Spatial Diversity in Local Government Revenue Effort under Decentralization: A Neural-Network Approach

Abstract

Decentralization reflects a global trend to increase the responsiveness of state and local governments to economic forces, but it raises the challenge of how to secure redistributive goals. Theoretically, as the equalizing impact of federal aid declines under devolution, we expect subnational state-level government policy to become more important, and geographic diversity in local governments' efforts to raise revenue to increase. In this paper we explore the impact of state fiscal centralization and intergovernmental aid on local revenue effort with the aid of Census of Governments data for county areas from 1987 for the Mid-Atlantic and East North Central region of the United States, with particular attention paid to rural counties. The 1987 period was chosen because it is the first year in which state policy trends diverged from federal decentralization trends and both state aid and state centralization increased while federal aid to localities continued to decline. Using a neural-network approach, we explore the spatially differentiated impact of state policy and find complementary responses in effort among some localities and substitution responses among others. Classification-tree analysis of this diversity suggests that decentralization and the competitive government it promotes are likely to exacerbate inequality among local governments.