한국생산성학회

According to the endogenous growth theory by Romcer(Romer, 1986, 1990), the stock of knowledge by research and development(R&D) and human capital has an increasing marginal productivity and growth rate can be increasing over time, in contrast to the models where capital exhibits diminishing marginal productivity. In line with this, this paper examined the effect of R&D investment and human capital on firm growth, focusing on the sustainable growth of small and medium-sized enterprises(SMEs) and large firms. In this study, we first found the human capital influenced positively on the sales growth of high-growth SMEs and high-growth large firms. In particular, we found the human capital of high-growth SMEs more influenced on firm growth than that of high-growth large firms. Second, we also found the human capital influenced on the sales growth of high-growth SMEs and high-growth large firms that exceed the average of R&D intensity more positively than their counterparts that do not. In particular, we found the human capital of high-growth large firms that exceed the average of R&D intensity more influenced on firm growth than that of high-growth SMEs that exceed the average of R&D intensity. The results of this study have a important implication. Firms could expect learning effect of human capital from the R&D activities and that could help the firms grow continuously in both high-growth SMEs and high-growth large firms. One of the limitations of this study is the time lag between R&D activities(innovation) and their commercial success. However, it is expected that R&D activities(innovation) will indeed pay off on average in the long-term(Coad & Rao, 2008). In line with this, many previous studies include one-year time lag between innovation activities(R&D) and firm growth(e.g., Coad & Rao, 2008; Coad et al., 2016; García-Manjón & Romero-Merino, 2012). In particular, Coad et al.(2016) performed additional test in order to examine the relationship between the R&D and firm growth with longer time lag, and found the results are generally similar to the results of one-year time lag. We also analyzed longitudinal data and multiple industry to overcome this limitation, as previously suggested and tested(e.g., Davidsson & Wiklund 2006; García-Manjón & Romero-Merino, 2012; Liu et al. 1999; Weinzimmer et al. 1998; Yang & Huang 2005).