Patent Trolls in the United States
- 1 Patent Trolls in the United States
- 1.1 Introduction
- 1.2 The Role of Intermediaries in the Patent System
- 1.3 “Functional Claiming” and Uncertain Infringement
- 1.4 The Economic Cost of Patent Troll Activity
- 1.5 Range of Victims
“PAE” was defined as “an entity that does not offer products/services” and makes a “demand” regarding patents (Chien 2012). See patent assertion entities, patent troll definition and patent troll legislation.
The purpose of the U.S. patent system, according to the Constitution, is “to promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries” (U.S. Const., art. 1, sec. 8, cl. 8). Giving inventors this right provides a powerful incentive for innovation.
Patent policy must navigate a fine line however, as excessive enforcement of that exclusivity—
such as through abusive litigation or overly broad patent claims—may dampen incentives for
future innovation. Innovators who fear inadvertently infringing existing patents may reduce
innovative activity or take costly steps to defend against lawsuits claiming infringement, leading to fewer resources available for wages, job creation, and innovation of new products and services.
Firms that own patents but do not practice (Firms that “practice” their patents use them to design or manufacture products or processes) them can play a useful role in the innovation ecosystem. Firms that aggregate and manage patents can play an important intermediary role, bringing value to society by more efficiently matching inventors to patent users in an otherwise
illiquid market, and by developing expertise in legitimately protecting patents from infringement. However, some litigation strategies may reduce incentives to transfer or commercialize technology by unwarrantedly raising potential innovators’ fears that they will be accused of patent infringement if they do so.
This entry looks particularly at firms who do not practice the patents they own and instead
engage in aggressive litigation to collect license and other fees from alleged infringers. A review of the evidence suggests that on balance, such patent assertion entities (PAEs) (also known as
“patent trolls”) have had a negative impact on innovation and economic growth.
The success of the PAE business model in part reflects patent policy challenges created by the
rapid growth of complex software products. Because of rapid technological change and the special characteristics of software, it has been hard to define clear boundaries for patents, and hard to set an appropriate bar for non-obviousness, leading to many opportunities that PAEs (and in some cases, non-PAEs) have exploited.
The Role of Intermediaries in the Patent System
“Functional Claiming” and Uncertain Infringement
To be awarded a patent, an inventor must disclose the invention in sufficient detail to enable
skilled practitioners in the relevant field to understand it and potentially build upon it. Patent
applicants must also articulate the specific claims as to the scope of the patent. Understanding
which products and processes are, in fact, protected by the patent is essential to avoiding
infringement upon that patent. Moreover, such clarity enables patents to serve the socially
beneficial purpose of promoting technology transfer. The Patent and Trademark Office grants
patents only if the claims are novel (have not been made before) and are not obvious to a person skilled in the relevant art.
Setting an appropriate bar for novelty and non-obviousness is particularly important in a new field; if the bar is not set high (something difficult to do in a new field), firms may well find themselves inadvertently infringing patents, both because of the sheer number of patents and because commercial need is driving many inventors to create similar inventions near-simultaneously (Lemley and Melamed 2013). Many practitioners of such technologies (such as railroads in the 19th century and software today) find it more profitable to focus on expanding the overall market for their products by technological cooperation with rivals, rather than working to clearly delineate property rights (Boldrin and Levine 2013).
An additional reason that the issue of overbroad patents is particularly salient in software is due to the prevalence of “functional claiming” in these patent classes (Lemley 2012). A claim term is “functional” when it recites a feature by “what it does rather than by what it is” (In re Swinehart 1971). (For example, functional language is often used to add further description to a structure or step, e.g., a claim may recite a conical spout (a structure) that allows several kernels of popped popcorn to pass through at the same time (a function). In re Schreiber, 128 F.3d 1473, 1478(Fed. Cir. 1997)).
Functional claiming involves claiming exclusive rights over any device that performs a given
function, regardless of how that function is performed.
Functional language can therefore lead to very broad and/or vague claims. These problems are
especially acute for software patents. For these patents, it has been argued that the code is the
function, with the implication that a software patent arguably excludes any other code that
performs that same function. In contrast, in pharmaceuticals, the distinction between a function
and the means used to perform that function is generally clear. For example, several patents have been awarded for the function of reducing cholesterol; each patent covers a different chemical compound—a different means of providing that function.
Compounding the problem is the fast-moving, interdependent nature of technical change in the
software industry. Functional claims can be used to ‘over-assert’ a patent by attempting to cover
products and processes that were never contemplated by the inventor or the examiner as being within the claim scope at the time of the invention. For example, a patent claim about a
programmed processor could be asserted broadly to cover any and all devices that achieve the
claimed result, rather than being limited to a device programmed with the specific software used by the inventor.
In addition, a single piece of software or website might have several thousand “functions” that
could be claimed in as many patents. It is also difficult for an outsider to judge what an inventor
meant by a claim and to know what sort of invention would be “obvious” to a skilled practitioner
and thus unworthy of a patent. For example, in the case discussed earlier, the appeals court had to consider detailed features of twenty-year-old technologies to determine whether the shopping basket patents in fact made novel claims.
Thus, it can be very difficult to know if one is infringing patents. These broad, functionally-defined, and intertwined patents are therefore a key part of the patent troll business model. These intermediaries acquire broad patents and threaten suit, in hopes of extracting settlements. If even one patent in a complex product is held to be infringed, the product cannot be legally sold (Lemley and Shapiro, 2005). This situation can lead to problems for practicing firms both large (note that a single smartphone may read on over 100,000 patents) and small (the basis of the demand letters discussed in the examples above is the alleged interaction between components of a computer network found in most offices). The stakes are particularly high when the venue for an infringement dispute is the U.S. International Trade Commission (ITC), given the ITC’s inability to award damages and reliance instead on exclusion orders barring import of products deemed infringing into the United States.
The Economic Cost of Patent Troll Activity
While most patents are not litigated and are properly practiced and enforced, the harassing
litigation tactics of some patent trolls, combined with substantial litigation costs (ranging from a median of $650,000 for smaller cases, to a median of over $5 million per case where the amount in controversy exceeds $25 million) (AIPLA 2013), have added significant costs to the innovation ecosystem and sapped investments in research and development, causing great harm to society. These costs are of several types, as showed below.
Direct costs to firms that practice patents
James Bessen and Michael Meurer (2012) find that defendants and licensees paid patent troll’s $29 billion in 2011, a 400% increase from 2005; they estimate that less than 25% of this money flowed back to innovation. In their papers, Bessen and Meurer define PAEs as firms with each of the following characteristics: they “do not produce goods, rather they acquire patents in order to license them to others,” they “seek to derive the majority of their income from the enforcement of patent rights,” and they file lawsuits.
In addition, in the majority of patent troll cases, the legal cost of the defense exceeds this settlement or judgment amount (Chien 2012c).
Private costs of lost opportunities to commercialize technology
One might argue that the losses to defendants accused of infringement would be offset by gains to the owners of patents. However, very little such transfer of value appears to take place. For example, in the years 2000 through 2010, a set of fourteen publicly-traded patent troll followed by Bessen, Meurer, and Ford (2011) had total revenues of $7.6 billion. These revenues may include revenues from sources other than litigation, and therefore may overstate the value of transfers from defendants to these patent trolls. Note that the $7.6 billion does not include payment streams received after 2010 related to settlements won during the study period. Future payment streams are unlikely to be large given that settlements tend to be paid in lump sums. Also, the “event study” method used by the authors controls for the impact of the recession on firm valuations, because the method looks at changes in a firm’s share value around the time of a lawsuit filing.
Patent suits initiated by those fourteen entities were associated with a decline of $87.6 billion in defendant company share value over the same period, implying that the financial award experienced by winning PAEs amounts to less than 10% of the lost share value in this sample.
While drops in the share value of a defendant companies may reflect other economic factors (e.g. the now-raised expectation of losing future suits or making settlement payments), the 90% of lost defendant share values that simply vanishes suggests considerable lost value to society from forgone technology transfer and commercialization of patented technology. Aggregating to all suits by patent trolls yields lost wealth of over $300 billion in four years starting in 2007. That is, the stock market values the lost opportunities for technology commercialization as significantly greater than the direct payments from defendants and licensees to PAEs.
Even if patent assertion entities do not prevail in the courtroom, their actions can significantly
reduce incremental innovation while litigation is ongoing, a situation that can persist for years. The reason is that such action could be viewed by courts as an evidence of “willful infringement” if the plaintiff’s patent is upheld, making the firm liable for treble damages. For example, one study found that during the years they were being sued for patent infringement by a PAE, health information technology companies ceased all innovation in that technology, causing sales to fall by one-third compared to the same firm’s sales of similar products not subject to the PAE demand (Tucker 2013).
Social costs of reduced innovation
A great deal of economic literature shows that firms do not capture all the value created by the research and development they do (Mansfield 1968). Thus, the losses caused by excessive litigation exceed even the large stock market losses described above, including lost value to consumers who are not able to buy innovative products, and reduced income for workers whose pay is lower because they are unable to work with more productive new processes.
Range of Victims
Although patent trolls often target major, household-name and deep-pocketed technology companies, they also target start-ups and small companies. In fact, though the most substantial settlements are often extracted from large entities, the majority of PAE suits target small and inventor-driven companies (Bessen and Meurer 2012).
Recent surveys provide evidence for the negative impact of PAE litigation on innovative
companies. The impact on smaller startups is particularly acute. In a recent survey of 223
technology company startups, 40 percent of PAE-targeted companies reported a “significant”
operational impact (e.g. change in business, exit from the market, delay in milestone, change in
product, etc.) due to the suit or threat thereof (Figure 2). “PAE” was defined as “an entity that does not offer products/services” and makes a “demand” regarding patents (Chien 2012).
In another recent survey of 116 inhouse counsels, primarily from firms with over $100 million in annual revenue, nearly all firms reported that patent troll demands had affected them financially or distracted them from their core business, with nearly 40 percent stating that patent troll activity had led them to make changes to an underlying product (McBride 2013).
PAEs have also sent infringement notices to “downstream users” of technologies, who are often
small companies, as in the scanner and Wi-Fi cases discussed above. Although the amount of
money extracted from each company is small, the number of potential defendants makes this
strategy potentially profitable overall.
Aggressive litigation tactics have also been adopted by some firms that practice their patents. The recent spate of patent litigation among large technology companies—termed the “smartphone patent wars” by the press—typifies this behavior, which also involves companies purchasing massive numbers of patents as a defense against of litigation, or as leverage in negotiating licenses with competitors.
Between $15-20 billion was spent on patent litigation and patent purchases in the smartphone
industry from 2010- 2012 (Lemley 2012). And in 2011, spending by Apple and Google on patent
litigation and patent acquisitions exceeded spending on research and development of new
products, according to public filings (Duhigg and Lohr 2012). In 2012, Google spent $12.5 billion to buy Motorola Mobility and its patents and $5.2 billion in 2011 on research and
development (R&D) Google, Inc., Annual Report (Form 10-K) 6 (Jan. 26, 2012), available at
http://www.sec.gov/Archives/edgar/data/1288776/000119312512025336/d260164d10k.htm. In 2011, Apple spent $2.4 billion on R&D but contributed more, approximately $2.6 billion, to a single transaction to buy patents from Nortel. Apple, Inc., Annual Report (Form 10-K) 7 (Oct. 26, 2011), available at http://www.sec.gov/Archives/edgar/data/320193/000119312511282113/d220209d10k.htm. Indeed, Google’s $12.5 billion purchase of Motorola, according to its own statements, was undertaken in large part to prevent patent suits from competitors (Womack and Tracer 2011).
“Defensive” purchase of patents has come under scrutiny by the Federal Trade Commission and
the Department of Justice for potentially anti-competitive behavior.(See justice.gov/atr/public/press_releases/2012/280190.htm). In one illustrative case, Apple and Motorola engaged in protracted legal wrangling over whether Motorola’s royalty
requests were reasonable given that the technology was “standard essential,” i.e. required for a
standardized technology to function. In some technical standards-setting situations in which a
patented technology is being considered for inclusion in a standard (such as Wi-Fi), a patentholder may agree to offer licenses for the technology on “fair, reasonable, and nondiscriminatory” (FRAND) terms, in return for gaining access to the broad market that having a standard potentially creates.
When standards incorporate patented technologies, owners of those patents benefit from
expanded marketing and licensing opportunities, while the public benefits from products
embodying the best technical solutions. However, a product that complies with such a standard
will necessarily read on these patents, creating a potential incentive for patent owners to raise the price of a license after the standard is set. In early 2013, the Department of Justice and Patent and Trademark Office issued a joint policy statement on the implications of this phenomenon for enforcement at the International Trade Commission (ITC). (justice.gov/atr/public/guidelines/290994.pdf). Also in 2013, the FTC settled with Google, issuing a consent decree in which Google agreed to honor Motorola’s prior commitments to license standard-essential technologies on FRAND terms (Federal Trade Commission 2013).