Does the law of innovation work against itself?
|Written by Ron Katznelson, Ph.D.
President, Bi-Level Technologies
Posted: July 22, 2014 @ 9:59 am
A recent research report by Professor Catherine Tucker, titled The Effect of Patent Litigation and Patent Assertion Entities on Entrepreneurial Activity has received wide publicity. The report purports to document suppression of Venture Capital (VC) investments due to litigation by so-called “patent trolls” – or Patent Assertion Entities (PAEs), which Professor Tucker identifies as “frequent patent litigators” – patentees that have caused more than 20 “instances of a patent being litigated” in a span of 17 years (p31-32). Professor Tucker arrives at a startling conclusion: VC investment would have been about $22 billion higher over the course of five years “but for litigation brought by frequent patent litigators” (p36) – suggesting that enforcement of patents can frustrate the goal of the Patent Act. In other words, the law of innovation ostensibly works against itself.
A footnote in the report’s front page discloses that the source of funding for this particular study was provided by an organization that has a substantial interest in the study’s outcome – the Computers and Communication Industry Association (CCIA). It may come as no surprise that Professor Tucker’s conclusions happen to fall in line with this organization’s previously-articulated beliefs that patents block new technology development. It is also noteworthy that Google, a prominent member of the CCIA, has also provided Professor Tucker with $155,000 of funding since 2009. In a June 2014 press release announcing this study, CCIA attorney Matt Levy proclaimed:
“For the first time, we have an economic model proving patent reform is not a zero sum game between protecting intellectual property and reducing abusive patent litigation … Professor Tucker’s research reveals the harms of skewing the patent system too far in favor of protecting low-quality patents.”
For her part in the press release, Professor Tucker stated:
“My analysis showed that litigation by patent trolls did not foster entrepreneurial investment at all. It is instead associated with significantly reduced venture capital investment in entrepreneurship, preventing startups from developing and suppressing job growth.”
If true, these purported economic findings are no trivial matter to be consigned to a mere working paper and an industry association press release – they must be of major significance and deserve substantial attention in our innovation policy discourse.
There is nothing wrong with academics obtaining research funding from corporate sponsors, and to Professor Tucker’s credit, she fully discloses her funding sources. It is important, however, to fully vet and scrutinize the resulting work product of such sponsored studies. The underlying data and methodology should be independently verified by subject-matter researchers and experts. Yet, no peer review of this paper had been made before Professor Tucker’s “results” were broadcast to the world. Because of the publicity this paper received, it is the purpose of my full paper to fill that void.
Indeed, closer scrutiny of this article reveals that it can support none of the claims of the author or sponsors quoted above. This is because the study is fundamentally flawed and is fraught not only with methodology and analytical errors, but with fatal factual errors due to the assumption of false information.
Before exposing these fatal factual errors, it is important to note as a threshold matter that Professor Tucker’s study lacks information or evaluation of the merits of the patent infringement cases included in her study. The patents, or their “quality” are not identified and cases are categorized solely based on whether they were brought by “frequent patent litigators.” The author purports to study the “harmful” effects of PAEs and does so solely through categorizing cases by the patentees’ frequency of litigation. The report defines a “frequent litigator” as an entity that has filed twenty or more patent lawsuits during the 17-year study period (p15), but admits that some cases were brought by practicing entities (p16).
Professor Tucker introduces the motivation for her study by presenting what she calls three “case studies,” involving startups that were sued for patent infringement. The introduction recounts the alleged setbacks and harms that befell these startups as a consequence of the patent litigation against them, including reduced investments in their product development. The study apparently seeks to establish a broad-based relationship between the intensity of patent lawsuits against identified startups and the identified VC investment levels in these startups. This, however, the study fails to accomplish because of fatal flaws in methodology. The basic flaws are in two categories: assuming false information, and modeling. In this summary I address further only the first category.
Assuming false information
In perhaps the most astonishing methodological failure, Professor Tucker’s study does not actually identify or connect any defendants in the lawsuit dataset with any of the small entity firms in the dataset that are potential recipients of VC investments. This is a basic relational infirmity at the firm level. The author goes to great length to isolate only small entities with 50 or fewer employees for the firm dataset. Yet, the paper provides no indication that any of these firms were in fact defendants in the patent suits collected in the dataset. For all we know, the patent suits in the study may have all been filed against entities with more than 50 employees – a class of defendants that is presumably out of the scope of the study.
Flawed reliance on counterfactuals
Realizing perhaps that there are relatively very few cases brought against startups and that she cannot make the connection on a firm level referred to above, Professor Tucker attempts instead a “regional” connection, but introduces a fatal methodology flaw by making a patently false assumption (pun intended): that the litigation venue and the defendant’s home district are one of the same; i.e. that a patent lawsuit filed in a given district names a defendant actually located in that district – a defendant that is potentially eligible for VC investments made in that district.
Professor Tucker actually acknowledges this flaw, explaining that “we are assuming there is a relation between the district where the patent lawsuit was filed and the region where VC investment is received by entrepreneurial start-ups” (p17). However, nothing is done to rectify this flaw and it is instead summarily dismissed by making another counterfactual assumption: that by excluding only the Eastern District Court of Texas and the District Court of Delaware from analysis, this somehow leaves a dataset with correct relations for all other districts. This assumption is fatally incorrect. My paper discusses much evidence that the disconnected relationship of litigation venue and defendants’ home region pervades litigation in all districts.
The relation between the district where the patent lawsuit was filed and the region where VC investment is received by start-ups is a central predicate of this study – the results are invalid due to an incorrect relationship. We know that this relationship is certainly wrong for one out of every five cases – cases with foreign defendants with their domicile in none of the U.S. districts. These cases were not removed from the analysis dataset. In conclusion, Professor Tucker’s regional framework – the crux of her thesis – is based on false assumptions and incorrect information; it can only produce random results that have no bearing on her thesis.
Study framework that contradicts the operating framework of the patent system
Even if one were to assume that a relation between a patent lawsuit venue and the defendant’s home district exists in her dataset, which is not the case, Professor Tucker’s hypothesis that one could expect regional VC investment preferences to be affected by patent litigation intensity in the region is predicated on another illogical counterfactual: that a kind of a patent enforcement “regional deterrence” effect can be detected – that even if filed against established large firms in the district, patent suits filed in a given district deter VC investments in startups located in that district but not in other districts.
This implicit patent “regional deterrence” hypothesis has no business or legal basis because patent litigation threats are not regional – patent law is a Federal law that applies across the nation. A patent lawsuit against a firm in Silicon Valley would subsequently be as threatening to a firm in Boston as it would be to another firm in Silicon Valley. There are no regional barriers for patentees’ enforcement rights; a legal determination by one Federal district court on a given patent and on an infringing instrumentality is binding on users of the same instrumentality in all other districts. Professor Tucker’s implicit notion of “regional deterrence” may apply to some other wars but not to patent wars.
Grossly incomplete accounting of entrepreneurial investments
Finally, it is worth stepping back to note that Professor Tucker’s focus solely on purported attenuated investments by defendants accused of patent infringement misses a major component of the total investment in entrepreneurial innovation: it is the major investments in patentable inventions that are largely incentivized by the exclusive right under the patent law. In many instances, these are risk investments that would not occur but for the assurances under patent law that exclusivity can be maintained and infringers can be stopped. Without investigating the legal merits of the claims brought against alleged infringers, Professor Tucker appears to adopt an unsupported baseline hypothesis: that the defendants in her patent cases brought by “frequent litigators” are subject to “abusive patent litigation” rather than meritorious patent claims to stop infringement of valid patents.
Thus, Professor Tucker’s working premise is that these patent infringers in her study were unjustly discouraged from investing in development of infringing products; that they were denied opportunities to raise more capital and hire more employs to continue to develop and sell infringing products. It may come as a shock to Professor Tucker, but stopping infringement is precisely what the patent system is designed to do. The patent system is designed to reduce and suppress investments in infringing products that are not coordinated with the patent holder; its purpose is to discourage infringement and encourage inventions around existing patents, thereby promoting the progress in useful arts. The patent system was created to do all that in order that investment capital continues to flow to develop and commercialize patentable inventions. Professor Tucker totally ignores this critical side of the entrepreneurial investment equation, ignoring the fact that much entrepreneurial investment is undertaken because of the power of patents to stop infringement. Apparently, some of Professor Tucker’s “case study” startup defendants are simply bona-fide infringers who carelessly failed to conduct the rudimentary patent due-diligence prior to entry and are consequently denying fair returns on these other investments made in entrepreneurial firms who hold patents. See the full paper for a discussion of an exemplary “case study.”
Professor Tucker’s study makes no distinction among patent lawsuits, the merit of their underlying claims, or whether they were ruled abusive. It focuses instead on patent owners – whether they are “frequent patent litigator.” It inexplicably presumes that alleged infringers sued by “frequent patent litigator” were somehow victims of baseless patent suits, despite the fact that all patents are presumed valid (35 U.S.C. § 282), whether litigated once, 20 times, or never. Professor Tucker’s effort to detect adverse patent litigation effects on VC investments is flawed because there is no relationship in her dataset between the defendants and startups eligible for the VC investments in her dataset. Nor is there even a relationship in her dataset between the district where the patent lawsuit was filed and the region where VC investment is received by start-ups. The results are invalid because they are based on a false assumption that the litigation venue and the defendant’s home district are one of the same. In addition, model employed is designed to produce a predetermined desired result because it does not include a key independent variable related to VCs preference for investment in their local region.
Professor Tucker concludes that she observed “high levels of patent litigation” – a “problem” for which she appears to have a solution – “one way to reduce litigation would be to decrease the number of patents granted” (p38). Based on her purported findings, Professor Tucker suggests that legislation with “policies directed at repressing the activities of PAEs or frequent litigators” may encourage VCs to increase investments in entrepreneurial activity (p40-41). It is perhaps significant that the VCs themselves strongly disagree with this notion: the National Venture Capital Association (NVCA) objected to patent “reforms” along the lines suggested by Professor Tucker or by her research sponsor the CCIA, precisely because they would harm investors’ portfolio companies. These objections arise perhaps because the NVCA and other innovation industry stakeholders recognize that, contrary to Professor Tucker’s thesis, the law of innovation does not work against itself.
Read the full paper at http://ssrn.com/abstract=2468193.