This paper analyzes a relationship between income inequality and economic growth using panel data for 48 US states and 1960-2000 Census years. First, we investigate the impact of inequality on growth using a generalized method of moments (GMM) estimator for dynamic panel data models. We depart from the previous literature by incorporating the state cost of living (COL) index instead of the national Consumer Price Index (CPI) to assess whether a differential price index matters in estimation. We also explore nonlinear relationships as recent theoretical literature pointed out. Our regressions suggest that using the state COL index produces more robust negative coefficients on inequality (higher inequality indicates less growth) as well as specification tests. The nonlinear estimations indicate that the relationship between inequality and growth is more likely to be nonlinear and with the state COL index, the coefficients on inequality terms are smaller. Second, we analyze whether economic growth affects inequality. Preliminary results indicate that a higher income level tends to increase inequality and an inverse U-shape relationship postulated by Kuznets (1955) is not strong enough using the state COL index as opposed to the national CPI.