From science to license: an exploratory analysis of the value

of academic patents

E. SAPSALIS*1 , B. van POTTELSBERGHE*²

2nd ExTra/DIME workshopEPFL, Lausanne, 29-30 Sept 2006

* ULB, 1 FNRS Research Fellow, ² Chief Economist at European Patent Office

This paper was partly written when E. Sapsalis was appointed Adj. Associate Research Scholar at Columbia Univ. (NY-USA). We thank the FNRS, the ULB and the foundation MC Adam for supporting financially this research stay

The views expressed in this article are purely those of the authors and may not in any circumstances be regarded as stating an official position of the EPO or of the ULB.

Contact: esapsali@ulb.ac.be

2

Objective

Analyse the value determinants of the technological, industrial, entrepreneurial impact of academic patents

3

Content

Academic patenting revolution Data Empirical Models Results Conclusions

4

Academic Revolution

5

The academic revolution

Academic Patenting at the EPO

0

500

1000

1500

2000

2500

3000

0,0%

0,5%

1,0%

1,5%

2,0%

2,5%

# EPO applications % EPO applications

Source: European Patent Office: own calculations

6

Academic Patenting Revolution

IP-oriented emerging technologies ; Biotech, Nanotech,…

Patent-oriented laws USA/European countries: Dayh-Dole Act like

legislations

More active role that academia has been asked to play in the Knowledge Economy

7

Pending questions

Management of Technology transfer Spin-offs; licenses; research contracts/ alliances

etc…

Debates related to the roles of university

Dissemination of science The balance between the different missions of

universities Quality of research; quality of patents

8

Patent Value

9

State of the art (1) The distribution of patents’ value is highly

skewed

Value Proxies Monetary value of a patent (e.g. Harhoff et al., 1999, 2002) Present value evaluated by experts (e.g. Reitzig, 2003) Forward citations (e.g. Lerner, 1994; Sapsalis and van Pottelsberghe,

2006, 2007) Composite indicator (e.g. Lanjouw and Schankerman, 1999) Creation of a start-up (e.g. Shane, 2001) Probability to get a patent …

Granted (e.g. Guellec and van Pottelsberghe, 2000) Opposed (e.g. Lanjouw and Schankerman, 1997) Renewed (e.g. Lanjouw and Schankerman, 1999) Licensed (e.g. Dechenaux et al., 2003 )

10

State of the art (2) Value determinants

Forward citations ++++ (ex: Shane, 2001) Patent family size ++ (ex: Lanjouw and Schankerman,

1999) Results of opposition & annulment procedures : + (ex: Harhoff et al, 2003) Backward patent citations: + (ex: Harhoff and Reitzig, 2000) Non-patent citations: (+) (ex: Harhoff and Reitzig, 2000) Claims: (+) (ex: Lanjouw and Schankerman,

1997) Patent scope: (+/-) (ex: Lerner, 1994) Time: + (ex: Guellec and van

Pottelsberghe, 2002) Technical field: (*) (ex: Harhoff et al, 1999) Ownership characteristics: (*) (ex: Guellec and van

Pottelsberghe, 2002)

11

Value determinants

12

Data sets

13

Data Data Source: DELPHION database

Patents’ priority date: 1985-2003 Assignees : 6 Belgian universities

UCL, ULB, Ulg, KUL, UG, VUB

EPO patents grouped in patent families 364 EPO patent families

334 EPO patent families with available exploitation data 142 licensed patents

53 licensed to spin-offs 89 licensed to established companies

14

Belgian academic patents

Academic patents and licenses (EPO priority date: 1985-2003)

0

10

20

30

40

50

60

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

No licensee Established Company Spin-off

15

Models

16

Empirical implementation (1) Negative Binomial

Model Number of patent

citations

Probit/Logit Model License

agreements With established

companies With spin-offs

17

Empirical implementation (2)

Dependent variables: Technological impact : # FPC Industrial impact: dummy variable

standing for a license given to industry

Entrepreneurial impact: dummy variable standing for a licence given to a spin-off

18

Empirical implementation (3)

Independent variables: Technological impact:

# FPC & origin Science knowledge :

# NPC & origin Technological knowledge:

# BPC & origin Ownership:

# Inv; # Ass & type IP protection:

Time; Patent Family; Claims; Scope

19

Setting hypotheses (1)

Technological impact Industry License = f+(FPC-self; FPC-PRI ;

FPC-Co) Startup License = f+(FPC-PRI ; FPC-Co)

Scientific base # forward patent citations = f+(NPC-self); f -

(NPC-non self) Industry License = f+(NPC-self); f -(NPC-non

self) Startup License = f+(NPC-self); f -(NPC-non

self)

20

Setting hypotheses (2)

Technological Base # forward patent citations = f+(BPC-PRI) ; f -

(BPC-Self) Industry License = f+(BPC-self; BPC-PRI ; BPC-

Co) Startup License = f+(BPC-PRI ; BPC-Co)

Ownership # forward patent citations = f+(Ass-PRI); Industry License = f+(Ass-Ind); Startup License = f+(NPC-self); f -(NPC-non

self)

21

Setting hypotheses (3)

IP Protection # forward patent citations = f+(Fam; Claim); Industry License = f+(Fam; Claim); f -(Scope) Startup License = f+(Fam; Claim; Scope);

22

Econometric Results

23

Impacts Technological, industrial and

entrepreneurial

24Obs 334 334 301

Log Lik. -453 -139 -120

25

LicensingEstablished Companies vs. Spin-offs

26129 Obs ; 13 Time dummies ; Log Lik. :-54

27

Concluding remarks (1)

Policy implications Focus on high level scientists

importance of tacit knowledge

Importance of collaboration Academic collaboration: + tech impact ; -

ind impact Industrial collaboration: + ind impact

importance of close contact with industry

28

Concluding remarks (2) Management implications

TTO and Funding bodies Management of IP co-developed by different

public research institutions Spin-offs and established companies exploit

different types of patents

Further research Analyse commercial impact Analyse the impact of academic inventions

on industrial portfolio