logit models and confirmed that the models were accurately predicting urbanization for 2010, we input the 2020 datasets by state into the binary logit models to produce a probability variable for each block. The probability value assigned to the block indicates how likely the block is to be urban in 2020. These values were used to create urbanization scenarios, described in step 5. 27 3.4.2 Step 4B: Estimate More Refined Binary Logit Models for Georgia Given the particular interest in identifying trending-urban implications for Georgia, we estimated other models for this state to incorporate additional information, particularly whether the block was part of the Atlanta MSA or the Savannah MSA in 2010. This updated utility function is shown below. UC or UA in 2000 Closest urban is an urbanized area in 2000 Rural and 0,1 miles from UA in 2000 Rural and 1,2 miles from UA in 2000 Rural and 2,3 miles from UA in 2000 Log of distance to road in 2010 Log of number of jobs in tract in 2010 population density 500,1000 in 2010 population density 1000,1500 in 2010 population density 1500,2000 in 2010 population density 2000,4000 in 2010 population density of 4000 or more in 2010 part of Atlanta MSA in 2010 part of Savannah MSA in 2010 0 (4) As a robustness check for Georgia, we estimated the model described above based on all blocks (defined as the global model) and rural-only blocks (defined as the rural-only 28 model). The first global model included all blocks in our study area (except for water blocks and blocks associated with large protected areas in rural communities). A total of 366,846 blocks were included in the global model, of which we excluded 704 due to missing values within our data. Because urban blocks were arguably more likely to remain urban between the 2000 and 2010 c, we estimated a second model that included just the 171,635 blocks that were rural in 2000 and located within 10 miles of an urban border. The second model provided a better assessment of the prediction accuracy, since the larger model gives credit for accurately predicting urban blocks that stay urban. 3.5 Step 5: Predict Urbanization Scenarios We used the results of the binary logit analysis to predict land use changes in 2020. There were slight differences in how we did this for the statewide models and Georgia-specific models. Fundamentally, we were able to do visual checks on the Georgia model, whereas we needed to use a more automated process for the state models. These processes are described below. 3.5.1 Step 5A.1: Predict Urbanization Scenarios for Each State Four urbanization scenarios for each state were created using the probability variables from the binary logit output and distances to surrounding UAs and UCs generated by the Near Analysis completed in ArcMap. The criteria for each scenario are described below: Merge IF: 1. The probability of the block being urban in 2020 is 50% or greater AND 29 A. Is classified as UC in 2010 AND within mile of a 2010 UC or UA (Scenario 1A); OR B. Is classified as UC in 2010 AND within 0 miles (contiguous) of a 2010 UC or UA (Scenario 1B); OR 2. The probability of the block being urban in 2020 is 75% or AND A. Is classified as a UC in 2010 AND is within mile of a 2010 UC or UA (Scenario 2A); OR B. Is classified as a UC in 2010 AND is within 0 miles (contiguous) of a 2010 UC or UA (Scenario 2B). To identify the nearest UA or UC, the 2020 binary logit model outputs were aggregated by their respective U.S. Census Bureau division (west, south, midwest, and northeast) and brought into ArcMap to conduct another Near Analysis, this time to obtain the distances between the UCs and UAs (these are the distances used to create each of the scenarios described above). The first step in the Near Analysis was to dissolve boundaries by the nearest UC/UA to assign blocks predicted to be urban to their closest UA or UC. Then, isolated slivers of a UC/UA were removed to prevent false merging. This could occur if an isolated portion of a UC/UA were contiguous to another UC/UA but the remainder of the UC/UA to which the isolate belongs may not be contiguous. These isolates were removed by selecting for shape areas greater than square mile. The maximum number of closest matches was set to 3, which yielded the three closest UCs/UAs to the input UC/UA, ranking each by its proximity. The output from this analysis was then brought back into R to generate the scenarios. 30 3.5.2 Step 5A.2: Identify Whether a UC Will Grow into a UA or Merge with Another UA in Each State In 2010, a total of 3,573 UCs and UAs existed in the U.S. (excluding Puerto Rico and the Island Areas). Each of the scenarios yielded a fewer number of UCs and UAs, meaning UCs/UAs had been absorbed by other UCs/UAs. Using the distances generated from the Near Analysis, each of the four merger scenarios (1A2B) was created. For example, under Scenario 1A (50% and within mile), the Arlington, TN UC is predicted to merge with the Memphis, TNMSAR UA. There were some instances in which a UC/UA was contiguous (within 0 miles) to more than one UC/UA. In this situation, the input UC/UA was assigned to merge with the contiguous UC/UA that had the highest population. 3.5.3 Step 5B: Predicting Urbanization Scenarios and UC Growth and Mergers for Georgia The process used for the national-level state model was repeated using the more refined Georgia global model. However, in addition to the automatic process described above, we visually inspected results to ensure that the UC mergers made intuitive sense. There were a few places near Atlanta where we "overrode" the automated process to assign blocks that were part of a UC to a more logical UA. For example, Carrolton, GA, was not assigned to the Atlanta urban area but as its own small urban area, as it was connected to Atlanta via a handful of blocks and the main growth in Carrolton was outside this "sliver" that connected the two. In other instances, we found mergers to be improbable based on an inspection of current land use patterns, and recoded those areas as non-urbanized (e.g., Jasper, GA). Finally, in some instances, we found mergers to be probable, such as in Jackson, GA, and coded it as part of the Atlanta area in the 50% model. 31 3.6 Step 6: Calculate New Urbanized and Non-Urbanized Population and Land Area Given an urbanization scenario, it is straightforward to calculate the new urbanized and non-urbanized population and land areas. At the state level, the urbanized and nonurbanized population and land areas are used as inputs for the FTA 5311 and 5307 funding formulas. However, to calculate funding gaps for individual transit operators in Georgia, these calculations need to be performed at the county level. This is the distinction in Steps 6A and 6B shown in Figure 4. The population and land area for each UC/UA was summed and assigned as non-urbanized if the population was less than 50,000; small urban if the population was between 50,000 and 200,000; and large urban if the population was greater than 200,000. Those areas assigned as small urban or large urban were then classified as UAs in each scenario. These scenario population and land area sums and classifications were joined back to the original 2020 block file containing population and land area. Then, the projected population and land area were summed at both the state and county levels for each scenario. This yielded a new urbanized population and land area for each of the scenarios. The percentages of urbanized and non-urbanized population and land area under each scenario were then compared to the 2010 percentages of urbanized and non-urbanized population and land area at both the county and state levels. The tables containing the percent changes for population and land area are included in Appendix C (see Table C1). 32 3.7 Step 7: Obtain Other Variables Used for FTA 5311 and 5307 Formulas To predict the FTA 5311 and 5307 appropriations after the 2020 decennial census, we needed to apply the funding formulas for each respective program, shown in Figure 5 and Figure 6. For the FTA 5311 program, we produced forecasts for the shaded boxes at the bottom right of Figure 5 using outputs from our state-specific binary logit models, the NTD, and other tables. We used a similar process in Figure 6 to predict the FTA 5307 appropriations. The inputs we used for each shaded box are shown in Table 3, and the specific tables we used are shown in Table 4 (along with references that contain their online links). We used the most recent data available that corresponded to the FY19 appropriations; thus, FY17 NTD data were used for the FY19 appropriations. 33 Source: FTA 2019h. FIGURE 5 5311 Formula Grant 34 Source: FTA 2019i. FIGURE 6 5307 Formula Grant 35 TABLE 3 Variables Used to Predict 5311 and 5307 Appropriations 36 Urbanized Status Funding Sub-Category 5307 LU Bus Tier Urbanized Areas Over 200,000 (Note: LU over 1,000,000 are calculated with the same data but different unit values) 5307 LU Bus Incentive 5307 LU Fixed Guideway Tier 5307 LU Fixed Guideway Incentive 5307 LU Low-Income Urbanized Areas Over 50,000 and under 200,000 5307 SU 5307 SU Low-Income 5307 STIC Based on Land Area and Population Non-Urbanized Based on Land Area, Areas Vehicle Revenue Miles, and Low- income Individuals LU= large urban and SU=small urban Variable Population Population Density Bus Revenue Vehicle Miles (Bus Passenger Miles)2 / Operating Costs Fixed Guideway Revenue Vehicle Miles Fixed Guideway Route Miles Commuter Rail Floor (binary) (Fixed Guideway Passenger Miles Fixed Guideway Passenger Miles) / Operating Costs Commuter Rail Incentive Floor (binary) Low-income Population Population Density Low-income STIC Factors / Qualifying Performance Category Population Land Area Land Area Vehicle Revenue Miles Low-income Data Source State-specific binary logit models FTA Table FTA Table FTA Table FTA Table FTA Table State-specific binary logit models FTA Table FTA Table State-specific binary logit models State-specific binary logit models FTA Table FTA Table TABLE 4 FTA Tables Used to Predict 5311 and 5307 Appropriations 37 Reference FTA Source Tables Application for National and Georgia Models* 2019a Census Data on Rural Population and Land Area (used for the Section 5311 Rural Area Formula apportionments) Substituted by binary logit model predictions for 50% -mile scenario and 75%, 0-mile scenarios. 2019b Census Low Income Population Data (used for the Section 5307 and 5311 apportionments) Kept static for calculations. 2019c Census Urbanized Area Population and Population Density Data (used for Section 5303, 5305, and 5307 apportionments) Substituted by binary logit model predictions for 50% -mile scenario and 75%, 0-mile scenarios. 2019d National Transit Database Data Used for the Section *Modified by adding future rural miles that will be in 5307 Urbanized Area Formula and Section 5339 Bus urbanized area using the percentage of the county that shifted Formula Apportionments to urbanized in terms of population. 2019e National Transit Database Data Used for the Section 5311 Apportionments *Modified by subtracting future rural miles that will be urbanized using the percentage of the county that shifted to urbanized in terms of population. 2019f National Transit Database Data Used for the STIC Apportionments Used to attribute STIC funding based on 2019 values. 2019g National Transit Database and Census Data Used for the Tribal Transit Apportionments Kept static for calculations. 2019j Table 3A. Section 5307 Operating Assistance Special Rule Operator Caps *Amount apportioned to Large Urbanized Areas used for "operating expense gap calculations," along with vehicle revenue hours attributed to counties in the UA. 2019k Table 5: FY 2019 Formula Apportionments Data Unit Used to translate each input (e.g., population) into a dollar Values (Full Year) amount. * Modifications to input NTD data for Georgia are shown with a *. Note Table 3A is only used for the Georgia-specific gap calculations. 3.8 Step 8: Predict Funding 3.8.1 Step 8A: Predict, by State, 5311 and 5307 Appropriations after 2020 Census Given the inputs for the formula funding, we can predict, by state, the 5311 and 5307 appropriations after the 2020 census. To describe this process, we present two examples: one for the 5311 appropriation and the other for the 5307 appropriation. Example 1: 5311 Appropriation for Georgia after the 2020 Census As shown in Figure 5, there are four inputs that we used to predict the 5311 appropriation (shown in the shaded parts of the figure). These are summarized in Table 5 and include the non-urbanized land area, non-urbanized population, non-urbanized VRM and nonurbanized low-income population. We obtained estimates of the 2020 non-urbanized land area and population using the state-specific binary logit model that we estimated; these include an "aggressive" forecast (based on the 50% probability, mile model) and a "conservative" forecast (based on the 75% probability, 0 mile model). We assumed that the non-urbanized VRMs will remain the same after 2020; this may not be the case for those rural areas that transition to a small urban area or get absorbed into a large urban area, but for the purposes of determining a range of potential 5311 funding at the state level after 2020, this effect will be small. We also excluded the VRM from tribes from the analysis. Finally, we used "FTA Table 5" to convert each of these inputs into a dollar amount (these are shown in Table 6). 38 TABLE 5 Example 5311 Appropriation Calculation for the State of Georgia Variable Source 2020 Nonurbanized land area 2020 Nonurbanized population 2020 Nonurbanized low income FY18 Nonurbanized VRM GA binary logit model GA binary logit model GA binary logit model National Transit Database Data Used for the Section 5311 Apportionments Aggressive Estimate (50%, mile model) 42,292 2,444,276 946,226 16,340,485 Conservative Estimate (75%, 0 mile model) 42,765 2,501,056 985,205 16,340,485 TABLE 6 FTA Values Used for 5311 Appropriation (FY19) Appropriation Formula Piece Based on Land Area and Population Population Land Area Based on Land Area, Vehicle Revenue Miles, and LowIncome Population Land Area Vehicle Revenue Mile Low-Income Data Value 4.720 30.53 9.18 0.054 1.915 Using the inputs shown in Table 5 and Table 6, we calculated the 2020 FTA 5311 appropriation for Georgia as follows (assume 50%, mile forecast is used): 2020 5311 Appropriation = 4.72 (2020 non-urbanized population) + 30.53 (2020 non-urbanized land area) + 9.18 (2020 non-urbanized land area) + 0.054 (2020 non-urbanized VRM) + 1.915 (2020 low-income) 39 2020 5311 Appropriation = 4.72 (2,444,276) + 30.53 (42,292) + 9.18 (42,292) + 0.054 (16,340,485) + 1.915 (946,226) = 15,910,807 Example 2: 5307 Appropriation for Georgia after the 2020 Census As shown in Figure 6, there are multiple inputs that we used to predict the 5307 appropriation (shown in the shaded parts of the figure). These are summarized in Table 7 and include the urban populations and urban population densities associated with: (1) small urban areas with populations of 50K200K; and (2) large urban areas of populations greater than 200K. We obtained estimates of the 2020 population and population densities for small and large urban areas using the state-specific binary logit model that we estimated; these include an aggressive forecast (based on the 50% probability, mile model) and a conservative forecast (based on the 75% probability, 0 mile model). We obtained estimates of the service characteristics reported to the NTD. These service characteristics include: fixed guideway passenger miles travelled, fixed guideway vehicle revenue miles, fixed guideway directional route miles, bus revenue vehicle miles, and operating costs. We assumed the service characteristics as of FY17 (the most recent data available/used for FY19 appropriations) will be the same after 2020. As noted earlier, this may not be the case for those rural areas that transition to a small urban area or are absorbed into a large urban area, but for the purposes of determining a range of potential 5311 funding at the state level after 2020, this effect will be small. We excluded tribal service characteristics from the analysis. We assumed that the amount of STIC funding the state received in the most previous year will continue in the future. Finally, we used FTA Table 5 to convert each of these inputs into a dollar amount (these are shown in Table 8). 40 TABLE 7 Inputs for the 5307 Appropriation Calculation for the State of Georgia Variable 2020 Population for Areas 1M+ 2020 Population for Areas <1M 2020 Population for Areas 50K200K 2020 Population Density for Areas 1M+ 2020 Population Density for Areas <1M 2020 Population Density for Areas 50K200K 2020 Low Income Population for Areas 50K200K 2020 Low Income Population for Areas 200K+ FY17 Bus VRM for Areas 1M+ FY17 Bus VRM for Areas <1M FY17 Bus Pax Miles for Areas 1M+ FY17 Bus Pax Miles for Areas <1M FY17 Fixed Guideway Pax Miles for Areas 1M+ FY17 Fixed Guideway Pax Miles for Areas <1M FY17 Fixed Guideway VRM for Areas 1M+ FY17 Fixed Guideway VRM for Areas <1M FY17 Fixed Guideway Directional Route Miles for Areas 1M+ FY17 Fixed Guideway Directional Route Miles for Areas <1M FG Operating Costs for Areas 1M+ FG Operating Costs for Areas <1M Bus Operating Costs for Areas 1M+ Bus Operating Costs for Areas <1M State-wide STIC points Source GA binary logit model GA binary logit model GA binary logit model GA binary logit model GA binary logit model GA binary logit model (2019b) (2019b) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019d) (2019f) Aggressive Estimate (50%, mile model) 5,347,854 963,793 1,202,540 1,485 1,359 925 367,383 1,403,328 47,335,932 7,175,086 357,733,553 17,118,846 469,323,071 256,504 22,405,959 15,550 100.4 1.4 196,339,074 875,235 310,820,674 28,148,985 15 Conservative Estimate (75%, 0 mile model) 5,194,605 923,404 1,101,036 1,799 1,652 1,249 338,647 1,363,766 46,833,207 6,963,124 357,230,828 16,906,884 469,323,071 256,504 22,405,959 15,550 100.4 1.4 196,339,074 875,235 309,716,092 27,681,785 15 41 TABLE 8 FTA Values Used for 5307 Appropriation (FY19) Appropriation Formula Piece Bus Tier for Urbanized Areas Above 1M Population Population Density Bus Revenue Vehicle Miles Bus Tier for Urbanized Areas Under 1M Population Population Density Bus Revenue Vehicle Miles Bus Incentive (PM Denotes Passenger Miles) (Bus PM)2 / Operating Cost Fixed Guideway Tier Fixed Guideway Revenue Vehicle Miles Fixed Guideway Route Miles Commuter Rail Floor Fixed Guideway Incentive (Fixed Guideway OM)2/Operating Cost Commuter Rail Incentive Floor Low Income Individuals for Areas Under 200K Low-income Low Income Individuals for Areas Over 200K Low-income Urbanized Area Formula Program for Areas Under 200K Population Population Density Small Transit Incentive Cities For Each Qualifying Performance Category Data Value 3.346 0.0008918 0.4301 2.884 0.001311 0.5354 0.01402 0.6248 38,861 9,748,729 0.0008806 447,620 2.353 4.231 6.775 0.003442 261,911 Using the inputs shown in Table 7 and Table 8, we calculated the 2020 FTA 5307 appropriation for Georgia as follows (assume 50%, mile forecast is used). This is a simplification of our process, as in our actual calculations we used UA-specific values for population densities and accounted for UAs that crossed state boundaries. As such, the value below is not the actual 5307 that Georgia receives but is meant to demonstrate the application of the funding formula. 42 2020 5307 Appropriation = 3.346 (2020 Population in LU 1M+) + 0.0008918 (2020 Population in LU 1M+ 2020 Population Density in LU 1M+) + 0.4301 (Adjusted FY17 Bus VRM in LU 1M+) + 2.884 (2020 Population in Areas <1M) + 0.001311 (2020 Population < 1M 2020 Population Density <1M) + 0.5354 (Adjusted FY17 Bus VRM in areas <1M) + 0.01402 (FY17 Bus Passenger Miles)2 / (FY17 Bus Operating Costs) + 0.6248 (FY17 Fixed Guideway VRM) + min { 9,748,729 or 38,861 (FY17 Fixed Guideway Route Miles)} + min { 447,620 or 0.0008806 (FY17 Fixed Guideway Passenger Miles)2} / (FY17 Fixed Guideway Operating Costs) + 2.353 (FY17 Low-income for Areas 50K200K) + 4.231 (FY17 Low-income for Areas 200K+) + 6.775 (2020 Population in Areas 50K200K) + 0.003442 (2020 Population in Areas 50K200K) (2020 Population Density in Areas 50K200K) + 261,911 (FY17 STIC Qualifying Criteria Met) 2020 5307 Appropriation = 3.346 (5,347,854) + 0.0008918 (5,347,854 1,485) + 0.4301 (47,335,932) + 2.884 (963,793) + 0.001311 (963,793 1,359) + 0.5354 (7,175,086) + 0.01402 (357,733,553+17,118,846)2 / (310,820,674+28,148,985) + 0.6248 (22,405,959 +15,550) + min { 9,748,729 or 38,861 (100.4+1.4)} + min {447,620 or 0.0008806 (469,323,071+256,504)2} / (196,339,074+875,235) 43 + 2.353 (367,383) + 4.231 (1,403,328) + 6.775 (1,202,540) + 0.003442 (1,202,540) 925 + 261,911 (15) = $100,604,554 Note: this example calculation differs from the actual calculations performed, which based the inputs on UA data. This calculation also does not include the FTA 5340 growing states portion, which for urban areas is approximately 4% (thus our total "5307" estimate for Georgia would be approximately $104.6M). 3.8.2 Step 8B: Predict, by County in Georgia, 5311 and 5307 Funding Levels After the 2020 Decennial Census and Funding Gaps for High Risk Transitions We replicated the analysis above for the state of Georgia and calculated the projected 5311 and 5307 levels for each county. To compare against the existing funding levels for each county, we had to include the FTA 5340 growing states component in the Georgia analysis. Thus, the analysis is identical to that in Step 8A but includes an additional piece to represent the FTA 5340 funding for growing states. 3.8.3 Summary For the national model predictions described in Step 8A, we only included population and urbanized status changes when calculating the predicted funding levels for 5311 and 5307 funding because of the difficulty of linking providers to geographical areas at the national level. That is, there is no national database linking counties to rural providers. However, we have this information for the state of Georgia and, as such, could calculate 44 county-level estimates of future FTA 5311 and 5307 funding. In addition, we could conduct a more detailed analysis of funding gaps that will likely occur after the 2020 decennial census if current transit systems operating in rural areas are absorbed into large urban areas. To do the latter, we adjusted the most recent NTD service data and "shifted" the service amounts currently in rural areas that we expect to be part of urban areas after 2020. These "Georgia-specific" methodology applications are shown in Table 3 (see p. 37). The calculations used to predict funding levels for 5311 and 5307 after the 2020 decennial census were identical to those used for the national level, with one key difference: we needed to use county-level data (versus state-level data) as inputs. To do this, we needed to: (1) associate providers with individual counties; and (2) make assumptions as to when a county will offer 5311 and/or 5307 service. With respect to the first assumption, in Georgia, the majority of service providers serve residents in a single county. For these providers, there is a one-to-one correspondence between the service provider and county. However, there are also several multi-county providers in Georgia. For these providers, we allocated the total service characteristics using one of two methods. For some providers, we knew the actual vehicles that were being used/had been assigned to individual counties, and could directly calculate the county-level service characteristics using these individual vehicles. For other providers, we only knew the total service characteristics across all of the counties in the service area. For these providers, we allocated the total service characteristics to individual counties proportional to the rural population in each of these counties. With respect to the second assumption, we assumed that it is feasible for a county to start providing 5307 service if at least 20% of its population is within an urbanized area. 45 The final step in the analysis was to calculate "funding gaps" associated with those counties trending urban in Georgia. A major challenge for areas transitioning from 5311 service to 5307 service is that FTA limits the amount of 5307 funding that can be used to support operating expenses. For most transit systems, no 5307 funding can be used toward operating expenses. However, there is an exception for small systems with fewer than 100 buses operating in urban areas, which is colloquially referred to as the "100 bus rule." This rule applies a cap on the amount of 5307 funding that can be used toward operating. For those systems that operate fewer than 75 buses, up to 70% of an operator's percentage of the UA's vehicle revenue hours can be applied toward operating expenses. For those systems that operate between 76 and 100 buses, up to 50% of the operator's percentage of the UA's vehicle revenue hours can be applied toward operating. We define the group of operators that would be most affected by urbanization and the 100 bus rule as "high-risk counties" that currently have exclusively rural service but are predicted to have at least 20% of their populations classified as large urban after the 2020 decennial census. These counties would face a two-year lag on 5307 funding related to NTD apportionment data, because the NTD data used are from two years prior to the fiscal year of the apportionments and the 100 bus rule is based on 5307 operating characteristics. To calculate the funding gap for these high-risk counties, we take the differences between the 5311 funding appropriations for FY19 and FY20. 3.9 Summary This chapter provided a detailed description of the data and methodology used in the analysis. The next chapter presents the key results from the analysis. 46 4 RESULTS This chapter discusses the results, which are organized by whether they are used to determine Georgia-specific or national-level trends and funding implications. The Georgia results that do not rely on national calculations are discussed in Section 4.1 and include the binary logit results, identification of areas in Georgia that are trending urban and at risk for losing 5311 funding, and implications on transit funding for individual transit operators, respectively. National-level results are discussed in Section 4.2 and include the statespecific binary logit results, identification of areas in the nation that are trending urban and are at risk for losing 5311 funding, predictions of state-level 5311 appropriations, and predictions of 5307 appropriations after the 2020 decennial census, respectively. Finally, in Section 4.3, the Georgia results that are derived from the national-level predictions of the 5311 and 5307 analysis are discussed. This includes a summary of Georgia's outlook on future 5311 and 5307 funding and insights gleaned from examining 5311 and 5307 appropriations at the county level in Georgia. 4.1 Results for Georgia 4.1.1 Georgia Binary Logit Model Results The results of the Georgia binary logit model are shown in Table 9. These include the global model (that contains all blocks) and the rural model (that contains only those blocks that were rural in 2000). The number of U.S. Census Bureaudefined urban blocks increased by 14% between 2000 and 2010 in Georgia, and our final model accurately predicts the results of rural-to-urban changes in Georgia between 2000 and 2010 with the 47 accuracy at 94% for the global model and 92% for the rural model. Our model shows that the strongest predictor variables for rural-to-urban conversion were in proximity to an existing urban border and population density. Specifically, "border blocks," which we define as those within a quarter mile of an existing urban border, increase the odds of ruralto-urban conversion by 25-fold, and even blocks up to 3 miles away from existing urban borders have significantly higher odds of conversion (see Table 9, column Exp()). The coefficients for our population density variables follow the U.S. Census Bureau guidelines, with blocks including more than 500 people per square mile having positive probabilities of conversion, and this trend extends with larger positive probabilities as densities increase. Furthermore, blocks within the fastest growing MSAs in the state (i.e., Atlanta and Savannah) also affect the odds of rural-to-urban conversion, as does a block being in a census tract with larger numbers of jobs. All other factors being equal, a location closer to primary and secondary roads also significantly increases the odds of urbanization, albeit with a relatively small effect on the odds of conversion. In sum, if a block is near existing roads, near existing urban area boundaries, has a population density over 500 people per square mile, and is in an area with more jobs, it has a high probability of urbanization. 48 TABLE 9 Binary Logit Model Results for Georgia Global Modela S.E. Exp() Urban (UC or UA in 2000) Closest urban is an urbanized area 6.11 0.03 448.20 1.12 0.02 3.05 In 2010 Atlanta MSA 0.95 0.02 2.59 In 2010 Savannah MSA 0.92 0.05 2.51 Urban area (reference) -- -- -- Rural and (0,1] miles from UA Rural and (1,2] miles from UA Rural and (2,3] miles from UA 3.23 0.02 25.20 1.39 0.03 4.00 0.78 0.04 2.18 Log of distance to roads -1.11 0.02 0.33 Log of number of jobs in tract 0.45 2010 population density (0,500] PSQM -- (reference) 2010 population density (500,1000] PSQM 1.98 2010 population density (1000,1500] PSQM 2.14 2010 population density (1500,2000] PSQM 2.27 2010 population density (2000,4000] PSQM 2.69 2010 population density of 4000 or more 2.90 PSQM 0.01 1.57 -- -- 0.03 7.21 0.04 8.49 0.04 9.65 0.03 14.74 0.04 18.14 Constant -7.12 0.06 0 a Accuracy 94%; R2 = 0.87; b Accuracy 92%; R2 = 0.54 All variables significant at a 0.01 significance level. Prob 1.00 0.75 0.72 0.72 -- 0.96 0.80 0.69 0.25 0.61 -- 0.88 0.89 0.91 0.94 0.95 0 Rural Modelb S.E. Exp() 1.10 0.02 3.00 0.94 0.02 2.57 0.97 0.06 2.64 -- ---- 3.17 0.02 23.89 1.36 0.03 3.89 0.77 0.04 2.16 -1.13 0.03 0.32 0.52 0.01 1.69 -- ---- 1.99 0.03 7.33 2.13 0.04 8.45 2.16 0.05 8.67 2.59 0.04 13.38 2.78 0.05 16.04 -7.78 0.07 0 Prob 0.75 0.72 0.72 -- 0.96 0.80 0.68 0.24 0.63 -- 0.88 0.89 0.90 0.93 0.94 0 4.1.2 Areas in Georgia that are Trending Urban Our model identified several low-risk transitions that occur when a rural system grows into a small urban system. As shown in Figure 7, within the greater Atlanta area, Winder is 49 expected to grow internally from a UC to a UA with at least a 0.75 probability, as is Carrollton with at least a 0.50 probability; these two areas may also merge with adjacent large UAs. *Note: Green outlines are urban clusters in 2010 (U.S. Census Bureau 2017). Map prepared with ESRI ArcMap 10.5. FIGURE 7 Probabilities of Urbanized Areas in Metro Atlanta 2020 Important changes will also occur through UCs being absorbed into larger UAs, with the assumption that if boundaries are touching, the urban areas will merge. Figure 8 illustrates the concept of urban areas merging and identifies the UAs and UCs that could potentially be absorbed into the Atlanta UA. If the UCs in Figure 8 (Winder, Monroe, Bremen, and 50 Jefferson) are absorbed into the Atlanta UA, they would transition directly from rural eligibility to large UA eligibility in terms of funding category. In fact, our model shows that in Georgia this merger-driven change (representing medium- and high-risk funding scenarios) is more common than population-driven shifts from rural to urban (representing low-risk funding scenarios). This is principally an issue in the Metro Atlanta region. Source: U.S. Census Bureau 2017. Map prepared with ESRI ArcMap 10.5. FIGURE 8 Urban Clusters and Urbanized Areas Expected to Merge with Atlanta After the 2020 Census Although most urbanization in Georgia is occurring around Atlanta, the cities of Macon and Savannah are also projected to expand outward and could potentially merge with 51 surrounding UCs and UAs (see Table 10). If Savannah merges with Rincon and Buckhead (the neighboring UCs) as our 0.50 model predicts, the total urbanized population would be such that Rincon and Buckhead would no longer be eligible for 5311 operating assistance. TABLE 10 Predictions of Areas in Georgia That Will Need to Transition to New FTA Funding Potential Mergers 2010 Urbanized Areas or Clusters at Risk for Merger Alternate Scenarios MaconWarner Robins (LU) Macon (SU) (Bibb County)** Warner Robins (SU) (Houston County, Peach County)** UA grandfathering criteria prevents merger Savannah (LU) Rincon (UC) (Effingham County)** Buckhead (UC) (Bryan County)* N/A Gainesville (SU) (Hall County)** UA grandfathering criteria prevents merger Atlanta Urbanized Area (LU) Winder (UC) (Barrow County)** Bremen (UC) (Haralson County, Carroll County)* Jefferson (UC) (Jackson County)* Merges to (LU) with AthensClarke Winder, GA, or Atlanta UA N/A N/A Monroe (UC) (Walton County)* N/A Albany (SU) Leesburg (UC) (Lee County)* N/A *Key: UC=urban cluster; SU=small urban area; LU=large urban area. Areas defined as urban clusters by U.S. Census Bureau are classified as rural by FTA and are eligible for 5311 funding. Model predictions in the 5074 probability range are denoted by * and those in the 7589 probability range are denoted by **. LU presumed not to merge unless grandfathering rule is changed. Further, as shown in Figure 9, growth in Atlanta is such that our model predicts that several existing UAs will also merge. For the 2010 decennial census, a set of "grandfathering" rules were used to keep distinct UAs (often located in distinct MPOs) from merging 52 (FTA 2015). We discussed this issue with a representative from the U.S. Census Bureau who noted that although "we have not begun to work on criteria for urbanized areas and urban cluster for the 2020 census,...at this time, we are not planning substantive changes to the criteria" (Ratcliffe, et al., 2016). Based on this expectation, we anticipate that Cartersville and Gainesville (which are not part of the Atlanta MPO) will remain separate small urban transit systems and not be absorbed into the large Atlanta UA. Ultimately, the grandfathering rules will help maintain operating transit funding in megaregions with growth expanding across multiple urban areas if one or more of these UAs can maintain a small urban designation. Source: U.S. Census Bureau 2017. Map prepared with ESRI ArcMap 10.5. FIGURE 9 Contiguous or Near-Contiguous MPOs and Urbanized Areas in Metro Atlanta After 2020 Census 4.1.3 Anticipated Funding Gaps in Georgia The anticipated funding gaps for the high-risk counties that are trending urban and currently do not offer 5307 operations is summarized in Table 11. These counties correspond to those shown as UCs on Table 10 that will potentially merge with an LU. 53 These counties will face a shortfall of $225K$1.11M in years one and two after the decennial census. As noted earlier, these operators will be prevented from using FTA 5307 funds for the first two years after the 2020 decennial census while they wait for their new 5307 operations to become certified. Among the counties shown in Table 11, Barrow is the only one that appears in both the conservative and aggressive forecast. All of the other counties shown in Table 11 appear only on the aggressive (and therefore less likely) forecast. TABLE 11 Operating Funding Gaps for Georgia Counties Trending Urban County Barrow Bryan Carroll Effingham Haralson Jackson Walton TOTAL FY19 5311 Appropriation 356,850 153,844 586,929 342,856 195,205 341,984 350,405 2,328,073 5311 Forecast 75% 0 mi 102,638 102,638 50% mi 62,673 134,336 241,717 241,808 139,926 221,705 187,394 1,229,559 Gap 75% 0 mi -254,212 -254,212 50% mi -294,177 -19,508 -345,212 -101,048 -55,279 -120,279 -163,011 -1,098,514 4.2 Results for States Nationwide 4.2.1 Nationwide Binary Logit Model Results Appendix A contains the results of the 50 state binary logit models. As part of the modeling process, we estimated a state-specific binary logit model that included all of the variables shown in Equation 3. However, for some states, this specification produced counterintuitive results. For example, the coefficient associated with jobs was negative (implying the more jobs, the more likely the block was to be rural) or the coefficient associated with 54 distance to roads was positive (meaning the farther you are from a road, the more likely you are to be urban). Given these are clearly counter-intuitive results, we excluded these variables from state-specific models that did not have a positive coefficient for jobs or a negative coefficient for distance to roads. In a similar way, we combined categories associated with the variables that measured the distance from a rural area for rural blocks to ensure that these coefficients were monotonically decreasing, meaning that as you moved farther from the rural area, you were less likely to be urban. Similarly, as the population density increases, we expect that a block would be more likely to be urban (thus, the relationship among population density coefficients should be monotonically increasing as density increases). An example of this process is seen with Rhode Island. Rhode Island combines the rural and (1,2], (2,4] and 4+ categories together (note 4+ is the reference category and assigned a value of zero). Similarly, Rhode Island combines the population density variables for (1000, 2000] and (2000, 4000]. The need to drop variables and/or combine categories was most prevalent in states that were very small (such as Rhode Island, Vermont, and New Hampshire), states that are predominately rural (such as Alaska, Oklahoma, Nevada, and Wyoming), or states that have distinct geographic features (including Hawaii, which is a set of islands). Overall, the accuracy of the nationwide binary logit models is at least 90%, with many states having a higher prediction accuracy. The accuracy of the 50 state binary logit models used to predict urbanization ranged from 90.3% (Delaware) to 98.6% (North Dakota). We concluded that the state-specific binary logit models are performing well, and used these for forecasting land use changes after the 2020 decennial census. 55 The general results from these models indicate that the three strongest predictors of a block being urban in 2020 were: (1) the block's urban or rural classification in the previous Census; (2) the block's population density; and (3) the distance to an existing UC or UA. These variables and their probabilities can be interpreted as follows: 1. Holding all other variables constant, if the block was classified as urban in the previous Census, it was 99% more likely to be urban in 2020. 2. Holding all other variables constant, if the block's population density was between 500 and more than 4,000 persons per square mile, then it was 88.2% to 94.4% more likely to be urban in 2020. 3. Holding all other variables constant, if the block was classified as rural in the previous Census and was less than 1 to 4 miles from an existing UC or UA, then it was 87.8% to 97.9% more likely to be urban in 2020. 4.2.2 Areas in the Nation that are Trending Urban Appendix B, Table B1 summarizes the areas in the nation that are trending urban and, specifically, the urban clusters that are expected to merge with other urban clusters or urban areas after the 2020 decennial census. Not all of these mergers will result in a change in funding eligibility, but this is an important first step for understanding where the growth is occurring and for calculating inputs in land areas and populations that are needed to estimate changes in 5311 and 5307 allocations after the 2020 decennial census. The UCs that are expected to merge with other UCs or UAs are shown for the four different prediction scenarios. The most aggressive scenario is 1A, which corresponds to the 50% probability model using a mile distance threshold, whereas the most conservative 56 scenario is 2B, which corresponds to the 75% probability model using a 0 mile distance threshold. The first remarkable result is that predictions are highly sensitive to the underlying rules that are used to determine when a UC merges. Under the most aggressive scenario, a total of 195 UCs/UAs are expected to merge after the 2020 decennial census, whereas under the most conservative scenario, only 20 UCs/UAs are expected to merge. Scenario 1B, which likely represents a "best guesstimate" of the future (with a 50% probability and 0 mile threshold) suggests 86 UCs are expected to merge, and Scenario 2A (with a 75% probability and mile threshold) suggests 112 UCs are expected to merge. Comparing the states, under the most conservative scenario, Florida (at 6) and Pennsylvania (at 5) have the largest number of UCs/UAs that are expected to be part of a large UA. Under the most aggressive scenario, Texas has the largest number of UCs/UAs that are expected to merge (at 32), followed by Florida (13), North Carolina (12), Arizona (10), and Pennsylvania (10). 4.2.2.1 Highest Risk Transitions for the Most Aggressive Scenario (1A) Scenario 1A allows for the merging of a UC with either another UC or a UA if the blocks within the UC have at least a 50% probability of being classified as urban after the 2020 census and the UC is located within mile of an existing UC/UA. This scenario is the upper boundary (excluding UA merger scenarios) estimate for urbanization (showing the maximum predicted urbanization) under the current U.S. Census Bureau urbanization rules. 57 Selecting for blocks with a 50% probability of urbanizing and assuming a merge if the UC is within mile of the UA, the total number of UCs/UAs was reduced to 3,427 from 3,573. A total of 102 UCs are predicted to undergo a "high-risk" transition from a UC to a large urbanized area. A total of 49 UCs are predicted to merge into another UC, and a total of 44 UCs are predicted to merge into a small UA. A full list of all mergers under this scenario is included in Appendix B, Table B1. The high-risk transitions are denoted by an asterisk (*) on the table. 4.2.2.2 Highest Risk Transitions for the Least Aggressive Scenario (2B) Scenario 2B allows for the merging of a UC with either another UC or a UA if the blocks within the UC have at least a 75% probability of being classified as urban after the 2020 census and the UC shares contiguous borders (a distance of 0.0 miles) with an existing UC/UA. This scenario is the lower boundary estimate for urbanization (showing the minimum predicted urbanization) under the current U.S. Census Bureau urbanization rules. Under this scenario, the total number of UCs/UAs was reduced to 3,560 from 3,573 UCs/UAs in 2010. A total of 6 UCs were predicted to undergo a "high-risk" transition from a UC to a large UA (see highlighted rows in Table 12). A total of 8 UCs were predicted to merge into another UC, and a total of 6 UCs were predicted to merge into a small UA. The list of UCs and their respective mergers are shown below in Table 12. A full list of all mergers under this scenario is also included in Appendix B, Table B1. 58 TABLE 12 Urban Clusters Predicted to Merge Under Scenario 2B (U.S. Census Bureau 2010) State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/UA LSAD Alabama Priceville, AL Decatur, AL UA Connecticut Willimantic, CT Hartford, CT UA Delaware Bridgeville, DE Salisbury, MDDE UA Florida Crystal River, FL Homosassa Springs UA Beverly HillsCitrus Springs, FL Fernandina Beach, FL Yulee, FL UC Four Corners, FL Winter Haven, FL UA Panama City Panama City, FL UA Northeast, FL Poinciana, FL Kissimmee, FL UA Yulee, FL Fernandina Beach, FL UC Georgia Winder, GA Atlanta, GA UA Louisiana Donaldsonville, LA Houma, LA UA New Jersey Newton, NJ New YorkNewark, UA NYNJCT Ohio Ashtabula, OH Conneaut, OH UC Conneaut, OH Ashtabula, OH UC Pennsylvania Jersey Shore, PA Lock Haven, PA UC Lock Haven, PA Jersey Shore, PA UC Lykens, PA Williamstown, PA UC Roaring Spring, PA Altoona, PA UA Williamstown, PA Lykens, PA UC Virginia Purcellville, VA Washington, DCVAMD UA Note: High-risk transitions (from rural to large urban) are shaded. 4.2.2.3 Internal UC/UA Growth An existing UC or UA also can urbanize without merging/gaining urbanized land area. Internal growth occurs through population growth inside the existing UC/UA boundaries. The following tables show sets of internal growth: (1) UCs growing into a small urban area (Table 13 and Table 14), and (2) small urban areas growing into large urban areas (Table 15 and Table 16). Even if a UC is not predicted to merge in one of the scenarios discussed 59 in this thesis, it could still be subject to urbanization through internal growth. UCs that are candidates for both merging and internal growth are of particular concern for this project. Urban Cluster to Small Urban Growth Table 13 and Table 14 include areas that were classified as UCs (under 50,000 people) in 2010, but are predicted to grow internally to have a population of greater than 50,000 in 2020. This growth would cause these areas to not only shift from classification as UCs to small UAs, but also puts these areas at risk for transitioning from FTA 5311 to 5307 funding. The highlighted rows in the tables are UCs that have a projected population of 47,500 or greater (within a 5% margin of the small UA threshold) in 2020. It is important to consider these areas in this scenario to adjust for potential under-prediction by any of the state regression models. For the most aggressive scenario (i.e., 50% probability of having a population of at least 50K), a total of 22 UCs are predicted to grow internally to become small UAs, with an additional 11 UCs within a 5% margin of the threshold. These areas are highlighted in Table 13. Five of the UCs listed in Table 13 are also listed to merge under the 50% model. These UCs include: 1. Bullhead City, AZNV; predicted to merge with the Laughlin, NV UC 2. Poinciana, FL; predicted to merge with the Kissimmee, FL UA; also listed as a rural to large UA transition 3. Winder, GA; predicted to merge with the Atlanta, GA UA; also listed as a rural to large UA transition 4. Carolton, GA; predicted to merge with the Atlanta, GA UA; also listed as a rural to large UA 60 5. Sandusky, OH; predicted to merge with the small UA LorainElyria, OH The UCs at risk for transitioning from UC to small UA under a more conservative model (i.e., 75% probability) are listed in Table 14. A total of 14 UCs are predicted to grow to over 50,000 people, with an additional 15 UCs that have a population within a 5% margin of the UA threshold. 61 TABLE 13 UCs Predicted to Grow into Small UAs Under the 50% Model 2010 2020 State Name 2010 UC Name Population Population Alaska LakesKnik-Fairview Wasilla, AK 44,236 59,230 Maricopa, AZ 42,337 52,364 Arizona SahuaritaGreen Valley, AZ 40,691 50,100 Bullhead City, AZNV 48,656 54,463 California Hollister, CA ReedleyDinuba, CA 42,002 46,247 47,984 53,208 Florida Poinciana, FL 41,922 50,426 Georgia Carrollton, GA Winder, GA 42,872 37,831 49,187 49,220 Idaho Twin Falls, ID 48,836 56,333 Kansas Salina, KS 47,493 48,714 Kentucky Paducah, KYIL 48,791 51,043 Michigan Traverse City, MI 47,109 51,396 Montana Bozeman, MT Helena, MT 43,164 45,055 53,030 51,073 North Carolina Morehead City, NC Wilson, NC 44,798 49,190 50,989 51,605 North Dakota Minot, ND 42,650 59,936 New Mexico Clovis, NM Roswell, NM 41,570 49,727 50,077 50,283 Findlay, OH 48,441 48,649 Marion, OH 46,384 47,978 Ohio New PhiladelphiaDover, OH 46,366 48,732 Sandusky, OH 48,990 48,157 Oklahoma Enid, OK Stillwater, OK 47,609 44,515 50,694 50,585 South Carolina BeaufortPort Royal, SC 48,807 56,087 Tennessee Cookeville, TN 44,207 50,567 Eagle Pass, TX 49,236 54,707 Texas Galveston, TX Lufkin, TX 44,022 44,927 47,782 49,527 Rio Grande CityRoma, TX 46,344 57,116 Virginia Danville, VANC 49,344 49,698 Source: U.S. Census Bureau 2010. Highlighted rows are just under the 50,000 population threshold for being classified as small UAs. 62 TABLE 14 UCs Predicted to Grow into Small UAs Under a 75% Model State Name 2010 UC Name 2010 2020 Population Population Alaska LakesKnik-FairviewWasilla, AK 44,236 54,630 Arizona Maricopa, AZ 42,337 51,773 SahuaritaGreen Valley, AZ 40,691 49,217 Bullhead City, AZNV 48,656 53,336 California ReedleyDinuba, CA 46,247 52,360 Florida Poinciana, FL 41,922 50,419 Georgia Carrollton, GA 42,872 47,695 Idaho Twin Falls, ID 48,836 55,300 Kansas Salina, KS 47,493 48,393 Kentucky Paducah, KYIL 48,791 50,349 Michigan Traverse City, MI 47,109 50,151 Montana Bozeman, MT 43,164 52,277 Helena, MT 45,055 49,311 North Carolina Morehead City, NC 44,798 49,031 Wilson, NC 49,190 50,709 North Dakota Minot, ND 42,650 58,639 New Mexico Clovis, NM 41,570 49,657 Roswell, NM 49,727 49,690 Ohio Findlay, OH 48,441 48,049 New PhiladelphiaDover, OH 46,366 47,974 Sandusky, OH 48,990 47,882 Oklahoma Enid, OK 47,609 49,753 Stillwater, OK 44,515 49,451 South Carolina BeaufortPort Royal, SC 48,807 52,770 Tennessee Cookeville, TN 44,207 48,756 Texas Eagle Pass, TX 49,236 53,822 Lufkin, TX 44,927 48,599 Rio Grande CityRoma, TX 46,344 55,519 Virginia Danville, VANC 49,344 47,998 Source: U.S. Census Bureau 2010. Highlighted rows are just under the 50,000 population threshold for being classified as small UAs. 63 Small Urban to Large Urban Shifts No small UAs were predicted to grow internally to become large UAs. However, there were 7 small UAs in the 50% model and 10 small UAs in the 75% model that were within a population margin of 9,000 people (4%). These UAs are listed in Table 15 and Table 16. TABLE 15 Small UAs Close to Growing into a Large UA Under 50% Model UA Name Erie, PA OlympiaLacey, WA Clarksville, TNKY Waterbury, CT Sioux Falls, SD North PortPort Charlotte, FL Cedar Rapids, IA Source: U.S. Census Bureau 2010. 2010 Population 196,611 176,617 158,655 194,535 156,777 169,541 177,844 2020 Population 198,502 198,491 197,088 194,196 193,979 193,968 192,891 TABLE 16 Small UAs Close to Growing into a Large UA Under 75% Model UA Name College StationBryan, TX Gainesville, FL Erie, PA OlympiaLacey, WA Salinas, CA Deltona, FL Waterbury, CT Waco, TX Clarksville, TNKY Sioux Falls, SD Source: U.S. Census Bureau 2010. 2010 Population 171,345 187,781 196,611 176,617 184,809 182,169 194,535 172,378 158,655 156,777 2020 Population 198,928 198,031 197,945 197,011 196,981 196,173 193,993 193,527 193,514 191,214 64 4.2.3 Forecasts of Nationwide 5311 Funding Levels After the 2020 Census The apportionment quotient for states in 2010 for the 5311 rural funding program is shown in Figure 10. The quotient represents each state's unconstrained share of the appropriated funds through the 5311 formula. This quotient was calculated by dividing each state's national share of non-urbanized land area and population over the total nonurbanized land area and population for the U.S. in 2010. Each state's land area portion was multiplied by 20% and the population portion was multiplied by 80%. These two percentages are used to determine the state's total apportionment. No state is eligible to receive more than a 5% share of their portion of non-urbanized land area (i.e., Alaska and Texas). This was not corrected for in the percentages reported below, and, therefore, they are unconstrained; however, this only affected 1.98% of funding nationally, from one state (i.e., Texas), in 2010. As depicted in Figure 10, those states that are eligible to receive the highest share of 5311 funding include Texas, California, North Carolina, Alaska, and Ohio. It could be expected that vastly rural western states, such as Montana, Wyoming, Nevada, etc., would receive a higher quotient of 5311 funding, but this is not so because those states' shares of nonurbanized population are low relative to other states. Since the highest weighted input into the funding formula is non-urbanized population, those states do not receive a large portion of 5311 funding. The maps provided in the next section show percent change in the 5311 population and land area quotient relative to the numbers presented in this 2010 map. 65 Sources: ESRI 2017; U.S. Census Bureau 2010, 2017. FIGURE 10 FTA 5311 Apportionment Quotient for 2010 by State The two key inputs into the FTA 5311 apportionment quotient is non-urbanized population and non-urbanized land area. To predict the FTA 5311 apportionment for 2020, we needed to estimate these two inputs. The predictions for the FTA 5311 apportionment for 2020 are presented below for two scenarios: Scenario 1A (corresponding to the most aggressive scenario we examined) and Scenario 2B (corresponding to the most conservative scenario we examined). Also, note that although the topic of this report is on the urbanization of rural areas in the U.S., the results are presented in the context on non-urbanized land area and population. 66 This is to show the non-urbanized land area and population deficits throughout the country to understand which states or regions are likely to be subjected to the aforementioned issues with urbanization and FTA 5311 funding because of urbanization. Those states with negative percent differences in apportionment will be presented with a funding gap. 4.2.3.1 5311 Allocation Forecasts for the Most Aggressive Forecast Scenario (1A) Using the ESRI population data, the national total non-urbanized population is predicted to decrease by 1,695,956 people, which would represent a 1.9% overall reduction in nonurbanized population between 2010 and 2020 for Scenario 1A (see Figure 11). For the remaining urbanization/merger scenarios, the national change in non-urbanized population is: Scenario 1B (50% probability and within 0 miles): An increase of 0.65% (or 573,835) Scenario 2A (75% probability and within mile): An increase of 2.73% (or 2,428,140 persons) Scenario 2B (presented in Section 4.2.3.2; 75% probability and within 0 miles): An increase of 3.18% (or 2,832,743 persons) So, the national percent change between the urbanization scenarios (excluding the UA merger scenarios), non-urbanized population is predicted to change between -1.91% and 3.18%. 67 Sources: Mapchart.net 2019; U.S. Census Bureau 2010, 2017. FIGURE 11 Percent Change in Non-Urbanized Population Under Scenario 1A by State Between 2010 and 2020 Urbanization is also modeled at the county level to show which counties within each state may be the drive behind the state's overall change (see in Figure 12). The blue outlines (50 counties in total) indicate that the county grew from less than to greater than 50% urbanized population between 2010 and 2020. A full list of these counties is included in Table 17. Further, under Scenario 1A, a total of 41 counties are predicted to become principally urban, with a total urbanized population surpassing the 89% threshold set forth by the U.S. Census Bureau (Ratcliffe et al. 2016). These counties in particular are of concern as they are predicted to become principally urban after the 2020 census. In Figure 12, these counties are represented by a yellow crosshatch. 68 *Notes: Counties predicted to grow to more than 50% urbanized population in blue; counties predicted to increase more than 10% in urbanized population in yellow crosshatch. Sources: ESRI 2017; U.S. Census Bureau 2010, 2014. FIGURE 12 Percent Change in Urbanized Population Under Scenario 1A by County Between 2010 and 2020 69 70 County Name MatanuskaSusitna Borough Crawford Lonoke Mohave Windham Barrow Bryan Oconee Walton Maui Twin Falls Jackson Boone McCracken St. James Parish Iberville Parish Berkshire Grand Traverse Lamar Gallatin Lewis and Clark Carteret Craven Johnston Haywood TABLE 17 Counties that Grew to over 50% Urbanized Population Scenario 1A State 2010 % Urban Pop. 2020 % Urban Pop. % Change Urban Pop. % Change Urban Land Area County Name 2010 % State Urban Pop. AK 0.0% 55.9% 55.9% 0.4% Wayne NC 49.8% AR 48.0% 52.0% 3.9% 0.6% Wilson NC 0.8% AR 45.2% 51.0% 5.8% 0.3% Ward ND 0.0% AZ 26.7% 52.5% 25.8% 0.4% Hunterdon NJ 45.6% CT 27.8% 53.8% 26.0% 3.8% Chaves NM 0.0% GA 16.7% 82.8% 66.1% 38.7% Curry NM 0.0% GA 30.6% 55.9% 25.4% 3.6% San Juan NM 40.8% GA 49.7% 65.1% 15.5% 12.8% Jefferson NY 49.8% GA 33.4% 63.1% 29.8% 9.4% Oneida NY 49.4% HI 36.1% 53.2% 17.1% 4.0% Ulster NY 48.8% ID 0.0% 65.1% 65.1% 1.0% Erie OH 8.4% IL 46.8% 65.7% 18.9% 3.7% Garfield OK 0.0% IN 38.4% 51.6% 13.2% 4.2% Payne OK 0.0% KY 0.0% 73.8% 73.8% 18.1% Rogers OK 20.6% LA 0.0% 61.6% 61.6% 7.2% Beaufort SC 42.5% LA 34.4% 52.2% 17.8% 1.9% Kershaw SC 20.4% MA 45.1% 59.2% 14.1% 1.1% Lincoln SD 49.5% MI 0.0% 52.5% 52.5% 9.1% Putnam TN 0.0% MS 49.6% 57.3% 7.6% 2.6% Comal TX 49.0% MT 0.0% 50.0% 50.0% 0.8% Johnson TX 29.4% MT 0.0% 73.3% 73.3% 0.8% Maverick TX 0.0% NC 0.0% 68.8% 68.8% 9.3% Starr TX 0.0% NC 48.8% 73.2% 24.4% 4.6% Box Elder UT 49.1% NC 22.2% 55.1% 32.9% 8.8% Albemarle VA 49.4% NC 44.6% 56.0% 11.4% 3.8% Prince George VA 46.6% 2020 % Urban Pop. 59.2% 63.9% 71.1% 50.3% 75.5% 86.2% 55.3% 50.5% 50.1% 51.4% 71.7% 78.3% 59.9% 54.5% 81.8% 53.1% 62.6% 64.5% 57.2% 68.9% 91.7% 80.6% 52.2% 52.2% 56.4% % Change Urban Pop. 9.4% 63.1% 71.1% 4.8% 75.5% 86.2% 14.5% 0.8% 0.7% 2.6% 63.3% 78.3% 59.9% 33.9% 39.3% 32.7% 13.1% 64.5% 8.2% 39.5% 91.7% 80.6% 3.1% 2.8% 9.8% % Change Urban Land Area 6.1% 7.2% 1.4% 1.3% 0.5% 1.6% 0.4% 0.0% 0.0% 0.5% 12.1% 2.4% 3.1% 5.8% 10.5% 4.5% 2.3% 11.6% 9.3% 10.4% 1.6% 2.3% 0.4% 2.1% 5.0% Additionally, non-urbanized land area was reduced by 7.13% nationally. The percent change by state is illustrated in Figure 13, with Florida, North Carolina, and South Carolina having the largest reduction in non-urbanized area. For the remaining urbanization/merger scenarios, the national change in non-urbanized land area is: Scenario 1B (50% probability and within 0 miles): A decrease of 7.05% (or 242,864 square miles) Scenario 2A (75% probability and within mile): A decrease of 6.52% (or 224,638 square miles) Scenario 2B (presented in Section 4.2.3.2; 75% probability and within 0 miles): A decrease of 6.55% (or 225,960 square miles) Between the four scenarios (excluding the UA merger scenarios), the national percent change in non-urbanized land area is predicted to be between -6.52% and -7.13%. Considering these predictions in conjunction with the non-urbanized population changes, an increase in urbanized land area does not always coincide with an increase in urbanized population. 71 Sources: ESRI 2017; U.S. Census Bureau 2010, 2017. FIGURE 13 Percent Change in Non-Urbanized Land Area Under Scenario 1A by State Between 2010 and 2020 As a result of these predicted shifts in non-urbanized population and land area, the overall FTA 5311 land area and population quotients for all but three states (Hawaii, Georgia, and Rhode Island) are expected to change under Scenario 1A (the most aggressive scenario). Figure 14 illustrates the percent shift in each state's quotient based on the 2010 quotient percentages presented in Figure 10. A total of 26 states are predicted to have an increase in their FTA 5311 population and land area quotients (ranging from 0.01 to 1.01%); this could likely lead to an increase in 5311 funding for these states after the 2020 census. Twenty-one (21) states are to have a reduced quotient for the 5311 apportionment formula (ranging from -0.01 to -1.84%), indicating a likely reduction in 72 the apportionment for these states after 2020. See Appendix C, Table C1 for additional details on the changes in urbanized populations, land areas, and FTA 5311 funding. Sources: ESRI 2017; U.S. Census Bureau 2010, 2017. FIGURE 14 Percent Change in Land Area and Population Quotient under Scenario 1A for the FTA 5311 Formula by State between 2010 and 2020 Florida is an example of a state that is expected to have a reduced quotient for the 5311 formula. Florida's binary logit model predicted urbanization correctly 93.5% of the time. Between 2010 and 2020, Florida is predicted to lose a total of 3% of its non-urbanized population and 4.36% of its non-urbanized land area under this scenario. This can also be interpreted that the state is predicted to gain both urbanized population and land area over 73 the 10-year period, or, simply put, the state is predicted to become more urban. In 2010, the state held 2.65% of the country's total non-urbanized population and 1.36% of the country's non-urbanized land area. In 2020, for Scenario 1A, these percentages were calculated to be 2.25% and 1.39%, respectively. Florida's total share of non-urbanized population is predicted to drop, although its share of the nation's non-urbanized land area actually was predicted to increase. So, even though the state's raw quantity of square miles is predicted to decrease (46,790 mi2 to 44,465 mi2), the percentage is predicted to increase because there was an overall loss in non-urbanized land area throughout the country (245,851 mi2 in total). The population component to Florida's 5311 quotient (80% of the state's relative national share of non-urbanized population) was reduced from 2.12% to 1.8% in 2020. The land area component of the 5311 quotient (20% of the state's relative national share of nonurbanized land area) increased by 0.01% for the reasons listed above. This goes to say that a state's portion of non-urbanized land area is a stronger determinant for its 5311 apportionment total. Further, almost all of the states that were predicted to experience a decrease in their overall 5311 quotients held a large share of the nation's non-urbanized population relative to the states that were not predicted to experience a decrease in 5311 quotients. In other words, states that hold a large share of the nation's non-urbanized population and experienced a decrease in both non-urbanized population and land area between 2010 and 2020 were modeled to have a decrease in the 5311 quotient. 74 4.2.3.2 Forecasts of 5311 Funding for the Most Conservative Scenario (2B) Scenario 2B allows for the merging of a UC with either another UC or a UA if the blocks within the UC have at least a 75% probability of being classified as urban after the 2020 census and the UC shares contiguous borders (a distance of 0.0 miles) with an existing UC/UA. This scenario is the lower boundary estimate for urbanization (showing the minimum predicted urbanization) under the current U.S. Census Bureau urbanization rules. Our models predict the national total non-urbanized population to grow by 2,832,743 people in this scenario, which amounts to a 3.18% overall increase in non-urbanized population between 2010 and 2020 for Scenario 2B. The states of Delaware, Maine, Rhode Island, and Hawaii all were predicted to experience increases in non-urbanized population between 2010 and 2020. Alaska, North Dakota, and Idaho were predicted to lose over 5% of their non-urbanized populations (see Figure 15). For Scenario 2B, non-urbanized land area across the U.S. was reduced by 6.55% between 2010 and 2020. This change is illustrated by state in Figure 16. Connecticut (-0.73%), Delaware (-0.59%), and Massachusetts (-0.52%) experienced a decrease in non-urbanized land area of more than -0.5%. At the county level (Figure 17), almost all of the counties in Connecticut, Delaware, and Massachusetts were predicted to experience an increase in urbanized land area between 1% and 10%. 75 Sources: Mapchart.net 2019; U.S. Census Bureau 2010, 2017. FIGURE 15 Percent Change in Non-Urbanized Population Under Scenario 2B by State Between 2010 and 2020 Sources: Mapchart.net 2019; U.S. Census Bureau 2010, 2017. FIGURE 16 Percent Change in Non-Urbanized Land Area Under Scenario 2B by State Between 2010 and 2020 76 As a result of these predicted shifts in non-urbanized population and land area, the overall FTA 5311 land area and population quotients for all but four states (Idaho, Maryland, Vermont, and Rhode Island) are expected to change. Figure 18 illustrates the percent shift in each state's quotient based on the 2010 quotient percentages presented in Figure 10. Similar to Scenario 1A, a total of 28 states are predicted to have an increased 5311 land area and population quotient (ranging from 0.01% to 1.01%), whereas 18 states are predicted to have a reduced quotient (ranging from -0.01 to -1.86%). Generally, the same states are predicted to have a reduced 5311 quotient as in Scenario 1A, although to a lesser degree, with the exception of New Mexico, Texas, Louisiana, Maryland, Delaware, and New Hampshire. Additionally, Mississippi and Iowa are predicted to have a reduced 5311 population and land area quotient in Scenario 2B, whereas in Scenario 1A, both states were predicted to have an increased 5311 quotient. 77 Sources: ESRI 2017; U.S. Census Bureau 2010, 2014. FIGURE 17 Percent Change in Urbanized Land Area Under Scenario 2B by County Between 2010 and 2020 As with Scenario 1A, all of the states predicted to have a reduced 5311 population and land area quotient after 2020 are concentrated in the eastern part of the U.S. (with the exception of Alaska). This could be correlated with the size of the counties in this area of the country, which are much smaller in terms of land area than counties in the western states. Counties with a large land area will inherently have smaller population densities relative to the population densities in counties with a small total land area. The eastern states shaded in dark orange (-0.01% to -0.97%) that have small land areas are more vulnerable to shifts in population. 78 Sources: ESRI 2017; U.S. Census Bureau 2010, 2014. FIGURE 18 Percent Change in Land Area and Population Quotient under Scenario 2B for the FTA 5311 Formula by State between 2010 and 2020 Under Scenario 2B, Michigan is expected to have a 0.14% decrease in its 5311 population and land area quotient. Similar trends to those in Florida under Scenario 1A are predicted to occur in Michigan under this scenario. Michigan's non-urbanized population is predicted to decrease by nearly 77,000 people, which equates to a 0.2% drop in the state's national share of non-urbanized population (dropping to 3.53% from 3.73%). Michigan's national share of non-urbanized land area is predicted to remain fairly constant (only losing 70 square miles of rural land area), but its percent share actually increased from 1.55% to 79 1.66%. This is due to the predicted overall loss of national non-urbanized area, which decreases the denominator. This is the same trend modeled under Scenario 1A, but with a lesser degree of change in quotients. 4.2.4 Forecasts of Nationwide 5311 and 5307 Funding Levels After the 2020 Census The analysis above showed that the decline in rural populations and land areas is occurring at about the same rate across the nation. In the event that transit funding is reauthorized at the current levels, we would not expect to see notable changes in 5311 funding levels. However, it is possible that transit funding for the 5311 and 5307 programs would change to reflect the shift in population trends that is occurring across the nation. We predicted the amount of funding that would be needed after the 2020 decennial census for each program by using the current appropriation formulas and FY19 FTA data values, and updating the inputs to reflect changes in rural and urban populations. We did this for the most conservative and least conservative scenarios. The results of this analysis are summarized in Table 18. Tables C1C5 in Appendix C contain more details and summarize these changes for each state. Note that modeling assumptions, our results differed slightly from the "actual" results. In particular, we predicted FY18 5311 and 5307 appropriations of $629M versus the actual of $659M and $4.6B versus the actual of $5.1B in our calculations, respectively. As seen in Table 18, there are large potential shifts across the programs, with the rural and large urban 1M+ areas losing allocations. Stated another way, there is an increased need for funding to support small urban areas, particularly those with populations between 50K and 200K. The funding needs of the small urban areas with populations between 50K and 80 200K is expected to increase by 37% to 51%, and the funding needs of small urban areas with populations between 200K and 1M is expected to grow 23% to 24%. Viewed in the context of the entire analysis, it is clear that the overall needs in transit funding by system size will be different in 2020 than they were in 2010. TABLE 18 Predicted Changes in 5311 and 5307 Funding After 2020 (Assumes FY19 FTA Data Values) Funding Source Current Predicted (Population) Appropriation* Appropriation Difference % Difference 5311 rural (<50K) 5307 small urban (50K200K) 5307 large urban (200K1M) 5307 large urban (1M+) 629M 402M 839M 3.38B 483 to 505M -124 to -146M 550 to 608M 148 to 206M 1.035 to 1.044B 196 to 205M 3.00 to 3.06B -316 to -358M -20 to -23 37 to 51 23 to 24 -9 to -11 TOTAL 5.25B 5.13 to 5.16B -118 to -71M -1.4 to -2.2 *Note: Actual appropriations differ slightly due to modeling assumptions. FY19 5311 (that includes the 5340 growing states) appropriation was $630M (vs. $628M) and FY19 5307 appropriation was $4.80B (vs. $4.62B in our calculations). See Federal Register (2019) for the appropriation amounts and Appendix C for more details. Note that the numbers reported on the table above do not include the 5340 growing states portion in the totals. 4.3 Additional Analysis Conducted for Georgia 4.3.1 Georgia's Outlook After the 2020 Decennial Census The overall outlook for future 5311 and 5307 appropriations for Georgia is quite positive, in part because the state has many areas that have been growing at a pace that is above the national average. As such, Georgia has one of the highest appropriations from the 5340 growing states program. According to the FTA website, "the Growing States and High Density States Formula Program (49 U.S.C. 5340) was established by SAFETEA-LU [Safe, Accountable, Flexible, Efficient Transportation Equity Act: A 81 Legacy for Users] to apportion additional funds to the Urbanized Area Formula and Rural Area Formula programs" (FTA 2016). If this program continues, it will help mitigate potential losses in funding from the 5311 appropriations for Georgia (and likely other states). Nationally, the predicted appropriation for the 5307 program was forecast to decline -1.4% to -2.2%. As shown in Table 19, Georgia's higher-than-average growth suggests an increase of 5.2% to 6.3% in the overall 5307 program (again, assuming the FY19 FTA data values are used to set future appropriation levels after the 2020 decennial census). TABLE 19 Georgia's Funding Outlook After 2020 (Assumes FY19 FTA Data Values) Funding Source 5311 rural 5311 and 5340 rural 5307 and 5340 urban Current Predicted Appropriation Appropriation Difference % Difference 21.2M 15.9 to 16.3M -4.9M to -5.3M -23 to -25 24.2M 22.0 to 23.6M -546K to -2.2M -2.3 to -8.9 101M 106 to 107M 5.28 to 6.34M 5.2 to 6.3 Table 20 presents more details on the 5307 program by showing the predicted appropriations for each of the large and small urban areas in Georgia. For each size category, the total national-level appropriation and percent difference in funding is shown to demonstrate how Georgia's urban areas are growing relative to urban areas of comparable size in the U.S. Atlanta is predicted to receive additional 5307 funding, despite the fact that, overall, the funding needs within this program may decrease. Georgia has also seen explosive growth around the Savannah large urban area, and Tennessee and Georgia have seen dramatic growth in the Chattanooga large urban area. Columbus and 82 AugustaRichmond have seen declines and are at risk of losing 5307 funding, even if the overall funding for this program increases by 23 to 24 percent. A large variation also exists within the small urban areas, with Rome and Cartersville seeing explosive growth; AthensClarke, Gainesville, and WarnerRobins experiencing above-average growth; Macon on par with national trends; and Dalton, Brunswick, Hinesville, Valdosta, and Albany showing below-average growth relative to the nation in 5307. For these latter small urbans, this may come as a surprise, as despite the fact these areas have grown since the 2010 decennial census, they may face cuts in their 5307 appropriations. TABLE 20 Funding Outlook for Large and Small Urban Areas in Georgia After 2020 (Assumes FY19 FTA Data Values) Urban Area (Population Size) Current Predicted Appropriation Appropriation Atlanta 69.1M 74.2 to 74.7M Avg. 5307 (1M+) 3.38B 3.00 to 3.06B Savannah 3.5M 13.5 to 13.5M Chattanooga TNGA 3.6M 13.6 to 13.7M Columbus GAAL 3.8M 3.58 to 3.61M AugustaRichmond 2.6M 2.44 to 2.49M Avg. 5307 (200K1M) 839M 1.04 to 1.04B Rome 765K 1.94 to 1.96M Cartersville 604K 1.30 to 1.32M AthensClarke 1.7M 2.85 to 2.94M Gainesville 1.5M 2.44 to 2.45M WarnerRobbins 1.7M 2.5M to 2.7M Macon 1.8M 2.50 to 2.57M Dalton 1.0M 1.27 to 1.30M Brunswick 631K 763 to 798K Hinesville 707K 847 to 854K Valdosta 1.02M 1.20 to 1.26M Albany 1.2M 1.38 to 1.40M Avg. 5307 (50200K) 402M 550 to 608M Note: Does not include the 5340 growing states portion. Difference 5.1 to 5.5M -316 to -358M 10.0 to 10.0M 10.0M to 10.1M -167K to -196K -142K to -196K 196M to 205M 1.18 to 1.19M 693 to 718K 1.2 to 1.3M 901 to 910K 860K to 1.0M 677 to 746K 255 to 283K 132 to 167K 140 to 147K 181 to 233K 128 to 147K 148 to 206M % Difference 7.4 to 8.0 -9 to -11 289 to 290 279 to 280 -4 to -5 -5 to -7 23 to 24 154 to 156 115 to 119 71 to 75 58 to 59 51 to 62 37 to 41 25 to 28 21 to 26 20 to 21 18 to 23 10 to 12 37 to 51 83 4.3.2 County-Level Funding in Georgia Appendix D, Tables D1D5 provide information on the current and expected 5311 and 5307 levels for every county in Georgia. We can identify counties that currently receive a large amount of appropriated funds through the 5311 or 5307 programs, yet do not have transit service for one or both of these programs. This is an indication of strong demand for transit within these counties that is not being met. Table 21 shows the FY19 5311 appropriation for counties in Georgia that currently do not have transit service. There are 37 counties in Georgia that currently do not have transit service; of these, 21 receive appropriations of more than $100K. As shown in Figure 19, there is no clear geographic pattern for these counties. The need to start transit service for rural communities is highest in Barrow, Laurens, and Coffee, which currently receive 5311 appropriations of $357K, $344K, and $290K, respectively. TABLE 21 FY19 5311 Appropriation Levels for Counties That Do Not Have Transit Service County Barrow FY19 5311 Appropriation 356,850 County FY19 5311 Appropriation Stephens 169,326 Laurens 343,592 Washington 162,573 Coffee 290,198 Franklin 150,781 Harris 205,599 Appling 139,727 Newton 199,352 Fayette 116,242 Toombs 190,041 Charlton 110,912 Tattnall 178,355 Oglethorpe 110,363 White 175,914 Jeff Davis 110,267 Emanuel 175,702 Jasper 104,453 Monroe 173,863 Oconee 102,671 Madison 173,755 Note: Only those counties with appropriations greater than $100K are shown. 84 Source: Mapchart.net 2019. FIGURE 19 FY19 5311 Appropriation Levels for Counties That Do Not Have Transit Service Table 22 and Figure 20 show the FY19 5307 appropriation levels for those counties that currently do not have transit service. There are 11 counties in Georgia that receive 5307 appropriations, but only the six shown in Table 22 receive an appropriation of more than $100K. Among the counties shown in Table 22, Houston is clearly the outlier that is eligible to receive $2.25M from the 5307 program, and is in arguably a strong position to initiate transit service. Fayette, Rockdale, and Newton also have federal funds of $467K, $456K, and $395K, respectively, that they would be able to receive if they decided to start transit service under the 5307 program. As shown in Figure 20, Fayette, Rockdale, and Newton are all near the Atlanta metro area. 85 TABLE 22 FY19 5307 Appropriation Levels for Counties That Do Not Have Transit Service County FY19 5307 Appropriation Houston 2,250,481 Fayette 466,882 Rockdale 456,127 Newton 395,258 Oconee 160,141 Chattahoochee 106,684 Note: Only those counties with appropriations greater than $100K are shown. Source: Mapchart.net 2019. FIGURE 20 FY19 5307 Appropriation Levels for Counties That Do Not Have Transit Service 86 One of the reasons that the counties shown in Table 21 and Table 22 may have not initiated transit service is because of economy of scale issues. That is, there may be a level of funding at which it becomes financially feasible for counties to offer transit service. To investigate whether this may be occurring, Figure 21 shows the FY19 5311 appropriation levels for those counties that do offer transit service under the 5311 program and not the 5307 program. A wide range of appropriations at the county level are shown in the figure, but the majority of the counties that offer service (either individually or within a multicounty operation) fall in the $100K$299K range. Source: Mapchart.net 2019. FIGURE 21 FY19 5311 Appropriation Levels for Counties That Have Only 5311 Transit Service Table 23 shows the percentage of counties that offer transit service under the 5311 program among those counties that receive at least $25K annually in FTA 5311 87 appropriations. As the appropriation level increases, so does the probability that the county will offer transit service. A total of 60% of counties that have FTA 5311 appropriations greater than $25K and less than $50K do not offer transit service. This decreases to 34% for those counties that receive between $50K and $100K and to 26% for those counties that receive between $100K and $200K. Above the $200K sample, this percentage levels off at 14% to 15%, suggesting that at above this appropriation threshold other factors other than the subsidy amount may be influencing the county's decision to provide transit. TABLE 23 Comparison of FY19 5311 Appropriation Levels Across Counties FY19 5311 Appropriation 25K 49K 50K 99K 100K 199K 200K 299K 300K or more # Counties That Offer 5311 Transit 4 25 53 18 11 # Counties that Do Not Offer 5311 Transit % Counties that Do Not Offer 5311 Transit 6 60% 13 34% 19 26% 3 14% 2 15% Table 24 sheds light on the difficulties that counties providing 5311 service may be facing in transitioning to a mixed transit service based on funding from both the 5311 and 5307 programs. The three counties of Lowndes, Whitfield, and Forsyth have 5307 appropriations of approximately $1M (specifically, $1.1M, $952K, and $922K, respectively). Four counties have funding around $500K, namely Glynn ($668K), Paulding ($649K), Columbia ($615K), and Coweta ($485K). Among these, Glynn County (where Brunswick, GA, is located) has announced plans to start transit service based on 5307 funds (Brunswick Area Transportation Study 2019). Seven counties in Georgia are eligible 88 for 5307 funding at the $123K$270K level and seven counties are eligible for 5307 funding at the $23K$66K level. TABLE 24 FY19 5307 Appropriation Levels for Counties That Only Offer 5311 Transit Service County Lowndes Whitfield Forsyth Glynn Paulding Columbia Coweta Catoosa Spalding Walker Lee FY19 5307 Appropriation 1,076,358 952,223 922,057 667,852 649,216 615,924 485,128 269,716 231,447 183,474 167,091 County Walton Murray Carroll Peach Bryan Jones Jackson Long Madison Dawson FY19 5307 Appropriation 153,032 133,565 123,215 66,984 53,704 53,219 41,301 33,585 25,064 22,585 Table 25 provides evidence of the difficulties that counties are facing in initiating transit service based on the 5307 program. Compared to rural transit, the threshold at which it becomes feasible for counties to offer transit under the urban program appears to be much higher. As seen in Table 25, with the exception of Liberty, all of the counties receive more than $1M in 5307 appropriations. This suggests that the current regulatory framework to help transit operators in trending urban areas transition from rural to urban funding is not working very well. Simply stated, counties and transit operators are making the decision not to draw down their 5307 funds until the appropriation level reaches around $1M. This could be because of increased reporting requirements associated with offering joint 5311 and 5307 service, or it could be due to the limitations surrounding the use of 5307 89 program funds for operating expenses. Understanding why the thresholds for initiating transit service are so distinct between the 5311 and 5307 programs in Georgia would be an interesting direction for future research. TABLE 25 FY19 5307 Appropriation Levels for Counties That Only Offer 5307 Transit Service County Fulton DeKalb Clayton Gwinnett Cobb Chatham Muscogee Bibb Clarke FY19 5307 Appropriation 25,680,194 19,785,158 7,268,028 6,621,237 6,500,720 3,646,778 3,541,921 2,412,707 2,391,818 County Hall Floyd Richmond Henry Bartow Cherokee Dougherty Douglas Liberty FY19 5307 Appropriation 2,194,364 1,861,312 1,702,365 1,231,829 1,220,368 1,185,905 1,172,357 1,134,652 723,315 4.4 Summary The potential impacts on transit throughout the nation hinge more on the overall appropriation levels that will be targeted to each system type, rather than the disruptions that will be caused from the high-risk transitions of rural transit systems into large urban areas. However, future authorizations and legislations can address both of these core issues by looking closely at how authorization levels should be set and by eliminating and/or adapting the 100 bus rule to allow for current 5311 VRH to be used in place of 5307 VRH. Eliminating the 100 bus rule and simply allowing small systems operating within large urban areas to use FTA 5307 funds toward operating costs would provide the smoothest transition for these high-risk systems and enable them to continue operations 90 with minimal disruptions--that is, there would be no two-year gap in funding and systems could continue to apply the same level of FTA funding toward operating expenses. Alternatively, allowing 5311 VRH to be used as part of the 100 bus rule will effectively eliminate the two-year gap in the use of 5307 funds for operating expenses, but would reduce the amount of FTA funding that can be applied toward operating expenses. The level of operating funding could be maintained by applying a factor to the 5311 VRH in the current formula4. In addition, we recommend that future authorizations explore ways in which subrecipients in areas trending urban can more seamlessly transition to the 5307 program. Our analysis of Georgia shows that many subrecipients are not making this transition until the 5307 appropriation reaches $1M. A possible alternative to the 100 bus rule would be to allow subrecipients with 5307 appropriations of less than $1M to use those funds toward operating expenses.. 4 Determining this factor is outside the scope of the study. However, the basic idea is the following: to calculate the percent of funds that can be used for operating as FACTOR 5311 VRM / TOTAL 5307 VRM for the large urban area, where the FACTOR is a number greater than 1. 91 5 FINDINGS AND RECOMMENDATIONS 5.1 Findings This report predicted changes in funding for the FTA 5311 and 5307 programs through modeling spatial and temporal changes in the U.S. population between 2010 and 2020. We used this model to identify--both for Georgia and the nation--non-urbanized areas that have a high probability of being reclassified as small urban areas and/or of being absorbed into small urban or large urban areas after the 2020 decennial census. This list will be particularly helpful for states and the Federal Transit Administration for identifying transit systems that are at high risk of losing the ability to use their FTA funding for operating expenses after the 2020 decennial census due to being absorbed into large areas. As part of the analysis, we used the predictions from our binary logit models, along with current service data available from the National Transit Database and FTA's Data Value Table (colloquially referred to as "Table 5") to predict if and how the funding needs of rural, small urban, and large urban areas would change after the 2020 decennial census. Our results show a dramatic need for increased funding for small urban areas and large urban areas with populations under 1M. While our analysis shows that the "funding needs" of rural systems would theoretically decrease after the decennial census, we are not advocating for a decrease in the FTA 5311 appropriation for non-urbanized/rural systems. Rural transit systems are predominately demand-responsive systems that operate in lowdensity areas, often transporting individuals over long distances to connect to urban centers with medical and other facilities. The ability to sustain transit systems in these low-density 92 and often isolated rural areas is contingent on appropriation levels remaining at the same or similar level as today. Another key finding from our research is that the 100 bus rule, while designed to provide some flexibility for small systems operating within a large urban area, will likely cause problems for small systems located close to large urban areas after the 2020 decennial census. This has two underlying causes. First, the appropriation language requires that to qualify to use 5307 urban funds for operating expenses, small systems must compare their vehicle revenue hours of service certified under the 5307 program to the vehicle revenue hours of the entire large urban area to which they belong. This effectively creates a two-year gap in operating funding for these systems, as they would not be able to start 5307 service until after the 2020 decennial census and then would have to wait two years to have their data certified and used in the appropriation formula. Second, because the vehicle revenue hours of the majority of small transit systems are miniscule compared to the large urban transit operator, the 100 bus rule effectively reduces the amount of FTA funding these systems will be able to use toward their operating expenses in years three and beyond. Our analysis shows that the overall needs in transit funding by system size will be different in 2020 than they were in 2010. As with any analysis, there are limitations. Many assumptions went into this analysis, and replicating the funding allocation process and the 100 bus rule was quite challenging. Our analysis is robust, though, in the sense that the general funding needs hold across a conservative and aggressive planning scenario. It is our hope that the results of our analysis will be used to help influence future appropriation 93 discussions and to help Georgia and other states identify areas of the nation that are trending urban and at risk for sustaining transit service after the decennial census 5.2 Recommendations Based on our analysis, we offer the following recommendations for future appropriations for the FTA 5311 and 5307 programs: Maintain current appropriation levels for the 5311 program. Significantly increase appropriation levels for the 5307 program on the order of $344M$411M per year and direct this increase in funding to small urban systems and large urban systems serving populations less than 1M. Provide small transit systems with more flexibility of using FTA funds for operating expenses that do not depend on whether these small transit systems serve rural, small urban areas, or large urban areas. One possibility is to allow the use of 5307 towards operating expenses for those systems that receive 5307 appropriations of less than $1M. If the 100 bus rule continues as part of future legislation, then allowing small transit systems that are transitioning to 5307 service to use their 5311 vehicle revenue miles as part of future appropriations would eliminate the "two year gap" in funding that these systems currently face. Sunset the 100 bus rule and tie the definition of "small" transit systems to a different metric, such as vehicle revenue hours. Given the advances in incorporating transportation networking companies (such as Uber and Lyft) into public 94 transportation offerings, the 100 bus rule can arguably result in transit systems not pursuing innovative solutions for fear of losing operating revenues. 95 REFERENCES Atlanta Regional Commission (ARC). (2018). Regional Transit Policy for FTA Urbanized Area Formula Fund Section 5307/Section 5340. Retrieved from https://atlantaregional.org/wp-content/uploads/fta-formula-fund-regional-transitpolicies-section-5307-2018.pdf. Boustan, L.P., Bunten, D., Hearey, O. (2013). Urbanization in the United States, 1800 2000 (Working Paper). Retrieved from https://scholar.princeton.edu/sites/default/files/lboustan/files/research21_urban_ha ndbook.pdf. Brunswick Area Transportation Study (2019). The Brunswick Transportation Study's Unified Planning Work Program: FY 2020 Draft. Available online at https://www.glynncounty.org/DocumentCenter/View/61729/Draft-BATS-FY2020-UPWP. Environmental Systems Research Institute (ESRI). (2017). ArcGIS Release 10.5.1. Redlands, California. Environmental Systems Research Institute (ESRI). (2015). 2015/2020 Population Projections. Available online at https://doc.arcgis.com/en/esridemographics/data/updated-demographics.htm. Federal Register. (2011). Urban Area Criteria for Census 2000 Proposed Criteria. 76 Fed. Reg. 53032-53034 (August 24, 2011). Retrieved from https://www2.census.gov/geo/pdfs/reference/fedreg/fedregv76n164.pdf. Federal Register. (2012). Small Transit Intensive Cities Program. Available online https://books.google.com/books?id=lHPvSk4p9g0C&pg=PA63683#v=onepage& q&f=false. Accessed November 5, 2019. Federal Register. (2019). FTA Fiscal Year 2019 Apportionments, Allocations and Program Information. A Notice by the Federal Transit Administration on 07/03/2019. Available online at https://www.federalregister.gov/documents/2019/07/03/201914248/fta-fiscal-year-2019-apportionments-allocations-and-program-information. Accessed October 12, 2019. Federal Transit Administration (FTA). (n.d. (a)). Grant Programs. Retrieved from https://www.transit.dot.gov/grants. Federal Transit Administration (FTA). (n.d. (b)). Rural Transportation Assistance Program 5311(b)(3). Retrieved from https://www.transit.dot.gov/funding/grants/ruraltransportation-assistance-program-5311b3. 96 Federal Transit Administration (FTA). (2015). Census 2010 and FTA Formula Grants. Retrieved from https://www.transit.dot.gov/sites/fta.dot.gov/files/Census%202010%20and%20FT A%20Formula%20Grants%2012-7-15.pdf. Federal Transit Administration (FTA). (2016). Metropolitan Planning Organization (MPO): Overview. Retrieved from https://www.transit.dot.gov/regulations-andguidance/transportation-planning/metropolitan-planning-organization-mpo. Federal Transit Administration (FTA). (2017). National Transit Database: Policy Manual. Office of Budget and Policy. Retrieved from https://www.transit.dot.gov/sites/fta.dot.gov/files/docs/ntd/57981/2017-ntd- policy-manual_2.pdf. Federal Transit Administration (FTA). (2018). FY18 Comprehensive Review Guide: 20. Section 5311 Program Requirements. Retrieved from https://www.transit.dot.gov/sites/fta.dot.gov/files/docs/regulations-and- guidance/safety/triennial-reviews/69531/fy18-comprehensive-review-guide- section-20-section-5311-program-requirements.pdf. Federal Transit Administration (FTA). (2019a). Census Data on Rural Population and Land Area (Used for the Section 5311 Rural Area Formula Apportionments). Available online at https://www.transit.dot.gov/funding/apportionments/census-data-ruralpopulation-and-land-area-used-section-5311-rural-area. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019b). Census Low Income Population Data (Used for the Section 5307 and 5311 Apportionments). Available online at https://www.transit.dot.gov/funding/apportionments/census-low-incomepopulation-data-used-section-5307-and-5311-apportionments. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019c). Census Urbanized Area Population and Population Density Data (Used for the Section 5303, 5305, and 5307 Apportionments). Available online at https://www.transit.dot.gov/funding/apportionments/census-urbanized-area- population-and-population-density-data-used-section. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019d). National Transit Database Data Used for the Section 5307 Urbanized Area Formula and Section 5339 Bus Formula Apportionments. Available online at https://www.transit.dot.gov/funding/apportionments/national-transit-database- data-used-section-5307-urbanized-area-formula-and. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019e). National Transit Database Data Used for the Section 5311 Apportionments. Available online at 97 https://www.transit.dot.gov/funding/apportionments/national-transit-databasedata-used-section-5311-apportionments. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019f). National Transit Database Data Used for the Small Transit Intensive Cities (STIC) Apportionments. Available online at https://www.transit.dot.gov/funding/apportionments/national-transit-databasedata-used-small-transit-intensive-cities-stic. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019g). National Transit Database and Census Data Used for the Tribal Transit Apportionments. Available online at https://www.transit.dot.gov/funding/apportionments/national-transit-databaseand-census-data-used-tribal-transit-apportionments. Accessed for FY19 on September 24, 2019. Federal Transit Administration (FTA). (2019h). Section 5307 Urbanized Area Formula Program. Available online at: https://www.transit.dot.gov/funding/apportionments/section-5307-urbanized-area- formula-program. Accessed September 24, 2019. Federal Transit Administration (FTA). (2019i). Section 5311 Rural Formula Program. Available online at https://www.transit.dot.gov/funding/apportionments/section5311-rural-formula-program. Accessed September 24, 2019. Federal Transit Administration (FTA). (2019j). Table 3A: FY 2019 Section 5307 Operating Assistance Special Rule Operator Caps (Full Year). Available online at https://www.transit.dot.gov/funding/apportionments/table-3a-fy-2019-section5307-operating-assistance-special-rule-operator-caps. Accessed for FY19 on September 24, 2019. Green, J., and Mahomoui, D. (2017). lehdr: Grab LEHD Flat Files. R package. Retrieved from https://rdrr.io/github/jamgreen/lehdR/. Kuhn, M., Wing, J., Weston, S., Williams, A., Keefer, C., Engelhardt, A., Cooper, T., Mayer, Z., Kenkel, B., the R Core Team, Benesty, M., Lescarbeau, R., Ziem, A., Scrucca, L.,Tang, Y., Candan, C., and Hunt, H. (2018). caret: Classification and Regression Training. R package version 6.0-80. Retrieved from https://CRAN.Rproject.org/package=caret. Long, J.A. (2018). jtools: Analysis and Presentation of Social Scientific Data. R package version 1.0.0. Retrieved from https://cran.r-project.org/package=jtools. Manson, S., Schroeder, J., Van Riper, D. and Rulles, S. (2019). IPUMS National Historical Geographic Information System: Version 14.0. Minneapolis, MN: IPUMS. Available at: https://www.nhgis.org/user-resources/geographic-crosswalks. Mapchart.net (2019). Online GIS tool. Available at: https://mapchart.net/. 98 Pebesma, E. (2018). sf: Simple Features for R. R package version 0.6-3. Retrieved from https://CRAN.R-project.org/package=sf. R Core Team. (2018). The R Project for Statistical Computing. Vienna, Austria. Retrieved from https://www.R-project.org/. Ratcliffe, M., Burd, C., Holder, K., and Fields, A. (2016). Defining Rural at the U.S. Census Bureau: American Community Survey and Geography Brief. Retrieved from https://www2.census.gov/geo/pdfs/reference/ua/Defining_Rural.pdf. Smith, S. K., Tayman, J., Swanson, D. A. (2001). State and Local Population Projections, Methodology and Analysis. New York, NY: Kluwer Academic/Plenum Publishers. Stapel, E. (2012). Exponential Functions: Compound Interest. Retrieved from https://www.purplemath.com/modules/expofcns4.htm. UCLA, Institute for Digital Research and Education. (2011). FAQ: What Are Pseudo R- squareds? Retrieved from https://stats.idre.ucla.edu/other/mult- pkg/faq/general/faq-what-are-pseudo-r-squareds/. U.S. Bureau of Economic Analysis (BEA). (2017). "SA25N Total Full-Time and PartTime Employment by NAICs Industry 1." Retrieved from https://apps.bea.gov/. U.S. Census Bureau. (2001). Census 2000 Summary File 1, United States. Prepared by NHGIS: https://www.nhgis.org. U.S. Census Bureau. (2010). U.S. Census Bureau. 2010 Population Estimates. Prepared by Social Explorer: http://www.socialexplorer.com. U.S. Census Bureau. (2011a). Census 2010 Summary File 1, United States. Prepared by NHGIS: https://www.nhgis.org. U.S. Census Bureau. (2011b). Population Distribution and Change: 2000 to 2010. Retrieved from https://www.census.gov/prod/cen2010/briefs/c2010br-01.pdf. U.S. Census Bureau. (2015). 2010 Census Tallies of Census Tracts, Block Groups & Blocks. Retrieved from https://www.census.gov/geographies/reference-files/timeseries/geo/tallies.html#par_textimage_1. U.S. Census Bureau. (2017). TIGER/Line Shapefiles. Retrieved from https://www.census.gov/geo/maps-data/data/tiger-line.html Walker, K. (2018). tigris: Load Census TIGER/Line Shapefiles. R package version 0.7. https://github.com/walkerke/tigris. Wickham, H. (2018). stringr: Simple, Consistent Wrappers for Common String Operations. R package version 1.3.1. Retrieved from https://CRAN.Rproject.org/package=stringr. 99 Wickham, H., Franois, R., Henry, L., and Mller, K. (2018). dplyr: A Grammar of Data Manipulation. R package version 0.7.5. Retrieved from https://CRAN.Rproject.org/package=dplyr. Wickham, H., Hester, J., and Chang, W. (2018). devtools: Tools to Make Developing R Packages Easier. R package version 1.13.5. Retrieved from https://CRAN.Rproject.org/package=devtools. 100 APPENDIX A: STATE BINARY LOGIT MODELS This appendix reports the binary logit models for each state that we estimated to predict whether an area would be rural or urban after the 2020 decennial census. 101 TABLE A1 Binary Logit Results 102 Alabama Alaska Arizona Arkansas California t-stat t-stat t-stat t-stat t-stat Urban (UC or UA in 2000) 6.25 191.1 8.06 51.6 6.31 181.2 8.91 78.9 7.65 211.0 Closest urban is an urbanized area 1.24 65.5 0.66 9.78 1.7 78.2 0.82 31.1 1.41 90.5 Rural and (0,1] miles from UA 3.56 113.8 5.62 36.3 3.75 119.9 6.27 55.8 4.46 124.7 Rural and (1,2] miles from UA 2.01 49.9 3.67 20.1 2.28 56.5 4.59 39.1 2.67 65.7 Rural and (2,4] miles from UA 1.49 38.3 3.38 18.0 1.05 25.1 3.23 26.6 1.25 26.7 Log of distance to roads -1.31 -60.1 - - - - -0.80 -18.9 -0.97 -64.9 Log of number of jobs in tract 0.34 41.4 0.06 2.56 0.08 11.1 0.25 19.4 0.24 41.3 2010 pop den (0,500] (ref.) PSQM* - - - - - - - - - - 2010 pop den (500,1000] PSQM* 1.71 49.1 1.96 16.4 2.35 52.3 1.84 36.0 2.24 63.8 2010 pop den (1000,2000] PSQM* 1.91 54.0 2.04** 23.8 2.77 62.7 2.00 38.2 2.55 64.2 2010 pop den (2000,4000] PSQM* 2.33 59.7 2.04** 23.8 3.18 72.0 2.51 43.5 3.20 76.5 2010 pop den of 4000+ PSQM* 2.30 49.1 2.04** 23.8 4.33 111.5 3.09 45.1 4.31 133.1 Constant -6.64 -102.3 -6.96 -32.5 -5.48 -111.1 -9.05 -64.0 -7.29 -131.8 R2 0.73 0.80 0.81 0.80 0.84 Accuracy 0.9291 Pred. Pred. No Yes 0.9545 Pred. Pred. No Yes 0.9479 Pred. Pred. No Yes 0.9527 Pred. Pred. No Yes 0.9613 Pred. Pred. No Yes Actual No 128,800 9,068 18,428 765 126,038 6,059 119,589 4,921 230,917 13,104 Actual Yes 7,670 90,576 460 7,267 6,333 99,534 3,284 45,652 13,670 433,755 *Population density in persons per square mile. **For Alaska, population categories estimated for (500,1000] and 1000+ PSQM. Colorado t-stat 7.68 148.0 1.18 48.9 5.11 102.5 3.70 65.1 2.94 51.5 -1.09 -34.7 0.07 6.71 - - 2.01 36.8 2.50 42.5 2.96 51.6 3.95 80.3 -6.15 -73.0 0.81 0.9479 Pred. Pred. No Yes 96,359 5,245 5,012 88,021 TABLE A1 Binary Logit Results (Continued) 103 Connecticut t-stat Urban (UC or UA in 2000) 7.71 33.2 Closest urban is an urbanized area 0.88 15.7 Rural and (0,1] miles from UA 4.20 18.1 Rural and (1,2] miles from UA 2.21 9.10 Rural and (2,4] miles from UA 1.61 6.42 Log of distance to roads -1.90 -24.7 Log of number of jobs in tract 0.18 9.61 2010 pop den (0,500] (ref.) PSQM* - - 2010 pop den (500,1000] PSQM* 2.21 31.3 2010 pop den (1000,2000] PSQM* 2.35 28.0 2010 pop den (2000,4000] PSQM* 2.78 26.9 2010 pop den of 4000+ PSQM* 3.07 29.4 Constant -6.90 -25.9 R2 0.72 Accuracy 0.9455 Pred. Pred. No Yes Actual No 12,954 1,817 Actual Yes 1,663 47,425 *Population density in persons per square mile. Delaware t-stat 6.55 41.0 0.25 4.42 3.97 25.8 2.65 16.4 1.72 10.8 -1.85 -20.4 0.11 4.48 - - 1.86 17.8 2.12 22.1 2.78 24.9 2.87 25.5 -4.46 -20.3 0.62 0.9033 Pred. Pred. No Yes 5,263 1,067 1,181 15,737 Florida t-stat 6.71 230. 0.85 57.7 3.83 142. 2.75 91.5 1.97 63.1 -0.49 -33.8 0.02 3.44 - - 2.01 74.4 2.32 83.1 2.60 84.3 2.83 90.5 -4.40 -96.5 0.72 0.9347 Pred. Pred. No Yes 102,550 13,725 15,997 322,735 Georgia t-stat 6.03 192. 1.40 74.6 3.22 108. 1.45 36.6 0.65 15.4 -1.43 -50.5 0.45 51.0 - - 2.08 62.9 2.31 68.1 2.89 72.1 3.11 61.7 -7.25 -106. 0.77 0.9382 Pred. Pred. No Yes 134,275 9,472 7,301 120,234 Hawaii t-stat 4.29 57.9 2.04 30.9 1.85 25.2 - - - - - - 0.33 17.4 - - 3.02 26.2 3.19 24.1 3.24 21.9 3.30 29.5 -5.82 -40.8 0.70 0.9393 Pred. Pred. No Yes 189,623 12,103 10,129 154,287 Idaho t-stat 6.86 124. 1.19 35.5 4.17 79.7 2.15 29.6 1.72 25.8 -0.94 -24.0 0.47 25.0 - - 1.88 27.0 2.05 31.4 2.60 42.4 3.37 59.6 -8.63 -59.8 0.81 0.9618 Pred. Pred. No Yes 106,125 3,183 2,238 30,188 TABLE A1 Binary Logit Results (Continued) 104 Illinois Indiana Iowa Kansas t-stat t-stat t-stat t-stat Urban (UC or UA in 2000) 8.54 147. 9.11 91.9 8.93 125 8.46 149. Closest urban is an urbanized area 1.57 87.3 1.38 66.2 1.29 41.9 1.28 40.6 Rural and (0,1] miles from UA 5.58 95.2 6.11 61.4 5.92 81.0 5.42 95.4 Rural and (1,2] miles from UA 3.31 49.1 4.14 39.7 3.73 41.0 3.04 39.4 Rural and (2,4] miles from UA 2.58 40.0 2.78 26.3 2.75 31.6 2.15 27.4 Log of distance to roads -1.28 -47.4 -0.51 -16.0 -1.19 -26.8 -0.92 -24.7 Log of number of jobs in tract 0.29 35.7 0.33 33.4 0.50 31.1 0.37 25.1 2010 pop den (0,500] (ref.) PSQM* - - - - - - - - 2010 pop den (500,1000] PSQM* 1.85 40.0 1.89 39.0 1.97 24.4 1.73 24.1 2010 pop den (1000,2000] PSQM* 1.91 46.9 1.97** 64.3 2.30 30.0 2.09 32.2 2010 pop den (2000,4000] PSQM* 2.09 59.3 1.97** 64.3 2.56 40.2 2.61 44.5 2010 pop den of 4000+ PSQM* 3.06 85.2 2.62 67.2 3.17 49.9 3.34 53.3 Constant -8.55 -107. -9.87 -81.4 -10.5 -77.8 -8.86 -74.9 R2 0.83 0.81 0.87 0.86 Accuracy 0.9571 Pred. Pred. No Yes 0.9519 Pred. Pred. No Yes 0.9685 Pred. Pred. No Yes 0.9691 Pred. Pred. No Yes Actual No 163,511 7,834 122,674 6,123 149,019 3,557 165,547 4,275 Actual Yes 10,875 254,053 6,353 124,278 3,081 55,193 2,870 58,558 *Population density in persons per square mile. **For Indiana, one coefficient was estimated for (1000,4000] PSQM. Kentucky t-stat 8.27 86.7 1.13 39.4 5.83 61.3 3.71 35.7 2.95 28.2 -1.41 -22.9 0.28 26.8 - - 1.69 30.8 1.88 34.8 2.10 37.2 2.79 44.1 -8.55 -72.0 0.79 0.9422 Pred. Pred. No Yes 86,571 4,423 3,987 50,402 Louisiana t-stat 8.27 110. 1.15 49.2 5.22 69.9 3.40 41.5 2.82 35.0 -1.73 -37.2 0.32 33.3 - - 1.89 37.0 2.09 39.2 2.39 41.3 2.78 48.2 -8.15 -82.7 0.79 0.9444 Pred. Pred. No Yes 89,115 5,579 4,962 89,906 TABLE A1 Binary Logit Results (Continued) 105 Maine Maryland Massachusetts Michigan Minnesota t-stat t-stat t-stat t-stat t-stat Urban (UC or UA in 2000) 7.21 70.3 6.72 113. 6.69 58.7 7.30 153. 7.07 168. Closest urban is an urbanized area 1.02 19.0 1.15 40.9 0.96 25.8 1.36 62.0 1.18 41.5 Rural and (0,1] miles from UA 4.17 40.2 3.26 56.2 3.36 29.4 4.26 88.4 4.25 107. Rural and (1,2] miles from UA 1.71 11.0 0.93 12.4 1.41 11.1 2.29 38.2 1.78 28.4 Rural and (2,4] miles from UA - - 0.92 13.3 0.96 6.94 1.60 27.3 0.29 3.57 Log of distance to roads - - -1.70 -36.7 -1.04 -31.1 -1.02 -40.8 -0.98 -28.1 Log of number of jobs in tract 0.85 28.9 0.27 23.4 0.36 27.2 0.23 26.6 0.56 42.2 2010 pop den (0,500] (ref.) PSQM* - - - - - - - - - - 2010 pop den (500,1000] PSQM* 1.45 15.1 1.36 24.4 1.77 30.9 1.89 43.0 2.00 35.7 2010 pop den (1000,2000] PSQM* 1.57 14.9 1.47 24.8 1.96** 42.3 1.96 44.7 2.20 40.9 2010 pop den (2000,4000] PSQM* 1.95 18.0 1.84 32.3 1.96** 42.3 2.23 54.0 2.64 56.5 2010 pop den of 4000+ PSQM* 2.07 21.1 2.17 47.8 2.78 43.0 3.09 67.7 3.25 70.5 Constant -12.2 -51.7 -6.11 -61.9 -7.12 -49.3 -6.97 -93.3 -8.95 -86.9 R2 0.83 0.75 0.68 0.84 0.84 Accuracy 0.9681 Pred. Pred. No Yes 0.9427 Pred. Pred. No Yes 0.9445 Pred. Pred. No Yes 0.9610 Pred. Pred. No Yes 0.9632 Pred. Pred. No Yes Actual No 48,157 920 40,921 4,125 21,692 4,108 151,263 5,995 153,597 5,487 Actual Yes 1,054 11,837 3,702 87,780 3,886 114,242 5,945 143,238 3,381 78,193 *Population density in persons per square mile. **For Massachusetts, one coefficient was estimated for (1000,4000] PSQM. Mississippi t-stat 7.37 113. 1.75 58.4 4.87 74.3 3.07 39.8 2.23 28.0 -0.99 -25.8 0.43 35.9 - - 1.61 30.5 1.68 32.3 2.27 41.3 2.37 41.4 -9.02 -84.8 0.76 0.9443 Pred. Pred. No Yes 115,251 4,440 4,771 40,989 TABLE A1 Binary Logit Results (Continued) 106 Missouri t-stat Urban (UC or UA in 2000) 6.64 197. Closest urban is an urbanized area 1.62 76.2 Rural and (0,1] miles from UA 3.71 113. Rural and (1,2] miles from UA 1.61 32.2 Rural and (2,4] miles from UA 0.50 8.82 Log of distance to roads -1.11 -37.9 Log of number of jobs in tract 0.43 42.2 2010 pop den (0,500] (ref.) PSQM* - - 2010 pop den (500,1000] PSQM* 1.75 39.0 2010 pop den (1000,2000] PSQM* 1.88 44.6 2010 pop den (2000,4000] PSQM* 2.21 57.4 2010 pop den of 4000+ PSQM* 2.69 66.3 Constant -7.80 -99.7 R2 0.82 Accuracy 0.9570 Pred. Pred. No Yes Actual No 200,021 7,827 Actual Yes 6,115 109,947 *Population density in persons per square mile. Montana t-stat 11.6 45.6 0.95 16.3 8.58 34.0 6.80 25.8 6.05 23.4 -1.91 -26.2 0.36 12.2 - - 2.16 19.6 2.62 23.2 3.12 26.7 3.73 29.5 -11.6 -36.4 0.90 0.9815 Pred. Pred. No Yes 103,663 1,412 923 20,061 Nebraska t-stat 8.40 126. 1.80 43.1 5.84 86.9 3.62 43.2 1.42 11.2 -1.16 -20.2 0.22 11.9 - - 1.28 13.8 1.77 22.0 2.09 30.6 2.98 43.5 -7.89 -53.9 0.87 0.9712 Pred. Pred. No Yes 139,536 3,253 2,156 42,910 Nevada t-stat 7.94 83.7 2.08 47.4 5.53 62.1 3.91 39.7 2.45 23.3 -0.68 -13.8 - - - - 2.71 27.6 2.53 24.9 2.57 27.4 4.20 48.1 -6.40 -71.8 0.86 0.9615 Pred. Pred. No Yes 47,357 1,526 1,689 32,937 New Hampshire t-stat 5.64 88.0 1.36 30.2 2.81 44.7 0.79 7.74 - - - - 0.65 27.4 - - 1.48 20.7 1.53 19.4 2.22 23.5 2.45 28.0 -9.21 -50.2 0.74 0.9352 Pred. Pred. No Yes 26,538 1,555 1,353 15,441 New Jersey t-stat 6.54 59.0 1.25 29.6 3.05 27.5 1.14 9.24 0.29 2.19 -0.69 -22.0 0.18 14.6 - - 1.99 31.3 2.24 34.4 2.74 40.6 3.77 51.0 -5.99 -40.4 0.73 0.9588 Pred. Pred. No Yes 20,596 3,753 2,824 132,617 TABLE A1 Binary Logit Results (Continued) 107 New Mexico New York North Carolina North Dakota Ohio t-stat t-stat t-stat t-stat t-stat Urban (UC or UA in 2000) 8.77 76.4 6.97 163. 5.91 207. 11.3 50.1 8.49 122. Closest urban is an urbanized area 0.90 28.1 1.28 64.6 1.09 69.1 0.77 12.6 1.34 76.5 Rural and (0,1] miles from UA 6.06 52.7 3.55 82.5 3.29 121. 8.34 37.3 5.46 78.4 Rural and (1,2] miles from UA 4.10 32.9 1.91 36.4 1.80 53.6 5.89 24.9 3.18 42.2 Rural and (2,4] miles from UA 2.20 14.2 1.12 20.9 1.19 36.0 4.83 19.9 2.39 32.2 Log of distance to roads -0.63 -16.5 -0.92 -30.8 -0.90 -41.9 -1.09 -12.7 -1.16 -38.6 Log of number of jobs in tract 0.25 23.2 0.14 17.3 0.30 42.3 0.57 19.0 0.36 43.6 2010 pop den (0,500] (ref.) PSQM* - - - - - - - - - - 2010 pop den (500,1000] PSQM* 2.12 33.9 1.67 44.7 1.86 70.1 1.82 11.4 1.80 46.0 2010 pop den (1000,2000] PSQM* 2.44 35.8 1.88 45.9 2.13 78.0 2.40 13.8 2.09 53.1 2010 pop den (2000,4000] PSQM* 2.64 35.8 2.23 53.5 2.55 81.0 2.89 16.6 2.41 66.4 2010 pop den of 4000+ PSQM* 3.33 48.4 2.86 78.3 3.01 76.2 3.75 21.4 2.80 85.6 Constant -8.95 -65.9 -6.26 -88.8 -6.26 -109. -13.0 -42.4 -9.32 -103. R2 0.84 0.82 0.70 0.90 0.80 Accuracy 0.9633 Pred. Pred. No Yes 0.9581 Pred. Pred. No Yes 0.9151 Pred. Pred. No Yes 0.9857 Pred. Pred. No Yes 0.9473 Pred. Pred. No Yes Actual No 119,250 3,143 137,533 8,283 127,932 13,084 114,095 910 153,970 9,497 Actual Yes 2,949 40,717 5,819 185,103 10,270 123,842 931 13,128 9,029 179,090 *Population density in persons per square mile. **For Oklahoma, one coefficient was estimated for (1000,4000] PSQM. Oklahoma t-stat 7.64 158. 1.66 60.6 4.79 98.5 2.32 33.9 1.82 27.4 -1.15 -32.2 0.31 27.5 - - 1.92 36.4 2.13** 60.1 2.13** 60.1 2.72 56.2 -7.96 -88.7 0.82 0.9597 Pred. Pred. No Yes 170,521 5,333 4,945 74,375 TABLE A1 Binary Logit Results (Continued) 108 Oregon Pennsylvania Rhode Island South Carolina South Dakota t-stat t-stat t-stat t-stat t-stat Urban (UC or UA in 2000) 7.47 120. 6.86 198. 0.64 4.86 6.62 148. 8.16 81.8 Closest urban is an urbanized area 1.25 39.1 0.88 53.9 8.96 25.8 1.53 67.2 1.43 24.6 Rural and (0,1] miles from UA 4.70 75.2 3.58 104. 5.16 14.9 3.84 88.2 5.78 59.4 Rural and (1,2] miles from UA 2.73 33.4 1.91 46.6 - - 2.45 48.5 3.76 29.5 Rural and (2,4] miles from UA 1.27 13.7 0.57 13.1 - - 1.53 29.5 3.26 27.5 Log of distance to roads -1.54 -36.6 -1.47 -53.4 -1.52 -10.5 -1.15 -34.6 -0.96 -12.0 Log of number of jobs in tract 0.59 34.9 0.18 23.4 0.36 8.45 0.33 31.7 0.15 6.85 2010 pop den (0,500] (ref.) PSQM* - - - - - - - - - - 2010 pop den (500,1000] PSQM* 2.46 38.1 1.98 63.7 1.54 8.23 1.83 47.3 1.72 13.9 2010 pop den (1000,2000] PSQM* 2.75 37.8 2.10 65.8 1.55** 11.2 1.92 48.8 1.82 16.2 2010 pop den (2000,4000] PSQM* 3.02 41.0 2.37 74.7 1.55** 11.2 2.36 50.7 2.50 25.2 2010 pop den of 4000+ PSQM* 3.70 60.2 2.80 99.7 3.09 13.2 2.89 49.8 3.51 32.7 Constant -9.77 -72.8 -6.04 -98.6 -8.43 -18.6 -7.27 -86.5 -7.55 -42.5 R2 0.87 0.79 0.79 0.75 0.85 Accuracy 0.9683 Pred. Pred. No Yes 0.9475 Pred. Pred. No Yes 0.9696 Pred. Pred. No Yes 0.9332 Pred. Pred. No Yes 0.9751 Pred. Pred. No Yes Actual No 119,520 3,103 156,942 12,648 2,829 345 84,039 5,719 70,916 1,196 Actual Yes 2,970 65,989 8,668 227,764 363 19,789 5,763 76,422 920 11,851 *Population density in persons per square mile. **For Rhode Island, one coefficient was estimated for (1000,4000] PSQM. Tennessee t-stat 6.42 169. 1.36 68.1 3.63 98.9 1.91 42.1 1.08 22.7 -1.38 -43.1 0.33 46.2 - - 1.83 51.1 2.26 58.7 2.77 63.3 3.31 59.9 -6.98 -113. 0.77 0.9373 Pred. Pred. No Yes 127,792 8,205 6,947 92,005 TABLE A1 Binary Logit Results (Continued) 109 Texas Utah Vermont Virginia Washington t-stat t-stat t-stat t-stat t-stat Urban (UC or UA in 2000) 7.11 311. 6.63 103. 8.18 42.4 6.84 176. 6.82 152. Closest urban is an urbanized area 1.68 144. 1.48 41.6 1.39 14.4 1.46 69.0 1.19 45.5 Rural and (0,1] miles from UA 4.33 195. 4.07 71.8 5.29 28.0 3.84 102. 3.74 86.0 Rural and (1,2] miles from UA 2.77 106. 2.25 29.2 3.39 15.2 2.36 51.4 1.89 31.2 Rural and (2,4] miles from UA 1.85 67.2 0.69 6.97 - - 1.48 30.7 1.27 21.3 Log of distance to roads -0.65 -53.3 -1.07 -22.5 -2.91 -16.7 -1.71 -56.0 -0.96 -33.2 Log of number of jobs in tract 0.23 49.4 0.11 8.58 0.29 6.79 0.29 32.9 0.34 28.2 2010 pop den (0,500] (ref.) PSQM* - - - - - - - - - - 2010 pop den (500,1000] PSQM* 1.86 72.8 2.83 35.4 2.08 14.2 1.56 37.4 2.06 42.2 2010 pop den (1000,2000] PSQM* 2.22 90.1 3.39 44.5 2.43** 20.3 1.94 44.8 2.27 42.5 2010 pop den (2000,4000] PSQM* 2.57 103. 4.18 53.4 2.43** 20.3 2.43 53.0 2.78 52.7 2010 pop den of 4000+ PSQM* 3.65 138. 5.11 59.6 2.65 15.4 2.86 64.8 3.71 73.9 Constant -6.92 -179 -5.92 -54.0 -8.58 -24.9 -6.93 -93.5 -7.16 -73.7 R2 0.82 0.84 0.81 0.80 0.81 Accuracy 0.9512 Pred. Pred. No Yes 0.9589 Pred. Pred. No Yes 0.9661 Pred. Pred. No Yes 0.9493 Pred. Pred. No Yes 0.9545 Pred. Pred. No Yes Actual No 414,635 19,729 70,054 2,756 22,722 506 142,203 7,040 76,880 4,878 Actual Yes 23,007 419,148 1,931 39,397 456 4,713 6,298 107,621 3,462 98,175 *Population density in persons per square mile. **For Vermont, one coefficient was estimated for (1000,4000] PSQM. West Virginia t-stat 7.17 113. 0.36 13.2 4.52 71.9 3.16 44.5 2.29 32.3 -2.01 -43.0 0.27 20.8 - - 1.48 28.9 1.64 32.1 1.72 33.4 2.07 42.9 -6.92 -64.4 0.75 0.9379 Pred. Pred. No Yes 82,652 4,926 2,890 35,299 TABLE A1 Binary Logit Results (Continued) Wisconsin Wyoming t-stat t-stat Urban (UC or UA in 2000) 6.91 175. 9.18 66.1 Closest urban is an urbanized area 1.00 43.7 0.51 8.57 Rural and (0,1] miles from UA 4.05 109. 6.40 46.5 Rural and (1,2] miles from UA 1.49 24.1 3.58 19.3 Rural and (2,4] miles from UA 1.07 19.4 - - Log of distance to roads -1.50 -41.9 -4.50 -19.2 Log of number of jobs in tract 0.26 24.4 0.92 12.6 2010 pop den (0,500] (ref.) PSQM* - - - - 2010 pop den (500,1000] PSQM* 2.25 47.2 1.95 13.8 2010 pop den (1000,2000] PSQM* 2.44 52.9 2.54** 24.3 2010 pop den (2000,4000] PSQM* 2.88 66.1 2.54** 24.3 2010 pop den of 4000+ PSQM* 3.17 71.4 3.78 27.3 Constant -6.65 -79.3 -9.79 -36.2 R2 0.81 0.87 Accuracy 0.9537 Pred. Pred. No Yes 0.9747 Pred. Pred. No Yes Actual No 137,391 6,587 67,169 1,265 Actual Yes 4,169 84,008 866 14,810 *Population density in persons per square mile. **For Wyoming, one coefficient was estimated for (1000,4000] PSQM. 110 APPENDIX B: RURAL AREAS PREDICTED TO MERGE WITH URBAN CLUSTERS OR URBAN AREAS Table B1 provides a list of rural (or, more precisely, non-urbanized) areas that are predicted to merge with other urban clusters or urban areas after the 2020 decennial census. Those rural/non-urbanized areas that are absorbed into large urban areas are at risk for losing the ability to use FTA funds for operating expenses for two years and may see significant reductions in the amount of FTA funds that can be used to support operating expenses in years three and beyond. 111 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas after 2020 under Different Scenarios 112 State 2010 Urban Cluster Athens, AL Name of UC/UA Predicted to Merge Into Huntsville, AL UC/ UA UA 50% 50% mi 0 mi 1* 0 Grand Bay, AL Mobile, AL UA 1* 1 Alabama Hazel Green, AL Huntsville, AL UA 1* 1 Priceville, AL Decatur, AL UA 1 1 Robertsdale, AL DaphneFairhope, AL UA 1 1 Buckeye, AZ AvondaleGoodyear, AZ UA 1* 1 Bullhead City, AZNV Laughlin, NV UC 1 0 Lake of the WoodsPinetop-Lakeside, AZ Show Low, AZ UC 1 0 Marana West, AZ Tucson, AZ UA 1* 0 Arizona Nogales, AZ Rio Rico Northeast, AZ Rio Rico Northeast, AZ Nogales, AZ UC 1 0 UC 1 0 Show Low, AZ Lake of the WoodsPinetop-Lakeside, AZ UC 1 0 Somerton, AZ Yuma, AZCA UA 1 1 Vail, AZ Tucson, AZ UA 1* 0 Vistancia, AZ PhoenixMesa, AZ UA 1* 0 AuburnNorth Auburn, CA Sacramento, CA UA 1* 0 Carmel Valley Village, CA SeasideMonterey, CA UA 1 1 Cottonwood, CA Redding, CA UA 1 0 Forestville, CA California Galt, CA Santa Rosa, CA Lodi, CA UA 1* 0 UA 1 1 Half Moon Bay, CA San FranciscoOakland, CA UA 1* 0 Mecca, CA IndioCathedral City, CA UA 1* 0 Nipomo, CA Arroyo GrandeGrover Beach, CA UA 1 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 75% mi 0 0 1 1 1 1 1 1 0 1 1 1 0 0 0 1 1 0 0 0 0 0 1 75% 0 mi 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas after 2020 under Different Scenarios (Continued) 113 State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ 50% 50% UA mi 0 mi Edwards, CO Vail, CO UC 1 0 Colorado FirestoneFrederick, CO Lochbuie, CO Longmont, CO DenverAurora, CO UA 1 1 UA 1* 1 Vail, CO Edwards, CO UC 1 0 Jewett City, CT Connecticut Willimantic, CT Worcester, MACT Hartford, CT UA 1* 0 UA 1* 1 Bridgeville, DE Salisbury, MDDE UA 1 1 Georgetown, DE Millsboro, DE UC 1 1 Delaware Middletown, DE Milford, DE Philadelphia, PANJDEMD Dover, DE UA 1* 1 UA 1 1 Millsboro, DE Georgetown, DE UC 1 1 Ocean View, DE Ocean Pines, MDDE UC 1 1 Crooked Lake Park, FL Winter Haven, FL UA 1* 0 Crystal River, FL Homosassa SpringsBeverly HillsCitrus Springs, FL UA 1 1 Fernandina Beach, FL Yulee, FL UC 1 1 Four Corners, FL Orlando, FL UA 1* 1 Golden Gate Estates, FL Bonita Springs, FL UA 1* 0 Jupiter Farms, FL Miami, FL UA 1* 0 Florida Panama City Northeast, FL Panama City, FL UA 1 1 Poinciana, FL Kissimmee, FL UA 1* 1 Rainbow Lakes Estates, FL Homosassa SpringsBeverly HillsCitrus Springs, FL UA 1 0 Santa Rosa Beach, FL Fort Walton BeachNavarreWright, FL UA 1* 0 Sugarmill Woods, FL Homosassa SpringsBeverly HillsCitrus Springs, FL UA 1 1 Wedgefield, FL Orlando, FL UA 1* 0 Yulee, FL Fernandina Beach, FL UC 1 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 75% mi 1 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 1 75% 0 mi 0 0 0 0 0 1* 1 0 0 0 0 0 0 1 1 1* 0 0 1 1* 0 0 0 0 1 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas after 2020 under Different Scenarios (Continued) 114 State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ 50% UA mi Buckhead (Bryan County), GA Savannah, GA UA 1* Georgia Lula, GA Monroe, GA Gainesville, GA Atlanta, GA UA 1 UA 1* Winder, GA Atlanta, GA UA 1* Hawaii HaleiwaWaialuaPupukea, HI Urban Honolulu, HI PukalaniMakawaoHaikuPauwela, HI Kahului, HI UA 1* UA 1 Idaho Rathdrum, ID Coeur d'Alene, ID UA 1 Lake Holiday, IL Chicago, ILIN UA 1* Illinois Murphysboro, IL Carbondale, IL UA 1 Wonder Lake, IL Round Lake BeachMcHenryGrayslake, ILWI UA 1* Indiana Charlestown, IN Lowell, IN Louisville/Jefferson County, KYIN Chicago, ILIN UA 1 UA 1* Kentucky Nicholasville, KY Wilmore, KY Wilmore, KY Nicholasville, KY UC 1 UC 1 Donaldsonville, LA Baton Rouge, LA UA 1* Louisiana GallianoLaroseCut Off, LA GramercyLutcher, LA Houma, LA New Orleans, LA UA 1 UA 1* Rayne, LA Lafayette, LA UA 1* Glenwood, MD Baltimore, MD UA 1* Maryland Manchester, MD Romancoke, MD Baltimore, MD Baltimore, MD UA 1* UA 1* Ocean Pines, MDDE Ocean View, DE UC 1 North Brookfield, MA Massachusetts North Adams, MAVT Worcester, MACT Pittsfield, MA UA 1* UA 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 50% 0 mi 0 0 0 1 1 1 0 1 1 0 1 0 0 0 1 0 1 0 0 0 0 1 0 0 75% mi 0 0 1 1 0 0 0 1 1 1 1 1 1 1 1 0 1 0 1 0 0 0 1 1 75% 0 mi 0 0 0 1* 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas after 2020 under Different Scenarios (Continued) State 2010 Urban Cluster Cedar Springs, MI Name of UC/UA Predicted to Merge Into Grand Rapids, MI UC/ UA UA 50% 50% mi 0 mi 1* 1 Fowlerville, MI South LyonHowell, MI UA 1 0 Michigan Goodrich, MI Detroit, MI UA 1* 0 Paw, MI Kalamazoo, MI UA 1* 1 Sparta, MI Grand Rapids, MI UA 1* 0 MonticelloBig Lake, MN Minnesota Stewartville, MN MinneapolisSt. Paul, MNWI Rochester, MN UA 1* 1 UA 1 0 Canton, MS Mississippi Gautier, MS Jackson, MS Pascagoula, MS UA 1* 1 UA 1 1 Branson, MO Forsyth, MO UC 1 0 Eureka, MO St. Louis, MOIL UA 1* 0 Missouri Forsyth, MO Platte City, MO Branson, MO Kansas City, MOKS UC 1 0 UA 1* 1 Smithville North, MO Kansas City, MOKS UA 1* 1 Willard, MO Springfield, MO UA 1* 0 Belgrade, MT Bozeman, MT UC 1 0 Montana Bozeman, MT Laurel, MT Belgrade, MT Billings, MT UC 1 0 UA 1 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 75% mi 1 0 0 0 1 1 0 0 0 0 0 0 1 1 1 1 1 0 75% 0 mi 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 115 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas after 2020 under Different Scenarios (Continued) 116 State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ 50% 50% UA mi 0 mi Nebraska Plattsmouth, NE Omaha, NEIA UA 1* 0 Nevada Laughlin, NV Bullhead City, AZNV UC 1 0 New Hampshire Concord, NH Epping, NH Manchester, NH Boston, MANHRI UA 1 1 UA 1* 1 New Jersey Newton, NJ New YorkNewark, NYNJCT UA 1* 1 Aztec, NM New Mexico Kirtland, NM Farmington, NM Farmington, NM UA 1 0 UA 1 0 Bedford, NY New YorkNewark, NYNJCT UA 1* 0 Chester, NY PoughkeepsieNewburgh, NYNJ UA 1* 0 New York Lockport, NY Maybrook, NY Buffalo, NY Walden, NY UA 1* 0 UC 1 0 Ravena, NY AlbanySchenectady, NY UA 1* 0 Walden, NY PoughkeepsieNewburgh, NYNJ UA 1* 0 Archer LodgeClayton, NC Raleigh, NC UA 1* 1 Fearrington Village, NC Durham, NC UA 1* 1 Grifton, NC Greenville, NC UA 1 1 Havelock, NC New Bern, NC UA 1 1 Lake Norman of Catawba, NC Charlotte, NCSC UA 1* 1 North Carolina Maiden, NC Oak Island, NC Hickory, NC St. James, NC UA 1* 0 UC 1 0 PinehurstSouthern Pines, NC Whispering Pines, NC UC 1 0 Smithfield, NC Raleigh, NC UA 1* 1 St. James, NC Oak Island, NC UC 1 0 WendellZebulon, NC Raleigh, NC UA 1* 0 Whispering Pines, NC PinehurstSouthern Pines, NC UC 1 0 North Dakota Lincoln, ND Bismarck, ND UA 1 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 75% mi 1 1 0 1 1 1 1 1 1 1 1 0 1 0 0 0 0 1 0 1 1 0 1 1 1 0 75% 0 mi 0 0 0 0 1* 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas After 2020 Under Different Scenarios (Continued) 117 State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ 50% 50% UA mi 0 mi Ashtabula, OH Conneaut, OH UC 1 1 Ohio Conneaut, OH Genoa, OH Ashtabula, OH Toledo, OHMI UC 1 1 UA 1* 0 Sandusky, OH LorainElyria, OH UA 1 0 Claremore, OK Tulsa, OK UA 1* 0 Oklahoma Collinsville, OK Tulsa, OK UA 1* 1 Harrah, OK Oklahoma City, OK UA 1* 0 Oregon Aumsville, OR Salem, OR UA 1* 0 Burgettstown, PA Pittsburgh, PA UA 1* 0 Fairdale, PA Masontown, PA UC 1 0 Jersey Shore, PA Lock Haven, PA UC 1 1 Lock Haven, PA Jersey Shore, PA UC 1 1 Lykens, PA Pennsylvania Masontown, PA Williamstown, PA Fairdale, PA UC 1 1 UC 1 0 Quarryville, PA Lancaster, PA UA 1* 0 Roaring Spring, PA Altoona, PA UA 1 1 Saw Creek, PA East Stroudsburg, PANJ UA 1 1 Williamstown, PA Lykens, PA UC 1 1 Camden, SC Columbia, SC UA 1* 0 Chesnee, SC Spartanburg, SC UA 1* 1 South Carolina Clover, SC Lake Murray North Shore, SC Seneca, SC Rock Hill, SC Columbia, SC Greenville, SC UA 1 0 UA 1* 0 UA 1* 0 Sun City Hilton Head, SC Hilton Head Island, SC UA 1 0 York, SC Rock Hill, SC UA 1 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 75% mi 1 1 1 1 0 1 0 0 1 1 1 1 1 1 1 1 0 1 1 0 0 1 1 1 0 75% 0 mi 1 1 0 0 0 0 0 0 0 0 1 1 1 0 0 1 0 1 0 0 0 0 0 0 0 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas After 2020 Under Different Scenarios (Continued) 118 State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ UA 50% mi 50% 0 mi South Dakota Brandon, SD Harrisburg, SD Sioux Falls, SD Sioux Falls, SD UA 1 1 UA 1 1 Arlington, TN Memphis, TNMSAR UA 1* 1 Atoka, TN Memphis, TNMSAR UA 1* 0 Jasper, TN Tennessee Norris, TN South Pittsburg, TNAL Knoxville, TN UC 1 0 UA 1* 0 White Pine, TN Morristown, TN UA 1 0 South Pittsburg, TNAL Jasper, TN UC 1 0 Aledo, TX Weatherford, TX UC 1 0 Alvarado, TX DallasFort WorthArlington, TX UA 1* 0 Anna, TX McKinney, TX UA 1* 0 Boerne, TX San Antonio, TX UA 1* 1 Canyon, TX Mescalero Park, TX UC 1 0 Cleburne, TX DallasFort WorthArlington, TX UA 1* 1 Cleveland, TX Houston, TX UA 1* 1 Deerwood, TX ConroeThe Woodlands, TX UA 1* 0 Texas Denton Southwest, TX DentonLewisville, TX UA 1* 1 Devine, TX Lytle, TX UC 1 1 Forney, TX DallasFort WorthArlington, TX UA 1* 1 Granbury, TX Pecan Plantation, TX UC 1 0 Grangerland, TX Houston, TX UA 1* 1 Hempstead, TX Prairie View, TX UC 1 0 Homesteads Addition, TX DallasFort WorthArlington, TX UA 1* 1 Justin, TX DallasFort WorthArlington, TX UA 1* 0 Lake Conroe Eastshore, TX ConroeThe Woodlands, TX UA 1* 1 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 75% mi 0 0 1 0 1 0 0 1 0 0 0 1 0 1 1 0 1 1 0 1 0 1 0 0 1 75% 0 mi 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas After 2020 Under Different Scenarios (Continued) 119 State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ 50% UA mi Lake Conroe Northshore, TX ConroeThe Woodlands, TX UA 1* Lake Conroe Westshore, TX ConroeThe Woodlands, TX UA 1* Lytle, TX Devine, TX UC 1 Magnolia, TX Houston, TX UA 1* Manor, TX Austin, TX UA 1* Mescalero Park, TX Canyon, TX UC 1 Texas (Continued) Odem, TX Paloma Creek SouthPaloma Creek, TX Pecan Acres, TX Corpus Christi, TX DallasFort WorthArlington, TX DallasFort WorthArlington, TX UA 1* UA 1* UA 1* Pecan Plantation, TX Granbury, TX UC 1 Prairie View, TX Hempstead, TX UC 1 Rio Hondo, TX Harlingen, TX UA 1 Seguin, TX San Antonio, TX UA 1* Springtown, TX DallasFort WorthArlington, TX UA 1* Weatherford, TX Aledo, TX UC 1 Park City, UT Summit Park, UT UC 1 Utah Santaquin, UT ProvoOrem, UT UA 1* Summit Park, UT Park City, UT UC 1 Vermont Milton, VT Burlington, VT UA 1 Virginia Purcellville, VA Washington, DCVAMD UA 1* Granite Falls, WA Marysville, WA UA 1 Washington Indianola, WA Bremerton, WA UA 1* Snoqualmie, WA Seattle, WA UA 1* Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 50% 0 mi 1 1 1 1 1 0 1 0 1 0 0 1 0 0 0 0 0 0 1 1 1 0 0 75% mi 0 1 1 0 1 0 0 1 0 1 1 1 0 0 0 1 0 1 0 1 0 1 1 75% 0 mi 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1* 0 0 0 TABLE B1 List of UCs Predicted to Merge with Other UCs or Urban Areas After 2020 Under Different Scenarios (Continued) State 2010 Urban Cluster Name of UC/UA Predicted to Merge Into UC/ 50% 50% 75% 75% UA mi 0 mi mi 0 mi Burlington, WI Milwaukee, WI UA 1* 0 1 0 Lake Geneva, WI Walworth, WI UC 1 0 0 0 Wisconsin Mukwonago, WI Union Grove, WI Milwaukee, WI Racine, WI UA 1* 0 0 0 UA 1 0 0 0 Walworth, WI Lake Geneva, WI UC 1 0 0 0 Hudson, WIMN MinneapolisSt. Paul, MNWI UA 1* 0 1 0 Note: A * represents a high-risk transition (rural to large urban); a 1 without a * represents a UC growing into a small UA. 120 APPENDIX C: SUPPORTING TABLES FOR PREDICTED CHANGES IN 5311 AND 5307 FUNDING ALLOCATIONS This appendix contains supporting tables for the following future scenarios: Scenario 1A corresponds to the 50% probability model using a mile distance threshold Scenario 1B corresponds to the 50% probability model using a 0 mile distance threshold Scenario 2A corresponds to the 75% probability model using a mile distance threshold Scenario 2B corresponds to the 75% probability model using a 0 mile distance threshold Table C1 summarizes the predicted changes in urbanized populations and land area that are the key inputs to determining the 5311 funding apportionments. Table C1 also shows the percentage of the total 5311 funding each state currently receives and compares this to the percentage each state is forecasted to receive after the 2020 decennial census. Note that the percentages shown in Table C1 do not take into account any overall change in the total 5311 apportionment, but instead represent an "apportionment quotient." The apportionment quotient allows for a comparison of relative changes in funding categories without relying on funding data. The quotient represents each state's unconstrained share of the appropriated funds through the 5311 formula. This quotient was calculated by 121 dividing each state's national share of non-urbanized land area and population over the total non-urbanized land area and population for the U.S. in 2010. Each state's land area portion was multiplied by 20% and its population portion was multiplied by 80%. These two percentages were used to determine the state's total apportionment. No state was eligible to receive more than a 5% share of its portion of non-urbanized land area (i.e., Alaska and Texas). This was not corrected for, but this only affects 1.98% of funding nationally, from one state (i.e., Texas). Tables C2 to C5 show the changes in 5311 and 5307 funding each state would experience if the FTA data values from FY19 were applied to the new population, population density, and other inputs used in the allocation formula after the 2020 census. Table C2 reports the forecasts for the 5311 program. Note that this includes only the 5311 portion and not the 5340 growing states portion. FTA typically publishes the combined total of the 5311 and 5340 programs, e.g., in FY2019 this combined appropriation was $24.5M for Georgia and $716.4M nationally for the 50 states. The last two columns on Table C2 show the combined total of the 5311 and 5340 programs by state5 and the percent of the total funding that is associated with the 5340 program. In a separate calculation, we determined that of the $24.5M for Georgia, approximately 13% corresponds to the growing states program and 87% to the 5311 program. Thus, our overall number for the 5311 program of $629M would be approximately equivalent to $723M at the national level for the combined 5311 and 5340 program, which is 5 Data for American Samoa, Guam, N. Marina Islands, Puerto Rico is not included in our analysis. In FY19, they received $3,856,817 in combined 5311 and 5340 funding. 122 consistent with the number published by the Federal Register (Federal Register 2019). That is, we estimate that our methodology to replicate the federal funding formulas are within 1 to 2 percent of the actual appropriation. Tables C3 to C5 report the forecasts for the 5307 program, which are categorized for large urban areas with populations of 1M or more, large urban areas with populations of 200K1M, and small urban areas with populations of 50K200K, respectively. Note that the total amount of 5311 and 5307 fund appropriated after the 2020 decennial census as of the time of this writing is unknown; as such, these results reflect the shift in funding needs from rural and large urban areas to smaller urban areas that would be needed after the 2020 decennial census to effectively "maintain" current levels of FTA funding that transit systems in each category currently receive. Similar to before, our 5307 calculations do not include the growing states portion. FTA typically publishes the combined total of the 5307 and 5340 programs, e.g., in FY2019 this combined appropriation was $101M for Georgia and $5.33B nationally. In our separate calculations, we determined that of this $101M, approximately 4% corresponds to the growing states program and 96% to the 5311 program. Thus, our overall number for the 5307 program of $4.62B would be approximately equivalent to $4.80B at the national level, which is consistent with the number published in the Federal Register of $4.83B (Federal Register 2019). That is, we estimate that our methodology to replicate the federal funding formulas are within 3 to 4 percent of the actual appropriation. 123 TABLE C1 Percent Change in 5311 Apportionment (20102020) 124 National Population Share (%) National Land Area Share (%) Total 5311 Apportionment (%) % Change in Apportionment Year 2010 2020 2010 2020 2010 2020 Scen- 50% 50% 75% 75% 50% 50% 75% 75% 50% 50% 75% 75% 50% 50% 75% 75% ario mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi AK 0.44 0.43 0.42 0.42 0.41 16.56 7.42 7.41 7.37 7.37 3.67 1.83 1.82 1.81 1.80 -1.84 -1.85 -1.86 -1.86 AL 2.76 2.74 2.70 2.75 2.74 1.42 1.51 1.51 1.52 1.52 2.49 2.49 2.47 2.50 2.49 0.00 -0.03 0.01 0.01 AR 1.98 2.10 2.04 2.02 2.00 1.49 1.60 1.60 1.59 1.59 1.89 2.00 1.96 1.93 1.92 0.11 0.07 0.05 0.04 AS* 1.43 1.46 1.43 1.40 1.43 3.24 3.49 3.49 3.47 3.24 1.80 1.87 1.84 1.81 1.79 0.07 0.04 0.01 0.00 AZ 0.06 1.19 1.20 1.34 1.35 0.00 3.46 3.46 3.47 3.47 0.05 1.64 1.65 1.76 1.78 1.59 1.60 1.71 1.73 CA 4.14 4.17 4.16 4.16 4.19 4.31 4.60 4.60 4.61 4.61 4.18 4.25 4.25 4.25 4.27 0.08 0.07 0.07 0.11 CO 1.31 1.31 1.28 1.28 1.31 2.97 3.19 3.18 3.18 3.18 1.64 1.68 1.66 1.66 1.69 0.04 0.02 0.02 0.05 CT 0.61 0.55 1.59 0.55 0.55 0.09 0.09 0.11 0.10 0.10 0.51 0.46 1.29 0.46 0.46 -0.05 0.79 -0.05 -0.04 DC* 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 DE 0.32 0.27 0.26 0.34 0.34 0.05 0.05 0.05 0.05 0.05 0.26 0.23 0.22 0.28 0.28 -0.04 -0.04 0.02 0.02 FL 2.66 2.26 2.24 2.43 2.45 1.36 1.39 1.39 1.44 1.45 2.40 2.08 2.07 2.23 2.25 -0.32 -0.33 -0.16 -0.14 GA 3.78 3.62 3.56 3.62 3.61 1.55 1.64 1.64 1.66 1.66 3.33 3.23 3.18 3.23 3.22 -0.10 -0.15 -0.10 -0.10 GU* 0.18 0.18 0.18 0.17 0.18 0.01 0.01 0.01 0.01 0.01 0.14 0.15 0.14 0.14 0.14 0.00 0.00 0.00 0.00 HI 0.44 0.43 0.42 0.46 0.46 0.18 0.19 0.19 0.19 0.19 0.39 0.38 0.37 0.41 0.40 0.00 -0.01 0.02 0.02 IA 2.00 2.04 1.98 1.97 1.95 1.60 1.72 1.72 1.71 1.71 1.92 1.97 1.93 1.92 1.91 0.05 0.01 0.00 -0.01 ID 0.87 0.86 0.85 0.84 0.83 2.39 2.53 2.52 2.51 2.51 1.18 1.19 1.18 1.17 1.17 0.02 0.01 0.00 0.00 IL 2.89 2.81 2.75 2.80 2.80 1.51 1.60 1.60 1.62 1.62 2.62 2.57 2.52 2.56 2.56 -0.05 -0.10 -0.06 -0.05 IN 2.98 2.95 2.89 2.87 2.86 0.98 1.04 1.04 1.05 1.05 2.58 2.57 2.52 2.50 2.50 -0.01 -0.06 -0.08 -0.07 KS 1.60 1.64 1.60 1.59 1.57 2.35 2.53 2.53 2.52 2.52 1.75 1.82 1.79 1.77 1.76 0.07 0.04 0.02 0.02 KY 2.88 2.95 2.87 2.85 2.82 1.12 1.20 1.20 1.20 1.20 2.53 2.60 2.54 2.52 2.50 0.07 0.01 -0.01 -0.02 LA 1.97 1.93 1.92 1.96 1.96 1.21 1.29 1.28 1.29 1.29 1.82 1.80 1.79 1.83 1.83 -0.02 -0.03 0.00 0.01 MA 0.71 0.66 0.67 0.66 0.68 0.14 0.15 0.15 0.15 0.15 0.60 0.56 0.57 0.56 0.58 -0.04 -0.03 -0.04 -0.02 MD 1.07 1.06 1.05 1.07 1.06 0.23 0.24 0.24 0.24 0.24 0.90 0.89 0.88 0.90 0.90 -0.01 -0.02 0.00 0.00 ME 1.10 1.16 1.13 1.12 1.11 0.89 0.95 0.95 0.95 0.95 1.06 1.12 1.09 1.09 1.08 0.06 0.03 0.03 0.02 MI 3.74 3.62 3.55 3.55 3.53 1.55 1.66 1.66 1.66 1.66 3.30 3.23 3.17 3.17 3.15 -0.08 -0.14 -0.13 -0.14 MN 2.51 2.52 2.47 2.45 2.47 2.28 2.44 2.44 2.43 2.43 2.46 2.51 2.46 2.45 2.46 0.04 0.00 -0.01 0.01 MO 2.93 2.97 2.91 2.89 2.89 1.95 2.08 2.08 2.09 2.09 2.73 2.79 2.75 2.73 2.73 0.06 0.02 0.00 0.01 *AS=American Samoa, DC=Washington DC, GU=Guam, MP=Northern Marina Islands, PR= Puerto Rico, VI=Virgin Islands. Other initials represent states. TABLE C1 Percent Change in 5311 Apportionment (20102020) (Continued) 125 National Population Share National Land Area Share Total 5311 Apportionment % Change in Apportionment Year 2010 2020 2010 2020 2010 2020 Scen- 50% 50% 75% 75% 50% 50% 75% 75% 50% 50% 75% 75% 50% 50% 75% 75% ario mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi MP* 0.06 0.06 2.37 2.38 2.36 0.01 0.01 1.43 1.44 1.44 0.05 0.05 0.05 0.05 0.05 0.00 0.00 0.00 -0.00 MS 2.42 2.43 0.75 0.81 0.80 1.34 1.44 4.54 4.51 4.51 2.20 2.23 2.18 2.19 2.17 0.03 -0.02 -0.01 -0.03 MT 0.82 0.77 4.61 4.77 4.75 4.22 4.54 1.37 1.39 1.39 1.50 1.52 1.51 1.55 1.54 0.02 0.01 0.05 0.04 NC 4.84 4.64 0.49 0.49 0.49 1.31 1.37 2.15 2.14 2.14 4.14 3.99 3.96 4.09 4.08 -0.15 -0.18 -0.05 -0.06 ND 0.45 0.50 0.95 0.93 0.93 2.00 2.15 2.38 2.37 2.37 0.76 0.83 0.82 0.82 0.82 0.07 0.06 0.06 0.06 NE 0.95 0.96 0.72 0.78 0.78 2.22 2.38 0.26 0.26 0.26 1.20 1.25 1.23 1.22 1.22 0.05 0.03 0.02 0.02 NH 0.78 0.74 0.70 0.70 0.70 0.24 0.26 0.14 0.14 0.14 0.67 0.64 0.63 0.68 0.68 -0.03 -0.04 0.01 0.01 NJ 0.77 0.72 0.98 1.06 1.07 0.13 0.14 3.77 3.75 3.75 0.64 0.60 0.59 0.59 0.58 -0.04 -0.05 -0.05 -0.06 NM 1.07 0.99 0.06 0.06 0.06 3.51 3.77 0.01 0.01 0.01 1.56 1.54 1.54 1.60 1.60 -0.02 0.00 0.00 0.00 NV 0.34 0.44 0.43 0.44 0.43 3.17 3.40 3.40 3.39 3.39 0.90 1.04 1.03 1.03 1.02 0.13 0.12 0.12 0.12 NY 3.78 3.73 3.76 3.63 3.65 1.27 1.36 1.36 1.35 1.35 3.28 3.26 3.28 3.17 3.19 -0.02 0.00 -0.10 -0.08 OH 4.51 4.43 4.37 4.30 4.32 1.08 1.15 1.15 1.15 1.16 3.82 3.77 3.73 3.67 3.69 -0.05 -0.09 -0.15 -0.12 OK 2.29 2.27 2.26 2.35 2.34 1.97 2.11 2.11 2.10 2.10 2.22 2.24 2.23 2.30 2.29 0.01 0.00 0.08 0.07 OR 1.62 1.68 1.64 1.64 1.62 2.76 2.97 2.97 2.96 2.96 1.85 1.94 1.91 1.90 1.89 0.09 0.06 0.05 0.05 PA 4.19 4.14 4.05 4.02 4.00 1.18 1.27 1.27 1.26 1.26 3.59 3.57 3.49 3.47 3.45 -0.02 -0.10 -0.12 -0.12 PR* 0.39 0.40 0.39 0.38 0.39 0.05 0.06 0.06 0.06 0.05 0.32 0.33 0.32 0.31 0.32 0.01 0.00 -0.01 0.00 RI 0.11 0.11 0.11 0.11 0.11 0.02 0.02 0.02 0.02 0.02 0.09 0.09 0.09 0.09 0.09 0.00 0.00 0.00 0.00 SC 2.30 2.08 2.10 2.12 2.17 0.82 0.85 0.85 0.87 0.87 2.01 1.83 1.85 1.87 1.91 -0.17 -0.15 -0.13 -0.09 SD 0.64 0.68 0.66 0.67 0.67 2.20 2.36 2.36 2.35 2.35 0.95 1.02 1.00 1.01 1.00 0.06 0.05 0.05 0.05 TN 3.26 3.26 3.21 3.28 3.27 1.13 1.20 1.20 1.21 1.21 2.83 2.85 2.81 2.87 2.86 0.01 -0.03 0.03 0.04 TX 6.92 6.76 6.68 7.04 7.09 7.37 7.82 7.82 7.87 7.87 7.01 6.98 6.91 7.20 7.24 -0.03 -0.10 0.20 0.26 UT 0.59 0.64 0.64 0.64 0.63 2.36 2.51 2.51 2.50 2.50 0.94 1.01 1.01 1.01 1.00 0.07 0.07 0.07 0.06 VA 2.72 2.82 2.75 2.77 2.74 1.08 1.15 1.15 1.15 1.15 2.39 2.49 2.43 2.44 2.42 0.09 0.04 0.05 0.04 VI* 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 VT 0.58 0.60 0.58 0.58 0.58 0.27 0.29 0.29 0.28 0.28 0.52 0.54 0.52 0.52 0.52 0.02 0.01 0.00 0.00 WA 1.89 1.92 1.91 1.88 1.91 1.87 2.01 2.01 2.00 2.00 1.89 1.94 1.93 1.91 1.92 0.05 0.04 0.02 0.04 *AS=American Samoa, DC=Washington DC, GU=Guam, MP=Northern Marina Islands, PR= Puerto Rico, VI=Virgin Islands. Other initials represent states. TABLE C1 Percent Change in 5311 Apportionment (20102020) (Continued) Scen- 50% 50% 75% 75% 50% 50% 75% 75% 50% 50% 75% 75% 50% 50% 75% 75% ario mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi mi 0 mi WI 2.83 2.77 2.79 2.72 2.75 1.53 1.64 1.64 1.64 1.64 2.57 2.55 2.56 2.50 2.52 -0.02 -0.01 -0.07 -0.04 WV 1.39 1.46 1.42 1.41 1.40 0.69 0.74 0.74 0.73 0.73 1.25 1.31 1.28 1.27 1.27 0.06 0.03 0.02 0.02 WY 0.48 0.53 0.52 0.51 0.50 2.82 3.03 3.03 3.01 3.01 0.95 1.03 1.02 1.01 1.00 0.08 0.07 0.06 0.06 *AS=American Samoa, DC=Washington DC, GU=Guam, MP=Northern Marina Islands, PR= Puerto Rico, VI=Virgin Islands. Other initials represent states. 126 TABLE C2 Comparison of Current (FY19) and Future 5311 Funding by State (Assumes Same FTA Data Values) 127 5311 Appropriation 5311 Forecast Forecast FY19 Appropriation % Diff (Forecast FY19) State FY19 75% 0 mi 50% mi 75% 0 mi AK 9,065,348 8,350,472 8,349,117 -714,876 AL 15,608,151 11,693,306 11,022,482 -3,914,845 AR 12,395,276 10,311,876 10,303,497 -2,083,400 AZ 12,337,581 9,455,187 8,379,374 -2,882,394 CA 27,739,204 22,313,491 21,189,201 -5,425,713 CO 11,798,741 9,565,234 9,252,016 -2,233,507 CT 2,908,207 1,107,710 958,179 -1,800,497 DE 1,666,013 1,435,343 1,013,291 -230,670 FL 15,731,428 10,225,239 9,606,752 -5,506,189 GA 21,190,416 16,275,942 15,914,485 -4,914,474 HI 2,612,904 2,673,114 2,370,647 60,210 IA 12,472,602 10,941,200 10,921,068 -1,531,402 ID 8,154,661 6,403,948 6,305,164 -1,750,713 IL 16,691,323 14,394,258 13,753,930 -2,297,065 IN 15,921,007 12,624,186 12,417,564 -3,296,821 KS 11,501,401 10,230,491 10,223,947 -1,270,910 KY 17,031,406 15,006,099 14,996,023 -2,025,307 LA 11,578,172 8,397,272 7,996,376 -3,180,900 MA 3,586,270 1,481,279 1,259,885 -2,104,991 MD 5,432,142 3,238,043 3,030,024 -2,194,099 ME 7,211,895 5,621,543 5,615,730 -1,590,352 MI 21,177,294 15,994,943 15,432,304 -5,182,351 MN 15,803,225 14,095,286 13,765,361 -1,707,939 MO 17,964,320 15,855,404 15,546,133 -2,108,916 MS 14,353,859 12,605,859 12,363,723 -1,748,000 50% mi -716,231 -4,585,669 -2,091,779 -3,958,207 -6,550,003 -2,546,725 -1,950,028 -652,722 -6,124,676 -5,275,931 -242,257 -1,551,534 -1,849,497 -2,937,393 -3,503,443 -1,277,454 -2,035,383 -3,581,796 -2,326,385 -2,402,118 -1,596,165 -5,744,990 -2,037,864 -2,418,187 -1,990,136 75% 0 mi -8% -25% -17% -23% -20% -19% -62% -14% -35% -23% 2% -12% -21% -14% -21% -11% -12% -27% -59% -40% -22% -24% -11% -12% -12% 50% mi -8% -29% -17% -32% -24% -22% -67% -39% -39% -25% -9% -12% -23% -18% -22% -11% -12% -31% -65% -44% -22% -27% -13% -13% -14% Actual 5311 and 5340 Appropriation FY19 9,427,438 17,799,272 14,000,223 13,678,385 31,257,249 13,042,941 3,372,123 1,941,404 18,257,477 24,524,576 2,972,961 14,097,605 8,967,295 18,863,416 18,324,478 12,765,647 19,346,765 13,158,010 4,185,221 6,317,468 8,063,252 24,084,118 17,918,414 20,285,797 16,215,551 % 5340 in Appr. FY19 4% 12% 11% 10% 11% 10% 14% 14% 14% 14% 12% 12% 9% 12% 13% 10% 12% 12% 14% 14% 11% 12% 12% 11% 11% TABLE C2 Comparison of Current (FY19) and Future 5311 Funding by State (Assumes Same FTA Data Values) (Continued) 128 State MT NC ND NE NH NJ NM NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV WY TOTAL 5311 Appropriation FY19 10,909,730 26,494,927 5,478,259 8,080,019 3,924,278 3,785,847 11,102,161 6,801,059 20,580,008 23,156,208 14,989,643 12,577,795 21,705,974 545,665 12,662,470 6,808,536 18,451,526 40,448,609 6,681,416 14,615,402 3,960,275 12,743,571 15,825,451 7,863,875 6,881,466 629,007,016 5311 Forecast 75% 0 mi 8,661,335 19,894,692 4,270,209 6,996,160 3,087,179 1,903,974 9,320,795 5,898,607 15,273,515 18,681,172 13,591,762 10,503,786 14,650,572 97,744 8,179,026 6,008,575 15,223,310 37,862,478 5,710,939 10,515,405 3,724,090 9,903,178 12,966,172 5,922,254 5,771,587 504,915,243 50% mi 7,783,432 18,209,622 4,094,386 6,923,017 2,579,186 1,899,890 8,201,067 5,893,845 14,754,511 18,274,971 12,189,552 10,463,056 14,350,343 90,953 7,150,048 5,882,646 14,248,797 34,983,058 5,595,263 10,497,043 3,585,952 9,484,247 12,387,348 5,914,288 5,771,338 483,194,130 Forecast FY19 Appropriation 75% 0 mi -2,248,395 -6,600,235 -1,208,050 -1,083,859 -837,099 -1,881,873 -1,781,366 -902,452 -5,306,493 -4,475,036 -1,397,881 -2,074,009 -7,055,402 -447,921 -4,483,444 -799,961 -3,228,216 -2,586,131 -970,477 -4,099,997 -236,185 -2,840,393 -2,859,279 -1,941,621 -1,109,879 -124,091,773 FY19 -3,126,298 -8,285,305 -1,383,873 -1,157,002 -1,345,092 -1,885,957 -2,901,094 -907,214 -5,825,497 -4,881,237 -2,800,091 -2,114,739 -7,355,631 -454,712 -5,512,422 -925,890 -4,202,729 -5,465,551 -1,086,153 -4,118,359 -374,323 -3,259,324 -3,438,103 -1,949,587 -1,110,128 -145,812,886 % Diff (Forecast FY19) FY19 -21% -25% -22% -13% -21% -50% -16% -13% -26% -19% -9% -16% -33% -82% -35% -12% -17% -6% -15% -28% -6% -22% -18% -25% -16% -20% 50% mi -29% -31% -25% -14% -34% -50% -26% -13% -28% -21% -19% -17% -34% -83% -44% -14% -23% -14% -16% -28% -9% -26% -22% -25% -16% -23% Actual 5311 and 5340 Appropriation FY19 11,618,598 30,794,235 5,908,127 8,879,328 4,551,832 4,384,421 11,944,762 7,116,819 23,503,142 26,668,523 16,898,264 14,025,727 24,945,192 632,431 14,739,811 7,373,772 21,241,675 47,163,642 7,247,225 16,935,907 4,403,771 14,510,893 18,067,921 8,873,281 7,262,958 712,559,343 % 5340 in Appr. FY19 6% 14% 7% 9% 14% 14% 7% 4% 12% 13% 11% 10% 13% 14% 14% 8% 13% 14% 8% 14% 10% 12% 12% 11% 5% 12% TABLE C3 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of 1M or More (Assumes Same FTA Data Values) 129 5307 Appropriation 5307 Forecast 5307 Forecast FY19 Appropriation % Diff (Forecast FY19) Urban Area FY19 75% 0 mi 50% mi 75% 0 mi 50% mi 75% 0 mi 50% mi Atlanta, GA 69,110,223 74,241,598 74,656,878 5,131,375 5,546,655 7.4% 8.0% Austin, TX 30,638,400 30,063,025 30,206,868 -575,375 -431,532 -1.9% -1.4% Baltimore, MD 52,915,892 50,356,876 50,485,494 -2,559,016 -2,430,398 -4.8% -4.6% Boston, MANHRI 123,060,958 94,020,265 94,108,958 -29,040,693 -28,952,000 -23.6% -23.5% BridgeportStamford, CTNY* 19,460,666 27,592,493 8,131,827 0.0% 41.8% Charlotte, NCSC 18,036,007 27,345,556 27,483,922 9,309,549 9,447,915 51.6% 52.4% Chicago, ILIN 253,006,909 194,318,318 194,495,177 -58,688,591 -58,511,732 -23.2% -23.1% Cincinnati, OHKYIN 18,306,261 27,162,793 27,229,400 8,856,532 8,923,139 48.4% 48.7% Cleveland, OH 26,801,584 32,039,594 32,087,579 5,238,010 5,285,995 19.5% 19.7% Columbus, OH 16,872,226 15,360,528 15,391,986 -1,511,698 -1,480,240 -9.0% -8.8% DallasFort WorthArlington, TX 74,345,825 67,579,177 68,353,428 -6,766,648 -5,992,397 -9.1% -8.1% DenverAurora, CO 55,762,916 54,788,647 54,842,789 -974,269 -920,127 -1.7% -1.7% Detroit, MI 41,029,661 45,322,534 45,387,832 4,292,873 4,358,171 10.5% 10.6% Houston, TX 76,957,061 79,430,407 79,758,027 2,473,346 2,800,966 3.2% 3.6% Indianapolis, IN 13,495,503 12,692,332 12,740,172 -803,171 -755,331 -6.0% -5.6% Jacksonville, FL 13,310,581 22,789,258 22,828,222 9,478,677 9,517,641 71.2% 71.5% Kansas City, MOKS 16,305,454 24,604,565 24,724,905 8,299,111 8,419,451 50.9% 51.6% Las VegasHenderson, NV 34,802,410 38,018,933 38,036,888 3,216,523 3,234,478 9.2% 9.3% Los AngelesLong BeachAnaheim, CA 300,863,182 220,990,487 220,997,230 -79,872,695 -79,865,952 -26.5% -26.5% Louisville/Jefferson County, KYIN 13,896,577 11,487,326 11,554,543 -2,409,251 -2,342,034 -17.3% -16.9% Memphis, TNMSAR 10,318,625 19,211,043 19,435,612 8,892,418 9,116,987 86.2% 88.4% Miami, FL 108,608,246 94,806,051 94,900,212 -13,802,195 -13,708,034 -12.7% -12.6% Milwaukee, WI 20,116,961 28,128,828 28,376,709 8,011,867 8,259,748 39.8% 41.1% MinneapolisSt. Paul, MNWI 54,842,322 55,899,637 56,232,339 1,057,315 1,390,017 1.9% 2.5% * BridgeportStamford, CTNY is part of the large urban area with a population of less than 1M and is included in Table C4. TABLE C3 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of 1M or More (Assumes Same FTA Data Values) (Continued) 130 Urban Area Nashville-Davidson, TN New Orleans, LA New YorkNewark, NYNJCT Oklahoma City, OK Orlando, FL Philadelphia, PANJDEMD PhoenixMesa, AZ Pittsburgh, PA Portland, ORWA Providence, RIMA Raleigh, NC Richmond, VA RiversideSan Bernardino, CA Sacramento, CA Salt Lake CityWest Valley City, UT San Antonio, TX San Diego, CA San FranciscoOakland, CA San Jose, CA Seattle, WA St. Louis, MOIL TampaSt. Petersburg, FL Virginia Beach, VA Washington, DCVAMD TOTAL LU W/ POP OF 1M+ 5307 Appropriation FY19 22,542,344 14,657,182 808,359,426 7,918,076 30,297,795 140,017,442 52,882,111 32,709,705 45,156,632 23,444,475 11,850,766 11,719,577 32,436,257 25,329,918 26,646,246 29,951,497 67,595,517 138,804,053 36,480,904 105,604,370 34,351,735 28,698,906 17,733,813 171,679,291 3,379,732,488 5307 Forecast 75% 0 mi 21,205,512 20,232,773 650,372,512 16,813,880 29,742,938 110,126,197 55,844,024 38,331,171 45,127,933 22,450,932 10,532,590 9,933,428 30,070,728 28,675,389 28,431,871 28,091,181 59,923,547 106,851,884 32,405,860 91,718,804 38,682,777 36,750,747 25,221,197 143,668,830 3,001,864,482 50% mi 21,280,647 20,331,264 672,060,065 16,900,381 30,077,503 116,233,633 56,054,148 38,382,666 45,176,348 22,492,535 10,958,225 9,979,666 30,079,432 28,857,631 28,436,252 28,435,769 59,943,903 106,970,578 32,407,510 91,890,764 38,810,743 36,840,773 25,237,979 143,735,736 3,063,481,815 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi -1,336,832 -1,261,697 5,575,591 5,674,082 -157,986,914 -136,299,361 8,895,804 8,982,305 -554,857 -220,292 -29,891,245 -23,783,809 2,961,913 3,172,037 5,621,466 5,672,961 -28,699 19,716 -993,543 -951,940 -1,318,176 -892,541 -1,786,149 -1,739,911 -2,365,529 -2,356,825 3,345,471 3,527,713 1,785,625 1,790,006 -1,860,316 -1,515,728 -7,671,970 -7,651,614 -31,952,169 -31,833,475 -4,075,044 -4,073,394 -13,885,566 -13,713,606 4,331,042 4,459,008 8,051,841 8,141,867 7,487,384 7,504,166 -28,010,461 -27,943,555 -358,407,340 -316,250,673 % Diff (Forecast FY19) 75% 0 mi 50% mi -5.9% -5.6% 38.0% 38.7% -19.5% -16.9% 112.3% 113.4% -1.8% -0.7% -21.3% -17.0% 5.6% 6.0% 17.2% 17.3% -0.1% 0.0% -4.2% -4.1% -11.1% -7.5% -15.2% -14.8% -7.3% -7.3% 13.2% 13.9% 6.7% 6.7% -6.2% -5.1% -11.3% -11.3% -23.0% -22.9% -11.2% -11.2% -13.1% -13.0% 12.6% 13.0% 28.1% 28.4% 42.2% 42.3% -16.3% -16.3% -10.6% -9.4% TABLE C4 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of (200K1M] (Assumes Same FTA Data Values) 131 5307 Appropriation 5307 Forecast 5307 Forecast % Diff FY19 Appropriation (Forecast FY19) Urban Area FY19 75% 0 mi 50% mi 75% 0 mi 50% mi 75% 0 mi 50% mi AberdeenBel Air SouthBel Air North, MD 1,535,872 1,321,535 1,333,438 -214,337 -202,434 -14% -13% Akron, OH 7,386,461 6,623,203 6,652,114 -763,258 -734,347 -10% -10% AlbanySchenectady, NY 10,337,470 9,402,159 9,454,494 -935,311 -882,976 -9% -9% Albuquerque, NM 18,455,944 19,138,597 19,141,634 682,653 685,690 4% 4% Allentown, PANJ 7,669,403 6,780,035 6,815,569 -889,368 -853,834 -12% -11% Amarillo, TX 3,200,793 932,132 953,132 -2,268,661 -2,247,661 -71% -70% Anchorage, AK 15,581,375 14,037,442 14,037,442 -1,543,933 -1,543,933 -10% -10% Ann Arbor, MI 6,997,149 6,525,796 6,539,440 -471,353 -457,709 -7% -7% Antioch, CA 6,265,137 15,296,864 15,299,240 9,031,727 9,034,103 144% 144% Appleton, WI 2,381,315 2,042,761 2,050,809 -338,554 -330,506 -14% -14% Asheville, NC 2,682,276 2,791,047 2,862,825 108,771 180,549 4% 7% Atlantic City, NJ 9,735,559 18,276,519 18,278,544 8,540,960 8,542,985 88% 88% Augusta-Richmond County, GASC 2,635,928 2,439,800 2,494,396 -196,128 -141,532 -7% -5% AvondaleGoodyear, AZ 3,032,209 1,035,771 1,153,071 -1,996,438 -1,879,138 -66% -62% Bakersfield, CA 7,583,180 5,583,293 5,586,045 -1,999,887 -1,997,135 -26% -26% Barnstable Town, MA 6,259,733 15,304,666 15,315,068 9,044,933 9,055,335 144% 145% Baton Rouge, LA 5,640,479 5,154,311 5,320,016 -486,168 -320,463 -9% -6% Birmingham, AL 6,606,267 6,109,967 6,190,789 -496,300 -415,478 -8% -6% Boise City, ID 4,242,303 3,527,851 3,534,463 -714,452 -707,840 -17% -17% Bonita Springs, FL 2,885,842 2,871,302 2,910,384 -14,540 24,542 -1% 1% Bremerton, WA 2,499,412 892,454 938,053 -1,606,958 -1,561,359 -64% -62% BridgeportStamford, CTNY* 19,460,666 21,593,430 2,132,764 -19,460,666 11% -100% Brownsville, TX 2,415,759 1,907,824 1,922,504 -507,935 -493,255 -21% -20% Buffalo, NY 13,497,649 21,142,159 21,327,960 7,644,510 7,830,311 57% 58% Canton, OH 3,689,186 3,371,504 3,379,904 -317,682 -309,282 -9% -8% *BridgeportStamford, CTNY grows to a large UA with population of more than 1M and is on Table C3 for the 50% scenario. TABLE C4 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of (200K1M] (Assumes Same FTA Data Values) (Continued) 132 5307 Appropriation 5307 Forecast 5307 Forecast FY19 Appropriation Urban Area FY19 75% 0 mi 50% mi 75% 0 mi 50% mi Cape Coral, FL 5,286,175 5,277,389 5,335,572 -8,786 49,397 CharlestonNorth Charleston, SC 5,445,476 4,979,962 5,034,276 -465,514 -411,200 Chattanooga, TNGA 3,596,876 13,626,796 13,662,160 10,029,920 10,065,284 College StationBryan, TX* 2,862,891 856,498 -2,862,891 -2,006,393 Colorado Springs, CO 6,522,379 5,088,388 5,111,031 -1,433,991 -1,411,348 Columbia, SC 4,595,822 4,302,655 4,432,367 -293,167 -163,455 Columbus, GAAL 3,778,265 3,582,596 3,611,466 -195,669 -166,799 Concord, CA 22,287,174 28,035,480 28,037,601 5,748,306 5,750,427 Concord, NC 1,881,707 1,804,150 1,857,408 -77,557 -24,299 ConroeThe Woodlands, TX 3,047,072 3,005,719 3,246,531 -41,353 199,459 Corpus Christi, TX 5,369,050 4,662,692 4,696,060 -706,358 -672,990 Davenport, IAIL 4,125,771 13,766,357 13,792,142 9,640,586 9,666,371 Dayton, OH 15,382,035 19,563,143 19,600,374 4,181,108 4,218,339 Deltona, FL* 2,648,804 982,025 -2,648,804 -1,666,779 DentonLewisville, TX 6,035,665 13,809,849 13,855,547 7,774,184 7,819,882 Des Moines, IA 6,088,524 5,426,282 5,449,633 -662,242 -638,891 Durham, NC 7,568,603 7,174,606 7,220,756 -393,997 -347,847 El Paso, TXNM 13,168,279 10,793,133 10,823,168 -2,375,146 -2,345,111 Eugene, OR 7,798,477 16,656,052 16,663,276 8,857,575 8,864,799 Evansville, INKY 2,426,679 2,121,806 2,137,787 -304,873 -288,892 Fargo, NDMN 2,887,390 875,950 882,844 -2,011,440 -2,004,546 Fayetteville, NC 3,058,932 2,862,576 2,911,073 -196,356 -147,859 FayettevilleSpringdaleRogers, ARMO 2,526,869 2,397,062 2,426,134 -129,807 -100,735 Flint, MI 6,258,934 5,855,901 5,874,481 -403,033 -384,453 Fort Collins, CO 4,159,303 13,560,462 13,579,236 9,401,159 9,419,933 *College StationBryan, TX, and Deltona, FL, do not grow to a large UA under the 75% scenario. % Diff (Forecast FY19) 75% 0 mi 50% mi 0% 1% -9% -8% 279% 280% -100% -70% -22% -22% -6% -4% -5% -4% 26% 26% -4% -1% -1% 7% -13% -13% 234% 234% 27% 27% -100% -63% 129% 130% -11% -10% -5% -5% -18% -18% 114% 114% -13% -12% -70% -69% -6% -5% -5% -4% -6% -6% 226% 226% TABLE C4 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of (200K1M] (Assumes Same FTA Data Values) ( Continued) 133 5307 Appropriation 5307 Forecast Urban Area FY19 75% 0 mi Fort Walton BeachNavarreWright, FL 2,511,235 930,829 Fort Wayne, IN 2,944,921 2,548,679 Fresno, CA 10,278,598 7,739,378 Gainesville, FL* 2,942,624 Grand Rapids, MI 9,373,351 17,983,305 Green Bay, WI 2,010,191 1,713,006 Greensboro, NC 4,445,507 4,149,019 Greenville, SC 2,947,640 2,851,373 Gulfport, MS 2,064,468 2,075,197 Hagerstown, MDWVPA 2,279,744 905,734 Harrisburg, PA 6,091,045 14,786,807 Hartford, CT 15,498,285 23,826,768 Hickory, NC 1,454,305 1,504,249 Huntington, WVKYOH 2,124,765 1,986,025 Huntsville, AL 2,086,788 1,993,552 IndioCathedral City, CA 4,605,488 3,882,733 Jackson, MS 2,461,588 2,181,069 Kalamazoo, MI 2,992,366 2,789,108 KennewickPasco, WA 6,250,085 5,999,162 Killeen, TX 2,102,789 1,680,674 Kissimmee, FL 5,661,493 5,671,072 Knoxville, TN 5,787,419 5,621,880 Lafayette, LA 2,145,059 2,016,425 Lakeland, FL 2,122,613 1,947,544 Lancaster, PA 7,958,742 14,886,578 * Gainesville, FL does not grow to a large UA under the 75% scenario. 50% mi 964,097 2,564,431 7,750,691 899,090 18,068,033 1,721,614 4,196,632 3,077,950 2,092,639 939,232 14,803,824 23,843,016 1,566,430 2,019,179 2,179,384 3,959,426 2,312,013 2,848,910 6,017,525 1,697,471 5,683,147 5,715,765 2,124,783 1,966,923 14,929,227 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi -1,580,406 -1,547,138 -396,242 -380,490 -2,539,220 -2,527,907 -2,942,624 -2,043,534 8,609,954 8,694,682 -297,185 -288,577 -296,488 -248,875 -96,267 130,310 10,729 28,171 -1,374,010 -1,340,512 8,695,762 8,712,779 8,328,483 8,344,731 49,944 112,125 -138,740 -105,586 -93,236 92,596 -722,755 -646,062 -280,519 -149,575 -203,258 -143,456 -250,923 -232,560 -422,115 -405,318 9,579 21,654 -165,539 -71,654 -128,634 -20,276 -175,069 -155,690 6,927,836 6,970,485 % Diff (Forecast FY19) 75% 0 mi 50% mi -63% -62% -13% -13% -25% -25% -100% -69% 92% 93% -15% -14% -7% -6% -3% 4% 1% 1% -60% -59% 143% 143% 54% 54% 3% 8% -7% -5% -4% 4% -16% -14% -11% -6% -7% -5% -4% -4% -20% -19% 0% 0% -3% -1% -6% -1% -8% -7% 87% 88% TABLE C4 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of (200K1M] (Assumes Same FTA Data Values) ( Continued) 134 5307 Appropriation 5307 Forecast Urban Area FY19 75% 0 mi 50% mi LancasterPalmdale, CA 8,121,638 14,618,628 14,622,560 Lansing, MI 6,003,951 5,486,559 5,498,984 Laredo, TX 3,211,709 2,381,783 2,384,372 LexingtonFayette, KY 4,390,200 3,462,818 3,472,564 Lincoln, NE 3,106,523 2,407,355 2,410,663 Little Rock, AR 4,447,622 14,206,395 14,220,205 LorainElyria, OH* 2,573,950 1,038,559 Lubbock, TX 3,054,915 2,604,721 2,640,825 Madison, WI 7,638,764 16,417,533 16,439,058 Manchester, NH* 2,201,835 882,516 McAllen, TX 6,134,770 5,221,777 5,269,842 McKinney, TX 2,602,279 960,898 1,052,489 Mission ViejoLake ForestSan Clemente, CA 9,126,615 15,278,651 15,279,370 Mobile, AL 2,809,173 2,521,428 2,582,319 Modesto, CA 5,038,276 3,567,039 3,575,950 Montgomery, AL 2,381,200 2,083,267 2,113,754 MurrietaTemeculaMenifee, CA 4,518,327 3,724,707 3,733,529 Myrtle BeachSocastee, SCNC 1,430,712 1,661,993 1,703,104 Nashua, NHMA 1,478,591 1,342,650 1,392,132 New Haven, CT 16,804,976 20,307,088 20,316,893 NorwichNew London, CTRI 1,717,657 1,547,618 1,556,813 OgdenLayton, UT 12,073,034 18,395,244 18,410,201 Omaha, NEIA 7,706,373 6,018,148 6,072,061 Oxnard, CA 9,297,121 15,430,826 15,436,316 * LorainElyria, OH and Manchester, NH do not grow to a large UA under the 75% scenario. 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 6,496,990 6,500,922 -517,392 -504,967 -829,926 -827,337 -927,382 -917,636 -699,168 -695,860 9,758,773 9,772,583 -2,573,950 -1,535,391 -450,194 -414,090 8,778,769 8,800,294 -2,201,835 -1,319,319 -912,993 -864,928 -1,641,381 -1,549,790 6,152,036 6,152,755 -287,745 -226,854 -1,471,237 -1,462,326 -297,933 -267,446 -793,620 -784,798 231,281 272,392 -135,941 -86,459 3,502,112 3,511,917 -170,039 -160,844 6,322,210 6,337,167 -1,688,225 -1,634,312 6,133,705 6,139,195 % Diff (Forecast FY19) 75% 0 mi 50% mi 80% 80% -9% -8% -26% -26% -21% -21% -23% -22% 219% 220% -100% -60% -15% -14% 115% 115% -100% -60% -15% -14% -63% -60% 67% 67% -10% -8% -29% -29% -13% -11% -18% -17% 16% 19% -9% -6% 21% 21% -10% -9% 52% 52% -22% -21% 66% 66% TABLE C4 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of (200K1M] (Assumes Same FTA Data Values) ( Continued) 135 5307 Appropriation 5307 Forecast Urban Area FY19 75% 0 mi Palm BayMelbourne, FL 5,040,688 4,561,829 Palm CoastDaytona BeachPort Orange, FL 4,519,302 4,269,598 Pensacola, FLAL 3,170,984 2,984,458 Peoria, IL 3,328,725 2,949,097 Port St. Lucie, FL 3,002,069 2,745,739 Portland, ME 10,294,729 13,044,560 PoughkeepsieNewburgh, NYNJ 20,138,533 28,686,869 ProvoOrem, UT 8,138,726 15,348,545 Reading, PA 3,410,942 2,789,058 Reno, NVCA 6,703,847 16,192,188 Roanoke, VA 2,593,590 2,399,784 Rochester, NY 8,690,291 7,357,388 Rockford, IL 2,913,732 2,461,773 Round Lake BeachMcHenryGrayslake, ILWI 4,978,522 12,817,027 Salem, OR 5,942,971 5,263,970 Salinas, CA* 3,913,695 Santa Barbara, CA 3,881,154 872,730 Santa Clarita, CA 4,790,953 13,339,407 Santa Rosa, CA 3,991,685 3,082,162 SarasotaBradenton, FL 7,572,223 7,255,072 Savannah, GA 3,463,471 13,460,447 Scranton, PA 4,490,894 3,792,444 Shreveport, LA 3,575,017 3,301,225 Sioux Falls, SD* 2,517,459 * Salinas, CA, and Sioux Falls, SD, do not grow to a large UA under the 75% scenario. 50% mi 4,600,623 4,330,560 3,051,926 2,978,791 2,764,761 13,081,079 28,791,525 15,403,931 2,800,161 16,216,630 2,421,640 7,387,445 2,475,254 12,855,393 5,299,172 843,046 878,564 13,349,189 3,125,185 7,313,893 13,513,340 3,801,500 3,318,409 825,798 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi -478,859 -440,065 -249,704 -188,742 -186,526 -119,058 -379,628 -349,934 -256,330 -237,308 2,749,831 2,786,350 8,548,336 8,652,992 7,209,819 7,265,205 -621,884 -610,781 9,488,341 9,512,783 -193,806 -171,950 -1,332,903 -1,302,846 -451,959 -438,478 7,838,505 7,876,871 -679,001 -643,799 -3,913,695 -3,070,649 -3,008,424 -3,002,590 8,548,454 8,558,236 -909,523 -866,500 -317,151 -258,330 9,996,976 10,049,869 -698,450 -689,394 -273,792 -256,608 -2,517,459 -1,691,661 % Diff (Forecast FY19) 75% 0 mi 50% mi -9% -9% -6% -4% -6% -4% -11% -11% -9% -8% 27% 27% 42% 43% 89% 89% -18% -18% 142% 142% -7% -7% -15% -15% -16% -15% 157% 158% -11% -11% -100% -78% -78% -77% 178% 179% -23% -22% -4% -3% 289% 290% -16% -15% -8% -7% -100% -67% TABLE C4 Comparison of Current (FY19) and Future 5307 Funding for Large Urban Areas with Population of (200K1M] (Assumes Same FTA Data Values) ( Continued) 136 5307 Appropriation 5307 Forecast Urban Area FY19 75% 0 mi South Bend, INMI 3,937,268 12,698,696 Spartanburg, SC 2,047,572 905,701 Spokane, WA 7,695,565 6,989,887 Springfield, MACT 8,587,556 7,830,670 Springfield, MO 2,397,699 2,088,681 Stockton, CA 7,320,446 13,600,594 Syracuse, NY 5,196,752 4,462,876 Tallahassee, FL 2,928,012 2,562,624 Thousand Oaks, CA 2,742,471 11,838,986 Toledo, OHMI 6,377,295 15,680,925 Trenton, NJ 9,662,607 16,073,903 Tucson, AZ 14,587,852 22,918,141 Tulsa, OK 6,189,272 5,397,957 Urban Honolulu, HI 27,322,881 33,452,311 VictorvilleHesperia, CA 7,793,840 7,222,852 Visalia, CA 5,558,718 4,842,457 Waco, TX* 2,561,739 Wichita, KS 4,346,546 3,477,161 Wilmington, NC 2,149,241 2,051,296 Winston-Salem, NC 4,560,128 4,415,389 Winter Haven, FL 1,534,845 1,690,699 Worcester, MACT 5,531,604 14,279,461 York, PA 2,942,865 2,678,208 Youngstown, OHPA 3,804,442 3,373,242 LU (200K1M] TOTAL 839,337,257 1,044,156,803 * Waco, TX, does not grow to a large UA under the 75% scenario. 50% mi 12,711,264 982,465 6,996,157 7,851,543 2,141,076 13,614,281 4,471,508 2,617,844 11,847,821 15,716,214 16,075,105 23,108,629 5,599,606 33,521,928 7,235,981 4,852,742 883,225 3,498,009 2,066,117 4,472,126 1,546,518 14,389,408 2,685,546 3,395,088 1,035,325,304 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 8,761,428 8,773,996 -1,141,871 -1,065,107 -705,678 -699,408 -756,886 -736,013 -309,018 -256,623 6,280,148 6,293,835 -733,876 -725,244 -365,388 -310,168 9,096,515 9,105,350 9,303,630 9,338,919 6,411,296 6,412,498 8,330,289 8,520,777 -791,315 -589,666 6,129,430 6,199,047 -570,988 -557,859 -716,261 -705,976 -2,561,739 -1,678,514 -869,385 -848,537 -97,945 -83,124 -144,739 -88,002 155,854 11,673 8,747,857 8,857,804 -264,657 -257,319 -431,200 -409,354 204,819,546 195,988,047 % Diff (Forecast FY19) 75% 0 mi 50% mi 223% 223% -56% -52% -9% -9% -9% -9% -13% -11% 86% 86% -14% -14% -12% -11% 332% 332% 146% 146% 66% 66% 57% 58% -13% -10% 22% 23% -7% -7% -13% -13% -100% -66% -20% -20% -5% -4% -3% -2% 10% 1% 158% 160% -9% -9% -11% -11% 24% 23% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) 137 Urban Area Abilene, TX Albany, GA Albany, OR Alexandria, LA Alton, ILMO Altoona, PA Ames, IA Anderson, IN Anderson, SC AnnistonOxford, AL Arroyo GrandeGrover Beach, CA AthensClarke County, GA Auburn, AL Bangor, ME Battle Creek, MI Bay City, MI BeaufortPort Royal, SC Beaumont, TX Beckley, WV Belgrade, MT Bellingham, WA Beloit, WIIL Bend, OR Benton HarborSt. JosephFair Plain, MI Billings, MT 5307 Appropriation FY19 1,665,452 1,248,067 920,160 1,048,265 1,104,778 1,217,359 1,041,553 1,112,092 874,749 902,288 795,891 1,674,247 1,024,526 797,878 1,031,837 989,937 2,009,112 744,520 - 1,849,599 920,128 1,260,055 769,591 1,739,633 5307 Forecast 75% 0 mi 2,002,373 1,395,280 1,088,213 1,327,580 1,174,489 1,431,188 2,476,530 1,253,255 1,120,163 1,338,597 1,749,190 2,936,877 1,495,046 1,436,393 1,448,165 1,373,021 771,169 2,420,196 968,025 - 3,254,092 1,016,799 1,691,084 1,143,704 2,197,495 50% mi 1,612,746 1,376,371 887,885 1,291,127 1,077,746 1,436,256 2,194,173 1,215,565 1,098,923 1,340,133 1,900,219 2,854,876 1,335,054 1,402,976 1,372,806 1,312,552 781,230 2,140,685 1,005,561 904,951 3,038,700 914,338 1,541,788 1,118,817 2,226,323 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 336,921 -52,706 147,213 128,304 168,053 -32,275 279,315 242,862 69,711 -27,032 213,829 218,897 1,434,977 1,152,620 141,163 103,473 245,414 224,174 436,309 437,845 953,299 1,104,328 1,262,630 1,180,629 470,520 310,528 638,515 605,098 416,328 340,969 383,084 322,615 771,169 781,230 411,084 131,573 223,505 261,041 - 904,951 1,404,493 1,189,101 96,671 -5,790 431,029 281,733 374,113 349,226 457,862 486,690 % Diff (Forecast FY19) 75% 0 mi 50% mi 20% -3% 12% 10% 18% -4% 27% 23% 6% -2% 18% 18% 138% 111% 13% 9% 28% 26% 48% 49% 120% 139% 75% 71% 46% 30% 80% 76% 40% 33% 39% 33% - - 20% 7% 30% 35% - - 76% 64% 11% -1% 34% 22% 49% 45% 26% 28% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 138 Urban Area Binghamton, NYPA Bismarck, ND Blacksburg, VA Bloomington, IN BloomingtonNormal, IL BloomsburgBerwick, PA Boulder, CO Bowling Green, KY Bozeman, MT BristolBristol, TNVA Brunswick, GA Bullhead City, AZNV Burlington, NC Burlington, VT Camarillo, CA Cape Girardeau, MOIL Carbondale, IL Carson City, NV Cartersville, GA Casa Grande, AZ Casper, WY Cedar Rapids, IA Chambersburg, PA Champaign, IL Charleston, WV 5307 Appropriation FY19 2,410,112 1,209,862 1,254,458 1,807,579 2,256,103 726,543 2,294,257 1,108,029 815,360 631,488 1,520,203 1,494,473 1,328,129 699,137 885,742 963,289 604,401 829,314 966,591 2,647,833 631,986 2,734,066 2,018,958 5307 Forecast 75% 0 mi 3,116,354 2,342,107 2,532,362 2,865,973 3,479,668 844,916 4,331,551 1,482,285 941,637 1,017,466 798,074 889,465 2,548,547 2,875,349 1,530,011 1,366,826 1,542,329 1,594,975 1,322,543 1,421,137 1,432,999 3,158,375 795,011 4,462,048 2,918,238 50% mi 3,124,063 2,082,729 2,415,878 2,786,991 2,922,218 850,377 3,659,501 1,331,358 1,029,659 763,391 - 2,439,512 2,938,642 1,184,140 1,280,481 1,596,024 1,345,878 1,297,097 1,120,871 1,352,022 2,823,984 800,926 3,710,601 2,928,406 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 706,242 713,951 1,132,245 872,867 1,277,904 1,161,420 1,058,394 979,412 1,223,565 666,115 118,373 123,834 2,037,294 1,365,244 374,256 223,329 941,637 - 202,106 214,299 166,586 131,903 889,465 - 1,028,344 919,309 1,380,876 1,444,169 201,882 -143,989 667,689 581,344 656,587 710,282 631,686 382,589 718,142 692,696 591,823 291,557 466,408 385,431 510,542 176,151 163,025 168,940 1,727,982 976,535 899,280 909,448 % Diff (Forecast FY19) 75% 0 mi 50% mi 29% 30% 94% 72% 102% 93% 59% 54% 54% 30% 16% 17% 89% 60% 34% 20% 25% 26% 26% 21% 68% 92% 15% 96% 74% 66% 119% 71% 48% 19% 26% 63% 45% 60% 97% -11% 83% 80% 40% 115% 35% 40% 7% 27% 36% 45% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 139 Urban Area Charlottesville, VA Cheyenne, WY Chico, CA Clarksville, TNKY Cleveland, TN Clovis, NM Coeur d'Alene, ID College StationBryan, TX Columbia, MO Columbus, IN Conway, AR Cookeville, TN Corvallis, OR Cumberland, MDWVPA Dalton, GA Danbury, CTNY Danville, IL DaphneFairhope, AL Davis, CA Decatur, AL Decatur, IL DeKalb, IL Delano, CA Deltona, FL Dothan, AL 5307 Appropriation FY19 1,571,856 1,090,927 1,771,536 2,061,780 822,667 - 1,468,154 - 1,875,838 807,416 898,185 - 1,147,927 695,960 1,018,323 7,490,579 643,255 1,890,510 862,013 1,259,222 1,174,401 1,487,583 858,971 5307 Forecast 75% 0 mi 2,766,730 1,359,151 2,603,808 2,965,336 1,102,699 2,013,131 3,603,455 2,974,042 1,083,129 1,239,398 2,656,660 837,200 1,273,110 3,964,666 1,563,667 3,708,201 1,051,579 2,167,351 1,321,376 2,119,809 3,349,611 1,082,671 50% mi 2,538,544 1,321,948 2,482,282 2,807,703 1,075,811 850,203 1,926,915 2,696,602 853,172 1,243,249 691,857 2,393,335 804,880 1,301,521 3,976,964 702,304 1,669,245 3,172,636 1,033,675 1,997,098 1,034,472 1,459,590 1,046,822 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 1,194,874 966,688 268,224 231,021 832,272 710,746 903,556 745,923 280,032 253,144 - 850,203 544,977 458,761 3,603,455 - 1,098,204 820,764 275,713 45,756 341,213 345,064 - 691,857 1,508,733 1,245,408 141,240 108,920 254,787 283,198 -3,525,913 -3,513,615 - 702,304 920,412 1,025,990 1,817,691 1,282,126 189,566 171,662 908,129 737,876 146,975 -139,929 632,226 -27,993 3,349,611 - 223,700 187,851 % Diff (Forecast FY19) 75% 0 mi 50% mi 76% 61% 25% 21% 47% 40% 44% 36% 34% 31% 37% 31% 59% 44% 34% 6% 38% 38% 131% 20% 25% -47% 108% 16% 28% -47% 143% 96% 22% 72% 13% 43% 159% 68% 20% 59% -12% -2% 26% 22% TABLE C5 140 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) Dover, DE Urban Area 5307 Appropriation FY19 1,428,623 5307 Forecast 75% 50% 0 mi mi 2,605,033 2,808,311 5307 Forecast FY19 Appropriation 75% 50% 0 mi mi 1,176,410 1,379,688 % Diff (Forecast FY19) 75% 50% 0 mi mi 82% 97% DoverRochester, NHME 1,070,043 1,556,643 1,546,728 486,600 476,685 45% 45% Dubuque, IAIL 990,867 1,463,386 1,297,011 472,519 306,144 48% 31% Duluth, MNWI 1,663,043 2,933,838 2,813,259 1,270,795 1,150,216 76% 69% Eagle Pass, TX - 1,033,411 1,041,396 1,033,411 1,041,396 East Stroudsburg, PANJ 617,681 1,030,371 1,139,893 412,690 522,212 67% 85% Eau Claire, WI 1,330,484 2,138,446 2,040,338 807,962 709,854 61% 53% El CentroCalexico, CA 2,226,423 4,049,127 3,433,064 1,822,704 1,206,641 82% 54% El Paso de Robles (Paso Robles)Atascadero, CA 957,202 2,224,525 2,127,419 1,267,323 1,170,217 132% 122% ElizabethtownRadcliff, KY 906,953 1,711,034 1,637,616 804,081 730,663 89% 81% Elkhart, INMI 1,926,787 2,381,886 2,211,498 455,099 284,711 24% 15% Elmira, NY 953,991 1,069,893 1,077,351 115,902 123,360 12% 13% Enid, OK - - 832,671 - 832,671 Erie, PA 3,180,265 4,322,413 4,326,632 1,142,148 1,146,367 36% 36% Fairbanks, AK 742,293 1,438,424 1,438,424 696,131 696,131 94% 94% Fairfield, CA 2,591,168 3,157,110 2,896,939 565,942 305,771 22% 12% Farmington, NM 693,871 804,783 1,024,565 110,912 330,694 16% 48% Flagstaff, AZ 1,082,819 2,643,220 2,356,255 1,560,401 1,273,436 144% 118% Florence, AL 946,798 1,171,739 1,129,697 224,941 182,899 24% 19% Florence, SC 1,112,285 1,393,024 1,328,865 280,739 216,580 25% 19% Fond du Lac, WI 782,680 898,283 788,869 115,603 6,189 15% 1% Fort Smith, AROK 1,759,263 2,173,284 2,036,292 414,021 277,029 24% 16% Frederick, MD Fredericksburg, VA Gadsden, AL 1,981,012 2,597,369 2,476,667 616,357 495,655 31% 25% 1,935,763 3,176,171 2,994,054 1,240,408 1,058,291 64% 55% 717,753 868,008 856,781 150,255 139,028 21% 19% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 141 Urban Area Gainesville, FL Gainesville, GA Gastonia, NCSC GilroyMorgan Hill, CA Glens Falls, NY Goldsboro, NC Grand Forks, NDMN Grand Island, NE Grand Junction, CO Grants Pass, OR Great Falls, MT Greeley, CO Greenville, NC Hammond, LA Hanford, CA Hanover, PA Harlingen, TX Harrisonburg, VA Hattiesburg, MS Hazleton, PA Helena, MT Hemet, CA High Point, NC Hilton Head Island, SC Hinesville, GA 5307 Appropriation FY19 1,543,006 2,082,756 1,486,911 853,979 745,924 1,013,968 711,827 1,731,135 746,692 985,301 2,112,891 1,712,133 760,041 1,676,929 925,729 1,961,169 1,016,726 987,991 815,966 3,239,783 2,195,454 763,685 707,151 5307 Forecast 75% 0 mi 4,700,098 2,443,793 2,624,919 2,010,050 1,033,272 947,125 1,553,557 852,305 2,090,600 937,592 1,154,362 2,871,409 2,247,291 1,071,347 3,185,975 1,130,924 2,400,890 2,354,356 1,269,404 955,964 - 3,909,402 2,997,275 1,117,958 854,261 50% mi 2,451,740 2,576,369 1,819,491 1,037,494 987,784 1,255,517 687,158 1,973,347 907,394 1,069,576 2,342,235 2,048,541 1,101,224 2,786,892 1,149,268 2,193,871 2,202,792 1,193,131 954,297 785,535 3,824,241 2,944,131 1,352,809 846,707 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 4,700,098 - 900,787 908,734 542,163 493,613 523,139 332,580 179,293 183,515 201,201 241,860 539,589 241,549 140,478 -24,669 359,465 242,212 190,900 160,702 169,061 84,275 758,518 229,344 535,158 336,408 311,306 341,183 1,509,046 1,109,963 205,195 223,539 439,721 232,702 1,337,630 1,186,066 281,413 205,140 139,998 138,331 - 785,535 669,619 584,458 801,821 748,677 354,273 589,124 147,110 139,556 % Diff (Forecast FY19) 75% 50% 0 mi mi 58% 26% 35% 21% 27% 53% 20% 21% 26% 17% 36% 31% 41% 90% 22% 22% 132% 28% 17% 59% 24% 22% 21% 32% 24% -3% 14% 22% 9% 11% 20% 45% 66% 24% 12% 117% 21% 17% 21% 18% 37% 34% 46% 77% 21% 20% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 142 Urban Area Holland, MI Homosassa SpringsBeverly HillsCitrus Springs, FL Hot Springs, AR Houma, LA Idaho Falls, ID Iowa City, IA Ithaca, NY Jackson, MI Jackson, TN Jacksonville, NC Janesville, WI Jefferson City, MO Johnson City, TN Johnstown, PA Jonesboro, AR Joplin, MO Kahului, HI Kailua (Honolulu County)Kaneohe, HI Kankakee, IL Kenosha, WIIL Kingsport, TNVA Kingston, NY Kokomo, IN La Crosse, WIMN Lady LakeThe Villages, FL 5307 Appropriation FY19 1,345,253 904,946 688,073 1,956,364 1,348,789 1,711,748 833,438 1,210,560 928,574 1,356,865 1,081,797 746,173 1,425,653 972,054 850,127 1,024,709 1,068,731 1,959,242 1,260,767 2,008,018 1,191,701 707,653 865,681 1,472,558 1,466,216 5307 Forecast 75% 0 mi 1,624,191 1,607,854 911,275 2,600,453 1,691,360 3,260,586 2,010,479 1,390,627 1,396,972 1,892,201 1,253,298 902,482 1,822,422 1,571,843 1,275,675 1,386,642 2,634,787 2,431,727 2,191,607 2,535,145 1,515,024 833,711 995,054 2,218,685 2,797,002 50% mi 1,520,153 1,890,934 929,246 2,558,310 1,540,787 2,855,762 2,011,236 1,329,394 1,315,536 1,846,779 1,089,181 821,896 1,818,686 1,584,234 1,288,560 1,282,714 2,831,971 2,033,069 1,950,709 2,278,682 1,518,649 850,801 983,624 2,064,308 2,520,169 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 278,938 174,900 702,908 985,988 223,202 241,173 644,089 601,946 342,571 191,998 1,548,838 1,144,014 1,177,041 1,177,798 180,067 118,834 468,398 386,962 535,336 489,914 171,501 7,384 156,309 75,723 396,769 393,033 599,789 612,180 425,548 438,433 361,933 258,005 1,566,056 1,763,240 472,485 73,827 930,840 689,942 527,127 270,664 323,323 326,948 126,058 143,148 129,373 117,943 746,127 591,750 1,330,786 1,053,953 % Diff (Forecast FY19) 75% 0 mi 50% mi 21% 13% 78% 109% 32% 35% 33% 31% 25% 14% 90% 67% 141% 141% 15% 10% 50% 42% 39% 36% 16% 1% 21% 10% 28% 28% 62% 63% 50% 52% 35% 25% 147% 165% 24% 4% 74% 55% 26% 13% 27% 27% 18% 20% 15% 14% 51% 40% 91% 72% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 143 Urban Area Lafayette, IN LafayetteLouisvilleErie, CO Lake Charles, LA Lake Havasu City, AZ Lake JacksonAngleton, TX LakesKnik-FairviewWasilla, AK Las Cruces, NM Laughlin, NV Lawrence, KS Lawton, OK Lebanon, PA Lee's Summit, MO LeesburgEustisTavares, FL LeominsterFitchburg, MA Lewiston, IDWA Lewiston, ME Lexington ParkCaliforniaChesapeake Ranch Estates, MD Lima, OH Livermore, CA Lodi, CA Logan, UT Lompoc, CA Longmont, CO Longview, TX Longview, WAOR 5307 Appropriation FY19 2,378,379 1,125,731 1,716,888 750,524 1,041,578 1,959,239 1,580,780 1,456,415 1,051,578 1,219,460 1,677,888 1,635,038 735,235 817,434 672,615 935,580 1,530,698 1,578,795 1,463,833 1,262,991 1,781,183 1,207,407 948,910 5307 Forecast 75% 0 mi 4,041,475 1,630,629 2,120,947 1,006,529 1,272,136 669,426 2,302,346 2,679,650 1,615,354 1,295,205 1,525,349 2,341,816 2,385,997 884,848 942,390 1,170,761 1,326,473 1,895,691 1,811,331 2,151,469 1,481,524 2,596,068 1,537,696 1,349,163 50% mi 3,777,130 1,409,908 2,056,309 859,311 1,038,052 722,497 2,254,958 909,568 2,366,237 1,442,122 1,320,400 1,324,386 2,280,916 2,388,730 825,292 874,679 1,183,010 1,283,672 1,690,345 1,900,922 1,919,702 1,410,988 2,611,017 1,451,185 1,303,322 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 1,663,096 1,398,751 504,898 284,177 404,059 339,421 256,005 108,787 230,558 -3,526 669,426 722,497 343,107 295,719 - 909,568 1,098,870 785,457 158,939 -14,293 243,627 268,822 305,889 104,926 663,928 603,028 750,959 753,692 149,613 90,057 124,956 57,245 498,146 510,395 390,893 348,092 364,993 159,647 232,536 322,127 687,636 455,869 218,533 147,997 814,885 829,834 330,289 243,778 400,253 354,412 % Diff (Forecast FY19) 75% 0 mi 50% mi 70% 59% 45% 25% 24% 20% 34% 14% 22% 0% 18% 15% 70% 50% 11% -1% 23% 26% 25% 9% 40% 36% 46% 46% 20% 12% 15% 7% 74% 76% 42% 37% 24% 10% 15% 20% 47% 31% 17% 12% 46% 47% 27% 20% 42% 37% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 144 Urban Area LorainElyria, OH Los Lunas, NM Lynchburg, VA Macon, GA Madera, CA Manchester, NH MandevilleCovington, LA Manhattan, KS Mankato, MN Mansfield, OH Manteca, CA Maricopa, AZ Marysville, WA MauldinSimpsonville, SC Medford, OR Merced, CA Michigan CityLa Porte, INMI Middletown, NY Middletown, OH Midland, MI Midland, TX Minot, ND Missoula, MT MonessenCalifornia, PA Monroe, LA 5307 Appropriation FY19 722,092 1,452,236 1,827,667 1,620,842 1,070,408 938,604 890,440 988,787 1,746,879 1,962,918 1,491,590 2,516,162 2,533,974 915,573 826,663 1,360,617 713,382 1,791,436 1,175,041 851,416 1,547,211 5307 Forecast 75% 0 mi 2,915,100 913,652 2,668,951 2,573,211 2,442,075 2,529,228 1,546,804 1,298,186 928,720 1,085,915 2,953,838 1,171,841 2,630,056 2,249,526 3,154,918 3,561,306 1,089,232 1,015,157 1,510,994 1,144,045 2,390,757 1,011,600 2,229,996 1,265,603 1,816,683 50% mi 918,611 2,662,227 2,504,189 2,023,801 1,429,622 962,374 860,782 1,024,247 2,464,161 1,139,416 2,621,243 2,139,627 2,812,342 3,241,393 1,037,612 1,025,294 1,483,604 1,087,067 2,115,635 956,558 2,165,288 1,299,272 1,789,215 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 2,915,100 - 191,560 196,519 1,216,715 1,209,991 745,544 676,522 821,233 402,959 2,529,228 - 476,396 359,214 359,582 23,770 38,280 -29,658 97,128 35,460 1,206,959 717,282 1,171,841 1,139,416 667,138 658,325 757,936 648,037 638,756 296,180 1,027,332 707,419 173,659 122,039 188,494 198,631 150,377 122,987 430,663 373,685 599,321 324,199 1,011,600 956,558 1,054,955 990,247 414,187 447,856 269,472 242,004 % Diff (Forecast FY19) 75% 50% 0 mi mi 27% 27% 84% 83% 41% 37% 51% 25% 45% 34% 38% 3% 4% -3% 10% 4% 69% 41% 34% 34% 51% 43% 25% 12% 41% 28% 19% 13% 23% 24% 11% 9% 60% 52% 33% 18% 90% 84% 49% 53% 17% 16% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 145 Urban Area Monroe, MI Morehead City, NC Morgantown, WV Morristown, TN Mount Vernon, WA Muncie, IN Murfreesboro, TN Muskegon, MI Nampa, ID Napa, CA New Bedford, MA New Bern, NC Newark, OH Norman, OK North PortPort Charlotte, FL Ocala, FL Odessa, TX OlympiaLacey, WA Oshkosh, WI Owensboro, KY Paducah, KYIL Panama City, FL Parkersburg, WVOH Pascagoula, MS Petaluma, CA 5307 Appropriation FY19 688,588 - 1,034,057 687,736 903,069 1,350,345 1,833,136 2,081,949 2,397,017 1,573,642 2,577,908 591,280 1,081,879 1,651,451 2,163,800 2,015,958 1,916,391 2,365,100 1,200,616 1,075,117 1,903,395 907,248 634,229 1,136,707 5307 Forecast 75% 0 mi 1,286,040 - 2,165,261 853,483 1,880,113 2,038,752 2,885,099 2,444,320 3,340,029 2,866,140 3,018,416 747,918 1,328,187 2,428,836 3,098,998 2,761,271 2,701,601 4,131,528 1,613,545 1,811,210 693,965 2,459,879 1,068,131 1,218,359 1,377,294 50% mi 1,262,140 702,384 2,156,025 905,351 1,783,011 1,977,824 2,740,267 2,428,227 2,909,180 2,646,080 2,982,232 1,041,470 1,232,230 2,152,277 2,885,488 2,624,430 2,152,410 4,081,212 1,481,066 1,629,774 698,449 2,184,556 1,053,965 1,196,957 1,150,927 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 597,452 573,552 - 702,384 1,131,204 1,121,968 165,747 217,615 977,044 879,942 688,407 627,479 1,051,963 907,131 362,371 346,278 943,012 512,163 1,292,498 1,072,438 440,508 404,324 156,638 450,190 246,308 150,351 777,385 500,826 935,198 721,688 745,313 608,472 785,210 236,019 1,766,428 1,716,112 412,929 280,450 736,093 554,657 693,965 698,449 556,484 281,161 160,883 146,717 584,130 562,728 240,587 14,220 % Diff (Forecast FY19) 75% 50% 0 mi mi 87% 83% 109% 24% 108% 51% 57% 17% 39% 82% 17% 26% 23% 47% 43% 37% 41% 75% 34% 68% 109% 32% 97% 46% 49% 17% 21% 68% 16% 76% 14% 30% 33% 30% 12% 73% 23% 52% 29% 15% 18% 16% 92% 89% 21% 1% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 146 Urban Area Pittsfield, MA Pocatello, ID Port Arthur, TX Port Huron, MI Porterville, CA Portsmouth, NHME Pottstown, PA Prescott ValleyPrescott, AZ Pueblo, CO Racine, WI Rapid City, SD Redding, CA ReedleyDinuba, CA Rio Grande CityRoma, TX Rochester, MN Rock Hill, SC Rocky Mount, NC Rome, GA Roswell, NM Saginaw, MI SahuaritaGreen Valley, AZ Salinas, CA Salisbury, MDDE San Angelo, TX San Luis Obispo, CA 5307 Appropriation FY19 821,051 1,103,220 1,979,404 1,121,515 1,456,001 926,389 1,306,240 1,149,843 1,981,436 2,286,292 1,175,469 1,624,169 1,601,817 1,220,576 916,227 765,477 1,810,903 1,254,522 1,371,084 1,072,489 5307 Forecast 75% 0 mi 1,444,053 1,288,692 2,358,490 2,265,547 2,709,318 1,174,267 1,625,759 1,620,005 2,363,399 2,439,728 1,545,375 1,917,502 1,355,153 990,089 2,762,620 1,642,969 1,306,252 1,956,737 2,185,850 5,510,418 2,380,709 1,688,708 2,786,567 50% mi 1,737,420 1,155,061 2,079,759 2,251,134 2,380,202 1,195,476 1,635,653 1,410,151 2,181,702 2,288,857 1,285,716 1,949,402 1,338,763 986,719 2,559,383 1,831,245 1,224,094 1,941,604 823,725 2,081,979 825,427 - 2,352,297 1,393,005 2,537,750 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 623,002 916,369 185,472 51,841 379,086 100,355 1,144,032 1,129,619 1,253,317 924,201 247,878 269,087 319,519 329,413 470,162 260,308 381,963 200,266 153,436 2,565 369,906 110,247 293,333 325,233 1,355,153 1,338,763 990,089 986,719 1,160,803 957,566 422,393 610,669 390,025 307,867 1,191,260 1,176,127 - 823,725 374,947 271,076 - 825,427 5,510,418 - 1,126,187 1,097,775 317,624 21,921 1,714,078 1,465,261 % Diff (Forecast FY19) 75% 0 mi 50% mi 76% 112% 17% 5% 19% 5% 102% 101% 86% 63% 27% 29% 24% 25% 41% 23% 19% 10% 7% 0% 31% 9% 18% 20% 72% 35% 43% 156% 21% 60% 50% 34% 154% 15% 90% 23% 160% 88% 2% 137% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 147 Urban Area San Marcos, TX Santa Cruz, CA Santa Fe, NM Santa Maria, CA Saratoga Springs, NY SeasideMonterey, CA SebastianVero Beach SouthFlorida Ridge, FL SebringAvon Park, FL Sheboygan, WI Sherman, TX Sierra Vista, AZ Simi Valley, CA Sioux City, IANESD Sioux Falls, SD Slidell, LA South LyonHowell, MI Spring Hill, FL Springfield, IL Springfield, OH St. Augustine, FL St. Cloud, MN St. George, UT St. Joseph, MOKS State College, PA StauntonWaynesboro, VA 5307 Appropriation FY19 830,113 2,837,032 1,230,908 3,076,557 778,802 2,029,566 1,963,278 785,158 1,065,765 872,755 724,249 2,625,241 1,558,458 1,205,416 1,346,432 1,828,601 2,228,815 1,206,630 927,394 1,718,414 1,525,247 1,187,661 1,613,613 728,453 5307 Forecast 75% 0 mi 2,437,908 4,913,633 2,049,082 3,522,923 1,017,338 3,444,597 2,676,144 1,054,344 1,173,720 1,304,645 923,676 3,028,277 1,720,860 3,559,375 1,514,744 2,006,313 2,732,240 2,841,725 1,361,672 1,311,834 2,754,778 2,356,589 1,610,366 3,407,390 947,257 50% mi 2,285,495 4,640,124 2,056,532 2,861,033 1,027,568 3,248,150 2,462,920 978,860 1,053,574 1,151,989 761,591 2,889,692 1,480,196 1,469,060 2,043,596 2,651,496 2,538,201 1,279,173 1,190,728 2,440,099 2,262,935 1,437,159 3,366,392 853,011 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 1,607,795 1,455,382 2,076,601 1,803,092 818,174 825,624 446,366 -215,524 238,536 248,766 1,415,031 1,218,584 712,866 499,642 269,186 193,702 107,955 -12,191 431,890 279,234 199,427 37,342 403,036 264,451 162,402 -78,262 3,559,375 - 309,328 263,644 659,881 697,164 903,639 822,895 612,910 309,386 155,042 72,543 384,440 263,334 1,036,364 721,685 831,342 737,688 422,705 249,498 1,793,777 1,752,779 218,804 124,558 % Diff (Forecast FY19) 75% 0 mi 50% mi 194% 175% 73% 64% 66% 67% 15% -7% 31% 32% 70% 60% 36% 25% 34% 25% 10% -1% 49% 32% 28% 5% 15% 10% 10% -5% 26% 49% 49% 27% 13% 41% 60% 55% 36% 111% 30% 22% 52% 45% 14% 6% 28% 42% 48% 21% 109% 17% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 148 Urban Area Stillwater, OK Sumter, SC Temple, TX Terre Haute, IN TexarkanaTexarkana, TXAR Texas City, TX Titusville, FL Topeka, KS Tracy, CA Traverse City, MI Turlock, CA Tuscaloosa, AL Twin Falls, ID Twin RiversHightstown, NJ Tyler, TX UniontownConnellsville, PA Utica, NY Vacaville, CA Valdosta, GA Vallejo, CA Victoria, TX Villas, NJ Vineland, NJ Waco, TX Waldorf, MD 5307 Appropriation FY19 870,749 1,237,833 1,319,858 969,544 1,369,129 767,859 2,148,951 1,837,089 2,133,605 1,867,821 861,818 1,705,730 649,521 1,697,100 1,777,453 1,024,294 3,517,573 987,194 657,906 1,263,177 1,414,104 5307 Forecast 75% 0 mi 1,106,195 1,690,207 1,560,233 1,170,657 1,732,190 1,186,562 2,465,839 2,784,440 674,556 2,468,995 2,393,451 1,119,246 1,076,254 2,239,219 1,285,294 1,940,546 2,585,828 1,257,182 4,338,503 1,761,822 779,544 1,521,206 3,231,146 1,974,099 50% mi 896,600 1,071,453 1,456,923 1,431,153 1,134,023 1,561,273 1,135,801 2,204,142 2,089,203 685,668 2,012,569 2,245,936 1,090,944 1,079,753 2,069,223 1,307,165 1,960,863 2,318,020 1,204,974 3,871,113 1,500,705 781,379 1,526,283 1,870,104 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi - 896,600 235,446 200,704 452,374 219,090 240,375 111,295 201,113 164,479 363,061 192,144 418,703 367,942 316,888 55,191 947,351 252,114 674,556 685,668 335,390 -121,036 525,630 378,115 1,119,246 1,090,944 214,436 217,935 533,489 363,493 635,773 657,644 243,446 263,763 808,375 540,567 232,888 180,680 820,930 353,540 774,628 513,511 121,638 123,473 258,029 263,106 3,231,146 - 559,995 456,000 % Diff (Forecast FY19) 75% 50% 0 mi mi 27% 23% 37% 18% 18% 8% 21% 17% 27% 14% 55% 48% 15% 3% 52% 14% 16% -6% 28% 20% 25% 25% 31% 21% 98% 101% 14% 16% 45% 30% 23% 18% 23% 10% 78% 52% 18% 19% 20% 21% 40% 32% TABLE C5 Comparison of Current (FY19) and Future 5307 Funding for Small Urban Areas (Assumes Same FTA Data Values) ( Continued) 149 Urban Area Walla Walla, WAOR Warner Robins, GA Waterbury, CT Waterloo, IA Watertown, NY Watsonville, CA Wausau, WI WeirtonSteubenville, WVOHPA Wenatchee, WA West Bend, WI WestminsterEldersburg, MD Wheeling, WVOH Wichita Falls, TX Williamsburg, VA Williamsport, PA Wilson, NC Winchester, VA Woodland, CA Yakima, WA Yuba City, CA Yuma, AZCA Zephyrhills, FL TOTAL SMALL URBAN 5307 Appropriation FY19 829,386 1,676,792 8,488,289 1,604,739 743,530 1,423,095 989,878 904,359 1,024,442 918,661 834,683 1,130,657 1,465,447 911,217 853,990 974,250 1,303,449 2,032,796 2,170,111 2,165,077 883,931 401,726,141 5307 Forecast 75% 0 mi 1,489,118 2,720,081 4,779,102 1,908,066 907,178 2,708,043 1,168,191 1,036,873 2,570,962 1,648,824 996,382 1,278,391 1,771,442 2,078,793 2,545,269 864,731 1,286,611 2,052,983 2,382,514 2,762,012 2,979,528 1,221,736 608,072,325 50% mi 1,391,064 2,537,309 4,781,643 1,702,558 909,395 2,409,307 1,082,226 1,032,167 2,444,707 1,533,497 986,942 1,256,663 1,509,023 2,001,495 2,551,217 872,659 1,163,577 1,552,273 2,114,629 2,572,352 2,700,259 1,122,781 550,150,188 5307 Forecast FY19 Appropriation 75% 0 mi 50% mi 659,732 561,678 1,043,289 860,517 (3,709,187) (3,706,646) 303,327 97,819 163,648 165,865 1,284,948 986,212 178,313 92,348 132,514 127,808 1,546,520 1,420,265 730,163 614,836 161,699 152,259 147,734 126,006 305,995 43,576 1,167,576 1,090,278 1,691,279 1,697,227 864,731 872,659 312,361 189,327 749,534 248,824 349,718 81,833 591,901 402,241 814,451 535,182 337,805 238,850 206,346,184 148,424,047 % Diff (Forecast FY19) 75% 0 mi 50% mi 80% 68% 62% 51% -44% -44% 19% 6% 22% 22% 90% 69% 18% 9% 15% 14% 151% 139% 79% 67% 19% 18% 13% 11% 21% 3% 128% 120% 198% 199% 32% 58% 17% 27% 38% 38% 51% 19% 19% 4% 19% 25% 27% 37% APPENDIX D: SUPPORTING TABLES FOR PREDICTED CHANGES IN 5311 AND 5307 FUNDING ALLOCATIONS BY COUNTY IN GEORGIA This appendix contains supporting tables for the following future scenarios: Scenario 1A corresponds to the 50% probability model using a mile distance threshold Scenario 2B corresponds to the 75% probability model using a 0 mile distance threshold Tables D1D5 report predicted changes in 5311 and 5307 funding for counties in Georgia. Unlike the analysis in Appendix C, the 5311 and 5307 totals reported here do include the growing states portion (of 5340 piece). The tables organize the counties according to the following classifications: Counties that Currently Do Not Have Transit Service Counties that Currently Operate Countywide 5311 Service and No 5307 Service Counties that Currently Operate Citywide 5311 Service and No 5307 Service Counties that Currently Operate Both 5311 and 5307 Service Counties that Currently Operate Only 5307 Service Similar to Tables C2C5, Tables D1D5 show the changes in 5311 and 5307 funding that each county in Georgia would experience if the FTA data values from FY18 were applied to the new population, population density, and other inputs used in the allocation 150 formula after the 2020 census. The key difference is that the numbers reported in Appendix D include the 5340 growing states portion for both the 5311 and 5307 amounts. 151 TABLE D1 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Do Not Have Transit Service 152 County Appling Atkinson Barrow Candler Charlton Chattahoochee Clinch Coffee Echols Emanuel Evans Fayette Franklin Harris Houston Irwin Jasper Jeff Davis Johnson Lanier Laurens Madison Marion Monroe Montgomery 5311 Appropriation FY19 139,727 70,248 356,850 82,389 110,912 30,882 81,943 290,198 44,480 175,702 77,378 116,242 150,781 205,599 98,435 76,767 104,453 110,267 75,337 72,093 343,592 173,755 71,335 173,863 66,243 5307 Appropriation FY19 64,187 106,684 466,882 743 2,250,481 25,064 4,309 - 5311 Forecast 75% 0 mi 141,462 78,140 102,638 83,695 118,680 33,195 82,813 305,999 46,431 180,103 81,161 96,611 151,556 224,836 87,112 75,794 103,696 112,854 75,034 76,886 351,720 177,972 74,590 180,446 70,358 50% mi 143,192 79,095 62,673 84,719 120,132 33,547 83,826 309,741 46,999 182,305 82,153 53,439 153,409 225,093 71,540 76,720 103,472 114,234 75,951 76,808 356,021 179,683 75,502 171,880 71,218 5307 Forecast 75% 0 mi 305,713 112,384 492,792 502 2,547,348 906 - 7,356 - 50% mi 325,985 111,723 500,578 2,879 2,300,627 2,075 3,359 36,036 20,906 - 5311 Difference ForecastFY19 75% 0 mi 50% mi 1,735 3,465 7,892 8,847 -254,212 -294,177 1,306 2,330 7,768 9,220 2,313 2,665 870 1,883 15,801 19,543 1,951 2,519 4,401 6,603 3,783 4,775 -19,631 -62,803 775 2,628 19,237 19,494 -11,323 -26,895 -973 -47 -757 -981 2,587 3,967 -303 614 4,793 4,715 8,128 12,429 4,217 5,928 3,255 4,167 6,583 -1,983 4,115 4,975 5307 Difference ForecastFY19 75% 50% 0 mi mi 241,526 261,798 5,700 5,039 25,910 33,696 -241 2,136 296,867 50,146 906 2,075 3,359 - 36,036 3,047 16,597 TABLE D1 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Do Not Have Transit Service ( Continued) 5311 5307 Appropriation Appropriation 5311 Forecast 5307 Forecast 5311 Difference 5307 Difference (Forecast FY19) (Forecast FY19) County FY19 FY19 75% 0 mi 50% mi 75% 0 mi 50% mi 75% 0 mi 50% mi 75% 0 mi 50% mi Newton 199,352 395,258 182,894 148,808 445,782 445,711 -16,458 -50,544 50,524 50,453 Oconee 102,671 160,141 94,694 72,770 205,074 219,724 -7,977 -29,901 44,933 59,583 Oglethorpe 110,363 1,026 113,968 114,891 1,034 1,796 3,605 4,528 8 770 Rockdale 81,003 456,127 63,759 37,823 482,190 479,565 -17,244 -43,180 26,063 23,438 Schley 39,148 - 42,003 42,517 - - 2,855 3,369 Stephens 169,326 - 170,702 172,790 - - 1,376 3,464 Tattnall 178,355 - 188,052 190,352 - - 9,697 11,997 Toombs 190,041 - 195,572 197,964 - - 5,531 7,923 Treutlen 51,556 - 51,060 51,685 - - -496 129 Washington 162,573 - 161,493 163,468 - - -1,080 895 Webster 26,762 - 28,543 28,892 - - 1,781 2,130 White 175,914 - 184,811 187,071 - - 8,897 11,157 TOTAL 4,786,535 3,930,902 4,591,333 4,442,383 4,601,081 4,450,964 -195,202 -344,152 695,243 545,126 153 TABLE D2 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Countywide 5311 Service and No 5307 Service 154 County Bacon Baker Baldwin Banks Ben Hill Berrien Bleckley Brantley Brooks Bryan Bulloch Burke Butts Camden Carroll Catoosa Chattooga Clay Colquitt Columbia Cook Coweta Crawford Crisp Dade 5311 Appropriation FY19 85,353 39,788 303,816 127,715 141,737 149,233 92,934 139,785 129,146 153,844 502,429 196,499 157,906 351,066 586,929 121,553 180,648 43,425 355,339 160,529 137,809 294,015 97,000 178,092 103,421 5307 Appropriation FY19 9,933 53,704 123,215 269,716 615,924 485,128 9,527 5311 Forecast 75% 0 mi 86,968 39,821 306,337 131,272 142,225 152,702 92,834 145,157 123,938 184,309 514,839 201,100 155,169 371,619 592,153 114,839 177,718 42,899 371,475 150,506 148,469 276,641 95,998 176,196 105,144 50% mi 88,031 40,308 310,083 132,877 143,964 154,569 93,969 146,932 124,839 134,336 521,135 203,559 152,634 376,163 241,717 113,176 179,891 43,424 376,017 143,481 150,284 183,605 97,172 178,350 101,163 5307 Forecast 75% 0 mi 12,560 104,027 1,291 160,175 346,800 1,052,167 624,565 10,444 50% mi 13,762 153,850 4,148 153,218 334,730 1,027,775 667,081 12,405 5311 Difference (Forecast FY19) 75% 0 mi 50% mi 1,615 2,678 33 520 2,521 6,267 3,557 5,162 488 2,227 3,469 5,336 -100 1,035 5,372 7,147 -5,208 -4,307 30,465 -19,508 12,410 18,706 4,601 7,060 -2,737 -5,272 20,553 25,097 5,224 -345,212 -6,714 -8,377 -2,930 -757 -526 -1 16,136 20,678 -10,023 -17,048 10,660 12,475 -17,374 -110,410 -1,002 172 -1,896 258 1,723 -2,258 5307 Difference (Forecast FY19) 75% 50% 0 mi mi 2,627 3,829 50,323 100,146 1,291 4,148 36,960 30,003 77,084 65,014 436,243 411,851 139,437 181,953 917 2,878 TABLE D2 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Countywide 5311 Service and No 5307 Service ( Continued) 155 County Dawson Decatur Dodge Dooly Early Effingham Elbert Fannin Forsyth Gilmer Glascock Glynn Gordon Grady Greene Habersham Hancock Haralson Hart Heard Jackson Jefferson Jenkins Jones Lamar 5311 Appropriation FY19 120,195 225,414 162,359 132,151 103,306 342,856 150,338 169,038 115,481 201,061 27,871 204,146 362,415 199,776 132,953 278,571 86,752 195,205 171,447 89,321 341,984 142,343 72,432 165,680 123,633 5307 Appropriation FY19 22,585 - 6,524 - 922,057 - 667,852 - 41,301 - 53,219 - 5311 Forecast 75% 0 mi 91,720 222,943 161,786 129,552 100,496 374,753 147,517 193,238 65,249 214,464 28,596 192,009 376,325 203,446 141,982 297,668 86,370 194,004 173,395 87,013 336,684 142,563 78,347 167,998 122,400 50% mi 65,326 225,669 163,764 131,137 101,725 241,808 149,321 195,601 22,120 217,087 28,946 160,653 380,927 205,934 143,718 301,308 87,426 139,926 175,515 87,970 221,705 144,306 79,305 163,988 123,896 5307 Forecast 75% 0 mi 49,032 - 21,309 - 1,321,238 - 720,823 - 1,422 - 93,661 - 48,094 129 50% mi 91,103 - 191,892 - 1,373,610 - 660,734 - 56,111 - 97 200,775 58,138 126 5311 Difference (Forecast FY19) 75% 50% 0 mi mi -28,475 -54,869 -2,471 255 -573 1,405 -2,599 -1,014 -2,810 -1,581 31,897 -101,048 -2,821 -1,017 24,200 26,563 -50,232 -93,361 13,403 16,026 725 1,075 -12,137 -43,493 13,910 18,512 3,670 6,158 9,029 10,765 19,097 22,737 -382 674 -1,201 -55,279 1,948 4,068 -2,308 -1,351 -5,300 -120,279 220 1,963 5,915 6,873 2,318 -1,692 -1,233 263 5307 Difference (Forecast FY19) 75% 50% 0 mi mi 26,447 68,518 14,785 185,368 399,181 451,553 52,971 -7,118 1,422 56,111 97 52,360 159,474 -5,125 129 4,919 126 TABLE D2 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Countywide 5311 Service and No 5307 Service ( Continued) 156 County Lee Lincoln Long Lowndes Lumpkin McIntosh Macon McDuffie Meriwether Miller Mitchell Morgan Murray Paulding Peach Pickens Pierce Pike Pulaski Putnam Quitman Rabun Randolph Screven 5311 Appropriation FY19 109,217 62,560 97,555 247,428 201,185 110,952 117,894 156,230 163,503 56,244 189,939 130,783 196,489 206,703 139,557 188,909 142,209 116,990 85,829 151,234 30,616 123,047 94,469 127,231 5307 Appropriation FY19 167,091 - 33,585 1,076,358 133,565 649,216 66,984 954 - 5311 Forecast 75% 0 mi 110,216 62,897 135,136 237,524 210,523 112,492 113,137 156,914 161,675 55,370 186,698 132,538 185,842 156,953 129,030 198,920 146,942 113,844 85,200 153,537 29,191 127,357 93,368 125,686 50% mi 80,106 63,666 134,301 228,695 212,607 113,867 114,521 158,833 163,485 56,047 188,981 134,159 168,111 118,370 116,957 201,353 148,739 102,461 86,242 155,415 29,548 128,914 94,510 127,223 5307 Forecast 75% 0 mi 192,431 - 61,112 1,196,358 35,857 163,020 814,427 80,934 4,824 - 50% mi 223,772 - 63,895 1,125,250 573 - 126 - 194,345 809,109 91,379 - 13,168 - 5311 Difference (Forecast FY19) 75% 50% 0 mi mi 999 -29,111 337 1,106 37,581 36,746 -9,904 -18,733 9,338 11,422 1,540 2,915 -4,757 -3,373 684 2603 -1,828 -18 -874 -197 -3,241 -958 1,755 3,376 -10,647 -28,378 -49,750 -88,333 -10,527 -22,600 10,011 12,444 4,733 6,530 -3,146 -14,529 -629 413 2,303 4,181 -1,425 -1,068 4,310 5,867 -1,101 41 -1,545 -8 5307 Difference (Forecast FY19) 75% 50% 0 mi mi 25,340 56,681 27,527 120,000 30,310 48,892 573 126 29,455 60,780 165,211 159,893 13,950 24,395 3,870 12,214 TABLE D2 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Countywide 5311 Service and No 5307 Service ( Continued) 157 County Seminole Spalding Stewart Talbot Taliaferro Taylor Telfair Thomas Tift Towns Troup Turner Twiggs Union Upson Walker Ware Warren Wayne Wheeler Whitfield Wilcox Wilkes Wilkinson TOTAL 5311 Appropriation FY19 72,738 178,815 78,594 72,558 22,137 83,154 125,394 333,441 268,249 73,881 442,159 74,070 76,992 145,535 188,702 274,768 273,977 52,712 237,040 62,975 211,382 86,544 94,561 84,129 16,025,544 5307 Appropriation FY19 231,447 183,474 952,223 6,800,646 5311 Forecast 75% 0 mi 74,453 172,344 77,534 71,596 22,433 82,385 127,013 347,403 269,824 82,607 467,006 71,212 72,244 158,815 187,808 256,458 268,808 50,489 239,893 67,755 177,180 86,399 93,791 84,030 16,082,374 50% mi 75,364 140,893 78,482 72,471 22,708 83,392 128,566 351,651 273,123 83,617 472,717 72,083 72,988 160,757 190,105 257,611 272,095 51,106 243,328 68,583 156,211 87,455 94,938 85,057 15,134,022 5307 Forecast 75% 0 mi 247,702 300,589 1,029,458 8,694,449 50% mi 265,604 154 301,984 1,013,004 9,137,954 5311 Difference (Forecast FY19) 75% 50% 0 mi mi 1,715 2,626 -6,471 -37,922 -1,060 -112 -962 -87 296 571 -769 238 1,619 3,172 13,962 18,210 1,575 4,874 8,726 9,736 24,847 30,558 -2,858 -1,987 -4,748 -4,004 13,280 15,222 -894 1,403 -18,310 -17,157 -5,169 -1,882 -2,223 -1,606 2,853 6,288 4,780 5,608 -34,202 -55,171 -145 911 -770 377 -99 928 56,830 -891,522 5307 Difference (Forecast FY19) 75% 50% 0 mi mi 16,255 34,157 154 117,115 118,510 77,235 60,781 1,883,010 2,337,308 TABLE D3 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Citywide 5311 Service and No 5307 Service County Calhoun Polk Sumter Terrell Walton Worth TOTAL 5311 Appropriation FY19 60,196 272,322 235,049 82,850 350,405 178,361 1,179,183 5307 Appropriation FY19 153,032 153,032 5311 Forecast 75% 0 mi 60,062 278,108 226,890 80,453 332,242 175,944 1,153,699 50% mi 60,797 281,509 229,665 81,436 187,394 178,096 1,018,897 5307 Forecast 75% 0 mi 195,497 195,497 50% mi 302,973 302,973 5311 Difference (Forecast FY19) 75% 0 mi 50% mi -134 601 5,786 9,187 -8,159 -5,384 -2,397 -1,414 -18,163 -163,011 -2,417 -265 -25,484 -160,286 5307 Difference (Forecast FY19) 75% 50% 0 mi mi 42,465 149,941 42,465 149,941 158 TABLE D4 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Both 5311 and 5307 Service County Bartow Chatham Cherokee Dougherty Hall Henry Liberty Richmond TOTAL 5311 Appropriation FY19 277,333 110,493 259,193 107,834 259,256 198,129 116,931 140,988 1,470,157 5307 Appropriation FY19 1,220,368 3,646,778 1,185,905 1,172,357 2,194,364 1,231,829 723,315 1,702,365 13,077,281 5311 Forecast 75% 0 mi 243,254 91,053 207,921 96,409 244,407 98,885 120,785 123,295 1,226,009 50% mi 191,218 62,680 131,575 91,353 231,928 39,332 118,257 110,685 977,028 5307 Forecast 75% 0 mi 1,268,757 3,751,271 1,407,028 1,127,522 2,417,680 1,455,091 771,419 1,703,093 13,901,861 50% mi 1,301,721 3,658,738 1,410,516 1,108,121 2,357,220 1,488,995 740,703 1,649,351 13,715,363 5311 Difference (Forecast FY19) 75% 0 mi 50% mi -34,079 -86,115 -19,440 -47,813 -51,272 -127,618 -11,425 -16,481 -14,849 -27,328 -99,244 -158,797 3,854 1,326 -17,693 -30,303 -244,148 -493,129 5307 Difference (Forecast FY19) 75% 0 mi 50% mi 48,389 81,353 104,493 11,960 221,123 224,611 -44,835 -64,236 223,316 162,856 223,262 257,166 48,104 17,388 728 -53,014 824,580 638,084 159 TABLE D5 Comparison of Current (FY19) and Future 5311 and 5307 Funding for Counties in Georgia (Assumes Same FTA Data Values and Includes Growing States): Counties that Currently Operate Only 5307 Service County Bibb Clarke Clayton Cobb DeKalb Douglas Floyd Fulton Gwinnett Muscogee TOTAL GRAND TOTAL 5311 Appropriation FY19 147,791 46,333 15,329 10,283 11,712 131,355 238,009 64,949 24,449 40,891 731,101 5307 Appropriation FY19 2,412,707 2,391,818 7,268,028 6,500,720 19,785,158 1,134,652 1,861,312 25,680,194 6,621,237 3,541,921 77,197,747 5311 Forecast 75% 0 mi 117,092 38,655 10,386 218 4,036 104,526 233,387 39,349 11,379 37,757 596,785 50% mi 80,610 33,160 2,182 361 74,528 216,505 16,275 5,864 29,741 459,226 5307 Forecast 75% 0 mi 2,431,907 2,631,717 7,311,963 6,964,350 20,012,098 1,208,721 1,883,167 26,396,943 7,522,623 3,733,106 80,096,595 50% mi 2,265,132 2,530,319 7,251,276 6,826,786 19,793,833 1,181,012 1,874,303 26,089,202 7,328,509 3,703,723 78,844,095 5311 Difference (Forecast FY19) 75% 0 mi 50% mi -30,699 -67,181 -7,678 -13,173 -4,943 -13,147 -10,065 -7,676 -11,351 -26,829 -56,827 -4,622 -21,504 -25,600 -48,674 -13,070 -18,585 -3,134 -11,150 -134,316 -261,592 5307 Difference (Forecast FY19) 75% 0 mi 50% mi 19,200 -147,575 239,899 138,501 43,935 -16,752 463,630 326,066 226,940 8,675 74,069 46,360 21,855 12,991 716,749 409,008 901,386 707,272 191,185 161,802 2,898,848 1,646,348 24,174,995 101,134,544 23,629,142 22,010,706 107,489,483 106,415,315 -545,853 -2,154,006 6,344,146 5,280,771 160