. Page 1 of 17 ... Henderson et al., 1998; Agrawal,. Henderson, 2
Co-patenting between evidence from Italy
University
and
Industry:
By Roberto Parente, Michele Petrone and Daniele Cerrato1
This paper focuses on academic patents owned by business companies and investigates the factors affecting the industry involvement in academic research aimed at obtaining new patents. The focus is on co-patenting, i. e. those patents in which university professors are inventors, while business companies are patent applicant. In particular, we investigate if the copatenting between academicians and companies is affected by two main factors: the academic research quality and the regional industry structure. The empirical analysis is based on a sample of 300 patents produced by university professors as inventors. We have selected professors belonging to seven Italian Universities. In summary, the results of the analysis, suggest that factors related to the regional context, rather than factors related to the academicians’ performance assume a major role in promoting business involvement in academic research activities that will end in a new patent.
Roberto Parente is Full Professor at University of Salerno. Michele Petrone Phd is Technology Transfer Consultant at Basilicata Innovazione (Area Science Park) Daniele Cerrato Phd is Assistant Professor at Università Cattolica del Sacro Cuore –Faculty of Economics, Piacenza Address corrispondence to Roberto Parente, Università di Salerno, Via Ponte don Melillo 84048 Fisciano (SA) email
[email protected] 1
1. Introduction The university-industry collaboration is considered a crucial factor for firms‟ competitiveness and economic development (Etzkowitz, Leydesdorff, 2000; Cooke et al., 2004; Bonaccorsi, Daraio, 2007). Many research efforts have been made to understand the factors affecting the transfer of research results to the industrial system (Rosenberg, 1990; Pavitt, 1991; Henderson et al., 1998; Agrawal, Henderson, 2002; Cohen et al., 2002; Jaffe et al., 2003; Laursen, Salter, 2004). Within literature on universityindustry collaboration, several studies focus on academic patents (Jaffe, 2000; Mowery et al., 2001; Sobrero et al., 2006; Meyer, 2006; Calderini et al., 2007; Fabrizio, Di Minin, 2008). In particular, the exploitation of academic patents through their transfer from university to industry is considered the most controversial and difficulty phase (Dechenaux et al., 2011). Data on US patents show that the growth in academic patents applications has been higher than the corresponding growth in academic patents licensed to the industry, resulting on a reduction of the commercial
appeal of the academic research (Thursby and Thursby, 2002). In Italy, according to the last survey of Netval (Network for the valorization of university research), universities hold a total portfolio of 1.949 patents; but only 245 license agreements have been concluded between universities and industry, equal to 12.5% of the patents stock (Netval, 2010). But along patents that have been invented by academicians and registered by the university (academic patents owned by universities) there are also a number of patents that are invented by academicians but are registered since the beginning by companies (academic patents owned by business companies) (Balconi et al., 2004; Fabrizio, Di Minin, 2005; Thursby et al., 2006). Empirical evidence suggests that academic patents owned by companies prevail in European Universities (Lissoni et al., 2008). In Italy, for example, Lissoni et al. (2008), show that almost 74% of Italian patents in which there is at least one Academician as inventors are owned by business companies. In these cases flows is very well known, university professors define research agreements with business companies, receiving research grants but transferring
the intellectual property that could result from the research project to the companies.
2. Literature review Hypothesis
and
2.1-Academic research quality and co-patenting. Many studies define a relation between the research quality and the probability of interaction with industry (Tornquist, Kallsen, 1994; Mansfield, Lee, 1996; Lissoni, 2010; Laursen et al., 2010). Abramo et al. (2010) demonstrate the importance of academicians scientific profile in the companies‟ choice about collaborations. Bonaccorsi and Daraio (2007) provide a number of explanations for this relation: the professors‟ international visibility, their ability in managing laboratories and capability in attracting the best students and Phds are valuable for business companies. Lissoni and Montobbio (2006) underline an effect of reputation, because the most productive professors have greater visibility and are more likely to be contacted by business companies. These studies support the hypothesis of a positive
relation between the quality of academic research and the probability to attract the industry at a very early stage of development of IP. In particular, we investigate if the quality of the research made by the professorsinventors influences the probability of co-patenting with business companies. Hypothesis 1: The probability that a patent is co-patented between academicians and industry is positively affected by the quality of research of the academicianinventor.
2.2-Regional industry and copatenting. A broad range of literature focuses on the relationships between the characteristics of the local environment and university– industry collaboration (Mansfield, Lee, 1996; Jaffe, 1989; Breschi, Lissoni, 2001; Arundel, Genua, 2004). Mansfield and Lee (1996) wrote a seminal paper on the role of geographical distance in the relationship between research and industry. They found that the probability of a firm funding academic R&D is negatively related to the distance between the firm and the university. This field
of literature includes studies at both macro and micro level. At macro level, several studies analyze the relation between regional indicators of innovation and the exploitation of research results in local universities (Siegel et al., 2003). In Great Britain, Chapple et al. (2005) have concluded that universities localized in Regions with higher GPL and industrial R&D are more effective in technology transfer activities. In Italy, Evangelista et al. (2002) have classified regions in four groups on the basis of the regional innovation system and the level of public research exploitation. These empirical studies support the hypothesis that the level of innovation in the regional industry influences the probability of co-patenting between university professors and business companies. Hypothesis 2: The probability that a patent is co-created between research and industry is positively affected by the innovation capacity of industrial system in the Region At micro-level, many studies analyze the role of clusters in explaining the relation between university and industry (Saxenian, 1994; Lester, 2005). Sub-regional effective clusters are characterized
by one or more universities, with an excellent research base in a specific scientific-technological field, and business companies that are able to develop and apply research results (Etzkowitz, Leydesdorff, 2000; Cooke, Piccaluga, 2004). Technological parks, incubators, liaison offices facilitate university-industry collaborations in local clusters. Public governments that invest in cluster building as a strategy to develop innovation can boost the development of University Indusrty relationship (Doloreux, Dionne, 2008). We investigate if the strength of a sub-regional cluster influences the probability of co-patenting between university professors and business companies. Hypothesis 3: The probability that a patent is co-created between research and industry is positively influenced by the existence of a Regional cluster focused on the technology stream that is attributable to the patent itself.
3. Descriptive statistics The empirical analysis is based on a sample of 282 academic patents, filed by university professors as inventors. We have selected
professors belonging to seven Italian Universities (University of Trento, Polytechnic of Milan, University of Florence, University of Salerno, University of Bari, University of Calabria, University of Catania). We have used names and surnames of university professors in the European Patent Office database (espacenet), for searching patents filed by university professors. Each patent has been coded on the basis of its applicant, i.e. the owner of intellectual property rights. Specifically, the sample have been classified, according to the type of applicant, in three categories: industry as applicant (academic patents owned by business companies, i.e. patents defined as „co-patented‟); university as applicant (academic patents owned by the university); professors as applicant (academic patents owned by the same professor). The sample consists of: 49.3% (139 observations) of patents that belonging to the category "firm as applicant"; 43.6% (123 observations) of patents belonging to the category "university as applicant"; and, finally, a residual proportion of 7.1% (20 observation), patents in which the researcher-inventor is also applicant (Table 1).
*** Table 1 about here *** About half of the patents produced by academic inventors, therefore, is filed directly by companies, which have cooperated with the university in the achievement of patented results. Typically, the cooperation starts as a financial grant to a project promoted by the same company, and regulated by an explicit agreement, according to which the company has the right to exploit any innovative results that will be reached through the project. In the majority of the remaining cases, patents belong to the university. It means that the university assumes, in these cases, the costs of filing and maintaining the patent. Finally, in a very limited number of cases, the professor is the exclusive applicant of the patent. The sample can be divided into two groups: a first group where the percentage of patents belonging to the category "Firm as Applicant" is prevalent (54.5% at Polytechnic of Milan; 57.4% at the University of Florence; 53.8% at University of Trento; 71.8% at the University of
Catania); a second group where the university or the same professor result as applicant of the patent as usual situation (76.7% at the University of Salerno; 64.5% at the University of Bari; 67.8% at the University of Calabria). 4. Variables and measures An econometric analysis was performed to test our three hypotheses about the relationships between the „co-patenting‟ level and a) the academic research quality, b) the structure of the regional industry with regard to innovation, c) the presence of a cluster around the university. The dependent variable of the study is the likelihood that an academic patent is owned by a business company. We have introduced a dummy variable, which assumes value 1 if the company is patent applicant and 0 if the university is patent applicant. There are three independent variables in our study: - the quality of academic research of professor-inventor, measured by the number of citations of the most important publication of the professor-inventor (Breschi et al., 2008) (source: google scholar);
- the innovation level of the industrial system in the Region, measured by the percentage of innovative companies on the total number of companies in the Region (source: Italian Institute of Statistics – ISTAT); -the presence of a cluster, in the same technological field of the patent, within the area where the university is located. This variable has been measured as a dummy that assumes value 1 if in the area there is a high concentration of companies in the same technological field of patents and a dedicated technology transfer organization, 0 otherwise. Table 2 shows the variables and measures. *** Table 2 about here *** 5 Results and discussion Table 3 shows correlations between variables included in the analysis, while Table 4 presents the results of the econometric analysis. *** Table 3, 4 about here ***
In our model the dependent variable is the probability of the patents to be co-generated between universities and firms, i.e. the probability that a firm acts as the Applicant of the patent. Since this variable takes values 1 (the firm is the patent applicant) or zero (the firm is not the applicant), the models estimation are appropriate univariate dichotomous (or binary choice) models are appropriate estimation models, which are intended to describe the 'choice' between two discrete alternatives (Verbeek, 2004). The estimates were, therefore, made using the probit method, which falls into this class of models. In addition, in the econometric results (Table 4) the variables 'Research Quality' and 'Number of professors' were logarithmically transformed in order to reduce the high asymmetry in the distribution of the variables. The results do not confirm the hypothesis about the positive relationship between cogeneration of patents and quality of research: the number of citations does not increase the probability that the firm acts as the Applicant of the patent. The quality of research made by the academician-inventor is not a key element to attract firms in joint research activities. In other terms,
the scientific credentials of the researcher / research team does not significantly arise the interest of the firms. This result may be interpreted as a controversial signal about the nature of the link between excellence in scientific research and development of partnerships between universities and firms. Obviously other factors, on the research side, impact on business involvement. Recent literature (D'Este, Patel, 2007, Giuliani et al., 2010), for example, focuses on the individual characteristics of the researcher to explain the levels of collaboration with firms. These characteristics also emerge from the analysis of our sample, specifically with respect to the so called „academic serial inventor‟. D'Este and Patel (2007) show that researchers with a history of collaboration (more or less structured), with firms are more likely to be involved in a variety of business interactions in the future. Similar empirical evidence are also provided by other studies. Schartinger et al. (2001), show that the decision of a firm to work with university departments depends on the level of satisfaction due to previous collaborations with the researchers themselves. Our findings seem to support the view that enterprises appear to be less sensitive to the driver coming
from the most advanced research results. Geographical proximity is the factor that mostly explains the probability of “co-patenting”. In fact, both hypotheses 2 about influence of regional industrial system and hypothesis 3 on the "cluster" effect are confirmed by the empirical analysis: co-generation of patents between university and firms is positively and significantly affected (p