Professors Barton Beebe and Jeanne Fromer’s empirical tour de force presents a strong challenge to the conventional wisdom that there are infinite potential trademarks.1 To be sure, the claim that potential trademarks, broadly defined, are inexhaustible is tautologically true: there are infinite combinations of letters and other symbols — including sounds and colors — any of which might serve as a mark.2 In this sense, the claim that we might run out of trademarks seems as absurd as John Stuart Mill’s concern that we might run out of music.3 But not all marks are created equal. Some combinations of letters are unwieldy or unmemorable; others have negative connotations. Beebe and Fromer argue that the most valuable marks are common words, short and pronounceable neologisms, and common U.S. surnames.4 The concern about running out of these marks is far from absurd — just as Mill’s anxiety about running out of music makes much more sense when confined to three-note melodies such as the trademarked NBC chimes,5 of which there are only 469 possibilities.6
In this short space, I wish to laud the remarkable descriptive contribution of Are We Running Out of Trademarks? while sounding a note of caution on the normative implications. Beebe and Fromer’s data convincingly demonstrate that short, common word marks are becoming depleted and congested, and they present a number of plausible hypotheses about the negative welfare impact of this trend.7 Their findings suggest that trademark policy has been based on false assumptions and should be closely reexamined. But their data cannot elucidate the actual costs of depletion or congestion — particularly without noting how the market will adapt to reduce these costs — and cannot reveal if there are countervailing benefits. Generating concrete evidence of these costs and benefits seems like a necessary next step before recommending any significant changes to the current trademark system. After offering a laudatory evaluation of the value of Beebe and Fromer’s descriptive work, I explore why reforms in reaction to their research should proceed cautiously, and I suggest important avenues for future empirical work to build on these results.
I. Trademark Depletion and Congestion: From Anecdote to Evidence
In a recent trademark case, the Second Circuit asserted that “[o]ne seller’s monopolization of a particular term does not deprive competitors of anything of value because the number of arbitrary or fanciful marks available for use is infinite.”8 No citation was given; the court merely echoed the conventional wisdom that trademarks are inexhaustible.9 Similarly, forcing an infringer to change its mark has been found to create little competitive harm because “there are infinite other names under which defendants may continue to operate.”10 In some cases where the trademark is well-known, the “choice of a confusingly similar mark, out of the infinite number of marks in the world, itself supports an inference [of] bad faith.”11
While a few courts have recognized that competitively effective marks might not be limitless,12 these judges have had no more data to cite than the many judges and scholars who have argued that trademark depletion and congestion are impossible. Beebe and Fromer’s Article will help move these discussions from anecdote to evidence. They show that existing registrations do in fact substantially constrain the ability of new competitors to use common words and short neologisms as marks. Their Article is teeming with striking statistics — for example, of the 1000 most frequently used nouns or adjectives in American English, all 1000 were claimed within an active registration in 2014, and the average word appeared within registrations by 745 distinct registrants.13 Other noteworthy findings are that single-word marks cover 79% of all word usage14 and 81% of all syllable usage,15 and that 55% of the U.S. population has a surname that has been claimed as a single-word mark.16 And the real problem is likely significantly worse than these statistics indicate because their dataset does not include unregistered marks or account for the broadening scope that trademarks seem to be acquiring.17
This evidence suggests that the Second Circuit and other courts quoted above are simply wrong: based on depletion and congestion of the most sought-after trademarks, the “monopolization of a particular term” does in fact deprive competitors of something that they appear to value. And inferring bad faith based on the choice of a similar mark makes less sense when one realizes that firms are focused on a finite subset of possible trademarks. The question is not: “Given the infinite supply of marks, what is the chance that defendant would have chosen this plaintiff’s mark?” Rather, courts should ask: “Given the finite supply of marks that are perceived to be effective, the large number of those marks that are already being used in this category, and the expected search effort, what is the chance that defendant would have chosen a mark that is similar to some existing mark?”18
For example, consider the more than 5300 breweries in the United States.19 If we assume that each uses an average of 10 marks and that there are one million effective marks for beer, the probability that a new brewer who hasn’t done a thorough trademark search would choose the same trademark as an existing brewer — say, FIREMAN20 — is not one in a million (0.0001%); it is [1 − (1 − 1∕1,000,000)10×5300] ≈ 5%. If there are only 100,000 effective marks, the probability jumps to 41%.21 Of course, if the new brewer does thoroughly search existing marks, then the likelihood of choosing an existing mark decreases — but at a corresponding increase in the costs of entry. If nothing else, Beebe and Fromer’s Article should give courts a much better understanding of the difficulty of finding an unused, simple mark, and the resulting constraints that might cause a firm to choose a mark that is similar to others already in use.
II. Costs of Trademark Depletion and Congestion
Beebe and Fromer’s primary contribution is descriptive, but they also argue that the costs of depletion and congestion are significant enough to require policy reforms such as higher fees or congestion pricing, and a more stringent use requirement coupled with more rigorous auditing of the trademark register.22 “[T]he ecology of the trademark system is breaking down,”23 they say, and there is no “excuse for inaction.”24 They emphasize the costs to both firms and consumers, contending that new entrants face higher costs of entry or expansion and must settle for less effective marks, and that consumers face higher search costs due to firms choosing less memorable marks and due to the blurred distinctiveness of individual words.25
These hypothesized costs are both concerning and plausible. But there is not yet rigorous evidence that these mechanisms actually create a significant negative social welfare impact in practice. Further empirical research could help hone in on the actual costs of depletion and congestion, particularly given that market adaptations may already be mitigating what costs exist.
First, Beebe and Fromer’s argument lacks empirical support for their assumption that depletion and consumption force entrants to bear significant financial burdens. Beebe and Fromer cite news articles that quote entrepreneurs who have trouble naming new firms or products as their primary evidence.26 But this evidence alone cannot show how trademark depletion actually affects the average costs of entry. There might be ways to at least approximate the financial costs: one branding guide estimates the low-end cost of name development and logo creation at $13,500.27 What remain untested are the different financial costs that may result depending on the particular trademark class, the size of the market entrant, and the depletion or congestion level of the class.
Further empirical research would likely fill in these gaps in the research. For example, interviews with branding agents (or branding scholars) might clarify how much of these trademark-related financial costs to entry is driven by searching through existing marks, whether the cost is higher in more depleted classes, and how it has changed over time. Empirical research should also take into account the possible ameliorating effect of technological changes, including the increasing ease of using search engines to quickly and cheaply test the distinctiveness of potential marks.28 In addition, technological developments have reduced branding-related entry costs by making it easier to search existing registered marks29 or pay for low-cost searches,30 and by allowing firms to crowdsource logo design,31 connect online with low-cost freelance designers,32 or design their own logos.33
Another way to empirically tackle the problem of measuring the costs of depletion and congestion may be through comparisons with other countries.34 The United States is an outlier in requiring trademark use to maintain registration, so if depletion and congestion “have reached chronic levels”35 in the United States, trademark systems in other countries should presumably be on life support. If Beebe and Fromer are correct, one would thus expect trademark development costs to be a significant barrier to entry outside the United States. There is a vast literature on how variations in regulatory barriers affect entrepreneurship across countries;36 investigating how trademarks affect these barriers seems like an important avenue for further research.
Empirical work is also needed to understand the other hypothesized cost to new entrants: that after bearing this “ever-greater cost” to find a usable mark, applicants are forced to settle for “ever-less benefit.”37 Beebe and Fromer argue with some support from the branding literature that the areas of greatest trademark depletion and congestion — common words and neologisms — are also those with the most desirable trademarks.38 But it does not necessarily follow that a firm’s inability to use a common word or short neologism, such that it is left with a more complex mark than it initially desired, will have a significant impact on the firm’s value. The branding literature on “good”39 trademarks that Beebe and Fromer discuss does little to answer this question.
Whether more complex marks actually reduce firm value is linked to the corresponding concern for consumer welfare: do these marks in fact significantly increase search costs for consumers? This is a cogent hypothesis, as is the idea that congestion will harm consumers by blurring the distinctiveness of nonfamous marks. But given the difficulty scholars have had finding any concrete harm caused by blurring of famous marks,40 it seems worth empirical study. Furthermore, to determine whether reforms such as decluttering the register of underused marks will have any impact, it is necessary to determine how much of any consumer harm from depletion and congestion is caused by these inactive marks as opposed to a growing number of active brands.
Unfortunately, the most concrete evidence of consumer harm Beebe and Fromer present seems insufficient to take as proof that the current trademark regime reduces consumer welfare. The Article argues that “trademark congestion can literally kill”41 because “between eight and twenty-five percent of medication errors are attributed to name confusion.”42 There is reason to doubt this statistic.43 But in any case, the pharmaceutical industry is not a good illustration of the harms Beebe and Fromer are focused on; as they note, it has low rates of congestion and depletion, and different branding practices.44 Given the high non-trademark-related barriers to entry, it seems unlikely that reforms related to trademark pricing or use requirements would have much effect. Rather, any harms from drug-name confusion seem better addressed by the computer systems that are already being used to tackle this concern.45
Another reason to exercise caution before implementing major trademark reforms is that markets may already mitigate at least some of the costs of depletion and congestion. For example, Beebe and Fromer suggest that “trademark law can more generally learn” from the Food and Drug Administration’s efforts in the pharmaceutical industry.46 Perhaps one lesson from the pharmaceutical industry is that many harms of congestion and depletion can be addressed without any change in trademark law itself. If consumers have trouble keeping track of the exploding number of craft beers,47 they can keep track of favorites with the Untappd app — including by scanning barcodes rather than searching by name.48 Consumers can scan barcodes or take pictures of other products to see reviews and prices with shopping tools such as the Amazon app.49 The drawbacks of longer brand names also are less significant if consumers simply search for the desired product type using websites like Amazon or Google and then choose a product based on its prominence in the search results, giving them little reason to pay attention to the brand name.
The market for domain names may be instructive. As Beebe and Fromer note, appealing domain names are limited.50 But they also recognize that technological developments such as the availability of internet search tools and the creation of “new top-level domains” have alleviated many of the problems associated with the limited supply of the most desirable domain names.51 Similarly, new technologies may be able to reduce the negative externalities of constraints on attractive trademarks: if consumers can find products with unusual trademarks through search engines, then firms can decrease attempts to find the most desirable trademarks because they know consumers have alternative effective means to identify their goods.
In sum, there is not yet any concrete evidence that trademark depletion and congestion impose any significant harms for either firms or consumers. Without such evidence, reforming the trademark system seems premature and likely unresponsive to the actual challenges firms and consumers face. Of course, one response might be that policy reforms are still warranted simply to counteract a plausible risk of harm. But that would only be true if there were no offsetting benefits — something Beebe and Fromer never consider.
III. Potential Benefits of Depletion and Congestion
If trademark depletion and congestion in fact create costs for firms, they will benefit incumbents over entrants. Fromer and Beebe criticize these “anticompetitive costs”52 and argue that hindering entry is “inefficient.”53 This assertion is of course generally true: the entire field of antitrust law is devoted to removing certain barriers to competition that harm consumer welfare.54 But while consumers generally benefit from increased competition, economists have long recognized that barriers to entry can enhance efficiency in some circumstances.55 Here, I give just two examples of how any entry-related costs created by trademark depletion and congestion could have positive effects.
First, in a competitive market that already offers consumers a variety of choices, adding another choice to the mix may diminish consumer welfare. There is a burgeoning literature on the benefits of limiting consumer choice.56 As one scholar summarizes, “increased choice decreases satisfaction with matters as trivial as ice cream flavors and as significant as jobs” and can induce “choice paralysis” that makes consumers less likely to choose anything at all.57 Adding yet another craft beer or smartphone app to the market may simply make it more difficult for consumers to identify their preferred products, or may lead to an increase in consumer search costs that outweighs any corresponding gain from increased competition.58 If a given market is in fact running out of competitively effective trademarks, this trend may have the benefit of reducing lower-value choices. A high-value entrant can presumably bear the entry cost of finding a less effective mark or purchasing a mark from an existing business.59 A low-value entrant that offers no significant advantages in terms of cost or quality compared with the existing market might be deterred by these costs — but that’s not necessarily a bad thing.
Second, consider a very high-value entrant: a truly new and innovative product that has registered a mark in a class of goods with high depletion or congestion. Initially, first-mover advantage will enable this innovative firm to charge above-marginal-cost prices to appropriate returns on its research and development investments.60 Is it a problem if trademark law further discourages entry and helps the innovative firm maintain its first-mover advantage? From a static perspective, the answer is yes: discouraging entry of competing products that could drive down prices is a net loss for consumer welfare in the short term. But this loss might be offset by dynamic innovation gains, as I explore more fully in separate work with Jason George.61 Intellectual property law has long recognized the benefits of temporary monopolies for innovative firms, and if trademark protection is part of what allows an innovative firm to maintain its first-mover advantage, then it can provide an incentive to create those innovations in the first place.62 As economist Harold Demsetz has explained, reducing trademark protections for incumbents “is likely to reduce prices and increase sales of products already produced, but it also reduces incentives to develop new ones.”63
Of course, just as for utility patent law, this potential innovation benefit does not mean that trademark rights should be expanded without bounds — benefits for first movers have costs, such as impeding follow-on innovations by second movers. On net, Beebe and Fromer’s proposed reforms might even cause a net innovation gain. My point is simply that the current entry-related costs of trademark depletion could plausibly create an innovation benefit. As for other intellectual property law doctrines, a full economic analysis of a given trademark doctrine should consider all the costs and benefits, including the effect on innovation.64 A full economic analysis would permit more tailored reforms that could address the precise costs of depletion and congestion without losing all of the possible benefits of the current trademark regime.
None of the above discussion is meant to diminish the outstanding empirical contribution that Beebe and Fromer have made; as described above, their herculean efforts at merging and analyzing U.S. trademark registrations with other datasets have given trademark commentators a much richer picture of the available space for new trademarks. We are not running out of trademarks, but we are clearly running out of some trademarks — common words, syllables, and surnames. This understanding should inform courts faced with trademark law issues such as whether to infer bad faith from a firm’s choice of a mark similar to ones already in use.
Any normative theory of trademark law now needs to grapple with this evidence of the high state of depletion and congestion of the most sought-after trademarks. Future empirical study may well demonstrate that the resulting harms to new entrants and consumers outweigh any corresponding benefits, such as the benefits of reducing consumer choice and increasing innovation incentives.65 But given the lack of rigorous evidence regarding either the costs or the benefits of either depletion or congestion, much less the welfare effects of any particular policy change, it seems premature to recommend significant action. Concerns about the exhaustibility of competitively effective marks might end up being no weightier than John Stuart Mill’s worries about the exhaustibility of musical combinations.
* Associate Professor of Law, Stanford Law School. For very helpful comments, thanks to Daniel Ho, Mark Lemley, and the editors of the Harvard Law Review.
Recommended Citation: Lisa Larrimore Ouellette, Response, Does Running Out of (Some) Trademarks Matter?, 131 Harv. L. Rev. F. 116 (2018), https://harvardlawreview.org/2018/02/does-running-out-of-some-trademarks-matter/.