The Environmental Kuznets Curve Hypothesis (EKC) which argues that an inverted U-shaped relationship exists between economic growth and environmental degradation has been examined by numerous studies for different countries or country groups. However, the validity of the EKC hypothesis at the sectoral level is mostly ignored. In addition, most of these studies have modeled the nexus between per capita income and CO2 emissions based on the assumption of non-linearity. Unlike previous studies, the main purpose of this paper is to investigate the validity of EKC hypothesis for sub-elements of carbon dioxide emissions (i.e. total CO2 emission, commercial CO2 emission, electrical CO2 emission, industrial CO2 emission, residential CO2 emission and transportation CO2 emission) in the United States for the annual data of 1973-2015. In doing so, the rolling window estimation procedure is employed to observe the effect of per capita income on sectoral CO2 emission for each sub-sample period instead of the non-linear assumption. The results of the rolling window coefficients show that inverted U-shaped EKC hypothesis is valid for total CO2 emission, industrial CO2 emission, electrical CO2 emission and residential CO2 emission. However, the inverted U-shaped relationship between economic growth and CO2 emission is not supported for commercial and transport sector of the US. (C) 2018 Elsevier Ltd. All rights reserved.