Plateau du Kirchberg Luxembourg
28 & 29 Novembre 1996
Résumés des interventions

Diffusion technologique par brevet



Bruno van Pottelsberghe de la Potterie,
Université Libre de Bruxelles, DULBEA - CP140
50 av. F.D. Roosevelt, B-1050, Bruxelles
Tel: 32 2 650 41 51 - fax: 32 2 650 38 25 - E-mail:


It is widely recognized that the impact of R&D investment performed in a particular industry goes well beyond the improvement of its own productivity. Through the existence of R&D externalities, an innovation created in a particular industry may benefit several other industries. There is a burgeoning, though still quantitatively limited, empirical literature on the evaluation of the direction and the amplitude of inter-industry technological spillovers. The lion’s share of this literature do estimate the impact of a stock of ‘external’ or ‘indirect’ R&D on output growth. For a given industry, the indirect R&D capital stock is a weighted sum of the R&D performed in all other industries. These weights are most often computed from input-output or technological flows matrices. So far, concerning this weighting methodology, three consequencial issues deserve to be examined.


First, although the analytical distinction proposed by Griliches (1979) between rent-spillovers and pure knowldege-spillovers is ubiquitously acknowledged, there are still some implicit disagreements about the nature of the estimated spillovers from a given weighting matrix. For instance, Griliches and Lichtenberg (1984) consider Scherer (1982)’s technological flows matrix as the best tool to approximate rent spillovers. However, rent spillovers reflect the fact that intermediate input prices do not embody completely the product innovations or the quality improvement resulting from R&D activities. Therefore, it seems that an input-output matrix might be more appropriate for their evaluation while pure knowledge spillovers would be better sized through the use of technological flows matrices.


The second issue is closely related to the first one. Since the construction of technological (and/or innovation) flows matrices is apparently labor and time consuming, one may wonder whether they could be proxied by input-output matrices. A positive answer would mean that rent spillovers are similar to pure knowledge spillovers. A negative answer would highlight the usefulness of the few existing technological flows matrices but would raise the question of whether they are similar despite the fact that they are built from different hypotheses and concepts.


The third issue pertains to the problem of international comparability. Capron and al. (1996), Verspagen (1996), and Englander and al. (1988) assume that a single technological flows matrix prevails for the various industrialized countries included in their analyses. This implicit hypothesis is inescapable since the available technological flows matrices have been inharmoniously built. However, the pertinence or the empirical implications of such presupposition have never been tackled.


This paper aims at clarifying these three issues which could be summarized as follows: (i) What is the likely nature of the spillovers estimated through the input-output and the technological flows weighting matrices? (ii) Is the input-output matrix a good proxy for the technological flows matrix? (iii) How relevant is the assumption that a single technological flows matrix might be used for different countries?


The first section describes the well-known weighting procedure and addresses the problem of concordance between the nature of the spillovers and the nature of the weighting matrices. The second section brings some clues towards the second and the third issues. It provides a comparison of the weighting components pertaining to four different technological flows matrices and to the input-output matrices of the USA, Canada, Japan, Italy, France, the UK, and Germany. The four technological flows matrix are from Scherer (1982), the Yale method of Putman and Evenson (1992), and computed from the EPO and USPO data. The third section extends the comparison empirically. Econometric estimates of the impact of external R&D on output growth allow to further clarify the different issues. Section four concludes.


Abstract References

Capron H., Odagiri H., van Pottelsberghe B. (1996), « Inter-Industry Technological Spillovers: an International Comparison », paper presented in the international conference on « The Economics and Econometrics of Innovation », June 3-5, 1996, in Strasbourg.

Evenson R.E., Putman J. (1994), « Inter-Sectoral Technology Flows: Estimates From a Patent Concordance with an Application to Italy », miméo, Yale University.

Griliches Z. (1979), "Issues in assessing the contribution of Research and Development to productivity growth", Bell Journal of Econometrics, Vol.10, N°1, pp.92-116.

Griliches Z. and Lichtenberg F. (1984), "Interindustry technology flows and productivity growth: A reexamination ", , The Review of Economics and Statistics, 65, pp324-329.

Scherer F.M. (1982), "Inter-industry Technology Flows and Productivity Growth", The Review of Economics and Statistics, vol.64, n°4, November, pp.627-34.

Verspagen B. (1996), « Measuring Inter-Sectoral Technology Spillovers: Estimates from the European and US Patent Office Database », MERIT, Working Paper, 17p.




GOMEZ, M. and MURUA, J.R., University of the Basque Country (Spain)


This paper firstly studies, in the framework of the conventional paradigm of the production function, the relationship between R&D (input) and patents (output) in several sectors of a regional economy (Basque Country), comparing this correlation in differents periods.

Second, it evaluates the results achieved (got) in the framework of the new theoretical approaches about the regional innovation. It shows that the correltion previosly mentioned is not significant for several sectors of the Basque economy ; nevertheless, it might have explanatory interest in other sectors.



Keld Laursen :

Keld Laursen, IKE Group Private:
Department of Business Studies
Aalborg University, Fibigerstraede 4 Svendsgade 5,1th.
9220 Aalborg O, Denmark 9000 Aalborg, Denmark
Phone: +4598154211#2916 Fax: +4598156013 Phone: +4598113016


The paper explores firstly the impact of technological change on trade growth at the country level, using trade statistics and statistics on patenting activity in the US, across 20 countries for 17 manufacturing sectors. Secondly, using structural decomposition analysis, the paper examine statistically, whether the degree to which countries get access to sectors with above average growth in technological opportunity has any impact on growth in aggregate market shares of exports. The results demonstrate that there is a positive relationship between change in trade performance and change in technological capabilities across countries for 8 ‘technology intensive’ sectors over the period 1965-1988. Furthermore, the relationship between the degree to which countries get access to sectors with above average growth in technological opportunity and growth in aggregate market shares is investigated. In this context, it is shown that there is a positive relationship between growth rates in trade performance and the individual ‘national innovation system’s’ ability to actively move into technological sectors offering above average technological opportunity.



Is there any persistence in innovative activities?

Elena Cefis
Trento University and European University Institute, Florence



This paper examines the persistence properties of patents data. Using non-parametric techniques I reject exponentiality for the distribution of patents for each year of the sample. The result suggests that there exists a threshold to patenting and the threshold is represented by the first patent. Also, using a Markov Chain approach, I show that there is little persistence in innovation activities, that there is evidence of bimodality in the transition probability matrix and that these tendencies are stronger the longer is the transition period. There is heterogeneity along several dimensions of the sample (industrial classificatiion and quoted {\em versus} non-quoted, with firms belonging to the chemical industry displaying stronger persistence in innovative characteristics. Sensitivity analysis demonstrates the robustness of these features to changes in auxiliary assumptions (expecially, relaxing the time invariant assumption and representing the data with second order Markov chains). We conclude that being an innovator is not a specific innate characteristic of a firm.