Helsinki University of Technology. Department of Industrial Engineering and Management. Institute of Strategy and International Business. INVENTIVE OUTPUT ...
Helsinki University of Technology Department of Industrial Engineering and Management Institute of Strategy and International Business
INVENTIVE OUTPUT OF ACADEMIC RESEARCH: A COMPARISON OF TWO SCIENCE SYSTEMS Martin Meyer, Helsinki University of Technology M. Du Plessis, Steunpunt O&O Statistieken, K.U.Leuven Tanja Tukeva, SYO - Finnish Institute for Enterprise Management Jan Timm Utecht, SYO - Finnish Institute for Enterprise Management
Working Paper No 2004/3 Espoo, Finland 2004
Helsinki University of Technology Department of Industrial Engineering and Management Institute of Strategy and International Business
Helsinki University of Technology Department of Industrial Engineering and Management Institute of Strategy and International Business PO Box 5500 FIN-02015 HUT, Espoo, Finland Telephone Fax Internet
int 358 9 4513 090 int 358 9 4513 095 http://www.tuta.hut.fi/isib
This working paper is also available on the web.
INVENTIVE OUTPUT OF ACADEMIC RESEARCH: A COMPARISON OF TWO SCIENCE SYSTEMS M. Meyer* a-d, M. Du Plessisa, T. Tukevab,c, J.T. Utechtb,c a
Steunpunt O&O Statistieken, K.U.Leuven, Dekenstraat 2, B-3000 Leuven, Belgium b SYO - Finnish Institute for Enterprise Management, Helsinki, Finland c Institute of Strategy and International Business, Helsinki University of Technology, Espoo, Finland d SPRU Science and Technology Policy Research, University of Sussex, UK This paper compares the inventive output of two science systems in small European countries. More specifically, we examine patented inventions of Finnish and Flemish university researchers. The comparison includes inventive output as such and its concentration on organizations, inventors, and corporate owners as well as foreign assignations and the degree to which individual inventors have retained the ownership of the patents. While there are commonalities between the Finnish and Flemish systems in terms of patent concentration on key institutions and corporate assignees, there are also pronounced differences with respect to the ownership structure of academic patents, which was expected in light of the different intellectual property regulations. Our observations seem to suggest that the total inventive output of a research system is not a function of the prevailing intellectual property system but rather in correspondence to overall national inventiveness thereby pointing to more general (national, cultural) drivers of academic inventive activity. From a methodological viewpoint, this research illustrates that tracing university-owned patents alone would leave considerable technological contributions of academics unidentified - also in countries where universities own the rights to their researchers’ patents. Another finding with potential methodological implications is that patents are highly concentrated on institutions. If such a distribution law applies to large countries as well, analysts could cover most of the national academic patent output by an intelligent selection of universities.
Introduction There has been growing interest in the economic utilization of research results. Along with this trend, the bibliometric community has developed a number of approaches to meet the increased need for information about the contributions of the science system in general and university researchers in particular to the development of the national or regional innovation system. There are various approaches as to how one can study exchange between science and technology (see e.g. Meyer, 2002a,b).
*
This paper was presented at the Nordic Workshop for Bibliometrics and Research Policy, Aalborg, Denmark, and has been accepted for publication in Scientometrics.
First and foremost, one has to mention linkage bibliometric approaches that establish a connection between science and technology by tracing the science references in patents, as pioneered by Narin and his colleagues at CHI Research (e.g. Carpenter and Narin, 1983). This approach has been replicated and further developed by many others (e.g. Verbeek et al., 2002; Tijssen, 2002; Acosta and Coronado, 2003). As Smith and colleagues (1998) pointed out, patent citation analysis is the most popular analytical approach to investigate how science and technology relate to each other.
A number of authors, such as Roberts (1991) and Debackere (2000), point to the critical role of academics in technological entrepreneurship. Developments towards the entrepreneurial university (see e.g. Clark, 1998; Etzkowitz et al., 1998; Etzkowitz, 2003) have revived the interest in the technological activity of university researchers (Henderson et al., 1998). Van Looy and his colleagues (2003, 54) point to patents in this context since they “seem relevant, not only because they reflect technical inventions (novelty) with potential market value, but also because of the widespread availability of databases covering all technological fields.”
There have been a number of studies using a variety of approaches to explore contributions of academics to technological change (e.g. Meyer-Krahmer and Schmoch, 1998; Schild, 1999; Caro et al., 2003, Meyer et al., 2003a,b). While there is considerable work on university patenting, cross-country comparisons are far less frequent. A recent OECD (2003) policy report on patenting and licensing activities at public research organizations can be seen as an attempt to close this gap. The report presents a number of case studies as well as results from a survey of technology transfer offices in a number of OECD countries. More specifically, it points to different regulatory frameworks with respect to the ownership of intellectual assets in universities and research organizations. While many countries have put regulations in place that allow universities and their transfer offices to claim ownership of their researchers’ inventive activity, other countries leave all rights of ownership with the university faculty.
This also explains why cross-country comparisons are difficult. While in the first type of IPR system, the university or its transfer organization will in many instances own technologies invented by their researchers, the university link of patented inventions is far more difficult to trace in the second type of IPR system. More recently, studies were launched that explored academic inventions beyond university-owned patents (e.g. Meyer, 2003). These studies 2
attempt to trace also inventions made by academics that are not owned by the universities or their funding agencies. Academic patents are defined as patented inventions made by at least one inventor who was employed by a university. Using such a criterion allows the analyst to compare the inventive activity of university researchers at a more even playing field. Also universities that do not own any patents can be linked with inventive activities and patented technologies.
This paper draws on Finnish and Flemish studies following this more inclusive approach. In particular, we explore the commonalities and differences of university-related patenting for these two innovation systems which are characterized by different intellectual property rules in the universities and also different transfer practices. Given the relative lack of a comparative cross-country literature, this effort is largely exploratory. However, previous research allows us to formulate some expectations regarding commonalities and differences. The following section will present an overview.
Expected Commonalities and Differences At first, our paper will explore the total inventive output of researchers in both systems. One may expect that in a country in which there has not been much of an institutional incentive for faculty to patent research results the lack of an incentive structure corresponds to a low level of inventive activity of researchers. While there is some exploration as to whether incentive mechanisms have an effect on the commercialization of research results (e.g. Sellenthin, 2003), there is little work on whether the (non-)existence of a transfer infrastructure affects inventive activity as such.
Other analysts (e.g. Shane, 1992) see inventiveness as partly a cultural phenomenon observing differences between countries with respect to per capita patenting output. One may argue that countries with a high per capita level of patenting will also exhibit a high level of inventive activity amongst their academics, irrespectively of their incentive structure or regulatory framework.
The next topic we examine in this paper is the way in which inventive activity is distributed at various levels. We know from other areas that creative activity is unevenly distributed. We
3
also know mostly from US studies of university patenting that most of the output is concentrated on a relatively small number of actors (e.g. Mowery et al., 2001). Therefore, we should expect a similar concentration of patenting on relatively few universities. At the level of individuals (within universities) we also assume skewed distributions following research on inventive activity in general (Huber, 1998). We also expect that these concentrations would not vary considerably between universities.
Ownership of academic inventions is an area where we expect considerable differences between the two systems. Given the legal situation, universities will own a large share of academic inventions in Flanders while only few academic inventions are likely to be assigned to universities and their public financers in Finland. A more interesting question with regard to academic inventions in countries with university ownership of research results regards the extent to which patents are assigned to companies. While the technology transfer operation is likely to be involved in most of the patenting activity, ownership of patents can be transferred to companies. This is a situation that may arise in collaborative research between research groups and their industrial partners.
Foreign assignation of patents and technological specialization of academic inventions are probably influenced by national characteristics. Patented research results will most likely be traced in R&D intensive areas in which patent protection is an effective means to protect intellectual assets. Foreign ownership of academic patents may also be influenced by the openness of the national economy in the international context (e.g. Pavitt, 1982). Flanders and Belgium can be seen as a small, open economy in the center of Europe whereas Finland is seen as an economy at the periphery of Europe.
The following section will introduce the methodology and discuss possible limitations. Then various patent indicators are compared for Finland and Flanders. Then conclusions will be drawn. Commonalities and differences will be summarized. Finally, methodological implications and possibilities for future research will be discussed.
4
Method, Data Sources & Limitations Our analysis is based on two sources of data: patents and personnel register records. We compiled personnel register information for all major universities in Finland and Flanders and utilized databases of Finnish/Flemish USPTO patents.† A patent was considered Finnish if Finland occurred either as the inventor country or as the assignee country in at least one instance. A parallel procedure was applied to identify the Belgian patents. The Flemish patents were selected tracing Flemish municipalities in the address fields. As the initial studies on which this paper is based were independent from each other, different time frames were used. In Finland a database of US patents granted between 1986 and 2000 was available for analysis. In the Flemish study, the database included the years 1991-2001. Both studies covered utility patents only; design or any other patents were excluded.
Personnel registry information was obtained from Finnish and Flemish universities. For the Finnish data we had personnel register information available for the most recent year (then 2000 or 1999) and for 1997, the year in which most patents were filed. In Flanders, personnel records were available for the past ten years. Every university tends to have its own format. Therefore, the personnel data was cleaned and standardized. Doubles were removed. The matching procedure was based on cleaned names from the personnel registries and inventor names as listed in our patent databases. The matching procedure is based on researchers’ last names and initials. As bibliometric matching procedures lead to inventor/researcher name pairs that may link different individuals with the same name, a manual validation procedure was carried out to ensure that the databases of university-related patents were cleaned of homonyms. The same validation procedure was applied to both Finnish and Flemish data.
A number of limitations need to be considered when comparing these two datasets. A direct comparison must necessarily suffer from different delineations of the data sets. Personnel
†
US patents were selected for a number of reasons. First of all, US or European patent data are more suitable for international comparisons than national application data. Secondly, they can be seen as an indicator of commercially more promising inventions than national applications. National patent applications suffer from often relatively high withdrawal rates. Finally, US patents were chosen over European applications for practical reasons. The Finnish study on which this paper is based was conceived initially as a pilot for a Nordic comparative study. Not all Nordic countries are members of the European Patent Organisation. In Belgium, the structure and organization of research and development is not under Federal jurisdiction anymore but primarily regulated at the level of regions and communities. In light of this situation we analyzed Flemish patents only. The Flemish and Walloon parts of the country differed somewhat in their regulatory approaches to universities and intellectual property.
5
register data was available in Flanders for an entire decade while Finnish universities made available only two years of personnel registries. Also, the Finnish and Flemish USPTO databases covered different periods: 1986-2000 (Finland) compared to 1991-2001 (Flanders). This will obscure the comparison to some extent. For the comparison of total inventive outputs, the analysis was restricted to years only when both Finnish and Flemish data was available.
As a further restriction, there are innovation system-related issues to consider when comparing the inventive technological output of the two science systems. While Finland and Flanders have approximately the same population (5-6 million), there are considerable differences with respect to overall patent output, venture capital investment, and the university system. For instance, Finland as a country achieves comparatively high patent rates.‡ There are also more researchers and young Ph.D.’s per capita than in Belgium and Flanders. Government investments in the science system as well as total R&D efforts appear to be higher in Finland. The country also appears to outscore Flanders/Belgium in most other STI indicators with the notable exception of venture capital provision (European Commission, 2003). These factors can all have an impact on the extent of academic patenting and their degree of utilization.
‡
According to data compiled by the European Commission (2003), Finland has between 92-130 US patents per year and million population, compared to 59-88 for Belgium [1995-1999])
6
Results This section presents a comparison between Finnish and Flemish academic patents in particular with respect to the following indicators:
total inventive output; concentration of inventive activity on key universities; concentration of inventiveness within the universities on key individuals; ownership of academic patents by technological area and university; assignation of patents to industrial companies or other organizations in relation to unassigned patents; foreign assignation of academic inventions; and specialization of university-related patents by technological area.
Total Inventive Output
The total inventive output of Finnish university researchers is relatively high. We identified a total of 530 patents which could be related to university researchers as inventors. This is more than 8 percent of all Finnish US patents in our database. A total of 362 US patents could be linked to Flemish university researchers. This is about 5 percent of all Flemish patents in the Flemish US patent database.§ If one restricts the comparison to the years for which both Finnish and Flemish data is available (1991-2000), the patent counts are 432 Finnish patents vis-à-vis 306 Flemish patents (by issue year).
The difference in absolute terms appears not to be surprising if one compares the total inventive outputs of both university systems, there are 1.4 times as many US patents in Finland as there are in Flanders. This corresponds approximately to the difference observed in per capita US patenting between Finland and Belgium. However, one should bear in mind that ten years of personnel registry information were available for analysis in Flanders and only two years in Finland. Even though the validation process appears to indicate that no inventors with above-average activity were missed in Finland, the total number of Finnish patents may be higher. This would correspond to the relatively larger Finnish science system. In any case, §
Note for the purpose of comparison that the share of university-owned US patents is 1.67% in Belgium (19902000; as calculated by Van Looy et al., 2003, Table III, 58).
7
these findings seem to suggest that in the academic context inventive activity is not necessarily a function of a pre-existing regulatory framework.
Patent Concentration on Universities
In spite of the considerable differences in terms of science, technology and innovation indicators and the intellectual property (IP) systems, individual academic activity appears to be highly concentrated on a small number of universities. One or two universities, or one sixth to one seventh of the universities in either country or region, seem to account close to, if not more than, half of all university-associated inventive activity. In Finland, almost half of all 530 patents can be related to researchers working in only two of all twelve universities. About three quarters are associated with researchers in four universities. In Flanders, academic patents are even more concentrated on researchers from key organizations than their Finnish counterparts. The Catholic University of Leuven (KU Leuven) accounts for about half of all patents. Figure 1 compares the concentration of patenting on universities. Figure 1
Patent concentration on universities
100% 90%
Patents (accum. %)
80% 70% 60% Finland
50%
Flanders
40% 30% 20% 10%
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0%
Universities (accum. %)
Patent Concentration on Inventors
While inventive activity is concentrated on a relatively small number of inventors, the Finnish universities seem to exhibit more variety in the concentration of academic patents on inventors than their Flemish counterparts. This is illustrated in Figure 2 for a selection of
8
Finnish and Flemish universities that represent those cases with the strongest and the least concentration of patents on academic inventors. The x-axis describes the accumulated percentage of inventors (in descending order) while the y-axis gives the accumulated percentage of patents. This way one can see how many percent of inventors invented a certain percentage of patents. KU Leuven in Flanders and Oulu University in Finland exhibit the strongest inventor concentrations in an almost similar pattern. The top ten percent of the inventors account for more than forty percent of the patents associated with the universities. While in Flanders, the case of least inventor concentration (State University of Ghent - RU Gent) shows at this level no difference, the top ten percent of inventors at Helsinki University and Turku University (here not displayed) in Finland accounted for only a little more than a quarter of the patents. All in all, KU Leuven and RU Gent concentrations are closer to each other than the concentrations observed for the Universities of Oulu and Helsinki, which seems to indicate more variety of Finnish universities with respect to inventor concentration.
So, what could explain the differences? Does the type of intellectual property regulation have a smoothing, leveling effect in Flanders? Are different technological (and scientific) profiles and hence different technological patterns of inventiveness a possible explanation? Or could varying IP policies of universities explain the differences between Finnish institutions? We have shown elsewhere (Meyer et al., 2003a,b) that the Finnish universities have varied technology profiles. In addition, observers point to different approaches and attitudes towards technology transfer and universities in Finland (Kutinlahti, 2003). In Flanders, universities tend to have a more ‘homogeneous’ patenting profile (Du Plessis, 2003) even though transfer activities still differ in their extent.
9
Figure 2
Patent concentration on academic inventors with selected universities 100% 90%
Patents (accum. %)
80% 70% 60%
KU Leuven (FL) RU Gent (FL)
50%
Oulu (FI) 40%
Helsinki Univ (FI)
30% 20% 10%
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0%
Inventors (accum. %)
Patent Concentration on Assignees and Ownership Structure
While the assignee concentration is about the same in Finland and Flanders (see Figure 3), there are considerable differences in the assignee structure. The differences are not unexpected because of the different intellectual property regulations in Finland and Belgium. In Finland the academic researcher retains the right to patent his or her research results whereas in Flanders academics have to file invention reports to the university’s transfer organization which then pursues patenting this seems reasonable.
Finnish university-related patents are concentrated on a relatively small number of assignees (i.e., owners). The top twenty percent corporate assignee firms can be associated with about 70% of the academic patents. Nokia is with 56 university-related patents the largest owner, followed by Orion, Valmet, Fortum, Instrumentarium, Kone and Ahlstrom are the next largest firms (see Table 1). All of the aforementioned companies are large, established firms. Helsinki University Licensing (now Licentia) is the next with a total of 11 patents in the database. This raises the question to what extent academic entrepreneurs use patents to protect their inventions.
10
In Flanders, university or other public research organizations feature much more prominently among the list of assignees. This observation is underlined by an individual assignee ranking (see Table 1). Universities own 120 of the Flemish academic patents. Co-patenting plays an important role. About 10% of Flemish academic patents are assigned to more than one owner. More than a third of these jointly owned patents involve a university or other public sector organization as assignee. In light of the IP regulations for Flemish universities this is not a surprising result.
What may be surprising is the relatively strong position of non-university assignees. Industrial companies owned about three in five academic patents exclusively. Table 2 provides an overview of the ownership structure of Flemish academic inventions. The data indicates that collaborative modes of technology transfer do play a considerable role also in systems in which universities own the intellectual property rights to their employees’ research. Another implication is that an indicator, such as the mere number of university-owned patents, does not fully appreciate the wealth of contributions university researchers make to technological development. This underlines Pavitt’s (1998) skeptical view of university patents as an indicator of useful university research and also illustrates that Meyer’s (2003) more inclusive notion of university-related or academic patents captures more comprehensively the contributions of academic science to a country’s technology base.
Patent concentration on assignees of university-related patents 100% 90% 80% 70% 60% Finland
50%
Flanders
40% 30% 20% 10%
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0%
Patents (accum. %)
Figure 3
Assignees (accum. %)
11
Table 1 # 1 2 3 4 5 6 7 8
Most prominent owners of academic patents
Finland Assignee Nokia Orion Valmet Fortum Instrumentarium Kone Ahlstrom HUL (Licentia)
Table 2
Sector Industry Industry Industry Industry Industry Industry Industry Public Sector
# 1 2 3 4 5 6 7 8
Flanders Assignee IMEC Exxon Plant Genetic Systems K. U. Leuven Corvas International ZENECA Rega Stichting Agfa-Gevaert
Sector Public Sector Industry Industry Public Sector Industry Industry Industry Industry
Ownership Structure of Flemish academic patents
1992-2001Before 19951995 and later Industry 61.0% 67.2% 57.9% University and Public Sector 28.5% 19.7% 32.9% Unassigned or in private ownership 5.5% 5.7% 5.4% Joint University/Industry ownership 3.3% 5.7% 2.1% Joint University/Private ownership 1.4% 0.8% 1.7% Joint Government/Industry Ownership 0.3% 0.8% 0.0%
Having said this, one needs to acknowledge the change in ownership structure that has taken place over the last decade. There is a strong trend towards a higher degree of universityownership of academic inventions. Two out of three academic patents with an application date prior to 1995 were owned by an industrial company; only one in five patents was assigned to a university or public sector organization. Academic patents with an application date of 1995 and later are characterized by a shift in ownership structure towards the university and other public sector institutions. While patented academic inventions are still predominantly assigned to the private sector, almost a third of them are assigned to university or other public sector institutions.
This reflects the more active role universities have taken towards patenting and intellectual asset management and corresponds with the ‘surge’ in university patenting Saragossi and van Pottelsberghe (2003) observe in their study of patenting in major Belgian universities. Nevertheless, the data shows that even in science systems where universities principally own the rights to their researchers’ work, a look at university-owned patents would present a very narrow view of the technological impact of academic research. As academic patents include also inventions made by researchers while not working in universities, they also cover partly the mobility of researchers.
12
Foreign Assignations
Expected differences can be traced with respect to foreign assignations of university-related patents. In Finland, about 13.8% of all university-related patents are assigned to foreign-based organizations. Biotechnology patents account for by far the biggest share (26%). The nextlargest fields are pharmaceuticals (13.7%), electrical devices/electrical engineering (9.6%) and organic, fine chemistry (8.2%). A specialization analysis confirms that, of the aforementioned areas, biotechnology and electrical devices - electrical engineering are the technological sectors in which university-related patents tend to be over-proportionally foreign-assigned. In these fields, relatively more patents have been assigned to foreign organizations than their overall share would suggest.
Flanders as well as Belgium are often characterized as an open, small economy in the center of Europe with a high share of patenting of foreign firms. The Flemish academic patents underline this idea. A considerable number of them can be related with international corporations, so it is not surprising that about 23% of all Flemish academic patents are foreign-assigned. The most prominent cases are Zeneca and Alcatel.
Individual Inventors as Assignees
There are clear differences between Finland and Flanders with respect to so-called unassigned patents, i.e., patents that are not assigned formally to anybody else than the individual inventors. While 16% of university-related patents are unassigned in Finland, only 5% are in Flanders. This is a finding that is most likely related to the different ways in which intellectual property is handled in universities. Figure 4 compares the shares of unassigned to assigned patents in Finland and Flanders and describes the technological association of the patents in more detail for the Finnish subset.
A total of 85, or 16%, of academic patents in Finland are neither assigned to a company nor to another organization. This does not mean individual inventors have not licensed the patent to one or several companies for further utilization. However, the data shows where individuals own the patents and a corporate user of the inventions is not immediately visible. Medical engineering, instruments and organic, fine chemistry are the fields that account for most of the unassigned patents. 13
There are considerable variations between the universities. On one hand, there are a few universities where none of the inventions remained unassigned. On the other, there are universities where a considerable number of patents are still owned by individual inventors. In one instance, more than 40 percent of the academic patents were unassigned (see Meyer et al., 2003a, for details). In three cases, the rate was between a quarter to a third of all patents. In comparison to all university-associated patents, the share of unassigned patents was overproportionally large in analysis/measurement/control, medical engineering, organic, fine chemistry.
A far smaller number of patents are assigned to university researchers in Flanders. Only 20, or 5%, of all Flemish academic patents are assigned to individuals. To some extent, these patents can be related to the university. As for technological areas, unassigned patents occur four times in the field of medical engineering, three times in optics. Instruments, biotechnology and materials processing are the other areas that account for more than one patent that is not assigned to the university or any other organization.
Again, this observation is not surprising if one considers the different intellectual property regulations that are in place in Flanders and Finland. Inventors, if they are employed can only obtain the patent right if the university or its transfer organization is not interested. In addition, one should mention occasional joint university-private ownerships of patents. For instance, a professor at KU Leuven jointly owns a number of patents with the university. He runs a foundation in his own name that is associated with the R&D division of the university.
With respect to the considerable differences between the two systems, one must be careful to infer less economic utilization of Finnish academic patents because of the higher share of unassigned patents. Further survey research on a set of unassigned academic patents in Finland indicated that 60% of the unassigned inventions were utilized in industry while about a third of these patents found no application; around 8% were utilized in a non-commercial context. Almost a third of the patents that were utilized in a commercial context were associated with start-up firms.
14
Figure 4
Unassigned vis-à-vis assigned patents
Flanders
Finland
Unassigned, 19 patents, 20 5%
Analysis, measurement, control 14 Organic, fine chemistry 11
Assigned, 445 patents, 84%
Telecommunications 10
Unassigned, 85 patents, 16%
Biotechnology 9 Pharmaceuticals, cosmetics 9
Assigned, 360 342 patents, 95%
Other 17
Specialization
The differences in specialization raise the question as to what extent some of the diverging observations can be explained by IP policy and regulation or by technological patterns. Telecommunications and instrument-related patents are the largest technological sectors in university-related
patenting
in
Finland,
with
more
than
12%
respectively.
Pharmaceuticals/cosmetics and biotechnology account for about 10% of the university-related patents. Research elsewhere that followed up Finnish academic patents in more detail (Meyer et al., 2003b) suggests that patterns of collaborations vary with certain technology sectors. Not surprisingly the picture is different for the Flemish patents. Here, the focus is not on telecommunications but on organic chemistry and life-science related technological areas. Both fields account for about 16% of all patents. These results correspond to what Saragossi and Van Pottelsberghe (2003) observed for the three Flemish universities in their study of European patent applications of selected Belgian universities.
15
Conclusions Differences between Finland and Flanders
Our research points to expected differences as well as similarities in academic patenting that may not have been anticipated. As for expected differences, the different assignee structures and the higher rate of unassigned patents in Finland are to be mentioned. Far more patents owned by universities and public assignees in Flanders than in Finland. Given the different intellectual property regimes, this is no surprise.
We observed that the number of unassigned patents is higher in Finland than in Flanders, which is an expected finding. In Finland, every researcher still has the right to patent and commercialize his or her research results. This does not always require transferring the ownership of the patent to a corporation or other organization. Proprietor-entrepreneurs and start-up academic entrepreneurs may choose to keep the patents in their name and not the company’s.
Our study has illustrated a certain variance in Finland with respect to the within-university concentration of patents on key inventors. The inventor distributions varied to a larger extent than in the Flemish case. This raises the question to what extent the general IP system has affected this link. In the Finnish case, the universities still have more room to shape their local approach towards IP while in Flanders the overall regulation may further a similar type of behavior among researchers. On the other hand, Finnish universities differ substantially in their technological profiles. This may also relate to the differences in inventor concentrations.
The difference in inventive output was within the range of expectations for both systems. Finnish academics seem to be more inventive than their Flemish colleagues to a similar extent as the difference between Finland and Belgium in terms of overall patent productivity would suggest. This could imply that inventiveness is rather associated with national innovation systems or cultural factors than with the incentive structures and IP regulations.
16
Commonalities between the Two Innovation Systems
A number of commonalities are also worthwhile mentioning. In both systems, industrial ownership of academic patents is high. This is not surprising in Finland. Universities do not to own patents to a large extent. In Flanders, however, the university sector plays a much more important and still growing role. Yet, the academic patent data presented in this paper places this increased activity of universities in managing their intellectual assets in the broader context of collaborative research and university-industry collaboration. Large firms play a key role in absorbing academic inventions in Flanders as well as in Finland.
In spite of different IP regulations in the higher education sector, academic inventive activity appears to be almost similarly distributed across universities in Finland and Flanders. Also, inventor and assignee concentrations are similar in both Finnish and Flemish cases. This seems to point to a more general pattern that a large share of the academic patenting is concentrated on a small number on institutions.
This and the earlier observations have considerable interpretational and methodological consequences.
Implications for Methodology & Interpretation of Indicators
Two findings are particularly relevant from a methodological perspective on academic patents as indicators of useful technological output of university research. First, the data on Flemish academic patents illustrates that university-owned patenting is a rather narrow indicator of inventive contributions of university researchers even in a system in which academics do not own the rights to the inventions and universities have established transfer and licensing organizations. Business firms owned more than 60 percent of all Flemish academic inventions. In other words, there are three industry-assigned patents on any two universityowned patents.
In light of these findings, Pavitt’s (1998) skepticism towards university-owned patents as a measure of useful research seems justified. The findings also show that the more inclusive approach of tracing patented inventions made by academics allows the analyst to identify a much broader field of academic technological activity, irrespective of the university IP 17
regulations in a country. The indicator also captures to some extent the mobility of university researchers. Of course, this more inclusive measure is just one of several indicators of useful academic research and potential entrepreneurial activity in higher education (see e.g. Salter et al., 2001).
Suggestions for Future Research
Suggestions for future research are related to the observation that the absolute level of university-related inventiveness does not seem to vary substantially between Flanders and Finland. Given that both Finland and Flanders are of approximately the same size, this result may not be surprising. On the other hand, the handling of intellectual property rights is different in both systems. Finnish university researchers own the rights to the results of their research whereas their Flemish colleagues transferred all the rights to the university. This raises the question whether or not the different IP systems influence the output level. The evidence presented in this study seems to be in favor of cultural or national factors.
Another issue future research needs to address is how to measure the further utilization of university patents. Here, a combination of informetric with survey approaches may be helpful. This is an important issue especially if one seeks to compare inventive activity and its commercialization in countries or regions with different intellectual property regulations at university. The much higher rate of individual academic inventors owning ‘their’ patents in countries where researchers have the privilege to patent their research results themselves, such as Finland, and the high share of university and other public research organizations as owners of their researchers’ intellectual property, as it is the case in Flanders, raise the question to what extent these patents are also commercially exploited. As resource allocation mechanisms geared towards increasing third-stream activities also include patent counts, universities may hold patents for another reason than commercializing valuable inventions.
Patent data alone cannot give an answer to questions about the use of patents. A survey of inventors may shed light on this. To warrant comparability a common methodology in data collection should be applied. Without any further information, the analyst would find it difficult to make any judgments about ‘transfer efficiencies’ of the different systems. Lastly, a problem this first exploration was not able to address is the extent to which cooperative research and use of the patents through the industrial partner relates to academic 18
entrepreneurship. Here, future research needs to complement pure patent analyses with a hybrid approach that takes academic patents as the basis.
19
References Acosta, M.; Coronado, D. (2003), Science-technology flows in Spanish regions - An analysis of scientific citations in patents. Research Policy, 32 (10) 1783-1803. Caro, J.M.A; de Lucio, I.F.; Gracia, A.G. (2003), University patents: output and input indicators ... of what? Research Evaluation, 12 (1) 5-16. Carpenter, M.P.; Narin, F. (1983), Validation study: patent citations as indicators of science and foreign dependence. World Patent Information, 5 (3) 180-185. Clark, B.R. (1998), Creating Entrepreneurial Universities, Pergamon, Oxford. Debackere, K. (2000), Managing academic R & D as a business at K.U. Leuven-context, structure and process. R&D Management, 30, 4, 323-328. Du Plessis, M. (2003), Academic patents in Flanders. Presentation, KU Leuven. Etzkowitz, H (2003), Innovation in innovation: the Triple Helix of university-industrygovernment relations. Social Science Information, 42 (3) 293-337. Etzkowitz, H., A. Webster, P. Healey (eds.) (1998), Capitalizing knowledge: new intersections of industry and academia. SUNY Press: Albany, NY. European Commission (2003), Science, Technology and Innovation Indicators for the European Research Area (STI-ERA). Henderson, R.; A. Jaffe; M. Trajtenberg (1998), Universities as a source of commercial technology: a detailed analysis of university patenting, 1965-1988, Review of Economics and Statistics, 13, 119-127. Huber, J.C. (1998), Invention and inventivity as a special kind of creativity, with implications for general creativity. Journal of Creative Behavior, 32 (1) 58-72. Kutinlahti, P. (2003) SMEs and the new role of academic research. The country report: FINLAND. VTT Technology Studies, Espoo. Meyer, M. (2000b), Tracing knowledge flows in innovation systems – An informetric perspective on future research on science-based innovation. Economic Systems Research, 14 (4) 323-343. Meyer, M. (2002a), Tracing knowledge flows in innovation systems. Scientometrics, 54 (2) 193-212.
20
Meyer, M. (2003), Academic Patents as an indicator of useful research? A new approach to measure academic inventiveness? Research Evaluation, 12 (1) 17-27. Meyer, M., T. Goloubeva, J.T. Utecht (2003a), Free Patent Information as a Resource for Policy Analysis. World Patent Information, 25 (3) 223-231. Meyer, M.; Siniläinen, T.; Utecht, J.T. (2003b) Towards hybrid Triple Helix indicators – a study of university-related patents and a survey of inventors. Scientometrics, 58 (2) 321-350. Meyer-Krahmer, F., U. Schmoch (1998), Science-based technologies: university-industry interactions in four fields. Research Policy, 27 (8) 835-851. Mowery, D. C.; Nelson, R. R. ; Sampat, B. N.; Ziedonis, A. A. (2001), The growth of patenting and licensing by U.S. universities: an assessment of the effects of the BayhDole act of 1980, Research Policy, 30, 99 – 119. OECD (2003), Turning science into business: Patenting and Licensing at Public Research Organisations. OECD, Paris. Pavitt, K (1982), R&D, patenting and innovative activities – a statistical exploration. Research Policy, 11 (1): 33-51. Pavitt, K. (1998), Do patents reflect the useful research output of universities? Research Evaluation, 7 (2) 105–111. Roberts, E (1991), Entrepreneurs in high technology. New York: OUP. Salter, A., P. D'Este, B. Martin, A. Geuna, A. Scott, K. Pavitt, P. Patel, and P. Nightingale (2001), Talent, not Technology: Publicly funded research and innovation in the UK. London, Committee of Vice-Chancellors and Principles and the Higher Education Funding Council of England. Saragossi, S; B van Pottelsberghe (2003), What patent data reveal about universities: the case of Belgium, Journal of Technology Transfer, 28 (1) 47-51. Schild, I. (1999), A regional patent study to investigate inventive activity in East Gothia. Tema-T Working Paper No.207, April 1999, Linköping University. Sellenthin, M.O. (2003), Who should own university research? An exploratory study of the impact of patent rights regimes in Sweden and Germany on the incentives to patent
21
research results. ITPS, Swedish Institute For Growth Policy Studies, Östersund, Report # A2004:013. Shane, S.A. (1992), Why do some societies invent more than others? Journal of Business Venturing, 7 (1) 29-46. Smith, K., A. Ekland, E. Iversen, A. Kaloudis, P. Patel, R. Narula (1998), Understanding Science, Technology and Innovation Indicators, IDEA-Report 5/1998, Oslo: STEP Group. Tijssen, R.J.W. (2002), Science dependence of technologies: evidence from inventions and their inventors. Research Policy, 31 (4) 509-526. Van Looy, B ; J Callaert; K Debackere, A Verbeek (2003), Patent related indicators for assessing knowledge-generating institutions: toward a contextualised approach, Journal of Technology Transfer, 28 (1) 53-61. Verbeek, A.; Debackere, K.; Luwel, M.; Andries, P.; Zimmermann, E.; Deleus, F. (2002), Linking science to technology: Using bibliographic references in patents to build linkage schemes. Scientometrics, 54 (3) 399-420.
22
