Does international trade improve environmental efficiency? An application of a super slacks-based measurement of efficiency

This study analyzes environmental efficiency, and its determinants, for 98 countries in terms of four typical air pollutants—SO2, NOx, particulate matter 10 micrometers or less in diameter (PM10), and CO2—for the period 1970–2008. For this purpose, I propose a super slacks-based measure and data envelopment analysis (DEA) model with undesirable outputs—which has higher discriminating power than previous DEA efficiency indices, modifying the ones proposed in preceding articles. Furthermore, I analyze the determinants of environmental efficiency in association with the environmental Kuznets curve hypothesis and the pollution haven hypothesis. The panel regression results reveal that there is no Kuznets-type relationship between environmental efficiency and per-capita income. The impact of trade on environmental efficiency depends on relative per-capita income and capital–labor ratio, i.e., the higher the relative income and the lower the capital–labor ratio, the higher the environmental efficiency. Overall, the elasticities of trade openness for NOx, PM10, and CO2 are significantly negative for an average country in the sample.