4iP News

A comparative study of quality across the five largest patent offices estimates that only 2-6 per cent of granted patents have “dubious validity”

09 March 2017

As part of a series of articles and publications on the topic of patent quality, 4iP Council recently spoke to Professor Gaétan de Rassenfosse, Holder of the Chair of Innovation and IP Policy at the Ecole polytechnique fédérale de Lausanne (EPFL), Switzerland.

Professor Gaétan de Rassenfosse and his co-authors, Professors Adam Jaffe and Elizabeth Webster, recently published a comparative study of quality across the five largest patent offices entitled ‘Low-Quality Patents in the Eye of the Beholder: Evidence from Multiple Examiners’. The magnitude of the difference between the figures presented in their study and the figures obtained using patent litigation data bear important implications for discussions about patent quality.

Gaétan developed the following summary of the aforementioned work for 4iP Council and responded to our questions about its implications.

Research Summary

A comparative study of quality across the five largest patent offices

A comparative study of quality across the five largest patent offices
“only 2–6 percept of granted patents have dubious validity in the specific sense that they appear to be inconsistent with the country’s own standard for patent grant”

“the percentage of ‘weak’ patents is in single digits for all offices”

“the EPO and JPO are the most accurate offices”


Interview with Professor Gaétan de Rassenfosse

Holder of the Chair of Innovation and IP Policy at Ecole polytechnique fédérale de Lausanne.

You have made a comprehensive analysis of more than one million patent applications for about 400,000 inventions submitted in different patent offices worldwide (IP5 offices, which together attract 80% of worldwide patent activity) to analyse whether there is a problem of weak patents. What is the outcome of this study?

The most important contribution of this study is to provide a thought-provoking way of measuring patent quality. The vast majority of academic studies on the topic measure patent quality by looking at whether a patent survives litigation. By contrast, our model seeks to quantify quality by looking at the inconsistency of examination outcomes across offices. Our model allows us to estimate the overall ease of obtaining a patent in an office and to quantify the prevalence of what the economic literature names “weak patents”, that is, patents that should not have been granted according to an office’s own standard. Our central finding is that the overall prevalence of low-quality patents is likely to be smaller than what litigation studies predict.

Japan, as you pointed out in your article, has the highest inventive standard, China the lowest. Do you see a need for more coherence between patent office standards regarding the inventive step?

To be precise, our measure captures more than the de jure inventive step in an office. It captures systematic differences in offices’ propensity to grant applications (capturing de facto policies and practices). Having said that, I am not convinced that we need greater coherence across offices. Countries at different stages of technological and economic development have different needs with regards to the protection of IP. More generally, we still don't know what the optimal “height” of the inventive step should be (assuming we can measure and implement it)—it does not seem wise to have a worldwide standard if we can’t agree on what it should be. However, we certainly need greater coherence within offices. That is, it would be desirable to eliminate the randomness of the examination process. Our results show that the examination process at the SIPO and the USPTO is characterized by a greater level of apparent randomness than the examination process at the EPO and the JPO.

As you mention, the USTPO has initiated an on going online “patent quality chat” and has now an Office of Patent Quality Assurance (OPQA). Could you tell us more about this?

I have never taken part in the USPTO patent quality chats, so I cannot comment on the general tone of the discussion. These chats consist of monthly webinars designed to provide information on patent quality topics and gather input from the various stakeholders. Regarding the OPQA, its tasks are to conduct work product reviews, administer perception-based surveys, and provide support to the patent examining Corps on matters related to patent quality. While we can always discuss the relevance on some of the quality metrics used by OPQA, these initiatives at least show that the USPTO is taking the issue of patent quality seriously.

The EPO is also working hard in the last years to improve its examination quality. Are there any recommendations for the EPO or other Patent Offices around the globe on how to improve patent quality?

In the conclusion section we identify various detailed implications and I will therefore not repeat them here. However, one point that we did not mention is that patent offices, and the EPO in particular, should be more transparent and share more data with the scientific community. The EPO did an excellent job a few years ago in building the PATSTAT database, which has spurred a community of researchers and has considerably raised the level of discussions about the patent system. The USPTO has now clearly become the new “data champion” in researchers’ eyes thanks to high-level commitment on data sharing. I often hear from colleagues in the community that the EPO is a very shy and “last-century” organisation in terms of data sharing. I understand potential privacy issues related to data sharing, but I believe that the EPO needs to do more on this front if it wants to further raise the level of debate in Europe. The more information that we have on the prosecution process, the greater the number and quality of the analyses and subsequent policy recommendations.

You raise also the issue of courts making potentially wrong decisions. Taking into account that some appeal courts seem to sympathise with the approach of patent offices, would it be useful to intensify the communication between patent offices and courts of first instance?

Both examiners and judges can make mistakes. Information sharing between patent offices and courts of first instance is desirable in order to ensure that judges have the most accurate information possible. However, judges should take the information for what it is worth and maintain the independent nature of their decision. The more fundamental question that our study raises, however, is whether the courts systematically apply a stricter standard for validity than the patent office. This would raise fundamental questions about administrative law, since judges are not supposed to apply different standards.

What related research of yours might be of interest to readers of this interview?

I have worked quite a lot in the past on various questions related to the pricing of patents. More recently, I have studied the issue of congestion at patent offices by establishing a parallel between traffic congestion and the patent office backlog. In short, we use insights from the traffic congestion literature to propose ways to limit congestion at the patent office. I have also worked on the extent to which the patent system is used as a “technology protectionism” policy tool, with a specific focus on China. I will be happy to tell you more about this topic in a year’s time. All my research papers are available on my website and I am always open to discussing research findings and ideas.

The views expressed in this feature are those of the interviewee and may not reflect the views of 4iP Council or its members. The purpose of this feature area is to reflect thinking on the topic of intellectual property and enable open discussion.