I utilize the I(2) cointegration model to assess the empirical relevance of the environmental Kuznets curve for $$\hbox {CO}_2$$
CO
2
emissions in the US between 1960 and 2014. This takes the non-linearity of $$\hbox {CO}_2$$
CO
2
emissions into account by directly incorporating data that are integrated of order two, I(2). As a result, it enables an extensive dynamic analysis of the relationship between emissions and economic growth, as postulated by the environmental Kuznets curve both in the short, medium, and long run. The results indicate that the primary drivers behind the non-linear shape of US $$\hbox {CO}_2$$
CO
2
emissions in the long run are an increase in emissions caused by energy use and a decrease caused by more trade and the utilization of less polluting energy sources. GDP only exhibits short run effects. Hence, I do not find evidence in favor of a long-run relationship between economic development and the concave shape of emissions, as suggested by the environmental Kuznets curve.