The ability of individuals to move freely from one jurisdiction to another is generally seen as a constraint on the amount of redistribution that each jurisdiction within a system of governments can undertake. In this paper, we look at this proposition by developing a positive analysis of income redistribution by local governments in a federal system. We ask how much redistribution occurs when only local governments can have tax/transfer instruments, individuals can move freely among jurisdictions, and voters in each jurisdiction are fully aware of the migration effects of redistributive policies. Local redistribution is shown to induce sorting of the population, with the poorest households located in the communities that provide the most redistribution. While the threat of out-migration affects the potential for redistribution, our results suggest that significant local redistribution is nonetheless feasible. Numerical computations indicate that the proportion of residents who are renters is a major factor affecting the local choice of level of redistribution.

The paper investigates whether compensation among local jurisdiction is, by itself, sufficient to ensure efficient provision of local public goods. Jurisdictions have fixed boundaries, and each has an entrenched government with the power to tax and supply the public good. Residents can move costlessly among jurisdictions. It is shown that competition among numerous jurisdictions is not sificient to guarantee public sector efficiency. Though residents can "vote with their feet," land is immobile. Hence, governments can unsurp some land rents for their own ends. Increasing the number of jurisdictions limits but cannot completely eliminate the ability to exercise discretionary governmental power.

A theoretical and computational model with tax-financed, tuition-free public schools and competitive, tuition-financed private schools is developed. Students differ by ability and income. Achievement depends on own ability and on peers' abilities. Equilibrium has a strict hierarchy of school qualities and two-dimensional student sorting with stratification by ability and income. In private schools, high-ability, low-income students receive tuition discounts, while low-ability, high-income students pay tuition premia. Tuition vouchers increase the relative size of the private sector and the extent of student sorting, and benefit high-ability students relative to low-ability students

Government may provide a good that can, if legally permitted, be supplemented by private purchases. Policy is determined by majority rule. Under standard assumptions on preferences, a majority voting equilibrium exists. A regime of positive government provision with no restriction on private supplements is shown to be majority preferred to a regime of either only market provision or only government provision. Combined public and private expenditure on the good is higher under this dual-provision regime than under either of the alternatives. Under some preference configurations, the median-income voter is pivotal; under others, a voter with income below the median is pivotal.

This paper derives an econometrically meaningful test of the Tiebout hypothesis and demonstrates that previous tests are inappropriate. The implications generated by two closely related models of voter-determined local fiscal variables and individual resident housing choices are compared. We demonstrate that when the Tiebout mechanism operates without interference, housing quantity and location choices are Pareto efficient, while they are not when frictions interfere with its operation. We show that the appropriate test requires joint estimation of a set of structural equations determining housing purchases and locational choices utilizing data for both median and nonmedian voters across metropolitan area jurisdictions.

Having discovered and neatly portrayed the conditions for efficient provision of public goods, Paul Samuelson concluded that there was no viable mechanism for eliciting the information about preferences required to determine optimal public goods supply. Much subsequent theorizing about public finance decisions may be viewed as a quest for some demand-revealing mechanism which would refute Samuelson's assertion. Charles Tiebout, in particular, countered with the claim that a viable mechanism did exist for determining the optimal supply of local public goods. In his view, selfinterested individuals reveal their preferences for local public goods by their choice of jurisdiction. Tiebout reasoned that each individual adopts, from the menu of local fiscal environments, that local community which most closely reflects his preferences. He further argued that the greater the number and variety of communities, the more efficient would the public goods provision be. Without empirical support, this ingenious idea lay dormant for several years. The question remained as to whether the mechanism described by Tiebout, even if it existed, operated so that local jurisdictions effectively elicited and satisfied preferences for local public goods. Fundamental to answering this question was the necessity of designing an experiment which would reveal the result of voting with one's feet. No such experiment existed until Wallace Oates (1969) reasoned that operation of the Tiebout mechanism would result in the capitalization of differentials in fiscal variables into property values. His test findings indicated that indeed such capitalization was evident. At first, the results obtained by others who repeated and extended his experiment were generally consistent with his findings and hence his interpretation of the link between the Tiebout mechanism and the capitalization of fiscal variables. It was not long, however, before an alternative interpretation was suggested by Matthew Edel and Elliot Sclar, who argued that in the long run the Tiebout mechanism would result in no capitalization of differentials in taxes and public service levels. Their interpretation was precisely the opposite of that offered by Oates, and their empirical work -which indicated little evidence of capitalization-supported their interpretation. In retrospect, conflicting interpretations of the Tiebout hypothesis were almost inevitable given the absence of a formal model embodying the Tiebout mechanism. In an earlier paper with Michael Visscher, we argue that Tiebout's hypothesis is as follows: If individuals are able to choose from among a multiplicity of local jurisdictions, the result will be a Pareto-efficient provision of local public goods. We present two alternative models which differ in the technology for production of local public goods. One embodies the efficiency properties claimed by Tiebout, the other does not. Using these models we demonstrate that 1) Tax capitalization cannot be tested by the procedure proposed by Oates because the tax capitalization parameter is not econometrically identified-a possible reason for the contradictory empirical findings of Oates and of Edel and Sclar, and 2) A test of capitali*Associate professor, Carnegie-Mellon University, and assistant professor, Tulane University, respectively.