Abstract: This study proposes a decomposition approach based upon data envelopment analysis that identifies various sources of CO₂ emission. In addition to the previously identified seven sources, we propose three new ones. As an empirical application, this study applies the proposed approach to examine ten sources of CO₂ emission across Chinese provinces from 2008 to 2015. In the empirical study, we overcome methodological difficulties related to (a) what methodological merits of technology change indexes are and how to measure them in a separated manner and (b) how to separate effects of various sources and how to identify the annual shift of those sources of CO₂ emission changes. This study finds three empirical implications. First, four sources may increase the amount of CO₂ emission. They include an economic activity, a technology change on a desirable output and a potential energy intensity change. Second, two sources are important in reducing the amount of CO₂ emission. They are an operational efficiency change on a desirable output and a change in energy saving technology. Finally, conflicting results exist in some sources in the manner that they increase CO₂ emission in some provinces but decrease it in the other provinces.

Keywords: Data Envelopment Analysis; Production-theoretical Decomposition Analysis; Index Decomposition Analysis; Carbon Dioxide Emission; Provincial Imbalance