IPR and Innovation

1
IPR and Innovation

• Ashish Arora
• Heinz School,
• Carnegie Mellon University

2
IPR and Innovation

• Socially efficient level of RD (conditional on
  existing firms)
• Efficient allocation of RD input
• Commercialization
• Cumulative innovation and dynamics
• Institutions of knowledge generation

3
Socially efficient level of RD

• Incentives to innovate what is the role of
  patents
• Patents less important than other first mover
  advantage and commercialization capabilities
  (e.g., Yale, CMU surveys)
• Median patent of low value patent renewal
  studies (e.g., Pakes, Schankerman) 25 ESR
• But average value of patents may be significantly
  higher (e.g., Gambardella et al.)

4
Patent Premium, selected industries
Average Conditional
Drugs and medicines 0.9 1.65
Biotech 0.97 1.66
Computers and other office equipment 0.7 1.58
Semiconductors 0.4 1.50
Medical instruments 1.12 1.72
Total 0.49 1.53

• Patents not used for a majority of innovations in
  the vast majority of industries
• Patent protection is valuable in all industries,
  even in industries that do not patent much.

Source Arora, Ceccagnoli and Cohen, RD and the
Patent Premium, NBER, 2004
5
Patents provides significant incentives for
RDElasticity w.r.t patent premium
RD Patent Applications Patent apps/ RD
Drugs and medicines 0.96 1.4 0.45
Biotech 1.02 1.46 0.44
Computers and other office equip 0.74 1.29 0.55
Semiconductors 0.46 1.07 0.61
Medical instruments 1.18 1.48 0.30
Total 0.57 1.16 0.59
Source Arora, Ceccagnoli and Cohen, RD and the
Patent Premium, NBER, 2004
6
Efficient allocation of RD effort

• Entry mixed evidence
• Interaction between IPR and market power retard
  entry (e.g., MSFT)
• French org. chemical industry destroyed by
  excessively broad patents.
• Patents used to coordinate cartels in chemicals
  in pre WWII
• Patents by non producers promotes entry in
  chemicals (Lieberman)
• Patents promote vertical specialization
• Pharma firms increasingly focusing on marketing
  rather than RD
• Patents (and IPR broadly) may also be used to
  prevent vertical entry (e.g., software)

7
Commercialization Patent Protection and
Licensing Mixed Evidence

• No Effect?
• Other studies cast doubts on the link between
  patent strength and the extent or form of
  international technology licensing
• Fink (WP, 1997), using German data
• Fosfuri (WP, 2003) in the chemical sector
• Cassiman and Veugelers (2002) do not find that
  the patent strength encourages Belgian firms to
  enter into collaborative RD arrangements.
• Branstetter, Fisman, and Foley (NBER WP, 2002),
  using US data, find no evidence of an increase of
  technology licensing to unaffiliated parties for
  countries that strengthened patent protection.
  (But they see an increase in transfer to
  affiliate.)

• Find positive effect
• Gans, Hsu, and Stern (RJE,2002) presence of
  patent increases the likelihood that startup
  licenses to an incumbent rather than enter
  commercializing
• Yang and Maskus (W.P., 2000) positive
  relationship between IPRs regimes and licensing
  by U.S. MNCs.
• Nagaoka (WP, 2002) using data for Japanese
  firms, find weak patent regimes associated with
  an greater fraction of transfers to an affiliate
  rather than unaffiliated firm
• Smith (JIE, 2001) US firms more likely to export
  or directly manufacture rather than license
  technology in countries with weak patent regimes.

Impact of patents on licensing conditioned by
presence of complementary capabilities
8
10 Change in Patent Effectiveness Leads to
Complementary Assets Low High Complementary Assets Low High
change in licensing propensity change in the propensity to license patented innovations 6 2
change in licensing propensity change in the propensity to license patented innovations 1 -3
Source Arora Ceccagnoli Patenting and
Licensing, Management Science, forthcoming
9
Dynamics

• Cumulative innovation (e.g., Scotchmer, Green and
  Scotchmer, Merges and Nelson)
• Biotech produces anecdotal evidence of problems
• BRCA Myriad
• Stem Cells and Geron
• Murray Stern Sampat
• Anticommons (Heller and Eisenberg)
• Empirical basis weak (e.g., Walsh et al, 2004 for
  commercial research and Walsh et al. 2005 for
  large sample academic research)
• Problem may arise in future with full genome
  screening and gene chips

10
Institutions Biggest problem area

• Major threat to non market based institutions
  (e.g., David Gambardella and Hall)
• Many (all?) of biomed problem patent in US are
  university / non profit based patents
• BRCA Univ of Utah
• Stem Cells Wisconsin
• Canavan Miami
• Oncomouse Harvard