Wednesday, May 16, 2018

Mark McKenna: Trademark Counterfeiting And The Problem Of Inevitable Creep

One of my favorite events at Akron Law this past school year was hearing Professor Mark McKenna deliver the Oldham Lecture on his fascinating paper, Criminal Trademark Enforcement And The Problem Of Inevitable Creep.  The completed article, forthcoming in the Akron Law Review, is now available on SSRN.

The story, in Mckenna's telling, is simple. There is a criminal remedy for trademark "counterfeiting" because, most people would agree, using an identical trademark for goods or services that are identical to the trademark owner's is an economically and morally worse act than ordinary trademark infringement. A modern-day example of this atrocious crime is the company that has been hawking dysfunctional "Philips Sonicare" toothbrush replacement heads on Amazon.com. Consumers buy them thinking they are the real thing, and are sorely disappointed when the brush heads do not work. But to deserve the classification as criminal, as a legal matter, the act of counterfeiting must be proven "beyond a reasonable doubt" to fit within the exact text of the relevant statute, the Trademark Counterfeiting Act. According to McKenna, courts have veered from the statutory text, and are instead expanding criminal counterfeiting beyond Congressional authorization. Thus, the article's reference in its title to "inevitable creep."

There are parts of this well-done article with which people are likely to agree, and other parts with which people are likely to strongly disagree.

An ordinary civil trademark infringement case brought under Lanham Act Section 32 or Lanham Act Section 43(a)(1)(A) can involve an infringing mark that is substantially different from the real mark, and products that bear little resemblance to the genuine ones. For example, using the name "VIRGIN" in the sale of retail wireless telephone services can infringe the trademark, VIRGIN, even though VIRGIN is registered for use in association with a major international airline and retail electronics stores that do not sell telephones or telephone services. Using the name "RICA" in the sale of salty crackers can infringe the trademark RICA as used on cookies, at least when both crackers and cookies are referred to in Spanish as "galletas."

The reason for these outcomes is that, for ordinary civil trademark infringement, all that is required is that an appreciable number of consumers are likely to be confused by defendant's use of plaintiff's mark. Whether consumers are likely to be confused depends on a number of factors, such as the degree of similarity between the two marks, the proximity of the products, and so on. The marks and the products on which the marks are used can be, but need not be, identical.

Criminal trademark counterfeiting is different. Generally speaking, counterfeiting is just ordinary trademark infringement, but where the defendant uses precisely the same trademark on precisely the same goods or services for which the mark was originally registered.  (McKenna, p. 847). Thus, neither of the infringement examples mentioned above would be considered counterfeiting. On the other hand, if a new airline were to open up under the name VIRGIN, and people started boarding planes thinking erroneously that they were flying the well-known airline, VIRGIN, you can bet the government would take action to stop them.

To address this scenario, Congress passed the federal Trademark Counterfeiting Act (TCA) of 1984, codified in 18 U.S.C. § 2320. The TCA gives federal prosecutors that authority to "[act] on behalf of mark owners, seizing infringing and counterfeit goods at the border and prosecuting counterfeiters under federal criminal law." ( p. 847). According to McKenna, the TCA's purpose was "to increase the penalties associated with the most egregious form of trademark infringement—use of an identical mark for goods identical to those for which the mark is registered and in a context in which the use is likely to deceive consumers about the actual source of the counterfeiter’s goods." (p. 849).

The remedy is harsh.  Penalties include, along with seizure and forfeiture of goods, fines of up to $2 million for an individual (up to $5 million for a person other than an individual) and imprisonment of up to ten years for a first violation. See § 2320 (b); see also 19 U.S.C. § 1526(e). The government can also order "restitution," potentially in the amount of the full retail value of the genuine goods, which as you can imagine can be quite expensive. (p. 853).

But actually imposing this penalty is hard. And McKenna thinks it should in many cases be harder than it is currently is, at least if courts wish to be true to the language of the statute and Congress' intent.

The Relevant Statutory Language

This is mainly a paper about statutory construction of the relevant sections of the TCA.  Sections 2320(a) and 2320(f), respectively, delineate the offense of counterfeiting and define what it takes for a mark to be deemed a "counterfeit." They are quoted below, with the key portions underlined.

(a) Offenses.—Whoever intentionally

(1) traffics in goods or services and knowingly uses a counterfeit mark on or in connection with such goods or services [etcetera] ... shall be punished as provided in subsection (b).

...

(f) Definitions —For the purposes of this section—

(1) the term “counterfeit mark” means—

(A) a spurious mark—

(i) that is used in connection with trafficking in any goods, services, labels, patches, stickers... [etc.] or packaging of any type or nature;

(ii) that is identical with, or substantially indistinguishable from, a mark registered on the principal register in the United States Patent and Trademark Office and in use, whether or not the defendant knew such mark was so registered;

(iii) that is applied to or used in connection with the goods or services for which the mark is registered with the United States Patent and Trademark Office...[etc.]; and

(iv) the use of which is likely to cause confusion, to cause mistake, or to deceive...[etc.]


The Substantive Limitations On Criminal Counterfeiting Liability

Drawing on this statutory language, as well as the legislative history and case law, McKenna identifies several substantive limits that distinguish criminal counterfeiting from civil liability.  In McKenna's interpretation, these limits should be strictly construed.

1. No mens rea; no crime

First, unlike civil trademark infringement, criminal counterfeiting under the TCA has a mens rea. The defendant must "intentionally" traffic in goods or services and "knowingly" use a counterfeit mark on or in connection with those goods or services.  "Thus," McKenna writes, "criminal liability should not attach to unintentional infringement or use of a mark with a good faith belief that the use is legitimate." (p. 855). In the civil context, bad faith is a relevant factor, but it is not required.  See Quality Inns Intern., Inc. v. McDonald's Corp.695 F. Supp. 198 (U.S. D. Ct. Maryland, 1988).

2. Only registered marks, and only exactly as registered

Second, to violate the TCA a defendant must use a counterfeit of a mark that is both registered with the USPTO and in use by the mark owner for the precise goods and services specifically identified in the registration. (p. 856). As I'll explain below, McKenna interprets the language of § 2320(f)(1)(A)(ii) to mean that, in order for a defendant's mark to be deemed counterfeit, the mark must be roughly identical with the trademark owner's registered mark exactly as it is registered. Mixing up multiple registered trademarks on counterfeit goods does not necessarily suffice.

3. "Identical with, or substantially indistinguishable from" the registered mark 

Third, defendant must use a mark that is “identical with, or substantially indistinguishable from” the registered mark. The degree of similarity between a plaintiff and defendant's marks is a factor in civil cases as well, and in practice similarity is usually the most important factor in judging likelihood of confusion. But similarity, let alone identity, isn't strictly required for civil trademark infringement.

McKenna argues it is "abundantly clear that Congress meant to require a much higher level of similarity for criminal counterfeiting liability than is necessary for a finding of civil infringement..."  (p. 857-858).  But he asserts that courts do not always remain true to this congressional intention.

For example, in U.S. v. Lam (4th Cir. 2012), a case in which McKenna served as an expert witness, the Fourth Circuit permitted a finding of criminal counterfeiting even though the defendants did not use precisely the same mark that the trademark owner, Burberry, had registered. Instead, McKenna explains,
"the defendants sold handbags with a composite pattern that consisted of a plaid pattern with an equestrian knight superimposed over it. Burberry owned a federal registration of a plaid pattern ... and it owned several other federal registrations for an equestrian knight device. But importantly, Burberry did not own a single registration for any mark consisting of both components."
(p. 867-868) (underlining added). A beautiful image of defendant's counterfeit bag containing an amalgam of the two different registered Burberry trademarks is printed on page 867 of McKenna's article.

The court's error, in McKenna's view, arose from the fact that the statute states in § 2023(f)(1)(A)(ii) that the defendant's mark must be "identical with, or substantially indistinguishable from, a mark registered on the principal register..." Yet here defendant was mixing together two different marks that Burberry had registered. (p. 868). Relatedly, McKenna objects to the court's jury charge, which permitted the jurors to simply ignore the presence of the defendants' knight device in comparing the composite pattern with the registered Burberry marks. (pp. 868-869). Basically, the court allowed the jurors to consider the defendants' plaid design alone, absent the knight, which meant the jury could more easily find that the defendant's mark was "identical with, or substantially indistinguishable from" Burberry's plaid pattern mark. (p. 869).  McKenna views this as a violation of the so-called anti-dissection rule: the notion that marks must generally be viewed in their entirety, not in individual segments. (p. 868) ("[T]o allow the jury to ignore the defendants' knight device was to invite it to compare a part of the defendants' mark to the allegedly counterfeited registered mark, a comparison that would flout the anti-dissection rule...").

One might object that this seems like an overly strict reading of the statute. Congress' use of the phrase "identical or substantially indistinguishable from a mark registered on the principle register..." implies that superimposing one registered mark on top of another registered mark would be sufficient, at least so long as likelihood of confusion prong in § 2320(f)(1)(A)(iv) is ultimately satisfied. The spirit of the "substantially indistinguishable" language seems to be that a mark need not be identical to be a counterfeit. Mashing up two registered trademarks seems to fit that definition.  (For McKenna's further explanation for precisely why he thinks this isn't good enough, see pp. 867-870.)

4. Likelihood of confusion is required  

The final limitation, which also obviously exists in the civil context, is that the defendant's use of the allegedly counterfeit trademark must cause likelihood of confusion amount the relevant consumers.  See § 2320(f)(1)(A)(iv). Importantly, McKenna stresses that this likelihood of confusion requirement is different from the similarity-of-marks requirement discussed above, and he chastises courts’ apparent "conflation of the similarity question (whether the defendant’s mark is identical with, or substantially indistinguishable from the genuine mark) and the separate likelihood of confusion question." (p. 875).

The upshot is that similarity of the marks alone is not enough. Consumers must be likely to be confused when viewing the marks in the marketplace.  McKenna thinks this requirement is very important.  If all that was required was identical marks, then it would not matter whether defendant was selling what are clearly knock-off goods that would not be likely to confuse anyone as to source or affiliation.

Should Post-Sale Confusion Be Covered?

McKenna identifies several procedural errors in courts' likelihood of confusion analyses. But his real problem is that courts have begun to apply a post-sale confusion theory in the criminal context. (pp. 861-866).

Post-sale confusion was developed to deal with situations where the purchaser of a knock-off item knows the item isn't genuine, so they are not confused at the "point-of-sale." Instead, courts hold that the confusion occurs "post-sale," when third party bystanders see the knock-offs and think they are real.

In McKenna's view, post-sale confusion is one "of the most controversial civil infringement doctrines," and should not be applied in the criminal counterfeiting context, especially not without express congressional authorization. (p. 864).

There is little information as to whether Congress explicitly intended the TCA to cover post-sale confusion. (p. 864). According to legislative history uncovered by McKenna's research, legislators were concerned that infringers saw civil liability as simply "the cost of doing business," and were "emboldened" by the "lack of sufficiently stiff civil sanctions" to sell counterfeit goods to "defraud consumers out of billions of dollars each year in the United States alone.” (p. 849) (quoting S. REP. NO. 98-526, at 5 (1984)).

Despite this sparse legislative history, it is hard to believe Congress did not intend post-sale confusion to be covered. Indeed, McKenna himself notes Congress' approving citation to a "fake Rolex" case, where it was clear buyers knew what they were getting. (pp. 858-859; p. 866). The situation has gotten a lot worse since 1984. In today's marketplace, savvy consumers knowingly buy fake Rolexes, Guccis, and Tiffany's products. Amusing videos on YouTube show shoppers testing out counterfeit products purchased on Amazon and eBay, and sharing their results with the world.

Surely, says the devil's advocate, if Congress wanted this type of conduct to be illegal, then the TCA would have to require moving to a post-sale confusion theory.  McKenna expressly recognizes this objection, conceding that "[p]erhaps courts’ acceptance of post-sale confusion in the counterfeiting context" is not all that surprising. (p. 865).
"[A]fter all, many people refer colloquially to knock-off luxury goods as 'counterfeits,' and the legislative history of the TCA makes specific reference to a case about fake Rolex watches. Thus, maybe it is too hard to imagine that a counterfeiting statute would not reach imitation luxury goods, and post- sale confusion is the only way to accomplish that end. In fact, to the extent most people think of imitation luxury goods when they speak of counterfeiting, it might even seem that the argument against criminalizing post-sale confusion is just an argument against criminal liability for counterfeiting."
(pp. 865-866) (emphasis added).

McKenna's response is not entirely satisfying to me. "[Rejecting post-sale confusion," he writes, "would not render the TCA meaningless, as there are many counterfeit products— including counterfeit drugs—that risk substantial point-of-sale confusion. And those counterfeit products are clearly the ones most likely to do harm, both to consumers and to mark owners." (p. 866).

But eliminating a post-sale confusion theory would still seem to leave out a huge proportion of the products sold today that the average person would consider to be "counterfeit goods." Without getting into the broader discussion of whether post-sale confusion should be actionable, it seems like the "fake Rolex" situation should be deemed legally, as well as colloquially, a counterfeit even if the purchaser knows the Rolex is fake. This type of act generally checks all three boxes of trademark policy: (1) deterring morally bad acts by defendants who sell look-alike goods; (2) preventing a reduction in the sales of original sellers; and (3) protecting consumers from being confused about the source of products they see in the marketplace. Assuming courts do their job in carefully finding counterfeit marks are "identical" or "substantially indistinguishable," and that an appreciable number of consumers, somewhere down the line, is likely to be confused regarding source, then there is likely to be harm on all three dimensions.

On the other hand, McKenna points to scholarship by others like Jeremy Sheff, highlighting how hard it is to prove that third-party bystanders will change their buying decisions based on seeing counterfeit goods. (p. 865). If, for example, third-parties buy more versus less of the genuine goods when counterfeits abound, then there's no harm at all to the trademark owner. (p. 862). However, in the civil context, the Lanham Act does not require that consumers actually change their buying decisions. As Rebecca Tushnet has discussed, false advertising claims under § 43(a)(1)(B) have a "materiality" element, wheres trademark infringement claims do not. So the fact that trademark owners may not experience a reduction sales due to post-sale confusion should not, on its own, prevent society from classifying the act as criminally counterfeit.

Long story short, deciding precisely where one stands on statutory interpretation of the criminal counterfeiting act requires, at least in part, deciding where one stands on post-sale confusion in the civil context. And that is a complicated and polarizing issue. I am just glad that McKenna wrote this fantastic article and chose to publish in the Akron Law Review.

No comments:

Post a Comment