Do firms that earn revenues from licensing their patent portfolios, rather than producing physical products – often referred to as patent assertion entities (PAEs) or, more disparagingly as ‘patent trolls’ – frustrate or facilitate innovation? According to one view, PAEs purchase specious intellectual property and then file frivolous lawsuits in order to extort revenues from operating companies that would rather settle than go through the expense of litigation. The revenues earned by PAEs are therefore a tax on innovation: dollars that would be spent on R&D by operating companies are diverted to non-productive uses. A business model based on using low value patents to file frivolous lawsuits for their nuisance value, with little risk of being countersued, should yield easy returns—and a high yield, low risk business model that is not characterized by barriers to entry should proliferate rapidly. It follows, accordingly, that PAEs are a systemic threat to innovation, economic growth, and consumer welfare, and thus the laws governing patents and civil procedure should be reformed so as to make their business model unsustainable
An alternative view is that PAE’s create value by developing or purchasing patents which they then license to operating companies. They therefore provide a mechanism by which inventors can appropriate the returns to their investments, which allows the stock of new technologies to expand, thereby increasing innovation, economic growth and consumer welfare. The policy implication of this “intermediation hypothesis” is that reforms designed to make the PAE business model unsustainable will hinder innovation” because, if there are no intermediaries, there will be fewer inventive firms creating pioneer technologies.
This paper, which can be found here, seeks to conduct an empirical analysis of these hypotheses, to determine which one is correct. In short, the authors conclude that the facts do not seem to support the ‘patent troll hypothesis’: while there may be patent trolls, such firms do not seem to be large or numerous enough to constitute a systemic problem.
The paper is structured as follows:
Sections 2 and 3 define PAEs and seek to determine whether the facts support the patent troll hypothesis
The paper does so in two stages. First, it draws on existing literature to define the features of a harmful PAE (a “patent troll”). The defining characteristic of a harmful PAE is a business model characterized by negligible spending on R&D and the filing of frivolous lawsuits designed to earn easy returns. Calling these firms “patent trolls” is a shorthand way of saying that they extort payments from operating companies for low-value intellectual property by threatening expensive litigation.
From these characteristics, three testable implications of the patent troll hypothesis follow. The first is that patents held by trolls have little value other than as a means by which they can extract payments from operating companies. The second is that patent trolls are high-return, low risk enterprises. The third is that such firms are large and numerous.
Since there is no consensus on the definition of PAEs, the authors rely on an external expert (the RPX Corporation, a defensive patent aggregator and patent litigation insurer that has a strong incentive to identify PAEs in order to build its business model) to demarcate the universe of publicly-traded PAEs. They use RPX’s “Public PAE Reports” to identify 26 firms as publicly-traded PAEs. The authors find that the RPX-identified public PAEs do not appear to operate in a manner consistent with the patent troll hypothesis. As a group, the RPX-identified public PAEs spent nearly twice as much on R&D (as a percentage of revenues) than the weighted average of the 153 largest American high technology companies over the period 2011-16. RPX-identified public PAEs also do not appear to earn easy returns from filing nuisance lawsuits. If they were filing nuisance lawsuits using valueless intellectual property, then they would be highly profitable. The data indicate, however, that most of the RPX-identified public PAEs lose money – $3.1 billion over the period 2000-16.
Section 4 seeks to determine the negative effects that PAEs can have on the economy.
This section estimates the magnitude of the “innovation tax” that would be produced if all 26 RPX-identified public PAEs were actually harmful PAEs (patent trolls). The RPX-identified public PAEs operate in high technology markets. The authors estimate both the size of the U.S. high technology market and the total annual revenues of the 26 RPX-identified public PAEs. Lastly, they also estimate the litigation costs that the RPX-identified public PAEs might have imposed on other firms.
In short, the paper finds that the RPX-identified public PAEs as a group are too small to have much effect—either positive or negative—on the U.S. high technology sector. The total size of the transfer from the U.S. high technology sector to the RPX-identified public PAEs (their revenues, plus an estimate of the litigation costs they might impose on other companies) averaged only 0.28 percent of the high technology sector’s revenues in 2011-16. Given standard assumptions about elasticities of supply and demand in consumer products, the deadweight loss imposed by the RPX-identified public PAEs would be about 0.003 percent of the revenues of the U.S. high technology sector. In other words, any intermediation PAEs may perform for the U.S. high technology sector is too small to have any meaningful effect.
Section 5 seeks to determine whether there are facts about privately-held PAEs that might challenge the conclusions reached above.
To overturn the conclusions arrived at in section 4, privately-held PAEs would have to behave and perform very differently from the RPX-identified public PAEs – they would have to (i) spend much less on R&D or patent acquisitions than the average high technology company; and (ii) be profitable, low risk enterprises. Additionally, privately-held PAEs would also have to be large and numerous to approach the size of the RPX-identified public PAEs as a group.
To assess this the authors, while acknowledge that their analysis on this point could go deeper, assess publicly-available evidence about three large, privately-held patent licensing firms that are often referred to as NPEs, PAEs, and/or patent trolls. They find that their behaviour and performance is not unlike the RPX-identified public PAEs. They also find that the addition of the estimated revenues of these three firms, plus an estimate of the litigation costs they might impose on operating companies, would not substantially change their estimates of the transfer and litigation costs imposed by the RPX-identified public PAEs. It follows that the introduction of privately-held PAEs into the analysis would not substantially change the magnitude of any potential deadweight loss that would occur if PAEs were levying, in fact, an innovation tax.
Section 6 concludes.
The first takeaway of this study is that the magnitude of the transfers to the RPX-identified public PAEs plus the legal costs that might be associated with them are small compared to the size of the market in which they operate. This fact is not consistent with the hypothesis that PAEs impose a significant “innovation tax”. However, the small size of these firms as a group is also inconsistent with the hypothesis that they are important intermediaries in the market for innovation.
The second takeaway is that the RPX-identified public PAEs, as a group, do not seem to be earning economic rents. The third takeaway of the study is that the RPX-identified public PAEs tend to spend heavily on R&D. It follows that the claims about the business model of PAEs made by FTC, as well as some of the academic literature, require revision.
This is one of those papers which is well beyond my ability to comment critically on. I can nonetheless say that it provides a useful overview of the debate regarding PAEs, and does something that I consider extremely valuable: it seeks to provide an empirical basis for ongoing (academic and policy) debates.