When less is more: optimal patent breadth under the threat of patent validity challenges.

• Author: Yiannaka, Amalia

1. Introduction

   Firms that choose patent protection as a means of protecting their intellectual property know that their effort to safeguard their technological territory through patent protection does not always conclude with the granting of the patent. Patents may be challenged after the patent grant through the launching of a direct patent validity attack in the Patent Office or in the courts or through an indirect patent validity challenge during an infringement trial. (1) While indirect patent validity challenges are usually launched by the patentee's competitors, direct validity challenges can be launched by competitors and by various interest groups, individuals, and nonprofit organizations. (2,3)

   A consequence of a direct or indirect validity challenge is that a patent can be revoked or its scope can be amended. Statistics on direct validity challenges show that 6% of all patents granted by the European Patent Office (EPO) are directly challenged within nine months after the grant (EPO 2001), (4) with 80% of these patents ending up amended (Cyranoski 2004), while more than 75% of the patents whose validity is directly challenged in the U.S. Patent and Trademark Office (USPTO) are either revoked or amended (Barton 2000). Statistics on indirect validity challenges tell a similar story. Allison and Lemley (1998) point out that, since the creation of the Federal Circuit in 1982, courts litigating the validity of patents have found approximately 45% of them to be invalid, while Miller and Davis (1990) state that almost half of the patents whose validity is indirectly challenged during an infringement trial are invalidated.

   The breadth of patent protection has been shown to be an important determinant of patent litigation in studies examining the factors that expose patents to litigation risk (Lerner 1994; Lanjouw and Schankerman 2001: Harhoff and Reitzig 2004). The patent literature suggests that the relationship between patent breadth and the occurrence and the outcome of patent validity challenges is such that the greater the breadth of patent protection, the greater the likelihood of both direct and indirect validity challenges and the greater the likelihood that the courts will invalidate the patent or narrow its scope. Specifically, Lentz (1988) and Merges and Nelson (1990) observe that the greater the patent breadth, the greater the probability that the validity of the patent will be challenged, while Lanjouw and Schankerman (2001) and Harhoff and Reitzig (2004) empirically find a positive relationship between the breadth of the patent and the probability of indirect and direct validity challenges, respectively. In addition, Merges and Nelson (1990) and Waterson (1990) state that courts tend to uphold narrow patents and revoke broad ones because the greater the patent breadth, the harder it is to show novelty, "nonobviousness," and enablement, and thus the harder it is to demonstrate that the patent is valid (Cornish 1989).

   Given the escalation in patent litigation during the last decades, especially in the field of biotechnology, and given the increasing number of patents that are invalidated after being challenged (Lenz 1988; Choi 1998: Barton 2000; Bessen and Meurer 2005), the determination of effective patent protection is becoming increasingly important for innovators. Since the breadth of patent protection is determined, to a large extent, by the innovator through the claims that she makes in the patent application, the innovator can affect, through her patent breadth decision, whether the patent will be challenged and the outcome of the challenge. (5) Consequently, the innovator's patent breadth decision determines the innovation rents that can be captured by the patent.

   The purpose of this article is to examine the breadth of patent protection claimed by an innovator who maximizes her ability to appropriate innovation rents. Specifically, the article models the innovator's optimal patent breadth decision for a drastic process innovation when the innovator/patentee faces the possibility of a direct patent validity challenge by noncompetitors and potential entry and infringement by a competitor who provides a technologically equivalent process. The focus is on drastic innovations because where drastic innovations are concerned the innovator cannot generally rely on the Patent Office for help in structuring her patent claims, as the Patent Office's role in refining patent claims is often limited. (6)

   The only other study that theoretically examines the innovator's patent breadth decision is by Yiannaka and Fulton (2006). Their study focuses on the optimal patent breadth chosen by an innovator who has developed a drastic product innovation and faces potential entry and infringement by a superior quality product, where entry, if profitable, occurs with certainty (i.e., the entrant's research and development [R&D] process is assumed to be deterministic). This article differs from that of Yiannaka and Fulton (2006) in two major ways: It examines the incumbent's optimal patent breadth decision in a market in which innovations are viewed as quality equivalents rather than quality improvements, and it models the entrant's R&D process as being stochastic. As will become evident in the analysis that follows, the type of innovations considered has important implications for the innovator's patenting behavior. Whereas in the quality ladder model of Yiannaka and Fulton (2006) the innovator faces a potential entrant with an incentive to locate as fat away as possible from the patented product, even under infringement, when innovations are viewed as quality equivalents, as in our model, the innovator faces an entrant that would be better off copying the patented process (i.e., locating as close as possible to the incumbent in the process innovation space). Moreover, the modeling of the entrant's R&D process as a stochastic one allows us to examine the effect of the innovator's patent breadth decision on the entrant's incentive to generate a competing process (i.e., on the entrant's R&D spending, his probability of succeeding in the R&D process, and the timing in which success is realized). In addition, this article examines the effect of both direct and indirect patent validity challenges on the incumbent's patent breadth decision.

   The innovator's profit-maximizing patent breadth decision is modeled as a sequential game of complete information between an incumbent innovator and a potential entrant. The game begins with the incumbent deciding on the patent breadth that she will claim for a process innovation she has generated, which will be used for the production of a nonpatentable product. (7) Once the patent is granted, it may be directly challenged by a noncompetitor/third party. The entrant observes whether the patent has been challenged and decides whether to enter into the incumbent's market and, if entry occurs, whether to infringe the patent or to invest in the generation of a non-infringing process. The assumption that a noncompetitor, and not the potential entrant, directly challenges the validity of the patent is consistent with the findings of Bessen and Meurer (2005), who show that about a quarter of patent lawsuits occur between firms that are technologically distant and do not compete in the same markets.

   The results of the model show that the greater the patent breadth, the smaller the entrant's incentive to develop a non-infringing process and the smaller the entrant's optimal flow of R&D spending. However, when the incumbent can use the breadth of her patent to deter market entry she maximizes expected returns by claiming the smallest patent breadth possible that could achieve entry deterrence. In addition, the model shows that when entry cannot be deterred, while the incumbent can always use her patent breadth to induce the desired behavior by the entrant (i.e., infringement or non-infringement), claiming the maximum patent breadth possible is never an optimal strategy. This result contradicts the conventional wisdom that the innovator has an incentive to claim as much as possible (Lentz 1988; Lanjouw and Schankerman 2001) and, in part, the findings of Yiannaka and Fulton (2006), who show that even though in most cases claiming the maximum patent breadth is not an optimal strategy for the incumbent innovator, there are conditions under which this strategy would be optimal (e.g., when infringement is not optimal for the entrant).

   The theoretical framework developed to examine the innovator's optimal patent breadth decision and the main assumptions of the model are presented in section 2, followed by a description of the analytical solution of the patent breadth game in section 3 and a conclusion in section 4.

2. Model Setting

   Patent Breadth and the Process Innovation Space

   The model considers two agents: an incumbent innovator who has developed a patentable process innovation and a potential entrant. At the beginning of the game the incumbent's process has been developed and she has decided to seek patent protection; these decisions are thus treated as exogenous to the game. (8) The incumbent determines the breadth of patent protection claimed that maximizes expected profits; the entrant in turns decides whether to enter into the incumbent's market and, if entry occurs, whether to infringe the patent or to invest in the generation of a non-infringing process. To capture the limited role the Patent Office plays in structuring patent claims when drastic innovations are concerned, (5,6) the model assumes that the Patent Office grants the patent as claimed; the Patent Office is thus not explicitly modeled.

   It is assumed that the patented process results in zero per-unit production costs and that if the entrant succeeds in generating his own non-infringing process, his process will be equally efficient. If the entrant is successful in generating his own process, he does not have to...