Skip Navigation
| Skip Navigation |
|
|
|
|
|
Madam Chairman:
Thank you for this opportunity to discuss with the task force the ways we might better ascertain the impact of the federal budget on states and localities. The issue is a difficult one, but it is especially important in light of recent evidence showing major geographic shifts in economic and social activity. The concentration of people and jobs appears to be shifting from central cities to suburbs, from larger metropolitan areas to smaller ones and to nonmetropolitan areas, and from the industrialized northern tier of states to the south and the west. Many observers question whether federal policies--particularly budget policies--have been contributing factors in prompting or retarding these changes.
The question on which these hearings focus is: How do federal budget decisions affect state and local governments? Generally, this question has been answered only in an aggregate sense: total funding levels for grant programs have been presented, and the impacts of fiscal policy for the overall state and local sector have been estimated with econometric models. Such information is useful in understanding changes in the federal system that have taken place over time and in formulating fiscal policy. However, these aggregate analyses miss the most significant dimension, which is that individual state and local governments are very different and face different problems. Rarely will a federal action have an equal impact on all jurisdictions. Some will be affected more than others. One may gain from a federal resource allocation or distributional decision while another loses. In short, we need to know how budget policy affects individual states and localities.
The formulation of a budget requires two types of decisions. First, the nation's fiscal policy is determined largely by the budget's overall spending and, revenue levels. This fiscal policy will affect different states and localities in different ways. Second, specific budget allocations to various spending programs and taxing instruments must be made. To determine how this second set of decisions affects state and local governments, we need to answer three questions.
We know something about the first question but very little about the second and third.
The Geographic Distribution of Federal Expenditures
The expenditure budget can be broken into two parts: direct federal outlays and tax expenditures. More effort has been directed toward the compilation of information on direct outlays than on tax expenditures. Tax expenditures, however, have impacts that are important and that differ on a geographic basis.
Direct Federal Spending. Direct spending programs can be divided into grant programs and other federal spending. It is useful to distinguish grant programs because their impacts on state and local governments are direct and immediate while other federal actions have less direct effects. When the federal government purchases services, builds infrastructure, subsidizes specific economic activities, or provides benefits directly to individuals, certain areas of the country tend to be favored over others. Over time, these actions can contribute to changes in a state or local economy and population, leading to an increased or decreased tax base and greater or lesser demands for public services.
A considerable amount of information is now available regarding the geographic flow of federal funds. There are gaps in these data, however, and certain factors have impeded the compilation of a more comprehensive data base.
In general, more is known about the geographic pattern of federal spending for grant programs than for other types of federal programs. Indeed, information on the flow of grant-in-aid dollars to state areas--that is, the combined total of aid directed to the state government and all local governments within a state--is quite good. The Treasury Department publishes an annual report entitled Federal Aid to the States, which provides an accurate listing of the monies distributed to each state; the list is broken down into over 90 categories of grants. The General Accounting Office is nearing completion of a longitudinal analysis of these data, which should be of great value to this task force.
States, however, are quite heterogeneous, incorporating local jurisdictions with diverse problems that specific federal programs are designed to solve. Tracing grant funds to the state level does not allow a judgment about the efficiency with which grant monies are targeted on the areas experiencing the problems or about the other local impacts of the federal government.
Unfortunately there is less information on the distribution of grants among substate areas, and whatever information we have tends to be less reliable than the state-level data. One reason for this situation is the diversity and complexity of government structure in the federal system, which causes reporting difficulties. States differ tremendously in the way they structure local governments and in the way they assign responsibility for the performance of specific functions. Different levels and types of governments may be involved in providing any given service. A service that is the responsibility of a state government in one state may the responsibility of a county government in another and a city government or special district in a third.
A second reason why substate area information on the distribution of grants is difficult to obtain arises from the multiplicity of local governments serving any area. In addition to a general purpose municipal government, the county government, a school district, a township and various special districts may be responsible for providing services. Rarely are the boundaries of these various jurisdictions coterminous.
The variation in the assignment of responsibility for services, the multiplicity of governments, and the fact that boundaries overlap complicate the interpretation of what information there is on the distribution of grants. In order to get an accurate picture of the level of federal effort in a given city or other substate area, it is necessary to examine grants received by all of the governments providing services in the area. But it is not easy to estimate how much of a federal grant received by a state or a county government benefits the residents of a particular city. However, unless such estimates are made, very misleading judgments can arise.
The debate on New York City's fiscal crisis provides an example of this. During the debate, many comparisons were made between the situation in New York and that in other large cities. From a comparison of the grant awards to New York City and the city of Detroit, one might conclude that New York was advantaged and Detroit shortchanged. However, much of the apparent difference could be attributed to the fact that New York City is responsible for the administration of welfare and education programs, while welfare programs in Detroit are the responsibility of the Michigan state government, and its education programs are administered by an independent school district. Thus federal grants for these functions would appear in the budgets of these other governments.
Another difficulty encountered in tracing the flow of grant dollars to substate areas arises because, in some grant programs, states determine either all or part of the distribution of federal dollars to local areas. In many instances there is no reliable information on the amount passed through to local governments. Furthermore, states may have little idea where some federal dollars end up because state money is combined with these federal dollars before the distribution, and breaking out the federal piece is difficult.
Despite all these problems the Bureau of the Census and the Community Services Administration (CSA) have made commendable efforts to trace federal grant dollars and, in the case of CSA, other federal spending as well.
The Bureau of the Census annually surveys the budgets of a sample of state and local governments. At the local level, these surveys provide information on those federal programs that channel funds directly to local governments. Grant programs that provide indirect aid (by way of a state government pass-through) are counted as state aid. Thus, funds received under the Comprehensive Employment and Training Act, general revenue sharing, and community development block grant program would appear, in most instances, as federal aid; but Law Enforcement Assistance Administration funds, HUD planning grants, and medicaid dollars would not be distinguished from the states' own aid programs.
The Bureau of the Census groups the data for all local governments serving selected county areas. Thus, for those grant programs that entail direct federal/local relations--accounting for 23 percent of all grants in 1975--the share going to governments serving residents of many metropolitan counties and some large nonmetropolitan counties can be determined.
CSA is a second source of information on dollar flows to substate areas. Federal agencies are required to allocate their entire budgets among geographic areas. Spending is reported for all cities with populations larger than 25,000, all counties, and all states. The data base covers spending in both grant and nongrant programs.
While this data base is very useful, one has to be aware of its limitations, which CSA carefully notes in the beginning of their publications. Among these are that:
Some of these methods seem to produce patterns that closely approximate the actual location of the spending. For example, spending for the year in the Aid to Families with Dependent Children program was allocated among counties based upon their share of total spending in February--the one month when states are required to file comprehensive reports. In other instances, the reporting method is much less adequate. For example, all spending for some grant programs that require a state pass-through--such as vocational or handicapped education--is credited to the state capital although much of this money is distributed by the states to individual localities.
These reporting problems are more serious at the local than at the state level. In conjunction with the study of the distribution of federal spending by county, which CBO is preparing for this task force, we made a judgment as to the reliability of the reporting methodology for each program in the CSA data base. While this review was based on limited information and was to some extent arbitrary, we concluded that the reporting methodology for some 29 percent of all federal spending was open to serious question.
Tax Expenditures. The Congressional Budget Act of 1974 defines tax expenditures as "revenue losses attributable to provisions of the federal tax laws which allow a special exclusion, exemption or deduction from gross income or which provide a special credit, a preferential rate of tax or a deferral of tax liability." In many instances these tax provisions are designed to achieve objectives that are or could be pursued through grant or other expenditure programs.
Unfortunately there is little or no analysis that examines the geographic distribution of the impacts of tax expenditures. Data are available that may allow some analysis of the tax expenditure provisions which most directly benefit state and local governments: namely the exclusion of interest on state and local debt (a tax expenditure of $5.3 billion in 1976) and the deductability of state and local taxes (a tax expenditure of $13.5 billion in 1976). The value of these provisions undoubtedly differs significantly by jurisdiction. The impact of the first varies with the amount of state and local government borrowing. Benefits from the deductability of state and local taxes vary depending on the income of residents and state and local tax rates. Crude calculations suggest that the per capita "grant" implied by the tax deductability provision is more than five times larger in some states than it is in others.
Other tax expenditures undoubtedly have had an impact on the geographic pattern of economic development. Virtually nothing has been done, however, to uncover the extent to which certain tax expenditures have fostered suburban growth or the growth of certain regions of the county.
Evaluating the Geographic Distribution of Federal Funds
Tracing dollar flows is only one step in a comprehensive review of the budget's impact on states and localities. The second question is whether this flow of federal dollars appears to be going in the "right direction." The answer depends not only on an ability to classify state and local areas by some meaningful measures of their social and economic conditions, but also on some difficult value judgments.
Clearly, not all cities, states, or regions are alike. They face different kinds of problems, and some have greater need for federal financial assistance than others. Houston and Detroit are central cities of comparable size, but in other respects they are very different. One is growing and attempting to cope with the problems growth brings, while the other is declining and struggling to retain jobs and provide necessary services. Similarly, El Paso and Providence are smaller cities but in the former, economic growth is strong; in the latter it is weak. Per capita income in El Paso, however, is 21 percent lower than that in Providence. Which of the two is in greater need of the stimulus implied by the federal spending is a value question--properly addressed by policy makers. Analysts can only array the data--characterizing the places most in need according to the various criteria. Unfortunately some of the data useful in characterizing social and economic differences are available for local areas only from the decennial census.
Various studies have attempted to provide a base for evaluating the distribution of funds for specific programs or groups of programs. The CBO study being prepared for this task force, will look at different types of economic problems besetting counties, and it will attempt to determine whether outlays are directed primarily at those counties that appear to have the greatest difficulties.
Impact Of Budget Changes
On the important question--How would changes in budget policy affect the social and economic conditions of specific state and local areas?--we know least.
There is great interest in the geographic distribution of federal funds on the presumption that federal expenditures are beneficial to the local area in which they take place. While this assumption is undoubtedly true in general, it is simplistic, static and possibly wrong in particular instances. The benefits that derive from various kinds of spending are undoubtedly different but largely unknown; tire local conditions that must exist for different spending strategies to be effective are likewise poorly understood. Our ability to answer such vital questions as, "What levels and kinds of spending could reverse the pattern of economic decline in the nation's central cities?" is limited. In large part, our ignorance in these areas stems from the complexity of the economic development process. There are many factors influencing local development patterns including changes in private demand, technology, and state policies. It is difficult to isolate the impact of federal spending.
Improving Information
Clearly the information that I have outlined falls short of what one might like to have to measure the budget's impact on states and localities. There are, of course, ways around these problems, but some entail significant costs and others require changes that may not be within the power of the federal government. For example, the structure of local governments could be simplified and standardized across states. This might make it easier to administer a rational program of federal assistance since the needs of various units could be assessed more clearly. Such a restructuring, however, is clearly unrealistic and may not be better suited to local needs and citizens' preferences.
Another possibility would be to impose additional reporting requirements on recipient governments. For example, states could be asked to supply information on where grant dollars are spent in a form compatible with the information system designed by the federal agency to trace geographic patterns of spending. Added reporting requirements impose costs and burdens on those who must comply, however.
A third approach to improving the information base would be for the federal government to provide states with incentives to standardize those programs from which needed data could be collected. For example, local area unemployment statistics would be more reliable if states instituted standard eligibility requirements for their unemployment insurance programs. There is a tradeoff, however, since existing variations may be the result of differing state needs and preferences.
Undoubtedly, administrative changes that would improve the ability of federal agencies to trace federal expenditures are possible. The other witnesses today are better able to provide suggestions on where changes to improve reporting might be made and what these changes might cost.
In closing, I would like to point out that the federal government affects patterns of development through activities that are only partially reflected in the budget. Tax, trade, regulatory, and foreign policies are important, but their impacts on states and localities cannot be measured in terms of the budget. Trade policy can determine whether goods produced in certain sections of the country remain competitively priced with imports. Foreign policy decisions can affect the price of essential raw materials that impact on the various regions differently. Regulatory agencies establish rate structures which favor some activities and areas at the expense of others. These policy areas can and ought to be subject to analysis from the perspective of state and local area needs.