Abstract:
In this paper, the convergence of income per capita across U.S. metropolitan statistical areas (metros) is examined over the period between 1969 and 2011. We initiate the analysis with multivariate tests for stability, and the existence of unit roots, performed on the ratio of income per capita of a specific metro relative to the cross-sectional average. The tests used were: Multivariate homogeneous Dickey–Fuller, Levin–Lin–Chu (2002), Im–Pesaran–Shin (2003), MKPSS, Bai–Ng (2004) Panic and Hadri-LM tests. The analysis is complemented by the use of the panel stationarity test accounting for structural changes, as proposed by Carrion-i-Silvestre et al. (2005), which our other tests do not incorporate. The study of convergence is important for both economists, as a means to test growth theories and distinguish between different models, as well as policy makers who seek to maximize the utility of their constituents by making use of all information available to them. While, standard unit root tests indicates stationarity of our metric, i.e., the ratio of income per capita of a specific metro relative to the cross-sectional average, the tests of stability suggests divergence, barring the Bai–Ng test. But, given that the tests of stability are more conducive to checking for convergence, we conclude that in the 384 U.S. metros there is a divergence of per capita income.
