Input-output analysis of embodied emissions: Impacts of imports data treatment on emission drivers

Input-output (I-O) analysis has been widely used in embodied emission studies. For country-level analysis, the non-competitive imports assumption is generally preferred. However, national I-O tables with the noncompetitive imports assumption are often not available. In empirical studies, the uniform imports share approach is usually adopted to give the estimated tables from the I-O tables with the competitive imports assumption. When data are available for two different years, structural decomposition analysis (SDA) can be applied to study the drivers of the embodied emission changes. We propose a SDA framework with two imports effects, i.e. input imports effect and demand imports effect, to evaluate the impacts of imports data treatment on the drivers. An empirical study using China's latest I-O datasets, for 2017 and 2018, show that the impacts are not significant at the national level but are significant at the sectoral level. For the latter, the impacts on sectoral embodied emission intensity range from -5% to 16%, while on sectoral embodied emissions range from -8% to 35%. Sectoral aggregate embodied intensity (AEI) indicators are found to be more robust to the imports data treatments than sectoral embodied emission intensity. In some cases a change in the sign of the input imports effect or demand imports effect is observed, which leads to opposite outcomes. Implications of the findings on embodied emission studies and the role of imports are discussed.

Automated summary

The study finds that data treatment of imports has insignificant impacts on emission changes at the national level but significant impacts at the sectoral level. The sign change of the input imports effect or demand imports effect can lead to opposite outcomes in emissions.