Evaluating the efficiency and productivity change within government subsidy recipients of a national technology innovation research and development program

This study analyzes the efficiency and productivity change within government subsidy recipients of a national technology innovation research and development (R&D) program. We examine 6,990 government‐sponsored, completed R&D projects during the last three performance follow‐up survey years from 2010 to 2012, and present a design of the sample of panel data to cope with the typical R&D performance time lag using a set of massive observations associated with completed R&D projects for the past 7 years from 2005 to 2011. In particular, data envelopment analysis is adopted to measure the efficiency and productivity change, which is measured in the Malmquist index. Parametric and nonparametric statistical tests are carried out to check for statistically significant differences among the characteristics regarding the types of government subsidy recipients. This study's major findings are as follows. First, during the entire period analyzed (2010–2012), there was a similar yearly pattern of statistically significant differences in the government subsidy means among the recipient types. In contrast, there were no obviously equivalent differences in the efficiency and productivity change. Second, the productivity had increased year on year, but the increments were reduced from year to year. Third, the productivity change was induced mainly by the Frontier‐shift, which indicates overall technology innovation progress, compared with the Catch‐up, which only indicates a simple increase in the efficiency. In particular, in this empirical analysis, the recipient types of ‘national laboratory’ and ‘large company’ had relatively larger sizes of government subsidies per project. However, the efficiency and productivity change of these types was not better than the others. This implies, therefore, that the government should control the ratio of the subsidy to the total R&D budget with an appropriate upper limit. I empirically evaluate the productivity change within a national technology innovation R&D program. I design a sample of panel data to cope with the typical R&D performance time lag using massive observations. There is no obvious relationship between the government subsidy size and R&D productivity change. Some particular types of government subsidy recipient are inferior in terms of R&D productivity change. It practically implies that the government should control the ratio of the subsidy to the total R&D budget.