Measuring and comparing the R&D performance of government research institutes: A bottom-up data envelopment analysis approach

•This paper measures the R&D performance of government research institutes (GRIs).•The bottom-up approach based on data envelopment analysis (DEA) is proposed.•DEA is run to measure R&D efficiency of projects carried out by GRIs.•Nonparametric statistical tests are run to compare the R&D performance of GRIs.•The bottom-up DEA approach is applied to the 10 Korean GRIs for applied research.

Government-funded research institutes (GRIs) have played a pivotal role in national R&D in many countries. A prerequisite for achieving desired goals of GRIs with the limited R&D budget is to be able to effectively measure and compare R&D performance of GRIs. This paper proposes the bottom-up approach in which the performance of a GRI is measured based on the efficiency of its R&D projects. Data envelopment analysis (DEA) is employed to measure R&D efficiency of projects, and nonparametric statistical tests are run to measure and compare the R&D performance of GRIs. We apply the bottom-up DEA approach to the performance measurements of 10 Korean GRIs conducting a total of 1481 projects. The two alternatives for incorporating the relative importance of the output variables – the assurance region (AR) model and output integration – are also discussed. The proposed bottom-up approach can be used for formulating and implementing national R&D policy by effectively assessing the performance of GRIs.