Mark Lemley (Stanford) and Bhaven Sampat (Columbia) have released a draft copy of a recent study entitled "Is the Patent Office a Rubber Stamp?" The research was partly inspired by earlier studies (Quillen & Webster, 2002) that claimed that as many as 97% of all applications are eventually granted as patents. While these numbers have since been refuted, criticism of the PTO's perceived laxness during examination abounds, despite the (surprising?) fact that no one actually knows what percentage of patent applications actually issue as patents.
However, because of recent changes in the law regarding publication and PTO administrative procedure, Lemley and Sampat were able, for the first time, to track what happens to the vast majority of patent applications during prosecution. Significantly, they also had access for the first time to patent applications that are abandoned without a continuation.
And what did they find?
We find that the PTO rejects a surprisingly high percentage of patents. While more than two-thirds of all applications result in at least one patent, a significant number of applications are rejected and then finally abandoned by the applicant. We also find that the likelihood of obtaining a patent varies significantly by industry in surprising ways. For example, patents are much more likely to be granted in the chemical and pharmaceutical industries than in software and computer fields, despite the fact that most of the complaints about bad patents show up in the IT industries.Also,
[D]espite a variety of reforms that might be thought to reduce the use and abuse of continuation applications, we find a high use of continuation applications of various types. The extent and nature of the use casts significant light on the purposes of continuation applications, suggesting that different industries use continuations for different purposes. Even given the existence of a new and quicker procedure for continuing to fight with the examiner, many applicants persist in using the older continuation procedure in order to delay issuance of their patents or because they are seeking to construct a multi-patent fence.Currently, the PTO claims the allowance rate historically has been about 66%, which dropped to 54% recently. However, this number doesn't account for continuations. Under Lemley's and Sampat's analysis, parent and child applications were separately analyzed, and it was found that, when continuation applications are taken into account, the PTO has a grant "range" between 68.4% and 78.1%.
Not surprisingly, the first reaction a patent examiner has to an application is a non-final rejection - 86.5% of the PTO’s first office actions are non-final rejections. Only a tiny percentage (0.04%) began with final rejections, and 13.5% granted patents on the first office action without any argument or negotiation.
After a final rejection, patents are much more likely to be rejected if the applicant won’t amend (13.6% abandoned if the applicant amended versus 50.0% abandoned if they didn’t amend). Nevertheless, more than half of applications that received a final rejection have ultimately resulted in a patent, and another 20% are still pending.
When looking at specific industries, the paper notes that applications in the IT industries – computer hardware, software, communications, semiconductors, and electronics - make up more than 50% of all published applications. However, they have one of the lowest allowance rates in the PTO. Broken down by art unit, the numbers come out this way:
(ab = abandoned, all = allowed, p = pending)
• 1600 Biotech and Organic Chemistry:
31.5% (ab), 58.1% (all), 10.4% (p)
• 1700 Chemical and Materials Engineering:
24.2% (ab), 73.3% (all), 2.5% (p)
• 2100 Computer Architecture, Software:
27.2% (ab), 46.7% (all), 26.1% (p)
• 2600 Communications:
22.6% (ab), 59.6% (all), 17.7% (p)
• 2800 Semiconductors, Optical Systems:
16.8% 8(ab), 81.8% (all), 1.4% (p)
• 3600 Transportation, Construction:
33.3% (ab), 57% (all), 9.7% (p)
• 3700 Mechanical Engineering:
26.9% (ab), 71% (all) 2.1% (p)
Regarding software patents, about 13 percent of the applications in the study's sample were considered "software." The odds that a software patent is finally abandoned are virtually the same as those for a non-software patent – 23.7% of software patents, and 24.7% of non-software patents. But strikingly, grant rates in software are lower than overall grant rates: 51 percent versus 69 percent.
The paper noted this "surprising" finding:
While few should be surprised at the existence of industry-specific differences in patent grants, many people should be surprised – as indeed we are – at which industries have the highest grant rates. The received wisdom is that the problem of bad patents "rubber-stamped" by the PTO is largely confined to the IT industries. In fact, however, what we find is rather more complex. IT is hardly monolithic, and indeed includes the art units with both the highest and the lowest grant rates. The industry in which the grant rate is lowest is the computer industry, the very one in which many people would have said bad patents were most common. Similarly, the PTO rejects the highest share of applications in the field of business methods, from which a disproportionate number of complaints about the patent system arise.While the study is in draft form, you can bet that this will be an oft-cited paper in the ongoing patent/PTO reform debate.
To view the study, click here (link).