The Implications of Inequality for Fiscal Federalism (or Why the Federal Government Should Pay for Local Public Schools)

45 Pages Posted: 22 Jun 2018 Last revised: 18 Sep 2019

See all articles by Brian Highsmith

Brian Highsmith

Arthur Liman Center for Public Interest Law; Harvard University, Department of Government

Date Written: April 1, 2019


In designing public policy, a question of first principle is the degree to which government services—and the mechanisms of collecting revenue to finance those services—should be centralized within and across political systems. To inform their assessments of where redistribution properly should occur, public finance researchers have to date worked backwards from different assumptions about the mobility of residents within the political community. Scholars have disagreed about the viability of local governments’ efforts to redistribute wealth—with traditionalists arguing that these efforts are made impossible by residential mobility, and recent reformists countering that limitations on mobility indeed allow for limited redistribution at the local level.

But these theoretical arguments have largely sidestepped questions about what level of centralization is optimal for redistributive programs. And by focusing on the empirical question of residential mobility, they have ignored a variable that—I seek to demonstrate—is at least as important. In this paper, I argue that those two deficiencies in the literature are connected. I introduce a simple model to show that economic redistribution becomes more difficult—indeed, approaches impossibility—as economic inequality increases, regardless of one’s assumptions about levels of mobility (by the rich or poor). That is because economic inequality has an inherent spatial dimension: so long as citizens exhibit anything short of perfect mobility (and perfect responsiveness to redistributive policy), its rise will result in an increasing geographic concentration of fiscal resources available to governments. For precisely this reason, higher levels of economic inequality strengthen the case for centralizing the financing of any program with redistributive goals—including the great bulk of what contemporary governments aim to do.

I introduce the concept of a “fiscal unit” to refer to the geographic scope of public financing—which might be, depending on the program, a school district boundary, a county, a state, or the entire country. In order to achieve an equitable allocation of public goods, policymakers should respond to rising income inequality by shifting the site of revenue collection to occur at widely drawn “fiscal units”. This can take two forms. It can be done by expanding the scope of fiscal boundaries—for example, by funding locally-administered programs at the state or federal level. Alternatively, policymakers could respond to inequality by increasing fiscal transfers from higher levels of government (wider fiscal units) to lower, geographically smaller governments.

Rather than an afterthought, the existing level of economic inequality within a political community may be the single most important question for this aspect of policy design. Where wealth is unequally distributed, the primary responsibility of assessing the revenues used to finance public goods should be assumed by levels of government representing the greatest number of people. This paper thus suggests that policymakers should respond to rising income inequality by shifting not only the burden but also the site of redistributive taxation.

Suggested Citation

Highsmith, Brian and Highsmith, Brian, The Implications of Inequality for Fiscal Federalism (or Why the Federal Government Should Pay for Local Public Schools) (April 1, 2019). Buffalo Law Review, Vol. 67, No. 2 (2019), Available at SSRN: or

Brian Highsmith (Contact Author)

Harvard University, Department of Government ( email )

United States

Arthur Liman Center for Public Interest Law ( email )

P.O. Box 208215
New Haven, CT 06520-8215
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics