Economic research on income's impact on environmental quality is well summarized by the environmental Kuznets curve (EKC), a relationship between environmental quality and per capita income (PCI). In this paper, a flexible nonparametric approach is adopted to study this relationship in which polynomial splines are employed to approximate the EKC function. New spline confidence bands via wild bootstrap are proposed to examine the adequacy of certain specifications of the EKC. In time series context, a novel method is developed to investigate the EKC by decomposing the countrywise time series into a polynomial trend and a stationary noise sequence. The trends are examined for conformance to the EKC hypothesis, while the impact of economic intervention on environmental quality change is analyzed via the proposed confidence bands. Simulation studies illustrate the properties of the proposed methodology. Copyright © 2010 John Wiley & Sons, Ltd.