R&D and productivity growth: Evidence from the UK

Abstract

Although the econometric evaluation of R&D has attracted wide interest in many countries, it has not attracted much in the UK. The main objective of this paper is to fill this void, i.e., to estimate the impact of R&D on productivity growth of the UK manufacturing sector. However, there are some additional objectives. Firstly, we estimate the impact of R&D on productivity growth of large and small firms and we discuss a number of theoretical arguments regarding the role of firm size. Secondly, given that the technological infrastructure influences the innovative capacity of a firm, we compare the impact of R&D on productivity growth of high-tech firms with the corresponding impact on productivity growth of low-tech firms. Thirdly, we investigate whether the contribution of R&D to productivity growth has changed over time.

Based on firm-level data (78 firms, 1989–2002), we find that the contribution of R&D is approximately 0.04. Although the R&D-elasticity of large firms (0.044) is higher than the corresponding elasticity of small firms (0.035), the difference is small. In contrast, the R&D-elasticity is considerably high for high-tech sectors (0.11), but statistically insignificant for low-tech sectors. Finally, the investigation of the elasticity of R&D over time revealed an interesting discontinuity showing that although until 1995 the R&D-elasticity was approximately zero, after 1995 it increased dramatically to 0.09. We investigate the potential causes of such non-linearity and we suggest a number of possible explanations.

Keywords:

• Innovation
• Internet
• Research and development
• Productivity growth

Acknowledgements

I would like to thank seminar participants at Canterbury Business School for their valuable comments, in particular John Sharp, John Mingers and Alison Dean. I would also like to thank Dietmar Harhoff and Pierre Mohnen for their suggestions and three anonymous referees of this journal for helpful comments on an earlier version of this paper.

Notes

¹ We can also refer to technical change (A) or productivity growth as total-factor productivity (TFP), multi-factor productivity or residual productivity.

² Wakelin (2001) using a sample of 98 firms and R&D data that cover a 5-year period (1988–1992) found that the ROR to R&D is about 0.28.

³ For a detailed description of this framework, see Mairesse and Sassenou (1991).

⁴ The record of capital services, however, is not so different from the net fixed capital stock.

⁵ Diversified firms with activities in more than one sector were included in their closest sector according to their sales.

⁶ Note that the ‘Other’ group includes the industries of metal manufacturing, motor parts and transportation.

⁷ The small-firm sample contains companies whose average sales were less that £200 million per year, whereas the remaining firms were included in the large-firm sample.

⁸ Scale factor is the sum of the capital, labor and R&D coefficients.

⁹ For more details on the test, see Bera and Jarque (1981).

¹⁰ Note that this version of the white heteroscedasticity test is calculated using Microfit. It initially calculates the regression of the squared residuals on the squared fitted values and then, it tests whether the squared fitted values in this regression are statistically significant. For more details, see Koenker (1981).

¹¹ For more details on the test, see Breusch and Godfrey (1981).