38CO2000 Economics of Intellectual Property Rights IPRs Spring 2006: Lecture 6

1. 38CO2000Economics of Intellectual Property Rights (IPRs) Spring 2006: Lecture 6

2. Suzanne Scotchmer 09/14/2004. Subject to Creative Commons NC-SA License Recap Horizontal competition: The consumer cost of raising money through monopoly pricing is DWL Should breadth cause price to be lower, and the IP right to last longer? The ratio test: the optimal patent policy maximizes the ratio of ex post profit to ex post social welfare

9. IPs improve the functioning of markets for technology They determine the terms of the use, e.g., bargaining position in licensing negotiations Problems caused by IPRs Used to block others, fragmentation, hold-up problems Is stronger IP good for R&D incentives?

10. The effects of patents on cumulative innovation: the case of basic technology and its application basic research with little commercial value can be essential for creating the scope for commercial applications Assume deterministic innovation of basic technology and its commercial application the costs of creating the basic technology cB > 0 the market value of the technology in itself is zero if the technology is made, it can be protected by a patent a firm other than the patent holder has an idea of how to make a commercial application of the basic technology the cost of making commercial application is cA > 0

11. the private value of the application is P(T)=?T whereT is the (discounted) patent life, P� > 0 the social value of the application is S(T)=W/r-TDWL S�<0 the social value of the basic technology is at least S(T)-cA in general the first innovator has too little incentives to invest - payoff zero, you have to pay cB>0 The problem of cumulativeness: how to transfer surplus from the second innovator to the first innovator? if the application infringes the patent covering the basic technology, the second innovator forced to acquire a license if no infringement, no way to transfer the profits! If no transfers of profit, no investment in the first innovation ? no investment in the second innovation ? IP creates the market for technology

12. Assume potential infringement Can the firms be certain about infringement ? Consider licensing negotiations between the patent holder (the first innovator) and the innovator/producer of the commercial application (the second innovator) Are negotiations made before or after the application is made? Which is more realistic? Why? Consider first ex post licensing, i.e., negotiations occur only after the application is ready for production (cA has been sunk)

13. the available cake is P(T), the first and second innovator should find a way to divide it b = the share of the first innovator 1-b = the share of the second b reflects the bargaining power of the first innovator if no good reasons to assume otherwise, set b=1/2 (Nash-bargaining solution, solution for Rubinstein alternating offer bargaining) ? the payoff of first innovator ??B =bP(T) - cB the payoff second innovator: ??A=(1-b)P(T) � cA

14. In practice b affected by patent quality and patent strength assume the first innovator has full bargaining power but the patent is of imperfect quality b = strength of forward protection = probability that both the patent validity and the infringement holds in the court ? ??B =bP(T) � cB & ? ??A =(1-b)P(T) � cA as before Note: this abstracts from costs of litigation. These costs are huge in practice. 1-3 million USD 50.000-500.000 EUR ? why infringement disputes ever reach courts?

16. The basic tradeoff of the cumulative innovation: increasing b increases the incentive to create the basic technology but decreases the incentive to create the application it is possible that there is no incentive to make the commercial application even if P(T)>cA , ? there is no incentive to make the basic technology! increasing T could be a solution: both ??B and ??A are increasing in T If T ?max {cB/b?, cA/(1-b)?) then both innovations are made But the basic tradeoff of the horizontal competition appears (recall S(T), S�<0)

17. This is a manifestation of a hold-up problem: the second innovator realizes that she will be held-up in the negotiations over the license the problem emerges from contract incompleteness and relation-specific investment these concepts underlie the modern theory of a firm (cf. Williamson, Hart) Contract incompleteness: impossible to write a verifiable contract on the investment to develop the application because of transaction costs it can be hard to identify the second innovator ex ante the investment is likely to be complicated and hard to measure Relation-specific investment makes the investment irreversible (little value outside the relationship) when the cake is divided, the investment is sunk unless the second innovator gets a license, nobody willing to buy the firm/technology

18. Consider next ex ante licensing: negotiations over the license can be conducted before the commercial application is made (cA is not sunk) ? There is no hold-up problem! the first innovator has an incentive to secure that the commercialization is made, i.e., that (1-b)P(T) � cA?0 the first innovator requires at most bmax??1- cA/ P(T) even if b>bmax the commercialization will be made, if it is profitable, even if the first innovator has full bargaining power & perfect forward protection! ? the available cake is P(T)-cA the payoff of first innovator: ??(P(T)-cA) - cB the payoff second innovator: (1-?)(P(T) � cA)

19. forward protection (b) increases ? only in so far b<bmax , i.e., ?(b), ?�>0 if b??[0, bmax] forward protection cannot be used to secure the incentive to make the basic technology if ??(bmax)(P(T)-cA) ?cB the patent term works, i.e., set T?Tmin?such that ?(P(Tmin)-cA) �cB=0 ? Tmin?=(cA+cB/?) /? but the basic trade-off of the static model looms� Notes 1) Optimal patent life solves P(T*) - cA - cB = 0 ? T*=(cA+cB)/? a longer patent life would unnecessarily prolong monopoly distortions a shorter patent life would not create incentives to innovate 2) T*<Tmin

20. if the innovators collude or if innovation is concentrated in the same firm, the patent term can be set at the optimal level The profit of a merged firm: P(T)-cA-cB ?T =(cA+cB)/? guarantees the incentive to innovate competition policy in �Schumpeterian� industries is complicated issue! More generally, when hold-up problems are severe, vertical integration works ? a reason why we have firms (cf. Holmstr�m and Milgrom, JEP-98, Hart-95) 4) IP reduces the hold-up fear of the first innovator, because the infringement forces licensing, creating the market for technology Suppose that there is no infringement. The first innovator fears the hold-up (after she has invested and created the basic technology, no body is willing to pay for it) and does not invest, Collusion or vertical integration would be the only way to induce the investment in the basic technology

21. With ex ante licensing, the second innovation will be made if in so far P(T)>cA. ? the hold up problem concerning the first innovation can remain even with IP ? the key problem is to compensate early innovators ? how to increase their bargaining power? 6) The hold-up problem also remains concerning the second innovation if ex ante licensing is not feasible Why ex ante licensing can be infeasible? management of IP when innovation is cumulative If ex ante licensing is not feasible, hold-up problem could be solved e.g, via reputation, long relationships, strategic alliances, reciprocity, hostages� ? management of IP when innovation is cumulative

22. The effects of patents on cumulative innovation: the case of improvements (quality ladder) Suppose the first innovation has commercial value in itself and the application is an improvement . The quality of the first innovation is q1 and the quality of the application is q2 assume unit mass of consumers with a unit demand (buyers buy only if q?p) & Bertrand competition if both products in the market, both the price and the profits on the application equal ?1=q2-q1. As to the first innovation, both of them are zero if the first innovation is in the market alone, its price and profits are q1 Add a third innovator with quality q3

24. Consider the second innovator: If all inventions infringe, she is both a licensee of the first innovation and a licensor of her own innovation for the innovator of the third innovation It is well possible that stronger patens increase the gains as a licensor less than losses as a licensee ?2<?1 ? stronger patents are not necessarily good for incentives to innovate complementary innovation (Bessen&Maskin, Hunt)