Lemley on Fixing the U.S. Patent Office

What WON'T work:

Preventing Fee Diversion:  "[S]topping fee diversion is hardly a panacea. In the last several years, the UPTO has been fully funded – that is, Congress didn’t divert fees.  Nonetheless, the backlog grew.  The addition of 10-­20% of operating revenue wasn’t enough even to enable the PTO to hold steady."

Increasing PTO Fee-­Setting Authority:  "[I]t is likely not a good idea simply to spend more money to weed out bad patents.  Most of that money will be wasted on applications that are of no consequence to anyone.  And because of the structure of the examination system, it might not even succeed in weeding out bad patent applications.  Even if it did, however, the current fee structure makes patent quality self-limiting . . . the more bad patents [the USPTO] rejects, the fewer patents will pay maintenance fees, and the less money it will have to conduct a detailed examination."

Retaining Patent Examiners:  "There may well be benefits to reducing examiner attrition. But the evidence suggests that weeding out bad patents is not among them. Empirical research by Lemley and Sampat shows that the longer examiners spend at the PTO, the less searching they do, the less likely they are to issue initial rejections or demand claim amendments, and the more likely they are to ultimately grant a patent. (Lemley & Sampat, 2010). It is the most junior examiners who are most likely to reject applications. The reason is not precisely clear, but may have to do with increased workloads on senior examiners, or with acculturation into a corps whose ethos is to grant rather than deny patents."

Outsourcing Patent Searches:  "[R]ecent empirical evidence suggests that it might not work. [A recent study] found that [examiners] rely almost exclusively on art they find for themselves, not art submitted by applicants. And that doesn’t appear to reflect either applicants drafting around the art they found or the weakness of that art; U.S. examiners largely ignored even art that was submitted because it was found important by a foreign patent examiner during examination of a counterpart application. (Cotropia et al., 2010). If examiners are psychologically primed to rely principally on things they find for themselves, it won’t help to have others provide them with the best art. And it might even hurt, causing examiners not to focus on the best prior art."

What MIGHT Work:

"Second Pair of Eyes": 

Shortly after the Federal Circuit held business methods patentable in 1998, the PTO was inundated with business method patent applications. Most of those applications went to class 705. Indeed, by 2001 class 705 had the largest application volume. In response to this flood, the PTO initiated a specific “quality control” measure in this class in March 2000: the “second pair of eyes” review (SPER), under which applications are subjected to mandatory assessment by more than one examiner before being allowed. (Allison & Hunter, 2006). Requiring two examiners to agree seems to have had a dramatic effect: a 2009 study found that class 705 has the lowest grant rate among high volume classes. (Lemley & Sampat, 2008). One possible explanation for the low grant rate in this class is that the second pair of eyes is working, and that the grant rate reflects better rigor during examinations, rather than application volume. The fact that SPER leads to more rejections in Class 705 doesn’t mean it is an unalloyed success . . . Interestingly, the PTO recently shut down the SPER program in business methods. Too much success, it seems, carries its own risks.

Changing Examiner Incentives:

[PTO] human resource policies could be brought to bear, training examiners to search better, or giving them more time, or finding other ways to debias them. And it seems obvious – though likely politically infeasible – that the rules should not treat allowances differently than rejections. (Katznelson, 2010). These are good ideas, and they are worth exploring further. But implementation may be politically difficult. And some of the possible explanations point in different directions: should we give examiners more time to search, or less, for example?

Tiered Review (AKA "Gold Plated Patents"):

The Patent Office should focus its examination resources on important patents and pay little attention to the rest. But it is difficult for the government to know ahead of time which patents are likely to be important. There are two groups, however, that have better information about the likely technological and commercial value of inventions: patent applicants and competitors. To harness information in the hands of patent applicants, we could give applicants the option of earning a presumption of validity by paying for a thorough examination of their inventions. Put differently, applicants should be allowed to “gold plate” their patents by paying for the kind of searching review that would merit a presumption of validity. (Lemley et al., 2005) An applicant who chooses not to pay could still get a patent. That patent, however, would be subject to serious—maybe even de novo—review in the event of litigation. Most likely, applicants would pay for serious review with respect to their most important patents but conserve resources on their more speculative entries.6 That would allow the Patent Office to focus its resources, thus benefiting from the signal given by the applicant’s own self-interested choice.

Read/download: Mark A. Lemley, "Can the Patent Office Be Fixed?  (link)