Achieving Goal 3: Raising Revenue Fairly and Efficiently
Most of the policy discussion related to the fairness of school financing systems focuses, as it should, on the pattern and level of educational spending or outcomes. However, another aspect of equity should not be ignored: How fair is the distribution of the burden of the taxes or fees used to generate revenue for schools? Some aspects of this revenue perspective on equity are obviously intertwined with the spending and outcome issues embobied in the concepts of school finance equity and adequacy discussed in Chapters 3 and 4. However, many other aspects are not. Moreover, the basic principles typically used to evaluate the fairness of revenue sources differ from the standard equity principles underlying the school finance literature.
In addition to equity, other aspects of the revenue system are also important, such as how much it costs the government to administer the system, how stable the revenue sources are over the business cycle, and how extensively the tax system distorts taxpayers' decisions in undesirable ways. Policy makers who are striving to achieve the goal of raising revenues in a fair and efficient manner need to pay attention to all of these issues. In addition, they need to consider any tradeoffs or complementarities with the other two goals of a good financing system: increasing achievement for all students and reducing the nexus between achievement and family background.
Two main aspects of revenue raising should be distinguished—the particular revenue source that is used (i.e., property, income, or sales taxes) and the level of government (i.e., school district, state, or federal government) that is responsible for raising revenue. Table 8-1 displays the main options for the United States.
In this chapter we argue that the local property tax remains the best way to
TABLE 8-1 Tax Options for the United States
Type of Tax
raise local revenue for education. That is, provided a decision is made to lodge significant responsibility for raising revenue at the local level, the local property tax is preferred to other local taxes for that purpose. That analytical conclusion is depicted in Table 8-1 by the designation of property taxes as the single revenue source for local school districts. While local sales and income taxes are potentially feasible and are currently used in some states, the property tax is the preferred option.
At the state level, the relevant major revenue sources are state income and sales taxes, with 41 states using income taxes and 45 using sales taxes. States traditionally have not made much use of state-level property taxes to finance education. However, several states have recently shifted toward statewide property taxes or their equivalent, and other states could do so as well. Hence, along with state income and sales taxes, we identify statewide property taxes as a larger potential source of revenue for education than in the past.
Although the federal government now provides only about 7 percent of the revenue for K-12 education, in principle it could do a lot more. In the absence of major tax reform at the federal level (in the form, for example, of a shift away from the federal income tax to a value-added tax or other form of consumption taxation), the relevant revenue source for education at the federal level is the federal income tax.
After evaluating the property tax in some detail, this chapter addresses four questions about the revenue system. The first question is whether the mix of local taxes should be altered by reducing reliance on the property tax and increasing reliance on other local taxes or a modified version of the property tax. The second is what state revenue source would be best if heavier reliance were to be placed on state revenue sources. The third is whether it would make sense to shift away from local revenue raising in favor of much greater reliance on state revenues. And the fourth is whether it would be desirable to increase significantly the federal role in revenue raising for K-12 education.
EVALUATION OF THE PROPERTY TAX
Evaluating the local property tax as a source of funding for local schools is significantly more complicated than evaluating a state or national tax because of the close relationship between the revenues collected and the amount of spending
in each school district. This close relationship suggests to some people that the property tax should be evaluated as a benefit tax—that is, as a tax that is specifically paid for the services provided by a community, rather than in more standard ability-to-pay terms, which examine the distribution of the tax burden separately from what the funds are used for. When looked at as a benefit tax, the local property tax appears to perform better on efficiency grounds and, according to some people, also better on equity grounds than when it is looked at through the lens of the more traditional tax literature. In the following sections we refer to both perspectives.
Most taxes induce some inefficiency by encouraging taxpayers to alter their behavior in ways that would reduce their tax liability. In the case of the property tax, the standard concern is that households will respond to the tax by investing less in housing or that firms will respond by investing less in property subject to the property tax or by shifting their investments to areas with low property tax rates (Mieszkowski and Zodrow, 1989).
In contrast to this emphasis on the inefficiencies of the tax, the benefit-tax approach emphasizes that the local property tax may generate efficient decisions, especially with respect to the level of education services. Particularly in large suburban areas where households can choose among many small, relatively homogeneous school districts, households gain access to the education services provided by a district in return for paying its local property tax. Presumably, people will have a tendency to sort themselves among districts in line with their preferences for education, so that those with stronger preferences for education will end up in districts with more education and higher property taxes than those with weaker preferences. In effect, the property taxes act more like prices that consumers willingly pay for education than compulsory taxes. This analogy is particularly apt, according to the advocates of this perspective, when local zoning regulations ensure that residents in each school district end up in houses that are similarly valued, so that they pay similar property taxes for their uniform public education (see Hamilton, 1975; Fischel, 1992; Hoxby, 1996b).
This view, which is generally consistent with the well-known strand of public finance literature initiated by Charles Tiebout in 1956, draws attention to the efficiency benefits that accrue from local provision and financing of public goods. Significantly, however, the benefit-tax approach would apply to any local tax, not just to the local property tax. That is, the efficiency claims for the property tax relate more to the governmental level at which revenues are raised than to the desirability of the property tax or any other specific local tax. For this reason, we defer to later in the chapter a fuller discussion of the potential trade-offs between efficiency and equity that might arise with a shift to a larger state role in education finance.
Furthermore, the benefit-tax argument applies most directly to small, relatively homogeneous suburban districts only when local zoning enforces housing uniformity within districts, and it is not particularly pertinent to central-city or rural districts. In light of these limitations, it seems reasonable to treat the property tax like other taxes and to recognize that it will induce some distortions in behavior. Because all taxes induce distortions, the relevant question from this more traditional perspective then becomes how the distortions associated with the property tax compare with those associated with other potential local taxes, such as the income or the sales tax.
There is little doubt that many taxpayers view the local property tax as unfair. In 1978, for example, voters in California shocked the nation by supporting Proposition 13, a statewide initiative to reduce drastically the level and rate of growth of local property taxes. Massachusetts voters followed with their own stringent tax limitation measure two years later. While the motivation for these and other measures to limit property taxes are obviously mixed and complex, voters appear to have been motivated in part by their perception that the local property tax was unfair.1 Acting on concerns of this type, legislators in many other states did not wait for statewide referenda to reduce property taxes, but instead jumped on the anti-property-tax bandwagon and introduced policies designed to provide property tax relief to all or to specific groups of taxpayers.
Further evidence of continuing taxpayer dissatisfaction regarding the property tax emerges from annual surveys administered between 1972 and 1994 by the U.S. Advisory Commission on Intergovernmental Relations. During that period, between 25 and over 30 percent of the respondents consistently rated the property tax as the least fair tax (compared with federal and state income taxes and the state sales tax) and typically much less fair than the main alternatives to property taxes for financing education, state income and sales taxes.
Taxpayers have lots of reasons for believing the property tax is unfair. Many of these relate to the way the tax is administered (e.g., see Netzer and Berne, 1995, for examples specific to New York State). Compared with other taxes such as income and sales taxes, the property tax is more difficult to administer fairly because it requires that property be assessed. Ideally, the assessed value of a property should reflect its market value. However, where there are few market transactions, the value of the property must be approximated by one of several imperfect methods. It is not surprising that the assessment of property is subject
to error and to political influence in some states and, consequently, frequently departs from market value.
Another complicating factor is that any given parcel of property is likely to be subject to taxation by several local authorities, such as a county government, a municipality, a school district and perhaps several special-purpose districts. Such complexity may well confuse local taxpayers and make them view the overall burden as unfair. Finally, the fact that property taxes are typically levied only once or twice a year makes them more visible than income taxes, which are largely collected through withholding, or sales taxes, which are collected in small amounts at the cash register.
Other reasons for concern about the fairness of the property tax arise in particular circumstances. In California in the late 1970s, for example, part of the concern arose from the rapidly rising housing values and the three-year assessment cycle, which combined to produce huge increases in property valuations and tax burdens in a single year, increases that bore no relationship to the taxpayer's ability to pay as measured by current income. In Massachusetts, the high levels of property tax burdens made the tax difficult for some taxpayers to accept.
Beyond some of these taxpayer concerns, many of which are clearly valid and reflect underlying problems with how the tax is administered, experts in public finance also have much to say about the fairness of the property tax based on the two basic principles of tax equity. However, as the committee discovered, not all economists agree.
Those economists who emphasize that the local property tax is like a benefit tax would typically defend it on fairness grounds by appealing to the benefit principle of tax equity. According to this principle, a tax is fair if the burden of the tax is distributed among taxpayers in line with the benefits they receive from the services funded by the tax. However, in the committee's view the benefit principle is not very applicable in this context, largely because it is based on an unacceptable ethical foundation. Because the demand for education is highly correlated with parental income and education, this approach to equity would accept as fair differences in education levels across jurisdictions that correspond to differences in preferences and family ability to pay for education. Although such a pattern increases the efficiency with which education is provided (in the sense that those who have greater willingness to pay for education get more) compared with a uniform state-wide level, the committee sees no reason to assert that such an outcome is fair, especially given the acknowledged significance of education to a child's life chances.
An additional complication arises in using the benefit principle to evaluate the equity of property taxes because of the phenomenon of tax capitalization. For example, if two school districts provide similar education services but one does so at a lower tax rate (perhaps because of the presence of a power plant in the district), homeowners are likely to have bid up housing prices in the low-tax
district enough so that new homeowners would receive no net financial benefit from living in the district with the lower tax rate. Thus what may appear as an inequity—the fact that households in one community pay a lower tax rate than those in the other for the same services—turns out not to be an inequity if one takes into consideration the fact that the households paying the lower tax rate pay more for their housing.2
More useful and appropriate for evaluating the equity of the property tax, in the view of the committee, is the ability-to-pay principle. From this perspective, the key question is: How regressive is the property tax?
It turns out that there is no simple answer to this question. The incidence of the property tax has, in fact, been one of the most controversial topics in the field of local public finance. The controversy centers around the so-called old view and the new view of the incidence of the tax. Those who apply the old view argue that the tax is regressive on the grounds that landlords and business firms are able to shift much of the burden of their real estate taxes onto renters and consumers in the form of higher prices. The new view, in contrast, emphasizes that the property tax is a tax on wealth, so that the ultimate burden of the tax is distributed in line with earnings from wealth. Since those earnings rise disproportionately with household income, the tax burden, according to this view, could be progressive—that is, the tax would place a heavier burden on higher-income households than on lower-income ones. Fortunately, as noted by McLure (1977), the two views can be reconciled by interpreting the old view as one component of the new view. In particular the old view is most applicable to differences in property tax rates across jurisdictions.
This reconciliation leads to the general consensus that what makes the property tax regressive is the differences in tax rates across jurisdictions. Regressivity emerges to the extent that higher property tax rates are levied in districts with above-average proportions of poor households and lower rates in districts with richer households. However, a different conclusion emerges for property tax rates that cover a broad geographic area like a large state. In this case, the new view would be more applicable, which would mean that the burden of an increase in the property tax rate is likely to be proportional or even relatively progressive. This observation suggests that a shift away from reliance on local property taxes (with their tax rate differentials that make the tax regressive) to a statewide property tax could well make the tax system fairer.
This discussion generates two conclusions about the fairness of the property tax. First, from an equity perspective, the administration of the property tax leaves a lot to be desired and inevitably leads to the unfair treatment of some taxpayers. The committee would support ongoing efforts to administer the tax as fairly as possible. As spelled out by Netzer and Berne (1995:39), a property tax system should be: (1) ''transparent and straightforward," so that it is comprehensible to voters and property owners; (2) it should be "systematic," in the sense of having few internal contradictions; and (3) it should be "reasonably related to the policy objectives that animate the various provisions."
Second, the property tax is not so regressive—and hence unfair—a tax as some people make it out to be.3 Its major failings are (1) the regressive elements that emerge because it is a local tax and (2) the inequities in spending that result from the wide variation across districts in the property tax base. However, these latter inequities result more from the fact that many states place such heavy reliance on local—in contrast to state—revenue sources rather than to the property tax itself.
This conclusion about regressivity implies that converting the local property tax with its variation in tax rates across jurisdictions to a statewide uniform-rate property tax could well improve the fairness of the revenue system. It would eliminate most of the regressive element that arises from the differential tax rates across jurisdictions and would eliminate the spending disparities that arise from the variation in local property tax bases.4 A logical next question would then be whether further gains in equity could be obtained by shifting away from the property tax completely to other statewide taxes, such as income or sales taxes.
However, the committee is well aware that some people would argue that any gains in equity (either in the fairness of the revenue system or in the form of a more even pattern of spending across districts) from a shift to statewide taxes could come at a potentially large cost, namely the loss of local control and more efficient decision making that flows from local school districts having access to their own source of revenue. In light of the concern about local control and efficiency, it is worth examining first how the property tax stacks up against other local taxes that might be used by local school districts.
SHOULD THE MIX OF LOCAL TAXES BE CHANGED?
The two major candidates to replace the local property tax are the local income tax and the local sales tax. Alternatively, one could imagine modifying the local property tax by applying it to residential property alone or by sharing the revenues across communities within a metropolitan area.
Local Income or Sales Taxes
A somewhat stronger case can be made for replacing the local property tax with a local income tax than with a local sales tax, but even here the argument is not compelling.
Strauss (1995) tries to make the case for a local income tax with particular reference to New York State. His starting point is that because education represents an important form of income redistribution, it should be financed out of broad ability-to-pay taxes, such as income or consumption broadly defined. The property tax is unfair, he argues, because it does not allocate burdens in line with any reasonable concept of ability to pay. For example, it often imposes heavy burdens on the elderly who are property wealthy but income poor, and it treats taxpayers in districts with a lot of business property favorably relative to those in districts that are primarily residential (with no reference, however, to the potentially offsetting impacts on housing prices that occur when differences in property taxes are capitalized). Few would disagree with the argument that the income tax is the superior tax judged in terms of ability to pay. Unlike the property tax, it applies directly to households and hence can be adjusted to take into account the circumstances of the family, such as the number of dependents. Furthermore, provided one accepts the view that current annual income is the appropriate measure of ability to pay, tax burdens will inevitably be more in line with household ability to pay than they would be with the property tax. Of course, in practice, in order to minimize administrative and compliance costs by taking advantage of the existing state administrative structure, the definition of income for a local income tax would most likely follow the state definition. Consequently, the fairness of any local income tax would depend on the fairness of the state income tax.
Although from an ability-to-pay perspective a local income tax could well be a fairer way to raise revenue within a district than the local property tax, the income tax may well be less desirable on other grounds. First, it could well lead to even greater disparities across school districts than those associated with the local property tax. This outcome would occur, for example, if the amount of business property (which is included in the property tax base but not the income tax base) were larger in areas with lower-income residents than in areas with wealthier residents, a pattern that would tend to mitigate the effects of large differences in household income or property wealth across districts. In practice,
which tax base varies more is an empirical question and the outcome is likely to vary from one state to another. As Oates (1991) has pointed out, the unequal distribution of the tax base is likely to be a problem for any local tax whether it be a property, income, or sales tax, and the problem could well be less for the property tax than for other taxes.
Second, there are likely to be greater behavioral distortions with the income tax than with the property tax. If a redistributive local income tax and expenditure package is weighted toward lower-income individuals and families, then inmigration of the poor into the local jurisdiction raises the cost of redistribution and may cause out-migration of wealthier families. The property tax generates fewer distortions because the mechanism of tax capitalization through which differences in tax burdens are reflected in the prices of housing makes it more difficult for taxpayers to avoid the burden of the tax by movement from one jurisdiction to another.
Other possible advantages of the property tax over the income tax include the fact that it is potentially more stable over the business cycle because of the relative stability of housing values (but offsetting this is the possibility of increasing tax arrears during recessions) and its broader tax base, which includes business as well as residential property and means that tax rates can be lower for any amount of revenue. Finally, in contrast to the income tax, for which the marginal tax rate for a local income tax could be quite high because the same base is used by the federal government and most state governments, the property tax has the advantage of not being used by higher levels of government.
The local sales tax is an even less promising alternative than the local income tax. Shifting away from the local property tax to a local sales tax is unlikely to make the revenue system more fair. First, the sales tax itself is generally a very regressive tax. Although many states have moderated the regressivity of their state sales taxes to some extent by choosing not to tax food, the state sales tax typically remains quite regressive. There is no reason to believe that a local sales tax would be any less regressive. Whether it is more or less regressive than the property tax is a more complicated question, but one on which there is no clear presumption in favor of the sales tax.
Moreover the local sales tax has some other significant disadvantages. The disparities across districts in sales tax bases are likely to exceed the disparities in property tax bases because of the uneven distribution of large retail shopping centers across school districts. Local sales taxes are relatively easy for taxpayers to avoid and hence can distort the shopping behavior of local shoppers in significant ways. Revenue from sales taxes is typically quite unstable over the business cycle, and finally, sales taxes are not deductible under the federal income tax.
Thus, the property tax appears to dominate the other main alternative broad based local taxes—income and sales—as a revenue source for school districts based on standard evaluation criteria for revenue sources such as equity, effi-
ciency, and stability. However, one final criterion remains—the effect of the tax on the willingness and ability of voters to raise funds for education.
Various constraints have recently been imposed by state governments on the local property tax, in the form of property tax caps and other limitation measures. Many of these constraints have been binding, and they would appear to be a serious indictment of the property tax as a source of revenue for education. Dye and McGuire (1997) and Rueben (1997) provide empirical evidence that these property tax limitation measures are effective at limiting the level and growth of property taxes. The evidence on the effects of these restrictions on school outcomes, such as student test scores, is more mixed. Downes et al. (1998) do not find strong evidence of a short-run effect of the Illinois tax cap on test scores, while Downes and Figlio (1997) find evidence that mathematics test scores are lower in states with binding limitation measures, and Figlio and Rueben (1997) find evidence that teacher quality is lower in states with strict property tax limitation measures. Typically, when tax limitation measures are imposed, the affected jurisdictions are allowed to circumvent the restrictions if they receive voter approval. The effectiveness of voter referenda as an escape valve for jurisdictions hard hit by limitation measures is in question.
However, one must be careful not to assume that taxpayers are revolting against the property tax per se. They may, instead, be concerned about the overall level of taxes and have chosen to protest against the one that is closest to them, the one at the local level. One particular characteristic of such a local tax provides some support for this view. In many jurisdictions, the local property tax is treated as the residual tax, in that its rate is the easiest one to increase when a jurisdiction finds that it has a shortfall between its planned expenditure and the revenue it will receive in the form of state or federal aid or from other tax bases subject to fixed tax rates. The rising property tax rates may well induce local voters to view the property tax as the culprit when the real problem is that expenditure demands are outstripping the growth in other sources of revenue. By this reasoning, one might expect to see limitations imposed on other forms of local revenues, should they become much more broadly used.
Possible Modifications to the Local Property Tax
Given that the reasons for shifting away from the local property tax to another broad based local tax are not compelling, it is worth considering whether various modifications of the local property tax might be desirable. The two most commonly discussed alternatives are shifting the taxation of nonresidential property to the state level and the introduction of local tax base sharing, as in Minneapolis-St. Paul.
The fact that the property tax applies not only to residential but also to business property raises a variety of policy issues for the financing of education. Ladd and Harris (1995) examine the case for shifting the nonresidential portion
of the local property tax base to the state level, with the funds redistributed back to local districts for education. Such a shift might be justified on the following grounds. First, it would recognize the fact that it is residents who receive the primary benefits of locally provided education services, either directly in the form of education services or indirectly through the capitalization of education services into house values. In contrast, because they typically recruit workers from a region larger than the local school district, firms generally receive fewer benefits from local education than do residents. Second, the inclusion of business property in the local tax base could distort local spending by lowering the tax price of education to residents and thereby inducing districts with large amounts of local business property to overinvest in education, relative to the local benefits received by residents. Third, locally differentiated taxation of business property could distort firms' location decisions, as firms seek the districts with the lower tax rates, thereby creating inefficiencies in production. Fourth and more speculative is that shifting to statewide taxation of business property for education with the proceeds channeled back to school districts could generate a fairer pattern of education spending across the state. Whether this outcome would occur depends on the location of the business property and the formula by which the state distributed the revenue back to the local districts.
Ladd (1976) and Ladd and Harris (1995) observe that the impact on distribution as a result of statewide taxation of business property depends on the location of the business property and the structure of the aid system. Ladd (1976) simulates the results of this type of policy change on resource equity in Massachusetts. She concludes that the state would need to provide additional redistributive aid to poorer districts to make up for the lost revenue from the smaller tax base. If business property is disproportionately located in poorer school districts and the state aid system does not compensate, then equity would be reduced by such a proposal. Ladd and Harris (1995) consider the impact of this policy in New York State. They concluded that in order to improve the distribution of resources, the revenues would have to be distributed to the 75 percent of the school districts with the lowest income or property wealth. They noted, however, that such a redistribution program would drain a considerable amount of revenue from the New York City schools. The results from these studies highlight the limitations of these statewide property tax programs.
A variation of this approach is a system in which the tax base generated from new business investment within a metropolitan area is shared among the local communities. Such an approach has been used in the Minneapolis-St. Paul area for the financing of general public services since 1975 (Luce, 1998). Under that program, 40 percent of the new property tax base is put into a regional pool, which is then distributed among municipalities in line with their population and inversely with the market value of their property relative to the rest of the region. While this approach appears to have reduced fiscal disparities in the Minneapolis-St. Paul region, its failure to offset the higher costs of providing services in
some areas has generated some anomalous outcomes, such as that the city of Minneapolis is a net loser under the program. Recent simulations for other areas such as Maryland, Milwaukee, and Chicago generate similar predicted impacts, including the anomalous results for specific cities (Luce, 1998).
WOULD STATE TAXES BE BETTER?
If a larger share of the financing for education were shifted to the state level, states would have to rely primarily on their individual income tax, on their sales tax, or possibly on a new statewide property tax. Alternatively they might try to generate additional revenue from lottery proceeds or selective sales taxes. In the following sections, we explore the equity and efficiency of each of these revenue sources.
Broad-Based State Taxes
The economic incidence of the state income tax is much better understood than that of the property tax, largely because of the typically unchallenged assumption that the burden of the tax is borne by the people who pay it. The typical state imposes a graduated-rate structure with a top rate of between 5 and 8 percent. A few states impose either a flat rate or a very compressed graduatedrate structure on incomes above a low threshold amount, so that the tax burdens in these states are distributed nearly proportionally. More than half of the states with an income tax have an optional standard deduction and most allow taxpayers to use the items reported as federal itemized deductions on their state tax form Because of these characteristics, state personal income taxes are generally considered to be progressive, with the degree of each state's income tax progressivity based on the level of exemptions, number of deductions, and marginal tax rates.5
Since the state income tax rates are in addition to the federal income tax, they can impose significant efficiency costs. After all, it is the total tax rate that affects individuals' decisions about the trade-off between work and leisure, not simply the portion that is paid to one level of government rather than another. Moreover, such efficiency costs rise more than proportionately with an increase in the tax rate, since the further individuals move from what their preferred option would have been, the less well off they are. Hence, a 5 percent state tax rate on top of a 15 percent federal income tax rate would increase the efficiency loss by significantly more than 33 percent. However, working in favor of the income tax is the fact that compliance and administrative costs of the tax will be relatively low provided the state follows the federal definition of taxable income and uses employer withholding to collect the bulk of the revenue.
The stability of the revenue source over the business cycle (short-run elasticity) and how the revenue source responds to economic changes over time (long-run elasticity) also need to be considered. A revenue source that is stable during short-run business cycle instabilities is important to local governments. In addition, revenue sources that grow with growth in the state and local economy are typically preferred to those that grow more slowly. Both of these effects affect the stability of the tax revenue both in the short run and the long run. Revenue growth is required because economic growth will probably increase demands on government services. It is generally believed that an important advantage of the property tax is its relative stability and, therefore, predictability. However, the state personal income tax also seems to perform well in terms of these attributes. Sobel and Holcombe (1996) found the personal income tax base to be fairly stable over the course of a business cycle. They also found that it is very responsive to state income growth.
As for general sales taxes, economists typically assume that they are shifted forward to consumers in the form of higher prices and consequently impose a regressive burden on taxpayers because low-income households devote a larger share of their income to taxable items than do households with higher income. This regressivity is mitigated somewhat in states that exempt various categories of purchases, such as food and drugs, which represent a larger share of spending in low-income households than in high-income households.6 However, this improvement in equity is bought at a loss of tax revenue, since food for home consumption comprises a large share of total potentially taxable sales. Even when all these exclusions from the sales tax base are considered, Phares (1980) estimates the general sales tax to be regressive in each of the states that has one.
The state sales tax also imposes efficiency costs because of the lack of uniformity in the rates and tax bases and because the tax is applied to mobile consumers. These inefficiencies are most certainly larger than with the state income or property tax. Sales tax rates tend to range from 4 to 6 percent. These rates are low, but because some local governments often impose additional sales taxes on the same base, and because the base is narrowly defined, the inefficiencies of the tax are likely to be high, at least in some states.
The state sales tax is also not as stable over the short run or the long run as property or income taxes. The variability and growth of sales tax revenue depend on what items are included in the tax base. On one hand, according to Sobel and Holcombe (1996), retail sales taxes including food purchases have about the same short-run responsiveness as the personal income tax. On the other hand, short-run variability increases when food items are exempted (as they are in many states). Neither sales tax base (with or without food items) performs as well as personal income tax revenue with respect to long-run state income growth.
Revenue instability has become a concern in some of the states that have reformed their education finance systems. In states like California, and more recently Michigan, in which the state has become the primary financier of local school districts, local school districts find that their fortunes are tied to revenue sources that are arguably more sensitive to the business cycle than the property tax, and to revenue sources that they do not have control over. This is because of the restrictions placed on assessments as well as the difficulty (both actual and political) of accurately reevaluating property for tax purposes. In addition, in states that have shifted total financing to the state level, education spending competes for funding with other large state programs.
In part because of these revenue uncertainties, policies that were intended to shift from local to state spending have been partially undone. Arizona and California (and Massachusetts less dramatically) present interesting cases as each enacted major limitations on the local property tax in the late 1970s, which were reflected in large declines in their local shares between 1974–75 and 1984–85. Their local shares increased over the ensuing 10 years so that by 1994–95 Arizona's local share was right at the national average, and California's share was half again as large as it had been in 1984–85. This pattern results from either slowly growing state revenue sources for K-12 education, rising property values (and thus taxable property), or ineffective limitation measures.
In sum, any switch away from local financing of education to the state level will probably require a greater reliance on the state's income or sales tax. Although the sales tax fares less well by standard criteria, especially that of fairness, state legislators often seem to like it, perhaps largely because it is paid in such small amounts along the way and hence is relatively invisible. However, the income tax would be the fairest. Moreover, a recent analysis of the optimal combination of taxes to use at the state level by Gentry and Ladd (1994) shows that the state income tax dominates other state taxes on several key dimensions.
Alternative State Revenue Sources Such as Lotteries
Although income and sales taxes are the workhorses of state revenue systems and are currently the primary generators of state revenue for education, many states also rely on a variety of smaller revenue sources for education. Included among these sources are lottery revenues and selected sales taxes on items such as cigarettes. In some cases, these revenue sources contribute to the financing of education simply as part of a state's general fund. In other cases, they are specifically earmarked for education, as is frequently the case, for example, with lottery revenues.7 Of course that earmarking does not ensure additional funds for
education since the earmarked funds may simply replace funds that otherwise would have been appropriated for that purpose.
Neither lottery revenues nor selective sales taxes are good sources of revenue for education in part because they are incapable of generating significant amounts of revenue relative to the amount spent on education and in part because they generally are quite regressive and often are unstable. Revenue from the lottery imposes a regressive burden because lower-income households on average spend much higher proportions of their income on the lottery than do higher-income households (Clotfelter and Cook, 1989). The regressivity of selected sales taxes varies with the particular item taxed but also can be quite significant for some items, such as cigarettes and alcohol (Poterba, 1989). In some cases, especially when they are first started, the revenue from lotteries may grow quite rapidly. However, over time, revenues level off and may well decline unless a state advertises aggressively and continually introduces new games.8 The revenue from selective sales taxes will vary over the economic cycle to the extent the taxed goods are luxuries, the demand for which rises with income. Although spending on other goods, such as cigarettes, subject to selective sales taxation may be more stable over the cycle, that spending—and hence taxes from that source—may well decline over time as the number of smokers declines. In sum, none of these revenue sources represents a good substitute for a broad-based tax.
A Greater Role for State Property Taxes?
It would nearly be impossible for the country and most states to replace the property tax as the primary revenue source for education. Local governments provide almost 46 percent of all government revenue for primary and secondary public education and the local property tax accounts for over 95 percent of the local tax burden in those states served by independent school districts (see Tables 2–1 and 2–7). Hence, if for equity or other reasons, it made sense to expand the state role, state-level property taxes will undoubtedly have to play a significant role. State property taxes are probably as good or better than state sales taxes, but may be less desirable than state income tax financing. However, in light of the continued reliance on some form of property taxation, states should devote more attention to ensuring that the property tax is fairly administered. Nevertheless, statewide property taxes would be a fairer way to raise revenue than the current system of local property taxes for education. However, we emphasize that whether or not a shift to heavier reliance on state taxes, including state property taxes, is desirable raises many other issues—including the link between state
financing and overall spending levels, the relationship between financing and governance structures, and the effects on the productive efficiency of the system—which are discussed in the following section.
SHOULD STATES PLAY A BIGGER ROLE IN REVENUE RAISING?
An initial consideration for determining whether states should play a bigger role in revenue raising is the impact of such a strategy on the fairness with which revenues are raised. However, potential trade-offs between fairness in revenue raising and other criteria that are embedded in the other goals for a good finance system also need to be considered. In particular, if states were to play a bigger role in funding education, what impact might that have on the system's success in increasing student achievement in a cost-effective way (goal 1), and on its ability to reduce the nexus between family background and student achievement (goal 2)? To the extent that there are trade-offs among the goals, policy makers will need to decide which goal or goals they value most highly in making the decision about the appropriate role of state governments in revenue raising.
Implications for the Fairness of Revenue Raising
We have already argued that the local property tax is not a fair way to raise revenue for education when the relevant equity principle is an ability-to-pay standard. In particular, in a system of local property taxes the distribution of the tax burden across households will be more regressive than if a single statewide property tax were used. In addition, any system of local taxes is likely to give some districts significantly more capacity to generate revenue for education than other districts, and these disparities in turn will translate into spending inequities unless they are offset by carefully designed state aid programs.
Together these observations would seem to suggest that shifting more of the revenue-raising responsibility away from local school districts to the states will increase equity. However, two qualifications are worth noting. The first is that if the increase in state taxes is achieved through heavier reliance on state sales taxes (rather than state income or property taxes), fairness across households could be reduced. This conclusion follows because of the regressive nature of most state sales taxes. If reliance on such taxes is increased, low-income households could well end up bearing a larger portion of the education tax burden than they do with a local property tax. The second is that whether the increase in state financing offsets the inequities associated with reliance on local taxes will depend heavily on how the state distributes state aid for education among school districts. The more equalizing the state aid formula is, the greater the interdistrict equity that emerges from the shift from local to state funding.
In general, we conclude that, although the outcome is not guaranteed, a shift to greater reliance on state revenue sources could well increase the fairness with
which revenues are raised for K-12 education. In addition, however, policy makers need to consider the extent to which this increase in fairness is bought at the cost of other goals for a good finance system or, alternatively, the extent to which it could be used as a way to promote those other goals.
Implications for Raising Achievement (Goal 1)
A larger state role in school finance could affect the education system's ability to raise overall student achievement and the efficiency with which it is produced in at least three ways. First, it could affect student achievement through its impact on the willingness of voters to support spending for education. Second, it could affect the stability of education revenues over the business cycle. Third, it could affect the efficiency with which education is provided.
Impact on the Level of Funding
Designing a finance system that is capable of generating an adequate level of funding for education is of primary importance. The question is the extent to which a shift to greater state financing of education would facilitate the objective of ensuring adequate revenue for education. This concern with the level of funding reflects the committee's view that money can matter—that is, money can affect student achievement if it is used wisely.
From a political economy perspective, one might predict that a shift toward a larger state role in revenue raising might reduce political support for education and thereby reduce the amount of funds available for education. One reason for this prediction is that education would have to compete with a broader array of services for funding at the state than at the local level. Especially in times of economic recession, that competition could potentially be detrimental to education, now that states have been given more responsibility for income support functions. In addition, this prediction reflects differing perceptions of the benefits of education. Voters making decisions about education spending in their local school district are likely to perceive greater benefits than when voters throughout a state make decisions about the level of state spending on education. In the case of the local decision, voters receive benefits in the form of higher-quality education for their own children or, alternatively, in the form of higher house prices associated with the increased desirability of the community to other families who value education. In contrast, for statewide decisions, the benefits to voters are much more diffuse and typically would not include the benefit of higher house prices. Consequently, political support for education could be lower when decisions were made at the state rather than the local level.
Silva and Sonstelie (1995) have developed a more formal political economy model that generates the same conclusion. The key to their model is the distinction between the median voter for local decisions and that for statewide decisions.
Under a decentralized system, the voters within each local school district choose their desired combination of taxes and expenditures. Each decision is the outcome of a majority-rule vote and, under certain assumptions, the relevant decisive voter in the school district is the one who prefers the median level of spending. This voter is generally identified as having the median income level in the school district. The weighted average of all school districts is the ''state average" per-pupil spending level. With many different local districts available, families "vote with their feet" by moving to the school district with their preferred tax and expenditure combination. This migration creates relatively homogeneous school districts and, as they become more homogeneous, the weighted average of the school district medians will be approximately equal to the overall state average spending level or the amount desired by the individual with the state's mean income level. Under a centralized finance system, one tax and expenditure decision is made by the state's median voter or the individual with median income. Because for any typical distribution of income among households, the median is less than the mean, the amount chosen by the median voter at the state level will fall short of the average amount that would be chosen by the separate school districts in a decentralized system.
In addition to this income effect, Silva and Sonstelie identify a price effect of the change in revenue structure. This price effect is caused by the shift of the relevant tax base from the local property tax toward either a state income or a state sales tax. However, the size and direction of the price effect are unclear, since they depend on tax deductibility provisions (e.g., property and income taxes are deductible from the base of the federal income tax, but sales taxes are not) as well as the relative tax progressivity of the various taxes. The higher the progressivity of a tax, the lower would be the price to the median voter and the greater would be the willingness of that voter to support education. Because state income taxes are likely to be more progressive than local property taxes (yet both are deductible), this price effect could conceivably lead to greater support for education by statewide voters than by voters at the local level. Whether or not it does is an empirical question.
Also working to increase expenditures at the state level could be the economies of scale (described by Heise, 1998) gained by education lobbying groups under centralization as they are able to focus on a single state legislature rather than the far more numerous local school districts.
For empirical evidence, many observers have turned to California, where court-ordered reform eliminated most of the disparities in education spending across districts and increased the state role in financing. In that state, the subsequent change in spending was clear: it fell quite dramatically relative to what it otherwise would have been. However, for reasons we explain below, it would be a mistake to generalize from the California experience. In fact, we conclude that, in practice, a larger state role in education finance has led to higher spending in more states than it has led to lower spending.
The California experience (see Chapter 3) is worth describing because that state was the first to face successful court challenge to its system of education finance. In that case, Serrano v. Priest, the court severely restricted the degree of spending inequality across districts (to be less than $100 in 1971 dollars), allowing only for inflationary effects and variations in categorical aid. In effect, th state would determine the level of school district spending. The change in spending in California in the post-Serrano era has been dramatic. Rubinfeld (1995) showed that in 1971–72 California spending per pupil was 98 percent of the national average and that California ranked 19th among the states; by 1991–92, California spending was only 86 percent of the national average and the state's rank had fallen to 39th.
Silva and Sonstelie (1995) used regression techniques to attempt to measure for California the size of the price and income effects described above. They found that prior to Serrano in 1969–70, spending in California was similar to other states (after adjusting for differences in family income and tax prices). In 1989–90, however, they found that spending was significantly lower in California than they would have predicted. They estimate that roughly one-half of the decline in per-pupil spending in California can be attributed to Serrano. They attribute the remainder of the decrease to the growth of the California student population during the 1980s.
For a number of reasons, however, one should be careful about generalizing the California experience to other states. Most important is that, in 1978, California voters passed Proposition 13, which severely limited the level and rate of growth of local property taxes; they subsequently passed Proposition 4, which limited the rate of growth of state taxes. Together these changes rendered it extremely difficult to raise revenue in the state. While some authors (e.g. Fischel, 1996) attribute the passage of Proposition 13 to the Serrano decision, other plausible explanations are readily available. Those explanations start from the observations that prior to Proposition 13 housing values were rising very rapidly, that property tax assessments were rising almost as rapidly but in an uneven manner as only one-third of the residential properties were reassessed each year, that local public officials did not lower nominal property tax rates in proportion to the increases in property tax assessments, that the rapid rise in housing prices resulted in some shifting of the local tax burden away from business property onto residential property, and that the state had a large surplus with which it could have provided tax relief to local taxpayers. Together, these facts provided voters with plenty of reasons to be angry about their rising property tax burdens and to be frustrated with the local governments, the state government, or both.
Other changes make California unique as well. Because it was the first state supreme court decision in school finance, the Serrano court decision clearly did not involve the California legislature. After the Serrano decision, legislatures in other states may now view the court as a partner for the purpose of changing the education finance system. In addition, California experienced a large increase in
its student population during much of the 1970s and the first half of the 1980s, while the rest of the country was still experiencing a decline in student enrollment.
At least one study (Joondeph, 1995) provides some evidence to suggest that the relative decline in California's spending after finance reform was not unique to that state. Comparing the growth rate in current expenditures for schools over a 20-year period in five states whose education systems had been found unconstitutional prior to 1984—California, Washington, Connecticut, Arkansas, and Wyoming—he found that, in four of the five states, spending on education grew at a slower pace than in the nation as a whole. His study also revealed that, with the exception of Connecticut, funding increased the least in those states that reduced interdistrict disparities the most (as measured by changes in the Gini coefficient). However, his analysis is based on simple correlations that do not capture the effects of alternative pressures on education spending. More sophisticated multivariate techniques are needed to sort out the causal impact on spending of the shift to a larger state role in the financing of education. In addition, his study focused roughly (depending on the availability of data) on the period from the time of the court case to the 1991–92 school year, giving him a somewhat different number of years to analyze across the five states.
Other researchers address more broadly the question of whether the California pattern is generalizable. Although the results from this literature are a bit mixed, on balance, the studies were somewhat more likely to find that finance reforms increased state spending than the reverse. For example, Downes and Shah (1995) document that the stringency of constraints on local discretion determines the effects of reforms on the level and growth of spending, and they used their results to estimate the effect of legislative and court-ordered reforms in California and Arkansas. They conclude that court-ordered reform in California reduced spending in real terms in 1990 by $640. In contrast, in Arkansas they concluded that court-ordered reform increased spending by about $40 above what they otherwise would have predicted.
Manwaring and Sheffrin (1997) examined similar issues in a dynamic model of state education spending between 1970 and 1990. In their model, successful litigation raised real per-pupil spending by $26 per year of 0.64 percent of expenditures, and education reform raised per-pupil spending by $106 per year or 2 percent of expenditures (in 1990 dollars). They used their empirical results to estimate the impact of legislative and legal reforms in individual states and find such reforms have raised real per-pupil spending in 14 states and reduced it in only 5 states. They note "our findings suggest that states do have some flexibility in choosing policies that move in the direction of equalizing spending per pupil without leading, inevitably, to an overall decline in per-pupil spending" (1997:123).
Similarly, Murray et al. (1998) concluded that successful state litigation increased state per-pupil primary and secondary education spending by $88 (in
1992 dollars) while decreasing significantly within-state spending inequality. Their results also suggest that the highest-percentage increase in spending would be in the poorest school districts and spending in the wealthiest districts would remain constant. Overall, they found that the state's share of total spending rose as a result of court-ordered reform, with funding for this increase in the poorest districts coming from higher taxes. Contrary to these results, Hoxby (1996b) estimated that the average level of per-pupil spending fell with more effective equalization, as a result of the large disincentives on high-demand school districts that are contained in these plans.
In light of this mixed evidence, the committee itself looked at the patterns of spending over time in all of the states that have had court-ordered reforms. Instead of following the Joondeph approach of picking a selected set of years, the committee examined annual average growth rates of spending in each state relative to the national average for specified numbers of years since the state's court case. We believe this approach better answers the question of how a court case affected spending, say, 5, 10, or 15 years after the court case. The results for all the states experiencing court-ordered reform are shown in Table 8-2. Up to three entries are provided for each state. The first one shows the difference in the average annual growth rate in per-pupil expenditures in that state from the U.S. average five years after the state's court case. The second entry for each state is the difference after 10 years, and the third is the difference after 15 years. The dates cited for the court cases are relatively straightforward except for California, for which we could use the 1971 date of Serrano I or the 1976 date of Serrano II,
TABLE 8-2 Annual Average Growth Rates (Relative to National Average) in Per-Pupil Expenditures Following Court-Ordered Reform
which led to the virtual equalization of spending across districts. 9 For the table we have used the 1976 starting date. Had we used the 1971 starting date for the California figures, the first entry for California would have been a positive number. That is, for the five years after the first Serrano case, per-pupil spending on schools in California grew faster than the national average.
The table shows that for the 10 states for which data are available five years after the court case, only two (California and Montana) experienced rates of spending growth that were below the U.S. average growth rates. For the seven states with reforms early enough for there to be 10 post-reform years, the results are more mixed: spending in three states (California, Washington, and Wyoming) grew more slowly than the U.S. average, and spending in four states (New Jersey, Connecticut, West Virginia, and Arkansas) grew faster. Finally, after 15 years, three states exhibited faster spending and three states slower growth.
In summary, the committee believes that there is sufficient evidence to reject the conclusion that greater equalization of spending across local school districts will necessarily reduce spending on education. While spending in California did indeed fall relative to the spending in the nation (at least after 1976), the case for attributing that to the Serrano decision is not compelling. Moreover at least half of all states experiencing court-ordered reform have increased their spending relative to the national average over time. Greater equalization of spending is not incompatible with higher state spending for education.
Impact on the Responsiveness of the Tax Base
An additional consideration regarding a greater state role in education finance is what it will do to the stability of revenues over the economic cycle and to the growth potential for revenues over time as the economy grows. The flexibility of the base matters from a political perspective because it is politically much easier to raise revenue through growth in the tax base than through increases in tax rates. In general, the property tax base is quite stable over the business cycle and its growth over time depends heavily on assessment practices. Sobel and Holcombe (1996) find the personal income tax base to be fairly stable over the course of a business cycle and that, of all the main taxes, it appears to respond the most to growth in state income. This responsiveness is beneficial when a state's economy is growing but could be harmful during periods of slow state economic growth. According to Sobel and Holcombe (1996), the variability and growth of sales tax revenue depend on what items are included in the tax base. Retail sales taxes, including food purchases, vary over the business cycle in a manner similar to the personal income tax, but they vary much more when food
items are exempted. In general, sales tax bases (with or without food) are not as responsive to state income changes as the personal income tax, largely because they exempt spending on most services, which are the fastest growing part of the economy. Thus, in states like California, and more recently Michigan, where the state has become the primary financier of local school districts, local school districts find that their fortunes are tied to revenue sources that are arguably more sensitive to the business cycle than the property tax, and to revenue sources that they do not have control over.
This consideration is not inconsequential in weighing how best to design a financing system that raises revenues fairly but also generates sufficient revenue to foster the goal of facilitating high learning for all students. While a shift to a larger state role is likely to enhance equity, the increased responsiveness of the revenue system to changes in the economy could present serious problems in the event of an economic downturn. The absence of a serious economic downturn in the past 8–9 years provides little basis for prediction, but given that downturns in the economy are inevitable, one should not ignore their potential effects on the education finance system. The lesson from the research is that shifting away from a local property tax to a narrowly defined state sales tax (one with many exemptions including food) would put education revenues in the most jeopardy.
Impact on Cost-Efficiency and Student Outcomes
Independent of the impact on funding levels, would we predict that a larger state role in financing would increase or decrease student outcomes? At least two conceptual arguments can be made that a larger state role will decrease the cost efficiency of the system. Fischel (1996) and Hoxby (1996b) argue that because school quality is capitalized into higher home values, all homeowners (even those without children) are interested in improving their local schools when schools are financed locally. As a visible outcome of schools, test scores serve as a measure of school quality. Local homeowners can hold school administrators accountable for using higher property taxes (which will reduce housing values) effectively to improve local public schools (which will raise housing values) when improvement is measured by higher test scores. With a shift to a greater reliance on state revenue or if local control over spending is otherwise reduced, the incentives for this monitoring by all homeowners are removed and test scores may suffer.10
A second argument is that greater state involvement in financing will bring with it greater state control over the mix of inputs to be used by local school districts. For example, states may require districts to have certain class sizes, to
hire certain types of teachers, or to have libraries of a certain size. Given our conclusion in Chapter 5 that knowledge about the education production process is imperfect and that the effectiveness of certain inputs may vary with the specific context, such central directives might well keep schools from producing education in a cost-efficient manner.
Working in the other direction, however, is the possibility that the shift to a greater state role in financing could be accompanied by changes in school governance that give the state more authority to hold schools accountable for high achievement standards while giving them more flexibility to manage themselves. Such would be the case, for example, if the shift to greater state financing were part of an overall standards-based reform strategy. In that case, the greater state role (in financing and governance) could well lead to greater student achievement.
In practice, how much control states exert over local schools and the form of that control varies greatly from one state to another. Some states (such as California) control the total amount of spending by imposing spending restraints. A number of recent state education finance reforms have included some type of maximum per-pupil spending level or some type of limit on the growth rate of spending. Alternatively or in addition, state governments impose various regulations such as mandating a particular curriculum that students must complete, setting a minimum number of days and/or hours that students must spend in school, requiring specific treatments for some groups of students, granting teacher certification, imposing work rules for teachers and other school employees, setting standards for the provision of transportation and meals, and influencing or limiting capital investment decisions through controls on borrowing or debt. All of these regulations reduce the autonomy of local schools and, in most cases, influence the financial decisions of the schools. In evaluating the following evidence, it is important to bear in mind that to the extent that a larger state role is associated with more state control, it is difficult to sort out empirically the effects of centralizing finance from centralizing of governance.
Several researchers have investigated the relationship between the state role in financing and student outcomes by including various measures of centralization as explanatory variables in a standard educational production function regression. They typically find that a larger state share reduces student outcomes.
In two regression studies, Peltzman (1993) found that an increase in the state's share of education revenue lowered state SAT scores in the 1970s, but these variables were unrelated in the 1980s. In his examination of the performance of noncollege-bound students, Peltzman (1996) found that an increase in the state expenditure share was associated with a decline in scores on the Air Force Qualification Test. Fuchs and Reklis (1994) found that math scores are higher in states in which the state share in education revenues was lower. Hoxby (1996a) argues that the state revenue share is an inaccurate proxy for state control. For example, even though local governments in California collect property
tax revenue, the Serrano decision took away local control over how much they spend. Irrespective of that lack of local control, California's state share in 1995 is listed at only 56 percent (Table 2-2). Rather than state share, Hoxby included a measure of the local government's tax price for an additional dollar of education expenditure (where the local tax price in California would be infinite since the spending of local districts is fully constrained). She found that an aggressive power equalization plan (i.e., higher tax price) had undesirable impacts on student outcomes in that it raised the student dropout rate by 3 percent and that a move to fully state-financed schools raised the dropout rate by 8 percent. Husted and Kenny (1998) included a measure of state education spending inequality in their educational production function and conclude that the mean SAT score is higher in those states with greater within-state variation in spending.
A second approach to exploring the relationship between a larger state role and student outcomes is to look at changes in outcomes that follow court-mandated reform. This second approach is more general in that it captures the effects not only of finance reform but also of any accompanying changes in governance. The disadvantage of this approach is that the effects of the finance changes cannot be separated from those of the governance changes. In some cases (for example, Kentucky after its recent court case that declared unconstitutional not only its school finance system but also its governance system), both changes were large. Downes (1992) looked at the California experience following Serrano and found that greater equality in spending was not accompanied by greater equalit in measured student performance. Using individual-level data from the National Longitudinal Survey of the High School Class of 1972 (NLS-72) and the National Educational Longitudinal Survey (NELS), Downes and Figlio (1997) estimated that court-mandated school finance reforms did not result in significant changes in either the mean level or the distribution of student performance on standardized tests of reading and mathematics. They do find, however, those legislative reforms that are not a result of a court decision lead to higher test scores in general; the estimated effect was particularly large in initially low-spending districts.
In sum, there is no reason to think that a shift to a greater state share of funding (with no change in average funding level) will lead to greater achievement unless it is connected with policy changes designed to encourage that end. If state funding is not connected with such changes, the shift to a greater state share of funding may reduce the productive efficiency of the system, as local school districts may have less incentive to use their resources carefully given that fewer of those resources are coming directly from local taxpayers. The bottom line is that states that shift to more state funding in the name of fairness need to ensure that they do not inadvertently reduce the productivity of their education system in the process. Only with appropriate policy changes can they avoid that outcome.
Considerations Related to Goal 2: Breaking the Nexus
Many of the considerations related to how increasing the state role would affect goal 1 also apply to goal 2: breaking the nexus between family background and student achievement. Here we consider more directly the extent which a larger state role in financing can help ensure both equitable and adequate funding for disadvantaged students, particularly those who are concentrated in districts with limited resources relative to their educational needs.
One of the major justifications for a larger state role in education finance is that the state revenues can be used to offset what otherwise might be undesirable disparities in education spending across school districts. Court-ordered education finance reforms have frequently increased the state financial role and led to a more equitable distribution of spending than would have occurred in the absence of the reform. Assuming that money can make a difference in student achievement, this redistribution of resources to low-spending districts could potentially increase their students' achievement.
Current state aid is often not designed to compensate fully for the differences in the costs of providing education across school districts. This observation suggests that states that increase the role of the state in the financing of education must work hard to ensure that the funds are distributed across school districts in ways that take account as fully as is technically possible any differences in the costs of educating students. Failure to do so will be detrimental to the districts serving disproportionate numbers of difficult-to-educate children and will hinder efforts to reduce the nexus between family background and student achievement.
A second consideration is the political economy of grants-in-aid and its implications for the state's ability to ensure adequate funding for all districts, particularly those that are unable to supplement state funds from local sources. As Courant and Loeb (1997) describe, the extent to which centralization of school finance permits local districts to supplement the state funding has potentially significant implications for the likelihood that the state funding will be adequate for the poor districts. Consider a situation in which local districts are free to supplement the state funds in an unlimited way. One possible outcome is that many high-demand voters in rich districts will try to keep state spending on education relatively low. They do so out of their own self-interest. The higher the level of the state funding is, the greater is the proportion of their tax dollar that goes to pay for the education of children in other districts. Hence, they prefer a lower state amount so that they can direct their taxes to providing services in their own local district. According to Courant and Loeb, whether there is support for a state foundation grant at a level that would truly provide an adequate level of basic education depends on the number of voters in high-demand school districts and their political influence. This problem is alleviated when local supplementation is limited so that voters from high-demand school districts are restricted in their ability to substitute greater local revenue for general state grants.
A more extreme reaction by high-demand families who are restricted by state financing programs designed to equalize spending across districts is to opt out of the public school system altogether and send their children to private schools. As Brunner and Sonstelie (1997) explain, such equalization plans restrict the variation in public school spending across districts and they keep some families from obtaining their preferred level of education through the public sector. The only way for families to achieve their preferred level of education is by choosing a private school. Since these families no longer use the public school system and the statewide financing system eliminates benefits in the form of rising housing values, their support for the public education system is likely to shrink.
The empirical evidence on this question is mixed. Most of the research has focused, once again, on the experience in California after Serrano . The raw data suggest that Serrano has not led many families to choose private schools. Brunner and Sonstelie (1997) observe that about 9 percent of California schoolchildren were enrolled in private schools in 1973–74, compared with roughly 10 percent in the rest of the country. By 1992–93, private school enrollment had increased to about 10 percent in California and 12 percent in the rest of the country. They conclude that private school enrollment in California basically followed the national trend. Downes and Schoeman (1998), however, come to a different conclusion. They argue that even if the supply of private schools did not increase, Serrano could account for nearly half of the actual movement from public to private schools in California over the 1970–80 period.
These considerations highlight the importance of variation in individual demands for education and the fact that families with high demands for education are able to behave in ways that may be beneficial to them but potentially harmful to the goal of raising sufficient revenue to promote higher levels of achievement for all children, and particularly for disadvantaged students. To minimize the negative side-effects of the decisions of high-demand families on the overall level of support for education, states could (1) prohibit or severely limit local districts from supplementing state aid from local taxes (as was done in California) and (2) prohibit families from opting out of the public system in favor of private schools. However, such policies would run counter to some deeply rooted values in American education related to freedom of choice. Hence, the challenge is to design a system that ensures adequate funding for disadvantaged students but does not run roughshod over other values that people hold dear.
SHOULD A GREATER SHARE OF FUNDING FOR EDUCATION COME FROM THE FEDERAL GOVERNMENT?
A larger state role in financing primary and secondary education will only partially address education resource inequities in the United States. A case can also be made for a larger federal role in revenue raising for education, based on either an equity or an adequacy rationale. The two rationales lead to two alterna-
Equity Arguments for a Greater Federal Role
Since enactment of the 1965 Elementary and Secondary Education Act (ESEA), the federal government has played a significant role in funding schooling for specific groups of at-risk students. The case for targeted support for such students rests largely on a redistributive or equity rationale. In such instances, the highest level of government is most appropriate as the source of revenues. In part, the argument for this is simply practical. It is much more feasible for the federal government to play that role than it is for lower-level governments because of the possibility of movement of households among subnational jurisdictions.
For example, consider a large city school district with a large percentage of economically and educationally disadvantaged students. Given that it costs more to provide a given level of education to those than to other students, the city would have to spend significantly more per pupil than other school districts. This condition is likely to lead to higher tax burdens. Those high taxes would provide an economic incentive for middle-and upper-income households as well as businesses to move out of the city (or to choose not to move there in the first place). Consequently, the city school district would become increasingly impoverished and unable to do much redistribution of funds away from wealthier residents to the disadvantaged students. In the extreme, there would be no nonpoor residents in the city to support educational services for the poor. This behavior is likely to occur at the state level as well as the local level, but is much less common at the federal level given that taxpayers would have to move out of the nation to avoid the burden of paying higher taxes to support needy students.
Related to this practical position is the ethical argument that poverty and its associated education characteristics are national problems that deserve national attention. While some districts would be wealthy enough to fund additional programs for disadvantaged students within the district, one can ask whether it is fair to make the people who happen to live in those districts or states bear the financial burden while others have opted out of such payments by moving to areas with low proportions of at-risk students. To the extent that poverty and other measures of educational disadvantage are national problems, the fair way to fund their alleviation is with a national tax, such as the federal income tax, which would spread the financing burden among all U.S. residents in line with their ability to pay, regardless of where they live.
This federal equity rationale is already reflected in categorical programs for specified groups of students, that is, students who need additional educational services in order to achieve to acceptable levels. Largely developed during the
second half of this century, these programs often were initiated at the federal level in part because the students involved were not being served sufficiently by states and local school districts. Such programs include Title I, focused on low-achieving students from economically disadvantaged backgrounds; the Individuals with Disabilities Act (IDEA), concentrating on students with physical and mental disabilities; ESEA Title VII, directed at students whose native language is other than English; and several other related programs. Today, these programs for students with special needs permeate education strategies at the school, district, state, and national level, nearly always complemented by additional state and local funding. For most professionals in education today, these federally funded programs are simply a part of the infrastructure.
However, as discussed in Chapter 7, funding of these programs is still problematic. Although the federal government took the lead role in creating these programmatic emphases, it has never funded them at their fully authorized level. For IDEA in particular, this underfunding has created large financial burdens for states and local school districts. Thus one option for the federal government is to assume a larger share of this particular educational burden. Another would be to fund Title I at the full amount authorized by Congress, $24.3 billion annually (Independent Review Panel on the Evaluation of Federal Education Legislation, 1999:12). These additional federal funds could be used to expand school services for students with special needs and would free up state and local funds to expand the overall level of school services to all students.
Adequacy Arguments for a Greater Federal Role
A second argument for a greater funding role for the federal government emerges directly from analyses of funding inequities (Chapter 3). In their examination of 16,000 school districts, Evans et al. (1997) found that while within-state inequality fell slightly between 1982 and 1992, between-state inequality rose sharply. State government policies that are designed to improve intrastate inequality are not likely to improve interstate spending inequality. Only if education finance reform in states that is intended to reduce intrastate variation also raises the state average spending level and if the states pursuing such policies are those with relatively low per-pupil spending levels, would state-specific reform efforts reduce interstate differences. But these conditions are not always satisfied. Serious proposals to correct interstate inequality are most likely to require an increased federal role in financing education.
Thus, this second rationale leads to an alternative policy option for a larger federal role, namely ensuring that all states can adequately fund their schools. The federal government will face the same challenges (described in Chapter 4) in determining an adequate level of per-pupil revenues for a district or school with the typical mix of students. Nevertheless, there are proxy measures that could be used in the meantime. Odden and Busch (1998), for example, suggest the na-
tional median level of basic education revenues per pupil. Although whatever measure is used at this point would be imprecise, it would represent an acknowledgement that only an enhanced federal role can address interstate funding inadequacies. Especially in an era when the nationwide education goal is to teach students in all states to high standards, the time may have come to consider a new federal role in education on the basis of educational adequacy.
Such an approach would call for a new federal foundation program. An adequacy rationale at the state level leads to a foundation type of school finance structure; the state ensures that each district has an adequate level of education revenues so a district can educate an average student to specified performance standards and then would adjust this foundation amount by a factor that accounts for the higher costs of both students with special educational needs and geographic price differences for the educational inputs purchased. Each district would need to contribute financially to such a foundation base by making a required maximum tax effort. Districts that could not generate an adequate level of resources with the required tax effort would receive state aid to subsidize the difference.
A new federal role could be similar. First, federal policy makers would have to define a federal foundation level of spending that would be adequate for the state or district with the typical student. They would then need to adjust this base level by a factor that accounted for differences in pupil needs as well as in educational input prices across states. The federal government could ensure that each state could generate the foundation level of funding by giving each state federal aid equal to the difference between the foundation level and the revenue that the state would generate based on a minimum tax effort. Provisions would also be needed to ensure that states distributed revenues to districts and schools so that they too had an adequate amount of revenues per pupil. To be sure, considerable analysis would be needed to determine nationwide, cost-and price-adjusted revenue per-pupil amount for each state and district and how the minimum tax effort would be defined. Nonetheless, the basic approach should be clear.
Either of these policy options would require a substantial increase in federal revenues. Because of the existence of large federal surpluses at the current time, these ideas might have arrived at a fortunate time. Nevertheless, the politics for funding such new initiatives can be expected to be contentious.
Increasing the fairness with which revenues are raised for education will almost certainly require a greater revenue-raising role for states and the federal government. However, as with any change, there are trade-offs to be considered. The good news from our analysis is that, in some cases, the trade-offs are not so stark as some people have suggested. For example, the concern that increased
centralization of financing at the state level will inevitably lead to lower state spending on education, an inference drawn from the California experience, is not supported by the evidence from other states. Nonetheless, some trade-offs remain. Of most importance for harnessing education finance to the broader goals of education policy is the need for policy makers to pay close attention to ensuring that changes in financing mechanisms do not weaken the incentives for districts or schools to be vigilant about the productive efficiency of the system and that intergovernmental aid programs are carefully designed to promote the goal of reducing the nexus between family background and student achievement.