Comparative analysis of the R&D investment performance of Korean local governments

•We not only used slack-based model DEA (SBM-DEA), which overcomes the limitations of traditional DEA, but also conduct bootstrapping to enhance the statistical reliability of our results.•We examined R&D efficiency at the local government level in order to suggest precise implications.•We found that the average R&D efficiency score obtained in this study was 67.7% throughout the study period for all local governments. This implies that there is potential for a 32.3% enhancement.•We provided each ineffective local government with an input target calculated by four reference sets; two cities (Gwangju and Daegu) and two rural provinces (Gangwon and Gyeongbuk). This study examines Korea's research and development (R&D) investment performance at the local government level using slack-based model data envelopment analysis (SBM-DEA). The SBM methodology, which has replaced the traditional DEA model, is expected to provide more reliable empirical results for R&D investment performance. We confirm the statistical reliability of our results by conducting bootstrapping. The average score of Korea's R&D investment performance is 67.7%, implying that there is a 32.3% potential for efficiency improvement. Among the 16 local governments examined, Seoul, Gwangju, Daegu, and Gangwon show better performance with an average value higher than 0.8. We decomposed R&D investment efficiency into pure R&D investment technical efficiency and scale efficiency and derived implications regarding the input scales. We also reported benchmark information from trend-setting local governments that indicate ideal input mixes for fast-following local governments. Since no local government was found to be in the CRS group, we suggest that all local governments should transform their R&D investment input mix toward upscaling or downsizing.