THE ECONOMY. Trolls disincentivize innovation and intra-firm collaboration, ensuring market collapse.

Innovation contributes to economic growth by creating new knowledge and increasing productivity in society (Arrow, 1962; Brown et al., 2009; Rivera-Batiz & Romer, 1991; Sorenson & Fleming, 2004). Intellectual property rights facilitate and increase firms’ incentives to innovate by providing formal legal protection against expropriation (Arora & Gambardella, 2010a, 2010b; Hu & Png, 2013; Levin et al., 1987). Over the past three decades, we have witnessed innovation-driven economic growth in developed economies like that of the United States (US), which is fueled by technological breakthroughs under the protection of the patent system. At the same time, observers have noted that patent litigations initiated by non-practicing entities (NPEs) have surged since the 2000s (Bessen & Meurer, 2013; Liang, 2010). NPEs, also commonly known as “patent trolls,” are firms that amass patents and build their portfolios to enforce their intellectual property rights and litigate against potential infringers but have little or no intention of further developing these patents.1 Therefore, NPEs often initiate patent infringement lawsuits with little or no genuine intent of developing or implementing their intellectual property rights. Their primary goal is to earn monetary profits through private remedies or out-of-court settlements (Cohen et al., 2016, 2019).

Recent studies have shown that NPEs harm the collective welfare of society and reduce firms’ research and development (R&D) investment. Bessen, Ford, and Meurer (2011) find that NPE litigations have cost defendants about half a trillion US dollars of wealth between 1990 and 2010. Cohen, Gurun, and Kominers (2016, 2019) find that firms reduce their innovation activities after being targeted by NPEs. Recent anecdotal evidence has also shown that NPEs may also hurt public health and downstream research by suing medical device testing companies that develop devices used in COVID19 testing.2 However, there is scant empirical evidence of the consequence of rising NPE litigation threats on firm innovation strategies and trajectories.

1 By contrast, practicing-entities (PEs), such as Intel and IBM, file litigations for real patent infringements so as to protect their rights in future development and application of the focal patent. 2 As an example of controversial NPE lawsuit, Labrador Diagnostics, a “patent troll”, sued a medical device testing company, BioFire, for patent infringement in March 2020, where BioFire was working on a coronavirus test and Labrador Diagnostics asked for an injunction against the sales and usages of BioFire’s testing device. See more details in https://www.patentprogress.org/2020/03/18/patents-in-the-time-of-coronavirus/. 2

In this study, we investigate the important question of how heightened NPE litigation risks impact the innovation strategies and trajectories of firms that are targeted by NPEs. We theorize that facing the rising threat of patent trolls, firms targeted by patent trolls (henceforth, target firms) would shift their innovation strategies to protect themselves from the advances of NPEs.

3 Specifically, we hypothesize that after target firms have become the defendants in NPE-initiated lawsuits, they will draw more upon their in-house technologies to reduce the potential legal ground for NPE lawsuits. Furthermore, we hypothesize that after observing the occurrence of NPE lawsuits, the peer firms operating in related areas will shift the locus of their innovation activities away from those of target firms to avoid areas with high NPE litigation risks. Lastly, these effects should be more pronounced when the patented technologies are costlier in terms of R&D expenses and when the product markets in which the firms operate are more competitive.

To test our predictions, we construct a longitudinal dataset of US public companies with comprehensive patent litigation data between 2008 and 2016 from the following datasets: (i) patent litigation cases from the Lex Machina database; (ii) NPE data from the Stanford NPE litigation dataset (Miller, 2018); (iii) patent and related patent data from the European Patent Office (EPO) Worldwide Patent Statistical Database (PATSTAT, 2018 edition); and (iv) firm financial data and other attributes from the Compustat database.

To perform our empirical analyses, we focus on the effect of the first NPE litigation filed against a target firm, given that the reaction of target firms to subsequent NPE litigations might be affected by the reaction to the first NPE litigation. The patent lawsuits initiated by NPEs are usually sudden and unanticipated by the target firms as target firms have little or no idea whether and when an NPE might initiate a patent lawsuit against them. Therefore, we perform a staggered difference-in-differences analysis using the timing of the first patent lawsuit by an NPE against a given target firm (treatment group) as a plausible exogenous event. We find that compared with their propensity-score matched control firms, which have never experienced any NPE litigation, the firms in the treatment group show 3 Shah IP Law, an IP law firm, indicated, “how can a startup deter patent trolls? Unfortunately, there is not much that you can do to prevent trolls from filing lawsuits against you.” https://www.shahiplaw.com/can-startups-deterpatent-trolls/. 3 an increase in the number of backward self-citations to their patented innovations post-NPE litigation. We also find that these firms experience a decrease in the number of forward non-self-citations. The first set of results is consistent with our predictions that, post-NPE litigation, these target firms draw more upon their in-house technologies to reduce the potential legal ground for NPE lawsuits. The second set of results suggests that the non-target firms in related technology areas will shift their locus of innovation away from those of the target firms. In other words, these non-target firms distance themselves from the existing innovation activities at risk of NPE litigation in an attempt to “escape” the patent trolls. These effects are more pronounced when R&D expenses for the patented technologies are higher and when the product markets are more competitive.

This paper makes the following contributions. First, we contribute to the heated debate on how patent trolling and litigation lawsuits affect the innovation activities of firms and potential firm collaboration. While proponents have argued that NPEs serve society by “policing” the patent infringement, the opponents believe that they exploit the legal process and disrupt innovation activities (Cohen et al., 2016, 2019). However, few or no studies have examined how NPE lawsuits affect target firms’ innovation strategy. Our study fills this important gap. Specifically, we show that NPE litigations hamper inter-firm collaboration and technology developments by making firms adopt more inwardlooking innovation strategies and distance themselves from the existing network of innovation activities. In other words, patent litigation threats deter potential collaboration among innovating firms in related areas and hinder the exchange and accumulation of knowledge.

Second, we document the potentially detrimental effects of NPE litigation on knowledge externality resulting from technological innovation and adoption (i.e., knowledge spillover). Prior studies have provided evidence on the importance of knowledge spillover. Jaffe (1986) shows that the productivity of a firm is affected by the R&D investment made by its peer firms. Fosfuri and Rønde (2004) show that, to benefit from technology spillovers, firms have stronger incentive to cluster in an industrial district when the growth potential of an industry is high. Bloom, Schankerman, and Van Reenen (2013) ascertain that the positive spillover created by corporate R&D investment outweighs its negative effects on product market competition. Chen, Chen, Liang, and Wang (2013) find that firms benefiting more from spillover experience higher post-R&D long-run stock performance. Our study 4 shows that the threat of NPE litigation induces firms to isolate themselves from their usual network of innovation, thereby reducing the positive knowledge spillover beneficial to their industries and society. Third, our study advances our understanding of the effects of US anti-patent troll legislation and its policy implications. Most states in the US have passed anti-patent troll laws between 2013 and 2016. While some prior studies in law have looked at the usefulness of the adoption of anti-patent troll law (e.g., Hrdy, 2018; Martyn, 2014; Mezzanotti & Simcoe, 2019; Thoman, 2014; Vogel, 2015), few or no studies have examined its impact on the innovation and business activities of firms. Consistent with our predictions, we show that after the passage of the state-level anti-patent troll laws, firms significantly lower the reliance on their own technology and increase technology collaboration among firms. This large-scale empirical evidence provides strong justification for the adoption and implementation of antipatent troll legislation across different states in the US, as well as support for policymakers imposing greater restrictions on NPE litigation. For example, Cohen, Gurun, and Kominers (2017) propose an automatic process of the administrative review of infringement lawsuits in US district courts that meets the minimum threshold of patent litigations as a way of providing feedback and information for the evaluation and adjustment of the patent system. Our paper provides the theoretical and empirical reasons for why these proposed policies are vital and needed.

2. Conceptual Framework and Hypotheses Development 2.1 Main Effects of NPE Lawsuits NPEs amass patents and build their patent portfolios to enforce their intellectual property rights and initiate patent infringement lawsuits against potential infringers, but, unlike practicing entities, they have little or no intention of further developing these patents. Their primary goal is to earn monetary profits through private remedies or out-of-court settlements (Cohen et al., 2016, 2019). NPEs frequently purchase unused patents and strategically claim the ownership of basic components of certain products of the target firms. Patent litigation cost consists of direct and indirect parts. The major component of direct cost is the legal fee. According to a survey of IP-related costs of the American Intellectual Property Lawyer’s Association, litigation costs per patent related to claim 5 construction could reach 2.4 million US dollars or more.

4,5 In addition, based on the PwC 2018 Patent Litigation Study, the median damages award for NPEs was 14.8 million US dollars from 2013 to 2017. 6 Indirect patent cost is typically even higher than direct cost, including the drop in stock price around the announcement of patent lawsuits and the reduction of revenue resulting from the risk that the court might rule in favor of the plaintiff. For instance, a court might order a damage payment or an injunction interrupting the production of products claimed by the patents (Bessen & Meurer, 2012). To avoid being trapped in a long, costly legal process, many target firms might accept settlement agreements when facing NPE-initiated lawsuits. NPE-initiated lawsuits are inconsistent with the commonly known deterrence model, which describes that the patenting system should function in a way that deters potentially unfounded patent lawsuits (Chen, 2013). This is because NPEs do not utilize their patents to produce goods and services, hence bearing little, if any, indirect patent litigation costs, such as prohibited sales and loss of customer trust (LaLonde & Gilson, 2017; Urbanek, 2008; Yun et al., 2017). To the extent that the occurrence of NPE litigation is unpredictable (Fishwick, 2013), the first NPE litigation against a target firm could alarm the firm, raising awareness of the possibility of litigation. Therefore, to reduce the risks of being targeted by NPEs again and incurring further direct and indirect patent litigation costs, firms that are targeted by NPEs are likely to adopt an inward innovation strategy after the initial NPE lawsuit to reduce future litigation threats from NPEs. In other words, target firms will draw more upon their in-house technologies to reduce the potential legal ground for NPE lawsuits in the future. Prior studies have used the proportion of citations to those patents belonging to the same firm to represent internalized knowledge use and proportion of the benefits captured by the original firm (Hall et al., 2001, 2005; Zhao, 2006). Indeed, the number of self-citations to a firm’s focal patents reflects the firm’s strategy of resources allocation between investing in internal development and the external acquisition of technologies. A higher number of self-citations indicates that a firm internalizes more knowledge created by its own R&D efforts, which suggests a strong competitive position in a 4 Claim construction is defined as “the process in which courts interpret the meaning and scope of a patent's claims.” Given that the claims in patent lawsuits define the invention to which the patentee is entitled the right to exclude, claim construction generally determines the possible outcome of patent litigations. See more detailed definition of claim construction at https://www.finnegan.com/en/insights/articles/claim-construction.html 5 https://blueironip.com/what-are-the-costs-to-enforce-or-defend-a-patent/ 6 https://www.ipwatchdog.com/wp-content/uploads/2018/09/2018-pwc-patent-litigation-study.pdf 6

particular technology area as the firm relies more on itself and less on other firms (Hall et al., 2005). Furthermore, firms with a high level of self-citations to their own patents can better appropriate their intellectual assets by reducing potential threats and hold-ups from outside patent owners (Alcacer & Gittelman, 2006; Zhao, 2006). Thus, we postulate that the target firms’ inward-looking innovation strategy would be reflected in an increase in the proportion of backward patent self-citations. This leads to our first hypothesis: