Regional impacts of R&D and public R&D funding

Introduction

Recent research on regional economics has also lately focused on R&D. It has been emphasized that R&D is even more concentrated than production. Kelly and Hageman (1999) found that in the USA R&D tends to concentrate in certain locations across industries. Audretsch and Feldman (1996) have pointed out that innovative activity has greater propensity to cluster especially in the industries with a high R&D intensity than in other industries. To what extent the observed clustering of these studies depend on the so-called natural advantages, discussed by Ellison and Glaeser (1997), and to what extent they depend on knowledge spillovers, is unclear. Locations of R&D may also be affected by such physical spillovers as transportation costs (see Krugman, 1991). Finally, besides the above-mentioned factors labour and financial market imperfections can also be responsible for concentration of R&D activity. Public intervention through public funding decisions may also affect the geographical distribution of firms’ R&D activities.

In this study we tend to focus on the effects of intellectual spillovers. The study recalls or resembles us of, in that respect, the papers by Adams and Jaffe (1996) and Orlando (1999), who have empirically tested the impacts of explicitly defined external R&D sorted into different pools along geographic and technological dimensions. More specifically, in this study we empirically test the importance of geographic distance in the determination of R&D investments, productivity and employment. But technological proximity is also examined. We use industrial proximity as a rough measure of actual technological proximity which is the proximity between a firm’s and other firms’ research space.

When geographic or technological distance shortens, the positive spillover effect strengthens, but the plant’s own R&D may then be more easily replaced by external R&D. Thus, it is not even clear on the basis of theory whether geographical or technological proximity, associated with other firms’ R&D, has a positive or negative influence on a firm’s R&D investments. In fact, the purpose of the empirical part of this project is to test empirically ambiguous theoretical implications. On ground of theory the industrial and geographical closeness should affect rather positively productivity. Also this is tested empirically. We consider also the interaction between the firms’ R&D activity and total employment. It is not clear how the other firms’ R&D affect on employment. The net effect could easily be zero, if some regions specialise in industries which use a lot of R&D and the other regions in activities which require hardly any R&D. How R&D affect total employment in investigated empirically in this study. The empirical analysis focuses also on the role of public R&D funding. The effects of the staff’s education are also evaluated.

One of the main contributions of this study is the constructing of a large R&D database which includes the approximates of R&D stocks for practically all the relevant firms in Finland who have been active in R&D during the years 1985-1998. In this data set the firms’ R&D is allocated to be located in certain municipalities. This also makes it possible to link R&D data to the firms’ plants in the same locations. This large data set of R&D stocks is constructed by using partly imputed values for R&D investments. But in the approximation of missing values we have utilized observations of R&D investments in other years, the rate of growth of the specific industry’s R&D and the business register which gives information about the plants’ and firms’ existence in various years.

R&D data set is also sorted at the three-digit NACE level. The countrywide data on R&D stocks makes it possible to specify external R&D stocks for every firm in a certain location in a consistent way. The external R&D also includes the same firm’s R&D in different locations as well as the other firms’ R&D. The external R&D is then divided according to its geographical and industrial dimensions.

This study relies on the data of R&D stocks which does not exclude any relevant segment of the considered countrywide market. Foreign spillovers, though, are not taken into account. Because of good covering, the specified pools of external R&D stocks, corresponding to observations, are consistent with each other. This confirms the reliability of the estimated elasticities for external R&D.

This project also includes a short theoretical note on the determination of R&D investments (in Appendix C) and a separate paper on cooperation in R&D. In Appendix C we focus on R&D investments and show that because the other firms R&D can be substituted for the firm’s own R&D, the total effect from external R&D need not be positive.

In the paper which considers cooperation in R&D we stress that external influence from other firms’ R&D can also be based on voluntary cooperation, besides involuntary spillovers. In this paper we focus on cooperative agreements and point out that free-riding can lead to a situation where closer cooperation decreases firms’ R&D. However, if the possibilities to monitor each other’s behaviour improves, cooperation may even increase R&D in so far as the firms are risk-averters. It is possible that the monitoring possibilities are better, if the firms are near each other. This gives us a new reason to believe that closeness reinforces R&D activity. The task of empirical analysis is to test this implication too.

• ISSN: 1236-7176
• ISBN: 952-5071-50-2
• Press Release in Finnish
• Publication in PDF-format