Federalism and the Size of Government

Abstract

Over the last few decades, the size of the public sector, in both absolute and relative terms, has increased in all of the Western economies. This has led to considerable investigation of the factors determining government size. Factors considered have included income elasticity of voter demand for public goods (i.e., Wagner's Law); relative price changes; threshold effects (e.g., wars or depressions); interest group demands (e.g., public sector employees or unions [17]; productivity differentials [1]; and redistributional motivations [18;10]. One factor which has received little more than passing interest is government structure, that is, centralization of government activities as opposed to decentralization or federalism. Authors such as Fratianni and Spinelli [7], Musgrave and Musgrave [12], Beck [2], and Buchanan [4] make no mention of government structure as a factor in the determination of government size. When considered, the impact of government structure on government size is treated more as an afterthought. For example, Mueller and Murrell [11] incorporate a centralization variable in only two of their twenty-two regression equations, and then only in their two-stage least squares equations, and offer little analysis of its expected impact on government size. Exceptions to this standard treatment of federalism include Brennan and Buchanan [3], Oates [15], Nelson [13; 14], and Marlow [9]. Brennan and Buchanan consider the structure of government and its impact on public sector size in the context of a revenue-maximizing government model. They argue that with mobile voters and local public goods, decentralization of government taxing and expenditure powers would serve to constrain the size of the public sector. In a Tiebout-like world, any attempt on the part of one jurisdiction to exploit its citizenry would result in instantaneous and complete migration on the part of the citizenry to an alternative, nonexploitive jurisdiction. This competition among jurisdictions for the patronage of the citizenry limits the extent of fiscal exploitation possible and, as such, the size of the public sector. Brennan and Buchanan note, however, that [W]ithin a constitutionally designed federal structure, one would predict that there would be constant by competitive lower-level governments to secure institutional rearrangements that would moderate competitive pressures [3, 182]. These institutional rearrangements would take the form of collusive agreements between the central and lower-level governments which assign to the central government what were previously lower-level revenue-raising powers subject to the revenues being returned to the lower-level governments by