R&D innovation efficiency of Chinese high-tech industries: Three-stage network approach with fairness consideration

Recently, innovation evaluation has played an increasingly important role in the development of high-tech industry and the enhancement of international competitiveness. Data envelopment analysis (DEA) has been viewed as an effective data-driven method in evaluating the performance of decision-making units (DMUs). However, the traditional DEA models treat the DMUs as “black box.” As an extension of traditional DEA, a three-stage network evaluation transforming from a “black box” to a “gray box” structure has appeared. Unfortunately, the problem of intermediate elements has reduced the usefulness of the extended method. Although some gaming approaches have been introduced in the DEA model to reduce internal conflicts, the issue of fairness in the three stages, which have a significant impact on the behavior of the decision-makers, has still not been considered. To fill this gap, Neumann-Morgenstern cardinal utility is adopted to express the basic utility of a certain stage by considering itself efficiency and additional fairness utility from the other’s efficiency. Afterward, several non-cooperative DEA models have been developed by considering the efficiency of the leader stage and fairness disutility from other stages. Finally, a numerical example is presented and an empirical application is given to verify the proposed method.