This article notes that the econometric literature on convergence has been critical of traditional growth regressions for studying economic convergence across countries and regions, based on the popular notions of β and ó convergences. This is because these methods fail to allow for unobserved (and persistent) differences across countries, and they are susceptible to measurement errors, endogeneity biases, and spatial autocorrelation. The article does two things: First, it investigates the convergence hypothesis among Indian states by using panel unit root tests that explicitly incorporate...

This article notes that the econometric literature on convergence has been critical of traditional growth regressions for studying economic convergence across countries and regions, based on the popular notions of β and ó convergences. This is because these methods fail to allow for unobserved (and persistent) differences across countries, and they are susceptible to measurement errors, endogeneity biases, and spatial autocorrelation. The article does two things: First, it investigates the convergence hypothesis among Indian states by using panel unit root tests that explicitly incorporate cross-sectional dependence (various socioeconomic variables in different regions in India are expected to be contemporaneously correlated). Second, two measures of well-being are used: per capita consumption—both rural and urban—and per capita state-level GDP (SGDP). Most of the studies on India are based on SGDP, which is a questionable indicator of welfare. The article also incorporates possible structural breaks that may occur and may lead to completely different outcomes of the convergence hypothesis test.