Measuring non-school fiscal disparities among municipalities.

INTRODUCTION

Fiscal disparities among parallel governments--municipalities within a state or states within the nation--represent a policy concern when the overlying government (state or nation) has an interest in levels of services and/or tax rates across its subdivisions. According to Yinger (1986), "fiscal disparities exist when some cities can provide a given level of local public services at lower cost or at less sacrifice than can other cities." In Massachusetts, residents are served by 351 mutually exclusive and exhaustive cities and towns; each municipality is responsible for the entire range of common local services, including schools, police and fire protection, public works, human services, libraries and recreation, etc. (counties and other overlying sub-state governments are virtually nonexistent). This very local character of city and town government makes Massachusetts prone to substantial inter-jurisdictional fiscal disparities. As other research indicates and as we show below, both the environment in which these services are produced and the resources available to finance them vary markedly across municipalities in ways that affect the cost or tax burden involved in providing common services of average quality. The resulting fiscal disparities lead to inequity and inefficiency, which are mostly outside the control of local governments (Yinger, 1986). Previous research suggests that the overlying government may counteract such disparities and improve social welfare through intergovernmental aid (e.g., Bradbury, Ladd, Perrault, Reschovsky, and Yinger, 1984; Yinger, 1986), and Massachusetts state government has a history of providing general-purpose municipal aid, in addition to earmarked school aid.

Partly because increases in state aid were concentrated on public schools during the 1990s and partly because state budget cuts during recent economic downturns have reduced non-school aid and generally made it unpredictable, cities and towns in Massachusetts are under considerable fiscal strain. In recent years, many communities instituted layoffs and cut services, and many raised property taxes (see Municipal Finance Task Force, 2005; Bluestone, Clayton-Matthews, and Soule, 2006). Because the two existing general-purpose municipal aid formulas are viewed as being unresponsive to communities' needs, (1) policymakers are exploring alternative formulas for distributing non-school aid.

Formulas to allocate local aid in most states-whether for schools or for general municipal purposes-typically distribute funds in inverse proportion to each jurisdiction's ability to raise revenue locally (that is, local revenue capacity) and/or in proportion to its needs or costs. (2) The revenue-capacity side of the calculation typically takes account of the size of the tax bases that local governments are authorized to tap. The cost/need--side is generally extremely simple or ad hoc, often using population alone (or enrollment in the case of school aid) to represent need. Our research develops new measures of both elements to reflect more accurately the realities and constraints of local government in Massachusetts.

To improve the measurement of revenue capacity, this paper focuses on accounting for the constraints of a tax limitation, for the first time in the literature. We incorporate the effects of Proposition 2 1/2--the local property tax limitation in Massachusetts--in our measure of local revenue capacity. Proposition 2 1/2 caps property tax rates and limits year-to-year growth in property tax revenues of each city and town, but allows local voters to override the growth limits. Because Proposition 2 1/2 differentially limits individual cities' and towns' ability to generate revenue from their property tax bases, many local officials argue that the existing lottery aid formula, which distributes aid in inverse proportion to communities' per capita property tax bases, is unfair. On the other hand, Proposition 2 1/2's annual levy limits cannot be treated as exogenous binding constraints because local voters can, and in many communities do, override them. Our research uses data on recent taxing patterns to develop a measure of the effects of the constraints on local property taxation--a measure not in the control of local officials or local voters--and uses that measure to adjust the size of the local property tax base in measuring revenue capacity. Also different from the standard approach, in which revenue capacity depends only on the size of the property tax base, our measure of revenue capacity takes account of the ability to raise revenue from other (non-property-tax) local sources. In addition, we explicitly remove the capacity not available for general municipal (non-school) purposes such as the funds the state government requires communities to dedicate to public schools.

On the cost side, unlike previous studies of either total spending or school spending, our research focuses on non-school spending. Following Bradbury et al. (1984), Yinger (1988a, 1988b), Ladd and Yinger (1989), and Ladd, Reschovsky, and Yinger (1991), (3) we use regression analysis and recent data to quantify the relationship between local non-school spending and environmental cost proxies that are outside the control of local governments, such as population density and commuters per capita. Our regressions also control for demand, efficiency, and institutional factors that influence spending.

The difference between these measures of municipal costs and capacity for each community represents a new measure of non-school fiscal "gap," which can provide the basis for a new aid formula to channel more aid to communities with larger gaps. For such a purpose, the measures of costs and capacity must be predetermined-not in the immediate control of local governments; otherwise, the formula would reward communities that are inefficient or put less effort into raising revenue, and would create incentives for local officials to change behavior in ways that increase the amount of aid-by adding to measured costs or reducing measured capacity. This search for predetermined--and still relevant--indicators is a critical challenge in measuring both costs and capacity. As discussed below, the estimates that are produced by our methodology may be subject to some of the same biases seen elsewhere in the literature. But Louis, Jabine, and Gerstein (2003) argue that developing rational and systematic estimates of cost and capacity, even if they are flawed, is still preferable to the ad hoc adjustments adopted in most formulas used to allocate intergovernmental aid.