A Resampling Slack-Based Energy Efficiency Analysis: Application in the G20 Economies

In order to have a sustainable economic and social development, it is important to balance economic growth and ecological environmental damage. In this article, we used the resampling model under the triangular distribution to evaluate energy efficiency, because the input/output value may have measurement errors, time lag factors, arbitrariness, and other problems, causing their own DMU to change. After these factors were taken into consideration, the resampled input/output was estimated because a super-SBM efficiency value was placed in the confidence interval. From the past-present data, for the estimated data change, the time weight was provided according to the Lucas series, and the super-SBM was time-weighted. We applied this model to a dataset of G20 economies from 2010 to 2014. To the best of our knowledge, very few studies have applied the DEA method with resampling to analyze energy efficiency. Thus, our study contributes to the methodologies for energy efficiency evaluation. We found that the overall average energy efficiency is 0.653, with substantial differences between developed economies and developing economies. The most important finding is that neither overestimation nor underestimation occurred when sampling was repeated one thousand times using 95% and 80% confidence intervals, confirming the robustness of the super-SBM model. The less energy-efficient economies should adjust their energy policies appropriately and develop new clean energy technologies in the future.