Friday, March 30, 2012

Professor Lemley - Fixing the Patent Office

I read Professor Lemley's paper Fixing the Patent Office and want to highlight several things tonight.

Professor Lemley states the PTO has issued many dubious software, electronic commerce, and Internet patents over the last two decades. He doesn't quantify this observation, but I think the patent community would mostly agree. Professor Lemley states a natural reaction is the PTO needs to award patents only when deserved. However, he pragmatically notes most bad patents merely protect technologies that fail in the marketplace, protect against non-existent competitors, are not infringed, licensed, or the subject of negotiation or assertion in court. In short, their issuance has no social cost.

As Professor Lemley states it, "The problem, then, is not that the Patent Office issues large number of bad patents. Rather, it is that the Patent Office issues a small but worrisome number of economically significant bad patents and those patents enjoy a strong, but undeserved, presumption of validity."

Having framed the problem this way, Professor Lemley states his solution: "The Patent Office should focus its examination resources on important patents and pay little attention to the rest."

Before turning to discuss that proposed solution, Professor Lemley asserts what probably won't fix the PTO:

Preventing fee diversion. For the last two decades Congress has diverted 10-20% of the PTO fees to general revenue. Yet in the last several years fee diversion did stop, while the backlog of pending applications grew. My comment is last year the backlog was slightly reduced.

Giving the PTO the authority to set own fees. The PTO may be in the best one to decide fees, but it is unclear which fees should be increased or decreased. For example, examining all applications more intensely is wasteful. Being stricter on allowances reduces the patent maintenance fees that make up for low filing fees. Raising filing fees raises costs on startups at the outset. Increasing a maintenance fee even by a factor of ten ($40,000) won't discourage patent trolls given the potential rewards of patent litigation.

Rejecting bad patents. The vast majority of bad patents are not litigated or licensed so rejecting them is wasteful. Applicants dissatisfied with a rejection can file requests for continued examination (note 53% of pending applications are RCEs) and/or continuations.

Retaining patent examiners. The PTO has a high rate of turn-over, and the median examiner has only been at the PTO for three years. The problem is junior examiners appear to be better at rejecting bad patents than senior examiners.

Outsourcing search. The thinking is examiners are not particularly good at finding relevant prior art so let's outsource this time-consuming task. However, research show examiners rely almost entirely on their own prior art search not the prior art submitted by others.

Next, Professor Lemley asserts what might work:

Second pair of eyes (SPER). Professor Lemley reminds after State Street the PTO initiated SPER to help deal with the flood of business method patents. He notes, however, SPER is asymmetric - two examiners must agree to allow a case while it takes only one examiner to reject a case. This led to 16% allowance rates in contrast to 72% on average. Professor Lemley finds it interesting that the PTO recently shut down SPER program on business methods applications. My comment is SPER over all technology areas would be very expensive, if not wasteful.

Changing examiner incentives. Research indicates much of the problem may be examiner's behavior due to incentives. This section seemed a bit muddled about how and what incentives should change. My experience over the past two decades is the quantity of cases processed sometimes appears to be more important to the PTO than the quality of examination. Whatever the change to incentives, you can count on the examiner's union to resist changes to how the examiners are evaluated and patent attorneys to resist any changes that make it more difficult to get patents.

Tiered review. Professor Lemley now presents his solution that the PTO should focus on important patents and pay little attention to the others, but it is difficult to know which patents are likely to be important. Professor Lemley concedes the government would find it difficult to know, but patent applicants and competitors have better information about the likely value of an invention. Thus, he proposes applicants "gold plate" their patents by paying for an intense search and examination process that would merit a strong presumption of validity. My comment here is (1) startup may be excluded from gold-plating; and (2) the tiered review appears to conflict with the Supreme Court's i4i v. Microsoft decision that the clear and convincing standard for invalidity applies to all US patents.

Oppositions and adversarial evaluations. I agree with Professor Lemley's expectation that the new post-grant and inter partes review will help fix the Patent Office's tendency to issue bad patents at times, but note this addresses bad patents rather than fixes the Patent Office, which to me is about the original examination. Also it will be interesting if the public input reduces the PTO fees for post-grant procedures.

Professor Lemley concludes that (1) reform proposals can never eliminate bad patents because examination is an imperfect process; (2) we will continue to rely on litigation as the primary means for weeding out bad patents; (3) courts should modulate the presumption of validity to match the facts; and (4) clear and convincing must be earned by patent owners.

Despite enjoying Professor Lemley's paper in the end I could only find agreement with Professor Lemley on conclusion (1). With regard to conclusion (2), if reexamination is not the primary means for weeding out bad patents I think it is becoming the best way to attack dubious patents. AIPLA economic survey data shows it's much less expensive than patent litigation even if you factor in the proposed increase in PTO fees. PTO statistics indicate it is effective in eliminating invalid patents as we saw in Oracle v. Google. Further, timely reexamination may result in stays of expensive litigation. With regard to (3) and (4), I don't see how "modulating" the standard of proving invalidity and concluding nothing was earned in the examination is not merely a litigator's delight plus runs counter to i4i v. Microsoft as mentioned earlier.

Copyright © 2012 Robert Moll. All rights reserved.