Innovation and "Professor's Privilege"

Transcription

1 Innovation and "Professor's Privilege" Andrew A. Toole US Patent and Trademark Office ZEW, Mannheim, Germany NNF Workshop: The Economic Impact of Public Research: Measurement and Mechanisms Copenhagen, Denmark April 28, 2017 The views presented are those of the author and do not necessarily represent the position of USPTO..

2 Background Policy makers are increasingly interested in stimulating entrepreneurial behaviors among academic researchers Their focus: change the incentives facing academic researchers so that entrepreneurial choices are more attractive Numerous policy levers are available including: tax policies, employment policies, subsidies, education, and intellectual property (IP) policies.

3 Intellectual Property (IP) Policy Key question: Who should own the patent rights to university-discovered inventions? U.S. Bayh-Dole Act of 1980 suggests university ownership Granted ownership rights to patentable inventions discovered using Federal funds (from the government) to universities, not individual inventors. University ownership (or institutional ownership) is the currently accepted "model" for how to increase university invention and technology transfer: Denmark (1999), Germany (2002), Norway (2003) and other countries have implemented similar policies.

4 But is university-ownership the correct model? How do the incentives for faculty to patent and engage in academic entrepreneurship change when moving to a university-ownership model? Minimal evidence is available in the economics and policy literatures supporting a university-ownership model over other alternatives.

5 Three Recent Studies Czarnitzki et al. (2015), "Individual versus Institutional Ownership of University-discovered Inventions" Czarnitzki et al. (2016), "Knowledge Creates Markets: The Influence of Entrepreneurial Support and Patent Rights on Academic Entrepreneurship" Hvide and Jones (2016), "University Innovation and the Professor's Privilege"

6 Individual versus Institutional Ownership of University-discovered Inventions Authors: Dirk Czarnitzki Thorsten Doherr Katrin Hussinger Paula Schliessler Andrew A. Toole

7 Contributions of Czarnitzki et al. (2015) Uses a natural experiment based on a 2002 law change in Germany from inventor-ownership to universityownership. The German Federal law change was exogenous to individual scientists The institutional structure of the German research system creates a natural treatment group and control group for evaluating the legal change Use scientist-level data and difference-in-difference methods are used to evaluate changes in the quantity and quality of patents due to the law change.

8 Basics of the German Law Change Until 2002, university researchers owned the rights to their inventions, called professor privilege This only applied to university professors. Other academic researchers in public research organizations (like Max Planck) are not affected. These institutions already followed the Bayh-Dole ownership model In 2002, the German Federal Government abolished professor privilege From 2002 onwards university researchers had to disclose their inventions to the university. If the university decides not to claim the invention, the IP right is transferred back to the researcher. The university pays all expenses related to patent process and will search for potential licensees. (Peckerman 2007) If the invention is licensed by the university, the researcher (inventor) will receive a 30% royalty on gross revenues.

9 Conceptual Framework I Patent indicator function: patents it = propensity to patent it * new knowledge it Propensity to patent may change due to legal/economic changes affecting the expected cost and benefits of patenting. Increment in knowledge is usually referred to as "knowledge production function". We hold increments to knowledge constant, and focus on how patents respond to changes in the propensity to patent because of the law change.

10 Conceptual Framework II Law change transferred propensity to patent from inventor to university "regime shift" Effect of revoking the prof privilege on patenting will depend on how benefit and cost schedules shift. The benefit schedule for the university after law change is lower than for inventor before the law change -- because of three way bargaining among univ, prof and licensee company (Hellmann 2007) Thus is cost schedules were the same, law change would lead to reduction in patents!

11 E(mb) E(mc) Sci: E(mc) P' P* Sci: E(mb) Univ: E(mb) Quantity of patents

12 E(mb) E(mc) Sci: E(mc) Univ: E(mc) Sci: E(mb) P' P* P" Univ: E(mb) Quantity of patents

13 Conceptual Framework III Two groups: Low cost inventors: had prior relationships with industry partners. All search costs were already incurred For these faculty, the regime shift led to higher cost schedule, as university had to renegotiate established relationships High cost: no prior industry connections.

14 Hypotheses H1: Faculty members with established connections with industry partners experienced a decrease in the volume of patenting (all else constant) H2: Faculty members who did not have established connections with industry partners experienced an increase in the volume of patenting (all else constant) Net effect of policy depends on shares of high cost and low cost faculty in the population before the law change.

15 Final sample Panel database of researchers over time Excludes people that switch between (or are employed at both) treatment and control group (1,800 researchers) Keep only researchers that were observed before law change Individuals entering after law change do not contribute to estimation of treatment effect. Panel: ,718 professors (treatment group) 8,294 PRO researchers (control group)

16 Patent Trends Diverge after 2002

17 Prior to law change ( ) After law change ( ) N = researcher-years Professors with industry connection N = 9141 researcher-years Variable Mean Std. Dev. Min Max Mean Std. Dev. Min Max # Patents # Citation-weighted patents Career age Avg. publications N = researcher-years Control group with industry connection N = 9854 researcher-years # Patents # Citation-weighted patents Career age Avg. publications

18 Prior to law change ( ) After law change ( ) N = 6633 researcher-years Professors without industry connection N = 8121 researcher-years Variable Mean Std. Dev. Min Max Mean Std. Dev. Min Max # Patents # Citation-weighted patents Career age Avg. publications N = researcher-years Control group without industry connection N = researcher-years # Patents # Citation-weighted patents Career age Avg. publications

19 Overall Down 16.8% Connected Down 23% Unconnected Up 32% Poisson DiD: Patenting Volume # Patents Overall With industry connection Without industry connection Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. Professor*NewPolicy *** (0.053) *** (0.067) 0.391*** (0.085) Career age ** (0.014) (0.019) *** (0.020) Career age squared/ (0.028) * (0.038) 0.721*** (0.065) Avg publications 0.028*** (0.005) 0.017*** (0.005) 0.045*** (0.007) Time dummies (base 1995) *** (0.033) 0.160*** (0.039) (0.062) *** (0.052) 0.304*** (0.066) (0.086) ** (0.075) 0.307*** (0.098) (0.113) (0.097) (0.129) (0.144) (0.118) (0.156) (0.175) (0.139) (0.186) (0.203) (0.157) (0.210) * (0.228) # obs. 108,263 44,604 63,659 # obs. PRO researchers 71,860 22,955 48,905 # obs. professors 36,403 21,649 14,754 # obs. Professors after policy change 17,262 9,141 8,121 Robust standard errors. Significance: * p < 0.1, ** p < 0.05, *** p < 0.01.

20 Robustness 1. Conditional difference-in-difference 2. Use forward citation-weighted patent counts 3. Account for the announcement period, starting in Placebo test use those academics who are jointed affiliated with a university and PRO. No treatment effect. 5. Induced Entry

21 Steady Decline after 1999

22 Conclusions The transition from an inventor-ownership to a university-ownership model: Decreased both the volume and quality (as measured by citations) of patented inventions by university professors Policy makers presumed that costs and risks of patenting were too high for individual inventors: Appears to be wrong: The evidence supports an inventorownership model. Our work highlights the importance of the nature and strength of faculty-industry relationships Must understand these before undertaking policy initiatives intended to foster technology transfer.

23 Results from Czarnitzki et al. (2016) The reform included new Patent Valorization Agencies and support for university TTOs, but these did not produce an increase in the number of university researcher start-ups The reform changed the distribution of ownership for university inventions: The share owned by firms and by faculty members decreased The share owned by universities increased While patents and citation weighted patents strongly stimulate faculty start-ups, the reform did not strengthen this relationship it stayed the same

24 Results from Hvide and Jones (2016) The abolishment of Professor's Privilege in Norway led to about a 50% drop in the rate of start-ups by university researchers University researchers curtailed their patenting by about 50% after the reform Not driven by exiting researchers, but "stayers" patent much less They found a decrease in the quality of start-ups and patents (slower growth and fewer citations)

25 Remaining Challenges Need linked data to TTO metrics Important outcomes such as licensing and other contractual relationships are still missing from the story. Need to understand whether the effect is due to (1) the loss of control rights; or (2) the loss of income If income, they simply raise the royalty rate to offset the negative effects Need to understand the trade-offs at the individual level How does losing patent rights influence the balance of researcher activities? More teaching, more leisure, more research, more consulting,

26 Thank you

27 Conditional DiD: Patenting Volume # Patents Overall With industry connection Without industry connection Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. Professor*New Policy -0.19** ** *** 0.13 Time dummies (base ) ** *** ** *** *** *** *** *** *** *** *** *** ** 0.14 Observations Robust standard errors. Significance: * p < 0.1, ** p < 0.05, *** p < 0.01.