A letter to the Wall Street Journal about "Patently Absurd" on March 1, 2006
Lawrence B. Ebert
March 2, 2006
Ebert, who holds a Ph.D. from Stanford University, is a registered patent attorney, admitted in New York and New Jersey. He wrote on the patent grant rate issue in 86 Journal of the Patent and Trademark Office Society [JPTOS] 568 (2004), and is the author of LESSONS TO BE LEARNED FROM THE HWANG MATTER: ANALYZING INNOVATION THE RIGHT WAY, appearing in JPTOS in March 2006.
The figures for patent approval rates were published in "Things are not always what they seem to be," Intellectual Property Today, October 2005, available on LEXIS. Director Dudas has mentioned the numbers in talks in the last few months.
Discussion of the erroneous nature of the adjusted grant rates of Quillen and Webster appears in 86 JPTOS 568 (2004), 4 CHI-KENT J. INTELL. PROP. 104 (2004), and 4 CHI.-KENT J. INTELL. PROP. 186 (2005), all available on LEXIS.
Lunney's paper is Glynn S. Lunney, Jr., Patent Law, the Federal Circuit, and the Supreme Court: A Quiet Revolution, 11 S. Ct. Econ. Rev. 1 (2004), available on LEXIS.
One reader stated the USPTO reported a grant rate of 61.3% for 2004, rather than the 62.5% I noted. This level of variation is comparable to the differences in the numbers used by Quillen/Webster on the one hand and Clarke on the other. While it is a bit disconcerting that we don't have one set of numbers, variations of a few percent are not what I am worried about. I am concerned with the difference between the raw numbers of 62 to 70% reported by the USPTO in comparison to the "about 90%" asserted by the Wall Street Journal.
The reader also asserted that the USPTO’s report of grant rates is disingenuous - and has long been disingenuous. I am not sure that I agree with this. There are several issues. The USPTO has been reporting raw data. Apart from concerns, as noted above, that there seems to be some variability in numbers that ought to be constant, there is little doubt that the raw numbers are, and have been, in the range 60 to 70%. Quillen and Webster obtained corrected grant rates by subtracting, from abandonments, the number of continuing applications filed, on the theory that continuing applications reflected repeated attempts to claim the same invention. This is totally wrong as to divisional applications. It is likely wrong for most continuations-in-part. Even as to continuations, it is likely that the continuations contain claims of different scope than found in the parent application. The central contention of Quillen and Webster that continuing applications reflect repeated attempts to claim the same invention is wrong in almost all cases.
Of the assertion: "When the public hears an announced grant rate of “61.3%,” the public presumes that over one third of all patent applicants are being turned away by the USPTO and exit the system empty-handed," one has to contemplate what is meant by a grant rate. If the applicant tries twice, in different applications with claims of different scope, and succeeds only once, with claims of narrower scope, and with correlative prosecution history estoppel, is his batting average 1.000 (Quillen/Webster) or .500? There is an arbitrary element here, in that the statistics are based on "application" rather than the narrower "amendment" or broader "family," but that is how the numbers are kept. Grant rate is defined in terms of applications.
The continuation issue has been discussed many times on IPBiz, including
discussion on January 13.
In the first paper by Quillen/Webster [11 Fed. Cir. B.J. 1], footnote 17 said
n17 The authors have been told that in some circumstances it is possible to file a continuation application without immediately abandoning the parent application, and for the unabandoned parent application subsequently to mature into a patent, even though the continuing application and the parent application (which subsequently becomes a patent) are for the same invention. It is not possible to tell the number of such patents, if any, from the PTO data. To the extent there may be such patents, the Allowance Percentages and Grant Rates, as calculated in Tables 2 and 6 respectively, could be affected. See infra tbl. 2, tbl. 6.
As one of ordinary skill in the patent art knows, and as Quillen and Webster disclosed in their second paper, it is well-known for a patent to issue on a continuation application based upon a parent application for which a patent also issues. Upon reading footnote 17, one of ordinary skill in the patent art would recognize that the 97% number in the first Quillen and Webster paper was wrong, in the year 2001 or in the year 2006.
One recent writer explained the statement of Quillen and Webster in the following manner.
I’ve had occasion to review many patent families created via the continuation process. My personal observation is that once the patent attorney finally develops a limitation that he or she can point to as distinguishable over the prior art of record before the patent examiner, then the trick thereafter is to carry that limitation through in claims of subsequent children continuations (or even divisionals, by the way) and manipulate the claims in other ways. Once that happens, then the allowance of subsequent continuations is guaranteed – which I think is the true point that all of us know and QW were attempting to establish by statistics – that is, continuation filings guarantee patent issuances.
Of course, this does not explain "why" Quillen and Webster used a model in the 2001 paper which assumed that one could not have a patent on both a parent and continuation application, an assumption counter the knowledge of one of ordinary skill in the patent art. The falsely made assumption led to false patent grant rate numbers which have been cited in the Harvard Law Review, in the University of Pennsylvania Law Review, in the eBay brief to the Supreme Court, and in the Wall Street Journal.
Separately, the predicate sentence for footnote 17 was: Since the second application is for the same invention as the prior application, the prior application is normally abandoned after the second application is filed. n17 As the Supreme Court said in Aro and the CAFC in Innova, the claims define the invention. A patent on a continuation application is by definition not for the same invention as in the parent, otherwise there would be a rejection in the continuation application for statutory double patenting. In the continuation application, there might be an obviousness-type double-patenting rejection (and a terminal disclaimer) but inspection of cases shows that this is not always true. One notes that a Supreme Court case cited by the Innova court directly negates the Quillen/Webster position:
Smith v. Snow, 294 U.S. 1, 11, 79 L. Ed. 721, 55 S. Ct. 279 (1935) ("The claims of the patent, not its specifications, measure the invention.")
Separately, from my second paper in the Journal of Intellectual Property, in a discussion of the Lemley/Allison data used by Quillen/Webster:
The probability of a parent patent on a continuation, 13% (=19/141), is significantly lower than a parent patent on a divisional/cip, 47% (=73/156). The text of QWII is silent on patent families with two or more issued continuations, whether or not the parent issued as a patent. The important number for the adjustment in QWII is the 31%, from which the authors inferred that in 69% of the cases (100% - 31%) there was no prior (or in fact other) patent associated with the continuing patent.
The bottom line: the Wall Street Journal should not have asserted that the USPTO grants patents on about 90% of submitted applications. There is NO DATA WHATSOEVER to support this, and there is data that shows this assertion is NOT TRUE.