the GAO


The purpose of this critique is to demonstrate quantitatively the inadequacy of the GAO's recent report on DC's 'structural imbalance. As an attempt to find a sound analytical basis for declaring that the nation's capital city deserves an annual federal subsidy of between $500 million and $1.2 billion, it fails completely. There are few if any urban jurisdictions that could not benefit from a substantial federal hand-out to make ends meet. But this report falls well short of proving that point, and, to its credit, the GAO stops well short of recommending such subsidization. The chapters that follow belabor this issue from several different standpoints:

o There is a substantial body of history behind this effort. The District's inability to quantitatively demonstrate the supposedly insurmountable burdens of hosting the nation's capital was noted in the GAO's interim report of September, 2002. As NARPAC's comments indicate, that led to additional pressures from the Congress and the city to "find other approaches". The GAO turned to some respected work commissioned and published a decade earlier, modified its focus and changed the prior definition of 'structural imbalance' in the process. This became the framework for the final GAO report recently released, and the subject of additional NARPAC Commentary.

o The pioneer background report on a "Representative (average) Expenditures System (RES)" was developed by Robert W. Rayfuse, Jr. and published by the Advisory Commission on Intergovernmental Relations (ACIR) in December 1990. It was a noble effort "to show how much it would cost each state to provide the national-average level of each service", as adjusted for specific circumstances, such as higher crime level, or lower percentage of kids in public schools, called "workload factors". This seminal effort acknowledged its shortcomings: the efficiency of carrying out various government functions is assumed to be average, and the associated performance achieved is simply not addressed. It certainly does not address the costs of being above average as a national icon.

o Together with equivalent techniques to assess average national capacity to generate revenues, these methodologies were intended to improve the measurement of relative state needs in the award of federal grants. The "revenue capacity" indicators had already progressed through four iterations: starting from Personal Income and/or Gross State Product per capita, to "Total Taxable Resources" (TTR) per capita, and then to a "Representative (average) Tax System (RTS)" also sponsored by the ACIR. From the carefully presented data tables and explanations in the Rayfuse report, these four estimates seldom vary (max-to-min) by more than 8% (except for DC and two other outliers). Furthermore the latest RTS approach varies from the average indicator by no more than plus-or-minus 4% (except for DC). But all these methodologies seek the perfect national average income from taxes. There is no suggestion of the ability to increase tax rates for preferred locations, such as a lobbyist's board room with an unobstructed view of the nation's capitol building. A schematic of this basic (but new) approach to estimating structural imbalance is shown below and will be elaborated in the sections that follow:

o Using this approach of matching the typical need to spend and the typical ability to tax, a significant share of the states will have financial outcomes either above or below average. Hence, in the original report, 28 states couldn't cut their own mustard, while 22 states and the District of Columbia were more than capable of paying their bills. In fact, using any of the four measures of revenue generating capacity, DC's surplus came in highest or second highest on a per capita basis. The new GAO report does indicate that, using their modified methodology, many states have much bigger total imbalances than DC. For instance, as shown bellow, the ten highest rollers now should get well over $50 billion in subsidies if the Federal Government awards $1 billion to DC.

o From this RES/RTS starting point, the GAO updates, adjusts and and applies the methodology producing average outcomes in which the District falls far short of its ability to pay its bills by from $470 to $9,783 per DC capita as of 2000. By comparison, DC had more than needed financial wherewithal to pay its way by anywhere from $785 to $4,455 per capita in 1987. This staggering shift in fortunes, despite the arrival of a fiscally conscious, if not conscientious, government, and the Federal assumption of $800M of DC's municipal/state bills in 1997, should have raised caution flags to any competent reviewer. Not so DC's CFO, who grandly concludes that the GAO's "innovative effort to adapt these methodologies is a significant achievement" and that the resulting comparisons "are as fair and objective as can be accomplished, given the unique nature of DC..." In fact, there are errors in judgment and, almost certainly, in arithmetic as well.

o Starting with the ACIR methodology, GAO develops new "workload factors" and revised revenue capacity indicators and arrives at these horrendous new shortfalls which, not surprisingly, are attractive to those seeking a substantial federal subsidy. It presents two expenditure alternatives, one based on national average government services spending, and a second one showing a worse imbalance, based on typical spending patterns of representative urban areas. Those more skeptical of the wisdom of increased dependence on federal largesse take a more skeptical view of both the details of the assumptions and the calculations made. Substituting NARPAC's "workload" assumptions , along with the statistical data that support them (completely missing in the GAO report), the imbalance can be made to disappear. The following table is elaborated in the sections which follow:

o Perhaps more telling, the GAO report does not compare any of its "should spend" estimates for DC with the "does spend" values contained in, say, the FY2000 DC budget. There are no such comparisons with budget "actuals" in the GAO work. When such comparisons are made, there is little if any correlation between "should spend" and "does spend". In addition, there appear to be some computational problems with GAO's revenue capacity analysisThis too is elaborated in the sections that follow.

o Despite the length and detail of this analysis, there are sevel items of unfinished analysis that deserve serious attention. The matter of DC's real capital investment deficit has not been critically analyzed in depth. And the nagging issues of how much it really costs the District (out of pocket) to host the national capital and the commuters that make it work are still up in the air, and appear to be getting worse! No appraisal of DC's real or imagined structural imbalance will be complete without vetting these matters. Since this report was prepared, NARPAC has taken a critical, albeit superficial, look at DC's capital improvement planning process, and the potential structural imbalance in DCPS facilities planning and finds that they both contribute to an unrealistic exaggeration of DC's out-year problems. A more thorough approach by DC's budget planners is in order before "crying wolf".

return to the top of the page HISTORICAL PERSPECTIVE

skip ahead to the GAO Interim Report

The concept that DC is categorically unable to raise the money to pay its own bills has been around for a long time. It first came to NARPAC's attention at about the time of its incorporation with a book entitled "DC by The Numbers", by Edmond and Keating, published in 1995, which showed literally hundreds of statistical parameters wherein DC ranked "worst" or "next-to-worst" in every conceivable category from fiscal indicators (taxes, revenues, and expenditures) and economic indicators (poverty level, aid recipients) to education, crime, and social indicators (single parents, teenage births, etc.). At that time, NARPAC focused on the difficulty of trying to compare an inner city with a large minority population to your average American state.

Two years later, a more authoritative book was published by the Brookings Institution Press, by Carol O'Cleirecain, entitled "The Orphaned Capital". She is a former New York City Finance Commissioner and Budget Director and close associate from Brookings of Alice Rivlin, (who soon thereafter became the chair of the DC Control Board which was obliged by the Congress to assume responsibility for DC's financial affairs). She concluded that DC's "long term fiscal problems stem largely from its very nature as the nation's capital", and that its "present revenue structure is not sustainable over the long term and will not ensure ongoing budget balance." The now-standard several reasons were explored in depth: the lack of a broader state revenue source that could funnel suburban tax revenues to the inner city; the need to perform "state functions" anyway; the failure to be able to tax federal properties to offset the costs of servicing them; and the Congressional diktat against taxing commuters. These led to the standard suggestions of continued annual federal payments in lieu of taxes; several new taxes recommended by a recent Tax Commission Study; and the always-ignored "potential for regional revenue solutions". Meanwhile, DC's new government succeeded in producing balanced budgets year after year.

Eventually, Alice Rivlin came out with her own Vision of DC's Future which was extensively reviewed by NARPAC as an attempt to get these issues debated in public, in NARPAC's Vision for DC. For several years now, NARPAC has found the arguments about DC's inability to pay its own bills to be largely specious, the result of a monumentally inefficient inner city governmental bureaucracy. The city has also done little to attract high-end tax-paying businesses and residents who are more than willing to pay a premium to be situated close to the seat of US and world power.

Along the way, other reports from accounting firms like McKinsey & Company, and KPMG Peat Marwick, and senior advisory organizations such as the Federal City Council had voiced concerns, although virtually all have been based on projections of current spending without regard for increased efficiency, and also without regard for quantifying the very dubious claims of budgetary pressures on the District from the Federal presence. These earlier efforts are summarized under NARPAC's heading of Official Studies. Eventually, the Brookings Institution's Greater Washington Research Center put out a report presenting the case for making it a federal responsibility to provide a sound footing for the nation's capital. NARPAC responded with a long rebuttal entitled Financial Structural Imbalance which will not be repeated here. It lays out a very substantial case for DC resolving its own fiscal problems.

It is of continuing fascination to NARPAC that on the one hand, DC's leaders continue to press for more control of its own political future, while at the same time demanding that the Federal Government provide larger and more permanent subsidies. As will be noted below, by the 1997 National Capital Revitalization Act, Congress assumed some $870M in "state functions" and pre- existing pensions, in return for discontinuing its (fixed) $670M annual subsidy. This has done little to satisfy the city's demand for more financial assistance.

In any event, these continued demands for more federal assistance, including strong pressures from DC's non-voting delegate, eventually led the Congress in the Spring of 2002 to ask its analytical arm, the Government Accounting Office, to assess the various elements of the District's reported fiscal structural imbalance. The GAO presented its preliminary findings to the Congress in an interim report dated September 4, 2002, was given revised guidance, and submitted its final report on May 22nd, 2003. NARPAC takes issue with many of these findings from several standpoints.

return to the top of the page THE INTERIM GAO RESPONSE

skip ahead to Narpac Commentary

NARPAC summarizes the interim GAO report entitled "DC: Fiscal Structural Balance Issues" by condensing and stringing together various excerpts from three key sections:

Letter of Transmittal:

The District of Columbia has historically faced many challenges due to its unique circumstances and role as the nation's capital. After several years of struggling with financial crises and insolvency in the early 1990's, the District has significantly improved its financial condition by achieving five consecutive balanced budgets, an upgraded bond rating, and unqualified, or "clean", opinions on its financial statements. More recently, however, District officials have sounded the alarm that the District faces an imbalance between its long-term expenditure needs for program services and capital investment, and its capacity to generate revenues over the long run. These officials assert that the District faces a fiscal structural imbalance as a result of several factors, many stemming from the federal government's presence in the city, the absence of a state to provide funding for the state-like services provided by the District, and restrictions on the District's tax base.

These GAO reports unfortunately do not contain their tasking memos from the Congress, but indicates that they were asked to provide information on the following:

o the District's definition of fiscal structural imbalance and its contributing factors; .the constraints on the District's revenue, including the prohibition of an income tax on nonresidents;

o the District's estimates of its spending requirements, including its cost estimates for providing services to the federal government and its spending for state-like functions;

o changes in the District's financial relationship with the federal government resulting from the National Capital and Self-Government Improvement Revitalization Act of 1997 (Revitalization Act); and

o alternative approaches to measuring structural imbalance in the District.

The information being presented in this report is based on GAO's work performed to date on this issue, but we have not yet completed the work necessary to conclude whether, or to what extent, a fiscal structural imbalance may exist in the District.

Results in Brief

Some of the more interesting comments in the summary of this preliminary report include:

o While federal tax exempt property does constitute a substantial amount of property in the District, the federal presence also draws substantial economic activity which provides the District with additional revenues from sales and income taxes generally not available to other cities of its size.

o On the spending side, the District's estimated costs associated with providing state-like services are not supported by detailed analysis and data, and are derived from cost allocation formulas based largely on the judgment of District officials. Moreover, while the District does have responsibilities similar to those of many states, it also has state-like types of revenues.

o Similarly, the District's estimates of its costs for providing services to the federal government lack detailed support, and do not consider the services provided by the federal government for its own property in the District.

o Perhaps most importantly, the District's estimates of its fiscal structural imbalance are premised on the maintenance of the existing level and costs of services now provided into the future. As a result, the estimates do not consider the potential for mitigating an imbalance with cost savings through management efficiencies, reassessing current policies, and restructuring of key programs that might bring the District's costs into line with those of comparable cities.

o The District received some federal relief through the 1997 Revitaljzation Act, which required the federal government to take over certain services in such areas as criminal justice, transferring their financing from DC taxpayers to the nation's taxpayers as a whole. In addition, the federal government assumed financial and administrative responsibilities for one of the District's largest fiscal burdens, which it inherited from the federal government as part of the transition to Home Rule in 1973-its unfunded pension liability for vested teachers, police, firefighters, and judges.

o Also, the federal government's share of the District's Medicaid payments was increased from 50 to 70 percent. At the same time, the Revitaljzation Act eliminated the federal government's annual payment to the District, which had reached $660 million per year. As a result of the above changes, the District estimates net financial benefits ranging from a low of $79.1 million to a high of $203 million per year during the period from 1998 through 2002.

o While the District's estimates point to many specific factors they do not constitute a comprehensive assessment of imbalances between expenditures and revenue capacity. The District has not performed the analysis to determine whether it has the capacity to provide a level of services comparable to those provided by other cities with similar needs and costs. As a practical matter, such an analysis is key to detern1ining the presence of an underlying structural imbalance in the District's finances.

o At the present time, however, comprehensive data and analysis are not readily available to say with confidence how the District's financial situation compares to that of other jurisdictions. Preliminary indications suggest that the District would have to sustain a high level of expenditures compared to other state and local areas to provide at least an average level of services after adjusting for its unique demographic profile (i.e., high poverty) and costs. However, when compared to other entities, previous studies suggest that the city also has among the highest revenue capacity, or the ability to raise revenues from its own sources, even accounting for the federally imposed constraints on the city's revenue-raising authority.

Importantly, the two sides of the equation need to be put together to address whether the District has the revenue capacity to provide for its unique workload and costs with an average tax burden. Such a comparative analysis would need to adjust for the fact that the District is not strictly comparable to any current jurisdiction in the nation, due to its unique combination of state and city functions and revenues and its role as the nation's capital. We are currently undertaking such an assessment and plan to report the results of our study in the future.


While it has made significant progress over the past several years, the District, similar to many other jurisdictions, continues to face a series of substantial, long-term challenges to its financial viability. Addressing these challenges requires continued dedicated leadership to make the difficult decisions and trade-offs among competing needs and priorities.

Presently, insufficient data or analysis exist to discern whether or to what extent the District is, in fact, facing a fiscal structural imbalance. On the revenue side, the District clearly has constraints in its ability to increase its tax base. However, the District's estimates of its possible fiscal structural imbalance have limitations and did not address the levels or costs of services for its citizens in the long term, whether such services could be supported by its present tax structure or tax base, or cost savings that can be achieved from management efficiencies. The available studies comparing revenue capacity and expenditures across jurisdictions are imprecise and some may not be applicable to the District.

As such, the Congress would benefit from more systematic information about the District as it considers proposals for addressing the fiscal structural imbalance that the District is currently asserting exists. A fundamental analysis of the District's underlying capacity to finance at least an average service level in relation to its needs can help determine if there is a fiscal structural imbalance. Such all analysis would provide a stronger foundation for decision makers at all levels to address the District's financial condition.

We currently have ongoing work in this area and plan to issue a future report with a more comprehensive analysis of the District's long-term financial condition. Therefore, we are not making any recommendations at this time.

return to the top of the page NARPAC Commentary

We draw three specific items from this interim report for different reasons. First the CFO's own estimate of the deficit in projected operating budgets is almost trivially small, and hardly a clarion call for federal financial intervention:

Second, without external coaxing, the GAO devotes a substantial portion of its interim report to the obvious inefficiencies in DC's government operations. Much of this flavor is suppressed in the final report. A cynic about government procedures might suspect that significant pressures were exerted on the Government's independent accounting office to turn on a green light and sell a green suit. In any event, the same kinds of data needed to generate the desired report can be used to highlight DC's basic problem.

NARPAC thinks the single scatter plot below depicts the real heart of the problem. Compared to the fifty states which comprise the "norm" for the total government work forces needed at all state and local levels, DC stands out as being uniquely inefficient even when adjusting for the evident trend line for increased workloads demanded by higher levels of poverty. There are 225 more DC government workers per 10,000 residents than the norm, 300 more if the US average is used, ignoring the impact of administering to the poor. At almost exactly $60,000 per employee (according to DC 2000 budget data by object classification), that equates to some $770 million DC dollars in the wind. To ignore this when building the case for a threatening structural imbalance, falls somewhat short of pure objectivity.

Third, the summary of financial burdens passed to the Federal Government are seldom pointed out despite their impact on some of the subsequent estimates of needed DC expenditures. This summary is provided in the interim report:

return to the top of the page THE FINAL GAO RESPONSE

skip ahead to NARPAC's comments

The final GAO report was submitted in May, 2003 to the Congress. As with the prior section, NARPAC has taken the opening sections of the final report and edited them down to what we believe are the essentials, and without editorial comment.

report purpose

At various times District officials have asked the Congress for additional funds and other measures to enhance revenues, based on budgetary shortfalls flowing from structural conditions that are beyond their ability to control. To help inform the debate, GAO was asked to:

o assess whether, or to what extent, DC faces a structural imbalance between its revenue capacity and the cost of providing residents and visitors with average levels of public services:

o identify significant constraints on DC's revenue capacity;

o examine cost conditions and management problems in key program areas; and

o study the effects of DC's fiscal situation on its ability to fund infrastructure projects and repay related debt.

defining structural imbalance:

Although there is no uniform definition of structural imbalance, there are two concepts that can be used to measure it: current services and representative services imbalances:

A current services imbalance addresses this question: if a jurisdiction were to maintain its current level of services into the future, would it be able to raise the revenues necessary to maintain that level of service under its current taxing policies? This type of longitudinal analysis compares a jurisdiction 's projected fiscal position with its current position and is independent of other similarly situated jurisdictions.

In contrast, a representative services imbalance addresses this question: if a jurisdiction were to provide a representative basket of public services at average efficiency, would it be able to generate sufficient revenues from its own taxable resources and federal grants to fund a representative basket of services if its resources were taxed at representative rates? This type of analysis uses a basket of services and tax structure typical of other jurisdictions with similar public service responsibilities as a benchmark against which to compare imbalances between the cost of providing public ices and revenue-raising capacity.

This approach attempts to compare differences in jurisdictions' fiscal positions under a common set of policies regarding levels of services and taxation. GAO employed a .representative services approach in performing this analysis. A fiscal system is said to have a structural balance if it is unable to finance an average ( or representative) level of services by taxing its funding capacity at average (or representative) rates. Because GAO defines structural imbalance in terms of comparisons to national averages, for any given period a significant proportion of all fiscal systems will have structural deficits.

Last year, GAO examined issues related to DC's reported structural imbalance and, in September 2002, concluded that DC had not provided sufficient data and analysis to determine whether, or to what extent, DC is in fact facing a structural imbalance between its revenue capacity and the cost of meeting its public service responsibilities.

Determining empirically whether DC has a structural imbalance a complex task that involves making judgments about (1) the appropriate set of governments to use when developing benchmarks for DC's spending and revenue capacity, (2) the influence that various workload and cost factors, such as the number of school-age children, and (3) the best way to measure revenue capacity. Given the lack of professional consensus and a limited empirical basis, GAO performed several sensitivity analyses to show how its estimates changed as it varied specific judgments and choices regarding key assumptions.


The precision of GAO's estimates is adversely affected by data limitations for various cost and tax bases. Consequently, uncertainty surrounds the specific numerical estimates GAO presents. Nevertheless, GAO believes that the consistency of its basic result over a broad range of alternative assumptions and approaches provides sufficient support for the concluding observations offered in this report. GAO's methodology was vetted among key experts, including individuals who designed the underlying methodology and District economists.


Determining the appropriate benchmarks for DC's spending is complicated by the fact that DC is a unique governmental entity. For this reason, GAO computed two separate sets of benchmarks--one based on a "state" services baskets, the mix of services typically provided by state fiscal systems ( each state and all of its local governments), and a second based on an "urban" services basket, the mix of services typically provided by governments in more densely populated areas.

To calculate the cost of providing a representative level of public service GAO used the national average per capita spending for each expenditure function as a benchmark. GAO adjusted for differences in workloads ( e.g., number of school-age children) across states. GAO also adjusted for the fact that the private sector wage rate varies across states.

GAO did not adjust for differences in preferences or policy decisions across states, nor did it adjust for differing degrees of efficiency in providing services. Governments that are relatively inefficient would have to spend more than the average amount to provide an average leveJ services. In addition, GAO made no adjustments for the unique public service costs associated with DC being the nation's capital.

revenue capacity:

To estimate the total revenue capacity of each state fiscal system, GAO combined estimates for the two principal sources from which those systems finance their expenditures: (1) revenues that could be raised from each system's own economic base ("own source revenue') and (2) the federal grants that each system wowd receive if it provided an average basket of services. In the past, two basie approaches have been employed to estimate the own source revenue capacity of states: (1) those that use income to measure the ability of governments to fund public services and (2) those that attempt measure the amount of revenue that could be raised in each state if an average set of tax rates were applied to a specified set of statutory tax bases "typically" used to fund public services. Because experts do not agree as to which approach is superior, GAO computed separate results using both methodologies

. No consensus exists regarding the "best" approach to estimating structural imbalance. Consequently, GAO performed several sensitivity analyses to show how its estimates changed as it varied specific judgments and choices regarding key assumptions. The consistency of GAO's basic result over a broad range of alternative assumptions led GAO to conclude that DC does have a substantial structural imbalance, even though considerable uncertainty exists regarding its exact size.

While DC's revenue capacity per capita is large relative to those of most state fiscal systems, it would be even larger in the absence of several existing constraints on DC's tax authority. Despite these revenue constraints, the per capita revenue capacities of DC's income and property taxes are higher than those for all but a few state fiscal systems, partly reflecting the indirect benefits of the federal presence for DC's economy.

GAO's quantitative analysis was not able to account for all special circumstances beyond the control of DC, such as the high cost of special education services, and extra police and fire services associated with the federal presence, including those for political demonstrations. In recognition of DC's high-cost environment, the federal government provides certain supplemental financial support to DC, such as an enhanced federal share of DC's spending on Medicaid.

Significant and costly management problems, mostly under DC's authority to control, further increase spending unnecessarily in Medicaid, elementary and secondary education, and police and fire protection. Various reports have estimated wasted resources to be at least in the tens of millions of dollars. By addressing such management challenges, DC could free up local funds and possibly gain additional federal funds for use in increasing the levels of services to its residents and closing its current budget gap.

However, addressing these management problems will not offset DC's underlying structural imbalance, which is due to factors outside its direct control. In recognition of DC's management problems, the federal government provides DC with special technical assistance.

While capital spending has increased in recent years, DC continues to defer infrastructure improvements because of constraints in its operating budget. Most of DC's infrastructure and capital improvement projects are financed by using general obligation bonds. The interest and principal payments (debt service) on those bonds are paid from DC's operating budget. Although DC is not close to its legal debt limit, it cannot take on additional debt without cutting services or raising taxes that are already higher than other jurisdictions.

Contributing to DC's difficulties is its legacy of deteriorated infrastructure and its responsibility for funding its 40 percent share of the metropolitan area's mass transit system. DC is attempting to address its backlog of infrastructure projects, but it DC continues to defer major infrastructure and capital investment in part because of its structural imbalance.

principal findings:

Using other state fiscal systems as a benchmark, GAO's analysis indicates that the cost of delivering an average level of services per capita in DC exceeds that of the average state fiscal system by approximately 75 percent ( or a total of $2.3 billion more annually than if it faced average cost circumstances) and is over a third more than the second highest cost fiscal system, New York. If state fiscal systems were to provide a basket of services typically provided in more densely populated urban areas, GAO estimated that DC would have to spend over 85 percent more (or total of $2.6 billion more annually) than average to fund an average level of services.

cost of key services

DC faces high cost circumstances, largely beyond its control, in key program areas, including Medicaid, elementary and secondary education, and police and fire services, that increase the fiscal burdens on its budget.

For Medicaid, GAO estimated that DC should spend well over twice the national average per capita. To provide an average level of services DC would have to spend a total of $437 million more than if it faced average cost circumstances.

Similarly, GAO estimated that DC's per capita cost of elementary and secondary education is 18 percent above the average state fiscal system, due to DC's disproportionately high percentage of low-income children. To provide an average level of services DC would have to spend a total of about $136 million more than if it faced average cost circumstances.

Likewise, for police and fire services, DC's per capita costs of providing an average level of services are well over twice the national average. To provide an average level of services DC would have to spend about $480 million more than if it faced average cost circumstances.

revenue capacity

GAO's analysis indicated that DC's per capita total revenue and own-source revenue capacities are higher than those of all but a few state fiscal systems. Its capacity is high even though DC faces some significant constraints on its taxing authority, such as the inability to tax federal property or the income of nonresidents who work in DC.

The two estimation approaches (RTS and TTR) GAO used to measure DC's revenue capacity yielded the same basic result: DC's own-source revenue capacity per capita ranked among the top five when compared to those of the 50 state fiscal systems. This high own-source revenue capacity, combined with federal grant funding that is over 2.5 times the national average, gives DC a higher total revenue capacity than any other state fiscal system. DC's total revenue capacity ranged from 47 percent above the national average (based on a conservative version of the RTS approach) to 60 percent above (based on the TTR approach).

structural imbalance

Using a representative services analysis GAO found that DC faces a structural deficit in the sense that the cost of providing an average level of public services exceeds the amount of revenue it could raise by applying average tax rates.

Data limitations and uncertainties surrounding key assumptions in GAO's analysis made it difficult to determine the exact size of DC's structural deficit. GAO obtained its lowest deficit estimate of about $470 million per year by combining its lowest (state-based) estimate of DC's costs with its highest revenue capacity estimate (TTR), and its highest deficit estimate of over $1.1 billion by combining its urban-based spending estimate and its RTS revenue capacity.

Among the contributing factors to the structural imbalance are high cost conditions largely beyond DC's control, such as high poverty rates. In addition to having a high revenue capacity, DC also imposes above-average tax rates; however, high taxes are only sufficient to fund an average level of services.

Because of its high tax rates, actual revenues collected by DC exceeded GAO's lower estimate of its own-source revenue capacity by 33 percent and exceeded GAO's higher estimate of that capacity by 18 percent. However, DC's actual fiscal year 2000 spending was only equal to the cost of an average level of public services, based on the basket of services provided by the average state fiscal system.

management inefficiencies:

Using the urban basket of services typically provided by urban governments as a benchmark, DC's spending is 5 percent below that needed to fund an average level of services. To the extent that DC does not deliver services with the assumed average efficiency, its actual-level of services may be below average. In three key program areas (Medicaid, elementary and secondary education, and police and fire services), GAO identified significant management problems, such as inadequate financial management, billing systems, and internal controls.

While DC has taken some actions to correct management inefficiencies, more improvements are needed. In the case of Medicaid, in fiscal year 2001 DC had to write off over $78 million for several years worth of unreimbursed claims for federal Medicaid matching funds. In the case of education, District officials were not able to track either the total number of employees or whether particular positions were still available or had been filled. In the case of police and fire services, DC does not adequately track the costs it incurs to support the federal presence. This hinders its ability to make a case for additional federal reimbursement, requiring it to spend more of its own resources to support the federal presence.

infrastructure maintenance:

Although DC is making some attempts to address its backlog of infrastructure projects, it has nonetheless continued to defer significant amounts of infrastructure projects because of constraints in its operating budget. From 1995 to 2002, DC's outstanding general obligation debt has changed little, totaling $2.67 billion as of September 30, 2002.. DC's projections assume that debt service costs will increase at a higher rate than local revenues. When compared to combined state and local debt across the 50 states, DC's debt ranks as the highest in the nation both per capita and as a percentage of own-source revenue.

concluding observations:

Due to a combination of its significant management problems and its substantial structural deficit, DC is likely providing a below-average level of services even though its tax burden is among the highest in the nation. In all likelihood, cutting an already low level of services to residents as well as businesses and visitors, which could also have undesirable consequences for DC's economy.

An alternative to raising taxes or cutting services would be for District officials to continue deferring improvements to its capital infrastructure. This strategy is not viable in the long run.

District officials could address a budget gap by taking actions such as cutting spending, raising taxes, and improving management efficiencies. In contrast, a structural imbalance is largely beyond District officials' direct control. Without changes in the underlying factors driving expenses and revenue capacity, the structural imbalance will remain.

In the near term it may be necessary to change federal policies to expand DC's tax base or to provide additional financial support. Given the existence of structural imbalances in other jurisdictions and DC's significant management problems, federal policymakers face difficult choices regarding what changes, if any, they should make in their financial relationship with DC.

Federal policymakers could choose not to address DC's structural imbalance and require local officials to deal with the difficult choices it faces to meet its obligations. This approach recognizes that other jurisdictions also face substantial structural deficits and local officials are in the best position to decide the most effective means of balancing needed trade-offs.

Alternatively, additional federal assistance (beyond the high level already provided) for DC could compensate for its structural imbalance. However, this assistance might suggest that officials of other fiscal systems with sizable structural imbalances, would have equally strong claims on additional federal assistance. Nevertheless, by virtue of being the nation's capital, justification may exist for a greater role by the federal government to help DC maintain fiscal balance.

This strategy is not without its own risks. Significant management problems in DC mean that the aid provided could result in more wasteful spending or in DC simply postponing many management reforms. Given its management challenges, it is important that DC achieve basic management performance and accountability standards to ensure an efficient use of any resources.

return to the top of the page NARPAC Commentary on Final Report

If one accepts as incontrovertible that the quantitative aspects of this report are properly and conservatively developed (even if not made transparent in the report), then the qualitative material summarized in the previous sections can be accepted as a very good summary of the financial situation facing the nation's capital, particularly with regards the potential problems in DC's future operating budgets.

On the other hand, if the reader concludes, as NARPAC has, that the unshown quantitative analysis and statistic data are seriously flawed, then the findings of the report seriously misrepresent the case for believing that DC's present circumstances are beyond their own control. This leads to a set of completely different conclusions, such as:

o DC's financial problems do not produce a structural deficit beyond its own control;

o Many of DC's current fiscal and demographic circumstances can be brought within its long-term control, including reducing its unconscionable poverty and illiteracy rates;

o Management inefficiencies are not dwarfed by an inescapable need for a federal bailout;

o DC's claims of deferred infrastructure are as likely over-inflated as its inability to pay its own bills;

o There is no reason to settle for a nation's capital with average performance, average tax rates, or a bigger than average federal subsidy;

o There are many healthier ways by which the Federal Government could assure DC's financial independence;

Unfortunately, NARPAC is in no position to make such claims without providing quantitative justification for its alternative views, particularly when choosing to challenge a respected and authoritative federal organization such as the Government Accounting Office. These alternative views are also based on four fundamental observations:

balanced budgets are good:

First, NARPAC fully accepts the basic approach of trying to establish a stable balance between what city services "should cost" and what the city "can afford". It is starkly inappropriate for DC to depend on the Federal Government for its fiscal solvency. Similar techniques are used wherever long-term budgets must be met, even if the terminology varies. "Representative Expenditures" is a nine syllable synonym for "should cost", and "Representative Revenue Capacity" is a twelve syllable alternative to "affordability".

being average isn't our national capital city's objective:

Second, however, the notion of "representativeness", or its simpler equivalent "averageness", is more difficult to accept at face value as a primary criterion. While it may be abundantly clear when handing out federal grants (the motivation behind these methodologies), that each jurisdiction should be treated with totally non-preferential equality, there should be absolutely nothing average about the capital city of the United States of America. Keeping it from being average (or less) is the fundamental reason for NARPAC's existence.

Americans have every right to expect their nation's capital to have better than average schools, better than average life expectancy (particularly from gun shot wounds and sexually transmitted diseases), better than average neighborhoods, and better than average businesses. Moreover, Americans should expect that those who live here (it is by choice, not necessity) will pay higher than average taxes for season tickets for the nation capital's box seats. If nothing else, Americans should make sure that their capital city betters their national and urban norms.

Ringside, orchestra, and box seats are all purchased at premium prices, and are no bargain for season tickets. For once, NARPAC has not collected statistics on the fair price of prestige, but considers it noteworthy that the average cost of office space inside DC now exceeds $40 per square foot, while the average cost of office space throughout the metro area is somewhat over $25. That represents a 60% 'mark-up' over the average, even though the suburbs have enough empty space to absorb one-third of all DC's office workers. In fact, DC's vacancy rate of 8.8% (July '03) is very much lower than the 15.0% being experienced by the Maryland and Virginia suburbs. DC must surely have a premium revenue capacity.

apples and oranges don't average:

Third, it is virtually senseless to compare statistics of large rural American jurisdictions with small urban areas, more crudely labeled "inner cities", with population densities weill overt 100 times the national 'average' and minority populations almost three time the national average. Unfortunately, comparable statistics are not gathered for central cities, nor are they generally expected to be financially self-sufficient entities. Equally unfortunate, America does not yet recognize that metro areas, which occupy only 20% of America's land, but are home to 80% of its people, are the most practical financially self-sufficient jurisdictional entities. In any event, testing inner cities by national averages is likely to be an exercise in futility.

performance may well be unrelated to cost:

Fourth, it must be continuously repeated (as do ACIR and GAO) that measures of average taxing and spending are in no way related to any performance measures such as student reading skills or police crime-solving prowess. Although the averages may provide some indication of relative effectiveness (in performance) or efficiency (in spending), they are often almost irrelevant in measuring performance. This is explored in other parts of this web site dealing with education, poverty, and ignorance (another of NARPAC's favorite topics). In another of NARPAC's key areas, for instance, meeting or exceeding average welfare payments connotes absolutely nothing about success rates in reducing welfare rolls. Equally important in the case of the nation's capital, many costs for businesses and residents are simply "non-elastic". For those few who pay the majority of DC's taxes, there are real and imagined benefits associated with a Washington, DC address and postmark that are not influenced by rental or tax rates.

The following sections will try to make this alternative quantitative case, at the expense of considerable detail, hopefully made slightly more interesting by extensive graphics. We will deal first with the matter of what government services "should cost" and then with the counterbalance of what governments can afford to generate in revenues. We will try to use the resulting averages where they are useful in illuminating circumstances, but to avoid applying them blindly.

return to the top of the page THE BACKGROUND RAYFUSE/ACIR REPORT

skip ahead to applying the RES approach

The Representative Expenditures System (RES):

The RES methodology was developed by Robert Rayfuse under a commission from the Advisory Commission on Intergovernmental Relations in the late '80s. It was intended to provide the ying to balance four closely related yangs previously developed to provide federal standards for states' revenue-raising capability. These are key to establishing relative needs for federal grants and will be discussed subsequently. By comparison, the RES system is relatively untested, and is best described by the 'Rayfuse Report' entitled "Representative Expenditures: Addressing the Neglected Dimension of Fiscal Capacity" a remarkably apt description of its objective.

The RES system essentially develops techniques for estimating the extent to which any one jurisdiction should depart from the national average cost of providing a particular major government service. As described in the GAO report, for instance, the average per capita cost of providing primary and secondary education across the US in 2000 was $1338. (NARPAC finds $1298). This is up from $614 in 1987, according to the Rayfuse report.

This national average is then modified by three fundamental "workload" variables in any particular surrounding. One is the relative number of school kids per capita, and the second is the relative number of kids living in poverty, and the third is the relative wage scale in different locales. Hence the representative cost of public education is modified to fit known individual jurisdictional circumstances, but the sum total of all these variations will even out to match the total national expenditures recorded in various census documents on government finances. In this case, GAO determined that DC's spending should be 18% above average, which amounts to $1576 based on national statistics, or $1645 based on a smaller urban statistical sample, but using the same national scale-up factor (more about this later).

It should be noted that the RES system does not require any reference to current spending by the jurisdiction is question. In fact, the current GAO report provides no comparisons between their findings with what DC is currently spending. NARPAC finds it hard to believe that this would not have provided a real-world "sanity check" on the credibility of their calculations. In the case of education, Census records indicate DC spent $1554 per capita (not per kid) in 2000. In this case, DC is judged to be spending just about the right amount on educating its kids, according to this methodology, not a popular position with the exception of NARPAC!

By applying these correction factors to the national average per capita costs across a "basket of services" the "should cost" value for any jurisdiction's budget can be determined, and subsequently compared to the equivalent representative revenue capacity measures also based on various average national statistics. That basket includes: elementary and secondary education; higher education; public welfare; health and hospitals; highways; public safety; and "all other".

In NARPAC's view, the credibility of the GAO conclusions rests almost solely on two of their estimates: the "should cost" number for public safety, and the "can afford" number related to DC's tax base. The public safety issue is unfortunately the most complex, with several different workload factors and wage scales to consider. It is also a classic case of pyramiding estimating factors which combine to yield an answer which defies rational explanation. To NARPAC's analytical dismay, it adds fuel to the fire of those who claim "you can prove anything with numbers". In the detailed explanation that follows, there are four separate threads that must be followed, because there are major issues with each:

o are the chosen key parameters and their assigned weighting factors appropriate?

o is the rationale for determining the off-average "workloads" transparent and reasonable?

o is the scale-up factor for local wage rates reasonable (and consistent)?

o how should the off-average "workloads" be adjusted when shifting from a "national" to an "urban" base?


Base Case - National Baseline

As mentioned above the most significant term in DC's "should spend" basket of services deals with public safety, and the following paragraphs and charts deal with this subject in detail. Here is an attempt to explain each step as briefly as possible. The Rayfuse and GAO reports start from the three applicable budget line items: police forces, the courts and corrections of the justice system, and fire and emergency services (FEMS). An average expected cost is developed separately for each, as discussed below, and summarized on the accompanying charts::

a) police costs are based on three separate aspects of police work. As devised by Rayfuse and accepted by GAO, and interpreted for simplicity by NARPAC, the average police force will split its efforts equally between three functions driven by different "average workloads". Hence each has a "weighting factor" of 0.33. One third of police costs will driven by the total population served; the second third by the emphasis on violent crimes, as typified by the average murder rate; and one third by the crimes committed by young residents between 18 and 23 years old. According to GAO calculations for their base case, the following "workload factors" need to be applied when comparing DC's circumstances to the national average. DC's "effective population" is increased by the daily influx of commuters by 19% (though this is not compared to any national average). GAO refers to this as the "daytime population". DC's murder rate is 7.56 times the national average, and there are 32% more persons in the 18-23yr old category than the national average. b) justice systems costs are split equally among the same three weighting factors and driven by the same factors as above, except that commuters are assumed not to contribute to DC's court costs and corrections population (Hence no 1.19 workload factor for the population term.)

c) FEMS costs will be driven by different factors, unevenly weighted. In this case 40% of the Fire and Emergency Services is devoted to the "effective" population (hence a 1.19 workload factor), 50% to apartment buildings with more than five units, and 10% to buildings more than 60 years old. Compared to national averages, GAO estimates that DC has 3.32 times as many multiple unit apartment houses, and 2.69 times as many of its buildings are over 60 years old. These factors significantly raise the "should cost" of DC's FEMS, though the effort clearly focuses on the "F" and not the "EMS".

To these scale-ups must also be added relative wage costs compared to the national average. GAO computes that police should be paid 41.5% more, corrections personnel 33.5% more, and FEMS personnel 38% more than the national average as shown below:

These factors are all multiplied together and summed for each separate cost element, to provide a total "mark-up" (NARPAC term) for each cost component from the national average baseline. Hence DC "should spend" on its police force 4.75 times as much as the national average of $116M for its 572,000 people; 4.40 times the representative $99M for corrections, and 3.31 times as much as the national average of $47M. Note these numbers are provided in dollars for DC's population rather than the "per capita" values often used in RES. This is solely for the purpose of making them relate more easily to actual DC budget numbers.

In short, and as is shown on the chart above, GAO asserts that using this approach, DC should be spending $1,143M on its public safety, i.e, 4.36 times the national average of $262M, and 69% more than the $676M DC actually budgeted in 2000. Similar calculations are made for other major budget elements, but this is the only one that by itself generates a theoretical 'deficit' in DC's ability to pay its bills of $467 million.

Alternative Case Urban Baseline

GAO recognized that the differences between the spending patterns of state and local governments combined could be significantly different that for DC, which is essentially an inner city with no accompanying counties or state. They therefore turned to look for equivalent data sources, both budgetary and "workload", from more urban sources, settling on a limited Census 2000 data base for the "Top 25 US Counties and Independent Cities (ICs) with the densest populations". These populations vary from 66,700 per square mile in New York City, to a mere 3,625 in Denver. They then ruled out those with populations less than 250,000 a step which, interestingly enough, eliminated Arlington County, and the independent cities of Alexandria, Falls Church, Manassas Park, and Charlotteville, all in neighboring Virginia.

The GAO urban sample boiled down to eleven big cities including New York, Boston, San Francisco, Philadelphia, Baltimore, Washington, DC, Chicago, St. Louis, Norfolk, Milwaukee and Denver, plus five New York counties outside of New York City (Kings, Bronx, Queens, Richmond, and Nassau), and four more in New Jersey (Hudson, Essex, Union, and Bergen). The population of this sample is 23.6 million people, 12.2 million in and around New York City, and 11.4 million around the country.

From Census 2000 financial data beyond the ready reach (or grasp) of NARPAC, the GAO calculated that the new sample averagecost of a police force for a city of 572,000 would be $205.9M, up 78% from the national sample, the average cost for justice/corrections would decrease 30 % to $69.2M, due to the smaller prisoner load at the local level, and the average FEMS cost would rise 75% to $82.4M, primarily due, one would guess, to living 10,000 to the square mile instead of less that 80 per square mile for the country as a whole. These values are used in the companion table below:

To NARPAC's surprise and consternation, however, although the average costs have been adjusted to reflect the higher density urban circumstances, GAO retains the same workload and wage factors referenced to the national baseline. The arithmetic impact is, of course, very large indeed. With the urban average public safety cost increased $95 million, and the 4.36 "mark-up factor" still in place, the proposed public safety RES cost rises an additional $413M to $1.656 billion, just about 2.5 times ($980M) more than DC is currently spending! The charts below show just how far off the GAO RES estimates would be from comparative realities. The first one presents Census data of "police protection" costs in 20 cities in 13 different states, showing the relative costs as captured at state and local levels. Note that the city level is about 22% higher. The three blue lines at the bottom show what DC is actually spending now, and what would comprise "representative expenditures" with GAO' state (national) or urban baselines. The insert indicates that while NARPAC agrees with GAO on the national baseline, their 78% increase to the urban case is not confirmed by NARPAC's (different) sources:

The second comparison is very similar. In this case total public safety costs are compared by component between 14 states, and current DC costs are already 2.6 times their average, and grow far more under gao assumptions. Note from the insert that a large contributor to these differences is GAO' assumed need for substantial wage increases to attract and maintain suitable personnel. NARPAC sees no need for more than a nominal increase from current levels:

Moreover, it is not difficult to find summary data that demonstrates the potential change in the workload factor based on different baselines. In the simple example below, drawn from the Census 2000 Statistical Abstract, and based on total violent crime data (not murder rates), DC's violent crime rate is 3.6 times as high as those US states with less than 100 persons per square mile, 3.0 times the US total, 2.75 times the average for more densely populated states, 2.35 for 35 large counties, and 1.55 for 35 large cities. Similar data are available from FBI UCR reports which show DC's murder rate to be 10.7 times that of 2092 rural counties, but only 3.2 times the average of 68 large cities. These figures translate to 7.1 times more violent crimes for DC than rural counties, but only 1.4 times those of major cities. Other comparisons in shifting workloads with changing baselines are shown below:

NARPAC reluctantly concludes that this excursion should be eliminated from consideration in trying to conjure up any reasonable 'structural imbalance'. This reduces the maximum estimated structural deficit by $413M.

return to the top of the page NARPAC'S OWN ANALYSIS

While the RES/RTS approach may yield interesting insights into the order of magnitude of DC's potential structural imbalance, it surely needs to avoid using input factors with little relation to reality. Analysts also need to check the methodology with "what if" changes to avoid excessive sensitivity. The two "click-up" tables of calculations shown above are actually on spreadsheets in which any input number(s) can be changed, and the impact registered immediately. The illustrative changes tallied below are made relative to the GAO Base Case table above, and are each one-at-a- time changes from baseline, not cumulative:

But it is also possible to look at real-world statistics and assess how realistic some of these judgments are. NARPAC cites six cases as examples:

1. overall use of police manhours and arrests

Not counting traffic violations, US police make well over ten million arrests a year in the course of their duties. The chart to the left here shows the distribution of some nine million arrests made by 551,000 police personnel in over 9000 police jurisdictions which cover some 182 million of the 281 million Americans. Of those nine million arrests, some 1.5 million are considered for "serious crimes". These in turn are divided between about 1.1 million property crimes such as larceny, burglary and vehicle theft, and 400-odd thousand violent crimes including aggravated assault, robbery, rape and murder. The tip of the iceberg is some 8700 murders, less than one-tenth of one percent of the total. Even if they involved 100 times the effort of other more numerous crimes, they would consume less than 10% of the "workload".

The question is whether murders themselves are a good indication of the police workload across a wider variety of serious crimes. It was originally argued by Rayfuse that murders had the distinct advantage of being properly counted, whereas lesser crimes might or might not get reported. But it is difficult to imagine that the broader-based crime reports are as uneven a measure as the homicides themselves. In DC, for instance, there are about 36 violent crimes per murder. In the 68 largest cities in the US, the average is 83 violent crimes per murder, and in the 6900 small towns the average is 121. This scatter is exacerbated by the fact that black Americans are 6.8 times as likely to kill or be killed than white Americans (the vast majority of murders are within their own race: 85% for whites, 93% for blacks), and 8.3 times as likely to be involved in robbery. To NARPAC it seems that the use of violent crimes themselves may be a better workload indicator. These are tabulated on the chart below:

Furthermore, the relationship between violent crimes and police force levels is less than stellar. In small towns there are about 1.2 violent crimes per police person, while in big cities there are about 2.7, and in DC somewhere between 2.0 and 1.3 depending on whether the 3000-odd federal police forces on duty in DC are counted along with the 4300-odd DC police.

Finally, it is clear that young adults 18 to 24 years old do commit more than their share of crimes resulting in arrests (see above). It is not clear, however that they commit more serious crimes, as opposed to a larger share of the lesser crimes. As shown to the left, NARPAC found no particular correlation between number of youths and the types of crimes across a sample of cities. Nevertheless, they do add disproportionately to the overall workload, and are involved in a third of all US arrests, as shown below.

It seems quite appropriate to use this indicator as one contribution to police workload, but probably weighted somewhat less than a full third of all police activity. Moreover, this is yet another area where the baseline is key. While DC may be home to 32% more 18-24 yr-olds than the national average, it is only 9% above the (non-weighted) average for 20 cities, and actually 11% below the average for 20 metro areas. Lowering the "mark-up" for the higher youth population from 1.32 to 1.1 reduces the "structural imbalance" by $22M.

NARPAC has spent considerable effort trying to find those parameters that might provide respectable correlation between cause and effect in crime rates with little success. As shown on the four charts below, there is some relationship between typical police spending across a twenty- city sample and either violent or total crime rates (left hand charts below). In both of those charts, DC's police spending seems almost twice as high per capita as it should be. On the other hand, when we venture into the touchy area of violent or total crime rates vs the percent of those cities' populations that are black or Hispanic, there is a significant (albeit far from precise) trend line, and DC falls well within the normal scatter. In other words, DC's crime rates are not abnormal for cities with large minority populations.

2. the impact of race and/or poverty on crime

The FBI's Uniform Crime Reports data make it possible to tabulate the number of arrests in all the major crime categories by black or white alone, and in this case, the numbers tell a pretty grim story. Blacks comprise some 12.3% of the population but are arrested for 27.9% of the crimes in all categories, with their most disproportionate share (37.8%) in the violent crime categories of murder, forcible rape, robbery, and aggravated assault. These are shown by category and in total on the upper set of charts below, comparing relative crime rates for blacks, non-blacks and the overall total population:

As always, however, it is educational to look further to see what other parameters are in play. In this case, It seemed to appropriate to explore those same crime rates based on poverty levels by race, both because the national poverty rate for blacks is just about 2.5 times that for non-blacks, and because there is plenty of crime in cities with very few blacks. The lower bar charts above show that poverty is almost certainly a better overall indicator of criminal behavior than skin color, since the black/non-black crime rates based on poverty are only modestly different.

NARPAC then returned to the overall scatter charts used above in the hopes of showing better correlation. To its surprise, the correlation between poverty rate and city crime levels turned out to be as non-existent as those for urban density and for youth involvement. Further enlightenment in this area will have to await better analysis. Meanwhile, although NARPAC cannot show it in readily available data, there is certainly a substantial connection between crime rates and poverty.

3. "effective" population

NARPAC agrees with the GAO that some consideration should be given to the differences in population that arise from visitors to the city, be they tourists, commuters, or demonstrators. However, GAO again seems to have selected a workload measure that exaggerates the impact, even though it simultaneously demonstrates perhaps an upper limit on the cost of commuters to the city. Comparing DC to the national baseline, they have used Census "Journey to Work" statistics to get a crude measure of how many people work other than where they live. While it may seem clever at the outset, using as a baseline statistics on how many Americans cross state lines every day to get to work is surely extreme. DC must be the only jurisdiction in the world whose entire boundary is comprised of state lines with either Maryland or Virginia. Not surprisingly, GAO found that the workload figure for DC is many times higher than for the entire American working public.

To partially offset this, GAO analysts also recognize that these intruders are only in the city for part of the 24 hour day, and part of the 7-day week. Hence the commuters count as only about 20% of a full-time resident. On the other side, however, they noted they did not account for tourists, who NARPAC believes contribute almost as many transients.

Nevertheless, even if this estimated increase in workload were proper compared to the national baseline, it is hardly appropriate for an urban baseline. Using the same journey to work Census data, the chart to the right indicates how the ratios change (from 4.1:1 to 1:1) when the baseline statistics relate to changing jurisdictions (such as county lines rather than state lines). In fact everybody does it, and DC commuters are only somewhat more numerous (22%) than the average of eighteen representative metro areas. When this number is further reduced by their part time presence, the workload term is reduced to perhaps 1.05 and has at most a modest impact on the outcome, even ifdoubled to 1.10 account for tourists.

4. estimating FEMS workloads Unfortunately, NARPAC finds several numbers to challenge in this RES formulation as well. There may well have been some confusion as to whether this term was intended to apply to emergency response teams as well as to firefighters, but certainly the budget line items to which these are related encompass both, and it is in fact difficult to separate them out, particularly for jurisdictions like DC that send a fire engine on every emergency call. There are three weighting factors in this RES component.

The first factor, given 40% of the total baseline weight, addresses the "effective population" (or "daytime" population, as GAO call it). This is the only factor that essentially addresses EMS rather than fires, and NARPAC believes it is strongly underrated. According to the national FEMS data taken from the Intenet and summarized to the center left, well under 20% of all emergency responses deal with fires, real or false alarms, and less than a third of those fires are in residential units: the rest are in businesses, vehicles, or "other" (which in this case appears to stand primarily for dumpsters!).

The upper charts indicate how DC FEMS personnel are paid relative to both national and metro area standards. They indicate that DC wages are good relative to the 20 metro area sample, and even better relative to the national average, again suggesting different "mark-ups" depending on the baseline being used. The issue of wage is addressed again further along.

Concern for housing conditions that may prompt fires (old age and multiple units) seems to be somewhat overdrawn, and again the national baseline is not particularly representative of urban housing units. Rummaging through Census data leads the inveterate analyst to the chart to the right which provides housing data for NARPAC's twenty-city sample. Hence where GAO finds DC cursed with 3.3 times as many multiple-unit apartments and 2.7 times as many old structures based on a national average, these ratios drop to 2.0 and 1.8 respectively. This only points up yet again that a national baseline is inappropriate for an urban sample. Eliminating the mark-ups for old and multiple housing units reduces GAO's estimated 'structural imbalance' by almost $80M.

But either way, the role of fighting fires is much smaller than suggested by these weighting factors. Confirmation of this can also be derived from available (but now dated) DC statistics, which show that fire responses amounted to less than 17% of total DC FEMS activity from 1993 to 1997 even though the fire apparatus was far more frequently in use. For whatever its worth, DC's 1300-odd firefighters doused almost 5400 fires per year during that time, limiting property damage to an average of $22M per year in a city with a total (taxable) property assessment of roughly $44 billion.

The DC Statistical Handbook (INDICES) gives enough data to make possible a rough estimate of the use by commuters for FEMS services. The division of emergency responses is provided by Ward, noting that Ward 2 (encompassing "downtown") has higher demand because of the concentration of daytime employment there. Assigning an average DC level of support for Ward 2 residents, the remainder can then be allocated to those who only work there, be they workers who live somewhere in DC, or commuters from the suburbs. Assuming that two thirds of the 660,000 people who work in DC commute, then two thirds of the residual workload can be allocated to them. An 8% increase in "daytime" population leads to a 4% increase in EMS workload.

5. representative wages

Across each of these factors contributing to developing an estimate of DC's "representative expenditures", adjustments are made to assure that wages properly represent local conditions. This is a sort of "should pay" factor, and in the GAO RES development it accounts for millions of dollars in the public safety component. Fearing that "Census earnings data may not adequately reflect the labor market in which the DC government seeks to hire and retain workers" they chose to use BLS average wage rates for all private industry excluding manufacturing. Using the latest earnings data available (1990) which showed DC's resident earnings per employee to be only 104% of the national average, GAO concluded that this would not be sufficient to" reflect a competitive wage to attract and retain workers". GAO therefore raised average national wages for police by 42%, for corrections personnel by 34%, and FEMS personnel by 38% to provide "competitive pay".

As shown by the four-part chart below, NARPAC seriously doubts that any wage increases for average performance should be made. Using 20 representative metro areas for a relevant baseline, 2000 BLS data shows that DC already pays 11% over the average police wage, and 2% over the average FEMS wage. An even closer balance was found for elementary and secondary school teachers. This is germane because the GAO analysis does not suggest increasing teacher wages (or any others) to keep them competitive. Eliminating these extraordinary wage increases reduces the projected structural imbalance by over $300M.

6. missing workload factors for shared public safety tasks

The uniqueness of DC's circumstances extends to the sharing of jobs with other agencies both regional and federal. There are at least two worthy of note in the public safety area. First, there are over 3000 uniformed police personnel on regional and federal payrolls that are available to assist the District in many of its unusual functions. These range from over a thousand federal officers on the Capitol Police Force, and several hundred with the National Park Service, to Secret Service that protects the White House and the Federal Protective Service which assures the security of foreign embassies. The GSA protects federal buildings, the Smithsonian has its own police force, and so does the area's outstanding Metro system. Personnel from the FBI, ATF and DEA are often involved in combined operations with them Metropolitan Police Force, and even the DC National Guard gets involved on occasion.

. In a completely different area, the Federal Government recently took over the responsibility for DC's felony prisoners previously confined at the Lorton facility. Along with ancillary jobs related to pre-trial and subsequent parole functions, the DC budget has been relieved of more than $200M in annual costs. It seems unrealistic for the GAO to propose that DC's corrections budget (and its structural deficit) should be increased by from $50M to almost $200M to account for tasks now being done in large measure by the Feds.

NARPAC strongly believes that the RES methodology should be modified to accept the benefits of workloads shared with relevant government agencies. In fact, such sharing could probably further reduce DC's supposed inability to go it alone.

return to the top of the page NARPAC's Public Safety RES Calculation

Using the statistical factors developed in the preceding discussions, NARPAC uses the same general RES approach to develop a very different picture of what DC should be spending on its public safety elements. The tabulations are shown below in the same format as used by the GAO, and the differences indicated in red.

In summary, the NARPAC alternative does the following:

o uses the 20-metro area sample as the baseline, not US-wide or high-density counties;
o weighs population as a more important indicator that crime or youth;
o lowers the effective population factor to 1.10 for commuters and tourists together;
o uses violent crime rather than murders as the key crime indicator;
o finds the youth population distribution normal, not off-average;
o uses a generous wage factor of 11% based on BLS police comparisons;
o assumes that DC police are responsible for only 90% of the city's security; and
o assumes that DC corrections personnel are now only responsible for 50% of DC's prisoners;

On this set of inputs, for which the underlying statistics appear plausible albeit not exact, DC's public safety RES value would drop to $305M, a full $1.2 billion less than GAO's higher figure, and somewhat less than half as much as DC actually spent in 2000 ($676M). What this demonstrates most clearly is that if the GAO's exaggerated workload factors are tempered by a more relevant baseline (metro areas vs national, bls wages) and by less volatile indicators (violent crimes vs murders), then the root inefficiency of DC's bureaucracy compared to equivalent jurisdictions shines through.

In short, a carefully controlled version of the representative expenditures system can show just how unrepresentative DC's expenditures have become, even without compounding the problem by introducing its non-competitive output performance. But in the final analysis, one cannot disguise the fact that the nation's capital city, as one of America's 68 biggest, richest, metro area core cities, has twice as many police on duty per resident as the rest, and still has about 10% more serious crimes per resident, and that's without counting 3000 additional uniformed police employed by federal authorities.


Now we turn from representative expenditures to representative revenue generation. The Rayfuse Report provides enough data in its tables to permit comparing the four different techniques now available to estimate a state's ability to pay its bills. Despite their somewhat different roots, they end up differing from one another by only a few percent, and somewhat randomly at that. These methodologies were all developed for the express purpose of assessing each state's ability to pay, or in fact, not to pay its own way. They have been used primarily as the basis for determining equitable federal grants. Presumably, if it becomes fashionable for the federal government to compensate states for their "structural imbalances", these methodologies will be applied.

The first stab at determining ability to pay was based on the relative per capita personal income of each state. This was subsequently refined to include the relative "gross state product" of each state, a somewhat broader base. This led to a somewhat more detailed approach involving "total taxable resources" (TTR) and thence to a "representative tax system" (RTS) approach as an even more refined estimate of the average ability of each state to raise revenues. The Rayfuse Report was designed to provide a companion approach for determining and equally fair representative (average) expenditures by any state.

state scatter

It is of some passing interest to note how much scatter there is between the revenue-collecting requirements of thevarious states. The two charts to right show the trend lines between per capita state and local revenues (from Census Bureau government statistics) on the horizontal axis with either per capita personal income (upper chart) or gross state product (GSP) (lower chart) based on the summary tables in the Statistical Abstract. The trend lines are clear, but in both cases DC appears as an outrider (as a city rather than a state). Offhand, these charts suggest that DC may be generating about the "normal" amount based on its per capita personal income, but too little based on its GSP, even excluding the very large (but mostly untaxed) government sector.

It is important to keep repeating that this system makes the not unreasonable underlying assumption that if all spending on education, for instance, is totaled up and divided by some relevant workload (such as number of students), that this can be accepted as the norm for the United States. The RES system says nothing about relative efficiency of different jurisdictions depending, say, on their size or ethnic composition. Nor does it say anything about whether the performance of school kids is adequate or influenced by other social characteristics such as religious beliefs or parental involvement. Likewise, the RTS system seems to accept average tax rates and average property assessment values with no evident consideration of the relative quality of life associated with, say, living near the water or on bare subsistence farmland. These techniques may be adequate for selecting the perfect impersonal average, but there is no indication of their relevance in discriminating between the box seats and the bleachers at the nation's ball game.

It is also informative to look at the sources of gross state product by major sector (on the chart below) and see how far DC (left edge) departs from the US norm (center). The government sector contributions are shown above the 100% lines for easier comparison. Clearly, like many cities, DC's wealth is generated by the financial and services sectors with relatively little trade or the classic American functions of construction, manufacturing, and transportation. Furthermore, the trifling contribution of the agricultural sector in all but a few states is somewhat surprising. Finally, the huge impact of the presence of the federal government sets DC aside from other jurisdictions.

NARPAC sees two other implications here though it is in no position to expand on them. First, it is not clear that these major economic sectors are taxed in the same way, and there may we be some skewing of the "ability to pay" depending on whether the state is producing cars or services. Second, the property requirements for these sectors are vastly different and can generate very different revenue streams. For instance, at the one extreme, farm land, according to the Department of Agriculture, occupies a full 25% of all US land. At the other extreme, 80% of Americans, and surely the wealthier ones, live in all the metro areas across the US which occupy only 20% of US land. Surely some of the scatter in these data points is generated by unavoidable anomalies in taxable assets.

The summary chart below indicates how the four different revenue capacity approaches compare when applied to DC. The four lefthand clusters refer to the 1987 calculations: the blue bars are as calculated by Rayfuse, and the green bars as re-calculated by NARPAC as a beginner's lesson. Note that the simplest approach of using gross state product (GSP) as the basic indicator can be converted to a dead-wringer for the TTR approach by removing the government sector of the state product. For DC, the personal income (PI) and RTS approaches yield somewhat lower values, while GSP and TTR are somewhat higher. This varies from state to state depending on the contribution of personal income to the state's total revenue capacity.

The pale background bar indicates the total US GSP for 1987 on the left and for 2000 on the right. It is evident from this that DC's growth in wealth has not kept pace with the country as a whole, though the four approaches keep the same relative ranks. The solid red bars are the GAO calculations: an upper and lower value for RTS, and a single value for TTR. GAO provided no estimates for PI or GSP while NARPAC dutifully generated all four again. The large difference between the GAO and NARPAC calculations of the same figures led to a further puzzle, forcing us to recalculate GAO's numbers, as explained below

doing the math

From a computational point of view, developing the revenue indices requires a few arithmetic steps, complicated by the (necessary) inclusion of direct federal payments. These are summarized on the adjoining chart for those few who may care. Unfortunately, the GAO analysts appear to have missed some of these steps, and NARPAC cannot present its objections to their conclusions without exploring the math. The guiding principle is simple: find one state's ability to generate revenues as a share of the national total (0.369% in the case of DC) on the basis of per capita income, GSP, or whatever, and then assume that it can afford to pay the same share ($4.68 billion) of the nation's total bills ($1.31 trillion in 2000). If its "should-spend" bills are higher than this, it has a structural imbalance i.e., a deficit in its ability to pay what it should be paying.

The calculation is somewhat complicated when the Federal Government provides direct payments to the states. In this case, the total federal payments ($291.5 billion) are subtracted from the nation's total bills to provide a new (lower) fair-share payment from "own-sources" ($3.638 for DC). To this is added the estimated local federal payment to get a revised afford-to-pay figure ($5.19 billion for DC in '00 by NARPAC's calculations). For lack of a basis for better choice, NARPAC uses the actual DC federal grant value ($1.750M) rounded from Census data.

From what NARPAC can divine, GAO did not go through the intermediate steps of reconciling the own-resources RTS to the national totals. The result is an "ability to pay" that is several hundred million dollars lower. If, as NARPAC concludes, this is an error in calculation, it has a very significant influence ($483 million) on any local structural imbalance.

where is the "representative federal payments" methodology?

The impact of federal grants seems to be an understated influence on the structural balance question. It is particularly important for DC where the federal grants are far higher than for any of the fifty states. DC's federal support accounts for almost 38% of its total expenditures. For the national as a whole, the fifty states average only 25%, and the largest ten states are a bit lower at 23%. As a pertinent aside, we also note that the average for Maryland and Virginia, DC's next door neighbors, is only 20% from federal grants. They clearly have a substantially smaller problem with poverty. The scatter plot below indicates the extent of variation in federal payments per person in poverty as a function of each state's ability to pay. While the almost random cluster to the left represents 49 states, the fiftieth point for DC is an analyst's challenge. Whatever is DC doing in a totally separate orbit?

There does not yet seem to be any "representative federal payments" (RFP) methodology comparable to those for representative expenditures and tax systems. The GAO simply states that they estimate that the District would have received $1551 million "if it provided and average level of services". NARPAC tried briefly and unsuccessfully to generate a methodology that would capture the primary variables for federal grants. There seem to be three driving conditions: the first is assistance for the poor in health, housing, education, nutrition, and unemployment. The second is the major federal program to help with transportation needs, and the third is the ubiquitous catch-all heading of "other". While the latter two factors (less than 20% on average) might be somewhat arbitrary, the major poverty-relief component should relate to the number of those requiring assistance in each state.

In short, and as shown on the two charts below, the share of DC's expenditures provided by federal grants appears abnormally high by almost any measure. Federal grants amount to at least $2175 per DC resident, compared to $950 as the national average and $715 for Maryland and Virginia with their relatively low poverty rates. But these differences are further exaggerated by the fact that the grants for each person in poverty in DC reaches almost $14,000 per capita, while it is under $8300 across the US, just about $7900 for the "big ten" states, although over $9800 for Maryland and Virginia.

The detailed break-out of the various elements of the federal payments are shown in the lower graphic. If this reads as an inconclusive analysis, it is because NARPAC has not found a rational RFP methodology to explain the differences. Nevertheless, if DC's federal grant rate just for those in poverty was restricted to that of its immediate neighbors, DC's potential structural imbalance would be increased by over $350 million. The stark differences in poverty burdens between the capital city and its suburbs deserve far greater visibility than has been provided to date. Such insight might well provide a more rational basis for the sharing of financial (and social) burdens than straight subsidies to DC.

In fact, there is every indication that DC's federal grants for non-poverty assistance (transportation and "other") are equally generous. In yet another unsolved mystery (at least by NARPAC), DC's "other" federal grants dwarf those to the fifty states. These other grants do not appear in DC's budget, but do appear in other Census 2000 documentation of state and local government finances, such as the Statistical Abstract. Using the lower numbers, federal largesse to DC is still over four times the national per capita average of about $200. Using the higher values from the Statistical Abstract, DC is receiving almost fifteen times the national average. However, these incredibly high numbers have not been used in the structural imbalance calculations. Nevertheless, DC activists would have virtually no foundation for claiming that DC gets less than its fair share from the Federal Government!.

return to the top of the page UNFINISHED BUSINESS

DC's Need for Increased Capital Investment

It is NARPAC's opinion that DC's most troubling condition is not its operating budget but the seemingly huge backlog in its capital budget. If there were to be any rationale for awarding DC some sort of federal subsidy for hosting the nation's capital, we would find the need for an exceptionally fine national capital infrastructure (from sewers to railroad bridges) to be the most compelling. Unfortunately, no rational is presented for judging the reality of any of DC's deferred capital needs. In fact, subsequent analysis of DC's FY03 Capital Improvement Plan indicates that a) the out-years of the current plan are significantly unfunded, and b) a more realistic estimate of future DCPS school needs could eliminate most if not all of that alleged deferred maintenance and acquisition (see below).

The big gorilla in this tent is DC's aging public school system. But typical urban school costs and performance can be readily measured (less than a day's work for an analyst with publically available data). DC student performance, albeit depressingly bad, is not out of line if the chosen primary variable is parental education level. What is out of line is the very high annual cost differential for operating too many partially filled, smaller-than-average schools, and planning to maintain or replace them all despite a continuously declining public school population. Again it is difficult to develop sympathy for DC's capital investment plight when it appears large self- generated. Nevertheless, it deserves far closer scrutiny and objective analysis. NARPAC has taken another pass at this issue and concluded that the 'structural imbalance' in DC's long-term facilities needs is largely of its own making, by planning for too many schools for too many students in schools substantially smaller than the urban norm.

The "Burdens" of Hosting the Nation's Capital and the People Who Make It Run

Finally, neither the GAO nor NARPAC analysis adequately treats DC's repeated claims over many years that the municipal government is burdened with the costs of quasi-statehood, of hosting the nation's capital and most of its work force without reimbursement or the ability to tax either the commuters or the federal lands. The most valuable part of this GAO report is its assertion that DC lacks the quantitative data to make its case for uncompensated costs, or to quantify the benefits of the presence of the federal government. NARPAC has repeatedly claimed that the only difference between DC and Camden, New Jersey, is the presence of the federal government. We fully support nailing down these un-supported allegations. It is high time for DC to look inward for the solutions to its financial problems, instead of outward to the only institution responsible for its existence and current prosperity.

Scattered through the GAO reports and the NARPAC analysis are references to the uncompensated costs of the nearly half-million commuters that occupy the city's commercial and government buildings. The range of estimates is grotesque. The ACIR methodology itself suggests that the annual public safety costs of the added "daytime" population may reach $16 million. Even if this were doubled by equivalent costs to DC's transportation infrastructure, this would put a price of $60 per year per commuter, which is more than balanced by the CFO's own estimates that commuters may generate sales taxes of almost $300 each. NARPAC, on the other hand, suggests that the property taxes generated by the offices in which the commuters work could bring in $260 each.

At the other end of the spectrum, in what may prove to be The Wildest Claim Ever Made, an as- yet unpublished report (referenced in the interim GAO study) asserts that the annual per commuter cost to DC is $3016, of which $90 is for public safety (police and FEMS). This apparently flows from a consultant's report prepared for the legal team advocating DC's right to impose a commuter's tax. For 450,000 commuters, that amounts to $1.36 billion, or 42% of DC's total locally raised budget of $3.215 billion. In a pig's eye!

In another wild guess, the CFO estimates that 27% of DC's combined budgets for public works and public safety are spent in support of federal properties simply because 27% of the city's total assessed property value derives from (untaxed) federal properties. This takes no account of the National Park Service's own contributions, or those of over 3000 uniformed police on various federal payrolls in the nation's capital over and above DC's own 4300 finest! Until these rather rudimentary problems can be given some credible analysis, there is little basis on which to judge DC's true financial conditions or valid needs.

This page was updated on Sept 5, 2003




© copyright 2007 NARPAC, Inc. All rights reserved