Is the Trademark Office a Rubber Stamp?

What proportion of trademark registration applications submitted to the U.S. Patent and Trademark Office (PTO) actually result in registration? Surprisingly, we have long lacked an answer to this very basic question. We thus lack information crucial to assessing, among much else, overall "trademark quality" at the PTO-that is, the reliability of registration status as an indication of actual trademark validity.

1 Furthermore, without some knowledge of trademark application grant rates at the PTO, we remain unable fully to understand the calculus that trademark holders apply in deciding to file-or not to file-a trademark registration application.

Working from a previously unstudied dataset observing each of the some five million trademark registration applications submitted to the PTO from 1981 through 2010, this Article seeks to address this fundamental gap in our knowledge. It reports and analyzes trademark application grant rates at the PTO along a variety of dimensions, including by the year and filing basis of the application; by the type of mark that is the subject of the application and the category of goods or services with which the mark is associated; and by the type and country of origin of the commercial entity applying for registration. While it must be recognized from the start that there is no realistic way of establishing "correct" trademark registration grant rates as baselines against which the reported results may be compared, 2 even a basic understanding of the PTO's grant rates, which has up to now eluded us, will aid in evaluating PTO trademark quality and the motives of trademark applicants.

Part I provides background. It first explains why trademark registration matters: federal registration confers on trademarks 1.

The term is adapted from patent law. See, e.g., R. Polk Wagner, Understanding Patent-Quality Mechanisms, 157 U. PA. L. REV. 2135, 2138 (2009) ("Patent quality is the capacity of a granted patent to meet (or exceed) the statutory standards of patentability-most importantly, to be novel, nonobvious, and clearly and sufficiently described."); see also Katherine J. Strandburg et al., Law and the Science of Networks: An Overview and an Application to the "Patent Explosion," 21 BERKELEY TECH. L.J. 1293, 1338 (2006) (finding an "increasing stratification of patent citability" and observing that "[a]necdotal and survey evidence suggests that patent quality has been decreasing in recent years, resulting in the issuance of a larger fraction of more trivial-and hence less citable-patents").

2. Cf. Mark A. Lemley & Bhaven Sampat, Is the Patent Office a Rubber Stamp?, 58 EMORY L.J. 181, 186 (2008) ("We do not, and cannot, know the 'right' grant rate, or how many continuations are being used for abusive as opposed to legitimate purposes. If the grant rate were 97%, that could still be 'too low' if it turned out that every application filed was meritorious. Similarly, a grant rate of 5% could be too high if none were meritorious."). substantial procedural and substantive advantages. It then briefly reviews the trademark registration process and describes the nature of the dataset. Part II reports the main findings of the study. Part III proposes an interpretation of these findings. Part IV concludes.

Use, rather than registration, is the basis of federal trademark rights in the United States. The Lanham Act will protect a trademark owner's exclusive rights in any trademark it is using in commerce, regardless of whether the mark is registered, provided that the unregistered mark otherwise meets the various substantive requirements for registration established by the Act. 3 Conversely, the Lanham Act will not protect a trademark registrant's exclusive rights in its mark if it no longer uses its mark in commerce and cannot prove an intent to resume use in the near future. 4 On this basis, it is often said that the U.S. trademark system is a "use-based" system in contrast to the "registration-based" systems more common around the world. 5 In the United States, registration merely records the preexistence of externally established rights, or so goes the conventional wisdom. 6 In fact, even in the American use-based system, registration confers significant, substantive advantages on the registered mark, advantages that run far beyond the exclusive rights established by mere unregistered, "common law" use. Most notably, under section 7(c) of the Lanham Act, federal registration provides the trademark registrant with a "right of priority, nationwide in effect," with respect to the registered mark against anyone else in the nation who uses the mark after the date of application. 7 This right of priority extends nationwide even if, as is often the case, the registrant has not itself used the mark throughout the nation.

8 Indeed, this nationwide right of priority obtains even if, as is explained in the next section, the registrant merely had at the date of application an "intent to use" the mark in the near future somewhere in the nation, but had not yet done so. By contrast, exclusive rights in an unregistered mark extend geographically only so far as the owner is in fact using the mark. 9 A second substantive advantage that flows from registration is also highly significant, especially for "descriptive" marks. 10 Sections 15 and 33 of the Lanham Act provide that at any time after the fifth year of the mark's registration, the registrant may file a "declaration of incontestability" that limits the grounds upon which the mark's validity may be contested for the remaining life of the registration (which may be renewed in perpetuity). 11 Sections 15 and 33 explicitly list out these grounds; any that are not listed are foreclosed.

12 As the Supreme Court noted in the notorious case of Park 'N Fly, Inc. v. Dollar Park and Fly, Inc., among those grounds not listed is the claim that the registrant's descriptive mark lacks "secondary meaning" 13 as a designation of source. 14 Accordingly, the Court held that under 7. 15 U.S.C. § 1057(c) (2006). 8. Id. 9. See Emergency One, Inc. v. Am. Fire Eagle Engine Co., 332 F.3d 264, 269 (4th Cir. 2003) ("[T]he owner of common-law trademark rights in an unregistered mark is not entitled to injunctive relief in those localities where it has failed to establish actual use of the mark."). 10. A descriptive mark "immediately conveys information concerning a quality or characteristic of the product or service." In re MBNA Am. Bank, N.A., 340 F.3d 1328, 1332 (Fed. Cir. 2003). For this reason, U.S. trademark law assumes that consumers unfamiliar with such a mark will perceive the mark as a mere description of the product to which it is affixed rather than as a designation of the source of that product. See id.

("The perception of the relevant purchasing public sets the standard for determining descriptiveness."). However, if the mark holder can show that the mark has developed over time a "secondary meaning" as a designation of source, then U.S. trademark law will afford protection-and registration-to the mark. (1982) ("To establish secondary meaning, a manufacturer must show that, in the minds of the public, the primary significance of a product feature . . . is to identify the source of the product rather than the product itself.").

14. Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189, 196, 201 (1985).

12/22/2011 11:41 AM 2011] TRADEMARK OFFICE, A RUBBER STAMP? 755 the clear terms of the Act, the incontestable-and clearly descriptive-mark PARK 'N FLY for airport parking services could not be challenged on the ground that it lacked secondary meaning, notwithstanding that the record below strongly suggested that the mark lacked secondary meaning at the time of registration in 1971 and still lacked secondary meaning when the case was being litigated in the early 1980s. 15 In his lengthy dissent, Justice Stevens expressed his dismay that the decision of a single trademark examiner in an ex parte proceeding a decade earlier, followed by the registrant's perfunctory filing of a declaration of incontestability, could somehow prevent the Court from striking from the Principal Register an "inherently unregistrable" mark. 16 He also added his own opinion of trademark quality at the PTO for good measure. 17 Registered marks also enjoy significant procedural advantages. Section 33 of the Lanham Act specifies that registration is "prima facie evidence of the validity of the registered mark[,] . . . of the registrant's ownership of the mark, and of the registrant's exclusive right to use the registered mark in commerce on or in connection with the goods or services specified in the registration."

18 While the various circuits apply Section 33 differently, 19 many hold that, in addition to shifting the burden of production on the issue of validity to the defendant, 20

Because the limited purpose of this Article is to report ultimate trademark registration grant rates at the PTO, we need not detail every aspect of the process that leads to registration. Nevertheless, a brief (and necessarily noncomprehensive) review of certain features of this process will aid the reader in interpreting the results reported below.

To apply for trademark registration, the applicant must specify the statutory "basis" for its application-and as we will see, grant rates vary significantly across certain of these bases. The Lanham Act currently sets forth five filing bases, of which the first two are by far the most commonly used: (1) current use of the mark in commerce under section 1(a); 22 (2) a bona fide intent to use the mark in commerce under section 1(b); 23 (3) a claim of priority based on an earlier filed foreign application under section 44(d);

24 (4) ownership of a registration of the mark in the applicant's country of origin under section 44(e); 25 and (5) extension of protection of a "Madrid system" registration under section 66(a).

26 Two details of this scheme are worth emphasizing. First, the first four filing bases are not mutually exclusive; 27 66(a), by contrast, may not be combined with any other filing basis.

28 Second, the Sections 1(b) and 66(a) filing bases are relatively recent innovations. While the Sections 1(a), 44(d), and 44(e) filing bases have been available since the effective date of the original Lanham Act on July 5, 1947, the 1(b) intent-to-use (ITU) filing basis became available with the effective date of the

Trademark Law Revision Act (TLRA) on January 1, 1989, 29 and the 66(a) filing basis became available with the effective date of the Madrid Protocol Implementation Act on November 2, 2003. 30 A successful application proceeds through at least five stages: application, examination, publication, opposition, and registration. 31 Since October of 1998, the PTO has accepted trademark applications filed electronically, 32 and applicants can now quite easily file online with a credit card. 33 In contrast to the patent application process, trademark registration filing fees are relatively low. A basic application might cost approximately $500. 34 This helps to explain the curious-and increasingly notorious-phenomenon of what might be termed "meme mark" filings, in which all manner of applicants rush to register various words and phrases (such as "Let's Roll" 35 or "Seal Team 6" 36 ) very soon after their appearance in the media. Upon receipt of the application, a PTO Trademark Examining Attorney examines the application to determine whether it meets the procedural and substantive requirements for registration under the Lanham Act. 37 Among these requirements are that the mark possess either inherent or acquired distinctiveness of source, 38 that it not be confusing with some previously used or registered mark, 39 that it not be deceptive 40 or "deceptively misdescriptive" 41 without secondary meaning, and that it not be either scandalous or disparaging of some third party.

If the examiner determines that the application fails to meet any of the Lanham Act's requirements, the attorney will issue an "office action," to which the applicant has six months to respond. 43 The examiner and applicant may continue to communicate via office actions and responses (which may result in amendments to the application) until the examiner either approves the mark for publication in the PTO's Official Gazette or issues a "final office action" denying registration to the mark. 44 The applicant may appeal the final action to the Trademark Trial and Appeal Board (TTAB), a tribunal within the PTO consisting of administrative judges. 45 If the appeal is successful, the application proceeds to publication.

In the third and fourth stages, the application is published in the Official Gazette, after which a thirty-day opposition period begins in which "[a]ny person who believes that he would be damaged by the registration of [the] mark" may file with the PTO a notice of opposition to its registration. 46 The TTAB adjudicates all oppositions. 47 In the event that no opposition is filed or all oppositions are dismissed by the TTAB, Section 1(a) use-based applications proceed directly to registration, as do Section 44(d), 44(e), and 66(a) applications. 48 In the case of Section 1(b) ITU applications, however, the PTO will issue a notice of allowance, which states that the PTO will register the mark after the applicant has filed a satisfactory statement of use of the mark. 49 The applicant must file its statement of use within six months from the date of the notice of allowance, though this period is extendable for good cause up to an additional thirty months.

The data reveal the amount of time the registration process typically takes. From 1981 to 2004, the process for successful 1(a) use-based applications took an average of 16.7 months (SD=9.8) from the date of application to the date of registration. From 1989 to 2004, successful 1(b) ITU applications took an average of 29.3 months (SD=13.1). The notice of allowance and statement of use process explains the longer period of time for ITU registrations. From 2005 through 2010, the PTO has dramatically quickened the speed of the registration process. For this six-year period, successful use-based applications took an average of 12.0 months (SD=6.0), while successful ITU applications took an average of 21.5 months (SD=9.5). This is a remarkable increase in office efficiency, particularly in light of the fact that the per annum ratio of trademark applications to trademark examiners from 1995 to 2004 was about 770 to 1, while the ratio from 2005 to 2010 was only slightly lower, at about 726 to 1.

In June 2010, Under Secretary of Commerce and Director of the PTO David Kappos announced that the PTO would make publicly available bulk patent and trademark data for free

Figure 1 shows general trends in the number of trademark registration applications filed at the PTO per year from 1981 through 2010. Over this thirty-year period, total trademark applications at the PTO have more than quintupled, with 47,633 applications in 1981 rising to a high of 306,370 in 2007. As Figure 1 shows, the upward trend in annual trademark applications tracks the expansion of U.S. GDP. The dramatic spike in total applications in 1999 and 2000 coincides with the Internet bubble. 59 As stated above, the Section 1(b) intent-to-use filing basis became available on November 16, 1989. By 1992, it had overtaken the Section 1(a) use-based filing basis as the most common basis for filing. 60 The annual number of Madrid System applications received at the PTO remains disappointing, with at best a meager 15,368 applications claiming a Section 66(a) filing basis in 2008.

The overall registration rate for all applications filed from 1981 up to November 16, 1989, was .77. The rate then dropped to .53 for all applications filed from November 16, 1989, through 2007. 61 As Figure 2 shows, the emergence of the intent-to-use filing basis explains this pronounced break. The overall registration rate for all use-based applications filed from 1981 through 2007 was .75; for ITU-based applications filed from November 16, 1989, through 2007, 62 the rate was .37. As I will discuss more fully in the next section, this difference results not from the PTO's rejection at the examination stage of a relatively large proportion of ITU applications, but rather 61. The period of study ceases in 2007 in order to filter out applications that were still active (i.e., not rejected, accepted, or abandoned) as of 2010. The registration rate portion of the study does not include marks that ended up on the Supplemental Register.

62. For purposes of strict comparison, note that the overall registration rate for use-based applications filed from November 16, 1989, through 2007 was .74.

TRADEMARK OFFICE, A RUBBER STAMP? 763 from the failure of a large proportion of ITU applicants to file a statement of use after the PTO has approved the application for publication. As for filing bases not shown in Figure 2, Section 44(e) applications based on a prior foreign registration boasted a registration rate of .87 for the period 1981 through 2007, while the proportion for 44(d) applications was .72. For the period November 2, 2003, through 2007, the proportion for 66(a) applications was .76. Trademark examination standards are roughly similar among the trademark registration offices of the world, so it makes sense that 44(e) applications would enjoy such success at the PTO. 63 Finally, Figure 2 shows a slight dip in registration rates for use-based and ITU-based applications filed in 1999 and 2000. This is not surprising. In these years, the PTO received a flood of dubious trademark registration applications. 64 Table 2 shows registration rates by mark type and filing bases for the period November 16, 1989, through 2007. In Table 3 shows the registration rates for use-based and ITU-based applications filed from November 16, 1989, through 2007 from the leading sixteen countries of origin listed in Table 1. Applications from a wide variety of foreign countries enjoy registration rates substantially higher than do applications from the United States. It is unclear why this should be the case. Perhaps foreign applicants filing in the United States are, on average, more sophisticated than domestic filers, given the effort, costs, and knowledge required to file from abroad. For related reasons, foreign filers may hesitate to apply to register marks that are less likely to succeed. Table 4 shows registration rates by the five most common entity types and by filing basis for applications filed from November 16, 1989, through 2007. Unsurprisingly, applications filed by individuals did comparatively poorly, but it is not clear why partnerships did equally as badly.

While the main purpose of this Article is to report trademark registration rates at the PTO, registration rate data do not tell the full story. This is because some applicants fail to register their marks even after the PTO has completed its substantive review of the application and approved the mark for registration. As described above, the trademark registration process proceeds through the substantive examination stage early on in the process. 69 If the examining attorney determines that the application is substantively sound, then the application is published in the Official Gazette to give notice to third parties who may wish to oppose the application. Generally speaking, the fact that the mark has been published in the Official Gazette shows that that mark has survived PTO examination. 70 For this reason, it is appropriate to 69.

ITU applications may fail to survive PTO review, however, after publication in the Official Gazette if the statement of use raises issues barring registration, such as that the mark is being used descriptively. See 2 J. THOMAS MCCARTHY, MCCARTHY ON TRADEMARKS AND UNFAIR COMPETITION § 11:51.50 (4th ed. 2011) ("[T]here is, in the vast majority of cases, no way for the applicant to . . . prove the acquisition of distinctiveness through secondary meaning.").

assess not simply applications' registration rates, but also their publication rates.

The overall publication rate for all applications filed from 1989 through 2007 was .76. This proportion holds across the two major filing bases: for use-based applications filed from 1989 through 2007, the rate was also .76, and for ITU-based applications filed from November 16, 1989, to 2007, the rate was again .76. Thus, in contrast to the pronounced difference in registration rates for use-based and ITU-based applications, their publication rates are essentially the same. I explain this parity in publication rates and divergence in registration rates in the next subpart, where I provide data on how applications failed.

Figures 3,4, and 5 compare the proportion of applications published for opposition during the indicated periods to the number of applications filed. Each figure shows a slight decline over time in publication rates against a marked increase in filings. HOUSTON LAW REVIEW [48:4 failed. We cannot therefore glean from the dataset what proportion of marks were rejected as, for example, generic, or merely descriptive without secondary meaning, or confusing with a previously registered mark pursuant to section 2(d) of the Lanham Act. 72 However, the dataset does record the final "status code" of each application, which may offer some hints as to why particular applications failed. Tables 6 and7 set forth the final status codes for use-based applications that failed to survive to publication or to registration, respectively. Tables 8 and9 do the same for ITU applications. To facilitate like-for-like comparison, Tables 6 through 9 address applications filed from the effective date of the TLRA through 2007.

Applicants filed 1,336,759 use-based applications during this period. 73 Of these, 1,014,386 (75.9%) were published. Table 6 shows that of the 322,373 applications that failed to survive to publication, 4,127 (1.3%) were abandoned after an ex parte appeal-in other words, after what was very likely an unsuccessful appeal to the TTAB of the examining attorney's refusal to publish the application. Of published use-based applications, 991,151 (97.7% of 1,014,386) were registered. Table 7 shows that of the relatively small number of applications that were published but failed to survive to registration, 55.0% were abandoned after what was very likely a successful third-party opposition to the application before the TTAB. Applicants filed 1,951,993 ITU applications from the effective date of the TLRA through 2007. Of these, 1,483,435 (76.0%) were published. With respect to the 468,558 ITU applications that failed to survive to publication, Table 8 shows that 5,971 (1.3%) were abandoned after an ex parte appeal. So far, the use-based and ITU data are remarkably similar. After publication, however, use-based and ITU applications produce very different data. Of ITU applications that were published, only 719,635 (49.0% of 1,483,435) were registered. Table 9 shows that of the large proportion of ITU applications that were published but failed to survive to registration, 84.4% failed because the applicant failed to file a statement of use. ITU applicants who failed to file a statement of use largely account for the severe differences in the publication and registration rates reported in Tables 1 through 5.

This study presents two main findings: First, the registration rate for all applications filed from 1981 up to November 16, 1989, was .77, and then the rate dropped dramatically to .53 for all applications filed from November 16, 1989, to 2007. Second, the publication rate for all applications filed from 1981 through 2007 was .76 and remained reasonably steady over the course of that period.

These findings raise a host of questions. Some of these questions go to the objective issue of trademark quality at the PTO. Most importantly, is the PTO's publication rate of .76 unduly high, or indeed, unduly low? Stated more generally, given the standards set by the Lanham Act and the courts, are the PTO's trademark examination practices too permissive or not permissive enough in applying these standards? A second, separate set of questions goes to the issue of applicant strategy at the PTO. Why is the proportion of applications that the PTO accepts for publication not closer to 50%? After all, given the significant advantages that flow from and the relatively low cost of registration, it might make sense for trademark holders to file for registration if they face a better than 50% chance of succeeding. Significantly complicating this question, however, is the fact that since 1989 overall registration rates at the PTO have, in fact, hovered just above 50%.

As to the first set of questions going to trademark quality, it is worth emphasizing that short of reexamining in some sufficiently objective way the validity of every trademark application submitted to the PTO over the past three decades (or some large sample thereof), it is very difficult to establish a

TRADEMARK OFFICE, A RUBBER STAMP? 775 baseline against which the 76% publication figure can be compared. For many years, the PTO has conducted its own internal quality control checks, in which a senior examiner reviews other examiners' processing of applications, and has reported the results. 74 Substantive or procedural errors are typically found in no more than about 3% of applications per year. 75 This internal review process, however, is open to the charge that the internal reexamination may suffer from the same misapplication of the Lanham Act and courts' rulings as the initial examination. At this stage of our understanding of publication rates at the PTO, it is not probable that a persuasive argument can be that the PTO's trademark application publication rate indicates an over-permissive or under-permissive examination process. On the evidence available to us, we simply do not know-or should at least presume, as I am prone to, that the PTO has gotten it about right. But, hopefully, this Article's reporting of the PTO's publication rates over time may at least help us to begin to understand the issue.

As to the second set of questions going to applicant strategy, here we may at least advance somewhat our understanding of the calculus that trademark holders apply to the decision of whether to file a registration application. In this connection, let us assume that the PTO makes more mistakes than it realizes and that, according to some objective measure, even as low as only 65% (rather than 76%) of recent trademark registration applications actually deserved publication. Even if we accept this extraordinary assumption-which would mean that in 2007, for example, the PTO erred in its examination of nearly 30,000 applications-we are still left with the question of why the publication rate remains significantly higher than 50%. In other words, taking the PTO's examination standards as given, have trademark registration applicants on the whole been too conservative in filing only those applications that are especially likely to meet these standards? Where, in short, is the speculative filing that would drive the publication rate closer to 50%?

The answer may derive from the common law basis of trademark rights in the U.S. system and from the adverse effect 74. See 2009 U.S. PAT. & TRADEMARK OFF. ANN. REP. 37 (discussing the PTO's "in-process review" as a measure for achieving examination quality).

75. See id. at 18 (reporting the results of the PTO's internal review of trademark examination quality for Fiscal Year 2009, in which 96.4% of first actions and more than 97% of final decisions "met statutory and compliance rates for quality of decision making and writing"). HOUSTON LAW REVIEW [48:4 of a PTO rejection of a trademark application. 76 Recall that federal trademark rights are ultimately based on use rather than registration, so that federal trademark law will protect marks that are used in commerce even if they are not registered.

Registration provides a variety of significant advantages, but it is not a prerequisite for protection. 78 Consider, at the same time, that if the PTO rejects an application on substantive grounds, this rejection may function as very strong, if not dispositive, evidence of the invalidity of a mark for which the mark holder might later attempt to claim federal common law protection.

Faced with balancing the risk of a rejected application against the benefits of registration, the mark holder may rationally forego the opportunity to register a mark unless it is reasonably certain that the application will succeed. Indeed, for marks on the borderline between validity and invalidity, the mark holder might very well prefer to litigate the issue of validity in the context of an infringement action, where the defendant's bad faith might influence the court's validity determination. This calculus, together with the effort and cost entailed in applying to register, may help to explain why the PTO's publication rates are substantially above 50%. What then of the overall 53% registration rate for all applications filed from 1989 through 2007? This figure would appear to be consistent with the proposition that applicants engage in speculative filing of even borderline applications. However, in light of the degree to which this figure is influenced by the large proportion of ITU applicants who fail to file a statement of use, it probably tells us little. To be sure, these ITU applicants appear to be engaged in a kind of speculative filing in that they may never actually use the applied-for mark in commerce, but their publication rates suggest that with respect to the consequences of a possible substantive rejection of the

TRADEMARK OFFICE, A RUBBER STAMP? 777 mark, they are on balance as risk averse as those who file on other bases.

As an empirical matter, this Article has applied straightforward statistical methods to a very authoritative dataset finally to answer the question of what proportion of trademark registration applications are accepted at the PTO, either in the form of the PTO's decision to publish the application in the Official Gazette or, more generally, in the PTO's actual registration of the mark. As a theoretical matter, however, this Article has raised many more questions than it has answered: Why is the publication rate so high and so uniform across the major filing bases? Why do such a large proportion of ITU applicants abandon their mark after it has been approved for publication? And why has this proportion varied so little over time? If these ITU applicants in particular are speculative filers, then why is their publication rate nevertheless the same as that for applicants filing under other bases? Hopefully, further research on the PTO's application data will provide persuasive answers to some of these questions.