University autonomy, the professor privilege and academic patenting: Italy, 1996-2007

Francesco Lissoni1,2, Michele Pezzoni2,3, Bianca Potì4, Sandra Romagnosi5

GREThA – Université Bordeaux IV - France

2CRIOS – Università "L. Bocconi", Milano – Italy

3 DEMS, University of Milano-Bicocca

4 CERIS-Consiglio Nazionale delle Ricerche, Rome - Italy

5 National Agency for the Evaluation of Universities and Research Institutes (ANVUR), Rome – Italy

Corresponding author: F.Lissoni, GREThA - UMR CNRS 5113, Université Montesquieu - Bordeaux IV, avenue Léon Duguit, 33608 Pessac cedex – FRANCE ( ; tel. +33 (0)5 56 84 86 04)

Abstract

Using data on patent applications at European Patent Office, we search for trends in academic patenting in Italy, 1996-2007. During this time, Italian university underwent a radical reform process, which granted them autonomy, and were confronted with a change in IP legislation, which introduced the professor privilege. We find that, although the absolute number of academic patents has increased, (i) their weight on total patenting by domestic inventors has not, while (ii) the share of academic patents owned by universities has more than tripled. By means of a set of probit regressions, we show that the conditional probability to observe an academic patent has declined over time. We also find that the rise of university ownership is explained, significantly albeit not exclusively, by the increased autonomy of Italian universities, which has allowed them to introduce explicit IP regulations concerning their staff's inventions. The latter has effectively neutralised the introduction of the professor privilege.

Keywords: academic patenting, university autonomy, professor privilege

JEL codes: I23, O31, O34

Acknowledgements: Inventor data come from APE-INV, the project on "Academic Patenting in Europe" sponsored by the European Science Foundation (http://www.academicpatenting.eu). Participants to the "NameGame" APE-INV workshop series have all contributed to establish a robust methodology for name disambiguation, which has been essential for the creation of the database. Several Italian colleagues have provided us with data they collected over the years: Cinzia Daraio and Andrea Bonaccorsi (Aquameth data), Andrea Piccaluga (NetVal data), Rosa Grimaldi, Riccardo Fini, and Maurizio Sobrero (universities' IP regulations). Paolo “Colchonero” Freri provided valuable research assistance.

1. Introduction

After being granted autonomy in 1989, Italian universities have generally moved towards a more active management of their sources of revenues and financial assets (Geuna and Rossi, 2013). Among other things, they have taken an active interest in the commercialization of research results, by introducing intellectual property (IP) regulations aiming at securing control of the relevant patents. These initiative clashed however, in 2001, with the introduction of piece of legislation (the “professor privilege”) that nominally transferred all IP rights over academic research results from the universities to their faculty (Granieri, 2010; Baldini et al., 2010 and 2012). Italy is therefore an interesting case study, which allows assessing the relative impact on academic patenting of two contrasting forces: (i) a change in the governance and funding system of universities (the autonomy) and (ii) a change in the IP legislation (the professor privilege). As such, it contributes to the emerging literature on European universities’ changing practices concerning IP management, as a result of broader policy changes in higher education and research (Della Malva et al., 2013; Mejer, 2012). It also contributes to the literature that has either re-considered the professor privilege as an alternative to IP policies prevailing in US universities, or assessed its historical impact in the European countries which first introduced it, and then, in most cases, abolished it (Valentin and Jensen, 2007; Iversen et al., 2007; Lissoni et al., 2009; Greenbaum and Scott, 2010; Kenney and Patton, 2011; Geuna and Rossi, 2011; von Proff et al., 2012; Damsgaard and Thursby, 2013; Schön and Buenstorf, in this special issue).

Our contribution is mainly empirical and exploit a new and original database in order to:

-  Build reliable estimates of academic patenting in Italy throughout the period 1996-2007;

-  Test for the existence of time trends concerning the weight of academic patenting over total patenting, and its ownership distribution between universities and other subjects;

-  Estimate the impact of both the autonomy of universities and the introduction of the professor privilege on the observed trends.

We find that the weight of academic over total patenting has remained stable overtime, while the share of university ownership has increased significantly. We also find that the adoption of internal IP regulations by universities, an immediate consequence of their newly conquered autonomy, appears to have increased university patent ownership and effectively neutralized the professor privilege.

We proceed as follows. In section 2 we put forward a number of propositions concerning the determinants of academic patenting and the distribution of its ownership between universities, their faculty, and industry. In section 3 we provide an historical account of Italian universities’ gradual implementation of autonomy in the 1990s and the abrupt introduction of the professor privilege. Section 4 describes in extreme synthesis our methodology and data. Section 5 contains our econometric exercise and the discussion of results. Section 6 concludes.

2. University autonomy, IP legislation and the ownership of academic patents

We define "academic" any patent signed at least by one academic scientist, while working at his/her university, whether the patent is owned by the university, a public research organization (PRO), the scientist, a business company or any other organization, either exclusively or jointly with other assignees (Lissoni, 2012; Dornbusch et al., 2013). This choice is now common in the literature and it reflects a fundamental theoretical issue, one which concerns: (i) the origin of academic inventions, (ii) the nature of IP legislation, and (iii) the governance model of universities in different countries.

With respect to the origin of academic inventions, two broad categories can be put forward (Franzoni and Lissoni, 2009) . One category coincides with what Jensen and Thursby (2001) call “proofs and prototypes for sale”, namely the inventive results of fundamental research, mostly public-funded, which necessitate further development in order to be commercialized. In the US, these inventions were the target of the Bayh-Dole Act of 1980, which assigned them exclusively to universities, as opposed to the funding agencies or the inventors, with the intent of pushing them to engage actively in their commercialization (Mowery, 2001). The second category consists of inventions arising from collaboration with industry, with three actors involved in negotiations over IP: the academic inventor, his/her business partner, and the university administration. Collaboration models range from consultancy to contract research, and to joint research activities. As for the resulting IP, we expect the inventor and the business partner to have a greater say when it comes to consultancy (which in principle does not impinge upon the university’s resources), while administrations should have a greater say when moving in the direction of joint research activities (which may involve students, staff, and other resources of the institution). Overall, these inventions may be as important, technologically and commercially, than those in the first category (see Link et al., 2007, and Jensen et al. 2010, on consultancy; and Colyvas et al., 2002, on results of joint research).

As for the IP legislation, historically this has intervened in two ways. One concerns the rules attached to public funding, as in the case of the Bayh-Dole Act (Mowery and Sampat, 2005). The other concerns the relationship between the academic inventor and his/her university. While general IP legislation assigns the property of employees’ inventions to their employer, several European countries were until recently characterized by a special IP regime for academic inventions, known as “professor privilege” (von Proff et al., 2012; Damsgaard and Thursby, 2013). Once common in all German-speaking and Scandinavian countries, and nowadays surviving only in Sweden and Italy, this regime prescribes that the university has no title over the faculty’s inventions, unless decided otherwise by the inventor. Historically, this explains the high level of individual ownership of academic patents in Sweden, Denmark, and Germany (Lissoni et. al., 2008 and 2009; von Proff and Buenstorf in this issue).

Finally, the distribution of IP between faculty, business partners, and universities is affected by the latter’s autonomy from governmental control, which determines their capability to steer and control their faculty’s activities, including inventive ones. A synthetic definition of autonomy is provided by Aghion et al. (2010), who consider two parameters: (i) whether a university’s budget needs to be approved by the state; and (ii) the percentage of the university’s budget associated to competing grants, as opposed to block grants. A more complete conceptualization is provided by the European University Association (EUA), which measures autonomy by looking « […] at the ability of universities to decide on:

• organisational structures and institutional governance […]

• financial issues, in particular the different forms of acquiring and allocating funding

• staffing matters, in particular the responsibility for terms of employment [which may include IP matters]

• academic matters, in particular the control over student admissions » (Estermann and Nokkala, 2009; p.7)

Historically, European universities have never enjoyed the same degree of autonomy of their US counterparts (Ben-David, 1977; Clark, 1993). This has limited their control over their finances (including IP assets and related revenues) and staff (including the freedom to set up clear rules over IP concerning their faculty's inventions, such as disclosure obligations and rewards). In the absence of such control, European universities have traditionally resisted being involved in IP management, and often took the shortcut of allowing scientists to take their own decisions, even in the absence of the professor privilege. When engaged in cooperative or contract research with third parties, the latter often signed blanket agreements leaving all IP rights in their partners’ hands.

As a result, a large part of academic patents in Europe has for long gone unnoticed by official statistics, which classify the origin of the patent according to the applicants' identity, not the inventors'. It is only when recent studies have moved to re-classifying patents by inventor, and matched inventors' names to the names of university scientists, that the reality of academic patenting in Europe has emerged (review by Lissoni, 2012; more up-to-date information in the other articles of this issue). In all the countries considered by the literature, a significant percentage (from 3% to 8%) of corporate patents has been found to cover inventions by academic scientists. However, universities are the least important category of assignees of these same academic patents, with shares around 10% in most countries. Everywhere, they are superseded by business companies, whose shares, in the 1990s and early 2000s, ranged from 61% (in France) to 80% (in Sweden). The highest shares of university-owned academic patents were found in the Netherlands (26%) and in the United Kingdom (22%), whose universities enjoy the highest degree of autonomy in Europe (Estermann et al., 2009). In the US, which host a large number of private universities and whose public universities are not under federal control, the university share of academic patents can be estimated at 70%, and suggestions have been made that only inventions from consultancy escape the administrations’ control (Thursby et al., 2009). From this discussion, we can formulate a simple research question:

Q1 – Does the share of university ownership of academic patents increase with university autonomy?

The ownership distribution of IP may have consequences both for the universities’ finances, and their incentives to promote technology transfer activities. A detailed discussion of the first issue goes beyond the scope of this paper. We limit ourselves to point out that patent ownership may affect a university’s finances in two ways: a direct one, through the net results of IP management costs and IP licensing and trading revenues; and an indirect one, to the extent that exhibiting a strong patent portfolio improves either the evaluation of the university by funding agencies or its general reputation for technological excellence. The direct effect appears to benefit only a handful of large universities in the US (Bulut and Moschini, 2009; MacDonald, 2011), while the second one is highly disputed (Leydesdorff and Meyer, 2010; Thursby and Thursby, 2011) .

As for the effects of ownership on incentives, we are mostly concerned with those leading to the introduction of inventions (for their development and commercialization, see Dechenaux et al., 2011; and Darmsgaard and Thursby, 2012). Our research question can be phrased as follows:

Q2 - Does autonomy affect positively universities’ contribution to inventive activity in their country?

The same two questions can be reformulated with reference to the introduction of the professor privilege. The expected profits may either push individual scientists to file a patent on their inventions, but also (and with opposite effects) shield them from the pressure for commercialization eventually exerted by the university administration. Both scholars and legislators, however, have placed emphasis on the first of these two effects-

Evidence related to our research questions is still in its infancy, and too US-centric for lending itself to a generalization. As for Q2, Aghion et al. (2010), Bonaccorsi and Daraio (2007), and Estermann et al. (2009) suggest that autonomy is associated to higher efficiency and productivity. However, these studies focus on universities’ general tasks (teaching, research, and technology transfer), and make no use of specific datasets on patents. Concerning Q1, some evidence exist on the effects of the abolition of professor privilege, which is however limited by lack of data.

The latter is explained by the persistence, until recently, of a methodological problem, best described as a trade-off between accuracy, scope, and longitudinal depth. Identifying academic patents on the basis of inventors requires matching the inventors’ names to the names of academic scientists. While the information on inventors come with patent data, for which very long time series are available, information on academic scientists is generally obtained through ad hoc requests to universities or governmental institutions, which almost always are satisfied for one year only. Most existing studies are therefore limited to either cross-sectional evidence for a single country (Meyer, 2003; Saragossi and van Pottelsberghe, 2003; Balconi et al., 2004), cross-country evidence (Lissoni et al., 2008) or university case studies. In one particular case, that of Germany, researchers have made use of a surrogate indicator of the academic provenance of the patent, such as the presence of the academic title of “Professor” in the inventor’s name, which allows to build time series (Schmiemann and Durvy, 2003; Gering and Schmoch in OECD, 2003; Czarnitzki et al., 2007; 2009a,b; 2011a,b; 2012) . However, Von Proff et al. (2012), who follow the same approach, point out its many limitations. A more innovative alternative has been recently proposed by Dornbusch et al. (2013), who verify the academic affiliations of inventors by means of bibliometric sources. In this case, we miss all patents by academic scientists whose publications cannot be retrieved (possibly due to deficiencies in bibliometric sources for remote years). Our own solution to this problem is illustrated in section 4.