We address the question of whether the returns to R&D differ between R&D projects funded by public grants and R&D in general. To answer this question, we use a flexible production function that distinguishes between different types of R&D by source of finance. Our approach requires no adjustment of the sample or data in order to include firms that never invest in R&D, in contrast to the standard Cobb-Douglas production specification. We investigate the productivity and profitability effects of R&D using a comprehensive panel of Norwegian firms over the period 2001-2009. The results suggest that the returns to R&D projects subsidized by the Research Council of Norway do not differ significantly from R&D spending in general. Our estimate of the average rate of return to R&D is about 10 percent. This estimate is robust with respect to whether firms with zero R&D are included in the estimation sample or not. In contrast, using a standard Cobb-Douglas specification and restricting the sample of firms to those with positive R&D, leads to implausibly high estimates of the rate of returns.
C33 - Models with Panel Data ; C52 - Model Evaluation and Testing ; D24 - Production; Capital and Total Factor Productivity; Capacity ; O38 - Government Policy
