The indicator "aggregate carbon intensity", such as that given by the ratio of a country's CO2 emissions to its GDP, is often used by researchers and policymakers in the study of climate change or environmental performance. A better understanding of the determinants of this indicator is useful in climate policy analysis. This paper introduces four different models to calculate a country's aggregate carbon intensity using the input output framework. The models are developed based on different combinations of input output models, country imports assumptions, and approaches to calculating GDP. Using the four models and multiplicative structural decomposition analysis, the determinants of the aggregate carbon intensity change of China from 2007 to 2010 are studied. The results given by the models are compared and attribution analysis on one of the determinants is conducted. The proposed procedure can be similarly applied to study changes in other aggregate intensity indicators, such as the ratio of energy consumption, water consumption or material use of a country to its GDP. (C) 2015 Elsevier Ltd. All rights reserved.