R&D, innovation and productivity relationships : evidence from threshold panel model

This study investigates the nonlinear relationship between R&D expenditures, innovation, productivity, high-tech export products. Previous empirical research used linear standard structures to deal with these kinds of specifications, and it has shown that the linearity is frequently conditioned by other macroeconomic factors such as the level of development and the financial openness. Based on these arguments, our study investigates this question in econometric specification using panel smooth threshold regression methodology proposed by Gonzalez et al. in 2005. Our findings suggest that there is a threshold effect within the links between R&D expenditures, innovation, and productivity. The effect of R&D expenditures, innovation, productivity, and medium and high technology product exports is mixed. However, both positive and negative effects are found, depending on which innovation indicators are used or on which level of threshold variable is the most appropriate. The results advocate that the level of economic development can be considered as target indicators to conduct an innovation policy.