7 United States J A0 General Accounting Washington, Office D.C. 20548 Genera3 Government Division B-277191 June 26, 1997 The Honorable Sam Brownback Chairman, Subcommittee on Oversight of Government Management, Restructuring, and the District of Columbia Committee on Governmental Affairs United States Senate The Honorable Lauch Faircloth Chairman, Subcommittee on District of Columbia Committee on Appropriations United States Senate The Honorable Tom Davis Chairman, Subcommittee on the District of Columbia Committee on Government Reform and Oversight House of Representatives The Honorable Charles Taylor Chairman, Subcommittee on District of Columbia Committee on Appropriations House of Representatives Subject: District of Columbia: Revenues Compared With Those of Selected Cities In response to your request, this letter updates information presented in our 1995 GAO letter that compared revenues of the District of Columbia with those of selected other cities.’ Specifically, in this letter we (1) compare the District’s revenues, on a per capita basis, with the per capita revenues of eight selected cities,” using alternative assumptions regarding the allocation of state- level revenue to cities; and (2) identify some potential reasons why expenditures per capita may differ across cities. ‘District of Columbia: Exnenditures & Revenues (GAO/OCE-95-2R, Aug. 1, 1995). ‘Baltimore, Boston, Honolulu, Indianapolis, Jacksonville, New York, Philadelphia, and San Francisco. GAO/GGD-97-135B District of Columbia Revenues B-277191 The District of Columbia is responsible for functions that are typically divided among state, county, city, school district, and other local units of government. Consequently, comparisons of the District’s revenues per capita to those of other cities, without adjusting for the revenues of other levels of government, are not meaningful. To partially mitigate the comparability problems, we compared the District to other large cities with combined city/county governments and allocated school district and state revenues to the cities. Nevertheless, our comparisons should be viewed as rough at best. In addition, our results should not be used to judge whether the District’s revenues are sufficient to meet its expenditure needs. A much broader analysis, which was beyond the scope of your request, would be needed to address the question of revenue sufficiency. SCOPE AND METHODOLOGY To avoid problems in allocating county revenues between cities and other jurisdictions that might be contained within a county, we limited our comparison to cities that have combined city/county governments. We obtained the financial, population, and school enrollment data needed for our analysis from the Bureau of the Census for fiscal year 1993-94. These data were the latest available, and they were readily available only for cities with populations that exceeded 500,000 in 1990. Only eight cities with combined city/county governments met this size criterion. We did not independently verify the accuracy of the census data. For the eight cities, we adjusted for large intergovernmental revenue flows to avoid double counting. Except in the case of Honolulu, we relied on Census coding to ensure that the school districts included in our analysis comprehensively covered students attending public schools within the counties that we compared and did not cover students attending schools outside those counties. For Honolulu, we had to estimate the city’s share of the revenues of the state’s single school district on the basis of its share of total student population. We did not try to adjust for differences in fiscal years across jurisdictions, because the Census data did not provide any basis on which we could make such an adjustment. To make the total revenues of the selected cities more comparable to the District’s total revenue, we allocated state-level revenues to those cities. We were able to determine the amount of state-level revenue that was provided to each city in the form of direct transfers. However, we had no empirical basis for allocating shares of the nontransferred revenues to each city.3 We made our comparisons using two alternative assumptions regarding the allocation of that revenue. Our first assumption was that each city received a share of nontransferred state revenues that was proportionate to that city’s share of total state population. Our alternative assumption was that each city received a share of “Nontransferred revenue is that portion of a state’s total revenue that has not been distributed to specific localities in the form of intergovernmental transfers. 2 GAO/GGD-97-135R District of Columbia Revenues B-27719 1 the nontransferred revenue that was proportionate to its share of total state personal income. Because we do not know what states actually spent in each of the eight cities, our allocation assumptions may result in an under- or overstatement of the amount of revenue that cities would have to raise themselves if they were assigned state-level responsibilities. For all of the cities except the District of Columbia, the amount of revenues that we had to allocate on the basis of assumptions was large compared to those cities’ other revenues (see table 2). We did not try to allocate revenue used by the federal government to provide services, such as national parks or police service, in some cities. Our computations do not incorporate the revenue of special districts, such as regional transportation districts or local housing districts, because the Census Bureau has not published revenue data for these districts since fiscal year 1991-92. Census also has not published recent revenue data for small independent townships contained within the boundaries of the counties that are largely consolidated with Boston, Indianapolis, and Jacksonville. The 1991-92 revenue data for the independent townships and large special districts associated with the District and the eight cities in our study are presented in enclosure I. We discuss the effect of excluding these revenues when we present our results. We also reviewed a variety of studies of the District of Columbia’s finances, some of which make comparisons to other state or local jurisdictions, to identify potential reasons why expenditures per capita may differ across cities. We did our work in Washington, D.C., between May and June 1997 in accordance with generally accepted government auditing standards. We did not audit the reliability of the data collection systems that produced the information we used. We provided a draft of this letter to the District of Columbia Financial Responsibility and Management Assistance Authority and to the Chief Financial Officer (CFO) of the District of Columbia. On June 23, 1997, officials from each of those offices provided written comments on this letter that are reprinted in enclosures II and III and discussed at the end of this letter. RESULTS In our analysis, the District of Columbia had the highest total revenue per capita among the cities we reviewed, under both of the nontransferred revenue allocation assumptions we made (see table 1). GAOIGGD-97-135R District of Columbia Revenues B-277191 Table 1: Total Revenue per Capita Under Two Alternative Assumptions Total revenue if Total revenue if nontransferred nontransferred state revenue was state revenue was allocated on the allocated on the basis of personal City/County basis of population income District of Columbia $8,286 $8,286 New York 7,673 7,851 San Francisco 5,982 6,740 Honolulu 5,965 6,218 Boston 5,455 5,764 Philadelphia 4,804 4,643 Baltimore 4,649 4,358 Indianapolis 4,086 4,360 Jacksonville 3,417 3,357 Source: GAO computations based on the Census Bureau’s Government Finances Series: 1993-1994, and Countv and Citv Data Book: 1994. Other assumptions could produce a different result. For example, in commenting on a draft of this letter, the Office of the Chief Financial Officer of the District of Columbia offered an analysis using a different assumption.4 Under their assumption, all nontransferred state revenue was allocated to cities in proportion to their share of the state’s poverty population (see enc. III). Under this assumption, the District ranked second, behind New York City. While the ranking of the cities did not change significantly from one assumption to the next, the amount of total revenue per capita for some of the cities differed significantly under the alternative assumptions. These results illustrate the sensitivity of the comparisons to the assumptions used. Comparisons of city total revenues per capita do not address the question of whether the revenue available to each city was sufficient to meet its expenditure needs and should not be used alone to assess whether these revenues are too high or too low. Our review of ‘We did not verify the analysis provided by the CFO’s Office. 4 GAO/GGD-97-135R District of Columbia Revenues B-277191 studies of the District of Columbia’s finances revealed that expenditures per capita may vary across cities owing to differences in (1) the types of services provided, (2) the share of each city’s population that receives particular services, (3) the quality of service provided to each recipient, (4) the costs of labor and other service inputs, and (5) the efficiency of service provision. COMPARISON OF DISTRICT REVENUES WITH THOSE OF EIGHT OTHER COMBINED CITY/COUNTY GOVERNMENTS For the purposes of this comparison, total revenues for the District of Columbia are defined as all general revenues5 that it collects itself, plus the transfer that it receives from the federal government. For the eight selected cities, we have defined total revenues to be the sum of (1) general revenues that the cities collect themselves; (2) revenues that independent school systems use to educate students residing in those cities;” (3) revenues that the cities receive in the form of direct transfers from the federal, state, and other local governments; and (4) a portion of their states’ revenues that are not transferred to specific local governments. We were able to determine the first three components of total revenue for each city from available data. However, we could not determine exactly how much of the nontransferred state revenues should be allocated to each city. The Census data on state revenues and expenditures that we used were constructed from accounting records that generally did not indicate which specific localities benefit from state resources that are not dispensed as intergovernmental transfers. Lacking an empirical basis for allocating nontransferred state revenue, we chose to make our comparisons using two alternative assumptions regarding the allocation of that revenue. Our first assumption was that each city received a share of nontransferred state revenues that was proportionate to that city’s share of total state population. Our alternative assumption was that each city received a share of the nontransferred revenue that was proportionate to its share of total state personal income. Our estimates are sensitive to the assumptions that we made and, therefore, are not precise. For all of the cities except the District of Columbia, the amount of revenue that “General revenues do not include revenues from local government-owned liquor stores; receipts from sales of government-operated water supply, electric power, gas supply, and public transit systems; and employee contributions to public employee retirement insurance systems. ?.n the case of dependent school systems, funds collected for education are included in the city’s own revenues. 5 GAOIGGD-97-135R District of Columbia Revenues B-277191 we had to allocate on the basis of assumptions was large relative to those cities’ other revenues. The importance of those allocations can be seen in table 2, which shows the revenues per capita of the District of Columbia and the eight cities by source of revenue. In most cases, the amount of nontransferred state revenue that we had to allocate exceeded the amount of revenue that cities raised themselves; in all cases, except for San F’rancisco, it exceeded the direct transfers that the cities received from other levels of government. The sensitivity of city revenue estimates to assumptions about allocating nontransferred state revenue is also shown in comments on this letter we received from the Office of the Chief Financial Officer of the District of Columbia. The CFO’s Office produced city revenue estimates based on the assumption that all nontransferred state revenues are allocated to cities in proportion to their share of their state’s poverty population. These estimates are included in enclosure III to this letter. Under the assumptions used by the CFO’s office, the District ranks second, behind New York City, in total revenues per capita, and the gap between the District and most other cities in our comparison is smaller than under our two assumptions. The exclusion of special district and independent township revenues also reduces the precision of our estimates. For example, if we included all of the 1991-92 revenue of the townships and those special districts that were completely contained within the boundaries of each city/county, total revenue per capita would have been higher in all of the cities, except the District of Columbia and Honolulu.7 However, the addition of these revenues would have had a negligible impact on the rankings shown in table 2 (Boston would have moved ahead of Honolulu under the second assumption). 7The 1991-92 additional revenue per capita for each city would have been $1 for New York, $122 for San Francisco, $502 for Boston, $279 for Philadelphia, $235 for Baltimore, $476 for Indianapolis, and $147 for Jacksonville. 6 GAO/GGD-97-135R District of Columbia Revenues B-277191 MULTIPLE FACTORS AFFECT REVENUE NEEDS The amount of revenue per capita that a city needs depends on both the demand for public services in the city and the efficiency with which the city delivers those services. Our analysis did not address the question of whether the revenue available to each city was sufficient to meet its expenditure needs. The following factors have been cited in previous studies’ as causes of variations in per capita expenditures across jurisdictions: - Tvnes of Services Provided. Cities differ in terms of the types of service they provide. For example, Honolulu does not have to provide snow removal services. Some cities may have privatized functions, such as trash removal and bus service. - Share of Ponulation Receiving Services. The costs of some services, such as Medicaid, Aid to Families with Dependent Children (AFDC), and foster care, depend on the number of beneficiaries receiving the service. The larger the share of a city’s population receiving these services, the higher the costs per capita will be. -- Qualitv of Service. City governments differ in the quality of service they provide to each service recipient. For example, there are variations across cities in the coverage provided to Medicaid recipients, in teacher-to-pupil ratios, and in the timeliness and quality of street repairs. - Costs of Service Innuts. Cities in which the cost of living is higher than average may have to pay higher salaries to attract qualified workers, pay more for construction or public housin,, 0 or provide higher benefit levels to public assistance recipients to provide a minimum standard of living. - Efficiencv of Service Provision- Inefficiencies resulting from surplus or unqualified employees, unproductive work processes, or higher salaries than necessary to attract qualified workers may result in higher costs of providing services in some cities. ‘District of Columbia Financial Responsibility and Management Assistance Authority, Toward a More Eauitable Relationshin: Structuring the District of Columbia’s State Functions (April 15, 1997). Phillip M. Dearborn and Carol S. Meyers, The Necessitv and Costs of District of Columbia Services, prepared for the District of Columbia Financial Responsibility and Management Assistance Authority, Greater Washington Research Center (Washington, D.C.: Jan. 1997). Carol O’Cleireacain, The Orphaned Capital: Adonting the Right Revenues for the District of Columbia (Washington, D.C.: Brookings Institution Press, 1996). 8 GAO/GGD-97-135R District of Columbia Revenues B-277191 AGENCY COMMENTS AND OUR EVALUATION We received written comments on a draft of this letter from the Executive Director of the District of Columbia Financial Responsibility and Management Assistance Author?@ (the Authority) and the Deputy Chief Financial Officer (CFO) of the District of Columbia. (See encs. II and III.) We also met with the Chief of the District’s Office of Economic and Tax Policy and staff on June 23, 1997, to discuss additional details concerning the Deputy CFO’s comments. Overall, both organizations expressed the opinion that our analysis was incomplete or flawed and expressed concerns that the results would be misinterpreted or were not comparable for the cities we studied. Their specific concerns and our responses are discussed below. A principal concern of both the Authority and the CFO’s Office was the limited scope of our analysis, in that we were asked to provide information on revenues alone. Both offices noted that questions about whether revenues are sufficient to meet expenditure needs in any jurisdiction cannot be addressed without additional information about other matters such as specific requirements for and costs of services, efficiency indicators, and other cost drivers for each of the selected cities. In addition, the CFO’s Office specifically cited additional characteristics such as poverty rates, the proportion of a metropolitan area’s population that lives within the city’s boundaries, and rates of change in cities’ populations as factors that influence revenue needs. They both recommended that we broaden the scope of our analysis to include such additional information to provide an appropriate perspective for the data reported. In making these observations, both offices were concerned that, despite our caution that our results should not be used to judge whether the District’s revenues are sufficient to meet its expenditure needs, readers would attempt to make that judgment. As our cautionary statements throughout this letter indicate, we too have been concerned that readers may attempt to draw conclusions regarding revenue sufficiency from our results. We were asked to compare per capita revenues across selected cities within a short time frame that did not permit us to analyze expenditure comparisons or revenue sufficiency. However, we had recognized from the outset that the factors such as those indicated by the Authority and the CFO’s Office may influence revenue needs and, for that reason, we had included a section in our draft letter that discusses the multiple factors that affect revenue needs. Both the Authority and the CFO’s Office questioned whether the assumptions we used to allocate nontransferred state revenues-proportional to the distribution of population or personal income-were the assumptions that should have been used. The Authority suggested that any discussion of allocated state revenues should be based on the state revenues and associated costs of state programs and services provided to the cities. The CFO’s Office produced a table (see enc. III) showing estimates of total revenue per capita GAO/GGD-97-135R District of Columbia Revenues B-277191 for each city if all nontransferred state revenues that were allocated using a different assumption, namely the “percent of total state poverty located in the city.” We agree that the results of an analysis such as the one we did will vary depending on the assumptions that are made. One of the reasons why we chose from the outset to use two different assumptions in our analysis was to illustrate this very point, and we emphasized this point throughout our letter. More specifically, we used the distribution of population and personal income as alternative assumptions (1) to make the point that different assumptions affect the revenue estimates for each city, (2) because we were asked to update our prior analysis, and (3) because we lmew that data on those distributions were readily available in the short time frame we had for this analysis. There is no widely accepted empirically-based set of assumptions for this kind of analysis. Moreover, we do not agree, however, that the poverty distribution offered by the CFO’s Office is necessarily a better choice as a basis for allocating nontransferred revenues. Based on an analysis of state spending for fiscal year 1991/92, we estimate that state revenues devoted to social services and income maintenance spending-revenues that the CFO’s Office suggest are allocated to cities based on the distribution of poverty-account for, on average, about 42 percent of the nontransferred revenues of the eight states in our analysis.g Thus other factors are likely to affect the distribution of the majority of nontransferred revenues. The Authority also recommended that we explain how federal transfers to states to fund programs such as Medicaid and AF’DC are reflected in our analysis, so that readers would be better able to judge the appropriateness of our comparisons. As stated in our letter, we adjusted for intergovernmental revenue flows. Specifically, all federal transfers to the state are included in the state revenues, both transferred and nontransferred, that we allocate to the cities. State funding for Medicaid (including federal funds channeled through the states) is included in the nontransferred portion of state revenue. The division of state funding for AFDC (including federal funds channeled through the states) between the direct transfer and nontransferred portions of state revenue in the data we used varies across the states. The CFO’s Office said that the data we used are not consistent across cities and, therefore, should not be used in raw form to compare cities. For example, they said that reported revenues for California and Pennsylvania include tuition paid to state universities while reported revenue for New York does not. In their oral comments, representatives from the CFO’s office offered another example of the lack of comparability of the data by ‘This figure is the total nontransferred spending on social services and income maintenance for the eight states as a percent of the total nontransferred spending of the eight states. 10 GAO/GGD-97-135R District of Columbia Revenues B-27719 1 observing that the Census data do not include special assessments as taxes. They noted that the District does not levy special assessments but other cities do. While we did not independently verify the accuracy of the Census data, we did explore the specific issues raised by the CFO’s Office with Census officials to determine whether these observations were evidence of a lack of comparability in the data we used. On the basis of that review, we do not believe these observations are supported. The Census Bureau, using the comprehensive annual financial reports and other financial statements supplied by cities and states, has developed a detailed system for categorizing state and local government revenues and applies this categorization consistently across all jurisdictions. Specifically, we were told that Census data do contain tuition paid to state universities in all states, including New York. Also, while the Census does not classify special assessments as taxes, it does include them in the miscellaneous revenue category, and thus they are reflected properly in our estimates. However, to clarify these matters in the letter, we have added more details to identify how certain revenue items are treated in the Census data and in our analysis. In a similar vein, the Authority was concerned that our results were not comparable because the school districts we included in our analysis may not overlap with the boundaries of the city/county governments, thus implying that for some cities we may have included revenue for students attending public schools outside the city/counties we studied or excluded revenue that should have been included. This did not occur in our analysis. Except in the case of Honolulu, we relied on Census coding to ensure that the school districts included in our analysis comprehensively covered students attending public schools within the counties that we compared and did not cover students attending schools outside those counties. For Honolulu, we had to estimate the city’s share of the revenues of the state’s single school district, which we did on the basis of its share of total student population. We have revised the letter to clarify our methodology for allocating school district revenues. Copies of this letter are being sent to the Ranking Minority Members of your Committees, the Mayor and the Chief Financial Officer of the District of Columbia, the Chairman of the District of Columbia Financial Responsibility and Management Assistance Authority, and other interested parties. We will also make copies available to others upon request. Please contact me at (202) 512-9110, or James Wozny at (202) 512-9084, if you or your James R. White Associate Director, Tax Policy and Administration Issues 11 GAO/GGD-97-135R District of Columbia Revenues ENCLOSURE I ENCLOSURE I REVENUES OF INDEPENDENT TOWNSHIPS AND LARGE SPECIAL DISTRICTS. 1991-92 1991-92 Revenues (dollars in Revenue City/County Independent Township or Special District millions)a per capitab i c C ( r t E , San Francisco 7 3 1 1 1 12 GAO/GGD-97-135R District of Columbia Revenues ENCLOSURE I ENCLOSURE I Special Service District of Philadelphia 6 4 Pennsylvania Convention Center Authority 11.3 7 Philadelphia Parking Authority 72.9 48 Delaware River Port Authority 117.0 Indianapolis Indianapolis Housing Authority 9.1 11 Indiana Municipal Power Agency 146.7 Indianapolis Utilities District 270.4 331 Indianapolis-Marion County Public Library District 18.1 Indianapolis-Marion County Building Authority 13.3 Lawrence Township High School 1 Independent smaller cities and towns 77.4 Jacksonville Beaches Public Hospital District 11.3 Independent smaller cities and towns 91.8 %pecial districts with total general revenue in 1991-92 of less than $1 million are not included. bRevenues per capita are given when special district service areas are coterminous or within a county. Census county population estimates are used to compute per capita amounts. Source: Bureau of the Census, Census of Governments, “Finances of Special Districts,” 1992 and special electronic Census of Government files “GIIND.DAT” and “FIN92lND.DAT.” 13 GAO/GGD-97-135R District of Columbia Revenues ENCLOSUREII ENCLOSURE II COMMENTS FROM THE DISTRICT OF COLUMBIA FINANCIAL RESPONSIBILJTY AND MANAGEMENT ASSISTANCE AUTHORITY District of Columbia Financial Responsibility and Management Assistance Authority Washington, D.C June 23,1997 Mr. James R. White Associate Director, Tax Policy and Administrative Issues United States General Accounting Office Washington, D.C. 20548 Dear h4r. White: Thank you for the opportunity to provide comments on the draft proposed correspondence entitled, “District of Columbia: Revenues Comnared With Those of Selected Cities.” - The D.C. Finsnciai Responsibility and Management Assistance Authority (“Authority”) affirms the admonition noted in the proposed correspondence, that the results should not be used to reach conclusions regarding whether the sufticiency of the District’s revenues to meet the District’s expenditures. This conciusion cannot be reached without a detailed anaiysis of the relative costs of the specific services snd actual costs of the services that each of the selected govemments provides and a comparison of those costs to the actual revenue base. Unfortunately, despite warnings to the contrary, the reader may attempt to draw a conclusion regarding the suEciency of District revenues based on the estimation of per capita revenue for. each government. To assist the reader in drawing an accumte conclusion, the Authority recommends that the analysis be broadened to include an examination of the tax rates, costs of labor, income rates, population, and other indicators of the revenue base of the sekcted cities. Since the District performs municipal, county, and state functions, the analysis in the proposed conclusion should include a detailed discussion of the revenues generated by states that are directly attributable to the selected governments. ‘Ibis discussion wilI present a clear picture of the revenue base of each of the selected governments. The Authority is also concerned that the assumptions used to compute the per capita revenue for each government may prevent a valid comparison. It is our understanding that to mitigate comparabiIity problems, school-district and state revenues were allocated to each city using two altemative approaches, population and personal income. We believe that these approaches do not provide a true estimate of the actual state revenue generated by the governments and these approaches do not adequately describe the actual revenue allocable to the sekcted governments. We note that the question of allocable state expenditures is not included in the proposed correspondence. Any discussion of allocated state revenues should be based on OneThomasCircic,N.W. . Stitc900 . Waskgto~D.CZC005 . (202)5@KMJ 14 GAO/GGD-97-135R District of Columbia Revenues ENCLOSUREII ENCLOSURE iI state generated revenues and the associated costs of state programs and services provided in the selected governments. Another possible problem is the use of school district revenues. Schooi district boundaries may not over lap with the boundaries of the selected governments and skew the results. Additionally, more analysis is necessary to explain the impact of Medicaid and AFDC on calculating the direct transfers from the state and the federal government. The revenue from these two federal aid programs flows to states not counties or cities. The federal contribution di&rs by state. Therefore, it must be explained how these revenues are reflected as transfers, in order to determine the appropriateness of comparisons with units of government that do not receive this type of federal assistance. Finally, the discussion of possible drivers of expenditures does not expiain the reasons for the District’s relative high per capita revenue rate. We would recommend that the analysis include an examination of the specific costs of municipal, county, and state services in each of the selected governments and compare the results to the District’s costs. Without discussing specific costs and adjusting for workloads, efficiency indicators, the presence of special districts and other specific cost drivers, the analysis cannot lead to an irrSormed policy discussion of the sufficiency of revenues and the revenue base that is necessary to support the optimal expenditure level. Again, we appreciate the opportunity to share our views on the proposed correspondence. If you have questions, please do not hesitate to contact my office. Sincerely, WEjrecutive obn W. Hill, Jr Director 15 GAO/GGD-97-135R District of Columbia Revenues ENCLOSURE III ENCLOSURE III COMMENTS FROM THE OFFICE OF THE CJ3IEF FINANCIAL OFFICER OF THE DISTRICT OF COLUMBIA GOVERNMENT OF THE DISTRICT OF COLUMBLA Office of the Chief Financial Of%icer *** Natwar M. Gandhi Deputy Chief Financial Officer Tax and Revenue June 23,1997 Mr. James R. White Associate Director, Tax Policy and Administration issues U.S. General Accounting OffLze Washington, DC 20548 Dear Mr White: On behalf of Anthony Williams, I am responding to your letter dated June 13,1997, requesting comments on your draft report entitled “District of Columbia: Revenues Compared with Those of Selected Cities.” As you know, the District of Columbia is engaged in the urgent task of bringing the District’s revenues and expenses into long-term balance. CIearly, we must take a hard look at any areaswhere District costs are out of line. However, we have serious disagreement with this report, which uses faulty data and methodology that exaggeratethe differences in revenue between the District of Columbia and eight other cities. Substantial revisions should be made before the report is released. The following are our concerns: 1. The report fails to recognize a primary function of state taxation-to generate revenue from better-off individuals (significantly concentrated in suburbs) and m-allocate funds according to need to worse-off individuals (significantly concentrated in inner-cities). This failure shows up in several places in the report: . Where state revenue is allocated either per capita or based on share of income, rather than based on a measurement of %eedn for each service supported by state expenditure. For example, by allocating Medicaid expenditures according to city population or income, the report assumesthat poverty is distributed in the same pattern as either population (but there are urban concentrations of the poor) or as income (which may well be m related to poverty). Poveay concentrations in the nine cities range from 8.9% (Honolulu) to 26.7% (Philadelphia) in 1993. e Where, by selecting only cities that have city/county governments, the report implies that the cities all have the same relative mix of urban and suburban populations. However, Jacksonville’s county includes 74% of the entire metropolitan area (and a 441 4th Street, N.W., Suite 4005, Washington, D.C 2&01 202/727-6083 16 GAO/GGD-97-135R District of Colnmbia Revenues ENCLOSURE III ENCLOSURE III Mr. James R. White Page 2 June 23,1997 significant part of the middle-dass population in the suburban ring) while the District only contains 14% of its metropolitan area population. When more population has private means,as in Jacksonville, there is less reason for higher per capita revenue distributions. 2. The data used in the report are not consistent across cities aad, therefore, should not he used in raw form to compare cities. For example, reported revenues for California and Pennsylvania include tuition paid to state universities, reported revenue for New York does not. 3. Because dynamic elements are omitted, the methodology of the report overstates revenue per capita in D.C. and understates revenue per capita in the other cities. Specifically, the District’s population declined sharply by 6.4% between 1990 and 1994, while the populations of Baltimore and Phiiadelphia declined less severeIy and the populations of the other cities increased. Because it takes time for governments at all levels to adjust expenditure patterns’, there is bias in the reported results. My staff has respondedin as much detail as possible, given the one-week time frame for a response. As your time is also limited, we would be pleased to work closely with you in addressing the points raised in the enclosed “Comment”. Please include these written comments in the report so they become part of the record and other users will have the benefit of our perspective. Thank you for the opportunity to review this drafi report. ax and Revenue Enciosure: Comment on GAO draft correspondence, “Diict of Columbia: Revenues Compared with Those of Selected Cities” cc: Anthony A. WiIliams ‘For example, populationdecfie lastyearis likely to be measured thii year, taken into account iu the subsequent budget process (next year), and to affect the budget the year sfter next- 17 GAO/GGD-97-135R District of Columbia Revenues ENCLOSURE III ENCLOSURE III REVIEW AND COMMENT: Draft Correspondence “District of Columbia: Revenues Compared with Those of Seiected Cities” Office of the Chief Financial Ofleer, Deputy CFO for Tar and Revenue June 20. 1997 The root difftcuky is that the “District of Columbia: Revenues Compared with Those of Selected Cities” (hereafter the “draft”) just scratches the surface of an area that already is the subject of a number of detaiIed studies. The draft, therefore, reflects very little from the lessons learned about making fair and accurate comparisons between the District of Columbia and other cities. It merely cites some of these studies in a footnote on p. 10 and lists without comment some of the reasons the topic is complex. we disclaimers.ost readdthe draft are iikelv to sm . . . . the&&&&~,,& Jhsa m anv other comnarable citv. and thr; ilaws i-the table -.d ~0 unnoticed. Specific issues with the draft are discussed below. 1Ahhough the discussion in the draft is stated in terms of revenues, it can just as well be viewed in terms of expenditures, since there is a virtuai equality between revenues and expenditures in cities. Unfortunately, there is no effort made to address in any serious way why the expenditures that need to be tinanced vary among cities. Ahhough it is essential to improve efficiency and . . reduce the cost of D.C. services, $ . . heaw w&lg.&&ted to its urban. 2. As the draft rem-s, a key methodological challenge of the analysis is how to allocate state revenues (including federal grants-in-aid made avaiiabie to the states) that are not directly paid to the city by the state. These indirectly aUocated revenues, which include most money for welfare, corrections, and higher education, are very substantial. These revenues account for well over half of all state revenues, and represent a substantial &action of the revenues calculated for many of the cities. For allocating these otherwise unaccounted for state revenues, the draft presents 2 altematives- proportionate shares of population and of income. However, both of these measures have substantial shortcomings because very large portions of the unallocated state spending are positively con&ted with neither of these bases. Indeed, income may be negatively correlated. The draft is especially flawed in faiImg to mention the traditionaI redisttibution role that state governments play in many service areas-namely, distributing resources from better-off areas (typically suburban areas), to cities or poorer rural areas where spending needs are high. An example of the shortcoming can be found in the Maryland welfare programs. Maryland spends from its tmaiIocated revenues more than 40% of its Medicaid and AFDC COSTSin Baltimore. Yet the GAO methodology allocates only about 14 percent on a per capita basis and even less on an income basis. Because these two programs constituted over $2.2 billion of 18 GAO/GGD-97-135R District of Columbia Revenues ENCLOSURE III ENCLOSURE III spending in 1995, the GAO understatement of Baltimore revenues is substantial. The differential between 14% and 40% of $2.2 billion is approximateiy $570 million, or more than $800 per person living in Baltimore. As mentioned above, the “right” way to allocate all of the state revenues (or expenditures) is a very complex task and we do not purport to have developed the complete answer in this short timename. However, with back-of-the-envelop estimates such as those contained in this report, and as a matter of professional integrity, every effort should be made to be sure that the ranges that are shown encompass the fitI1 range of plausible approaches. This is particularly true, given the strong possibility that those reading the report may draw important conclusions fkom it regarding the adequacy or lack thereof of D.C. revenues. I&&v. the ahocatton would be ma& based on e work-w for each &pe of state or federal exnendiue. e.9. the share of Jhc state’s uovertv DODUI~~~~Dfor welfarethe state, s um‘versitv-a her educati so forth. Accordingly, given the cleariy demonstrated redistributive role of state governments and of federal programs operated through the states, it would be entirely proper for the report to include at least one allocation formula based on need, such as the percent of the state’s poverty population that is located in the city. Based on information from the Bureau of the Census, the following table shows (1) the percentage of each city that is below the poverty line, and (2) the percentage of total state poverty that is located in that city. City % of poverty in the city % of total state poverty located in the city District of Columbia . 20.5% 100.0% New York I-24.6% I 61.1% I San Francisco 13.4% 1.8% Honoiuht 8.9% 68.2% Boston 19.0% 17.9% Philadelphia 26.7% 26.4% Baltimore 25.4% 34.9% Indianapolis 15.1% 18.1% Jacksonville 15.2% 4.9% Applying the percentages of state poverty to the unallocated amounts of state revenue used in the dr& results in an estimate that raises per capita revenues in some cities and lowers them in 2 19 GAOIGGD-97-135R District of Columbia Revenues ENCLOSUREIlI ENCLOSUREITI others. As a result, the ranking of cities changes, with New York higher than the District, and several other states get much closer to the District. The listing of per capita revenues under tbis assumption for the cities in the report is as follows: City Total revenue per capita if nontransferred state revenue is allocated on the basis of the incidence of poverty* Disnict of Columbia $8,286 New York 8,841 San Francisco 5,659 Honolulu 5,642 Boston 7,278 Philadelphia 6,777 Baltimore 7,652 Indianapolis 4,560 l Estimates make no adjustments for the dynamics of how population changes will tend to raise reported per capita revenue calculations in D.C. relative to other cities. D.C. population decliied 6.4% from 1990 to 1994, only two other cities (Philadelphia and Baltimore) dehned at ail (but by a smaller percentage), and the others all increased. .. . . 3. Bvsonfining~s over 500.000 whwe or no we between .. es It a-t true co- has been achieved. but -Indianapoiis and Jacksonville consist of merged cities and counties that encompass 64% and 74% of their metropolitan populations. In contra& the district represents only 13% of its . . metropolitan population. yith m of the better-off- the suburban w we would meet lower public expen&ures because- have pljvate mws to provide for manv t3-axbL -The proportion of city residents living in poverty varies considerably from city to city, as show in the first table above. Honolulu has only 8.9% of its population in poverty, while 26.7% 3 20 GAO/GGD-97-135R District of Columbia Revenues ENCLOSURE III ENCLOSURE III of Philadelphia’s residents are in this category. -Of the 9 cities that are compared, Census estimates show that only the District experienced a sharp (6.4%) decline in population from 1990 to 1994. Only two others (Philadelphia and Baltimore) experienced any decline at all, and these were smaller than D.C.‘s. (4.0% and 4.5% respectively). The other 6 cities all grew, 3 of them (Boston, Jacksonville, and Indianapolis) by more than 11%. Because all of the information in the report is shown on a per capita basis, the relatively sharp decline in D.C. population is likely to bias the results by tending to push per capita amounts relatively higher in the District. There is a long lead time in the budget formulation and execution process, and it also takes time for service levels to adjust to population changes. For example, if population declined in 1991, this likeiy was observed and measured in 1992, taken into account in the subsequent budget process in 1993, and reflected in the actual budget for 1994 or 1995, depending on when planning for the fiscal year takes place. 4. won in the reuort is not swized .- . therefore comoarisons should across aties. 6‘ the data are cle& ” Our experience, and that of others, has been that despite the best intentions of. the Bureau of the Census, it is diEcult to place full confidence in comparisons made with this data source. As au example, the census data counts college tuition in state universities as revenue for California and Pemsylvania, but not for New York. Further, the data in the report are confusing to follow. Are items such as special assekments, utility funds, retirement funding, public housing, highway funding, taxes on utilities and liquor, etc. given comparable treatment? In the time available, we were obviously not able to check into this. However, other researchers have advised us not to use these data for comparisons among jurisdictions. 5. For those using the report, it would be helpll if there were adequate tables in the appendices so that users couid have a better idea about the nature of the background data, how the data are researchedand “cleaned,” if necessary, and how the values in the tables are derived. 4 (268804) 21 GAO/GGD-97-135R District of Columbia Revenues Ordering Information The tist copy of each GAO report and testimony is free. Additional copies are $2 each. Orders should be sent to the following address, accompanied by a check or money order made out to the Superintendent of Documents, when necessary. VISA and Mastercard credit cards are accepted, also. Orders for 100 or more copies to be mailed to a single address are discounted 25 percent. Orders by mail: U.S. General Accounting Office P.O. Box 6015 Gaithersburg, MD 20884-6015 or visit: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW) U.S. General Accounting Office Washington, DC Orders may also be placed by calling (202) 512-6000 or by using fax number (301) 2534066, or TDD (301) 413-0006. Each day, GAO issues a list of newly available reports and testimony. To receive facsimile copies of the daily list or any list from the past 30 days, please call (202) 512-6000 using a touchtone phone. A recorded menu wilI provide information on how to obtain these lists. For information on how to access GAO reports on the INTERNET, send an e-mail message with “info” in the body to: info%vww.gao.gov or visit GAO’s World Wide Web Home Page at: http://www.gao.gov United States General Accounting Office Bulk Rat? Washington, D.C. 20548-0001 Postage& FeesPaid GAO 1 Permit No. G lOO 1 Official Business Penalty for Private Use $300 Address Correction Reauested
District of Columbia: Revenues Compared With Those of Selected Cities
Published by the Government Accountability Office on 1997-06-26.
Below is a raw (and likely hideous) rendition of the original report. (PDF)