The determinants of investment in research and development in a panel of UK manufacturing industries

Bettina Becker

Research output: ThesisDoctoral Thesis


This thesis attempts to identify some of the determinants of R&D investment in UK manufacturing industries in the 1990s, a decade in which R&D intensity in the UK declined. We begin with a summary of the key driving factors that emerge from the current empirical literature on R&D determinants, most of which has utilized firm level data. We then estimate an econometric panel model of R&D investment. Industry characteristics such as sales and profitability, product market competition, macroeconomic factors such as the real interest and exchange rates, and the composition of R&D expenditure and funding are found to be important. The thesis examines the sensitivity of the results to specification, functional form, and estimation method. In a next step, we acknowledge that pooled estimates will not provide reliable estimates of individual effects when the parameters are heterogeneous. Moreover, they may not even be consistent estimates of the average in this case. Evidence from a principal components analysis suggests there may be significant differences between high-tech and low-tech R&D determinants, and hence we move on to test the validity of the pooling assumption in our panel. We find that government funding is significant only for low-tech R&D, whilst skilled labour and foreign R&D matter only in high-tech sectors. Foreign R&D accounts for a substantial amount of total R&D. Considering its importance, we move on to investigate its determinants in the remainder of the thesis, where we take particular account of potential effects from exchange rate uncertainty.
Original languageEnglish
Awarding Institution
  • University of London
  • Smith, Ron P, Supervisor, External person
Publication statusPublished - 2006


Dive into the research topics of 'The determinants of investment in research and development in a panel of UK manufacturing industries'. Together they form a unique fingerprint.

Cite this