The Inverse Cournot Effect in Royalty Negotiations with Complementary Patents
The concept of “royalty stacking” is based on the so-called Cournot complements problem. Specifically, it is argued that the decision of a large number of inventors to license complementary patents necessary for the development of a product leads to excessively large royalties. In their paper, Gerard Llobet and Jorge Padilla amend the standard Cournot complements model by taking into account that patents are only valid and infringed with some probability, and can be challenged in court. This new economic model shows that royalty stacking is not an issue.
The arguments in favor of “royalty stacking” rely on theoretical models which reformulate the well-known Cournot complements problem in a licensing framework. Cournot showed that consumers are better off when all complementary product inputs are produced and marketed by a single firm. In industries where each single product is covered by multiple patents, a patent holder considering the royalty to charge may not fully take into account that an increase in this royalty is likely to result in a cumulative royalty rate that may be too high according to the licensees. The results derived by the two authors demonstrates that although the benchmark model has received substantial attention in the policy debate, it lacks not only practical evidence but also a proper theoretical foundation.
The new model departs from the existing literature in the following dimension: it is assumed that manufacturers of products covered by multiple patented technologies may challenge in court the patents that cover these products and, crucially, the likelihood that a judge rules in favor of the patent holder is increasing in the number and quality of its patents.
The model starts with the observation that, from the point of view of a technology implementer that produces in the downstream market, the decision to litigate is based on three important aspects: the strength of the patent portfolio under consideration, the legal costs of going to court and the additional profits that the producer expects to make if one or more patents are invalidated. Lower legal costs, weaker patents and higher expected gains from not being required to license the portfolio of a patent holder foster the decision of a technology implementer to litigate.
Patent holders take into account the implementers’ option to litigate when setting their royalties. In the new model, litigation will impose an upper bound on the royalty rate that firms can demand. This cap varies depending on the strength of the patent holder’s portfolio. A patent holder with a stronger patent portfolio will demand a higher royalty rate in the licensing negotiations than one with a weaker one. This is a first and important difference with the standard model underpinning the royalty-stacking benchmark. It provides a description of the actual royalty rates that firms negotiate, because stronger portfolios command higher royalty rates, irrespective of whether those negotiations involved essential or non-essential patents.
Moreover, the authors find that two patent holders with a small portfolio may command a lower royalty rate than a unique patent holder with a portfolio that corresponds to the sum of both and that this will be the case when the legal costs involved in litigating validity and/or infringement are low. This result is contrary to the implications of the standard Cournot complements model, in which the two patent holders would commit to high prices.