Although not yet out as a published law review, the paper by Mark A. Lemley and Carl Shapiro, Probabilistic Patents, is already cited in the law review literature, given below. Further, the discussion by Lemley and Shapiro on the patent grant rate work of Quillen and Webster merits commentary.
Papers citing Lemley and Shapiro:
Joseph Farrell and Robert P. Merges, Incentives to Challenge and Defend Patents: Why Litigation Won't Reliably Fix Patent Office Errors and Why Administrative Patent Review Might Help,
19 Berkeley Tech. L.J. 943 (2004)
-->Thus, for instance, if there are five infringers of equal size, each gets only a fifth of the gains from a successful challenge because each is paying only a fifth of the patentee's total royalties. Therefore, the patentee has five times more incentive to prevail in litigation than any one challenger has. Professor Miller and others have noted this problem and suggested policies to address it such as permitting infringers who compete with one another to
coordinate a legal challenge to a patent,n24 offering a bounty to a successful challenger, n25 and relying on fee-shifting. n26
n24. Mark A. Lemley & Carl Shapiro, Probabilistic Patents, 19 J.
Econ. Persp. (forthcoming 2005) (manuscript at 17, on file with authors), available at
Joshua D. Sarnoff, Abolishing the Doctrine of Equivalents and Claiming the Future After Festo, 19 Berkeley Tech. L.J. 1157 (2004)
-->Even when patents do not convey market power, patentees may exploit uncertainty regarding the scope of patents to deter competition by posing the threat of high-cost infringement litigation. n174
n174. Injunctions to remove products from the market are rare. See
Mark A. Lemley & Carl Shapiro, Probabilistic Patents, at 21 (July 14, 2004), available
at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=567883 (citing
Carl A. Shapiro, Antitrust Limits to Patent Settlements, 34 Rand J. Econ. 391 (2003)).
Carl Shapiro, Patent System Reform: Economic Analysis and Critique, 19 Berkeley Tech. L.J. 1017 (2004)
One may fairly ask whether market forces, in the form of patent
licensing agreements, can correct for poor patent quality, even if the underlying patent system is not well designed. Certainly, ceteris paribus, the "weaker" a patent, or the more likely that a patent will be found invalid if litigated, the lower
the royalties it can command in licensing negotiations. After all,
licensing takes place in the shadow of litigation, with licensing rates determined by the relative bargaining power of the patent holder and potential licensees/alleged infringers. Unfortunately, however, there is no reason to believe that the costs
of a flawed patent system are eliminated through the operation of
technology markets. To begin with, due to problems of free-riding and pass-through of royalties, we cannot count on litigation to invalidate patents that should never have been issued. n58
n58. See Joseph Farrell & Robert Merges, Incentives to Challenge and
Litigate Patents, 19 Berkeley Tech. L.J. 943 (2004); see also Mark Lemley & Carl Shapiro, Probabilistic Patents, J. of Econ. Perspectives (forthcoming 2005) (manuscript
at 25, on file with authors), available at
The paper by Shapiro also discussed the Quillen and Webster work on patent grant rate.
-->Turning to patent approval rates, or the percentage of patent
applications that ultimately result in patents, the FTC and NAS differ as to how to properly measure these rates. According to the FTC Report, one witness calculated the USPTO's grant rate at 98% in 2000, compared with 67% in Europe and 64% in Japan.
n46 However, a USPTO witness criticized these calculations and stated that, properly measured, the USPTO grant rate was about 75%. n47 The NAS cites a study by Quillen and Webster finding that, after certain corrections are made, "the USPTO eventually issued patents on between 85 percent and 97 percent of the applications filed between 1993 and 1998 - 20 to 30 percent higher than official
estimates, which have ranged between 60 percent and 70 percent for 20 years." n48 Recognizing that many factors influence patent approval rates, the NAS Study states, "The committee believes that high acceptance rates, especially if increasing over time relative to comparable rates in other industrialized countries would be reason to look more closely at examination quality." n49
[*1031] The NAS Sudy then observes that patent approval rates at the European Patent Office and the Japanese Patent Office declined more rapidly since 1998 than did the patent approval rates at the USPTO. n50<--
The content of footnote 48 is
-->n48. NAS Study, supra note 6, at 43 (citing C. Quillen and O.
Webster, Continuing Patent Applications and Performance of the U.S. Patent Office, 11 Fed. Cir. B.J. 1, 1-21 (2001), and R. Clarke, U.S. Continuity Law and its Impact on the Comparative Patenting Rates of the U.S., Japan, and the European Patent Offices, 8 J. Pat. & Trademark Off. Soc'y 335, 335-49 (2003)).<--
The Shapiro text is accurate as far as it goes; unfortunately, the cited NAS text is not entirely accurate, although analysis of more text than Shapiro presented would have attenuated the problem.
The NAS report in fact cites to both papers by Quillen et al., and in fact notes certain issues with the grant rate studies. [see below]
The paper by Lemley and Shapiro has some difficulties.
Page 7 of "Probabilistic Patents" mentions the work by Quillen and Webster: "As a result, the overwhelming majority of patent applications in the United States, perhaps 85%, ultimately result in an issued patent--far more than in Europe and Japan." citing Quillen et al. (2003) and NAS (2004).
Lemley and Shapiro cite the "right" Quillen et al. article here; they merely get the year wrong (they say 2003 even though the correct year is 2002).
The cite to the NAS report is tricky. The NAS report made no independent evaluation of patent grant rates. NAS merely referred to the work of Quillen et al. and to the work of Clarke. Both Lemley/Shapiro and NAS mistakenly reference the Quillen work as discussing "issued" patents; it doesn't. It follows the definition at the Trilateral Website. Some factors that enter into differences between "granted" patents and "issued" patents involve time delay AND withdrawal from issue [which I discuss in a recent paper].
**The NAS report itself
A Patent System for the 21st Century, Stephen A. Merrill, Richard C. Levin, and Mark B. Myers
Of the NAS report, page 53 has text that "the USPTO eventually issued patents on between 85 and 97 percent of applications filed between 1993 and 1998." The source for this would seem to be the first paper by Quillen and Webster, as noted in footnote 48 of the Shapiro paper. This first paper discusses a grant rate (not an issue rate) of 97%; the second paper discusses a grant rate (not an issue rate) of about 85%. This downward correction arose because of an erroneous assumption by QW in the first paper in obtaining the 97% number: that one could not have a patent on both a parent and on a continuing application. In reality, the first paper of Quillen et al. (2001) discusses grant rates of 97%, 87%, and 80%, although the 80% number arises from assumptions stated to be "contrary to fact." [see below]
The use by the NAS of the word "issued" is most unfortunate, because Quillen et al. made abundantly clear that their work pertained to "granted" as defined on the Trilateral Website. Curiously, the later work of Clarke was criticized as being related to "issued" patents (eg, Lemley and Moore, 2004). In fact, both Quillen et al. (2002, not 2001) and Clarke (2003) made certain corrections based on numbers derived from ISSUED patents.
The NAS report at page 54 has text "For this reason arriving at a consensus on a precise patent approval rate may be elusive."
The NAS report at page 54 also refers to a study by the Organisation for Economic Cooperation and Development (OECD, 2003) that showed that the grant rate for US priority applications with at least one subsequent EPO application was higher than the EPO grant rate for US priority applications. This is an interesting observation but one which does not prove that the USPTO is granting patents at a higher grant rate than the EPO for applications with the SAME claims. It is interesting that the NAS report relies on comparative data between regimes with different laws at the same time we have had a mini-debate between Justices Scalia and Breyer on the propriety of US courts citing to decisions of non-US courts.
***Text from Quillen et al. (2001) ["first paper"]
Calculations that do correct the PTO Grant Rates for continuing applications are in Table 6. n77 And, as can be seen from this table, the corrected Grant Rates are quite different from the uncorrected Grant Rates. For example, if it is assumed that all of the continuing applications represent a renewed attempt to patent the subject matter of their parent applications, then the number of net abandonments is the total number of abandonments less the total number of continuing applications filed. The overall Grant Rate for the 1993-1998 fiscal years on that assumption is 97%
If instead it is assumed that continuation and CIP applications, but not divisional applications, represent renewed efforts to patent the subject matter of their parent applications, the number of net abandonments is the total number of abandonments less the number of continuation and CIPs filed. The resulting overall Grant Rate for the 1993-1998 fiscal years is 87%. n79
If it is assumed, contrary to fact
, that only continuation applications represent a renewed effort to patent the subject matter of parent applications which are abandoned, then the number of net abandonments is the total number of abandonments less the number of continuation applications filed. The resulting overall Grant Rate for the 1993-1998 fiscal years in this case is 80%. n80
***Text from Quillen et al. (2002) ["second paper"]
Corrected Grant Rates (Appendix V) are calculated by adjusting the number of applications reported as abandoned by the number of refiled continuing applications so as to determine a net number of abandoned applications. Corrected Application Disposals are the sum of allowances and corrected applications abandoned (reported as Net UPR Applications Abandoned). n39 Net Disposals are the sum of Application Allowances and Net Abandonments, and the Grant Rate is the number of allowances divided by Net Disposals. The determination of the Grant Rate from Table 6A of the earlier paper is reproduced as Calculation 1 in Appendix V. The overall corrected Grant Rate for the six-year period was determined to be 97%. n40 Calculation 2 in Appendix V is a determination of the Corrected Grant Rate, adjusted to take into account the continuation applications on which a patent was granted on both the parent application and the continuation application (6.4% of the total number of continuing applications). Adjustment of the Grant Rate to take such continuation applications into account reduces the Grant Rate from 97% to 95%. Calculation 3 is the determination of the adjusted Grant Rate taking into account all continuing applications in which a patent was granted on both the parent application and the continuing application (31% of all continuing applications). The adjusted corrected Grant Rate on this assumption is 85%
, which is substantially above the Grant Rates reported for the EPO, JPO and USPTO on the Trilateral Website.