Factor complements and substitution for the translog knowledge production function considering ownership structure and firm financial condition

This paper uses stochastic frontier analysis to evaluate R&D efficiency and investigate factor substitutability for the translog knowledge production function. We find that most substitution elasticities are positive and support substitute relationships between the four R&D inputs. Ownership structure plays an important role in R&D activity and efficiency. Our findings suggest that higher foreign and manager shareholding will promote R&D efficiency. The little protection hypothesis and unproductive R&D hypothesis are supported by means of negative relationships for return on equity (ROE)–R&D efficiency and cash flow (CF)–R&D efficiency. R&D experience and past R&D quality appears to improve R&D efficiency. Finally, small firms have higher R&D efficiency.