In this paper, we investigate the validity of the environmental Kuznets curve (henceforth, EKC) hypothesis for 8 OECD countries. To this aim, we decompose the per capita GDP series into its increases and decreases and consider only increases by excluding decreases from the model. Therefore, this method may enable us to test the EKC hypothesis more accurately, in accordance with the original theory. Following decomposition, we apply the fixed-effect regression model with Driscoll-Kraay standard errors, and the common correlated effects mean group (CCEMG) estimator. Empirical findings indicate that while the undecomposed model with undecomposed per capita GDP series supports the EKC hypothesis for 4 out of 8 countries, the decomposed model with decomposed per capita GDP series does not do so for any country. Hence, these mixed results reveal a need to employ different alternative techniques, such as the data transformation/decomposition applied in this study, for testing the EKC hypothesis.