Remarks of Commissioner for Trademarks Lynne Beresford

From: Lynne Beresford, Commissioner for Trademarks

To: April 26, 2010 Roundtable participants

What future for the "use-based" Principal Register?

After the In re Bose Corporation  decision,  580 F.3d 1240, 91 USPQ2d 1938 (Fed. Cir. 2009), there has been serious discussion within the USPTO about the future topography of the US Principal Register. In Bose, the CAFC held that it is insufficient for a party alleging fraud to prove only that another "should have known" of the falsity of its statement, and that there must be proof of a false, material representation made "with the intent to deceive the PTO." In the months since the decision issued, there has been ongoing discussion in the trademark community about the impact of the decision on the federal trademark registration system.

The upcoming panel discussion will focus on the following questions: First, is there any concern about indirect effects of the Bose decision on the U.S. trademark register? Might the higher threshold for proving fraud reduce the risks associated with filing inaccurate identifications of goods or services? Put another way, has the Bose decision raised concerns that applicants and registrants will be more likely to claim use?

At least anecdotally the folks who work in the Office believe that the identifications of goods and services are getting longer and longer and include more and more goods and services. Of course, that in itself is not a bad thing. The worry is that a good number of the very long recitations of goods and services that are being used in applications no longer accurately reflect the "use" or "intent to use" of the applicant for all the goods and services recited.

Why does this matter? In terms of relative examination, the use based system coupled with rather narrow identifications, has enabled the register to grow, there are now more than 1.5 million registered marks. When registrations have very broad and encompassing identifications, they naturally can be cited against more applications.

Further, looking at the experience of Denmark and the United Kingdom, when identifications become too broad and too inclusive, relative examination becomes far more difficult if not impossible. As a result of the European Union's practice of accepting very broadly written identifications, the U.K. and Denmark found that relative examination became impracticable because too many marks were unregistrable due to blocking applications and registrations. When goods and services are described narrowly and precisely, many times there is no likelihood of confusion as to source, and similar marks can co-exist. However, when many or most registrations contain broad terms or long lists of items, the number of blocking marks increases dramatically. Examiners in Denmark and the U.K. continue to provide trademark applicants with a search report, but do not refuse registration, instead putting the burden on trademark owners to remove conflicting marks from the register through inter partes  proceedings. Given the present trend, this same change could take place here sometime in the future. Is this a result that the U.S. trademark community wants?

It's obvious that very broad identifications limit access to the register. By claiming a very broad range of rights, the applicant/registrant creates a very broad area that will be avoided by anyone making a wise business decision and wanting to avoid needless litigation. That is an appropriate result where there is bona fide intent to use or actual use for each and every good or service listed, but otherwise is not appropriate in the U.S. use-based system. Broad and inaccurate identifications present an unnecessary opportunity cost for new entrants to the market.

The Trademark Organization has conducted historical research in its databases to identify changes in the size of identifications of goods and services over the years. The attached charts and graphs are meant to answer some of the questions that trademark owners might have about trends in identification practice. Inasmuch as the USPTO does not allow class titles or similar broad terms to be used, applicants who want the same broad protection that is provided by a class title or general terms that cover a wide variety of items must use very long identifications for goods and services.

Chart 1 looks at the number of classes in registrations for the past 20 years. It's clear that the number of multi-class applications is growing.

Chart 2 shows the number of characters in identifications of goods and services in 2009 by bases. Clearly, the majority of large identifications are in registrations based on the Paris convention (section 44) and the Madrid Protocol (section 66).

Chart 3 shows the change in the length of identifications filed based on use in commerce (section 1(a)) for registrations issued between 1985 and 2009. Here one sees a slight rise in percentage of long identifications.

Chart 4 shows the length of identifications in registrations issued between 2005 and 2009 where the filing basis was the Madrid Protocol (section 66). The graph makes it plain that more than a third of the Madrid applications have ID lengths greater than 600 characters. Comparing Chart 4 to Chart 5, it's clear that registrations during the same period where the initial USPTO application was filed through TEAS Plus, have less than 5% of identifications that exceed 600 characters.

Chart 6 shows the percentage of all registrations, by year, with their identification character lengths.

Charts 7, 8, 9, 10 and 11 compare by basis certain classes and the length of characters in their identifications in 2009. These charts were added to give participants in the roundtable an idea of how class lengths vary from class to class within the system.

Chart 12 looks at the absolute number of registrations per class in 2009 that have character lengths in excess of 600 characters. One should note that this chart looks at a single class, not the whole registration, and looks at the character lengths in that class.

Chart 13 looks at the character lengths in Class 35 in 2009.

Finally charts 14 through 16 show the deletions from registrations on a year-by-year basis. These charts attempt to answer the question of whether, if identifications in applications do not reflect actual use or intent to use, there will be an increase in the numbers of deletions from the registrations when actual use must be shown at the time the section 8 or renewal is filed. This series of charts shows that there is an upward trend in identification size but no trend toward more deletions from the registrations. The Trademark Organization continues to do research in this area.

Some questions for the panelists to consider:

Whether a register with accurate lists of goods and services and minimal "deadwood" is a critical feature of the U.S. registration system;

Whether the Bose decision will result in less vigilance by applicants and registrants regarding the accuracy of the goods and services listed in their USPTO applications, registrations, and registration maintenance documents;

Whether the  Bose decision makes it more difficult to force trademark owners to remove inaccurate goods and services from their USPTO applications, registrations, and registration maintenance documents;

Whether the potential adverse effects of the Bose decision on the USPTO register and on trademark practice is acceptable to the trademark community;

What steps, if any, should be taken by the USPTO or Congress to change or further clarify the standard for proving fraud; and

Whether there should be alternative mechanisms or penalties, other than fraud claims, to ensure the accuracy of IDs and the integrity of the register.

There is clearly a trend towards larger identifications of goods and services. Should the bar and trademark owners be concerned?

What does this trend mean for small businesses, large trademark owners, the bar and the trademark system as a whole?