Bad patent info causes a litany of hurdles to diffusion –patent trolls, anticommons, and increased litigation.

Third, bad patent information quality, since it makes freedom to operate analyses prohibitively costly, favors patent trolling and the risk of patent litigation. Trolling behaviors consist of companies trying to provoke infringement and place infringing firms in hold-up situations (Pénin, 2012). Hold-up means that a firm has made sunk investments, such as research and development investments, marketing investments or specific assets investments, and is therefore committed to its technological trajectory. It can hardly stop using the technology if it is accused of infringement. Therefore, firms trapped in hold-ups are in a very uncomfortable and unbalanced position to negotiate licenses on infringed patents. Consequently, the strategy followed by patent trolls is to try to provoke hold-up situations by hiding their patents or by refusing to grant licenses before potential infringers have made sunk investments. This enables patent trolls to capture a share of the value that is much higher than the intrinsic value of the technology they own (Lemley, Shapiro, 2006; Farrell et al., 2007; Shapiro, 2010). The use of patents to provoke hold-up situations may also occur during the implementation of standards, where firms that own essential patents might be tempted to hide them during negotiations to increase their value ex-post (Tassey, 2000). This is particularly the case for ICT (Maskus, Merill, 2013; Lemley, Shapiro, 2013). Patent trolling entails important costs for society. According to a White House patent report (2013), trolls sued approximately 100,000 companies in the United States in 2012. Similarly, Bessen et al. (2011) found that “NPE lawsuits are associated with half a trillion dollars of lost wealth to defendants from 1990 through 2010, mostly from technology companies. Moreover, very little of this loss represents a transfer to small inventors. Instead, it implies reduced innovation incentives and a net loss of social welfare”. Furthermore, patent trolls have a deterring effect on R&D investment since, given the risk of encountering a patent troll, innovative companies might be more reluctant to invest (Pénin, 2012). In the same vein, due to the risks of trolling, firms might be more reluctant to collaborate and perform open innovation. Finally, patent trolls might also reduce the diffusion of a given technology (Tucker 2013). The existence of patent trolls is clearly linked to the bad quality of patent information. In a world where patent information is perfect, i.e. where patent users can identify at a minimum cost patents that they might infringe, firms that are trapped in hold-up situations are those that have not done clean freedom to operate analyses. Firms that perform these analyses properly can identify relevant patents and either invent around them or negotiate a license. But in a world with bad patent information, freedom to operate analyses are too costly and it is impossible for a manufacturing firm to evaluate ex-ante its freedom to operate. In such a world patent trolling is pervasive.

Fourth, bad patent information also massively increases the problem of “the tragedy of the anticommons” (Heller, Eisenberg, 1998) and the proliferation of costly patent litigations. A tragedy of the anticommons occurs when a technology is protected by a large number of patents. Therefore, ownership over this technology is fragmented. It is shared between a large number of patent holders, each having the right to refuse that others exploit the technology. In this case, firms that want to use the technology must obtain the agreement of all the patent owners. This fragmentation of ownership over a single technology therefore entails important transaction costs and induces a risk that the technology does not diffuse as it should. Due to the proliferation of patents, a promising technology might remain underused because it is too costly to negotiate and buy licenses from all the patent owners. This problem of anticommons is also known in the literature as a problem of royalty stacking (David, 2011). Empirically, the anticommons problem has been put forward by Murray and Stern (2007) and Von Graevenitz et al. (2011; 2012) [6]

If the anticommons problem is a possible consequence of the proliferation of patents, another possible consequence is the multiplication of patent litigations. Indeed, when a large number of patents protect a single technology, potential users of this technology must negotiate licenses with all the patent holders. If this is too costly, they can either decide not to use the technology, inducing an anticommons effect, or they can decide to use it without the permission of the patent owners. This second option might thus provoke a multiplication of costly and wasteful patent litigation.

Obviously, the main cause of the anticommons problem lies in the proliferation of patents in a given sector. The more patents, the more likely it is to experience anticommons or litigations. This is linked to the well-known problem of multiple marginalization already put forward by Cournot almost two centuries ago. But the risk of anticommons and costly litigations is also largely reinforced by the low quality of patent information, which makes it even more difficult to navigate through the patent thicket. Indeed, when patent information is of good quality, firms can identify relevant patents and negotiate a license for each of them. In this case, the proliferation of patents only makes it much more costly for manufacturing firms to use the technology since they have to buy a license for each patent they use (thus leading to the phenomena of royalty stacking). But when patent information is of low quality, firms have difficulty identifying relevant patents, thus facing the risk of hold-up and multiple litigations in addition to the problem of royalty stacking. In other words, the low quality of patent information worsens the problem raised by the proliferation of patents.