Table of Contents
List of figures
List of tables
List of appendices
List of abbreviations
2. Literature review
2.1 Patents and patent policy change in scientific literature
2.1.2 Patent policy changes in the history of the US patent system
2.1.3 The impact of patenting strategies on the patenting surge
2.1.4 Evidence from universities and federally financed laboratories
2.2 Markets for technology
2.2.1 Patent transactions
2.2.3 The emergence of patent trolls
3. Leading the way towards the weakening of the patent system
3.1 The problems of the pro-patent regime
3.2 The America Invents Act heralding the reversal of patent policy
5. Data and methodology
5.1 Dataset and sample construction
6. Results and discussion
6.1 Evidence from time series examinations
6.2 Robustness check
7.1 Theoretical and practical implications
7.2 Limitations and future research
List of figures
Figure 1: Patent transactions between firms on the US market for technology
Figure 2: Patent transactions involving non-US companies as assignees
Figure 3: Transactions to assignees from selected geographical areas
Figure 4: Patents acquired by patent trolls
Figure 5: Patents sold by patent trolls
Figure 6: Troll-to-troll patent transactions
List of tables
Table 1: List of Variables
Table 2: Chow breakpoint test
Table 3: Descriptive statistics
Table 4: Probit regression
List of appendices
Appendix 1: US patent applications from 1995 to 2015
Appendix 2: Overall patent-only transactions on the US market for technology
Appendix 3: Troll-to-troll transactions of Acacia Research and its shell companies from 2012 until September 2015
Appendix 4: Troll-to-troll transactions of further patent trolls from 2012 to 2015
Appendix 5: Correlation matrix and variance inflation factors
Appendix 6: Bibliography of literature on changes in the modern US patenting environment
Appendix 7: Bibliography of related articles on markets for technology
Appendix 8: Data points of figures and appendices
List of abbreviations
Abbildung in dieser Leseprobe nicht enthalten
This study investigates the impact of the recent US patent policy weakening around the America Invents Act, analyzing the US market for patent transactions. In an investigation of transactions derived from the USPTO’s Patent Assignment Database, we examine patent transfers between companies from 1995 until September 2015. Due to declining acquisitions, but increasing sales of patents conducted by patent trolls after 2012, our findings suggest that the business of patent trolls largely suffered from recent policy measures. Furthermore, we find that large patent trolls have changed their strategy towards splitting risk, as they recently tend to distribute their patent portfolio across different shell companies. However, we are unable to identify a decline concerning the activity of firms on the market for patent transactions, due to the weakening of patent protection. We ascribe the consistent level of participation on the market to a redistribution of intellectual property rights after the policy changes and to purchases motivated by speculations on a further policy reversal. Additionally, we observe an unexpected increase in patent sales to non-US companies in 2014 and 2015, as the slight decrease in patent purchases by other non-US companies is overcompensated by a growing number of transactions to Asian assignees
Keywords: Markets for technology - Weakening of patent rights - America Invents Act - US patent policy - Patent trolls - Patent transactions
Abbildung in dieser Leseprobe nicht enthalten
In dieser Studie werden die Auswirkungen der Abschwächung von Patentrechten in den USA anhand des Patenttransaktionsmarktes analysiert, indem alle von 1995 bis September 2015 getätigten Transaktionen, die in der Patent Assignment Database des USPTO festgehalten sind, untersucht werden. Der Umschwung des politischen Klimas im Patentsystem hat besonders Patenttrolle beeinflusst, was an einer Verringerung der Anzahl an gekauften Patenten und einer Erhöhung der Anzahl an verkauften Patente seit 2012 erkennbar ist. Weiterhin können empirische Belege für eine Veränderung in der Strategie von großen Patenttrollen gefunden werden. Diese verfolgen zunehmend eine Strategie der Risikostreuung, indem sie ihr Patentportfolio auf Tochterfirmen verteilen. Was in dieser Studie nicht identifiziert werden konnte, ist ein genereller Einbruch des Marktes für Patenttransaktionen aufgrund der legislativen Veränderungen im amerikanischen Patentsystem. Die weiterhin hohe Beteiligung auf diesem Markt lässt auf eine Neuverteilung von einigen Patenten, sowie Spekulationen auf eine weitere Umkehrung der Patentpolitik zurückschließen. In Hinsicht auf die Zahl der Transaktionen an ausländische Unternehmen wird weiterhin ein unerwarteter Anstieg der von diesen Firmen gekauften Patente in den Jahren 2014 und 2015 identifiziert. Die hohe Anzahl an Transaktionen an asiatische Unternehmen überkompensiert hierbei die rückläufigen Transaktionen an andere, nicht-amerikanische Firmen.
Abbildung in dieser Leseprobe nicht enthalten
In the history of research, various sources describe variances in the strength of intellectual property protection as a pendulum swinging back and forth over time (Cohen, 2004; Jaffe & Lerner, 2011; Ziedonis & Hall, 2001). Taking into consideration the interests of various players in the field, most importantly the diverse interests of large industries whose businesses are massively affected by patent policy, it is understandable that the strength of patent protection is challenged in an ongoing manner. From a research perspective, legal reforms have enabled scholars to conduct important studies and contribute various findings to the understanding of economic developments (Cohen, Nelson, & Walsh, 2000; Hall & Ziedonis, 2001). Recently, the severe changes in the US patenting environment around the Leahy-Smith America Invents Act (AIA) have led to important new implications for all actors in the field, providing a fertile ground for scholarly investigation.
Although the US patent system was regarded to provide the worldwide highest level of patent protection for several decades (Park, 2008), this preeminence is changing since the proclamation of the America Invents Act in September 2011. Due to a variety of measures weakening the rights of patent holders, patent protection in the US is now at the lowest point since a long time. Besides the rigorous legislative changes implied by the AIA, several decisions of the Supreme Court additionally contributed to the weakening of patent rights, especially by limiting the scope of patentable subject matter. Besides other legislative adjustments, the America Invents Act introduced several post-grant proceedings, offering a variety of possibilities to challenge the validity of a patent, after it is granted. An important example for an impactful decision of the Supreme Court, on the other hand, is displayed by the case of Alice Corp. Pty Ltd vs. CLS Bank International, in which abstract ideas were generally declared as non-patentable subject matter. Since this decision, the validity of software and business method patents is shrouded in uncertainty. More generally, practitioners in the field are facing a high level of insecurity concerning the validity of patents and the outcomes of patent litigation (Lloyd, 2016a; Wild, 2016). A huge debate about the design of the patent system is taking place in the US, while the uncertainty in the field limits the scope of actions for many firms.
The last major changes in the US patenting environment comparable to the recent developments were decided on over 30 years ago. The establishment of a specialized appellate court, the Court of Appeals for the Federal Circuit (CAFC), was initiated early in the 1980s and at that time, the political climate towards patents was going through fundamental changes. Researchers agree that the establishment of the CAFC and its favorable decisions towards patent holders, the establishment of the Bayh-Dole Act facilitating university patenting and important Supreme Court rulings displayed a major strengthening of patent rights throughout the 1980s (Gallini, 2002; Jaffe, 2000; Mazzoleni & Nelson, 1998). This pro-patent regime caused a shift in the patenting strategies of companies, especially in certain industries described as cumulative technology industries. This shift in patenting strategies is regarded as the main driving force behind an ongoing increase in US patent applications, a surge in patenting which emerged in the early 1980s (Cohen et al., 2000; Hall & Ziedonis, 2001). But despite promoting innovative output, the pro-patent policy in the US also lead to several problems that are now, more than 30 years later, addressed by recent reforms.
As products were becoming increasingly complex and built on several patentable inventions, the threat of holding up competitors progressively allowed patent holders to obtain substantial licensing revenues. Companies like IBM started to leverage their large patent portfolios by forcing other firms to license their technologies, which caused a patent portfolio race (Grindley & Teece, 1997; Hall & Ziedonis, 2001). Especially in cumulative technology industries, firms started amassing patents in order to improve their bargaining position in cross licensing agreements. Strong patent enforcement transformed patents into strategic weapons and caused the generation of patent thickets, where overlapping patents cover similar inventions and the holdup of competition is strongly facilitated (Shapiro, 2001). Especially in those industries, companies focusing on the exploitation of revenues from patents entered the field.
Patent trolls, which are also referred to as non-practicing entities (NPEs), are characterized by the sole exploitation of the patent portfolio, without actively producing goods or offering services. In cross-licensing negotiations patent trolls are invulnerable against producing companies, since cross licensing agreements do not affect firms that do not produce. Several researchers argue that patent trolls are holding up practicing companies to an extent that harms the US economy and regard patent trolls as the ultimate proof that the patent system relied on overly strong patent protection for too long (Bessen, Meurer, & Ford, 2011; Reitzig, Henkel, & Heath, 2007). These problems caused by an overly long period of strong patent protection, especially the emergence of patent trolls, are addressed by the recent weakening of patent protection around the AIA.
This study is about to examine the impact of the recent weakening of patent rights on the US market for patent transactions. Due to the timeliness of the events, empirical research on the topic is still scarce. To the best of our knowledge, there exists no study investigating the effect of the weakening of patent rights on the market for patent transactions, except for a small number of online articles (Andreyev, 2016). Research from the law area provides insights concerning the development of patent litigation cases filed at US district courts in the context of the weakening of patent rights (Barry, Arad, Ansell, Cartier, & Lee, 2015; Howard, 2016; Unified Patents, 2015), or analyses of the outcomes of the review programs that were introduced by the AIA (Dolin, 2014), but an economic interpretation of recent reforms’ implications is lacking. Therefore, we do not only address the question, if the legislative changes achieved the goal of curbing the activity of patent trolls, but moreover examine the implications for the overall US market for patent transactions. Four years after the signing of the AIA, we construct a dataset of patent transactions derived from the Patent Assignment Database of the USPTO, to provide empirical insights on the developments after the establishment of the AIA.
In this study, the US market for patent transactions is examined in three different contexts. We investigate the impact of the legislative changes on the overall participation of companies on the US market for patent transactions, analyze patent purchases by non-US companies and examine the reactions of patent trolls to the changing policy. Our results identify a strong decline of patent purchases conducted by patent trolls, while the number of patents sold by trolls is considerably increasing, after the reforms. Therefore, we argue that the changes in the legal environment significantly impeded the business of patent trolls, as trolls are increasingly selling patents that might have become unenforceable, but on the other hand acquire a significantly lower amount of new patents. Moreover, we find evidence in favor of the notion that the patent policy changes influenced the strategic behavior of some large patent trolls. Large patent trolls seem to have adopted a strategy of risk splitting by distributing their patent portfolio among shell companies. Although this practice is not particularly new, our data show an unprecedented increase in transactions between patent trolls after 2011, whereby the vast majority of these transactions are identified to be transactions between patent trolls and their subsidiaries. Concerning this strategy change, the most active company in the data is Acacia Research, one of the most notorious patent trolls in the US that transferred a substantial amount of patents to a variety of shell companies.
Further, we are unable to find support for the hypothesis that the participation of firms on the overall market for patent transactions is declining due to the patent policy intervention. Rather, our data suggest that transactions on the market for patents are increasing since 2014. Since the value of IP was lowered by the reforms, we explain this finding by the fact that participators on the market for patent transactions speculate on a further policy reversal. Moreover, the legislative changes might have triggered a reallocation of patents on the market, as some companies might have lost their interest in patents due to the lower level of protection, while others might be prone to increasingly acquire patents due to lower prices on the market. Concerning patent purchases of non-US companies, we are unable to observe a sustaining decline of transactions to foreign firms. Although we identify a breakdown of the number of patents acquired by non-US firms in 2013, the years 2014 and 2015 show a substantial increase in transactions to foreign assignees. A deeper look into the data reveals that this increase is in particular driven by transactions to Singaporean and Japanese companies, whereby the number of patents sold to Asian companies overcompensates the slight decrease in the number of purchases conducted by other non-US companies The remainder of this paper is structured as follows: Section two reviews literature on patents, the economic impact of changes in the US patent system and markets for technology. Section three provides an overview of the weakening of patent rights and describes recent developments in the US patenting environment, while the hypotheses are derived in section four. Section five describes the generation of the dataset, outlines the empirical approach and explains the variables included in the study, while chapter six presents and discusses the results. The last chapter concludes with a summary of the most important results, presents implications for theory and practice, explains the study’s limitations and offers avenues for future research.
2. Literature review
In the literature review, we place our study in the context of theoretical and empirical papers on changes in the patenting environment, as well as markets for technology. The literature review is organized in two sections. The first section genuinely describes the nature of patents and continues with a thorough review of studies on the impact of legal changes in the US patenting environment. A special focus is put on studies describing developments beginning in the 1980s, around the establishment of the Court of Appeals for the federal Circuit, which provide several important implications for the understanding of the processes surrounding recent reforms. The second part of the literature review describes extant literature on markets for technology. Since the empirical part of our study addresses the US market for patent transactions, we thoroughly review studies on markets for technology, including the areas of licensing, patent transactions and the activity of patent trolls in markets for technology.
2.1 Patents and patent policy change in scientific literature
This chapter begins with a general description of patents and the ways in which patents are used. The following sections discuss studies examining the impact of patent policy changes, both in a general context and later in the specific context of university patenting and patenting by federally subsidized laboratories. The central question addressed in this part of the literature review is, how patent policy changes affect the actions of participants in the market, since the developments emerging from the patent US policy of the last decades are now the cause for recent legislative changes. Therefore, it is important to review studies that examine, how patent policy intervention influences market developments in subsequent years. We predominantly review studies on changes in the US patent system, since this study’s focus lies on the US market. Moreover, the vast majority of research on the topic is conducted using US patent data.
A patent is a document granted to an inventor by an authorized governmental agency, crediting the right to exclude others from using, making or selling the patented invention. This right is obtained for a limited time period. Intellectual property rights represent the incentive for innovation that a state provides inventors with for the disclosure of new solutions to the public. An invention requires to be a new, non-trivial and commercially useful application, in order to be patentable. After being granted, a patent includes comprehensive information regarding the inventor, his or her employer and the invention’s technological antecedents. This publicly available information is accessible in computerized form. The references or citations included in a patent link it to other patented inventions upon which the original invention builds. The explicit patents that have to be cited for this purpose are determined by the patent examiner, whose duty it is to investigate the validity of patent applications and to concurrently ensure that the scope of a patent’s claims represent the technological improvement that the patent declares.1 As the USPTO is often referred to as being understaffed in the face of the amount of filed patent applications in the waiting line, patent examiners face a trade-off between a thorough examination and working off of the constantly growing number of patent applications2. Thus, the validity and scope of patents are frequently challenged in patent litigation, as a patent’s validity is not certain, even after the patent is granted (Teece, 2000).
A patent is usually valid for 20 years in the designated geographical area (WIPO, 2015a). However, it has to be kept in mind that patents do not practically provide a monopoly for the use of a technology, as it is the duty of the patent holder to keep an eye on the market and react to infringing behavior. Furthermore, patent litigation can be extremely expensive and time intensive, although the outcome is not always predictable (Barry et al., 2015; Lanjouw & Schankerman, 2001). Patents can have offensive and defensive strategic purposes, whereby the first cover activities to prohibit competitors from the use of certain inventions, while the latter describe efforts to prevent itself from being sued by others. While a patent displays an extremely effective mechanism to protect innovations in some industries, as in chemicals or pharmaceuticals, patents in other branches oftentimes offer less certainty and protection. Small firms and individual patent holders are thus often overstrained controlling the infringement of their patents, as patent protection in general is admittedly uncertain and difficult to handle (Cohen et al., 2000; Lemley & Shapiro, 2005).
2.1.2 Patent policy changes in the history of the US patent system
Patent data have been used to measure innovative output for a long time. Among a variety of research using patent data exists a stream of research which examines fluctuations in the number of patent applications. These variations are oftentimes examined in the context of governmental interference in the patenting environment. Griliches (1990) provides a comprehensive overview across the application of patent statistics in research, identifying a peak of patenting towards the late 1920s. This peak is followed by a massive decline over the following years. Schmookler (1966) names different reasons for the shortfall of patenting beginning in the 1930s: the negative attitude towards patents in the late 1930s, but as well the holdup in the Patent Office and the emergence of rising industries, which rely more on other appropriation mechanisms, such as secrecy and first mover advantages. His first explanation hypothesizes that increasing hostility towards patents and the weak enforcement of patent rights caused the decline in patenting activity, which was partly driven by the strong antitrust enforcement of corporate patenting at the time. Schmookler (1966) additionally argues that adverse compulsory licensing strongly decreased the incentives for companies to patent inventions, whereas even those companies unaffected by antitrust decrees reduced their patenting activity, as patents were losing their effectiveness. Although Griliches (1990) comments that this explanation appears “less cogent today” (Griliches, 1990, p. 1697), it remains unclear what the true reason behind the decline in patenting after the 1920s was. (Griliches, 1990) ascribes the lack of research on the impact of early patent reforms to the unavailability of computer readable data and statistical methods.
In contrast to the lack of research examining the impact of patent reforms during the 1930s, there is a considerable body of research investigating an unprecedented surge in US patenting starting in the early 1980s (Gallini, 2002; Jaffe, 2000; J. Kim & Marschke, 2004; Kortum & Lerner, 1999). Griliches (1990) mentions that the judicial climate towards patents has moved towards a pro-patent attitude since 1980s. The establishment of the Court of Appeals for the Federal Circuit, which certainly upheld a pro-patent policy, as well as the impact of judgments and rulings in favor of patents contributed to the evolvement of a historically beneficial era for patent rights (Gallini, 2002; Mazzoleni & Nelson, 1998; Shapiro, 2001). Hall (2004) gives an overview of the most important changes in the legal environment that took place in the 1980s and 1990s.3 Moreover, universities were allowed to retain the returns from patented inventions after the Bayh-Dole act of 1982 (Henderson, Jaffe, & Trajtenberg, 1998; Mowery, Nelson, Sampat, & Ziedonis, 2001). The establishment of the Court of Appeals for the Federal Circuit as an appellate court was introduced by the Congress to counteract the non-coherent and inconsistent sentences of district courts. Before the CAFC, legislation concerning patents was highly dependent on the district (Henry & Turner, 2006). Appeals of district courts’ decisions were from then on revised by the CAFC, which convoked an era of pro-patent attitude due to several favorable decisions for patent owners (Cohen et al., 2000; Gallini, 2002; Henry & Turner, 2006).
We put particular emphasis on the review of studies examining changes in the US patenting environment since the 1980s, since the recent weakening of patent rights is the first major subsequent governmental interference. Moreover, although there were several other important patent reforms and changes in the judicial climate towards patents in the history of the USA, there is no comparable body of research to those extensive studies examining the patenting surge in the 1980s. Looking back to the shortfall of patents in the 1930s, from the observation that the hostile patenting environment of the 1930s was accompanied by a decline in overall patenting, one could easily assume that a favorable governmental policy towards patents increases the amount of patenting. The other way round, one might assume that a disadvantageous patent policy leads to a decline in patent applications, as the pro-patent era of the 1980s was followed by a strong increase in patenting. Yet, these conclusions turn out to be premature.
The unprecedented surge in US patenting in the 1980s is examined in different large scale empirical studies, leading to diverse conclusions. Whereas Kortum and Lerner (1999) are reluctant to find evidence supporting the hypothesis that the increase in patenting was triggered by favorable patent policy and strong patent enforcement, Cohen et al. (2000) as well as Hall and Ziedonis (2001) do find evidence supporting this view. Conventional wisdom of the time regarded patent policy to be the reason for the surge in patenting, but the data of Kortum and Lerner (1999) suggest that the surge in patenting is explained by a worldwide increase in patenting, rather than a particular development in the US. Further, an alternative hypothesis called the fertile technology hypothesis is found to be more consistent with the data by the authors. This hypothesis suggests that new technologies in strongly growing fields, like biotechnology or information technology, raised the yield of patentable discoveries.
2.1.3 The impact of patenting strategies on the patenting surge
Cohen et al. (2000) come to a different explanation for the unprecedented increase in patenting, arguing that a change in the appropriability mechanisms and strategic behaviors employed by firms was caused by the strengthening of patent rights. The strategic behavior resulting from stronger patent rights could in turn have triggered increased activity in patenting. Comparing their data from 1994 with the results of a similar survey on firms’ appropriability mechanisms conducted in 1983, commonly known as the Yale survey (Levin et al., 1987), the authors find that strategic purposes of patents have gained in importance for patent holders. Interestingly, both Cohen et al. (2000) and Levin et al. (1987) find that patents are perceived as rather ineffective, compared to other appropriability mechanisms, such as lead time and secrecy. Further, the perceived ineffectiveness of patents is also confirmed by other researchers (Mansfield, 1986; Taylor, Silberston, & Silberston, 1973).4 That means, although patents are not necessarily required to appropriate returns from innovation, several firms nevertheless rely on patenting for different reasons.
The most prominent motives to patent identified by Cohen et al. (2000) include patent blocking, the use of patents in negotiations and defensive patenting to prevent the company from being sued by others. In this context, patent fences are highlighted as a frequently used mechanism.5 Hence, although patents are regarded as relatively ineffective in terms of appropriability purposes, the strategic uses of patents, such as blocking competitors or the use of patents in cross-licensing negotiations, have gained importance. In other words, Cohen et al. (2000) find that the motives to patent have broadened, since inventors do not only patent in order to obtain a monopoly for a technology, but also in order to gain strategic advantages, like the possibility to block competitors. It is important to mention that Levin et al. (1987) do not identify these motives in their study, which was conducted at an earlier time, when patents just began to gain in strength. This leads Cohen et al. (2000) to the conclusion that strategic motives emerged due to the new conditions in the patenting environment. Concluding, the era of strong patent rights might have triggered an increase in patent applications due to the fact that firms act strategically in the face of competition, whereby the strategic value of patents was raised. A further finding in their study supporting this view is that larger firms more heavily rely on patents since the establishment of the CAFC in 1982.
Consistent with the findings of Cohen et al. (2000), Hall and Ziedonis (2001) identify a dramatic increase in the propensity to patent since the early 1980s, which they term “patent portfolio race”. Their study focusses on the semiconductor industry, which is characterized by cumulative invention and is identified to be one of the thicket industries driving the upsurge in patenting (J. Kim & Marschke, 2004; Mazzoleni & Nelson, 1998). Cumulative technology industries are characterized by the fact that innovations usually build on various previous ones, which leads to large overlaps and has crucial impact on the possibility to hold up competitors (Scotchmer, 1991). Hall and Ziedonis (2001) argue that between 1982 and 1992, patenting in the semiconductor industry doubled, whereby this increase is not accounted for by higher R&D spending alone. Apart from the increasing strategic benefits of patents, the authors identify a second explanation for the increased patenting activity in the industry, as specialized technology firms progressively entered the market due to facilitated entry conditions, which are enabled by stronger patents. Not only the business model of specialized technology firms, but the business of patent trolls, as well, was assisted by the legislative changes6. Hall and Ziedonis (2001) relate this second explanation to the strengthening of the patent system, since the business model of selling or licensing specialized technology is greatly facilitated by stronger patent rights.
A further empirical study is addressing the same topic from an R&D perspective, rather than examining the surge in patent applications. J. Kim and Marschke (2004) provide further evidence supporting the hypothesis that stronger patent enforcement triggered the increase in patenting. Moreover, J. Kim and Marschke (2004) quantify the extent to which changes in the legal environment influenced the ratio of patents and R&D expenditures. They estimate that about 20 percent of the increase in patenting over the 1980s arose from the favorable patent policy, whereby increased R&D spending accounts for 70 percent of the increase in patenting. However, the authors admit that their model might understate the effect of legal changes if those legal changes as well lead to an increase in R&D spending. J. Kim and Marschke (2004) suggest that the increase in patent-R&D ratios they observe is greatest in cumulative technology industries, which they in turn relate to legal changes. Their data suggest that 60 percent of the increase in patenting across the US comes from only the electronics and computer industries, which supports the findings of Hall and Ziedonis (2001). The biotechnology sector, on the other hand was not found to rigorously contribute to the patenting surge, which is consistent with the findings of Kortum and Lerner (1999). Altogether, the study of J. Kim and Marschke (2004) supports the view that changes in the legal environment had a significant impact on the patenting surge. They moreover identify cumulative technology industries as the main driver of the patenting increase.
Hall (2004) further acknowledges the view that the CAFC and its decisions led to patenting among industries in which patenting was not particularly important before. Testing for structural breaks among the number of patent applications, a jump in the growth rate of patent applications could be identified in 1984, from 0.3 percent per year to an average of 6.9 percent. Controlling for structural breaks in Europe, Asia and developing countries, Hall (2004) comes to the conclusion that inventors resident in the US primarily account for the increase, with some help from foreign countries. This finding is again contrary to the findings of Kortum and Lerner (1999), who argue that the US patenting surge was rather driven by an overall increase in patenting, than a particular trend in the US. Hall (2004) suggests that the creation of the CAFC in 1982 and the litigation success of Texas Instruments against a number of firms in 1985 and 1986 were the most important influences from the side of the legal environment.
To put it in a nutshell, the majority of empirical research agrees that the pro-patent policy encouraged by the legal environment and the establishment of the CAFC had little direct impact on the overall patenting surge by US firms, but increased firms’ propensities to use patents as strategic weapons, which indirectly triggered an increase in US patenting. Although Jaffe (2000) calls the evidence from studies examining the surge in patenting inconclusive, later research agrees that the favorable patent policy indirectly promoted the increase in patenting (Gallini, 2002;
Hall, 2004; Somaya, 2012). The evidence from Cohen et al. (2000) and Hall and Ziedonis (2001) identify the growth in the strategic use of patents, especially in cumulative technology industries, which was indirectly triggered by the strengthening of patent rights. Although Kortum and Lerner (1999) highlight that an increase in patenting already started before 1982, the year in which the CAFC was established, Hall and Ziedonis (2001) argue that Kortum and Lerner might have dismissed the influence of the pro-patent shift in the patenting environment to easily.
Jaffe (2000) provides an overview including some of the previously discussed studies, which further involves studies examining the impact of the reforms on university and national laboratory patenting. The surge in patenting can be observed in an even more dramatic manner in the area of federally subsidized research. In the period from 1965 to 1992, overall patenting grew by only 50 percent, whereas for example the number of university patents multiplied by 15 in the same period, with a huge increase throughout the 1980s (Henderson et al., 1998). Moreover, the share of university patents in the US rose from less than 1 percent in 1975 to about 2.5 percent in 1990 (Trajtenberg, Henderson, & Jaffe, 1997) and the number of universities with licensing offices grew from 25 to 200 in 1990 (Association of University Technology Managers, 1996). Since the patenting environment for universities and other federally promoted research institutions went through comparable, significant changes in the early 1980s, this related stream of research will be reviewed in the next section.
2.1.4 Evidence from universities and federally financed laboratories
Further evidence concerning the changing US patenting environment of the early 1980s is provided by studies examining university patents and the patenting behavior of federally funded laboratories. Although these patents only account for a small fraction of the patenting activity in the US, the existence of a variety of studies shows particular scholarly interest in this specific topic. In 1980 the Bayh-Dole Act, which provided universities and other non-profit institutions the right to retain the property rights from their inventions, became law. Before the Bayh-Dole Act, the US government was allowed to claim all royalties from patents emerging from federally funded research. Additionally, in 1984, some limiting provisions of the Bayh-Dole Act were removed, further expanding the rights of federally funded research institutions. These fundamental changes were followed by an immense increase in university patenting, which increased fifteen times faster than the overall number of patents in the US. Hence, the right to obtain the revenues from own patents was one of the incentives for universities and federally funded research laboratories to increase their patenting activity (Henderson et al., 1998).
Henderson et al. (1998) identify two components leading to increased patenting by universities. The first component they suggest is the increasing fraction of university patents coming from smaller institutions, the second is a general decline in the average quality of university patents. The authors suggest a shift from basic research to applied research in the patenting of universities. Further research in the area takes on some of the findings of Henderson et al. (1998), while multiple other ideas of the authors are reflected with criticism. Coupe (2003) indicates that the strong difference between the patenting activities of firms and universities has decreased since the introduction of the Bayh-Dole Act. As the rates of university patenting increased towards the level of the growth rate of overall patenting, the author supports the assumption of a shift in the focus of universities towards applied research.
Mowery et al. (2001) challenge the view that the law reform triggered the surge in university patenting, indicating that Bayh-Dole had minor influence, but was not the main reason for the surge in patenting and licensing. Indeed, the impact of the Bayh-Dole Act is described as incremental, since a shift in university research independent of the legislative changes is observed by Mowery et al. (2001). Moreover, increased federal financial support for basic biomedical research in universities and a general rise in biotechnology research are responsible for the increase in university patenting and licensing, which indicates that other legislative changes were identified as the reason for the patenting surge. Mowery et al. (2001) highlight the Supreme Court decision in the case Diamond vs Chakrabarty7 as one of the most influential causes for the surge in university patenting, accounting for a large part of increased patenting among universities. As some of these developments predate the Bayh-Dole Act, there is evidence suggesting that university patenting and licensing would have increased, even without the Bayh-Dole Act, due to other changes in the legal environment.
Rafferty (2008) uses R&D expenditures in order to describe the influence of the Bayh-Dole act on university patenting. His results additionally confirm the dominant view, suggesting that important changes in university research predate the passage of the Bayh-Dole act. Hence, the author ascribes Bayh-Dole only little effect on the surge of university patents. Like in previously discussed studies, Rafferty (2008) is incapable to separate individual effects influencing the patenting and licensing surge, but suggests a shift towards greater industry financing of university research, which already started in the 1970s, to be another important influence for the surge in university patenting during the 1980s.
Although most research in the field concentrates on universities, the Bayh-Dole Act did not only affect university patents, but more generally patents that emerged from research financed by the federal government. There is a further study by Jaffe and Lerner (2001), who examine the influence of the changing patenting environment since 1980 on patenting and technology transfer at national laboratories in the US. The authors argue that the patenting environment of the 1980s definitely influenced the commercial activities of national laboratories, as their data significantly support their hypothesis.
All in all, researchers agree that legislative changes in the patent system had an important influence on the increased patenting activity by universities and federally financed laboratories since the early 1980s. However, it remains unclear if Bayh-Dole was the main driving force of the patenting surge in this area. Some scholars regard the Bayh-Dole Act as the most important influence for the surge in patenting in those areas (Coupe, 2003; Henderson et al., 1998; Jaffe & Lerner, 2001), other studies regard the Act of minor influence (Mowery et al., 2001; Rafferty, 2008), but in identify find other important changes in the patenting environment of universities as the influence for these developments, like the Diamond vs Chakrabarty decision (Mowery et al., 2001). In the end, the surge of patents from federally funded research projects was strongly driven by changes in the patenting environment. The favorable patent policy seems to have promoted the entry of less experienced universities in the area of patenting and additionally increased the incentives to patent inventions of lower quality.
2.2 Markets for technology
Literature on markets for technology is receiving increasing attention in research. The emergence and nature of markets for technology are progressively discussed in literature, as markets for technology have been growing to a considerable extent in recent years (Arora, Fosfuri, & Gambardella, 2001; Lichtenthaler & Ernst, 2007a; Teece, 1998). Some decades ago, the transfer of knowledge between firms was regarded as complicated and costly.8 Firms are increasingly trading knowledge on the markets for knowledge or technology, both in national and international contexts, whereby occasionally even the most central inventions of businesses are licensed (Keller & Chinta, 1990; Kline, 2003). Arora, Fosfuri, and Gambardella (2004) estimate the worldwide market for technology to account for $35 to $50 billion per year. Especially for small firms and individual inventors, licensing and sales of patents often represent some of the most important revenue generators (Galasso, Schankerman, & Serrano, 2013). As researchers start to focus on the question of how openness influences firms’ ability to innovate and appropriate the returns of innovation, markets for technology are increasingly moving into the spotlight (Chesbrough, 2006; Laursen & Salter, 2006). Consequently, researchers regard IP, especially patents, to be the key enabler for technology transfer (Dahlander & Gann, 2010; Gans & Stern, 2003; West, 2006).
While historically, firms almost exclusively relied on the internal exploitation of knowledge, recent developments show an increase in the external commercialization of knowledge (Lichtenthaler & Ernst, 2007e). This means that firms actively seek to generate revenues from existing intellectual property portfolios, most importantly, from patents. Texas Instruments for example received licensing revenues of $1.8 billion from 1986 to 1993 (Arora & Fosfuri, 2003; Grindley & Teece, 1997; Lichtenthaler & Ernst, 2007e), while IBM’s licensing revenues grew from about US$ 30 million in 1990 to US$ 1.9 billion in 2001 (Kline, 2003). Considering patent transactions, major arrangements have become more frequent in recent years. The $4.5 billion dollar deal where Microsoft, Apple and other companies bought 6000 patents from Nortel displays a notable example (Figueroa & Serrano, 2013). While intangibles accounted for only 38 percent of U.S. firms’ valuation in the beginning of the 1980s, the extent had increased to 70 percent by only the mid-1990s (Serrano, 2006; WIPO, 2003). Indeed, the external commercialization of knowledge assets is increasingly pursued by firms to appropriate the full revenues from intellectual property, which allows companies to cover the costs of technology development (Chesbrough, 2003; Ford, 1988). However, there are further reasons to participate on the market for technology, as demonstrated in the following section.
Arora et al. (2001, p. 423) provide a definition, whereby “a market for technology refers to transactions for the use, diffusion and creation of technology” and argue that the benefits of markets for technology lie in the reallocation and redistribution of knowledge. Teece (1986) suggests that complementary assets are an important precondition for appropriating the returns from technological innovation. Since “the company investing in research and creating new technologies isn’t necessarily the best one to commercialize every product opportunity” (Kline, 2003, p. 91), a firm with the appropriate complementary assets might be a better candidate to introduce an invention to the market, whereby the original inventor is likely to gain revenues exceeding the profits from self-exploiting the technology. Nevertheless, there are various difficulties when it comes to managing the activities of technology transfer, as well as certain imperfections within the markets for technology. Especially small firms have difficulties participating on markets for technology, which demand a certain level of legal expertise, know- how in the estimation of the value of patents, as well as financial resources (Hagiu & Yoffie, 2011). Hence, several authors suggest that technology transfer is especially complex, compared to transfers on conventional product markets (Arora et al., 2001; Bidault & Fischer, 1994). One example for an imperfection of markets for technology is the lack of market transparency (Gans, Hsu, & Stern, 2008; Hagiu & Yoffie, 2011). One result of this imperfection is that the efficiency of the market suffers from a lack of participation. Gambardella, Giuri, and Luzzi (2007) estimate a growth potential of 70% for the European patent market, in terms of licensing. Abolishing these obstacles to the market might be desirable, as Teece (1998) indicates that the efficiency of a market increases with market transparency and the number of participators.
In markets for knowledge, Lichtenthaler and Ernst (2007a) identify a necessity for firms to develop reputation as a knowledge provider, in order to generate a market pull effect for technology licensors. The distribution of profits in technology transfer is highly skewed, as small firms regularly struggle with the external exploitation of their patent portfolios, while a few large firms are earning immense revenues (Lichtenthaler, 2005). In fact, Lichtenthaler and Ernst (2007a) explain the difference between the success of pioneering companies and the remaining majority of firms with a higher level of reputation as knowledge provider. Their findings are supported by Sine, Shane, and Gregorio (2003), who identify institutional prestige as an important driver of university licensing. All in all, technology transfer is gaining increased importance for the competition of firms in globalized markets (Keller & Chinta, 1990; Teece, 1998). Firms whose business is based on technology or important knowledge assets that are protected by patents, as well as firms already holding patents that are only used to a minor extent, can draw important benefits from the participation on markets for technology.
2.2.1 Patent transactions
A key mechanism to realize the gains from invention is efficient technology reallocation by patent transactions (Galasso et al., 2013). When a patent is transferred from one party to another, the title of the patent and the rights associated with it are assigned to the seller. In the US, patent transactions are captured by the USPTO’s Patent Assignment Database under the term “assignor change”. Research analyzing these data reveals information on patent transfer, for example which patents are most likely to be traded. Serrano (2010) identifies several characteristics of frequently traded patents, as his findings suggest that high quality patents, e.g. patents with a high number of forward citations at a given age or a high level of generality, are more likely to be traded.
Moreover, he concludes that the age of a patent negatively affects the probability of the patent being traded. On the other hand, the fact that a patent has been previously traded or that its renewal fee has been paid recently positively influences the probability of the patent being traded. Bresman, Birkinshaw, and Nobel (1999) further acknowledge that the probability of a patent to be traded rests on the recency of a former transfer.
The nascent field of study focusing on patent transactions contains a few studies focusing on markets for technology in the late 19th and early 20th centuries (Lamoreaux & Sokoloff, 1999, 2001), while several other studies, examining data from the USPTO’s Patent Assignment Database, emerged after the works of Serrano (2006, 2010). Patent transactions are used to quantify the gains from trading patents and the costs of technology transfer in the market for innovation (Serrano, 2011), to study the value of patents in patent auctions (Odasso, Scellato, & Ughetto, 2015), to estimate the effect of patent transactions on litigation (Chien, 2011; Galasso et al., 2013), to examine the differences of patent transaction behaviors of small and large firms (Figueroa & Serrano, 2013), or to measure the size of an industry’s market for technology (Mani & Nandkumar, 2015). Galasso et al. (2013) further argue that patent transactions decrease the risk for follow-up litigation, due to a redistribution of patents on the market. Arora and Ceccagnoli (2006) state that firms expect an increase in competitive advantage from the acquisition of patents, which is partly driven by the opportunity to gain licensing revenues, besides being able to use the technology. Due to the strong interrelation of patent transactions and licensing, research in the area of licensing is reviewed in the following section.
Another important stream of research with huge importance in markets for technology is licensing, a more established topic in research, compared to the nascent field of patent transactions. The appropriability mechanisms on markets for technology are twofold: First, a patent holder can generate revenues by selling a patent, as discussed in the previous section. Second, a technology can be licensed to others. The advantage of licensing, compared to selling the patent, is that the patent holder does not lose control over the technology. Licensing means allowing a licensee to use a technology for a certain period of time in return for a previously bargained licensing fee. Other conditions of the contract are often established in advance, as well. These previously negotiated conditions can include several different measures, from the exchange of physical goods and employees to after sales services. Important employees or physical resources are increasingly licensed together with a patent, as “the wording of a patent is generally not sufficient to reproduce the technical work that is patented” (Bessy & Brousseau, 1998, p. 461). Out-licensing technology has become more popular among practicing firms in recent decades (Grindley & Teece, 1997; Lichtenthaler, 2005). Texas Instruments displays one of the most prominent examples highlighted by multiple researchers, as the company rigorously conducted reverse engineering to estimate the extent of infringement of their own patents by other firms and thereby calculated appropriate licensing fees (Grindley & Teece, 1997). Moreover, IBM’s licensing revenues grew from about US$ 30 million in 1990 to US$ 1.9 billion in 2001, whereby these revenues accounted for nearly 20 percent of the company’s net income in 2014, with a 98 percent profit margin (Kline, 2003). The latter example shows, how profitable licensing in markets for technology can turn out to be.
Due to the growing importance of the topic, there is a respectable body of research on licensing, covering various aspects of the field. Teece (1986) identifies complementary assets as an important factor influencing the licensing decision, since a patent holder who cannot, without substantial investment, appropriate the returns from an invention is more likely to license. Moreover, Teece (1986) regards IP protection as facilitator for licensing, since a weakly protected technology is easy to invent around. Additionally to the work of Teece, the licensing literature contains a huge body of theoretical contributions. The work of Arrow (1962) represents an important foundation for modeling the resource allocation in a licensing contract between a licensor and licensee and was followed by several studies modeling the right choice of the number of licensees (Katz & Shapiro, 1986), the division of the value created between licensor and licensee (Gallini & Wright, 1990; Kamien & Tauman, 1986), the likelihood of ex-ante or ex-post licensing (Gallini & Winter, 1985; Shapiro, 1985) and various other constellations in the licensing context.
Arora and Fosfuri (2003) introduce an important implication for licensing on markets for technology, as they consider not only a monopolist licensor, but acknowledge the competition between various licensors on the market for technology. Arora and Fosfuri (2003) argue that two effects influence the rate of technology licensing by a firm: the profit dissipation effect and the revenue effect. These effects demonstrate the trade-off between royalty revenues and the market share loss due to increased competition. Their theory suggests that competition on the markets for technology leads to more aggressive licensing behavior, especially if holders of patents do not operate in the product market. To test the hypotheses developed in Arora and Fosfuri (2003), Fosfuri (2006) employs an empirical setting. Her findings indicate that competition on markets for technology, as observable in the case of multiple licensors, indeed triggers more aggressive licensing behavior. Several authors build on the framework of Arora and Fosfuri (2003), studying licensing in markets for technology (Y. Kim & Vonortas, 2006; Motohashi, 2008). Empirical studies on licensing are rather scarce, compared to the variety of theoretical contributions, but there are a number of insightful papers. Gambardella et al. (2007) list several factors influencing the possibility of a patent being licensed, of which we review some selected: The presence and strength of patent protection, firm size and complementary assets and the industry context.9 First, the presence and strength of patent protection are found to encourage licensing in various studies (Anand & Khanna, 2000; Gallini, 2002; Teece, 1986, 1998). Arora (1995) especially highlights the importance of IP for the protection for tacit knowledge. Arora and Ceccagnoli (2006) empirically address the question of how the effectiveness of patents affects the patenting behavior of patent holders. It is possible to license a technology without protecting it by IP rights, but licensors usually want to know the quality of the patent, before they license it. Since disclosing an invention without IP protection enables imitation, inventions must be protected by IP rights. The data of Arora and Ceccagnoli (2006, p. 4) suggest that firms who do not patent rarely license their technology to others, as less than 10 percent of licensors in their sample do not patent. Further, nearly 33 percent of non-licensors do not patent. On the other hand, only 12 percent of non-patentees license their technologies, while about 40 percent of the patentees decide to license, which leads the authors to the conclusion that the presence of IP protection strongly influences licensing decision. Thereby, the strength of patent protection indirectly influences the licensing decision.
Secondly, various researchers regard the size of firms and complementary assets as an important indicator for licensing. Research suggests that small and very small ventures are especially prone to license (Arora & Ceccagnoli, 2006; Arora & Fosfuri, 2003). Analyzing a large scale dataset of European firms, Gambardella et al. (2007, p. 1180) find that the size of the firm is by far the most important determinant of a firm’s propensity to license, since the probability that a large firm offers a patent for licensing is 19 percentage points smaller than the respective probability of a small firm, while the conditional probability of an actual license is about fifteen percentage points lower. Interestingly, analyzing a large scale survey of Japanese firms, Motohashi (2008) regards large firms as well as small firms to be especially active in licensing, due to the heavy cross-licensing of technologies by large companies on the Japanese market. This U-shaped characteristic is confirmed by Zuniga and Guellec (2009), examining the same topic in a mixed survey of 600 European and 1600 Japanese firms. While large firms are usually predicted to be rather inactive in terms of licensing, as suggested by the findings of Gambardella et al. (2007), the contrary findings of Motohashi (2008), as well as Zuniga and Guellec (2009) can be explained by increased competition on the Japanese market for technology, which triggers more aggressive licensing behavior (Arora & Fosfuri, 2003; Fosfuri, 2006). Small firms are found to practice licensing extensively, since they lack complementary assets and therefore cooperate with larger firms to enter the market (Gans & Stern, 2003; Teece, 1986).
Thirdly, several authors argue for different licensing behaviors of firms in different industries, with patenting being especially effective in high-technology industries like chemicals, electronics or computers (Grindley & Nickerson, 1996; Grindley & Teece, 1997). Anand and Khanna (2000) provide the first large-scale empirical study on licensing, identifying a considerably high concentration of licensing in three two-digit SIC industries, chemicals, electronics and computers (SIC 28, 35 and 36 respectively). Although the authors are unable to explain the strong difference in licensing between industries within their data, they suggest differences in appropriability regimes between industries to be the reason.10
Nevertheless, there are certain deficiencies and downsides of licensing on markets for technology. Technology licensing can turn out to be unprofitable, since the process of identifying potential licensees is time consuming and costly (Lichtenthaler & Ernst, 2007a), whereas the transaction costs can grow to considerable amounts (Arora & Fosfuri, 2003; Arora et al., 2001; Teece, 1988). Caves, Crookell, and Killing (1983) identify several market failures in terms of licensing on markets for technology, including small-numbers bargaining, appropriability problems, uncertainty, transaction costs, and impacted information coupled with opportunism. However, it has to be mentioned that some of these early problems might have improved due to the increase of participating firms on the markets. Gambardella et al. (2007, p. 1180) also highlight that transaction costs in the area of licensing emerge from “harder-to-observe elements”,11 which they regard in line with the findings of Razgaitis (2004). Further, in a study focusing on organizational learning, Anand and Khanna (2000b) find that learning effects are stronger in joint ventures than for licensing contracts, which suggests ineffective communication between licensing partners.
2.2.3 The emergence of patent trolls
An increasingly important phenomenon on markets for technology representing one of the reasons for the proclamation of the AIA and the recent policy change towards weaker patent rights is certainly the activity of patent trolls, commonly known as non-practicing entities (NPEs), patent sharks or patent monetizers. These firms are characterized by their mere activity in enforcing patents against infringers to receive damages awards or settlement payments (Reitzig et al., 2007). Trolls are usually not actively using their patented technologies for “real” production or other practical purposes. Especially in cumulative technology industries, the patent thicket created by overlapping patents provides a fertile ground for the operations of patent trolls, as the thicket makes inadvertent infringement of patents likely (Reitzig, Henkel, & Schneider, 2010; Shapiro, 2001).
It is important to keep in mind that various authors define patent trolls differently. Bessen and Meurer (2008, p. 17) generally define patent trolls (not NPEs) as “individual inventors who do not commercialize or manufacture their inventions”, while Reitzig et al. (2007, p. 137) define patent sharks or trolls as “individuals or firms that seek to generate profits mainly or exclusively from licensing or selling their (often simplistic) patented technology to a manufacturing firm that, at the point in time when fees are claimed, already infringes on the shark’s patent and is therefore under particular pressure to reach an agreement with the shark.” The important difference between these two definitions is that Bessen and Meurer (2008) include pure research institutions, such as universities, research labs, or individual inventors in their concept of patent trolls, although these actors usually want to license their technology ex ante and actively conduct research. Reitzig et al. (2007) on the other hand exclude such actors from their definition. In our study, we follow the definition of Reitzig et al. (2007). We acknowledge the fact that the word troll includes a negative connotation, whereas for example “Non-Practicing Entity” (NPE) would be a more neutral expression. Nevertheless, as the definition of NPEs usually includes the actors discussed above, we decide to use the term patent trolls to preclude this misunderstanding.
Critics of patent trolls argue that they impede innovation12, cause a plethora of patent litigation cases and hold up other firms, while providing only minor contributions in terms of social welfare (Bessen et al., 2011; Davis, 2008; Reitzig et al., 2007). Reitzig et al. (2007) identify three strategies used by patent trolls. Firstly, in the damages-based strategy patent trolls try to obtain high damages awards in court.
1 Patent claims in this sense constitute the scope of a patent, it denoting what is and what is not protected by the patent.
2 According to Lemley (2001) and Merges (1999), patent examiners have, on average, a total of eighteen hours for the complete examination of a patent. Moreover, their salary increases with the number of granted patents, due to the huge amount of applications at the PTO, which incentivizes a fast granting process.
3 See Table 1, “Major Changes to the U.S. Patent System” in Hall (2004).
4 An exception to this finding is presented by the field of pharmaceuticals (Taylor et al., 1973).
5 A patent fence consists of multiple patents that cover similar innovations to a core patent, which “surround” the core patent and are filed to prevent competitors from “inventing around” or imitating the core invention.
6 The business model of patent trolls, which is strongly facilitated by these conditions is about to be discussed in detail under 2.2.3.
7 This decision from the 16th of June in 1980 enabled patenting of life forms, explaining a huge increase in biotechnology patents.
8 While Teece (1988) identifies high transaction costs for knowledge and technological information transfer, Nelson and Winter (1982) argue that organizational routines connected with technologies are difficult to transfer between companies. In a later paper, Teece (1998) requests the formation of more robust markets for intellectual property, as restrictions on knowledge transfers relaxed.
9 The remaining factors are the generality of the technology, the economic value of the technology, the fact if a technology is core or non-core to the venture, and competition.
10 This finding is in line with the results of Hall and Ziedonis (2001) and Cohen et al. (2000), as discussed in 2.1.2.
11 These elements include factors such as the inability to identify potential buyers, the difficulty in receiving internal approval to conclude the deal, disagreements on exclusivity or geographical restrictions (Gambardella et al., 2007, p. 1180).
12 The term “putting a tax on innovation” is frequently used in this context (Bryant, 2012; Magliocca, 2007).
- Quote paper
- Alexander Mosen (Author), 2016, Patent policy change in the USA. An empirical analysis of the effect on patenting strategies, Munich, GRIN Verlag, https://www.grin.com/document/380366