Length, breadth, and height of the “optimal patent” 26

There are numerous studies which have focused on the optimal design of a patent system that maximizes social welfare by optimally balancing negative and positive effects. Earlier studies often focused solely on patent life (duration of protection) while later studies in- cluded considerations about the optimal patent breadth (scope of protection) as well. A minority of studies investigate the height of patent protection (i.e. novelty requirements for protection).

One of the earliest studies of this kind was Nordhaus’ work on the optimal patent life (Nordhaus, 1967). The study focuses on “run-of-the-mill” inventions under perfect competi- tion, which are cost reducing inventions,35 and which leave the pre-invention price and quantity unaltered36. Furthermore Nordhaus assumed that during the patent term the pat- entee gains the entire monopoly profit while immediately after expiration of the patent term competition eliminates any rents (Scherer, 1972, p. 423). Maximizing the benefits to pro- ducers and consumers subject to any given patent term, Nordhaus obtained the welfare- maximizing patent term, which depends on the price elasticity of demand, the “importance of the invention”, the discount rate, and on the elasticity of cost reduction with respect to R&D investment. Furthermore he found that the optimal patent life is shorter for more im- portant inventions37, since they have a greater social impact and thus produce a greater deadweight loss.

35 _{His findings can be applied to the more general case of product inventions as well (Kaufer 1989: p. 25).}

36 _{This is due to the fact that the cost reduction is not big enough to induce a decrease in prices and an associated output} increase (Scherer 1972: p. 23).

37 _{This is quite contradictory to the conventional argument, which would award longer patent lives to more important} inventions since they are arguably riskier and probably require higher research investments. However, this surprising result of Nordhaus’ model only holds for run-of-the-mill inventions and not in the case of drastic inventions (Nordhaus 1967: p. 9).

Gilbert and Shapiro (1990) and Klemperer (1990) expanded the analysis of the optimal patent system by including patent scope in their models. Patent scope corresponds to “the size of the region of technology space from which a patentee may exclude others from operating” (Jaffe 1999, p. 23). In the case of an “independent single invention” the analysis of patent scope does not differ a great deal from the analysis of patent length. The broader the patent scope the higher the incentives for research and the higher the welfare loss from market power. In the case of cumulative research however, there are further aspects that have to be considered. On the one hand a broader patent promises higher returns to in- vestment and thus stronger research incentives. On the other hand, an inventive firm an- ticipates that its potential invention might infringe on an existing patent and thus research becomes riskier and less valuable. Consequently broader patents can not only spur but also deter the rate of invention (Jaffe 1999, p. 23).

Gilbert and Shapiro (1990) and Klemperer (1990) both focus in their work on the optimal combination of patent length and breadth. While Gilbert and Shapiro (1990) come to the conclusion that a longer patent life and a narrow breadth are optimal as long as broader patents are associated with increasing costs to society through deadweight losses for any given size of reward for the inventive firm, Klemperer (1990) identifies a case where ex- actly the opposite is the case. Klemperer identifies two sources of social costs: Firstly, con- sumers who substitute to another product class, and secondly consumers who substitute within the product class to less preferred products. A wider patent increases the first form of distortion by permitting higher prices, which leads consumers to switch to other product classes and thus increases deadweight loss, but a wider patent also reduces the possibili- ties for consumers to switch to unpatented, competitively provided varieties of the product thus reducing the second form of distortion. Thus he concludes that “..., infinitely lived, nar- row patents are typically desirable when substitution costs between varieties of the product are similar across consumers, but very short-lived, wide patents are desirable when valua- tions of the preferred variety relative to not buying the product at all are similar across con- sumers” (Klemperer 1990, p. 127). In short, Klemperer’s results indicate how the optimal combination of patent length and breadth varies with product classes.

While both Gilbert and Shapiro (1990) and Klemperer (1990) advocate the idea that the optimal patent system should include some kind of trade-off between patent length and breadth,38 Gallini (1992) produces somewhat contrasting results. This is attributable to the fact that Gallini introduces imitation costs to the model. While Gilbert and Shapiro (1990) and Klemperer (1990) both assume that the imitation cost is zero, Gallini (1992) by intro- ducing imitation costs endogenized the imitation decision. Since longer patent terms in- crease the incentives for imitation Gallini finds that if only patent length is considered, it should be short in order to discourage imitation. In the case where both patent length and

38 _{Gilbert and Shapiro (1990) find the optimal patent to be narrow and long, while Klemperer (1990) finds the optimal} patent to be either narrow and long or broad and short depending on elasticity of demand.

breadth are considered however, Gallini (1992) finds that the scope of protection should be broad while the length can be adjusted in order to reach the requested level of patent re- wards (pp. 52-53).

All three papers are however subject to limitations. In particular, their focus on a single independent invention implies that the potential negative effects of a broad patent on the rate of invention are not considered. In the case of cumulative research the question of the best way to distribute the rewards for inventing between the initial inventor and the follow- on inventor arises. An invention can be stimulated by an earlier invention in three different ways. The initial invention can spur the development of the second invention, in which case society profits by getting the second invention sooner. Alternatively the initial invention can reduce the research costs of the second invention in which case society profits by getting the second invention cheaper. The initial invention may also make inventions possible that could not have been developed otherwise. These externalities are not considered by the initial inventor’s research decisions unless they can appropriate some of the returns of the follow-on inventions however. The second inventor also needs to be rewarded in order to induce research investment in the follow-on invention (Scotchmer, 1991, p. 31).39

Recent studies by Green and Scotchmer (1995), O’Donoghue (1998) and O’Donoghue et al. (1998) are mostly consistent with Kitch’s40 conclusion that strong patent protection should be awarded to the initial inventor in the case of cumulative research. Scotchmer (1996) even claims that patent protection for follow-on inventions is not necessary at all to induce their development if there is no impediment to ex-ante licensing, since this maxi- mizes the reward for the initial inventor. Several researchers have raised doubts about this support for broad patent protection however. Merges and Nelson (1990) as well as Maz- zoleni and Nelson (1998) expressed their concerns about a broader patent scope with re- gard to the danger of blocking up the entire research field. Mazzoleni and Nelson argue that the risk of this effect is higher than most studies anticipate since they do not account for possible high transaction costs for licensing agreements (pp. 280-282). Merges and Nelson (1990) argue on the basis of historical evidence from the radio, aircraft and phar- maceutical industries that in the case of cumulative research broad patents can lead to blockages. Further they conclude that: “[t]here is no evidence [...] that firms coordinate the development of a prospect by licensing the cultivation of particular applications of a broad technology to particular licensees; indeed, patents were often pooled and cross licensed en masse to all firms seeking to enter the field” (pp. 908-909).

39 _{In this setting a third characteristic in addition to patent length and breadth becomes important, namely the height of} IPR protection. The height of protection determines the novelty requirements and thus together with patent breadth determines the level of protection for the initial invention from follow-on inventions (Gallini and Scotchmer, 2002, p. 66). Regarding the height of protection Scotchmer and Green (1990) argue with some reservations for weak novelty requirements.

40 _{Kitch (1977) argues in the case of sequential invention for a broad initial patent in order to prevent over investment and} wasteful duplication and thus secure orderly development of the follow-on invention. Licensing could assure the consideration of other research ideas and capabilities for follow-on invention.