Many empirical studies try to test whether there is income convergence across metropolitan areas in the continental United States. Drennan et al. (Journal of Economic Geography 4(5), 2004) claim that income among metropolitan economies is diverging for the period 1969–2001, after applying univariate unit root tests to the time series data. This paper brings new information to this area of study by using the nonlinear panel unit root test of the Exponential Smooth Auto-Regressive Augmented Dickey–Fuller (ESTAR-ADF) unit root test on the time series data for the period 1929–2005. Our results find evidence of stationarity for time series and thereby support beta and sigma convergence among states in a nonlinear setup. However, when the non-linear test encompasses cross section dependence as advocated by Cerrato et al. (2008), the evidence is attenuated.