In September, the Sunwater Institute published a report analyzing patent quality in the United States. With significant implications for ongoing policy debates on Capitol Hill and beyond, it is important to assess whether the data presented is consistent with the conclusion and recommendations that the report advances—and whether the report draws conclusions worth paying attention to.
Understanding Type 1 and Type 2 Errors
In its assessment of patent quality, the Sunwater Institute makes a distinction between “Type 1” and “Type 2” errors at the USPTOUnited States Patent and Trademark Office. See also PTO.. Type 1 errors, analogous to false positives, are patents which are approved despite having invalid claims, while Type 2 errors are valid patents which are wrongly rejected or abandoned and are analogous to false negatives.
As the report authors acknowledge, Type 1 errors “impose welfare costs through higher prices and litigation costs associated with enforcing these (incorrectly granted) patents.” While not quantifiable, Type 2 errors may indirectly hamper innovation incentives and prevent inventors who are denied their rightful patent protections from receiving the benefit of the patent bargain.
Based on an analysis of about 20 million independent claims across 980,000 patent applications filed between 2011 and 2013, the report finds that the rate of Type 1 error is roughly 7%, while the rate of Type 2 error is roughly 18%. However, this Type 2 rate is quite misleading. The Sunwater Institute defines Type 2 error as valid patent claims which were both wrongly rejected or abandoned. A patent claimant at the USPTOUnited States Patent and Trademark Office. See also PTO. can abandon their claim for a whole host of reasons, including to pursue it further in a continuationA second application for the same invention claimed in a prior nonprovisional application and filed before the first application becomes abandoned or granted. Continuations are unique to the U.S. and are criticized for giving applicants endless opportunity to wear down patent examiners. See Lemley and Moore, Ending the Abuse of Continuation Practice. application.
In fact, one academic study analyzed the PTO’s data from 1983-2018 and concluded that at least from 2009-2018, more than twice as many abandoned applications were refiled as truly abandoned. In other words, of that 18% categorized as Type 2 errors, it is likely that at most 6% of the abandoned applications are true Type 2 errors.
Further, choosing to abandon a valid claim does not necessarily mean that the claim would have been rejected had the PTOPatent and Trademark Office, informally used interchangeably with USPTO. actually decided on it. During my time as a patent prosecutor, clients would occasionally—not frequently, but occasionally—instruct me to abandon patent applications. Sometimes I was even instructed to abandon applications which had received approval to be issued. This might be because of a desire not to pay the issue and maintenance feesFees for maintaining in force a patent based on an application filed on or after December 12, 1980. See MPEP 2500 for more. due to budgetary constraints, or it might have been because the applicant determined that the patented technology just wasn’t worth anything—but it wasn’t an error on the part of the USPTOUnited States Patent and Trademark Office. See also PTO., either.
The report’s omission of these essential attributes of abandonment suggest that, without additional clarity here, their comparison of Type 1 and Type 2 error rates is likely to be flawed.
Rejecting the False Binary
Just as importantly, the report wrongly frames the levels of Type 1 and Type 2 error as a “tradeoff.” The report assumes that “attempts to singularly reduce Type 1 error will likely increase Type 2 error [and] efforts to reduce Type 2 error may also result in higher Type 1 error.” So, it argues, “the policy community, and the U.S. Congress in particular, should reframe its debate about patent quality” to end its “myopic focus on Type 1 errors.”
But ensuring quality at all stages of the patent review process need not be an “either-or” situation. For one, the most immediate way to reduce both Type 1 and 2 errors is to give patent examiners at the USPTOUnited States Patent and Trademark Office. See also PTO. the tools and resources they need to make accurate validity determinations. Examiners only receive an average of 19 hours of time to examine an application. Even doubling that, a step well within the resources of the USPTOUnited States Patent and Trademark Office. See also PTO. and applicants, would significantly reduce both Type 1 and Type 2 errors. With thousands of increasingly complex and novel technology claims being filed each year, this is a step without any tradeoffs which advocates on all sides of patent policy can agree upon.
Indeed, no one advocates singularly reducing Type 1 error without paying any attention to valid claims which are wrongly denied. But perhaps the reason for the allegedly “myopic” focus on Type 1 error in the patent policy debate is that only Type 1 errors can be addressed once a patent has been issued. Unfortunately, groups like the Council for Innovation Promotion, the report’s co-sponsor, are currently advocating for policies that would limit the PTO’s ability to mitigate Type 1 error and that would have have no impact on Type 2 error.
The Risk of Weakening Quality Review Mechanisms
Ill-conceived congressional legislation and agency rules threaten quality review procedures post-issuance. Already, the controversial Fintiv rule has made it more difficult for challengers to use the inter partes review (IPRIntellectual Property Rights. Usually associated with formal legal rights such as patent, trademark, and copyright. Intellectual property is a broader term that encompasses knowledge, information, and data considered proprietary whether or not it is formally protected.) process to challenge low-quality patents by directing any patent challenge already involved in parallel litigation away from the PTABPatent Trial and Appeal Board. Reviews adverse decisions of examiners on written appeals of applicants and appeals of reexaminations, and conducts inter partes reviews and post-grant reviews. The Board also continues to decide patent interferences, as it was known as the Board of Patent Appeals and Interferences (BPAI) before the AIA.. To make matters worse, the PREVAIL Act, a bill which may see an attempt to push it through Congress in the lame duck session, would institute a number of policies to make it more difficult for members of the public to challenge invalid patents.
Policies like PREVAIL and Fintiv make Type 1 error harder to combat, while also having no effect on Type 2 error rates. But there’s no reason for policymakers to adopt this “either-or” mindset.
By investing both in both accuracy at the initial examination stage, and ensuring that interested parties can correct invalid patents that have been issued using post-issuance review processes like IPRIntellectual Property Rights. Usually associated with formal legal rights such as patent, trademark, and copyright. Intellectual property is a broader term that encompasses knowledge, information, and data considered proprietary whether or not it is formally protected., the USPTOUnited States Patent and Trademark Office. See also PTO. can reduce both types of error while also addressing Type 1 error in instances where it has been made. This “both-and” approach is the only way to truly protect innovators and ensure that the patent system is being used to defend, not harm, legitimate developments in technology and competition.
