New concepts for bootstrap-based cross-efficiency and relative weight analysis and an application to China's governance-finance-innovation-sustainability system

Abstract

As for small (or medium-sized) data sets, there are limited number of samples. In such contexts, the cross-efficiency method can hardly be used for statistical tests. To overcome the difficulty, this study introduces a new method (bootstrap cross-efficiency). Our method contributes to the DEA literature, since it produces sufficient weight samples and can be used for genuine hypothesis tests. Meanwhile, this study makes another methodological contribution by proposing a new method of relative weight analysis. The new method provides us with insightful information on stage-based priority. Empirically, this study contributes to the sustainability analysis by providing an extended governance-finance-innovation-sustainability system and focuses on the analysis of Chinese provinces from 2009 to 2019. To the best of our knowledge, such work cannot be found in the existing studies. There are three main conclusions. First, China experienced considerable progress in overall efficiency measures during the examined periods. Such progress was mainly induced by government sector, R&D sector and industrial sector. Second, in terms of sector priority, R&D sector ranked first, which was followed by government sector and industrial sector. R&D sector and government sector accounted for about 64% of total relative weights. Financial sector played a minor role, but its importance grew over time. Finally, significant heterogeneity was confirmed in efficiency measures and coordination levels across groups and provinces.