In 2010, the Manhattan Strategy Group completed a report for the patent office studying around 1,000 randomly selected patent examiners to see how well they were doing their jobs. The paper cited one junior patent examiner who claimed that “rather than doing what I feel is ultimately right, I’m essentially fighting for my life.”
Is that hyperbole, or reality? To find out, Cornell University's Michael Frakes and the University of Illinois' Melissa Wasserman looked at individual patent examiners who made decisions on 1.4 million patent applications over the course of a ten-year period, from 2002 to 2012. (It's worth noting a lot has changed at the patent office in those years; it's added more staff and reduced its backlog significantly.) Matched against the USPTO's official roster of patent examiners, which the researchers obtained through Freedom of Information Act requests, Frakes and Wasserman were able to track the patent examiners' careers as they progressed up the ladder at the patent office. Their theory? That the way the patent office promotes people may be partly responsible for a tendency to grant patents too easily.
"Our data finds that as examiners are given less time to review applications, the less prior art they cite, the less likely they are to make time-consuming prior art rejections, and the more likely they are to grant patents," the researchers report.
There's another negative side-effect of this problem, too, the researchers observe. Because patent applications are somewhat randomly assigned once they've arrived in the office of a particular discipline, you never really know if your application is being assessed with a fine-toothed comb or if it's cruising by. More importantly, you don't know whether your competitor is getting the same or different treatment.
Citing "prior art," or patents that have already been granted, is one major way that the patent office tells applicants that their invention is too obvious or not new. Patents get rejected for other, non-substantive reasons. But prior art rejections are one way to judge an application on its merits.
Patent examiners have an incentive to approve or reject a patent as quickly as they can. That's because every time they do, they get two "credits" that go toward a progression system that determines when they get promoted (and how much they make). One interesting feature of the progression system, though, is that as you rise in rank (from, say, a GS-7 to a GS-14) the amount of time you're expected to spend on each patent decreases.
"I've long advocated for the need for examiners to have more time to examine," said Todd Dickinson, former executive director of the American Intellectual Property Law Association. "The problem over the years, of course, is that maybe until last year there kept being a diversion of funds away from the office."
Again using FOIAed data, Frakes and Wasserman found that each government pay grade was associated with a 10- to 15-percent decrease in the time allotted to each patent examiner per patent. Although the average patent application gets 19 hours of attention, that figure drops to as little as 10 hours at higher pay grades and in less complex disciplines (inventions dealing with artificial intelligence, for example, are going to be more complicated than inventions dealing with compound tools).
As patent examiners increase in rank, they also tend to process more applications than their less experienced colleagues. This makes some intuitive sense; Frakes and Wasserman discovered that new patent examiners cleared just 8.8 percent of the patents studied while those who'd been patent examiners for 14 years or more accounted for 15.7 percent.
It could be that, as the system seems intended to work, patent examiners just get more experienced over time and are therefore able to process more patents, more quickly. Or, the researchers admit, it could mean that the examiners are getting more laid back about giving approvals over time.
There are also more sophisticated metrics the patent office uses internally that could offer more specific findings about patent examiner performance, said Dickinson.
"I'd feel a little more comfortable if they were using that kind of quality data as opposed to a big, bold number like obviousness rejections, which can have a lot more variables associated with it. But as I said, nothing beats giving examiners more time."