Deepa Varadarajan's new article, The Trade Secret-Contract Interface, published in the Iowa Law Review, explores the role of contracts in trade secret law. This article returns to an issue that remained unresolved following rich exchanges between Robert Bone and other scholars such as Michael Risch and Mark Lemley. Varadarajan's article is a welcome follow up.
The Prior Debate: Does Trade Secret Law Have An Independent Justification? Prior scholarship probed whether trade secret law is justified separate and apart from contract law. Contract law operates on the notion that parties who enter into an enforceable agreement with one another can sue a counter-party who breaches and obtain damages in an amount that approximates the benefit of the bargain. In contrast, the owner of a "trade secret"—information that is not generally known and has been the subject of reasonable efforts to maintain its secrecy—can sue another who misappropriates the trade secret, regardless of whether they are in a contractual relationship. The trade secret owner can obtain an injunction to protect the sanctity of the secret, not just money damages.
In his 1998 article, A New Look at Trade Secret Law: Doctrine in Search of Justification, Robert Bone argued that trade secret law lacks a theoretical justification distinct from other legal theories, in particular contract law. When trade secret law reaches beyond contractual liability, Bone wrote, trade secret law subsists in a "normative vacuum that continues to remain unfilled." Within that vacuum, judges "view trade secret law as a relatively open-ended delegation of authority to police the morality of commercial relationships." (Bone, 245).
In the wake of Bone's bombshell, several academics sought to justify trade secret law. Two of the most oft-cited papers seeking to do so were those of Michael Risch (2007) and Mark Lemley (2008).
Several justifications for trade secret law emerged from these discussions. The seeds of most of them can also be found in the Supreme Court's crucial holding in Kewanee Oil Co. v. Bicron Corp. (1974), where the Court held Ohio's trade secret statute was not preempted by federal patent law.
(1) Promoting Innovation Beyond Patents First, trade secret protection, like patent protection, serves to promote innovation, including patentable as well as unpatentable innovation. As the Court put it in Kewanee, "[t]rade secret law encourages the development and exploitation of those items of lesser or different invention than might be accorded protection under the patent laws, but which items still have an important part to play in the technological and scientific advancement of the Nation."
(2) Promoting "Disclosure" Second, trade secret law, like patent law, promotes disclosure of information. But while patent law encourages public disclosure in patent specifications, trade secret law encourages "disclosure" within firms (between employers and their employees and between employees) and between firms (especially between licensors and licensees of protected information). As the Court explained in Kewanee, without trade secret protection,
"[t]he holder of a trade secret would not likely share his secret with a manufacturer who cannot be placed under binding legal obligation to pay a license fee or to protect the secret. The result would be to hoard rather than disseminate knowledge."
(3) Limiting Expenditures On Costly Precautions Third, trade secret law reduces the costs of self-help. In Risch's words, trade secret law provides "incentives for businesses to spend less money protecting secret information or attempting to appropriate secret information." (Risch, 44). Back in 2007, Risch gave a helpful "real world example" from his law practice.
"A client proudly told me about how his company’s new development center in China was set up to protect trade secrecy: fingerprint scanners, almost no Internet access, expensive network filtering appliances to scan outgoing email, special locks on the computers... and so forth. These expensive and potentially efficiency-reducing measures are not installed in the United States (though here there are other more “standard” precautions). The only difference in the client’s decision making is the perceived inability to enforce trade secret rights in China."
(4) Channelling Subject Matter Into, And Out Of, The Patent System Lastly, trade secret law, in Lemley's telling, helps to "channel" subject matter between the domains of trade secret protection, which requires retaining near-absolute secrecy, and patent protection, which requires near-total public disclosure. Lemley implied that channelling can only effectively occur if trade secret law is conceived as a property-like regime with strict subject matter restrictions, including first and foremost, secrecy.
But none of this convinced Bone. In his 2014 follow-on article, The (Still) Shaky Foundations of Trade Secret Law, Bone remained convinced of the fundamental soundness of his earlier critique, writing:
"I have learned much from this work. But it does not convince me that broad legal protection for trade secrets is justifiable. I remain skeptical that there is a normative basis for a freestanding trade secret law that is not parasitic on other legal norms."
(Bone 2014, 1804).
Bone's response (and thanks to Jeanne Fromer for reminding me to read it!) is particularly relevant as we move into a new world of federal civil trade secret law under Title 18. A few of Bone's key critiques are worth highlighting, as they could prove useful areas for empirical study:
"[I]t is not clear how much additional incentive trade secrecy adds [as an incentive to innovate], given that patent law already provides indirect incentives for nonpatentable inventions ... and .... firms already have market incentives to develop nontechnological, commercial information." (1812).
"[Trade secret protection] diverts investment away from patentable inventions by enhancing the private value of nonpatentable ones...[and] frustrates the disclosure goals of the patent system when firms opt for trade secrecy to protect patentable inventions." (1813)
"[Contrary to what the Supreme Court stated in Kewanee,] trade secrecy does not publicly disclose inventions. It discloses to a contracting party but only under confidentiality constraints. That party learns the invention and might profit from the general knowledge in future work. But it cannot teach the information to others." (1814).
"Defenders of trade secret law continue to insist that trade secrecy can be justified by its salutary effect on the precaution-stealing arms race. However, none of the more recent arguments add all that much to the analysis. Indeed, some treatments simply reassert the benefit without addressing any of the detection or litigation-cost problems [i.e. that the option to sue for trade secret misappropriation comes with its own costs.]."
All of these critiques remain subjects of controversy. If Bone is (still) right that trade secret law is not justified except as a "parasite" on other legal norms, or at least not quite enough to justify the costs of introducing it, then the new federal trade secret law subsists in a "normative vacuum" too.
Varadarajan: Maybe Contracts Are The Boogeyman, Not Trade Secrets
And now we come to Varadarajan's new article, The Trade Secret-Contract Interface. It's not exactly a "bombshell," but it does pry open a very important sliding door that hopefully others will enter. (Forgive the metaphors; it's been a hot summer).
The subject of the prior debate was whether trade secret law is justified independently from more traditional legal regimes like contract. In her new article, Varadarajan builds on this scholarship. But she enters the discussion from a different angle. Whereas Bone and the others debated how to justify trade secret law apart from contract law, Varadarajan's purpose is to probe, and ultimately criticize, the pervasive role of contracts within, and around, trade secret law.
As mentioned above, contract law can be far weakerthan trade secret law. This is because trade secret law acts in places where no contract exists, and provides a far blunter remedy. "Efficient breach" is not possible in trade secret; courts may impose injunctions even in cases where this leaves the trade secret owner with more than she bargained for. (To give one example, courts, including in my state of Ohio, may impose lengthy "headstart" injunctions "in order to eliminate commercial advantage that otherwise would be derived from the misappropriation.").
Take the example of the high profile Waymo/Uber lawsuit, in which Waymo sued Uber in order to stop Uber from acquiring and using trade secrets that Uber employee Anthony Levandowski allegedly obtained while working for Waymo. (The case subsequently settled.) Waymo never got Levandowski to sign a non-compete agreement and could not have done so given the California ban on non-competes, but using trade secret law Waymo might have obtained a permanent injunction preventing Levandowski from working at other firms like Uber, or even from working on self-driving cars period. Waymo never entered a contract with Uber at all; they were not in privity. Yet under trade secret law, Waymo could reasonably argue that Uber misappropriated trade secrets by hiring away Levandowski with knowledge that he came onboard with "14,000 highly confidential and proprietary files" containing Waymo information. If Waymo had succeeded at trial, it could potentially have enjoined Uber's entire self-driving car operation for years into the future.
Try doing that through contract law.
Yet at the same time, contract law can offer stronger protection for information than trade secret law. A principle reason is that contract protects a broader scope of information than trade secret law does. Waymo, for instance, could have accomplished a good deal using contracts. What if the information Levandowski took wasn't really a trade secret? Waymo still could have sued Levandowski for breach of contract with respect to "confidential," if not actually "trade secret," information he obtained while working at Waymo. Even without the ability to make Levandowski sign a non-compete agreement, Waymo could, and presumably did, have Levandowski enter a variety of contracts as a precondition to employment, ranging from non-disclosure to non-solicitation agreements. (On this point, see Jonathan Barnett and Ted Sichelman's working paper, arguing that companies in states that don't enforce non-competes have plentiful alternatives for limiting post-employment competition.) In addition, Waymo could have contracted with Uber and other competitors to agree not to "poach" each others' employees. True, the legality of anti-poaching agreements is seriously questioned. But the New York Times just reported that until recently major fast food companies like Arby's, Carl's Jr., and McDonald's used controversial "no-poach" clauses all the time, at least among franchises.
Varadarajan identifies two distinct functions that contracts play within trade secret law.
The Evidentiary Function of Contracts The first function of contracts play within trade secret law is evidentiary. When courts assess a trade secrecy claim, they use contracts in two ways. (Varadarajan, 1557-1559).
First, contracts inform precisely what subject matter is protected. U.S. trade secret law requires the owner of a trade secret to take "reasonable measures to keep such information secret[.]" 18 U.S.C. § 1839(3). Contractual protections are often crucial for establishing reasonable secrecy precautions. As Varadarajan notes, citing studies by David Almeling and his colleagues, “confidentiality agreements with employees and business partners are the most important factors in the courts’ analysis of reasonable measures.” (1557) (citing David S. Almeling et al., A Statistical Analysis of Trade Secret Litigation in State Courts, 46 GONZ. L. REV. 57, 69 (2011)). Thus, when putative trade secret owners fail to use contracts to protect their information, by requiring non-disclosure agreements when exchanging information with manufacturers and by requiring confidentiality contracts when hiring employees, they are more likely to lose in court.
Second, contracts inform what conduct is permissible or impermissible under trade secret law by helping courts determine whether there is a duty of confidentiality for purposes of establishing misappropriation. U.S. trade secret law does not impose liability for disclosing or using another's information unless the person does so using improper means (e.g. breaking into a factory) or in breach of a "duty to maintain the secrecy of the trade secret." 18 U..S.C. § 1839(5). There is not always a written contract indicating that a person owes a duty to maintain the secrecy of given trade secrets. In those cases, courts may nonetheless "imply" a contract, either as a matter of fact ("implied in fact") or as a matter of law ("implied in law"). Varadarajan argues courts are "wary" of implying contracts in the absence of good evidence, though, perhaps notably, she only cites a around five cases for this proposition, and one "but see." (1559-1560, nn. 88-89). This is important to her argument because if courts find "implied" duties to maintain secrecy everywhere, then contract isn't doing as much work as Varadarajan suggests in establishing liability.
This first part of the article (Parts I, II, and III) is largely descriptive. Varadarajan's contribution is simply to explain what trade secret lawyers and students already know: that in practice courts look to contracts as evidence that trade secrets exist and that defendant was under a duty with respect to those secrets. But Varadarajan also provides a simple explanation for contracts' utility in both establishing the existence of a trade secret and in establishing whether misconduct occurred. Drawing on her prior work, she argues that contracts, assuming they are clearly drafted and suitably represented to all relevant parties, provide "notice" as to what is protected, and what is not. "The pivotal evidentiary role of contracts in trade secret cases," she concludes, "stems largely from [contracts'] notice function." (1560).
Varadarajan is right to focus on notice. In patent law, patent disclosure and the claims in particular play an important notice function, alerting others, including potential infringers, of the patentee's rights and of the possibility of future litigation. See Nautilus, Inc. v. Biosig Instruments, Inc., 134 S.Ct. 2120, 2129 (2014) ("[A] patent must be precise enough to afford clear notice of what is claimed, thereby "`appris[ing] the public of what is still open to them.'") (citations removed). But in the trade secret context, the people who encounter trade secrets may not always know or even suspect that a trade secret exist. "Unlike patent law," Varadarajan writes, "trade secret law does not require putative trade secret owners to formally apply for or describe the boundaries of asserted trade secrets." Thus, the trade secret defendant "does not even learn of the existence or precise boundaries of the trade secret she is alleged to have violated until she becomes the subject of litigation." (1560).
Her theory, in short, is that clear contracts "can help alert recipients of information about the existence and scope of claimed trade secrets[,]" (id.), thus avoiding the uncertainty and unfair surprise that comes with invisible boundaries. Interestingly, while Varadarajan does not appear to take a stance on the debate mentioned above about which legal norms best justify trade secret law, but her focus on notice as a sign of "possession" sounds more in property theory than in contract. The "Evasive" Function of Contracts The second function Varadarajan identifies for contracts is substantive rather than evidentiary. Along with serving as ex post evidence of a viable trade secret claim, contracts can be used by sophisticated parties to get around or "evade" the boundaries of trade secret protection. Varadarajan calls this (I think far too leadingly) contracts' "evasive" function.
Varadarajan's identifies several "evasive" uses of contract in the trade secret context:
use of contracts to expand the subject matter of protectable information beyond a legal trade secret;
use of contracts to firm up an otherwise weak case for reasonable secrecy precautions;
use of contracts to eliminate potential reverse engineering defenses, for instance by publicly selling software subject to a license containing an anti-reverse engineering clause; and
use of contracts to prevent employees from working for competitors after they leave, i.e., non-compete agreements.
If the prior part of the article is descriptive, this part (Part IV and V) is highly normative. Drawing on the work of others like Orly Lobel, Pam Samuelson, and Rochelle Dreyfuss, Varadarajan asserts that contracts are being used to inappropriately "evade" the doctrinal limits of trade secret law. "While contracts can perform a valuable notice function in trade secret law," she writes, firms "use contracts strategically to evade trade secret law’s requirements and limitations." (1563). She thinks this is problematic because, in her words,
firms’ pervasive use of contracts to subvert trade secret law’s requirements and limitations can negatively impact cumulative innovation and employee mobility.
(1576) (emphasis added by me).
The most basic example of this problem (though, as I will explain, I am not sure it is a problem) is when firms use contracts to protect relatively confidential information that may not qualify as "secret." Varadarajan asserts that "[p]utative trade secret owners use contract law to evade [the] ongoing 'secrecy' requirement and chill uses of non-secret information—in effect, enlarging the scope of protectable information." Quoting Lobel, she makes the semi-empirical claim that "'[c]ontractually, it has become standard to include broad and open-ended lists of confidential information that goes beyond the statutory definition of trade secrets.'" (1565) (quoting Lobel).
To be clear, Varadarajan does not contend that any use of contracts to protect information should be viewed with suspicion. Rather, she argues contracts should be scrutinized more or less strictly, depending on where they fall on the "negotiability spectrum." (1573-1574). On one end, are business-to-business contracts involving sophisticated parties; on the other end are corporation-to-user and employer-to-employee contracts, where one contracting party may have unequal levels of information and bargaining power. (1574).
Since most contracts at issue in trade secret litigation are likely to fall on the latter end of this spectrum (primarily, contracts signed by employees), this stance would cast suspicion on many of the contracts courts currently see in trade secret disputes. For instance, non-compete agreements, which prevent employees from working for competitors for a set period following departure, would be more strictly scrutinized, even in states where they are not banned, by virtue of the fact that, as compared to licensing contracts, they involve parties with unequal bargaining power. (1577) ("[C]ourts should be more skeptical of employee and consumer contracts that depart from trade secret default rules.").
Varadarajan discusses various doctrinal ways to weaken these contracts: preemption, misuse, and traditional non enforceability of contract. (1576-1590). Without going into these solutions, I'll just note that none of them appears to have much of a bark. (See, e.g., 1588) ("Even in the employment context, where employees often assent under conditions of information asymmetry and disparate bargaining power, unconscionability arguments have held little sway.").
From Contract To Trade Secrets To ... Contracts
I am not convinced that this boogeyman exists. The earlier debates surrounding Bone's 1998 article implied that, where there is a contract in place, a legal remedy is more justified than when no contract exists. Yet Varadarajan suggests that relying so heavily on contracts is problematic, and that now contracts need to be justified when they go beyond the confines of trade secret protection.
Like Varadarajan, I also see irony in the trajectory her article describes: trade secret law has its origins partly in contract law; yet now companies use the existence of trade secrets to justify broad enforceability of their contracts.
[F]irms try to legitimize non-competes by invoking their connection to trade secret law—yet ironically, non-competes enable firms to bypass trade secret law’s requirements and limitations altogether. ... Indeed, “courts across the country point to the protection of trade secrets as the primary justification for enforcing non-competition covenants.”
(1572) (quoting Charles Tait Graves' article in Rochelle Dreyfuss and Katherine Strandburg's outstanding handbook on trade secrecy).
But there still has to be more discussion of why using contracts instead of or on top of trade secret law is inefficient, immoral, or what be it. In deciding..
An interesting study about Orange Book patents challenged both under Hatch-Waxman and Inter Partes Review caught my eye this week, but perhaps not for the ordinary reasons. One of the hot topics in drug patent challenges today is reverse payments: when the patentee pays the generic to stop a challenge. The Supreme Court has ruled that these payments can constitute antitrust violations. Though the drug companies give reasons, I'll admit that I've always been skeptical of these types of payments.
One of the key questions is whether the patent was going to survive. Most seem to assume that if a company pays to settle, then the patent was likely going to be invalidated. That's where the draft, Maintaining the Balance: An Empirical Study on Inter Partes Review Outcomes of Orange Book-Listed Drug Patents and its Effect on Hatch-Waxman Litigation, by Tulip Mahaseth (a recent Northwestern Law grad) comes in. Here is the abstract from SSRN:
The Hatch-Waxman Act intended to strike a delicate balance between encouraging pioneer drug innovation and promoting market entry of affordable generic versions of pioneer drugs by providing a streamlined pathway to challenge validity of Orange Book patents in federal district courts. In 2012, the America Invents Act introduced Inter Partes Review (IPR) proceedings which provide a faster, cheaper pathway to challenge Orange Book patents than Hatch-Waxman district court litigation. IPRs also have a lower evidentiary burden of proof and broader claim construction standard, which should make it easier, in theory, to obtain patent invalidation in IPRs as compared to Hatch-Waxman litigation. This empirical study on IPR outcomes of Orange Book patents in the past six years shows that both generic manufacturers and patent owners obtain more favorable final decisions in IPRs as compared to their Hatch-Waxman litigation outcomes because the rate of settlement in IPRs is much lower than in Hatch-Waxman litigation. Moreover, generic manufacturers do not appear to be targeting Orange Book patents in IPRs during their drug exclusivity period. Only 2 out of more than 400 IPRs against Orange Book patents were filed by generic petitioners during the patents’ New Chemical Entity exclusivity period. About 90% of the 230 Orange Book patents challenged in IPR proceedings were also challenged in Hatch-Waxman litigation. It is likely that generic manufacturers are not deterred from Hatch-Waxman litigation because of the lucrative 180-day exclusivity period, which gives the first generic filer 180 days to exclusively market their generic version without competition from other generics when the Orange Book drug patent is successfully invalidated in a subsequent district court proceeding. Therefore, IPR proceedings do not appear to be disrupting the delicate balance sought by the Hatch-Waxman Act. Instead, the IPR process has provided generic manufacturers a dual track option for challenging Orange Book patents by initiating Hatch-Waxman litigation in district courts and also pursuing patent invalidity in IPRs before the Patent Trial and Appeal Board, which has reduced rate of settlements resulting in more patents being upheld and invalidated.
There's a lot of great data in this paper, comparing Orange Book IPRs with non-Orange Book IPRs, including comparison of win rates and settlement rates.
But I want to focus on one seemingly minor point: as the number of IPRs has increased, the rate of settlement has decreased. And, more important, the decreasing rate of settlement has led to more invalidation and more affirmance of patents.
This result gives a nice window into how we might view settlements. Traditional Priest-Klein analysis says that this is exactly what we should see - that the previously settled cases were 50/50. But proving this is harder, and this data set would allow for a nice differences-in-differences analysis in future work.
Additionally, a split among outcomes implies that the settlements were not necessarily because the patentee believed the patent was at risk. If anti-competitive settlements were ruling the day, I would have predicted that most of the (recent) non-settlements would have resulted in patent invalidation. Then again, it is possible that a 50% chance was risky enough to merit a reverse payment settlement in the past. Regardless of how one comes out on this issue, this study provides some helpful details for the argument.
In Recalibrating Patent Venue, Colleen Chien and I did a nationwide study of forum shopping in patent cases (shocker - everybody did it, and not just in Texas), and predicted that many patent cases would shift from the Eastern District to the District of Delaware. And, lo, it has come to pass. Delaware is super busy. This has been good for us at Villanova (only 30 miles away from the court), as our students are getting some great patent experience in externships and internships.
But how much did firms value not being sued in Texas? The TC Heartland case is a clear shock event, so an event study can measure this. In Will Delaware Be Different? An Empirical Study of TC Heartland and the Shift to Defendant Choice of Venue, Ofer Eldar (Duke Law) and Neel Sukhatme (Georgetown Law) examine this question. The article is forthcoming in Cornell Law Review and a draft is on SSRN. Here is the abstract:
Why do some venues evolve into litigation havens while others do not? Venues might compete for litigation for various reasons, such as enhancing their judges’ prestige and increasing revenues for the local bar. This competition is framed by the party that chooses the venue. Whether plaintiffs or defendants primarily choose venue is crucial because, we argue, the two scenarios are not symmetrical.
The Supreme Court’s recent decision in TC Heartland v. Kraft Foods illustrates this dynamic. There, the Court effectively shifted venue choice in many patent infringement cases from plaintiffs to corporate defendants. We use TC Heartland to empirically measure the impact of this shift using an event study, which measures how the stock market reacted to the decision. We find that likely targets of “patent trolls”— entities that own and assert patented inventions but do not otherwise use them—saw their company valuations increase the most due to TC Heartland. This effect is particularly pronounced for Delaware-incorporated firms. Our results match litigation trends since TC Heartland, as new cases have dramatically shifted to the District of Delaware from the Eastern District of Texas, previously the most popular venue for infringement actions.
Why do investors believe Delaware will do better than Texas in curbing patent troll litigation? Unlike Texas, Delaware’s economy depends on attracting large businesses that pay high incorporation fees; it is thus less likely to encourage disruptive litigation and jeopardize its privileged position in corporate law. More broadly, we explain why giving defendants more control over venue can counterbalance judges’ incentives to increase their influence by encouraging excessive litigation. Drawing on Delaware’s approach to corporate litigation and bankruptcy proceedings, we argue that Delaware will compete for patent litigation through an expert judiciary and well- developed case law that balances both patentee and defendant interests.
As I discuss below, I have a like/dislike reaction to this paper.
I like the empirical setup - comparing companies that will be sued in Texas with those that will be sued in Delaware and with those that will be sued outside of Delaware (because they are incorporated elsewhere). The most robust results are that being sued anywhere else increased value, though there are some barely significant results that firms did even better in Delaware. This was interesting and useful.
But I dislike event studies generally, for a variety of reasons. The big issue here is uncertainty about how districts will change behavior in the long term. Thus, there is no way the authors could have waited for a correction (they note this, of course). But this means (which they also note) that this is a market prediction about firm value from not being sued in Texas - but that tells us very little about the actual impact of actual litigation or changes in the future. It may be that Delaware is already perceived as better than Texas, or it could be that the market is simply wrong about the cost of being in Texas.
I like the theoretical setup - that Delaware has attracted incorporation through its court behavior and laws, and using bankruptcy as an example. This adds a lot, though the empirical results only weakly support the theory (in a predictive sense) from a patent point of view. But the thought that Delaware will move to be pro-defendant (or at least less pro-plaintiff in their view) is an area for explanation.
But I dislike the theoretical conclusions. I believe the paper relies too heavily on the "forum selling" motif without critical analysis. There is at least one article (Colleen's and mine) that discusses other reasons why local scheduling winds up looking bad for defendants, such as congestion. Of course, we discuss selling as well, but we make the explicit point that the jury is out as to whether Delaware will be so friendly once the volume of cases doubles with its only four judges.
To that end, one would think that Delaware always had this pro-defendant (or balanced) approach if we buy the forum selling theory. But then why did so many plaintiffs (including NPEs) sue there even before TC Heartland? Doesn't that seriously crimp either a) a selling theory, or b) a post-TC Heartland change?
But I do like their further discussion of this about a two-stage venue selection game: 1) select where to incorporate or have facilities, and 2) select where to sue. This would be an interesting theoretical paper in the future, or even something to be tested. I wonder whether Delaware's incentives to treat plaintiffs/defendants a certain way will have much of an effect if the result is plaintiffs suing in other districts (which will often not have very helpful scheduling rules).
In the end, none of these are answered by the interesting empirical results. So, while I like the empirical analysis, I dislike that the analysis is not tied to the conclusions in a meaningful way.
The Supreme Court issued its opinion in WesternGeco last week. The holding (7-2) was relatively straightforward: if an infringer exports a component in violation of 35 USC 271(f)(2) (that is, the component has no substantial noninfringing use), then the presumption of extraterritoriality will not bar damages that occur overseas. And that's about all it ruled. It left harder questions, like proximate cause, for another day.
I spent the end of the week and weekend reading commentary on the case (and tussling a bit on Facebook and Twitter). A couple blog posts worth checking out are Tim Holbrook's and Tom Cotter's. I had just a few thoughts to add. 1. This was an easy case, and it's surprising it took this long to get here. 35 USC 271(f)(2) allows for infringement due to exporting a component of an infringing good because it was an end run around exporting the whole product. It stands to reason that if you lost a sale of the whole good in a foreign market because of that export, then you should be compensated for it. Why else have the (f)(2)? Justice Gorsuch's dissent describing the history of the Deepsouth case only makes this clearer. In Deepsouth, the Court held that exporting a part didn't count, and Congress said, "Oh, yes it does." It is telling that the majority doesn't even cite Deepsouth, yet the dissent and many lower court extraterritoriality opinions do.
2. That doesn't mean that every type of damage in the foreign jurisdiction should count. Causation matters, and as I'll discuss below, proximate causation is where the action will be in the future. It's frankly where the action should have been all along. The only reason causation wasn't the centerpiece is that the courts used the extraterritoriality question as a heuristic to avoid the harder causation questions. Thus, I suppose much of the angst about this case is that causation for the "lost contracts" type of damages is so bound up with extraterritoriality--and that extraterritoriality is just a cleaner way to deny damages that people think should not be allowed on causation grounds. But I've never been one for clean ways to deal with hard policy questions. Have you read my Section 101 scholarship lately?
3. In any event, I'm surprised by the surprise that proximate cause was not addressed. The question presented was very straightforward: is extraterritoriality an absolute bar to damages in this case? And the Court answered that question: No. In this era where really big partisan gerrymandering cases are decided on lack of standing (and other narrow opinions), it comes as no surprise at all that the Court answered the question asked and only question asked, notwithstanding the desperate pleas of amicii everywhere.
4. Finally, with respect to causation and expansion of damages, I don't think this case spells doom, but I do think that we'll see partial reversal of some cases. Take what I say with a grain of salt - I think Rite Hite v. Kelley is correctly decided as a causation matter. If you don't, well, you're head is probably going to explode, because broader causation will likely lead to a bit more foreign damages, even in 271(a) cases. But I don't think it will be crazy foreign damages. Let's look at the two 271(a) type cases everyone is talking about:
Power Integrations v. Fairchild: Power Integrations claimed damages for foreign infringing sales based on U.S. sales due to foreseeability (e.g. becoming a market leader and thus selling more.) The Federal Circuit mixed extraterritoriality with causation (won't be able to do this so much anymore), but winds up echoing WesternGeco's holding that the damages must be tied to U.S. infringement, and that there's no causation:
Power Integrations has not cited any case law that supports an award of damages for sales consummated in foreign markets, regardless of any connection to infringing activity in the United States. To the contrary, the entirely extraterritorial production, use, or sale of an invention patented in the United States is an independent, intervening act that, under almost all circumstances, cuts off the chain of causation initiated by an act of domestic infringement.
So, Power Integrations comes out the same way.
Carnegie Mellon v. Marvell: Two types of foreign damages were alleged. First were foreign sales of chips practicing a patented method. Second were foreign sales of chips designed using a patented method. The first types of damages don't change after WesternGeco. The chips were made and sold outside the U.S., and have no tie to domestic infringement, as required by WesternGeco.
But the more intriguing claim is to the chips designed by a patented design process, but made overseas. I frankly don't know how this comes out. Under WesternGeco, there is potentially no extraterritoriality defense because the foreign chip manufacture and sale flows from the domestic infringement. Or does it? Probably, if the question is compensating someone for harms. But causation is the more interesting question. I think this is more difficult than the straw person Justice Gorsuch creates (an infringing U.S. prototype); I agree with Tom Cotter that prototype damages causation is weak. But design method causation? I don't know, and I haven't done the research to see if there are any cases on this.
But my takeaway is the same - the answer to this question rests where it should: on causation. If you're outraged that damages would be owed due to infringement in the design process, then why should that outrage be lessened for U.S. product sales? Either damages that flow from such infringement should be allowed or they shouldn't, regardless of where they take place.
5. The final question is whether this affects business location. How many design and sale activities will move offshore to avoid damages? We'll see if this case has that effect, though it wouldn't be too surprising to see at least some shift.
When I first started practice, the place to go for patents was the Patent Depository Library at the Sunnyvale Public Library. Not only did they have copies of all the patents, they had other disclosures, like the IBM Technical Disclosure series. For those who wonder whether people actually read patents, I can attest that I never went to that library and found it empty. Many people, mostly individual inventors who did not want to pay for Delphion or some other electronic service, went there to look at the prior art. Sadly, the library ceased to be at the end of 2017. Widespread free availability on the Internet, plus a new USPTO center in San Jose siphoned off all the traffic.
Rather than rely on my anecdotal evidence, a new NBER paper examines the role of Patent Depository Libraries as evidence of patent disclosure. Jeffrey Furman (Boston U. Strategy & Policy Dept), Markus Nagler, and Martin Watzinger (both of Ludwig Maximillian U. in Munich) have posted Disclosure and Subsequent Innovation: Evidence from the Patent Depository Library Program to NBER's website (sorry, it's a paywall unless you've got .gov or .edu rights). The abstract is here:
How important is information disclosure through patents for subsequent innovation? Although disclosure is regarded as essential to the functioning of the patent system, legal scholars have expressed considerable skepticism about its value in practice. To adjudicate this issue, we examine the expansion of the USPTO Patent and Trademark Depository Library system between 1975 to 1997. Whereas the exclusion rights associated with patents are national in scope, the opening of these patent libraries during the pre-Internet era yielded regional variation in the costs to access the technical information (prior art) disclosed in patent documents. We find that after a patent library opens, local patenting increases by 17% relative to control regions that have Federal Depository Libraries. A number of additional analyses suggest that the disclosure of technical information in the patent documents is the mechanism underlying this boost in patenting: the response to patent libraries is significant and of important magnitude among young companies, library opening induces local inventors to cite more geographically distant and more technologically diverse prior art, and the library boost ceases to be present after the introduction of the Internet. We find that library opening is also associated with an increase in local business formation and job creation, which suggests that the impact of libraries is not limited to patenting outcomes. Taken together, our analyses provide evidence that the information disclosed in patent prior art plays an important role in supporting cumulative innovation.
The crux of the study is the match to other, similar areas with Federal Depository (but not patent) Libraries. The authors acknowledge that the opening of a patent library might well be a leading indicator of expected future patenting, but the authors discount this by arguing that the Patent Libraries would have had to somehow predict the exact year of increased patenting, and then apply in advance of that date and get approved just in time. The odds of this seem low, especially when the results are localized to within 15 miles of the library (and no further).
The first core finding, that patenting increased, is ambiguous normatively. The authors discuss enhanced innovation, but equally likely alternatives are that people just got excited about patenting or that innovation already occurring was more easily patented. That said, they find the same quality, which implies that the patenting wasn't simply frivolous.
The second finding is more important: that the types of citations and disclosures changed (and that those changes disappeared when patents were more readily available on the Internet). This finding implies that somebody was reading these patents. The question is who. A followup study looking at how the makeup of inventorship changed would be interesting. Were the additional grants solo inventors or large companies? Who used these libraries?
Even without answering this question, this study was both useful and interesting, as well as a bit nostalgic.
Those interested in the patent system have long complained of patent thickets as a barrier to efficient production of new products and services. The more patents in an area, the argument goes, the harder it is to enter. There are several studies that attempt to measure the effect of patent thickets, with some studies arguing that thickets can ease private ordering. I'd like to briefly point out another (new) one. Charles deGrazia (U. London, Royal Holloway College), Jesse Frumkin, Nicholas Pairolero (both of USPTO) have posted a new draft on SSRN, called Embracing Technological Similarity for the Measurement of Complexity and Patent Thickets. Here is the abstract:
Clear and well-defi ned patent rights can incentivize innovation by granting monopoly rights to the inventor for a limited period of time in exchange for public disclosure of the invention. However, when a product draws from intellectual property held across multiple firms (including fragmented intellectual property or patent thickets), contracting failures may lead to suboptimal economic outcomes (Shapiro 2000). Researchers have developed several measures to gauge the extent and impact of patent thickets. This paper contributes to that literature by proposing a new measure of patent thickets that incorporates patent claim similarity to more precisely identify technological similarity, which is shown to increase the information contained in the measurement of patent thickets. Further, the measure is universally computable for all patent systems. These advantages will enable more accurate measurement and allow for novel economic research on technological complexity, fragmentation in intellectual property, and patent thickets within and across all patent jurisdictions.
The authors use natural language processing to determine overlap in patent claims (and just the claims, arguing that's where the thicket lies) for both backward and forward citations in "triads" - patents that all cite each other. Using this methodology, they compare their results to other attempts to quantify complexity and find greater overlap in more complex technologies - a sign that their method is more accurate. Finally, they validate their results by regressing thickets against examination characteristics, showing that the examination factors more likely to come from thickets (e.g. pendency) are correlated with greater thickets.
This is an interesting study. The use of citations (versus technological class) will always be a limitation because not every patent in a thicket winds up being cited by others. However, the method used here (using forward and backward citations) is better than the alternative, which is using only blocking prior art.
The real question is what to do with all this information. Can it be applied beyond mere study of which areas have thickets? I suppose it could be helpful for portfolio purchases, and maybe to help decisions about whether to enter into a new technology.
I've done a few interviews about the latest Apple v. Samsung design patent jury verdict, but journalistic space means I only get a couple sentences in. So, I thought I would lay out a couple points I see as important. We'll see if they hold up as predictions.
There's been a lot written about the case, so I won't rehash the epic story. Here's the short version. The design patent law affords the winning plaintiff all of the profits on the infringing article of manufacture. The Supreme Court ruled (reversing about 100 years of opposite practice) that the article of manufacture could be less than the entire accused device for sale. Because the original jury instructions did not consider this, the Court remanded for a determination of what the infringing article of manufacture was in this case (the design patents covered the shape of the phone and the default screen). The Federal Circuit remanded, and the District Court decided that, yes, in fact, the original jury instructions were defective and ordered a retrial of damages.
The District Court adopted the Solicitor General's suggested test to determine what the article of manufacture was, determined that under that test it was a disputed fact question, and sent it to the jury. Apple asked for $1 billion. Samsung asked for $28 million. The jury awarded $533 million, which is more than $100 million more than the damages were before the Supreme Court ruled.
After the trial, one or more jurors stated that the entire phone was the article of manufacture because you can't get the screen without the rest of the phone. I suppose that the half a billion is deducting expenses that Apple didn't want to deduct.
So, here are my points: 1. Samsung is unlikely to win on appeal (despite the hope many seem to be holding). The reason is twofold:
A. Samsung appears to have agreed to the jury instruction on what determines an article of manufacture. Given that fact, the instruction won't be thrown out.
B. Samsung is thus left to argue that, as a matter of law, the whole phone cannot be the article of manufacturer, and no reasonable juror could so find under the test at issue. This is not an unreasonable argument, but it's a difficult one, because:
i. The post-trial jury statements about what happened in the room are inadmissible,
ii. Apple submitted some evidence under the jury instruction that the whole phone can be the article of manufacturer, and all inferences are determined in their favor, and
iii. There are explanations for an amount less than $1 billion that are consistent with the evidence (e.g. the jury assessed additional expenses). This is similar to the first trial, when Apple asked for more money on top of the jury verdict because Apple said it definitively showed more profits and the jury gave only 40% of what it asked for. The District Court said then - we just don't ask how the jury calculated if there was some evidence. I can easily seeing the court doing the same thing here.
2. The jury verdict is not a huge precedent, or a precedent at all. A different jury could have gone a different way on these same facts. Different cases will have different facts. This wasn't even the first design patent damages jury verdict since the Supreme Court ruled. Don't remember any other? See Microsoft v. Corel ($272K awarded, remitted to $74K because there really was insufficient evidence, apparently).* Sarah Burstein has a good post on that case. That's my point - jury verdicts are highly fact intensive. To the extent there are precedents here, they are:
A. The use of this particular jury instruction. To the extent the jury said they were confused, litigants and courts might be wary of it in the future.
B. Whether the District Court and/or Federal Circuit rule that, as a matter of law, a phone infringing these patents can never be the article of manufacture. In Microsoft, the Court went the other way (sort of) and said that the whole software could be the article of manufacture.
If the answer winds up being B - the whole phone can never be the article of manufacture, one wonders (well, I wonder) why this ever went to the jury with the instruction it did. Perhaps the District Court was hoping the jury would get it right and avoid the need for an appeal on this issue. But this is a murky space, where precedent would be helpful if the courts think it warranted. As such, in my view, it may be that the court simply believed that there could be evidence under the jury instruction that would lead to the whole phone being an article of manufacture. And, if that's true, we're back to where we started--Samsung having difficulty on appeal.
The parties are apparently mediating now; perhaps we'll never get an answer to these questions. Or perhaps a final settlement will reveal Samsung's view of what the answer was likely to be.
*Side note: it's pretty sad that the total profits of Corel Home Office products for two years was $74K. I weep for this once mighty franchise and it's view codes functionality.
Jonathan Ashtor (now at Paul, Weiss) has completed a few quality empirical studies in the past. His new foray is a new and creative way to determine novelty. It's on SSRN, and the abstract is here:
I construct a measure of patent novelty based on linguistic analysis of claim text. Specifically, I employ advanced computational linguistic techniques to analyze the claims of all U.S. patents issued from 1976-2014, nearly 5 million patents in total. I use the resulting model to measure the similarity of each patented invention to all others in its technology-temporal cohort. Then, I validate the resulting measure using multiple established proxies for novelty, as well as actual USPTO Office Action rejections on grounds of lack of novelty or obviousness. I also analyze a set of pioneering patents and find that they have substantially and significantly higher novelty measures than other patents.
Using this measure, I study the relationship of novelty to patent value and cumulative innovation. I find significant correlations between novelty and patent value, as measured by returns to firm innovation and stock market responses to patent issuance. I also find strong correlations between novelty and cumulative innovation, as measured by forward citations. Furthermore, I find that patents of greater novelty give rise to more important citations, as they are more frequently cited in Office Action rejections of future patents for lack of novelty or obviousness. I also investigate how novelty relates to the USPTO examination process. In particular, I find that novelty is an inherent feature of a patented invention, which can be measured based on the claim text of either an issued patent or an early-stage patent application.
Next, I use this measure to analyze the characteristics of novel patents. I find that novelty is an effective dimension along which to stratify patents, as key patent characteristics vary significantly across the distribution of novelty measures. Moreover, novel patents are closely linked to basic scientific research, as measured by public grant funding and citations to non-patent scientific literature.
Finally, I use this measure to observe trends in novelty over a forty-year timespan of American innovation. This reveals a noticeable, albeit slight, trend in novelty in certain technology fields in recent years, which corresponds to technological maturation in those sectors.
I'm skeptical of measures of patent quality by claim language only, but I like how he has used office actions to validate the measure. I think people will have to study this to see how it holds up, but I think it's an interesting and creative first step toward objectively judging quality.
As noted in my last post, one of the most quoted lines in copyright law is from Justice Holmes's 1903 opinion in Bleistein: "It would be a dangerous undertaking for persons trained only to the law to constitute themselves final judges of the worth of pictorial illustrations." This aesthetic neutrality principle has found purchase far beyond copyright law. But in a compelling new article, Aesthetic Judgment in Law, Professor Brian Soucek challenges this dogma: "Almost no one thinks the government should decide what counts as art or what has aesthetic value. But the government often does so, and often, it should." Soucek's article may have flown under the radar for most IP scholars because he does not typically focus on copyright law, but it is well worth a look.
Soucek's first point is that despite the "widespread aversion to aesthetic judgment" by government decisionmakers, such judgments are ubiquitous both at the "retail" level of individual artworks and at the "wholesale" level of "what constitutes art or aesthetic value in the first place." A number of scholars have made similar points in the IP space; see, for example, Andrew Gilden's argument that courts are more likely to consider images of women and racial minorities to be "raw materials" that are free to use. But the point holds even more strongly in other areas of law. Perhaps most obviously, there is significant direct spending on the arts that falls on the "government-set" side of the who decides? spectrum, including grant decisions by the National Endowment for the Arts, book purchases by public libraries, and hiring and curriculum decisions by humanities departments at public universities. Other examples are easy to find: Tariff and tax laws embody Congress's decision to benefit only certain types of art, and they require government officials to make judgments such as whether abstract art is art. Land-use laws ban conduct that is "offensive to the visual sensibilities of the average person" like front-yard clotheslines. The test for obscenity asks whether the work "lacks serious literary [or] artistic" value.
The primary arguments in favor of aesthetic neutrality have been that judges lack sufficient expertise to make aesthetic judgments and that taste is hopelessly subjective (or more precisely: unpredictable or relativist). Soucek argues that neither argument is persuasive. On judicial incompetency, he writes that "it is hard to understand why aesthetic judgments should be any harder to make than judgments involving, say, complex technology in patent disputes or judgments requiring deep familiarity with the economics of a particular industry in antitrust cases." Judicial ignorance could be mitigated by expert testimony—unless there is no expertise in aesthetics, which leads to the second argument for aesthetic neutrality: there's no accounting for taste. On the relativism of aesthetic judgments, Soucek states that (1) "even if we were to grant the relativism of retail aesthetic value judgments, this might still leave untouched most of the other substantive aesthetic judgments canvassed in Part I"; (2) "much of the leading work in philosophical aesthetics in the past three centuries has been spent explaining the universality of our judgments of taste"; and, perhaps most significantly, (3) "acceding to relativism is itself a substantive aesthetic judgment."
Soucek contends that the most convincing limits on aesthetic judgments by government actors come not from institutional (in)competency or aesthetic relativism, but from the Free Speech Clause's restrictions on viewpoint discrimination:
We might imagine a spectrum of First Amendment applicability. On one end, the Free Speech Clause applies in fullest force whenever the state is trying to shut down expression that diverges from state orthodoxy. Here, the aesthetic nondiscrimination principle is needed to keep the government from encroaching on private beliefs about substantive aesthetic matters. When private aesthetic judgments are regulated or, worse, disallowed—as they are, for example, in blight determinations and obscenity law—aesthetic neutrality becomes essential. But as we move across the spectrum from government regulation toward government subsidy of speech, aesthetic judgment becomes less concerning. Here we find cases, like tax and tariff exemptions and many areas of intellectual property, where private aesthetic judgments are being endorsed or subsidized by the government. Finally, at the far end of the spectrum, we move from government subsidies for speech to government speech itself. There the Free Speech Clause imposes no limit at all and the aesthetic nondiscrimination principle should cease to apply.
Locating a particular aesthetic judgment along this spectrum is intertwined with the question of whether the state action constitutes "government speech," a doctrine that Daniel Hemel and I attempt to elucidate in a forthcoming Supreme Court Review article. As Soucek notes, these distinctions are not easy, and while his framework "largely turns on the distinction between subsidizing versus regulating or limiting . . . . the word 'subsidy' is not some magic tailsman that causes First Amendment worries to disappear." But this article's nuanced discussion should be required reading for courts or commentators grappling with these issues across any of these areas of law. Highly recommended.
Just this morning, an interesting new literature review came to my mailbox via SSRN. In Is There a Role for Patents in the Financing of New Innovative Firms?, Bronwyn Hall (Berkeley economics) provides an extremely thorough, extremely helpful literature review on the subject. It's on SSRN, and the abstract is here:
It is argued by many that one of the benefits of the patent system is that it creates a property right to invention that enables firms to obtain financing for the development of that invention. In this paper, I review the reasons why ownership of knowledge assets might be useful in attracting finance and then survey the empirical evidence on patent ownership and its impact on the ability of firms to obtain further financing at different stages of their development, both starting up and after becoming established. Studies that attempt to separately identify the role of patent rights and the underlying quality of the associated innovation(s) will be emphasized, although these are rather rare.
This paper caught my eye for a few reasons. First, I'm working on a paper on this topic right now, using a high quality dataset that nobody has been able to exploit for this question. I hope my coauthor (David Ratigan, an economist here at Villanova) and I can do so! Hall's paper lays out some of the challenges we face, and the primary criticism of prior papers: whether the benefit of financing is simply the patent right, or instead the underlying quality of the invention. Professor Hall suggests that the best approach may be a detailed study of companies with unpatented inventions as compared to companies with patented inventions. I think it would be great, but really difficult, to do such a study. But I'm not convinced it is necessary with the proper random sample and controls. We'll find out, because that's what we're trying to do. Even if we fail, I think there is value in knowing the role of the patent right even if it is simply a proxy signal - more on this theoretical question below.
Second, I think it would be good for law folks to read this. This is not a literature I hear discussed or cited very often.
Third, Professor Hall and I seem to agree that the best in class for these studies is What is a Patent Worth? Evidence from the U.S. Patent Lottery. Professor Hall cites to a prior draft, called the Bright Side of Patents, which I blogged about here. Related, and more personally, I have all the same papers in my literature review, so I'm feeling glad that I didn't miss any.
In the end, there are a few theoretical divides that affect how these studies are done: Do we look only at those firms who received venture capital? Do we look at the fact of investment or the amount? Does final outcome matter, or just financing? Is signaling sufficient, or should we care if the signal is accurate?
This last question is the most important, and the one highlighted in this literature review. Must we separate the patent right from the patent innovation in order to determine that the patent system has value? Whenever I have propounded this theory of patenting, that's the pushback I get - that the patent is just a correlated signal with firm quality, so the patent doesn't have any real value on its own (this pushback even implies that the patent right has little value). But imagine a world where there is no patent system and firms innovate. How would they signal their quality? The method doesn't really matter, except to note that those very same firms that don't patent now can signal their quality in the exact same way. And so the question becomes threefold in my view:
If firms are able to signal their quality without patents, then shouldn't we expect to find random distribution of financing among firms with patents and firms without?
If firms are not able to signal their quality without patents, then is that a sufficient justification for the patent system--a quality signal?
Do we live in a world where firms could signal their quality without patents, but the patent system drowns them out as a "supersignal" (for better or worse)?