Analyzing Income Convergence at the County Level: The Case of Development in Central Appalachia

Abstract

Theoretical modeling and analysis of long-run economic growth have enjoyed a renaissance, spurred by innovations in theory and the development of large sets of historical data (Barro 1991; Barro and Sala-i-Martin 1995; Baumol 1986; Dowrick and Nguyen 1989; Maddison 1982; Romer 1994; Summers and Heston 1991). The data sets have been used to test the convergence hypothesis that poorer economies will experience faster rates of income growth. To date, model building and empirical analysis have focused primarily on nations, with a few examples of analysis at the state and local level (Barro and Sala-i-Martin 1991; Klein and Barkley 1993.) The counties of Appalachia have been the subject of focused development initiatives because they constitute a poorer region within the United States. The Appalachian counties of Virginia, West Virginia, and Kentucky, in particular, are known for having historically lower levels of income than the rest of the nation and for the boom/bust nature of short-run economic activity that results from dependence on coal and other natural resources (Caudill 1962; Hibbard 1987). In order to promote faster, sustainable economic growth in these counties and enable them to catch up, the Appalachian Regional Commission (ARC) has provided $2.5 billion in development assistance from its inception in 1965 to 1998. Complemented by funds from other federal, state, and local agencies, spending on economic development has totaled $5.25 billion in these three states (ARC Annual Report 1998). In this paper, the convergence hypothesis is tested using county level economic data for these three states from the Regional Economic Information System (REIS) database from 1969 to 1997. Further analysis of income growth in the region focuses on two economic issues pertaining to growth at the local level: (1) the dependence of local economies on natural resources and (2) the pro-