Knowledge Disclosure, Patents and Optimal Organization of Research and Development
Discussion: Alan MorrisonSaïd Business School, University of Oxford and CEPR
Sudipto BhattacharyaSergei Guriev
Motivation
• The agents who create ideas are frequently independent of the agents who develop them
• It is impossible to sell knowledge without leaking it
• What are the implications of this statement for – The best way to sell information– The best way to structure the research and development
units
Model
RU: effort eEx ante
Knowledge KDistn Fn G(K,e)
Interim Ex postDU1, DU2: effort E 0.5
P{Invention} = P(K,E)= KE2 Bertrand Compete: Payout 1 iff soleinventorInterim Sale of K
Open sales: Patent K to one party Knowledge LK disclosed to both parties
Closed sales: For lump sum + shares s Bilateral Rubinstein-style bargaining Outside option for both parties is open sale 1. Disclose the nature of K and let LK go at the same time 2. Bargain for K … Then decide whether to sell to other DU. This is unobservable: use s to render it incentive incompatible
Incentive Effects at Interim Date
Open Sales
DU get reservation utility increasing utilityMore ex post competition lower devt. effort from licenseeInterim E{surplus} dropping
Closed SalesOpportunistic sales less attractive lower s, higher devt. effortInterim E{surplus} increasing
STRONG IPRPROTECTION
L = leakage0 1WEAK IPR
PROTECTION
K=
kn
ow
led
ge
0
1Open Sales
Other DU more likely to Succeed less effortInterim E{surplus} dropping
Closed SalesOpportunistic sale less attractive lower s higher effortRU royalties droppingInterim E{surplus} increasing
Low K: closed sales incentive infeasible
Closed sa
les incre
asingly
attracti
ve
OpenClosed
Overall R&D Expenditures
STRONG IPRPROTECTION
L = leakage0 1WEAK IPR
PROTECTION
K=
kn
ow
led
ge
0
1Open
Closed
Open Sales Lower development effort from licenseeHigher development effort from non-licensee
Closed Sales Higher development effort
Lemma 4: Total either increasing or U-shaped
• In open region, L4 implies expenditure could be increasing
• Moving from O to C causes a drop in expenditure: could cancel closed sales effect
• For appropriate distn on K this happens … Inverted U shape
Ex Ante Structuring
• First best may violate RU budget constraint in closed sale case
• Could bring in a VC partner– VC buys some of the shares in the development
project– Further disclosure by RU IC, but harms VC/RU
coalition: hence a monitoring role for the VC
• But RU royalties dropping in K with closed sales– Deleterious incentive effect
• Could overcome with corporate venture which prevents the introduction of a VC
• For some K distns, ex ante incentive gains from venture outweight interim efficiency costs
Comments
• This is a rich and subtle model• Some of its many predictions are rather
surprising• While the paper is not a simple read, it repays
perseverance and I enjoyed working through it• I had some general questions about the set-up
and its interpretations
Questions
1. What in this model can be influenced by policy, and what is technological?• In the closed sale, L appears to describe
technological limitations to the disclosure process• In the open sale, this may be true, but may similarly
reflect the amount of information it is acceptable to steal
• This has an impact for the interpretation of the inverted U: what appears on the x axis may sometimes be hard to describe as “IPR protection”
Questions
2. The use by DUj of leaked information to invent when DUi has a patent might be regarded as a breach of patent• At present breach is costly to DUi because it
generates ex post competition• Its interesting in this case to think about whether
there should be a role for the courts: strength of IPR then has something to do with ex post enforcement
Questions
3. Is this really what the VC does? • In general, we expect the VC to have something to
do with selecting managers and monitoring them while they are actually working
4. I like the fact that the model can explain the inverted U-shape in terms of modes of sale. It would be interesting to read some speculation as to the effects of all of this upon consumers, so that a more complete picture of welfare could emerge, and with it some policy suggestions